text
stringlengths 87
880k
| pmid
stringlengths 1
8
| accession_id
stringlengths 9
10
| license
stringclasses 2
values | last_updated
stringlengths 19
19
| retracted
stringclasses 2
values | citation
stringlengths 22
94
| decoded_as
stringclasses 2
values | journal
stringlengths 3
48
| year
int32 1.95k
2.02k
| doi
stringlengths 3
61
| oa_subset
stringclasses 1
value |
|---|---|---|---|---|---|---|---|---|---|---|---|
==== Front
BMC Evol BiolBMC Evolutionary Biology1471-2148BioMed Central London 1471-2148-5-431609822410.1186/1471-2148-5-43Research ArticleThe genomic environment around the Aromatase gene: evolutionary insights Castro L Filipe C [email protected] Miguel M [email protected] Maria A [email protected] CIIMAR – Centre of Marine and Environmental Research, Rua dos Bragas 289, 4050-123, Oporto, Portugal2 ICBAS – Institute of Biomedical Sciences Abel Salazar, Largo Professor Abel Salazar, 2, 4099-003, Oporto, Portugal2005 12 8 2005 5 43 43 8 3 2005 12 8 2005 Copyright © 2005 Castro et al; licensee BioMed Central Ltd.2005Castro et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
The cytochrome P450 aromatase (CYP19), catalyses the aromatisation of androgens to estrogens, a key mechanism in vertebrate reproductive physiology. A current evolutionary hypothesis suggests that CYP19 gene arose at the origin of vertebrates, given that it has not been found outside this clade. The human CYP19 gene is located in one of the proposed MHC-paralogon regions (HSA15q). At present it is unclear whether this genomic location is ancestral (which would suggest an invertebrate origin for CYP19) or derived (genomic location with no evolutionary meaning). The distinction between these possibilities should help to clarify the timing of the CYP19 emergence and which taxa should be investigated.
Results
Here we determine the "genomic environment" around CYP19 in three vertebrate species Homo sapiens, Tetraodon nigroviridis and Xenopus tropicalis. Paralogy studies and phylogenetic analysis of six gene families suggests that the CYP19 gene region was structured through "en bloc" genomic duplication (as part of the MHC-paralogon formation). Four gene families have specifically duplicated in the vertebrate lineage. Moreover, the mapping location of the different paralogues is consistent with a model of "en bloc" duplication. Furthermore, we also determine that this region has retained the same gene content since the divergence of Actinopterygii and Tetrapods. A single inversion in gene order has taken place, probably in the mammalian lineage. Finally, we describe the first invertebrate CYP19 sequence, from Branchiostoma floridae.
Conclusion
Contrary to previous suggestions, our data indicates an invertebrate origin for the aromatase gene, given the striking conservation pattern in both gene order and gene content, and the presence of aromatase in amphioxus. We propose that CYP19 duplicated in the vertebrate lineage to yield four paralogues, followed by the subsequent loss of all but one gene in vertebrate evolution. Finally, we suggest that agnathans and lophotrocozoan protostomes should be investigated for the presence of aromatase.
==== Body
Background
The cytochrome P450 aromatase (CYP19) is a member of a large superfamily of enzymes named cytochrome P450, which are involved in many physiological functions, such as steroid biosynthesis [1]. CYP19 is a steroidogenic enzyme which catalyses the aromatisation of androgens to estrogens. Thus, aromatase activity is essential for maintaining a physiological balance between androgens and estrogens, a critical aspect in the reproductive function of vertebrates; in humans, a P450 aromatase mutation leads to sterility [2]. For many other vertebrate groups, it as been demonstrated that CYP19 plays a key role in sex differentiation [3].
While several CYP450 genes are universally distributed, CYP19 is so far restricted to the vertebrate lineage. In mammals [4], birds [5], amphibians [6], reptiles [7] and cartilaginous fishes [8], a single gene has been isolated. In most Actinopterygii, however, two genes, Cyp19a and Cyp19b, encode two different transcripts expressed in the ovary and brain respectively [9]. Linkage data from zebrafish clearly suggests that these two genes are most likely the result of a genome duplication in the ray-finned bony fish lineage [9]. Despite intensive research, the ancestry of CYP19 genes is yet to be deciphered. No orthologue has been described from fully sequenced invertebrate genomes, like Drosophila melanogaster, Ciona intestinalis or Caenorhabditis elegans [10]. Thus, it has been suggested that the CYP19 gene arose at the origin of vertebrates [10,11]. Nevertheless, there is now strong evidence indicating that these model invertebrate species have experienced extensive gene loss [12]. Significantly, the estrogen receptor which was thought to have emerged in vertebrate ancestry, has now been documented in the lophotrocozoan protostome Aplysia californica [13].
Paralogy regions (or paralogons) consist of a series of linked genes (unrelated) on one chromosome, many of which have linked homologues (or paralogues) on at least another chromosome (typically four) [14,15]. Two main scenarios have been put forward to account for their presence: evolutionary remnants of chromosomal "en bloc" duplications or genome duplication, followed by gene loss and inversions [16,17]; or they reflect independent tandem duplications of each gene family followed by adaptive groupings of genes on different chromosomes [18]. The term "en bloc" duplication is used here in the context described by Abi-Rached et al. [17]. One of the best characterised examples of a paralogon includes the genes around the MHC complex on human chromosome 6, with homologues on chromosomes 1, 9, 5, 15 and 19 [19,20] (figure 1). Despite other views [18], the most recent findings indicate that an ancestral MHC-like region/chromosome duplicated "en bloc" twice in early vertebrate ancestry to yield a four-array paralogon (figure 1) [17,19-22]. Accordingly, vertebrate genes within these regions should have multiple copies (up to four) equally related to single invertebrate orthologues [23]. However, this is not the case for a substantial proportion of gene families. Gene loss, genomic rearrangements and insertions in both vertebrate and invertebrate genomes will obscure the correct evolutionary pattern of gene families within paralogons [23]. This is particularly evident for gene families which are vertebrate single copy. If a vertebrate one-member gene family maps to a paralogon and no orthologue is found in invertebrate model species what should we conclude regarding the ancestry of such a gene family? The human CYP19 gene follows this pattern. It maps to HSA15q, a region proposed to be part of the MHC-paralogon [20] (figure 1). This genomic location is highly suggestive for a invertebrate origin of CYP19. Nevertheless, since no paralogues are found in other chromosomal regions, it could well be the case of a genomic location with no evolutionary meaning.
Figure 1 The MHC-paralogon. Two sets of paralogy regions are intact (A and B) while the remaining 2 are broken (C and D) (adapted from [20]). The map location of the MHC, CYP19 and the surrounding genes (and paralogues) is shown. In parentheses the genomic distance in megabases to the p telomere.
Phylogenetics, paralogy and comparative genomics can be a particularly powerful tool to address issues of gene ancestry. Here, we analysed the evolutionary history of the genes in close physical proximity to the aromatase gene(s) in several vertebrate species (Homo sapiens, Tetraodon nigroviridis and Xenopus tropicalis). Through phylogenetic analysis we demonstrate that the CYP19 region was structured most likely by "en bloc" genomic duplication (as part of the MHC-paralogon). Most importantly, we also determine that this region has retained the same gene content and overall organisation (CYP19 included) without any gene insertion in the three lineages. A single inversion of gene order has occurred in the mammalian clade. Finally, we describe for the first time an invertebrate CYP19 partial sequence from Branchiostoma floridae. We propose that the aromatase gene family is much older than previously hinted.
Results and discussion
In this study, we sequentially addressed three questions. First, we determined the duplication pattern (pre or post vertebrate radiation) of the gene families in close proximity to the human CYP19. A further test analysed the ancestry of the human aromatase genomic location (ancestral versus derived). Finally, we investigated the presence of CYP19 in other invertebrate species (B. floridae), other than those previously explored.
Phylogeny and paralogy
The human CYP19 maps to one of the proposed MHC-paralogon regions (HSA15q) (figure 1) [20]. Despite this proposal, no phylogenetic analysis has been performed to confirm that gene families at HSA15q are part of the MHC-paralogon. This strategy aimed at defining the presence/absence of invertebrate orthologues and the duplication timings of the selected gene families (both these questions are key predictions of paralogy regions). Therefore, we undertook the task of analysing the "genomic environment" surrounding the CYP19 human gene at HSA15q within a DNA sequence of 1.0 Mb (figure 2). Besides the CYP19, eight other ORFs are annotated within this DNA sequence, corresponding to the following genes: AP4E1, FLJ41287, COL, DMXL2, SCGIII, MGC35274, TMOD2, and TMOD3 (figure 2). This DNA module is outflanked by members of the Tropomodulin gene family, TMOD2 and TMOD3, which have been proposed to support the vertebrate genome duplication hypothesis [20]. Other members of the TMOD gene family map to expected regions of MHC paralogy (figure 1). Furthermore, a single orthologue is found in invertebrate species (e.g. D. melanogaster – sanpodo).
Figure 2 Physical maps of the genomic environment around CYP19 in Homo sapiens, Tetraodon nigroviridis and Xenopus tropicalis.
We began by investigating the gene complement for each gene family in vertebrate and invertebrate species through BLAST search. Phylogenetic analysis was then performed when no previous study was available to determine duplication timings (if duplication had occurred).
AP4-E1
Adaptor protein complexes function as vesicle coat components in different membrane traffic pathways [24]. AP4E1 is a recently described subunit of a 4th complex [24]. Up to the present day the AP4E1 gene has been found solely in vertebrate genomes. Through BLAST search we have found the first invertebrate sequence in the C. intestinalis genome (scaffold 11). As shown in the phylogenetic tree, CiAP4E1 is basal to the vertebrate genes with 1000 of bootstrap support (figure 3A). No further homologues were uncovered in vertebrate genomes.
Figure 3 Neighbor-joining phylogenetic trees from alignment of the putative protein sequences of AP4E1 (A), SSC-2S (B), COL (C), DMXL (D), PPBP (E) and TMOD (F). Figures at nodes are scores from 1000 bootstrap resamplings of the data. Arrow denotes duplication timing. Ag – Anopheles gambiae; Ce – Caenorhabditis elegans; Dm – Drosophila melanogaster; Ci – Ciona intestinalis; Hs – Homo sapeins; Mm – Mus musculus; Gg – Gallus gallus; Xt – Xenopus tropicalis; Fr – Fugu rubripes; Tn – Tetraodon nigroviridis.
SSC-S2
The ORF identified in Ensembl as NP_997264 (FLJ41287 protein), presents significant sequence similarity to three other GenBank entries. One of those is a novel tumor necrosis factor-α inducible gene, SSC-S2 [25], which maps to HSA5. SSC-S2 contains a motif in the amino terminus that shows a significant similarity to death effector domain II of cell death regulatory protein, Fas-associated death domain-like interleukin-1β-converting enzyme-inhibitory protein (FLIP) [25]. Through phylogenetic analysis we showed these four sequences to be paralogues (figure 3B). The four genes were named as follows: SSC-S2A (HSA5), SSC-S2B (HSA1), SSC-S2C (HSA15) and SSC-S2D (HSA19). Invertebrate orthologues were found in Anopheles gambiae, D. melanogaster, C. elegans (not used in the phylogeny) and C. intestinalis (scaffold 92). The duplication events date to early vertebrate origin, as indicated by the branching pattern of the tree (figure 3B). The invertebrate sequences are basal to the vertebrate genes with a significant bootstrap support (figure 3B). No homologue of SSC-S2A is found in Actinopterygii, possibly due to gene loss. Additionally, the human genes are all located in regions of MHC paralogy – HSA1, HSA19, HSA5 and HSA15 (figure 1), as expected from two rounds of "en bloc" duplication in early vertebrate ancestry. However, the tree branching pattern is not of the (A,B)(C,D) type, but sequential which is not in agreement with the "en bloc" scenario.
COL
The Ensembl annotation identifies this gene as Collomin. This as been renamed to Colmedin (COL1) [26]. Colmedin is a phylogenetically conserved type II transmembrane protein with collagen repeats and a cysteine-rich olfactomedin domain, with members described in C. elegans (two genes), Drosophila and vertebrates [26]. No orthologue was detected in C. intestinalis. Colmedin has been found to be a single-copy gene in several vertebrate species. BLAST search to Danio (not shown), Fugu and Tetraodon genomes uncovered a new Colmedin gene, which we name COL1b (figure 3C). COL1b is a specific paralogue of Actinopterygii. The genomic location of this new gene is explained most likely by an extra genome duplication (see following section).
DMXL
RAB-3 is a 12 WD domain protein which binds both GDP/GTP exchange protein and GTPase-activating protein for Rab3 small G protein family [27]. These domains are found in a variety of proteins and are likely to be involved in protein-protein interactions [28]. It shows a domain structure similar to that of DMXL1 which has 10 WD domains, and has been renamed DMXL2 [27,29]. Our phylogenetic analysis confirms that both genes are paralogues, with invertebrate sequences basal to vertebrate DMXL1 and DMXL2 (figure 3D). Moreover, DMXL1 maps to an expected region of MHC-paralogon in HSA5 (figure 1). Thus, the duplication event which originated DMXL2 and DMXL1 resulted most likely from two rounds of "en bloc" duplications in early vertebrate ancestry.
SCGIII
Secretogranin III (SCGIII) is a member of the granin protein family, that is a component of intracellular dense core vesicles. Through BLAST we found this gene family to be restricted to vertebrates (not shown).
NM_699205-PPBP
The ORF identified in Ensembl as NM_699205 codes for the hypothetical protein MGC35274. Sequence features (Lysin domain) indicate that it might be involved in cell wall catabolism. The C. elegans orthologue has been named Predicted peptidoglycan-binding protein (PPBP). Thus, we named the human gene PPBP1. A second PPBP gene can be found in vertebrate genomes, which we designate PPBP2. The phylogenetic tree indicates that both ORFs are paralogues (figure 3E). The second PPBP gene is present in Fugu (Tetraodon also has a second PPBP gene but due to the partial sequence was kept out of the phylogeny), amphibians and mammals. The tree pattern indicates that a duplication of an ancestral PPBP gene occurred specifically in the vertebrate lineage. Moreover, the second gene maps to an expected region of MHC paralogy in the human genome – HSA1.
TMOD2 and TMOD3
Popovici et al. [20] proposed that the TMOD gene family duplicated in the vertebrate lineage. However, no phylogenetic analysis was performed to support this assumption. In the human genome 4 tropomodulin genes have been annotated: TMOD1 (Erythrocyte tropomodulin), TMOD4 (Skeletal muscle tropomodulin), TMOD2 (Neuronal tropomodulin) and TMOD3 (Ubiquitous tropomodulin). In invertebrates a single tropomodulin gene is observed. The phylogenetic analysis by Almenar-Queralt et al. [30] suggests that TMOD1, 2 and 4 duplicated in the vertebrate lineage. Nevertheless, the origin of TMOD3 is still unclear. The genomic location of both TMOD2 and 3 is highly suggestive for a tandem duplication. Our phylogenetic analysis, supports this scenario (TMOD2 and 3 are tandem duplicates from an ancestral TMOD2/3 gene). The duplication event post-dates the divergence of fish and amphibians, since a single TMOD2/3 is found in both Fugu and Tetraodon (figure 3F). On the contrary, Xenopus, chicken, mouse and human have two distinct genes (mapping side by side) (the Xenopus TMOD3 orthologue has not been used in the phylogeny). Thus, we propose that a single tropomodulin gene existed in vertebrate ancestry. It duplicated to yield three TMOD genes (1, 2/3 and 4) as a result of "en bloc" duplication (probably as part of genome duplications). Later, a tandem duplication in the ancestor of Xenopus, chicken and mammals, originated the TMOD2 and TMOD3 genes.
The phylogenetic analysis of the full set of gene families within the human aromatase DNA segment, reveals that four of those have specifically duplicated in the vertebrate lineage. Only the SSC-2S gene family shows four paralogues. The tree branching pattern is not of the (A,B)(C,D) type (expected under an "en bloc" duplication) but sequential (expected under the adaptive duplication scenario). This observation has been interpreted as evidence against an "en bloc" scenario [31]. Thus, the phylogeny (branching patterns) per se does not support the "en bloc" duplications. In this context, the suggested duplicated regions could have resulted from a complex duplication, loss and rearrangement pattern, and not from "en bloc" duplications [18]. However, Furlong and Holland [23], have recently disputed this assumption.
Our analysis confirms the previous suggestion by Popovici et al. [20] that genes within HSA15q are part of the MHC-paralogon. We find that the paralogues for each gene family map to expected regions of MHC paralogy (figure 1). That is the case of SSC-2S (4 genes), DMXL (2 genes), PPBP (2 genes) and TMOD (3 genes). The physical proximity between these genes is also observed in other regions of paralogy besides HSA15. For example, paralogues SSC-S2B, PPBP2 and TMOD4 map closely in chromosome 1 (200 kb), while DMXL1 and SSC-S2A are separated by just 300 kb in chromosome 5 (figure 1). Furthermore, of those genes which are found to be single copy in vertebrates, only for CYP19 and SCGIII we have not found invertebrate orthologues (either in C. intestinalis, C. elegans or D. melanogaster).
Comparative genomics
Our phylogenetic analysis and paralogy study strongly suggests that the DNA segment harbouring the human aromatase gene resulted from an ancestral "en bloc" duplication of the MHC-paralogon. However, the question remains regarding the ancestry of human CYP19 genomic location. It could well be case that the human CYP19 location is of no relevant evolutionary meaning. In order to explore these possibilities, we compared the gene content and organisation around CYP19 genes in three vertebrate species: H. sapiens, T. nigroviridis and X. tropicalis (figure 2). The comparison indicates a striking pattern of conservation in both gene content and gene order in the three species (figure 2). No gene insertion occurred since the divergence of these lineages. Two genomic events took place: genomic inversion and tandem gene duplication (TMOD2, TMOD3). The DNA module containing TMOD3, TMOD2, PPBP1 and SCGIII is differently located in both fish/amphibian and humans (boxes figure 2). By comparing the gene order in the three clades it is possible to infer the ancestral configuration. Given that both Tetraodon and Xenopus have an identical gene order, it is more parsimonious to conclude that the H. sapiens configuration is derived (figure 4). The data indicates that prior to the divergence of these three lineages the aromatase gene was already at this precise location (figure 4). In the case of Tetraodon, this region has duplicated further onto chromosomes 5 and 13, originating CYP19a and b, most likely as the result of a further genome duplication on the teleost lineage (figure 4). Later in evolution, a genomic inversion has taken place in the human lineage (possible mammalian) (figure 4).
Figure 4 Evolutionary model for the origin of the aromatase gene family. Symbols from the two conserved DNA blocks from figure 2. Dotted line boxes denote conservation of synteny but not gene order. Black bars denote "en bloc" (or genome) duplications. Star indicates tandem duplication of TMOD2/3. Horizontal curve arrow indicates gene inversion of the DNA block containing TMOD3, TMOD2, PPBP and SCGIII.
CYP19 in amphioxus
The phylogenetic analysis, paralogy study and the conservation of gene organisation around aromatase in three vertebrate species suggest that CYP19 (and the immediate outliers) was present in the invertebrate unduplicated MHC-paralogon prior to vertebrate radiation. However, they do not rule out the possibility that CYP19 emerged either just before or after the divergence of lamprey/hagfish and gnathostomes (being inserted into the MHC paralogy regions) (arrows-figure 4). The exact timing of the proposed two rounds of genome duplications which structure the vertebrate genome is still contentious. However, the consensus points to one duplication prior to the divergence of lamprey/hagfish and gnathostomes, and the second after the divergence of these lineages [32] (figure 4). Furthermore, in the genome sequence of the sea squirt despite the evidence of an MHC unduplicated paralogon [33], no orthologue of aromatase has been found [10]. On the contrary, amphioxus has proved a more favourable model to address these issues [23]. In order to investigate the presence of CYP19 in B. floridae, a BLAST search to the trace archives of the Whole Genome Sequence using the CYP19 sequence from the stingray (Dasyatis sabina) was performed. A single hit with a significant E-value was retrieved. This information was subsequently used to isolate a partial sequence (494 bp) from DNA extracted from a cDNA 5–24 h embryo library. In figure 5A, we show the DNA sequence and the predicted amino acid translation. When run on BLAST, it clearly emerges that it belongs to the CYP19 gene family. We designate this AmphiCYP19. Vertebrate aromatase genes display typical putative structural domains. These include the I-helix region, Ozol's peptide region, Aromatic region and Heme-binding region [34] (figure 5B). The alignment provided in figure 5B indicates the presence of similar motifs in the AmphiCYP19 sequence we now describe. The orthology of the retrieved sequence was determined through phylogeny. Overall, this analysis strongly indicates that AmphCYP19 is part of the aromatase evolutionary clade (bootstrap of 1000) (figure 5C). Our result suggests that that a single CYP19 gene was present at the base of the chordate lineage. We propose that it has been independently lost in the urochordate C. intestinalis. We argue that the ancestral invertebrate chordate CYP19 gene underwent two rounds of "en bloc" duplication like many other gene families in the MHC-paralogon to yield four paralogues. Three of these have been lost. However, we cannot rule out that no CYP19 duplications occurred in the vertebrate lineage, regardless of whether "en bloc" duplications occurred in other gene families. We favour the first scenario.
Figure 5 Nucleotide and predicted amino acid sequence of the AmphiCYP19 (partial sequence) (A); alignment of CYP19 sequences; dashes denote insertions; black line, I-helix region; red line, Ozol's peptide region; green line, Aromatic region and blue line, Heme binding region. Bf, B. floridae, Ds, D. sabina, Am, Alligator mississippiensis, Gg, G. Gallus, Hs, H. sapiens, Mm, M. musculus, Xt, X. tropicalis, Tn, T. nigroviridis, Dr, D. rerio, Fr, Fugu rubripes (B); Neighbor-joining phylogenetic tree from the alignment of the putative protein sequences of CYP19 genes, figures at nodes are scores from 1000 bootstrap resamplings of the data ; an insertion of the TnCYP19b predicted protein sequence was kept out of the alignment (C).
The present results imply a significant theoretical change regarding the ancestry of the CYP19 gene family. This investigation started with the observation that the human aromatase gene maps to the MHC-paralogon. Nevertheless, two opposite scenarios could be draw from the phylogenetic analysis, paralogy and comparative genomics. Either the CYP19 locus was present in the invertebrate chordate unduplicated MHC-paralogon, and the presence of a single paralogue resulted from gene loss; or CYP19 originated early on in vertebrate evolution in its present position in the MHC-paralogon (figure 4). We went on to test these hypotheses. Our results can be summarised as follows: (1) vertebrate CYP19 containing regions are indeed part of the MHC-paralogon as demonstrated by the phylogenetic analysis of the gene families in close proximity; (2) comparative genomics of the aromatase region between fish, amphibians and humans shows a striking pattern of conservation without any gene insertion; (3) following the previous analysis, we found that CYP19 is not restricted to the vertebrate clade, given the description of AmphiCYP19.
Our model determine the loss of three aromatase paralogues upon duplication of the ancestral MHC-paralogon. For the vast majority of paralogy regions it is difficult to precisely determine the amount of gene loss (due to the absence of large sets mapping data from crucial organisms). In the case of the MHC-paralogon an estimate can be calculated, given the previous work of several authors [17,21,35]. The sequencing and mapping data of the MHC anchor genes in amphioxus, shows a significant proportion of gene families which are single-copy in both lineages (seven out of eighteen, excluding the anchor genes and those of unknown orthology; e.g. frequenin-like) [17,35]. Thus, the return to a single copy status following the "en bloc" duplication was not a rare event upon the duplication of the MHC-paralogon. At the moment we do not known whether AmphiCYP19 maps along with the MHC anchor genes in a single chromosome, but this hypothesis can be tested in the future [36].
Finally, we speculate that the ancestry of CYP19 genes could be more ancient than invertebrate chordate origin. Two reasons support this scenario. First, the sex steroid receptors (estrogens and androgens/progesterone/corticoids) are older than previously proposed [13]. The estrogen receptor found in Aplysia indicates that the duplication event from a sex steroid precursor receptor pre-dates the divergence of protostomes and deuterostomes [13]. Also, the phylogenetic analysis and paralogy studies of androgen, progesterone and corticoid receptors suggests that a single receptor was present in the ancestral Urbilateria [37]. Thus, the receptor gene kit for sex steroid hormones was already present in the primitive Bilateria (albeit not necessarily with a similar function). The second reason comes from Lophotrocozoan molluscs. These organisms respond to steroid hormones (e.g. estradiol) during their reproductive cycle [38]. Furthermore, biochemical analysis in mollusc tissue extracts reveals the presence of an aromatase-like activity [39,40]. In light of these findings and observations, we argue that the presence of CYP19 should be investigated in lophotrocozoan protostomes (e.g. molluscs).
Conclusion
We present here a detailed study of the genomic region containing the aromatase gene in three vertebrate lineages. The gene families found in close proximity to CYP19 show a clear pattern of vertebrate specific duplication, as expected from a paralogon. A key prediction from paralogy regions is their unduplicated presence in pre-vertebrate genomes. Significantly, we have also found that the genomic organisation of the human CYP19 genomic region mimics that of Tetraodon and Xenopus. Overall our analysis suggested the existence of aromatase in invertebrates. In agreement with this hypothesis we have found a CYP19 orthologue in the invertebrate chordate amphioxus. Contrary to previous suggestions, our data implies that CYP19 was present in the primitive chordate (and probably even earlier).
Methods
Phylogenetics and paralogy studies
The gene content around the CYP19 human gene in chromosome 15 comprises eight open reading frames (ORFs) within a DNA sequence of 1.0 Mb (figure 2). These are annotated as follows: AP4-E1 [GenBank: NP_031373], FLJ41287 (SSC-2SC) [GenBank: NP_997264], Collomin [GenBank: NP_861454], DMXL2 [GenBank: NP_056078], SCGIII [GenBank: NP_037375], MGC35274 [GenBank: NP_699205], TMOD2 [GenBank: NP_055363], and TMOD3 [GenBank: NP_055362] (figure 2). In order to find invertebrate orthologues and vertebrate paralogues, protein sequence from each gene was extracted and used for BLAST search (TBLASTN) against GenBank and Ensembl. Accession numbers for each gene are given in table 1.
Table 1 Accession numbers for the genes used in the phylogenetic analysis.
Gene family Gene Accession number Species
AP4E1 NP_031373 H. sapiens
ENSMUSP00000002063 Mus musculus
ENSGALP00000007610 Gallus gallus
SINFRUP00000132389 Fugu rubripes
GSTENG00025592001 T. nigroviridis
ENSXETP00000044864 X. tropicalis
ci0100133154 (scaffold11) C. intestinalis
SSC-S2 ENSANGP00000018260 A. gambiae
CG4091 AAF47048 D. melanogaster
ci0100140958 (scaffold 92) C. intestinalis
A NP_055165 H. sapiens
ENSMUSP00000034810 M. musculus
ENSGALP00000007631 G. gallus
ENSXETP00000044865 X. tropicalis
B NP_078851 H. sapiens
NP_081482 M. musculus
ENSXETP00000017843 X. tropicalis
SINFRUP00000149825 F. rubripes
GSTENT00011674001 T. nigroviridis
C NP_997264 H. sapiens
ENSMUSP00000034810 M. musculus
ENSGALP00000007631 G. gallus
ENSXETP00000044865 X. tropicalis
SINFRUT00000165316 annotated as pseudogene F. rubripes
GSTENT00018057001 T. nigroviridis
D NP_689575 H. sapiens
ENSMUSP00000076961 M. musculus
ENSGALP00000006772 G. gallus
SINFRUP00000156999 F. rubripes
GSTENT00028868001 T. nigroviridis
Colmedin CG6867 NP_573262 D. melanogaster
COF-2 AY494975 C. elegans
unc-122 AY494976 C. elegans
COL1 NP_861454 H. sapiens
NP_796324 M. musculus
ENSGALP00000021668 G. gallus
ENSXETP00000044876 X. tropicalis
COL1a SINFRUP00000165325 F. rubripes
GSTENG00018059001 T. nigroviridis
COL1b SINFRUP00000132381 F. rubripes
GSTENT00025590001 T. nigroviridis
DMXL XP_314464.1 A. gambiae
DmX NP_572302 D. melanogaster
CAB01916 C. elegans
ci0100136505 (scaffold149) C. intestinalis
DMXL2 ENSGALP00000007662 G. gallus
NP_056078.1 H. sapiens
XP_358382.2 M. musculus
GSTENG00018060001 T. nigroviridis
SINFRUP00000166673 F. rubripes
ENSXETP00000044880 X. tropicalis
DMXL1 ENSGALP00000003481 G. gallus
CAA06718 H. sapiens
ENSMUSP00000045559 M. musculus
GSTENG00030946001 T. nigroviridis
SINFRUP00000162769 F. rubripes
ENSXETP00000038136 X. tropicalis
PPBP PPBP1 NP_699205 H. sapiens
GSTENG00018055001 T. nigroviridis
SINFRUP00000165314 F. rubripes
ENSGALP00000007561 G. gallus
NP_081585 M. musculus
PPBP2 CAI16380 H. sapiens
ENSMUSP00000067811 M. musculus
ENSXETP00000017826 X. tropicalis
SINFRUP00000149824 F. rubripes
ENSXETP00000044860 X. tropicalis
ci0100137161 (scaffold34) C. intestinalis
XP_321699 A. gambiae
NP_650352 D. melanogaster
TMOD TMOD AAL13319 C. elegans
TMOD (spdo) AAC04506 D. melanogaster
TMOD ci0100140287 (scaffold487) C. intestinalis
TMOD1 NP_003266 H. sapiens
NP_068683 M. musculus
NP_990105 G. gallus
ENSXETP00000022442 X. tropicalis
GSTENG00033528001 T. nigroviridis
SINFRUP00000157299 F. rubripes
TMOD2/3 GSTENG00018054001 T. nigroviridis
SINFRUP00000165312 F. rubripes
TMOD2 NP_055363 H. sapiens
Q9JKK7 M. musculus
ENSGALP00000007536 G. gallus
ENSXETP00000044853 X. tropicalis
TMOD3 NP_055362 H. sapiens
ENSMUSP00000072087 M. musculus
ENSGALP00000007492 G. gallus
ENSXETP00000044845 X. tropicalis
TMOD4 NP_037485 H. sapiens
NP_057921 M. musculus
NP_990105 G. gallus
SINFRUP00000149822 F. rubripes
GSTENG00011677001 T. nigroviridis
ENSXETP00000044853 X. tropicalis
CYP19 NP_112503 H. sapiens
B. floridae
AAF04617 D. sabina
AAK31803 Alligator mississippiensis
NP_001001761 G. gallus
P28649 M. musculus
ENSXETP00000044866 X. tropicalis
Cyp19a GSTENP00025591001 T. nigroviridis
Cyp19b GSTENP00018058001 T. nigroviridis
Cyp19a NP_571229 D. rerio
Cyp19b NP_571717 D. rerio
Cyp19a SINFRUP00000132385 F. rubripes
Cyp19b SINFRUP00000165318 F. rubripes
Others AP3D AAC14585 D. melanogaster
AP3D1 NP_003929 H. sapiens
OLFM2 ENSP00000264833 H. sapiens
CiOLF ci0100131069(scaffold4) C. intestinalis
Cyp4d2 Z23005 D. melanogaster
CYP4 ENSMUSP00000003574 M. musculus
CYP4 ci0100146084 (scaffold15) C. intestinalis
Putative protein sequences for each gene family were aligned using the CLUSTAL X program (version 1.8). The produced alignments were further edited by eye to maximise the homologous regions (conserved domains) (given upon request). The phylogenetic reconstruction was based on conserved domains. If no domains were identified, reconstruction's were performed using the full-length alignment (without taking into account gaps or ambiguous sites). The phylogenetic trees were constructed using neighbor-joining from the CLUSTAL X program, on an amino acid distance matrix calculated with the Dayoff PAM option. Confidence on each node was assessed by 1000 bootstrap replicates. Trees were visualised with the Treeview program (version 1.6.6).
Mapping data was retrieved from H. sapiens [41], T. nigroviridis [42], X. tropicalis [43] and C. intestinalis [44].
Polymerase chain reaction (PCR)
A BLAST search to the B. floridae trace archives of the Whole Genome Sequence using the DNA sequence from the stingray (D. sabina) CYP19 was done. Clone AFSA830540 presented a significant E-value (data not shown). Clone walk 5' and 3' allowed the determination of further regions of CYP19 homology. To obtain a CYP19 sequence fragment an hemi-nested PCR approach was followed using DNA purified from an 5–24 h embryo cDNA library (J. Langeland, Kalamazoo, USA). Three oligonucleotides (2 forward and 1 reverse) were designed in conserved regions using the available genomic sequence: CYPF1 5' CTGGCTAACATCCGGGACAT 3'; CYPF2 5' CAGTGCGTGACAGAAATGCT 3'; and CYPR1 5' GACGGGCTCAGTTGGTACAT 3'. PCR was carried out in 50 μl reaction mixture consisting of 10 mM Tris-HCl, pH 8.0, 1.5 mMMgCl2, KCl 50 mM, TritonX 0.1%, 10 μM of each primer, 2 mM each of dATP, dCTP, dGTP, and dTTP, 1U DNA polymerase (Appligene Oncor). The first round of PCR (oligonucleotides CYPF1 and CYPR1) had the following profile: initial cycle of denaturation, 94°C 2 min, and forty amplification cycles with denaturation at 94°C for 45 s, annealing at 55°C for 30 s, and extension at 72°C for 1 min. An hemi-nested PCR was carried out afterwards using the first PCR product as a sample. A similar PCR profile was used with the exception of the extension time – 45 s (oligonucleotides CYPF2 and CYPR1). The PCR product was separated through 2% agarose gel and purified by using the QIAquick Gel Extraction kit (QIAGEN, Germany). The product was directly sequenced in both strands using the PCR oligonucleotides by STAB VIDA (Portugal). The sequence was deposited in Genbank DQ085624.
Authors' contributions
LFCC performed all sequence and phylogenetic analysis, comparative genomics, laboratory experiments and drafted the manuscript, MMS participated in phylogenetic analysis, design and co-ordination of the study, and MARH participated in the co-ordination of the study.
Acknowledgements
This study was supported by the project PDCTM/MAR/15284/99 from the Fundação para a Ciência e a Tecnologia, Portugal. LFCC is funded by the Fundação para a Ciência e a Tecnologia, Portugal (BPD/19608/2004). We acknowledge Daniel Rokhsar and the Department of Energy Joint Genome Institute for the unpublished shotgun data of the amphioxus genome project, and the Holland lab, University of Oxford, UK for the amphioxus cDNA. We acknowledge also three anonymous referees for their suggestions and comments.
==== Refs
Hall PF Cytochromes P-450 and the regulation of steroid synthesis Steroids 1986 48 131 96 3328326 10.1016/0039-128X(86)90002-4
Morishima A Grumbach MM Simpson ER Fisher C Qin K Aromatase deficiency in male and female siblings caused by a novel mutation and the physiological role of estrogens J Clin Endocrinol Metab 1995 80 3689 98 8530621 10.1210/jc.80.12.3689
Carreau S Bourguiba S Lambard S Galeraud-Denis I Genissel C Levallet J Reproductive system: aromatase and estrogens Mol Cell Endocrinol 2002 193 137 43 12161013 10.1016/S0303-7207(02)00107-7
Simpson ER Michael MD Agarwal VR Hinshelwood MM Bullun SE Zhao Y Expression of the CYP19 (aromatase) gene: an unusual case of alternative promoter usage FASEB J 1997 11 29 36 9034163
McPhaul MJ Noble JF Simpson ER Mendelson CR Wilson JD The expression of a functional cDNA encoding the chicken cytochrome P-450arom (aromatase) that catalyzes the formation of estrogen from androgen J Biol Chem 1988 263 16358 63 3182796
Miyashita K Shimizu N Osanai S Miyata S Sequence analysis and expression of the P450 aromatase and estrogen receptor genes in the Xenopus ovary J Steroid Biochem Mol Biol 2000 75 101 7 11226826 10.1016/S0960-0760(00)00164-3
Gabriel WN Blumberg B Sutton S Place AR Lance VA Alligator aromatase cDNA sequence and its expression in embryos at male and female incubation temperatures J Exp Zoo 2001 290 439 48 10.1002/jez.1087
Ijiri S Berard C Trant JM Characterization of gonadal and extra-gonadal forms of the cDNA encoding the Atlantic stingray (Dasyatis sabina) cytochrome P450 aromatase (CYP19) Mol Cell Endocrinol 2000 164 169 81 11026568 10.1016/S0303-7207(00)00228-8
Chiang EF-L Yan Y-L Guiguen Y Postlehwait J Chung B-C Two Cyp19 (P450 Aromatase) genes on duplicated zebrafish chromosomes are expressed in ovary and brain Mol Biol Evol 2001 18 542 50 11264405
Baker ME Co-evolution of steroidogenic and steroid-inactivating enzymes and adrenal and sex steroid receptors Mol Cell Endocrinol 2004 215 55 62 15026175 10.1016/j.mce.2003.11.007
Nelson DR Metazoan cytochrome P450 evolution Comp Biochem Physio C Pharmacol Toxicol Endocrinol 1998 121 15 22 10.1016/S0742-8413(98)10027-0
Kortschak RD Samuel G Saint R Miller DJ EST analysis of the cnidarian Acropora millepora reveals extensive gene loss and rapid sequence divergence in the model invertebrates Curr Biol 2003 13 2190 5 14680636 10.1016/j.cub.2003.11.030
Thornton JW Need E Crews D Resurrecting the ancestral steroid receptor: ancient origin of estrogen signalling Science 2003 301 1714 17 14500980 10.1126/science.1086185
Coulier F Popovici C Villet R Birnbaum D MetaHox gene clusters J Exp Zool 2000 288 345 51 11144283 10.1002/1097-010X(20001215)288:4<345::AID-JEZ7>3.0.CO;2-Y
Popovici C Leveugle M Birnbaum D Coulier F Homeobox gene clusters and the human paralogy map FEBS Lett 2001 491 237 42 11240134 10.1016/S0014-5793(01)02187-1
Lundin LG Evolution of the vertebrate genome as reflected in paralogous chromosomal regions in man and the house mouse Genomics 1993 16 1 19 8486346 10.1006/geno.1993.1133
Abi-Rached L Gilles A Shiina T Pontarotti P Inoko H Evidence of en bloc duplication in vertebrate genomes Nat Genet 2002 31 100 5 11967531 10.1038/ng855
Hughes AL Phylogenetic tests of the hypothesis of block duplication of homologous genes on human chromosomes 6, 9, and 1 Mol Biol Evol 1998 15 854 70 9656486
Kasahara M The chromosomal duplication model of the major of the major histocompatibility complex Immunol Rev 1999 167 17 32 10319248
Popovici C Leveugle M Birnbaum D Coulier F Coparalogy: physical and functional clusterings in the human genome Biochem Biophys Res Commun 2001 288 362 70 11606051 10.1006/bbrc.2001.5794
Castro LFC Furlong R Holland PWH An antecedent of the MHC-linked genomic region in amphioxus Immunogenetics 2004 55 782 4 14749904 10.1007/s00251-004-0642-9
Danchin EG Pontarotti P Towards the reconstruction of the bilaterian ancestral pre-MHC region. Trends Genet 2004 20 587 91 15522451
Furlong R Holland PW Were vertebrates octoploid? Phil Trans R Soc Lond B Biol Sci 2002 357 531 44 12028790 10.1098/rstb.2001.1035
Hirst J Bright NA Rous B Robinson MS Characterization of a fourth adaptor-related protein complex Mol Biol Cell 1999 10 2787 802 10436028
Kumar D Whiteside TL Kasid U Identification of a novel tumor necrosis factor-α-inducible gene, SCC-S2, containing the consensus sequence of a death effector domain of fas-associated death domain-like interleukin-1β-converting enzyme-inhibitory protein J Biol Chem 2000 275 2973 8 10644768 10.1074/jbc.275.4.2973
Loria PM Hodgkin J Hobert O A conserved postsynaptic transmembrane protein affecting neuromuscular signaling in Caenorhabditis elegans The Journal of Neuroscience 2004 24 2191 201 14999070 10.1523/JNEUROSCI.5462-03.2004
Nagano F Kawabe H Nakanishi H Shinohara M Deguchi-Tawarada M Takeuchi M Sasaki T Takai Y Rabconnectin-3, a novel protein that binds both GDP/GTP exchange protein and GTPase-activating protein for Rab3 small G protein family J Bio Chem 2002 277 9629 32 11809763 10.1074/jbc.C100730200
Neer EJ Schmidt CJ Nambudripad R Smith TF The ancient regulatory-protein family of WD-repeat proteins Nature 1994 371 297 300 8090199 10.1038/371297a0
Kraemer C Enklaar T Zabel B Schmidt ER Mapping and structure of DMXL1, a human homologue of DmX gene from Drosophila melanogaster coding for a WD repeat protein Genomics 2000 64 97 101 10708522 10.1006/geno.1999.6050
Almenar-Queralt A Lee A Conley CA Ribas de Pouplana L Fowler VM Identification of a novel tropomodulin isoform, skeletal tropomodulin, that caps actin filament pointed ends in fast skeletal muscle J Biol Chem 1999 274 28466 75 10497209 10.1074/jbc.274.40.28466
Hughes AL Phylogenies of developmentally important proteins do not support the hypothesis of two rounds of genome duplication early in vertebrate history J Mol Evol 1999 48 565 76 10198122
Escriva H Manzon L Youson J Laudet V Analysis of lamprey and hagfish genes reveals a complex history of gene duplications during early vertebrate evolution Mol Biol Evol 2002 19 1440 50 12200472
Azumi K De Santis R De Tomaso A Rigoutsos I Yoshizaki F Pinto MR Marino R Shida K Ikeda M Ikeda M Arai M Inoue Y Shimizu T Satoh N Rokhsar DS Du Pasquier L Kasahara M Satake M Nonaka M Genomic analysis of immunity in a Urochordate and the emergence of the vertebrate immune system: "waiting for Godot" Immunogenetics 2003 55 570 81 14530883 10.1007/s00251-003-0606-5
Graham-Lorence S Amarneh B White RE Peterson JA Simpson ER A three-dimensional model of aromatase cytochrome P450 Protein Sci 1995 4 1065 80 7549871
Vienne A Shiina T Abi-Rached L Danchin E Vitiello V Cartault F Inoko H Pontarotti P Evolution of the proto-MHC ancestral region: more evidence for the plesiomorphic organisation of human chromosome 9q34 region Immunogenetics 2003 55 429 36 14530884 10.1007/s00251-003-0601-x
Castro LFC Holland PWH Fluorescent in situ hybridisation to amphioxus chromosomes Zoolog Sci 2002 19 1349 53 12520094 10.2108/zsj.19.1349
Bertrand S Brunet FG Escriva H Parmentier G Laudet V Robinson-Rechavi M Evolutionary genomics of nuclear receptors: from twenty-five ancestral genes to derived endocrine systems Mol Biol Evol 2004 21 1923 37 15229292 10.1093/molbev/msh200
Di Cosmo A Di Cristo C Paolucci M Sex steroid hormone fluctuations and morphological changes of the reproductive system of the female of Octopus vulgaris throughout the annual cycle J Exp Zool 2001 289 33 47 11169491 10.1002/1097-010X(20010101/31)289:1<33::AID-JEZ4>3.0.CO;2-A
Le Curieux-Belfond O Moslemi S Mathieu M Seralini GE Androgen metabolism in oyster Crassostrea gigas: evidence for 17beta-HSD activities and characterization of an aromatase-like activity inhibited by pharmacological compounds and a marine pollutant J Steroid Biochem Mol Biol 2001 78 359 66 11717006 10.1016/S0960-0760(01)00109-1
Santos MM ten Hallers-Tjabbes CC Vieira N Boon JP Porte C Cytochrome P450 differences in normal and imposex-affected female whelk Buccinum undatum from the open North Sea Mar Environ Res 2002 54 661 5 12408632 10.1016/S0141-1136(02)00150-2
Homo sapiens Genome
Tetraodon Genome Browser
Xenopus tropicalis Genome
Ciona intestinalis Genome
|
16098224
|
PMC1215479
|
CC BY
|
2021-01-04 16:29:16
|
no
|
BMC Evol Biol. 2005 Aug 12; 5:43
|
utf-8
|
BMC Evol Biol
| 2,005 |
10.1186/1471-2148-5-43
|
oa_comm
|
==== Front
BMC Evol BiolBMC Evolutionary Biology1471-2148BioMed Central London 1471-2148-5-461611149510.1186/1471-2148-5-46Research ArticleRibosomal intergenic spacer (IGS) length variation across the Drosophilinae (Diptera: Drosophilidae) Mateos Mariana [email protected] Therese A [email protected] Center for Insect Science and Department of Ecology and Evolutionary Biology, University of Arizona, BioSciences West 310, Tucson, AZ 85721, USA2005 19 8 2005 5 46 46 22 4 2005 19 8 2005 Copyright © 2005 Mateos and Markow; licensee BioMed Central Ltd.2005Mateos and Markow; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
The intergenic spacer of the ribosomal genes in eukaryotes (IGS) contains duplications of the core transcription promoter. The number of these duplicated promoters, as measured by the IGS length, appears to be correlated with growth rate and development time in several distantly related taxa. In the present study, we examined IGS length variation across a number of species of Drosophila to determine the amount of variation in this trait across different evolutionary time scales. Furthermore, we compared the usefulness of two methods commonly used to determine IGS length: Southern Blot Hybridization (SB) and Polymerase Chain Reaction (PCR).
Results
Our results show broad variation in IGS length across the genus Drosophila, but closely related species had similar IGS lengths. Our results also suggest that PCR tends to underestimate the true IGS size when the size is greater than 5 kb, and that this degree of underestimation is greater as the IGS size increases.
Conclusion
Broad variation in IGS length occurs across large evolutionary divergences in the subfamily Drosophilinae. Although average IGS length has been shown to evolve rapidly under artificial selection, closely related taxa generally have similar average IGS lengths. Our comparison of methods suggests that without previous knowledge of the DNA sequence of the IGS and flanking regions, both methods be used to accurately measure IGS length.
==== Body
Background
Due to the importance of ribosomes in protein synthesis, cellular growth, and organismal development, ribosomal genes are highly transcribed; with ribosomal RNA accounting for nearly half of all cellular transcription and 80% of the RNA content of growing cells [reviewed by [1,2]]. To achieve these high levels of ribosome production, eukaryotes have multiple copies of ribosomal (r)DNA, arranged in tandem in the Nucleolus Organizer Regions (NORs) of one or more chromosomes. In addition, eukaryotes sustain high levels of transcription per rDNA copy [1].
The structure of the ribosomal intergenic spacer (IGS; Figure 1) appears to be important for achieving these high transcription levels. IGS varies in length from about 2 kb in yeast to 21 kb in mammals, and is also highly variable among and even within individuals of the same species [reviewed by [1]]. These length polymorphisms are mostly due to variation in the numbers of different internal subrepeats present in the IGS (Figure 1). In eukaryotes, some of these repetitive regions contain duplications of the core promoter [reviewed by [1]]. These promoter duplications have been shown to enhance rDNA transcription. For example, in Drosophila melanogaster, activity of the rDNA promoter is directly correlated with the number of IGS subrepeats that contain a promoter duplication [3,4]. Selection studies also support the idea that IGS structure is important for rDNA transcription and consequently for growth rate. Cluster et al. [5] found a relationship between IGS length and development time in D. melanogaster, where lines selected for fast development had on average, longer IGS variants (attributed to more copies of the promoter duplication) than lines selected for slow development. Similarly, under selection for rapid growth rate, average IGS length increased in Daphnia pulex [6], and after selection for high yield, the frequency of long spacers increased in maize [7]. Furthermore, longer spacers are associated with higher growth rates in several species of Daphnia [8]. Although these studies suggest that IGS length may have significant effects on life history traits, the evolutionary significance of IGS length remains a mystery. An initial step to understanding the evolutionary role of IGS length is to characterize its variation across a group of related taxa that could then be used to test hypotheses about the adaptive significance of IGS length using the comparative method. The main goal of the present paper is to characterize IGS length variation across a wide range of Drosophila species (subfamily Drosophilinae) and determine the amount of variation observed across different evolutionary time scales.
Figure 1 Ribosomal DNA array in Drosophila melanogaster. Diagram shows position of PCR product and primers, Hae III restriction sites at the ends of the intergenic spacer (IGS) region, and hybridization probe used in this study [modified from 9, 14].
Studies of IGS length variation have commonly relied on Southern Blot hybridization (SB) for inference of IGS length. However, more recently, several studies have used PCR to determine IGS length [6,8-10]. Each of these methods has advantages and disadvantages. The main disadvantage of PCR is that certain fragments (particularly the smaller ones) may be selectively amplified, and that it may not amplify large (> 4 kb) fragments. Thus, the amplified products may not represent the actual size frequency distribution of the IGS. The main disadvantage of SB is that it requires more DNA to begin with. Both methods require previous knowledge about the DNA sequence; either for the design of primers for PCR or for the selection of restriction digestion enzymes for Southern blot hybridization. The secondary goal of this study is to compare the usefulness of each method in estimating IGS length across a group of related taxa when knowledge about sequence of IGS and flanking regions is restricted to a small subset of the taxa under examination. Therefore, design of PCR primers and selection of restriction enzymes is based on this limited number of sequences.
Results
We examined IGS length variation based on Southern Blot hybridization SB and/or PCR in 71 species of the Drosophilinae representing 20 species groups in the genus Drosophila and four other genera (Table 1). Of these, only 52 yielded PCR product. Therefore, for the remaining 19 species we were able to infer IGS length based only on SB. Of the 52 species for which we obtained PCR product, 29 (representing 11 species groups in two subgenera of the genus Drosophila) had a single restriction site near one of the ends. This was the restriction digestion pattern we expected based on the D. melanogaster map (i.e., Hae III site No. 2; Figure 1). From the remaining 23 species, nine did not have a restriction site within the PCR fragment; thus SB-based inferences of IGS length in these nine species would have overestimated the true IGS length. In contrast, the remaining 14 species had more than one restriction site within the PCR fragment. Therefore, SB-based inferences of IGS length in these 14 species would have underestimated the true IGS length.
Table 1 IGS length results for each of the species examined in this study. IGS size index, IGS size range, three most dominant IGS sizes (% relative proportion), and number of bands. Presence and distance of the Hae III restriction site(s) from the end of PCR-amplified IGS products are indicated for each species.
Species Size Index Size Range Size of lost fragments (bp) after digestion of PCR products with HaeIII Dominant 1 Dominant 2 Dominant 3 No. of bands No. of females Strain ID or source locality
Based on Southern Blot Hybridization (SB) and PCR
Subgenus Sophophora
Species group: melanogaster
D. yakuba (SB) 5.30 2.63–7.76 5.22 (33.2) 6.98 (15.9) 4.6 (13.0) 12 9 14021-0261.0
D. yakuba(PCR) 3.88 2.68–4.64 150 4.64 (30.5) 4.01 (23.4) 3.62 (21.7) 6 1 14021-0261.0
D. melanogaster (SB) 6.37 3.87–12.89 5.5 (17.6) 5.2 (13.8) 6.9 (11.4) 15 1 Tucson, AZ
D. melanogaster (SB) 5.94 2.64–14.79 5.22 (11.8) 4.89 (10.2) 5.66 (9.0) 19 6 Tucson, AZ
D. melanogaster (SB) 5.73 5.08–7.84 5.62 (54.7) 5.36 (21.2) 6.29 (10.2) 9 1 Tucson, AZ
D. melanogaster(PCR) 2.72 1.98–3.11 240 3.11 (63.2) 2.12 (23.7) 1.98 (13.2) 3 1 Tucson, AZ
D. eugracilis (SB) 3.64 2.37–7.32 4.28 (22.8) 3.94 (15.0) 4.61 (10.4) 12 5 14028-0451.3
D. eugracilis(PCR) 4.83 4.83 470, 170 4.83 (100) 1 14028-0451.3
D. varians (SB) 7.49 7.44–8.24 7.44 (93.2) 8.24 (6.8) 2 8 14024-0431-0
D. varians(PCR) 3.11 2.64–3.96 210 2.64 (64.2) 3.96 (35.8) 2 1 14024-0431-0
D. kikkawai (SB) 5.19 3.18–11.38 4.8 (17.6) 4.36 (13.1) 4.61 (9.6) 17 7 14028-0561.0
D. kikkawai(PCR) 4.21 3.55–4.89 435 4.48 (29.5) 3.55 (26.4) 4.89 (23.5) 4 14028-0561.0
D. auraria (SB) 4.81 3.5–4.91 4.91 (90.6) 4 7 14028-0471.0
D. auraria(PCR) 2.86 2.86 520 2.86 (100) 1 1 14028-0471.0
D. kanapiae (SB) 5.69 5.58–5.78 5.78 (54.9) 5.58 (45.1) 2 6 14028-0541.0
D. kanapiae(PCR) 5.44 4.73–5.95 540 5.95 (58.1) 4.73 (41.9) 2 1 14028-0541.0
D. parvula (SB) 4.48 3.44–6.35 5.14 (18.9) 4.88 (17.9) 4.4 (14.2) 11 6 14028-0621.0
D. parvula(PCR) 5.46 5.46 620 5.46 (100) 1 1 14028-0621.0
Species group: obscura
D. pseudoobscura (SB) 3.30 1.68–4.27 3.65 (30.2) 3.36 (25.4) 3.93 (11.5) 10 10 TE 1198-2
D. pseudoobscura(PCR) 3.30 3.30 150 3.30 (100) 1 5 TE 1198-2
D. persimilis (SB) 3.35 1.94–4.5 3.67 (17.0) 3.34 (16.5) 3.02 (15.1) 10 3 14011-0111.24
D. persimilis(PCR) 3.23 2.71–3.99 150 3.40 (36.1) 3.11 (23.9) 2.90 (18.6) 6 5 14011-0111.24
D. miranda (SB) 4.54 3.78–5.29 4.86 (36.4) 4.56 (28.6) 4.28 (18.3) 6 5 14011-0101.11
D. miranda(PCR) 3.78 2.93–4.45 180 4.14 (17.5) 3.85 (17.4) 3.54 (16.0) 7 1 14011-0101.11
D. affinis (SB) 4.73 2.82–7.46 4.32 (10.4) 4.75 (10.2) 5.08 (10.02) 16 6 14012-0141.0
D. affinis(PCR) 3.43 2.93–4.11 180 3.14 (29.6) 3.45 (18.8) 2.93 (18.1) 6 14012-0141.0
D. bifasciata (SB) 5.54 3.19–8.68 5.58 (9.7) 5.24 (9.2) 4.91 (8.9) 15 3 14012-0181.1
D. bifasciata(PCR) 3.62 2.84–4.77 200 3.76 (18.2) 3.29 (16.5) 3.54 (16.2) 8 1 14012-0181.1
Species group: willistoni
D. paulistorum (SB) 11.60 7.08–16.56 10.53 (34.1) 11.67 (33.4) 15.67 (13.6) 6 8 14030-0771.11
D. paulistorum(PCR) 3.55 3.38–3.82 340 3.38 (62.0) 3.82 (38.0) 2.87 (7.2) 2 1 14030-0771.11
D. equinoxialis (SB) 6.77 5.4–10.04 6.22 (34.2) 7.93 (18.3) 6.9 (17.7) 6 7 14030-0741.0
D. equinoxialis(PCR) 2.74 2.52–3.07 170 2.52 (57.9) 3.07 (34.9) 3 1 14030-0741.0
D. willistoni (SB) 5.90 5.65–6.5 5.65 (70.4) 6.5 (29.6) 2 7 14030-0811.3
D. willistoni(PCR) 1.32 1.31–3.07 200 1.31 (99.0) 2.59 (0.3) 1.52 (0.3) 4 1 14030-0811.3
Subgenus Drosophila
Species group: virilis
D. novamexicana (SB) 5.11 3–8.18 5.33 (16.9) 6 (16.1) 4.96 (15.9) 11 5 15010-1031.4
D. novamexicana(PCR) 4.33 1.51–5.11 200 5.11 (42.0) 4.70 (41.0) 1.51(17.0) 4 1 15010-1031.4
D. lummei (SB) 5.56 4.88–7.64 5.53 (56.7) 5.17 (19.5) 4.88 (7.5) 6 4 15010-1011.2
D. lummei(PCR) 4.74 3.73–4.97 190 4.97 (70.2) 4.47 (15.4) 3.73 (5.0) 5 1 15010-1011.2
D. virilis (SB) 5.53 2.53–7.39 5.5 (27.7) 6.08 (21.6) 4.86 (16.2) 15 3 15010-1051.12
D. virilis(PCR) 1.57 1.42–3.56 110 1.56 (60.5) 1.45 (39.2) 3 1 15010-1051.12
D. kanekoi (SB) 8.01 6.37–8.9 8.15 (47.1) 7.33 (30.1) 8.9 (20.2) 4 1 15010-1061.0
D. kanekoi(PCR) 6.75 6.04–7.60 185 6.04 (45.4) 7.60 (54.6) 2 1 15010-1061.0
Species group: repleta
D. nigrospiracula (SB) 5.25 4.75–6.17 4.75 (34.9) 5.65 (33.8) 5.02 (14.2) 5 1 15081-1503.1
D. nigrospiracula(PCR) 4.32 3.35–5.28 290, 210 5.28 (25.4) 4.04 (25.2) 3.78 (25.0) 6 1 15081-1503.1
Species group: dreyfusi
D. camargoi (SB) 5.01 3.51–8.21 4.57 (11.01) 4.33 (10.9) 3.94 (10.1) 12 3 15060-1221.2
D. camargoi (PCR) 3.74 2.90–4.32 150 4.32 (21.64) 3.76 (16.98) 4.00 (16.3) 7 1 15060-1221.2
Species group: mesophragmatica
D. gaucha (SB) 6.67 5.06–8.18 6.05 (21.6) 6.64 (21.5) 7.25 (21.2) 5 4 15070-1231.0
D. gaucha(PCR) 4.41 2.49–5.23 160 2.49 (41.9) 5.23 (69.1) 2 1 15070-1231.0
Species group: nannoptera
D. pachea (SB) 9.00 6.98–14.35 8.04 (32.4) 6.98 (25.2) 8.76 (16.6) 6 7 Ejido
D. pachea(PCR) 7.89 7.89 620 7.89 (100) 1 1 Ejido
D. nannoptera (SB) 4.90 3.35–11.5 3.35 (23) 7.22 (12.7) 3.94 (11.8) 13 6 15090-1692.2
D. nannoptera(PCR) 4.12 3.39–4.70 670 3.39 (26.3) 4.70 (24.5) 4.27 (22.1) 5 1 15090-1692.2
Species group: immigrans
D. nasuta (SB) 6.30 5.16–7.09 6.49 (40.0) 6.13 (24.8) 7.09 (13.5) 6 1 15112-1781.9
D. nasuta(PCR) 2.98 2.79–3.67 250 3.67 (74.9) 3.49 (19.1) 3.67 (6.0) 3 1 15112-1781.9
D. albomicans (SB) 6.54 2.36–10.48 7.22 (14.5) 8.89 (13.28) 6.73 (13.18) 12 10 15112-1751.2
D. albomicans(PCR) 1.34 1.20–1.61 240 1.61 (67.3) 1.20 (13.0) 1.29 (7.6) 3 1 15112-1751.2
Species group: tripunctata
D. tripunctata (SB) 4.62 1.4–10 6.47 (14.3) 5.61 (13.2) 7.51 (12.3) 8 4 15220-2401.2
D. tripunctata(PCR) 4.12 2.98–5.32 180 3.83 (36.4) 5.32 (24.8) 3.71 (19.1) 6 1 15220-2401.2
Species group: testacea
D. putrida (SB) 3.73 3.73 3.73 (100) 1 7 15150-2101.1
D. putrida(PCR) 3.84 3.84 590 3.84 (100) 1 1 15150-2101.1
Based on PCR only
Subgenus Sophophora
Species group: melanogaster
D. ananassae 4.15 4.15 220, 830 4.15 (100) 1 1 14024-0371.3
D. pallidosa 4.33 3.16–4.82 400, 520, 860 4.17 (78.5) 4.82 (24.9) 3 1 14024-0433.1
D. greeni 2.72 2.27–4.18 160, 1066 2.27 (65.5) 4.18 (20.9) 2.51 (7.1) 4 1 14028-0712.0
D. seguyi 2.40 2.40 842, 490, 450 2.40 (100) 1 1 14028-0671.0
D. lini 3.38 3.33–3.64 120, 710, 1700, 1880, 2080 3.33 (81.2) 3.64 (18.8) 2 1 14028-0581.0
D. mayri 3.65 3.54–3.72 130, 210, 440 3.72 (60.1) 3.54 (39.8) 2 1 14028-0591.0
D. birchii 2.42 1.88–2.93 no Hae III site 2.93 (49.6) 1.98 (25.1) 1.88 (25.3) 2 1 14028-0521.0
D. baimaii 4.65 3.38–5.49 500, 640 4.92 (51.9) 3.38 (25.7) 5.49 (22.4) 3 1 14028-0481.1
Subgenus Dorsilopha
D. busckii 4.04 3.6–4.87 1547, 1016 3.60 (65.5) 4.87 (34.5) 2 1 Anza Borrego, CA
Subgenus Drosophila
Species group: repleta
D. bifurca 4.77 2.90–5.91 140, 190, 900 5.91 (44.7) 4.64 (29.2) 2.90 (26.1) 3 1 15085-1621.0
D. mojavensis 4.21 3.25–7.09 no Hae III site 3.72 (37.0) 4.00 (19.9) 4.99 (9.1) 6 1 San Carlos, Son., Mexico
D. mojavensis 4.81 4.81 no Hae III site 4.81 (100) 1 1 Ensenada de los Muertos, B.C.S., Mexico
D. arizonae 4.31 4.05–5.39 no Hae III site 4.05 (80.7) 5.39 (19.3) 2 1 Tucson, AZ
D. mettleri 7.09 7.09 220, 370, 750, 2950, 3250 7.09 (100) 1 1 Organ Pipe National Monument, AZ
D. aldrichi 5.47 5.47 no Hae III site 5.47 (100) 1 1 Tucson, AZ
Species group: bromeliae
D. bromeliae 4.39 1.63–4.50 200, 250, 440, 1640 4.50 (75.4) 4.05 (24.6) 4 1 10585-1682.0
Species group: funebris
D. funebris 4.20 3.93–4.36 no Hae III site 4.36 (76.1) 3.93 (23.9) 2 1 15120-1911.3
Species group: calloptera
D. ornatipennis 4.33 3.29–5.00 210, 320, 630 5.00 (45.3) 3.29 (31.3) 4.72 (23.4) 3 1 15160-2121.0
Species group: cardini
D. dunni thomasiensis 4.04 2.27–4.31 150, 560, 1000, 1060 4.03 (49.1) 3.85 (34.2) 4.31 (24.0) 4 1 15182-2301.0
D. parthenogenetica 3.54 3.33–4.16 no Hae III site 3.33 (53.3) 3.53 (27.4) 4.16 (19.3) 3 1 15181.2221.0
Other genera
Scaptodrosophila lebanensis 6.13 6.13 no Hae III site 6.13 (100) 1 1 11020-0021.0
Hirtodrosophila duncani 5.03 4.23–6.07 no Hae III site 4.23 (35.9) 6.07 (34.0) 4.75 (30.0) 3 1 92000-0075.0
Zaprionus ghesquerei 4.98 3.52–8.45 880, 600, 290 5.31 (35.7) 6.35 (16.6) 3.51 (15.2) 6 1 50000-2743.0
Based on Southern Blot Hybridization (SB) only
Subgenus Sophophora
Species group: melanogaster
D. teissieri 4.86 4.6–6.08 4.75 (58.3) 5.04 (23.7) 4.6 (13.5) 4 5 14021-0257.0
D. teissieri 5.20 4.95–6.27 4.95 (58.4) 5.22 (28.6) 6.27 (12.9) 3 12 14021-0257.0
D. barbarae 5.61 3.31–10.68 4.78 (30.3) 6.2 (18.5) 5.26 (14.1) 10 10 14028-0491.1
D. punjabiensis 7.16 4.14–8.64 7.24 (15.3) 7.68 (14.2) 8.05 (14.0) 11 7 14028-0641.0
D. bicornuta 10.77 4.88–16.67 11.91 (56.8) 9.67 (15.7) 8.57 (11.6) 6 6 14028-0511.0
Subgenus Drosophila
Species group: virilis
D. montana 6.42 1.39–18.83 10.55 (9.8) 8.38 (9.8) 6.69 (9.4) 16 4 15010-1021.24
D. borealis 7.84 5.38–8.8 7.82 (46.5) 7.23 (26.6) 8.8 (23.7) 5 4 15010-0961.0
Species group: repleta
D. hydei 5.16 3.44–8.66 4.5 (18.3) 5.88 (17.3) 4.81 (17.0) 10 3 15085-1641.28
D. navojoa 7.18 3.1–11.73 6.7 (26.63) 7.18 (20.2) 6.08 (14.6) 13 5 15081-1374.0
D. micromettleri 5.67 4.1–11.55 5.15 (38.0) 6.2 (35.7) 5.52 (11.0) 8 7 15081-1346.0
D. eremophila 5.02 2.45–13.47 4.9 (23.1) 5.91 (13.5) 7.06 (11.4) 13 6 15081-1292.0
D. wheeleri 7.18 6.47–9.76 7.03 (21.9) 6.69 (21.4) 6.47 (19.9) 8 2 15081.1501.1
Species group: robusta
D. robusta 8.14 2.91–14.59 9.04 (16.1) 10.2 (14.0) 7.19 (14.0) 16 3 15020-1111.5
Species group: melanica
D. melanica 5.13 3.78–5.71 5.4 (35.2) 5.71 (23.4) 5.06 (14.4) 7 4 15030-1141.3
Species group: nannoptera
D. acanthoptera 16.61 8.59–19.97 19.97 (63.8) 11.95 (18.2) 9.89 (11.7) 4 4 15090-1693.0
D. wassermani 13.11 8.77–18.25 12 (33.1) 14.72 (24.9) 8.77 (23.2) 4 5 15090-1697.10
Species group: picture wing
D. grimshawi 5.01 4.65–5.25 5.25 (60.8) 4.65 (39.2) 2 1 15287-2541.0
Species group: tumiditarsus
D. repletoides 9.72 1.68–18.35 8.36 (12.8) 18.35 (12.8) 15.3 (11.9) 22 4 15250-2541.0
Species group: polychaeta
D. polychaeta 3.61 2.38–4.81 3.62 (23.3) 3.33 (20.0) 3.06 (15.2) 10 4 15070-1231.0
Repeated SB of the same DNA extracts revealed very similar patterns of IGS length variation. Similarly, repeated PCR amplifications of the same DNA extracts revealed similar patterns. However, in some cases, examination of different numbers of individuals or DNA amounts of the same species resulted in slightly different patterns of IGS length variation. Nevertheless, the size index for each species was very similar across different numbers of individuals and different DNA amounts (results not shown).
Comparison of methods
We compared IGS size index based on SB and PCR for the 29 species for which presence of one restriction site was confirmed (Figure 2). Average IGS length (i.e., the size index) ranged from 3.3 kb in D. pseudoobscura to 11.6 kb in D. paulistorum (Figure 2; Table 1). It is unlikely that the large fragments resulted from incomplete digestion because use of different amounts of restriction enzyme and of DNA resulted in similar patterns (results not shown). In most cases, the size index based on PCR was smaller than the one inferred from SB (Figure 2; Table 1), although in a few cases they were almost equal and in three cases (i.e., D. eugracilis, D. parvula, and D. putrida), the PCR-based estimates were actually larger. The difference between the size index based on SB and the one based on PCR ranged from zero in D. pseudoobscura to 8 kb in D. paulistorum.
Figure 2 Intergenic spacer (IGS) size index (i.e., weighted average length) based on Southern Blot hybridization and PCR in females of 29 species. Species group to which they belong is indicated below each species group.
To compare the results from both methods for the 29 species for which presence of the expected single restriction site was confirmed, we performed least squares regression analysis as implemented in JMP [11] of the following variables: (1) PCR-based IGS index on SB-based IGS index (Figure 3a); (2) the IGS size difference between the two methods (i.e., SB minus PCR) on SB-based IGS index (Figure 3b). Our results suggest that although in most cases PCR-based sizes were smaller than SB-based sizes (i.e., most data points fell below the dashed line; Figure 3a), no relationship exists between the IGS size index inferred from SB and that inferred from PCR (i.e., the regression is not significant). In other words, the difference between the two methods is not consistent across taxa. The differences observed between the two methods can be attributed to: (1) the possibility that the Hae III site No.1 (Figure 1), which occurs upstream of the forward primer (and thus, not within the PCR amplified product), was lost and therefore the IGS size based on SB was overestimated; (2) measurement error; (3) differences in the length of the sequences adjacent to the IGS that are targeted by each method (i.e., the PCR fragment is not the same as the SB fragment; see Figure 1); or (4) the tendency of the PCR to amplify smaller fragments. Given the highly conserved nature of the region where Hae III site No.1 is found, we do not expect this site to have been lost often within Drosophila. However, loss of this restriction site is a concern in the case of D. paulistorum due to the large difference between PCR and SB results. Although, we lack an estimate of measurement error, our results based on multiple IGS-length estimates of the same taxon with a single method never differed by more than 1 kb (see Table 1). Similarly, based on the known sequence of the regions adjacent to IGS in several Drosophila species, the expected difference between the PCR and SB estimates should not exceed ~200 bp. Thus, we adopt the criterion that the difference between the two methods should be greater than 1 kb (dashed line Figure 3b) for it to be regarded as a true difference due to the method used. Based on this criterion, in about half of the comparisons the PCR-based estimates were smaller than SB-based estimates; most of which had a SB-based estimate of 5 kb or more. This is further illustrated by the observation that the size difference between the two methods increased as the size based on SB increased (Figure 3b), suggesting that the larger the IGS fragment, the greater the degree of underestimation based on PCR. This relationship is still significant after removing the results from D. paulistorum (not shown; P = 0.0019). An alternative, explanation is that the PCR-based results were accurate and thus the degree of overestimation by the SB method increases as the true IGS size decreases. This is unlikely however, because in D. melanogaster for example, the true length of the most common variant is known based on sequence data, and PCR-inferred IGS lengths of this species were always smaller.
Figure 3 Comparison of PCR and SB methods. Subset of 29 species for which presence of a single restriction site within the intergenic spacer (IGS) could be confirmed (see text). a. Relationship between IGS size index estimated from PCR and from Southern Blot hybridization (y = 0.1465x +3.012; r2 = 0.0324; P = 0.3498). For reference, dashed line represents equal PCR and SB values. b. IGS size index difference between Southern Blot hybridization and PCR vs. IGS size index based on Southern Blot hybridization. (y = 0.8535x - 3.012; r2 = 0.5321; P < 0.0001). Dashed line indicates SB-based index minus PCR-based index = 1 kb.
IGS length variation across the genus Drosophila
Unless otherwise noted, we discuss IGS length variation in Drosophila based only on the 29 species for which presence of Hae III site No.2 could be confirmed and for which no evidence of additional restriction sites within the IGS was observed (see above). We observed broad variation in IGS length across the genus Drosophila. The IGS size index ranged from 3.3 kb in D. pseudoobscura to 11.6 kb in D. paulistorum (Figures 2 and 4; Table 1). Even if we exclude D. paulistorum from our interpretation (see above), the IGS size index range is still broad, with D. pachea (9 kb) representing the species with the largest IGS index.
Figure 4 Intergenic spacer (IGS) size index (i.e., weighted average length) based on Southern Blot hybridization in females of 29 species of the Drosophilinae subfamily. Cladogram of phylogenetic relationships among species groups is based on Remsen and O'Grady [31]. Relationships within species groups are based on: melanogaster group [32]; virilis group [33, 34]; obscura group relationships [35]; and willistoni group [36]. Numbers above nodes indicate approximate date of divergence (in million years) based on Tamura et al. [19], unless otherwise noted. 1 based on divergence of the melanogaster subgroup versus the montium and ananassae subgroups [19]; 2 based on Pitnick et al. [18]; 3 based on Russo et al. [27].
Within species groups
IGS size index variation within species groups based on SB was lower than across the subfamily (Figure 4). For example, among the eight species examined from the melanogaster species group the largest difference between species was 3.9 kb (i.e., D. eugracilis vs. D. varians). Similarly, among the four species from the virilis species group, the largest difference was 3 kb (i.e., D. novamexicana vs D. kanekoi). The largest difference among five species in the obscura group was even smaller; 1.4 kb between D. pseudoobscura and D. affinis. The largest difference was observed in the willistoni species group; 5.7 kb between D. willistoni and D. paulistorum, but with the caveat that the result for D. paulistorum may be an overestimation (see above). The difference between D. nannoptera and D. pachea (nannoptera group) also was relatively large; 4.1 kb.
Comparisons between more closely related species suggest that they tend to have very similar IGS indices: D. parvula (4.5 kb) vs. D. kanapiae (5.7 kb); D. novamexicana (5.1 kb) vs. D lummei (5.6 kb); and D. persimilis (3.4 kb) vs. D. pseudoobscura (3.3 kb). The only exception was the comparison between D. paulistorum (11.6 kb) and D. equinoxialis (6.8 kb), but as mentioned above, the value for D. paulistorum may be an overestimation.
Within species variation
Based on SB, all of the species except one (i.e., D. putrida) had more than one IGS length variant (Table 1). In most cases, the length difference between the shortest and longest IGS length variant was at least 3 kb. However, for species in which more than one individual was used, we cannot distinguish between intra- and inter-individual variation. Nevertheless, we found differences among species in the number of fragments present in species where we examined single individuals; Drosophila robusta and D. melanogaster had the highest number of bands per individual (16 and 9–15, respectively), whereas D. grimshawi had only two bands and D. putrida had only one (Table 1).
Results based on PCR only
IGS sizes based on PCR were generally smaller than IGS sizes based on SB (Figure 2). Nevertheless, it is interesting to point out patterns of IGS size variation in the species for which SB could not be used due to the absence of one or both of the Hae III restrictions sites or to the presence of additional restriction sites in the IGS region. For the subset of species that were examined only by PCR, the size index ranged from 2.4 kb in D. seguyi and D. birchii, to 7.1 kb in D. mettleri (Table 1). Very close relatives or sister species tended to have similar lengths. For example, D. ananassae (4.1 kb) vs. D. pallidosa (4.3 kb); D. greeni (2.7 kb) vs. D. seguyi (2.4 kb); and D. arizonae (4.3 kb) vs. D. mojavensis (4.2–4.8 kb). Although these PCR-based values may not represent the true IGS size index (see below), they may provide a minimum estimate for IGS size.
Results based on SB only
For the remaining 19 species, we only report IGS sizes based on SB because we were unable to obtain PCR product. However, these results should be considered with caution because the presence of Hae III site No. 2 or of additional restriction sites could not be assessed. There are several possible explanations for our inability of obtain PCR product in these species. First, it is possible that IGS fragments were not amplified because they were too large (i.e., the largest fragment we were able to amplify was 7.9 kb in D. pachea). It is important to note that amplification of the IGS region in Drosophila is not trivial because of the length (i.e., usually above 3 kb), and the high degree of secondary structure present in this region [12]. Second, the priming sites could have diverged, although our PCR primers target highly conserved regions of the 28S and 18S ribosomal genes. Finally, despite having tried a large variety of amplification conditions, we may not have found the appropriate ones for that particular species.
Based on the SB results, the largest IGS size index observed was 16.6 kb in D. acanthoptera. One of its relatives in the nannoptera group, D. wassermani, also had a large IGS size index, 13.1 kb. Although it is possible that these results based on SB are an overestimation of IGS size (i.e., loss of a restriction site), it is interesting to note that D. pachea, another member of the nannoptera group (for which we were able to confirm the presence of Hae III site No.2), also had a relatively large IGS size of 9 kb. Interestingly, the most basal member of this group, D. nannoptera, had a much smaller IGS index of 4.9 kb. These SB-based results also suggest broad variation in the melanogaster species group, with D. bicornuta having the largest IGS index of 10.8 kb.
Discussion
Our study showed that in about half of the taxa examined IGS length estimates based on PCR were smaller than those estimated with SB, particularly when IGS sizes exceeded 5 kb, suggesting that PCR tends to underestimate the true IGS length because of selective amplification of smaller fragments. A comparison with results from previous studies suggests this. For example, studies of IGS length variation in D. melanogaster based on SB show that this species has many fragments larger than 5 kb [5,13]. In contrast, the studies that used PCR to infer IGS length in this species found that amplified fragments were always smaller than 4 kb [9,10]. Therefore, SB seems to be the most appropriate method for IGS length inference. However, knowledge about the sequence is required, or at least the presence of the appropriate restriction sites on the ends of the fragment of interest should be confirmed. For example, in the present study, we were able to confirm the presence of one of these restriction sites (i.e., Hae III site No. 2), by PCR amplification of IGS, followed by restriction digestion of PCR products. Nevertheless, the PCR fragment ideally should span the region that contains both restriction sites, because in at least one case (i.e., D. paulistorum), we suspect the other restriction site may have been lost. Unfortunately, we were unable to obtain amplification with PCR primers that spanned the region that contained both restriction sites.
IGS size variation
Our study revealed that IGS size index variation among species of the subfamily Drosophilinae is broader than previously reported [5,13-17]: from 3.3 kb in D. pseudoobscura to 9 kb in D. pachea and possibly 11.6 kb in D. paulistorum. Considering that D. pseudoobscura diverged from D. pachea, 40–63 million years ago, and from D. paulistorum 35–62 million years ago (Figure 4), the large IGS length differences are not surprising; particularly in light of the observation that average IGS length has been demonstrated to change rapidly after artificial selection in D. melanogaster [i.e., 24 generations to shift the average size from 5.54 to 5.8 kb and 15 generations to shift the average size from 5.54 to 5.12 kb; [5]].
Despite the speed at which IGS has been shown to evolve under selection, comparisons between very closely related taxa, including sister species pairs, suggest that they tend to have very similar IGS indices. For example, the close relatives, D. parvula and D. kanapiae differ from each other by 1.2 kb; and D. novamexicana and D. lummei [i.e., ~6 million-year-divergence; [18]] differ by 500 bp. An even more closely related species pair [i.e., ~ 0.85 million-year-divergence; [19]], D. persimilis and D. pseudoobscura, differ only by 100 bp. The only exception was the comparison between D. paulistorum and D. equinoxialis, who differ by 4.8 kb, but as mentioned above, the value for D. paulistorum may be an overestimation. Although the PCR-based results should be interpreted with caution, they may offer additional insight regarding patterns of IGS length across closely related taxa. For example, D. ananassae differs from its close relative D. pallidosa by 200 bp; D. greeni differs from D. seguyi by 300 bp; and D. lini differs from D. kikkawai (based on PCR) by 800 bp. Finally, the IGS index of D. arizonae is within the range of values reported for two populations of its sister species D. mojavensis, from which it diverged approximately 1–1.2 million years ago [20,21].
Comparisons of more distantly related taxa, even within the same species group show less clear patterns. For example, the four members of the montium subgroup of the melanogaster species group examined in this study (i.e., D. kikkawai, D. auraria, D. parvula, and D. kanapiae) differ from each other by a maximum of 1.2 kb. On the other hand, based on our results and the ones of Coen et al. [14], IGS size index ranges from 3.6 to 6 kb among eight members of the melanogaster subgroup (D. melanogaster, D. mauritiana, D. simulans, D. erecta, D. yakuba, D. teissieri, D. orena, and D. eugracilis) another subgroup within the melanogaster species group. Similarly, based on our results and the ones from Rae et al. [17], IGS size index ranges from 4.2 to 8.0 kb across nine members of the virilis species group; thus, the largest IGS index observed in this group almost doubles the smallest one. A large difference (i.e., 4.1 kb) is also observed between D. pachea and D. nannoptera (nannoptera group). Nevertheless, despite being members of the same species group, these two taxa may be up to 32 million years divergent [18], providing a long period for the accumulation of such differences. Unfortunately, similar comparisons of IGS length and divergence time are not possible for many of the taxa examined in this study because we lack divergence time estimates.
The observation that close relatives tend to have similar IGS indices, whereas more distant relatives may not, is consistent with the observation that closely related taxa exhibit a high degree of DNA sequence homology of the IGS region [as observed among members of the melanogaster subgroup; [22]], whereas more distantly related taxa exhibit no DNA sequence homology [as reported between the subgenera Sophophora and Drosophila; [16,23]], despite showing similarities in structure such as promoter duplications.
The ecological and evolutionary implications of the broad variation in IGS size observed across members of the Drosophilinae are largely unknown. However, several evolutionary mechanisms appear to play a role in the evolution of IGS variation. First, as a member of the ribosomal DNA multigene family, IGS is subject to concerted evolution [24]. The pattern of concerted evolution appears to be the result of unequal crossing over taking place both, at the level of the subrepeat arrays within the intergenic spacers, and at the level of the complete rDNA units (i.e., genes plus spacers) [25]. The former would create new IGS length variants while the latter would spread a particular variant across the chromosome(s). In addition, the high within-species specificity of the RNA polymerase I complex [26] suggests that IGS coevolves with components of transcriptional machinery. Finally, individual IGS variants may be adaptive, particularly with regard to developmental rate, as suggested by studies of two unrelated taxa, Drosophila melanogaster [5] and Daphnia pulex [6,8]. This observation has led to the suggestion that IGS length alone makes a considerable contribution to growth rate differences and hence life history evolution among related species [6]. Although in long evolutionary time scales, IGS length is highly variable across Drosophila, it does not appear vary broadly in shorter time scales. However, examination of other species may reveal additional variation in shorter time scales that may provide the necessary variation for testing this hypothesis. Nevertheless, tests of this hypothesis will only be informative if IGS length is accurately measured. Furthermore, even if IGS length is found to be adaptive, the crucial assumption that IGS length represents the number of promoter copies should ultimately be tested by DNA sequencing.
Conclusion
Broad variation in average IGS length occurs across large evolutionary scales in members of the subfamily Drosophilinae. However, despite the potential for rapid changes in IGS length shown by artificial selection studies, closely related taxa tend to have similar IGS sizes. Our comparison of methods suggests that PCR-based estimations tend to underestimate the true IGS size when the IGS size is greater than 5 kb and thus, in the absence of DNA sequence information for all the taxa under examination, both methods should be used.
Methods
Taxon selection
To examine the extent of IGS length variation across the subfamily Drosophilinae, where possible, we examined at least one species per major species group. Our taxon sampling scheme spanned divergences of at least 40 [27] or 63 [19] million years based on the estimated average divergence between members of the subgenus Sophophora and the subgenus Drosophila (genus Drosophila). To assess the amount of IGS length variation present in shorter evolutionary time scales, we examined closely related species, including sister species pairs.
Southern Blot
We extracted DNA from 1–10 individual female flies per species. We only examined female flies to prevent any sex bias in our interpretation. In D. melanogaster, and a few other species, Nucleolar Organizer Regions (NORs) are found on the X and Y chromosomes but in other species the locations are not entirely clear [28,29]. Furthermore, previous studies have shown that the Y-linked IGS variants of D. melanogaster are not related to differences in development time [13]. Whole flies were homogenized in 250 μl of DNAzol (Invitrogen, Carlsbad, CA) and 0.1 mg of Proteinase K, and incubated overnight at room temperature. Following centrifugation to discard cellular debris, DNA was precipitated by addition of 125 μl of 100% ethanol and overnight incubation at -20°C. The DNA pellet was recovered by centrifugation, then washed twice with 70% ethanol. DNA was resuspended in 20–100 μl of sterile deionized water and incubated 2–3 hr at 65°C.
We digested DNA extracts overnight with Hae III (New England Biolabs (Beverly, MA) following manufacturer's instructions, and treated with 0.4 μg/μl RNAse A for 5 min at room temperature. The Hae III enzyme was selected because Hae III sites are found on either end of the IGS in Drosophila melanogaster (Figure 1), and its distant [i.e., 40–63 million years divergent; [19,27]] relative D. virilis [17]. Hae III site No. 1 is also present in a distant relative of D. melanogaster, D. hydei, but sequences of the 3' end of IGS are lacking for this and other species, so presence of Hae III site No. 2 has not been confirmed. We then treated the digested DNA with SDS to a final concentration of 0.1% and proteinase K to a final concentration of 20 μg/ml followed by a 30 min incubation at 37°C. This treatment improves the migration of DNA during electrophoresis by removal of contaminating protein [30].
Samples were electrophoresed on 0.9% agarose gels and blotted onto positively charged Nylon membranes (Roche Applied Science, Indianapolis, IN) with the VacuGene XL (Amersham Biosciences Corp, Piscataway, NJ) according to manufacturer's instructions. DNA on the membrane was then UV crosslinked with the Stratalinker (Stratagene, Cedar Creek, TX) according to manufacturer's instructions.
We used a ~300 bp portion of the highly conserved 3' end of the 28S gene as a hybridization probe (Figure 1). We first amplified the fragment of interest in a solution containing ~1 μl of template in a final concentration of 1% DMSO, 20% Betaine, 0.2 mM dNTPs, 5 mM MgCl, 0.2 μM primers (28S-R3665 5'-TTATTTATCATTGCAGTCCAGCACGG-3' and 28S-F3349 5'CATAGCGACGTCGCTTTTTGATCC-3'), 2 units of Taq Polymerase (Invitrogen, Carlsbad, CA) and 1X of Buffer provided by manufacturer. The PCR template contained a mix of genomic DNA from one species per species group examined. The temperature profile had an initial denaturation of 2 min at 95°C, followed by 35 cycles of 1 min at 95°C, 1 min at 58°C and 1 min at 72°C, and a final extension of 7 min at 72°C. The amplified product was electrophoresed on an agarose gel and the fragment of interest was excised and used as template for an asymmetric PCR. This reaction was identical to the first one with the exception that we used less 28S-R3665 primer (final concentration of 0.002 μM) and we substituted regular dNTPs with those contained in the DIG DNA Labeling Mix (Roche Applied Science, Indianapolis, IN) to a final concentration of 0.4 mM. The labeled product was purified by Ethanol precipitation with Sodium Acetate, resuspended in ~100 μl, and added to hybridization buffer (below).
Pre-hybridization, hybridization, and high stringency washes were performed in a hybridization oven. All other incubation/washes were performed with slight agitation. Hybridization and washing solutions were prepared from two stock buffers: 20X SSC (3M NaCl, 300 mM sodium citrate, adjust with Citric Acid to pH 7.0) and 1X Maleic Acid (0.1M Maleic Acid, 0.15M NaCl adjusted with NaOH to pH 7.5). We incubated blotted membranes 2 hr at 68°C in prehybridization buffer [5X SSC; 2% (w/v) blocking reagent (Roche Applied Science, Indianapolis, IN) dissolved by heating; 0.1% N-lauroylsarcosine; and 0.02% SDS (w/v)]. We then incubated membranes overnight at 68°C in hybridization buffer (same as prehybridization buffer plus probe). Hybridization buffer (with probe) was boiled for at least 10 min prior to incubation. We washed hybridized membranes (five 5-min washes at room temperature) with a low stringency buffer (2X SSC containing 0.1% SDS). We then washed membranes (three 10-min washes at 68°C) with a high stringency buffer (0.1X SSC containing 0.1% SDS). Membranes were then equilibrated 2 min in washing buffer (1X Maleic Acid; 0.3% (v/v) Tween 20). We incubated membranes 45 min at room temperature in blocking solution (2% (w/v) blocking reagent in 1X Maleic Acid; dissolved by heating). We then incubated membranes 45 min at room temperature in Antibody solution (i.e., blocking solution and 1:10,000 Anti-Digoxigenin-AP, Roche Applied Science, Indianapolis, IN). Membranes were washed (two 10-min washes) in washing buffer and equilibrated (2–5 min) in detection buffer (0.1M Tris-HCl, 0.1m NaCl, pH 9.5). We added the chemiluminescent substrate CSPD (Roche Applied Science, Indianapolis, IN) following manufacturer's protocol and exposed the membrane to Kodak Biomax light-1 X-ray film for 15–180 min).
Analysis
X-rays were photographed with a Kodak Edas 290 digital camera and analyzed with Kodak 1D 360 software to determine molecular weight of each band observed as well as its relative intensity (with respect to other bands in the same lane). We use relative intensity as a proxy of relative copy number of each band.
We estimated the weighted average spacer length index (I) for each lane based on the length (i.e., molecular weight) and proportion (i.e., relative intensity) of each band as in Cluster et al. [5]:
where n is the number of spacer bands in a lane, Si is the fragment size (or molecular weight) of each band, and Pi is the relative intensity. Si was estimated by comparison with standards.
PCR amplification of IGS
To evaluate the consistency of PCR and Southern Blot hybridization (SB) in estimation of IGS length, we amplified the IGS region from females of the same strains examined by SB. Our PCR reactions (25 μl total volume) contained ~2 μl of template in a final concentration of 1% DMSO, 20% Betaine, 0.4 mM each dNTP, 3 mM MgCl, 0.2 μM primers IGSF2 5'-GTGCTGGACTGCAATGATAAATAAGG-3' (K. Glenn, unpublished) and IGSR1 5'-AAGCATATAACTACTGGCAGGATCAACC-3' (Y-C. Li, unpublished), 2 units of Taq Polymerase (Invitrogen, Carlsbad, CA) and 1X of Buffer provided by manufacturer. The IGSF2 primer is located in a conserved region at the 3'-end of the 28S gene; approximately 300 bp downstream of Hae III site No. 1 in two distantly related species [i.e., D. melanogaster; 23 and D. hydei; GenBank Acc. Nos. M21017 and AF465783, respectively; see Figure 1]. The IGSR1 primer is located in a relatively conserved region of the 18S gene; approximately 200 bp downstream of Hae III site No. 2 in D. melanogaster and D. virilis [23]; two distantly related species. Therefore, the amplified IGS fragments were expected to have a single restriction site near the 3'end. Following PCR, half of the amplified product was treated with Hae III to establish whether Hae III sites existed within the PCR amplified fragments. The Hae III-treated and untreated PCR products were run side by side on 1% agarose gels.
Authors' contributions
MM designed and conducted the experiments and analyses, and drafted the manuscript. TAM conceived the study, and participated in its design and coordination and helped to draft the manuscript. Both authors read and approved the final manuscript.
Acknowledgements
We thank the Tucson Stock Center for providing fly stocks. The laboratories of M. Wells and M. Kidwell provided access to equipment. J. Isoe, K. Glenn, Y-C Li, and L.J. Weider provided valuable technical advice. L.A. Hurtado and L.J. Weider provided helpful comments on earlier versions of this manuscript. This work was funded by NSF-IRCEB (#9977047) to TAM.
==== Refs
Moss T Stefanovsky VY Promotion and regulation of ribosomal transcription in eukaryotes by RNA polymerase I Prog Nucleic Acid Res Mol Biol 1995 50 25 66 7754036
Paule MR Lofquist AK Zimmermann RA and Dahlberg AE Organization and expression of eukaryotic ribosomal RNA genes Ribosomal RNA Structure, Evolution, Processing and Function in Protein Biosynthesis 1996 Boca Raton, CRC Press 395 419
Grimaldi G Di Nocera PP Multiple repeated units in Drosophila melanogaster ribosomal DNA spacer stimulate rRNA precursor transcription pnas 1988 85 5502 5506 2840664
Grimaldi G Fiorentini P Di Nocera PP Spacer promoters are orientation-dependent activators of pre-rRNA transcription in Drosophila melanogaster Mol Cell Biol 1990 10 4667 4677 2117701
Cluster PD Marinkovic D Allard RW Ayala FJ Correlations between development rates, enzyme activities, ribosomal DNA spacer-length phenotypes, and adaptation in Drosophila melanogaster PNAS 1987 84 610 614 3099299
Gorokhova E Dowling TA Crease T Weider LJ Elser JJ Functional and ecological significance of rDNA IGS variation in a clonal organism under divergent selection for production rate Proceedings of the Royal Society of London B 2002 269 2373 2379 10.1098/rspb.2002.2145
Rocheford TR Osterman JC Gardner CO Variation in the ribosomal DNA intergenic spacer of maize population mass-selected for high grain yield Theoretical and Applied Genetics 1990 79 793 800 10.1007/BF00224247
Weider LJ Glenn KL Kyle M Elser JJ Associations among ribosomal (r)DNA intergenic spacer length, growth rate, and C:N:P stochiometry in the genus Daphnia Limnology and Oceanography 2004 49 1417 1423
Polanco C González AI de la Fuente A Dover GA Multigene family of ribosomal DNA in Drosophila melanogaster reveals contrasting patterns of homogenization for IGS and ITS spacer regions: a possible mechanism to resolve this paradox Genetics 1998 149 243 256 9584100
Polanco C González AI Dover GA Patterns of variation in the intergenic spacers of ribosomal DNA in Drosophila melanogaster support a model for genetic exchanges during X-Y pairing Genetics 2000 155 1221 1229 10880483
SAS Institute JMP 1989 4.0.2 Cary, North Carolina, SAS Institute
Baldridge GD Dalton MW Fallon AM Is higher-order structure conserved in eukaryotic ribosomal DNA intergenic spacers Journal of Molecular Evolution 1992 35 514 523 1474605 10.1007/BF00160212
Clark AG Szumski FM Lyckegaard EM Population genetics of the Y chromosome of Drosophila melanogaster: rDNA variation and phenotypic correlates Genetical Research 1991 58 7 13 1936985
Coen E Strachan T Dover GA Dynamics of concerted evolution of ribosomal DNA and histone gene families in the melanogaster species subgroup of Drosophila J Mol Biol 1982 158 17 35 6811755 10.1016/0022-2836(82)90448-X
Williams SM DeSalle R Strobeck C Homogenization of geographical variants at the Nontranscribed Spacer of the rDNA in Drosophila mercatorum Mol Biol Evol 1985 2 338 346 3870866
Renkawitz-Pohl R Glätzer KH Kunz W Characterization of cloned ribosomal DNA from Drosophila hydei Nucleic Acids Research 1980 8 4593 4611 6255425
Rae PMM Barnett T Murtif VL Nontranscribed spacers in Drosophila ribosomal DNA Chromosoma 1981 82 637 655 6266780 10.1007/BF00285773
Pitnick S Markow TA Spicer GS Evolution of multiple kinds of female sperm-storage organs in Drosophila Evolution 1999 53 1804 1822
Tamura K Subramanian S Kumar S Temporal patterns of fruit fly (Drosophila) evolution revealed by mutation clocks Mol Biol Evol 2004 21 36 44 12949132 10.1093/molbev/msg236
Matzkin LM Population genetics and geographic variation of alcohol dehydrogenase (Adh) paralogs and glucose-6-phosphate dehydrogenase (G6pd) in Drosophila mojavensis Molecular Biology and Evolution 2004 21 276 285 14660699 10.1093/molbev/msh019
Matzkin LM Eanes WF Sequence variation of alcohol dehydrogenase (ADH) paralogs in cactophilic Drosophila Genetics 2003 163 181 194 12586706
Hayward DC Glover DM The promoters and spacers in the rDNAs of the melanogaster species subgroup of Drosophila Gene 1989 77 271 285 2502472 10.1016/0378-1119(89)90075-9
Tautz D Tautz C Webb D Dover GA Evolutionary divergence of promoters and spacers in the rDNA family of four Drosophila species J Mol Biol 1987 195 525 542 3116264 10.1016/0022-2836(87)90181-1
Ruiz Linares A Bowen T Dover GA Aspects of nonrandom turnover involved in the concerted evolution of intergenic spacers within the ribosomal DNA of Drosophila melanogaster Journal of Molecular Evolution 1994 39 151 159 7932779
Dover GA Brown S Coen ES Dallas J Strachan T Trick M Dover GA and Flavell RB The dynamics of genome evolution and species differentiation Genome evolution 1982 New York, Academic Press 343 372
Grummt I Roth R Paule MR Ribosomal RNA transcription in vitro is species specific Nature 1982 296 173 174 7063022 10.1038/296173a0
Russo CAM Takezaki N Nei M Molecular phylogeny and divergence times of Drosophilid species Mol Biol Evol 1995 12 391 404 7739381
Ashburner M Drosophila: A Laboratory Handbook and Manual 1989 Cold Spring Harbor, New York, Cold Spring Harbor Laboratory Press 1331
Stuart WD Bishop JG Carson HL Frank MB Location of the 18/28S ribosomal RNA genes in two Hawaiian Drosophila species by monoclonal immunological identification of RNA.DNA hybrids in situ Proc Natl Acad Sci USA 1981 78 3751 3754 6943580
EPICENTRE® company homepage [http://www.epicentre.com/f5_1/f5_1tt.asp]
Remsen J O'Grady PM Phylogeny of Drosophilinae (Diptera: Drosophilidae), with comments on combined analysis and character support Mol Phylogenet Evol 2002 24 243 264 10.1016/S1055-7903(02)00226-9
Schawaroch V Phylogeny of a paradigm lineage: the Drosophila melanogaster species group (Diptera: Drosophilidae) Biological Journal of the Linnean Society 2002 76 21 37 10.1046/j.1095-8312.2002.00044.x
Spicer GS Reevaluation of the phylogeny of the Drosophila virilis species group (Diptera, Drosophilidae) Annals of the Entomological Society of America 1992 85 11 25
Spicer GS Bell CD Molecular phylogeny of the Drosophila virilis species group (Diptera : Drosophilidae) inferred from mitochondrial 12S and 16S ribosomal RNA genes Annals of the Entomological Society of America 2002 95 156 161
O'Grady PM Reevaluation of phylogeny in the Drosophila obscura species group based on combined analysis of nucleotide sequences Mol Phylogenet Evol 1999 12 124 139 10381316 10.1006/mpev.1998.0598
Gleason JM Griffith EC Powell JR A molecular phylogeny of the drosophila willistoni group: conflicts between species concepts? Evolution 1998 52 1093 1103
|
16111495
|
PMC1215480
|
CC BY
|
2021-01-04 16:29:16
|
no
|
BMC Evol Biol. 2005 Aug 19; 5:46
|
utf-8
|
BMC Evol Biol
| 2,005 |
10.1186/1471-2148-5-46
|
oa_comm
|
==== Front
BMC EcolBMC Ecology1472-6785BioMed Central London 1472-6785-5-61608349610.1186/1472-6785-5-6Research ArticleStoichiometric estimates of the biochemical conversion efficiencies in tsetse metabolism Custer Adrian V [email protected] Department of Environmental Science, Policy and Management, 201 Wellman Hall #3112, University of California, Berkeley, Berkeley, CA 94720, USA2005 5 8 2005 5 6 6 10 2 2005 5 8 2005 Copyright © 2005 Custer; licensee BioMed Central Ltd.2005Custer; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
The time varying flows of biomass and energy in tsetse (Glossina) can be examined through the construction of a dynamic mass-energy budget specific to these flies but such a budget depends on efficiencies of metabolic conversion which are unknown. These efficiencies of conversion determine the overall yields when food or storage tissue is converted into body tissue or into metabolic energy. A biochemical approach to the estimation of these efficiencies uses stoichiometry and a simplified description of tsetse metabolism to derive estimates of the yields, for a given amount of each substrate, of conversion product, by-products, and exchanged gases. This biochemical approach improves on estimates obtained through calorimetry because the stoichiometric calculations explicitly include the inefficiencies and costs of the reactions of conversion. However, the biochemical approach still overestimates the actual conversion efficiency because the approach ignores all the biological inefficiencies and costs such as the inefficiencies of leaky membranes and the costs of molecular transport, enzyme production, and cell growth.
Results
This paper presents estimates of the net amounts of ATP, fat, or protein obtained by tsetse from a starting milligram of blood, and provides estimates of the net amounts of ATP formed from the catabolism of a milligram of fat along two separate pathways, one used for resting metabolism and one for flight. These estimates are derived from stoichiometric calculations constructed based on a detailed quantification of the composition of food and body tissue and on a description of the major metabolic pathways in tsetse simplified to single reaction sequences between substrates and products. The estimates include the expected amounts of uric acid formed, oxygen required, and carbon dioxide released during each conversion. The calculated estimates of uric acid egestion and of oxygen use compare favorably to published experimental measurements.
Conclusion
This biochemical analysis provides reasonable first estimates of the conversion efficiencies for the major pathways used by tsetse metabolism. These results now enable a deeper analysis of tsetse ecology based on the construction of a dynamic mass-energy budget for tsetse and their populations.
==== Body
Background
The dynamics of mass and energy flows in tsetse (Glossina Wiedemann), the time-varying rates of biomass ingestion and use, influence most aspects of the ecology of these flies including the nutrition and growth of individuals, the intrinsic growth rate of tsetse populations, and the transmission dynamics of the tsetse vectored trypanosomiases. A particularly effective approach to the examination of these dynamics involves the assembly and evaluation of a dynamic mass-energy budget [1-3] constructed specifically for tsetse. Such a budget could capture the essence of the time varying flows of biomass and energy in tsetse, could explain these dynamics mechanistically based on the internal physiology of the flies, and could form the core of a physiologically based model of tsetse population dynamics [4-6]. The budget and population model could, in turn, form the core of an epidemiological model of trypanosomiasis dynamics [7,8] because the tsetse vectored trypanosomiases are transmitted when tsetse feed to acquire biomass and energy.
At the center of any dynamic mass-energy budget applicable to heterotrophs lies the conversion of food substrates into usable products, either body tissue or metabolic energy. The efficiencies of these conversions influence the rates of mass and energy flow since the rate of food intake must at least equal, but may exceed, the rate of end product use divided by the conversion efficiency. Food conversion involves, first, the breakdown of food substances leading either to their absorption or to their egestion and, then, the post-absorptive processing of nutrients leading either to their assimilation as body mass, to their consumption as energy, or to their excretion either because they are too costly to process or because they could potentially accumulate to toxic levels. The ultimate yield from biomass conversion will reflect the richness of the food source, the loss of food which is not digested, the loss of resources which are either directly egested or absorbed but then excreted because they are too costly to assimilate, the costs involved in conversion, the efficiency of the conversion process, the cost of removal of potentially toxic substances in the food, the costs of transportation and growth necessary to digestion, and the energetic richness of the resulting substances.
This paper develops quantitative estimates of the efficiencies of biomass conversion in tsetse. These conversion estimates will provide the working numbers needed to elucidate tsetse energetics and thereby construct a dynamic mass-energy budget which can quantify the need for food of individual flies based on their rates of respiration, development, growth, and reproduction.
Approaches to estimation
The fundamental principles of mass and energy conservation constrain the efficiency of conversion in metabolic reactions to lie at or below unity. Other principles from thermodynamics, physics, chemistry, and biochemistry then successively lower the maximum theoretical yield by including the impact of more inefficiencies and costs, thereby achieving more realism and constraining the estimate sequentially closer to the actual biological efficiency experienced by the organism.
The balance of mass can be treated separately from the balance of energy since biological systems do not involve nuclear transformations but the two balances are coupled so they can be leveraged jointly for estimation. The physical law of mass conservation, when applied at the organismal level, requires that, across any time interval, the original mass of the organism plus the mass of food material ingested must equal the final mass of the organism plus the mass of material egested. This was the analytic approach of van Helmont's famous 1648 willow growth experiment [9]. Chemistry refines the law of mass conservation into a law of element balance in which the masses of each element can be treated separately and each must balance across all the biological reactions. This approach is widely used in ecology notably in the field of ecological stoichiometry [10] and in the recent studies of isotope flows. The biochemical approach further refines the estimate by considering the fate of specific structural molecular fragments. This is the approach that is required for nutritional studies of the trophic flow of vitamins or of necessary and non-synthesizable amino acids. The biological approach refines the estimate even further by considering explicitly indigestible tissue structures such as the seeds in fruit. The strategy of mass balance provides a useful component for the estimation of conversion yields because each of these approaches defines a strict equality between the start and the end of the conversion, albeit with successively more detail in the entity being balanced.
The balance of energy in the conversion process can also be used in estimating the efficiency of conversion. While a full energetic budget would consider the radiation balance of organisms as well as the balance in the potential energy of chemical bonds, this study focuses on the balance of chemical energy only, because the biologically useful energy is exclusively that of chemical bonds and because, in heterotrophs, the radiation budget only affects the heat balance of the organism. The omission of radiation from the energy budget alters the analysis from one of a strict equality between the inputs and outputs to one of a constraining inequality because the energy lost as heat is not quantified. The laws of thermodynamics place an upper limit on the efficiency of energetic conversion: the first law requires that the conversion product have at most the same amount of energy as the substrate while the second law states that the efficiency of conversion must be less than unity. Since the potential energy of chemical bonds is only accessible by conversion of the original bonds to new bonds, the approach of physical chemistry examines the potential energy available from the complete oxidation of the food and body tissue with a suitable oxidant. This is the approach taken by the studies which use calorimeters to burn food, exuvium, and body tissues in oxygen and then calculate the metabolic efficiency of energy conversion based on the difference in potential chemical energy. This approach has been used to obtain a static energy budget for tsetse [11]. The approach of biochemistry reduces the estimate of conversion efficiency from that provided by physical chemistry because it considers explicitly the energy recovered through the reactions of biochemical conversion rather than the energy potentially recoverable from the overall chemical reaction. The difference between the two can be significant with the core reactions involved in pyruvate catabolism only capturing, at a maximum, 33% of the chemically available energy [12]. The biological approach further reduces the estimate of energetic recovery efficiency by considering the inefficiencies of conversion due to the imperfection of biological structures and by including the biological costs of digestion, of transport, and of storage. The biological approach would, for example, consider the inefficiencies in the production of ATP due to leakage of the mitochondrial membrane and the costs of maintenance of the digestive lumen, formation of digestive enzymes, manufacture of metabolic enzymes, formation of new cells for the growth of internal organs used in digestion [13-15], and synthesis of molecular complexes for transport such as the joining of glycerides to proteins. However, these biological inefficiencies and costs do not necessarily need to be considered separately. Biological inefficiencies, if they are static or have reasonable average values, can be incorporated directly into the stoichiometric estimation process. Biological costs, to the extent that they involve inhalation of oxygen or exhalation of carbon dioxide, can be accounted for as separate costs in an energetic budget rather than explicitly included in the estimate of conversion efficiency.
Experimental approaches to the estimation of conversion yields have not yet obtained estimates for the major biomass conversions in multicellular animals. Experiments on whole organisms have examined the conversion costs as part of the 'work of digestion', alternatively called the 'specific dynamic action', 'feeding heat increment', or 'postprandial thermogenesis'. These studies use calorimetric measurements and measurements of post-digestive respiration to derive estimates of the costs of digestion in metabolic energy production. A study of this kind was able to estimate that the costs of digestion in a python species account for around a third of the energy available in the food [13]. More detailed experiments are attempting to resolve the different costs but these studies have not yet separated biological costs from the biochemical costs nor, apparently, have such experiments yet distinguished the costs of body tissue formation from the costs of metabolic energy production. Experiments on single cells are achieving remarkable results and now aim to model the full chemical kinetics to estimate the actual production yields of specific metabolic products, especially those which are industrially important. Such experiments are beginning to determine quantitatively the realized conversion efficiencies based on measurements of the flux of the 13C isotope and drawing on the use of mass spectrometry, gas chromatography, and nuclear magnetic resonance detection [16,17]. However, these conversion estimates depend extensively on coupled stoichiometric modeling [18,19] so these are not simply experimental approaches but rather integrations of experimental and theoretical strategies.
This analytic approach
This paper derives quantitative estimates of upper theoretical yields for each of the major conversions in tsetse metabolism using essentially a biochemical approach. The approach was chosen because it could rapidly obtain a complete set of estimates based on a mechanistic interpretation and do so with more accuracy than alternative approaches. The biochemical approach promised to be rapid since it could be performed solely through calculation; indeed, the early success of an initial quick study led to its expansion into this complete analysis. Because the scientific literature now provides all the data required for the analysis, including the composition of tsetse food and tissues, the metabolic pathways, and the reaction sequences, the biochemical approach is able to obtain a complete set of estimates for each of the major conversions. The biochemical approach also provides a mechanistic basis which can explain the yield estimates as the sum of yields and costs in component reactions rather than being simply an empirical measure derived from experiment. The biochemical approach is expected to provide a more accurate estimate than would a calorimetric approach because it considers explicitly more of the costs and inefficiencies of the conversion process.
The calculations presented in this work are necessary because no applicable preexisting results could be found. Tsetse metabolism is unusual in combining a protein diet, the use of uric acid for the disposal of nitrogen, and the use of the amino acid proline as the primary substrate for energy production during flight. These peculiarities bring difficulties to the reuse of existing stoichiometric calculations. The reference values which would be necessary to avoid the calculations presented in this paper could not be found despite consultations with physiologists specializing in bioenergetics and a search of the general literature. The literature on tsetse energetics also fails to provide suitable values. The most complete analysis of tsetse metabolic energetics is the static energy budget presented by Bursell and Taylor [11] which integrates most of the earlier research. That paper presents empirical estimates for both the conversion efficiency of blood to body fat and the consumption of oxygen during fat catabolism but does not provide estimates for the conversion efficiencies of the other major metabolic pathways.
This calculation of conversion yields involves several steps. Tsetse metabolism is simplified down to its core pathways. Tsetse food and body tissues are reduced to simple representative compositions. Each metabolic pathway is expanded into a single series of reactions which lead from the composition of the substrates to that of the products and by-products. Metabolic reactions are considered to occur in a steady state system in which a single input is entirely converted and the system restored to its original state. Based on this view, the yields can be derived from simple stoichiometric calculations. This approach derives estimates of the upper theoretical yields for each of the metabolic conversions considered important in tsetse metabolism and obtains estimates of the uric acid formed to dispose of nitrogenous waste, of the oxygen required, and of the carbon dioxide released in each conversion.
This biochemical approach to the calculation of conversion yields has several conceptual consequences. This approach implicitly considers conversion as a fixed process with static conversion efficiencies since the approach describes, for each conversion, only the single, most efficient pathway, and any such reaction sequence will follow Avogadro's principle of fixed numerical proportions between the substrates and products of chemical reactions. Since tsetse feed on a single, invariant food substrate, namely vertebrate blood [20,21], this analysis does not consider changing efficiencies of metabolism due to food switching or nutritional variation. This greatly simplifies the analysis and budgeting of tsetse energetics by allowing a fixed conversion rate between food and products, for example, an adult fly body mass can be explicitly quantified in terms of its bloodmeal equivalents. Another consequence of the approach is that the conversion pathways are implicitly considered independent of each other.
A simplified description of tsetse metabolism
This analysis simplifies tsetse metabolism to the conversions which account for the bulk of the transformations of food into metabolic energy, of food into body mass, and of storage body mass into metabolic energy.
Tsetse feed uniquely on vertebrate blood which is composed of approximately 80% water and 20% protein. The amounts of lipids, carbohydrates, and nucleic acids in blood are negligible, between 0.8% and 1% of wet blood or 4% and 5% of the dry weight [22,23], so tsetse can be assumed, as a first approximation, to feed exclusively on protein. Tsetse egest haematin unprocessed but digest the remaining blood protein, absorbing the constituent amino acids into their haemolymph [24,25]. Therefore, the tsetse food source consists essentially of a pool of amino acids which are converted into metabolic energy or into tsetse body tissue.
Metabolic energy is provided by a number of molecules including the reducing molecules nicotinamide-adenine dinucleotide (hereinafter NADH) and flavin-adenine dinucleotide (FADH2) and including adenosine triphosphate (ATP). These each have other closely related molecules with essentially identical roles in metabolism. In this paper, all these energetic molecules are treated for simplicity as their equivalent amount of ATP.
Tsetse body tissue composition is variable but is approximately 70% water, 10% ether soluble dry weight, and 20% residual dry weight (RDW) [26,27]. The ether soluble fraction of dry weight is assumed to consist entirely of fat (Fat), i.e. triglycerides, since this is the component of body mass which is soluble in ether. The non-fat dry weight of most insects consists of around 70% protein, 10% structural carbohydrates, 10% other carbohydrates, 6% minerals, with the remainder miscelaneous substances such as nucleic acids [28]. Glycogen and other free carbohydrates, which are important for energy storage in most insects, are essentially absent from tsetse [25] so that protein should contribute around 78% of tsetse RDW. As a first approximation, used for simplicity, this study considers all of the RDW to be protein, i.e. polypeptides.
In pregnant female tsetse a significant portion of the bloodmeals are used to produce the milky secretion (milk) of the milk gland which nourishes the offspring in utero. This secretion comprises the major gain in body mass of females during pregnancy. This milk consists entirely, by dry weight, of fat and protein [29-31] and is passed on directly to the embryo which has a similar composition. The metabolic pathways of pregnancy therefore involve the conversion of blood to fat and of blood to protein, which are the same pathways as those involved in the manufacture of tsetse body mass.
Tsetse fat acts as an energy reserve which is catabolized for metabolic ATP following two pathways: one pathway, common to all organisms, produces energy for general metabolism and another pathway, specific to tsetse, powers flight by converting alanine to proline and then back [24,32-37].
The major metabolic pathways of tsetse therefore consist of the conversion of amino acids from blood, into the the energy of the outer bi-phosphate bond in ATP, into body or milk fat, or into body or milk protein, and the conversion of fat into ATP along two seperate pathways, one for resting metabolism and one for flight. These pathways are presented schematically as the five arrows of figure 1.
Figure 1 A schematic overview of the major metabolic pathways in Glossina. The major metabolic pathways in tsetse involve the conversion of vertebrate blood obtained by feeding into energy (ATP), fat, or protein. Stored fat is catabolized for energy along two pathways, one for general metabolism and another for flight.
Crude estimates of the conversion efficiencies of these pathways can be derived from calculations based on the chemical energy content of the end products. Each milligram of blood protein could produce at most between 4.098 × 10-4 and 5.465 × 10-4 moles of ATP from adenosine diphosphate (ADP), between 0.33 and 0.44 milligrams of Fat, or 1 milligram body protein, since the change in free energy during protein combustion in oxygen is between -3 and -4 kcal/g, the change from the release of the third phosphate of ATP is -7.32 kcal/mol, and the change due to fat combustion is around -9 kcal/g. Each milligram of fat could produce at most 1.230 × 10-3 moles of ATP from ADP. These estimates are high since they do not include the energetic costs and inefficiencies of the conversion process, the costs of disposal of toxic by-products, or the biological costs of digestion and growth. These crude estimates do not provide the amount of respiratory gases consumed or produced in the conversion process and do not distinguish between the two pathways tsetse use to catabolize fat. More detailed estimates require the use of a stoichiometric approach.
The efficiency of conversion along each of these pathways depends, first, on the detailed chemical composition of the reactants, both substrates and products, in these pathways and, second, on the actual sequences of reactions used in each pathway. These two issues are discussed next.
The detailed chemical composition of the reactants
The exact chemical composition of the reactants involved in these conversions define the end points of the metabolic reaction pathways. This composition constrains the overall yield of the pathway, determines the mass balance of the reactions, and determines the amounts of toxic by-products produced by conversion.
Water, despite forming the major component of the blood food source and of tsetse body tissue, is ignored throughout this analysis because it has no influence on the calculated efficiencies of conversion. Water participates in the biochemical reactions of energy production only as the aqueous reaction medium and as the end product of full catabolism. Since biochemical reactions occur in an aqueous medium, water is never a limiting factor for the reactions. This analysis of the net efficiencies of biochemical conversion can therefore focus entirely on the non-aqueous, 'dry,' components of the tissues.
The biochemical composition of bloodmeal protein is presented in Bursell [22] based on the molar fraction of each amino acids in vertebrate blood. These data are displayed in the second column of table 1. Part of the bloodmeal is not used by tsetse but is egested immediately in the exuvium. The composition of the exuvium is presented in Bursell [22]. The oxygen binding haematin in the bloodmeal is egested without being split into constituent amino acids and both arginine and histidine are absorbed from the gut but then are excreted unmodified, presumably due to their high nitrogen content [22]. Columns 3 and 4 of table 1 show the total nitrogen and carbon mass contained in a pool of one hundred moles of bloodmeal amino acids; both must be accounted for completely in the stoichiometric analysis.
Table 1 Amino acids in blood and muscle tissue
Amino Acids Blood Protein Muscle Protein Milk Protein
Bursell, 1965 Bursell, 1965 Agosin, 1978 Cmelik, 1969
Amino Acid N C Amino Acid Amino Acid Amino Acid
numeric % mol/100 mol mol/100 mol numeric % numeric % numeric %
Alanine 8.3 8.3 24.9 7.4 10.14 5.9
Arginine 3.5 (Excreted) 7.3 6.08 2.6
Asparagine (as Aspartic Acid) (below) 10.66
Aspartic Acid 9.5 9.5 38.0 9.8 (above) 5.7
Cysteine 2.3 2.3 6.9 0.0 0.00 4.5
Glutamine (as Glutamic Acid) (below) 18.70 18.3
Glutamic Acid 8.3 8.3 41.5 20.2 (above)
Glycine 3.8 3.8 7.6 6.8 5.96 0.9
Histidine 6.9 (Excreted) 3.3 1.66 0.8
Isoleucine (as Leucine) (below) 4.12
Leucine 12.9 12.9 77.4 12.5 10.38 8.5
Lysine 9.4 18.8 56.4 9.6 8.59 3.4
Methionine 1.2 1.2 6.0 5.7 2.11 w/ cysteine
Phenylalinine 6.5 6.5 58.5 5.6 2.97 5.7
Proline 4.9 4.9 24.5 4.1 3.03 10.5
Serine 4.3 4.3 12.9 (w/ Glycine) 4.87 4.3
Threonine 5.2 5.2 20.8 3.9 4.25 6.3
Tryptophan 0.2 0.4 2.2 0.0 0.00
Tyrosine 3.1 3.1 27.9 3.1 2.38 16.9
Valine 9.6 9.6 48.0 (w/ Methionine) 4.10 5.7
Totals 99.9 99.1 453.5 99.3 100.00 100.0
The amino acid composition of vertebrate blood and tsetse muscle. The composition is reported as number of molecules of each amino acid in 100 moles of blood or muscle amino acids (numeric %). The nitrogen and carbon composition is the number of those atoms contributed by each amino acid in 100 moles of blood amino acids (mol / 100 mol).
The composition of tsetse body fat is estimated in Langley and Pimley [30]. The fatty acid composition of tsetse fat consists approximately of 40% palmitic acid (16 carbons:0 double bonds), 30% palmitolic acid (16:1) and 20% oleic acid (18:1) [30]. For this analysis, all fat is assumed to consist of triesters of palmitic acid. This approximation should only lead to a minor error in the estimated yield since the sixteen carbon chain represents a reasonable average chain length [22], because the energetic difference between equal masses of different fatty acids are minor, and since the energtic cost of double bond formation is minor.
The composition of tsetse muscle protein is presented in Bursell [22] and can be recalculated from Agosin [38]. Those results are repeated in columns 5 and 6 of table 1. This study considers the amino acid composition of tsetse flight muscle protein to be identical to the composition of vertebrate blood both based on the similarity of the second, fifth, and sixth columns of table 1 and because the differences between the two estimates of tsetse muscle protein composition are on the scale of their differences to vertebrate blood. A more accurate analysis would require a series of dedicated experiments providing estimates of the composition of these tissues and assessing the stability of this composition over time. Based on the simplification used here, the formation of tsetse muscle protein is considered to require a bulk transfer of amino acids obtained from the bloodmeal along with the energetic cost of peptide bond formation.
The composition of the fat produced in the milky secretion of females and incorporated into the developing embryo was analyzed by Langley and Pimley [31]. They found that the fat in milk consists of approximately 65% palmitic acid (16:0), 27% palmitolic (16:1), 7% linoleic acid (18:2), and 2% myristic acid (14:0). As stated above, this analysis considers all fat to be triesters of palmitic acid, an assumption which should not lead to major errors in the final estimated efficiencies of conversion.
The composition of the protein in tsetse milk is presented in an article by Cmelik et al. [29] and repeated here in column 7 of table 1. The amino acid composition of the milk differs from the composition of blood and muscle principally in the elevated fractions of proline, glutamate, and tyrosine [39]. This analysis was unable to integrate this difference in amino acid compositions; instead, the composition of milk protein is assumed identical to the composition of the protein in vertebrate blood and in tsetse muscle. This simplification should lead to an overestimate of the yield of milk mass from a given amount of blood because the three amino acids present in higher proportions have a higher energy content than the average amino acid so the production of these amino acids would require the conversion of part of the blood to energy. However, the simplifying assumption is required by the limitations in the methodology of the study and in the data used for this analysis. A more refined approach is left to future studies.
Tsetse dispose of the excess nitrogen in the bloodmeal by egesting haematin, by excreting arginine and histidine, and by forming uric acid [22,40]. There is no published evidence of the excretion, by tsetse, of ammonium so all disposed nitrogen is assumed to be excreted as uric acid. The formation of uric acid requires both carbon for the molecule's rings and ATP to drive the reaction. These costs of nitrogen disposal must be included in any realistic estimate of the net conversion yields along each of the major metabolic pathways.
In this paper, the five major metabolic pathways presented in figure 1 have been simplified to the following conversions. Tsetse food is assumed to consist exclusively of amino acids in the proportions presented by Bursell [22]. This pool of amino acids is converted into ATP, into triesters of palmitic acid, or into protein with identical amino acid composition to the blood. The triesters of palmitic acid are themselves catabolized to form ATP along either of two metabolic pathways.
The reaction sequence along each pathway
The sequence of reactions used by tsetse for the conversions described by figure 1 can be obtained by linking general descriptions of the metabolic pathways used by tsetse, obtained from the tsetse literature, with specific descriptions of the biochemical reactions which occur in these pathways, obtained from the biochemical literature.
The metabolic pathways used by tsetse combine standard pathways used by most organisms with pathways specific to tsetse. The formation, from amino acids, of ATP and of fat proceed through standard pathways, respectively, the Krebs cycle followed by the electron transport chain and the triglyceride synthesis pathways. The pathway used by tsetse to form uric acid proceeds via glycine [22,24]. Tsetse use both the standard pathway of fat catabolism through the Krebs cycle and use a special pathway for flight energy which combines the common path used to convert fat to proline [24,33,35,41] with the common path used to convert proline to ATP [24,34,36,37] while alanine acts as the transamination reactant [42].
The specific reactions involved in each step of the conversion pathways are similar for most organisms and have been established by biochemists and presented in the biochemical literature. The book Biochemical Pathways by Michal [12] provides an integrated presentation of the biochemical reactions which occur in living organisms. This book describes every reaction involved in each of the metabolic pathways presented in figure 1.
Biologically realistic calculations of the proportions of conversion between substrates (blood or fat) and products (ATP, fat, or protein) in the five pathways presented in figure 1 must include the energetic costs incurred in the reactions of conversion and the costs of disposal of the toxic by-products of each pathway. ATP is required for the formation of both fat and protein and for certain steps in most reaction pathways. To generate the required ATP, a proportional quantity of the reaction substrate must be catabolized. The amino acid diet of tsetse is high in nitrogen whose accumulation could potentially reach toxic levels. The formation of uric acid to excrete this excess nitrogen requires both energy and carbon which must be derived from some of the amino acids in the bloodmeal.
An overview of the biochemical conversions considered here is presented in figure 2. The dark area (gray) includes the Krebs cycle, the lighter area (blue) outlines the triglyceride (fat) metabolic pathway, and the lightest shaded area (green) includes the elements involved in the formation of uric acid. The hexagonal labels denote the amino acids and the rectangular labels identify Krebs cycle products. The entry points of amino acid into the Krebs cycle are taken from McCabe and Bursell [41](see figure 3) and Michal [12](see figure 3.8-1). Threonine (Thr), capable of entering the cycle either as Acetyl-CoA or as Succinyl-CoA, is assigned entirely to Succinyl-CoA following McCabe and Bursell [41]. The estimates of the energetic yields of Krebs cycle constituents are presented in table 2. These are used to obtain the estimates of the energetic yields of amino acid catabolism given in table 3.
Figure 2 The biochemical pathways involved in digestive conversion in tsetse. The biochemical conversion pathways used by tsetse. Amino acids from the bloodmeal (hexagons) enter the reactions of conversion as Krebs cycle constituents (rectangles), as shown by the arrows. The dark area (grey) includes the Krebs cycle, the lighter area (blue) outlines the pathways of fat creation and catabolism, and the lightest area (green) shows the reactions of uric acid formation.
Table 2 Krebs cycle yields and gas exchange
Substrate ATP Created CO2 Released O2 Consumed
Pyruvate 12.5 3 2.5
Acetyl-CoA 10 2 2
2-oxoglutarate 21 5 4
Succinyl-CoA 18.5 4 3.5
Fumarate 16 4 3
Oxalacetate 13.5 4 2.5
The gross energy yield and gas exchange in the catabolism of Krebs cycle elements in atoms per amino acid converted.
Table 3 Amino acid catabolism
Amino Acid Substrate Intermediate Product ATP Created CO2 Released O2 Consumed
Alanine Pyruvate 12.5 3 2.5
Arginine (Excreted)
Aspartate Oxalacetate 13.5 4 2.5
Cysteine Pyruvate 12.5 3 2.5
Glutamate 2-oxoglutarate 21 5 4
Glycine 1/2 Pyruvate 7.5 2 1.5
Histidine (Excreted)
Leucine 3 Acetyl-CoA 33 6 7
Lysine 2 Acetyl-CoA 29 6 6
Methionine Succinyl-CoA 18 5 4
Phenylalinine Fumarate & 2 Acetyl-CoA 36 9 10
Proline 2-oxoglutarate 26 5 5
Serine Pyruvate 12.5 3 2.5
Threonine Succinyl-CoA 20 4 4
Tryptophan 2 Acetyl-CoA & Pyruvate 22.5 11 10.5
Tyrosine Fumarate & 2 Acetyl-CoA 36 9 9
Valine Succinyl-CoA 25 5 5
The gross energy yield and gas exchange in amino acid catabolism based on the estimated energy yields of Krebs cycle elements given in table 2, in atoms per amino acid converted.
Overview of this analysis
The results section presents the estimated yield, uric acid production, and gas exchange in each of the five major metabolic pathways. First, the analysis examines the pathway of conversion of blood into the energy in the outer bi-phosphate bond of ATP, requiring an estimate of the concomitant cost of uric acid formation. Next, conversion estimate of blood to protein is calculated as a bulk transfer of blood amino acids to protein coupled with the cost of protein formation. Third, the conversion of blood to fat is analyzed. Finally, the last two sections consider the catabolic breakdown of fat for the formation of ATP from ADP, first, following the pathway used for resting metabolism, and then, following the pathway used for flight.
The discussion section evaluates the yield estimates by comparing them to empirical measurements described in the literature, then assesses the results, examines the assumptions, and considers the validity of the approach.
Results
The conversion efficiency of each of the five pathways of figure 1 is considered separately below. The spreadsheet file used to calculate the estimates is included with the manuscript both in the open Gnumeric file format (a GNU Zip compressed XML file) [see Additional file 1] and in the unpublished Microsoft Excel file format [see Additional file 2].
Blood catabolism for ATP formation
The first major metabolic pathway involves the conversion of the amino acids obtained from the protein in a bloodmeal into the high energy bonds in ATP used to power cellular processes and muscular motion. All adult tsetse use part of their bloodmeals for this purpose. The net estimate of the yield of this conversion pathway must incorporate the costs of uric acid formation required to excrete all of the nitrogen in the bloodmeal.
The proportion of conversion of blood to ATP is calculated starting, for convenience, with a pool of blood which consists of 100 moles of amino acids in the molar proportions of human blood presented in [22] and repeated in the second column of table 1. The starting pool of amino acids, after accounting for the excretion of arginine and histidine and for the amino acids which contain several nitrogen atoms, contains 99.1 moles of nitrogen and 453.5 moles of carbon, as presented in the totals of the third and fourth columns of table 1.
When bloodmeal amino acids are converted to ATP, all of the absorbed nitrogen is excreted as uric acid [22]. Uric acid formation requires the use of a glycine carbon backbone, the input of carbon from both carbon dioxide and formate, the inclusion of nitrogen from the transamination of glutamine to glutamate and the input of energy. This analysis assigns all of the glycine, serine, and cysteine and part of the alanine in the blood meal to form the glycine precursors to uric acid. These four amino acids are present in sufficient concentrations in the bloodmeal to provide all the glycine required to form enough uric acid to dispose of all the absorbed nitrogen. Both the carbon dioxide and the energy required in the manufacture of this uric acid can be obtained from the catabolic degradation of some of the amino acids in the original pool.
The overall costs of uric acid formation are calculated as conversions of alanine and cysteine to serine, serine to glycine and glycine to uric acid based on the reaction sequences in Michal [12], with the parentheses in this paragraph indicating the figure from that work documenting the reactions. The conversions of alanine to pyruvate (4.2-1) and cysteine to pyruvate (4.5-3) are followed by the conversion of pyruvate to serine (4.4-1). Serine is converted to two glycine molecules (4.4-3) and, finally, the glycine is converted to uric acid (8.1-2 and 8.1-4). The overall costs of uric acid formation from each amino acid precursor are as follows:
1 Alanine + 7 CO2 + 1 O2 + 23 ATP → 2 Uric Acid
1 Cysteine + 7 CO2 + 1 O2 + 23 ATP → 2 Uric Acid
1 Glycine + 3 CO2 + 0.5 O2 + 11.5 ATP → 1 Uric Acid
1 Serine + 7 CO2 + 0.5 O2 + 25.5 ATP → 2 Uric Acid
The numbers of ATP molecules required are higher than the results presented for avian biochemistry by Stevens [43](see page 75) because this analysis includes the two NADH molecules needed to recreate formate for the activation of N10-Formyltetrahydrofolate [12](see figure 9.6-1) and includes a NADH molecule released during the conversion of Inosine-5'-Phosphate to Xanthosine-5'-Phosphate [12](see figure 8.1-4).
Taking 3.89 moles of alanine, 2.30 moles cysteine, and 4.30 moles serine, each of which produce twice the amount of uric acid, and 3.80 moles glycine, 24.78 moles of uric acid are produced. Since each uric acid molecule holds four nitrogen atoms, this quantity of uric acid accounts for the 99.1 moles of nitrogen in the original pool of 100 moles of amino acids (the total of the third column of table 1). This process requires approximately 296 moles of ATP, 85 moles of carbon dioxide, and 10 moles of oxygen.
The remaining amino acid molecules can be fully catabolised through the Krebs cycle to form high energy phosphate bonds in ATP. The reactions of catabolism are described in the text and figures of Michal [12], as follows. The conversions of each amino acid into pyruvate, acetyl-CoA, or Krebs cycle elements are described in the figures of chapter 4 which covers amino acid biochemistry. Additionally, the conversions of pyruvate and acetyl-CoA to elements in the Krebs cycle are presented in figure 3.3-1, the Krebs cycle is shown in figure 3.8-2, and the ATP yields of the oxidation of both NADH and FADH2 are given in section 3.8.3. Using this information, the calculated yields of catabolism are presented in the tables of this paper, table 2 for the Krebs cycle products and table 3 for the amino acids themselves. This conversion requires oxygen for the reduction of NADH and FADH2. The carbon contained in the amino acids is released as carbon dioxide. The sulfur in cysteine and methionine is simply ignored; it is likely excreted in the pigmented fraction of the exuvium [22].
The yield of converting the amino acids remaining from the original pool of 100 moles of bloodmeal amino acids after accounting for the costs of uric acid formation are presented in table 4. The catabolic degradation of the remaining amino acids generates approximately 1900 moles of ATP and 414 moles of carbon dioxide while using about 408 moles of oxygen. The 414 moles of carbon dioxide released in the catabolic breakdown of these amino acids, when combined with the carbon contained in the amino acids used as precursors to uric acid (3 per mole of alanine, cysteine, and serine, 2 per mole of glycine), account for all of the 453.5 moles of carbon in the original pool of amino acids.
Table 4 Yield of available amino acids
Amino Acid Remaining
mol ATP Gain
mol CO2 released
mol O2 used
mol
Alanine 4.41 55.16 13.24 11.03
Aspartic Acid 9.50 128.25 38.00 23.75
Glutamic Acid 8.30 174.30 41.50 33.20
Leucine 12.90 425.70 77.40 90.30
Lysine 9.40 272.60 56.40 56.40
Methionine 1.20 21.60 6.00 4.80
Phenylalinine 6.50 234.00 58.50 65.00
Proline 4.90 127.40 24.50 24.50
Threonine 5.20 104.00 20.80 20.80
Tryptophan 0.20 4.50 2.20 2.10
Tyrosine 3.10 111.60 27.90 27.90
Valine 9.60 240.00 48.00 48.00
Totals 1899.11 414.44 407.78
The results of the catabolic conversion of the amino acids remaining from a pool of 100 moles of blood derived amino acids after accounting for the total disposal of the nitrogen through uric acid formation.
The overall catabolism of the bloodmeal for energetic production is simply the combination of the uric acid disposal pathway and the catabolism of the remaining amino acids. The overall conversion of 100 moles of blood derived amino acids results in 25 moles of uric acid, 1600 moles of ATP, and 330 moles of carbon dioxide while using 420 moles of oxygen. The full catabolism of the blood meal has a calculated respiratory quotient of 0.79, close to the 0.8 value usually used for protein.
These numbers become more meaningful when considered in terms of mass. For convenience we can start with 1.0 mg of dry blood. Dry blood is approximately 20% of the wet weight of blood [22,44] so 1.0 mg dry blood would correspond to 5.0 mg of wet blood meal. Of the 1.0 mg dry blood, approximately 0.88 mg is non-haematin protein available for digestion [22]. In the separation of protein into amino acids, water of hydration is absorbed giving the amino acids 1.14 times the mass of the protein [22] so the resulting mass of amino acids is once again around 1.0 mg. The amino acids in this pool have a proportional average mass of 131.51 g per mole calculated from the proportions of amino acids in blood given by Bursell [22] and the molecular mass of each amino acid. The milligram pool of amino acids therefore contains 7.61 × 10-6 moles of amino acids. Using this number of molecules of blood derived amino acids and the yields established above, a 1.0 mg dry bloodmeal yields 1.22 × 10-4 moles of ATP while requiring the formation of 0.32 mg uric acid. Based on the ideal gas law at standard temperature (25°C) and pressure (1 atmosphere), 1.0 mg of dry blood requires 782 mm3 oxygen and yields 615 mm3 of carbon dioxide when degraded entirely to create ATP.
Blood conversion to protein
The second major metabolic pathway involves the conversion of blood into protein. Adult flies which have just emerged from the puparial stage create protein to increase their flight musculature [45,46]. Mature pregnant females form protein as part of the milk which feeds the developing offspring. Based on the assumption of equivalence between the amino acid composition of blood, of muscle, and of milk protein, the formation of muscle or milk protein consists simply of the direct transfer of a portion of the bloodmeal amino acids coupled with the use of 4 ATP per amino acid [12](see page 135) to form the protein chain. This estimated energetic cost of protein formation ignores the possible energetic recovery from the pyrophosphate released in the reaction linking each amino acid to the transfer RNA molecule [47](see page 963). The other costs of protein manufacture, including all of the cellular processes required to develop and maintain the protein synthesis infrastructure, are also ignored so that this analysis underestimates the costs and therefore overestimates the amount of protein generated from a given bloodmeal.
Starting again with 100 moles of original amino acids, the conversion of this entire pool of amino acids to protein requires 100 moles of peptide bonds or 400 moles of ATP. By expanding the original pool with an additional 25 moles of amino acids, the 400 moles of ATP required can be obtained since the previous section showed that 100 moles of amino acids yield 1600 moles of ATP. Therefore the pool of 100 moles of amino acids which are used for the protein biomass must be combined with the 25 moles of amino acids used to generate energy thereby requiring a total of 125 moles of amino acids from the blood to create protein containing 100 moles of amino acids. Equivalently, 100 moles of blood amino acids can be partitioned into 80 moles to make protein while burning the remaining 20 moles for metabolic energy and uric acid formation. This process would generate 1/5th of the uric acid and carbon dioxide from the previous section and require 1/5th of the oxygen.
In mass terms, 1 mg of dry blood can be used to create 0.80 mg of muscle or milk protein requiring the formation and excretion of 0.06 mg uric acid, the inhalation of 156 mm3 of oxygen, and the exhalation of 123 mm3 of carbon dioxide.
Blood conversion to fat
The third major metabolic pathway in tsetse involves the use of the bloodmeal to create triglycerides in the fat body. All adult flies must convert part of the bloodmeal to energetic reserves and pregnant female flies will use part of the blood allocated to reproduction to create the triglycerides in the secreted milk.
Starting again with an original pool of 100 moles of blood derived amino acids, a first allocation will be to the formation of uric acid. Fat does not contain nitrogen so all of the nitrogen in the bloodmeal must be formed into uric acid and excreted before fat formation. This process is identical to the excretion of nitrogen involved in the pathway converting blood to ATP so the numbers presented in that section work for this pathway as well.
The remaining amino acids provide both the carbon precursors of the fat molecules and the energy required for synthesis. The reactions are taken from Michal [12] and listed here in parenthesis. The synthesis of triesters of palmitic acid requires the combination of one pyruvate molecule to make the glycerol head (figure 3.1-1) with 24 acetyl-CoA molecules to make the fatty acid chains (section 6.1-4). Linking the acetyl molecules into fatty acid chains and then to the glycerol head (section 6.2-1) requires 137 ATP molecules to drive the synthesis. The amino acids remaining after accounting for uric acid formation are therefore assigned proportionally to pyruvate, acetyl-CoA, and ATP. Carbon dioxide is released and oxygen consumed only in the process of catabolic degradation for ATP production.
The balanced use of the amino acids in the bloodmeal after uric acid formation comes from the creation of 4.08 moles of fat, necessitating 4.08 moles of pyruvate, 98 moles of acetyl-CoA (4.08 × 24), and 560 moles of ATP (4.08 × 137). The required pyruvate can be obtained by combining 3.88 moles of alanine, the amount remaining after uric acid formation, with 0.2 moles of pyruvate generated from tryptophan degradation. The 98 moles of acetyl-CoA, shown in table 5, column 3, can be generated with the quantities of amino acids presented in table 5, column 2. The 560 moles of ATP can be generated through the catabolic conversion of the amino acids remaining after accounting for the amino acids used to form uric acid, pyruvate, and acetyl-CoA (shown in table 6). This ATP is the sum of the 654 moles from amino acid catabolism (shown as the total of the third column of table 6) plus the 200 moles arising from the creation of acetyl-CoA (shown as table 5, column 4), minus the 296 moles incurred in the creation of uric acid.
Table 5 Amino acids used to form acetyl-CoA
Amino Acid To Acetyl-CoA
mol Acetyl-CoA
mol ATP Gain
mol CO2 released
mol 02 used
mol
Alanine 0.50 0.50 1.25 0.50 0.25
Aspartic Acid 9.50 9.50 33.25 19.00 4.75
Glutamic Acid 3.04 3.04 33.42 9.11 6.08
Leucine 12.90 38.70 38.70 0.00 12.90
Lysine 9.40 18.80 84.60 18.80 18.80
(Phenylalinine) 6.50 13.00 (Added to energetic catabolism.)
Threonine 5.20 7.80 9.10 5.20 3.90
Tryptophan 0.20 0.40 0.80 0.80 0.80
(Tyrosine) 3.10 6.20 (Added to energetic catabolism.)
Totals 97.94 200.32 52.61 46.68
The amino acids from the bloodmeal assigned in this analysis to the synthesis of acetyl-CoA for the fatty acid chains of the fat molecule. Phenylalinine and tyrosine yield a mixture of Krebs cycle products as shown in figure 1 and only the fractions yielding acetyl-CoA are considered here, the other fractions are considered in the synthesis of ATP presented below. Tryptophan is catabolized into both the pyruvate described above and an acetyl-CoA fraction included in this table.
Table 6 Amino acids used for ATP formation
Amino Acid To ATP
mol ATP Gain
mol CO2 released
mol O2 used
mol
Alanine 0.03 0.41 0.10 0.08
Glutamic Acid 5.26 110.50 26.31 21.05
Methionine 1.20 21.60 6.00 4.80
(Phenylalinine) 6.50 104.00 32.50 39.00
Proline 4.90 127.40 24.50 24.50
(Tyrosine) 3.10 49.60 15.50 15.50
Valine 9.60 240.00 48.00 48.00
Totals 653.51 152.91 152.93
The amino acids catabolised to fuel triglyceride assembly. Only the parts of phenylalinine and tyrosine not used in the creation of pyruvate and acetyl-CoA are consumed to generate ATP.
The net effect of the conversion of 100 moles of blood amino acids into fat is the creation, as before, of 25 moles of uric acid, the creation of 4.1 moles of fat, the release of 120 moles of carbon dioxide, and the consumption of 210 moles of oxygen.
In mass terms, 1 mg of dry blood yields 0.25 mg fat, 0.32 mg uric acid, 224 mm3 carbon dioxide, and requires 391 mm3 oxygen.
Fat catabolism
The two remaining major metabolic pathways both involve the catabolic degradation of fat to obtain ATP. The first pathway is used by tsetse during all stages of life. This pathway proceeds directly through the Krebs cycle. The second pathway is used by adult tsetse to provide energy to flight muscles and involves the transformation of alanine to proline and back. These two pathways are considered below.
The first pathway of triglyceride degradation involves, first, the split of the triester into the glycerol head and the three fatty acid chains [12](see figure 6.2-1), then, the sequential split of an acetyl group from each fatty acid [12](see figure 6.1-9), and, finally, the degradation of both the glycerol and acetyl molecules via the Krebs cycle [12](see figures 3.1-1, 3.3-1, and 3.8-2). The conversion of one triglyceride made of three palmitic acid chains proceeds as follows
1 Triglyceride + 8 ATP → 25 Acetyl-CoA + 24 NADH + 21 FADH2 + 1 CO2
or
1 Triglyceride + 72.5 O2 → 333.5 ATP + 51 CO2
for each triester of palmitic acid.
The second pathway of fat catabolism starts with the degradation of the triglyceride as above to acetyl-CoA. The acetyl-CoA molecules are then used to convert alanine through the Krebs cycle to proline. This proline is transported to the flight muscles and degraded, through the other reactions of the Krebs cycle, back to alanine during flight. The formation of proline, based on the pathway presented in McCabe and Bursell [33](see figure 3) and the reactions in Michal [12], is:
1 Alanine + 1 Acetyl-CoA + 2 NADH + 2 ATP → 1 Proline
This equation differs from that presented in Bursell [35](see figure 3) by accounting for the molecule of ATP required to activate E-Biotin with the carbon dioxide molecule as part of the pyruvate carboxylase pathway [12](see figure 3.3-1) and by accounting for a slight difference between the modern understanding of the Krebs cycle and the interpretation of Bursell's. In the flight muscle sarcosome, the proline is degraded back to alanine yielding:
1 Proline + 3 O2→ 1 Alanine + 2 CO2 + 15 ATP
according to the pathway presented in McCabe and Bursell [41](see figure 3) and the reactions in Michal [12]. This equation differs from that presented by Bursell [37] but only in one atom of oxygen and one of ATP. When these two equations are combined, the degradation of each acetyl radical derived from the fat molecule results in the overall equation:
1 Acetyl-CoA + 2 O2→ 2 CO2 + 8 ATP
This pathway generates two fewer ATP per acetyl radical degraded than does degradation through the Krebs cycle. The amount of oxygen released has been calculated based on offsetting the NADH molecules needed to create proline with those generated in the degradation of proline. Since these processes may occur at different times, this estimate of oxygen use only works as a long term average.
Generating energy through this pathway, using the modern interpretation, would result in the overall degradation of one triglyceride molecule as follows.
1 Triglyceride + 72.5 O2→ 283.5 ATP + 51 CO2
This pathway provides less energy than the general pathway of oxidation presented above, entirely due to the differing yields in the degradation of the acetyl radicals.
Converting these numbers to mass terms, 1 mg of tsetse fat can be used to generate 4.13 × 10-4 moles of ATP for the first pathway or 3.51 × 10-4 moles for the second. Both pathways release 1550 mm3 of carbon dioxide and, despite differing in energetic yield, both require 2200 mm3 of oxygen. The calculated respiratory quotient of fat catabolism is 0.70, which is the value generally used for fats and the value presented by Michal [12](see page 408).
Overall results
The yields in each of the five major metabolic pathways of figure 1 are presented in table 7. The table includes estimates of the proportions of conversion of blood to ATP, fat, and protein and of fat to ATP through the two pathways used by tsetse. The table also includes the proportion of uric acid formed to excrete excess nitrogen and the volumes of carbon dioxide released and oxygen consumed in each conversion process.
Table 7 Overall Results
Substrate Product Uric Acid
(mg) CO2
(mm3) O2
(mm3)
Blood (1 mg dry) ATP 1.221 × 10-4 mol 0.3199 614.5 782.1
Fat 0.2458 mg 0.3199 223.5 391.1
Protein 0.8008 mg 0.06398 122.9 156.4
Fat (1 mg) ATP 4.131 × 10-4 mol - 1545 2197
ATP (via Proline) 3.512 × 10-4 mol - 1545 2197
The calculated yield estimates of the five major metabolic pathways in tsetse and the associated uric acid formed, carbon dioxide released and oxygen used. ATP indicates the additional formation of one phosphate to phosphate bond, Fat refers to triesters of palmitic acid, and Protein indicates polypeptide chains. Volumes were calculated assuming a standard temperature of 25°C and pressure of 1 atmosphere. The accuracy of these numbers exists only to report calculated results; these numbers are probably only accurate to the first significant figure. The estimated volume of oxygen released during fat anabolism is probably too high as explained in the discussion.
Discussion
The estimates presented in table 7 can be evaluated by comparison against experimental results.
Bursell [22] estimated the amount of uric acid excreted following a bloodmeal to be directly proportional to the size of the meal. He estimated that a 1 mg dry mass bloodmeal would lead to the formation of 0.280 mg uric acid for the first meal and 0.342 mg in subsequent meals. These measurements compare favorably with the calculated values presented here of 0.32 mg uric acid per mg dry bloodmeal for the production of either ATP or fat and of 0.06 mg uric acid per mg dry bloodmeal in the production of protein. Flies feeding for the first time are investing much of their bloodmeals into the formation of flight musculature [45,46] so the value measured by Bursell should be due to the use of the bloodmeal for both conversion to ATP and manufacture of protein. The measured value for the first bloodmeal can be explained from the calculated values by assuming 15.4% of the first bloodmeal is used to make protein and the rest is used to make energy or fat. The measured value in later meals, while greater than the calculated values for the production of energy and fat, is nonetheless close to the value calculated here. The calculated results account for all of the nitrogen in the bloodmeal so that it is not necessary to consider the possibility of nitrogen excretion through ammonium.
Rajagopal and Bursell [48] measured the rate of respiration in resting tsetse in the days after feeding. They estimate a three day total respiration of 2184 mm3 oxygen following a bloodmeal of 25 mg wet mass. Using the calculated rates presented here, the same bloodmeal would consume approximately 3900 mm3 of oxygen if all the blood were converted to energy, 2000 mm3 of oxygen if the blood were made into fat, and 800 mm3 of oxygen if the blood were made into protein. Since these numbers bracket the measured result, it is possible to explain the measurement as a combination of the calculated values, demonstrating the plausibility of the estimates presented here.
Puparia of tsetse do not feed but depend on accumulated energetic reserves. Puparia of Glossina morsitans orientalis Vanderplank consume roughly 4200 mm3 oxygen total during this phase of their lifecycle [49] and use approximately 2.0 mg fat [50]. From the calculations presented here, the conversion of 2.0 mg fat to energy would require 4400 mm3 oxygen which falls within 5% of the measured result, again validating these estimates.
These comparisons indicate that the yield proportions calculated in this paper appear plausible as initial estimates despite the disparate sources of analytic data, the many assumptions and simplifications required by the analysis, and the issues with the methodological approach of this work. The stoichiometric approach seems to account for the bulk of experimental observation based on its mechanistic explanation of chemical reaction yields.
While the overall results appear to be plausible, certain of the assumptions on which the yield calculations were based appear especially problematic and should be revisted. The assumption that all metabolic energy can be treated as equivalent to ATP correctly quantifies the energetic costs but can lead to a significant over-estimate of the oxygen required by a conversion pathway. This assumption implicitly assumes that all of the energy obtained from NADH and FADH2 involves the generation of ATP through the electron transport chain with the concomitant use of oxygen. However, some of the conversion pathways involve reactions which use NADH and FADH2 directly. The assumption has a trivial effect on the estimate of the oxygen use in the conversion of blood protein into ATP. Part of the energetic cost of uric acid formation can be met directly by NADH, accounting for 2.5 of the 11.5 ATP equivalents per uric acid molecule. Since this accounts for roughly 22% of this energetic cost, the estimate of required oxygen for this path should be lowered by this amount. However, the effect of the assumption on the estimate of the overall oxygen requirement is insignificant, leading to an increase of two moles out of the 420 moles total oxygen use estimate, an effect of less than 1%. The assumption does have a significant effect on the estimate of oxygen use in the conversion of blood protein into fat. Partly this is due to an overestimate of oxygen use in the formation of uric acid, but that effect is tiny. Predominantly, the effect comes from the energetic cost of triglyceride formation in which, of the 137 ATP equivalents required per triglyceride, 110 can be met with NADH directly. The overall estimate of oxygen use should therefore be lowered by around 77% to approximately 90 mm3, leading to respiratory quotient of around 2.5. The assumption should not have any significant effect on the estimate of oxygen use during conversion of blood into body protein production. A trivially small error does arise due to the assumption because protein formation requires ATP for peptide bond formation and ATP production overestimates oxygen use by around 1% due to the error in uric acid formation. The assumption should not have any effect on the estimates of oxygen use during fat catabolism to ATP along either pathway.
The assumption that tsetse body mass can be treated simply as a combination of water, palmitic acid, and muscle protein ignores the contributions of carbohydrates, such as chitin, of nucleic acids, such as transcription RNA, and of other non-proteinaceous body mass components but this omission should not lead to any major error. A more accurate representation of the body fat fraction of tsetse mass should not greatly alter the results obtained here since the energetic difference in the manufacture of equal masses of different fats is relatively small. The assumption that RDW is composed entirely of muscular protein will affect the results of the analysis depending on the relative proportions of these other tissues, on differences in the costs and efficiencies of the reactions of conversion, and on the differences in the energy content of these tissues compared to protein. Structural carbohydrates account for only around 10% of typical insect body mass [28], have a similar energy content to protein, and differ in formation costs primarily due to uric acid formation which requires only 10% by mass of the original pool of amino acids. The treatment of chitin as protein should therefore only have a small effect on the yield estimate. Non-structural carbohydrates do not form a significant part of tsetse body mass [25]. Nucleic acids account for only around 4% of typical insect body mass [28]. The assumption that RDW can be treated as comprised entirely of protein therefore appears to lead only to a small error. Both the assumption of the equivalence of the amino acid composition of vertebrate blood, tsetse muscle, and female milk and the simplification of protein manufacturing costs to peptide bond formation lead to overestimates in the amount of protein produced from the blood. Preliminary analysis suggests the correct yield might be as low as 0.56 mg milk protein per milligram blood instead of the 0.80 mg reported here. However, accurate estimates incorporating differences in amino acid composition require both better data to demonstrate the consistency of the amino acid composition in different species, samples, locations, and time periods and better methods of calculation able to obtain a true maximum yield for the pathways described. This analysis ignores the pathway of carbohydrate synthesis and catabolism which could be an important, albeit temporary, metabolic pathway. Tsetse do not use carbohydrates to store energy, as was stated earlier, but might nonetheless use the pathway in their metabolism. The cyclic creation and use of carbohydrates from Krebs cycle precursors requires no oxygen, releases no carbon dioxide, and only consumes small amounts of energy. Ignoring this pathway, even if it were used, should therefore not have any significant impact on the calculated yields. These assumptions were necessary in this paper to obtain the estimates of table 7 but are not inherent to the biochemical approach. Future studies could therefore refine these estimates.
The yield estimates presented in table 7, while they improve on the current understanding, could nonetheless be improved. The biochemical approach leads to estimates which are significantly lower than the crude estimates which were presented in the background section and which were based on the energy content of the tissues. The yield estimates of ATP in the catabolic pathways are between 22% and 30% of the crude estimate. The yield estimate of fat is between 56% and 74% of the crude estimates, while the yield estimate of protein is 80% of the crude estimate. These modifications are due to considering the energetic costs of the transformation, to considering the biomass and energetic costs of uric acid formation for the disposal of nitrogen, and to an approach which incorporates the inefficiencies of the biochemical reactions involved in conversion. However, all of the biological costs of digestion, growth, and transcription are still ignored. Because both nitrogen and carbon are tracked explicitly in the stoichiometric equations, the estimated amounts of uric acid formed and of carbon dioxide released should be essentially exact. The estimated amount of oxygen required is still inaccurate. The oxygen requirement depends primarily on the amount of ATP formed through the electron transport chain. As was discussed above, the treatment of all energy as ATP equivalents leads to a significant overestimate of the oxygen requirement in the conversion of blood protein to body fat.
The biochemical approach and the calculation strategy used in this paper succeed in providing working numbers for future research but suffer from a number of limitations. The derivation of tissue composition from data obtained by disparate studies leads to an inconsistent basis for the rest of the analysis and lacks detail such as the variation in these tissue compositions between species and over time. This could be improved through direct experimental analysis. The derivation of the reaction sequence from the figures of Biochemical Atlas [12] does not guarantee obtaining either the real or an optimal pathway. This could be improved using an approach based on stoichiometric matrices [51,52] which would include multiple alternative pathways and might allow numerical determination of the optimum. The determination of the stoichiometric balance of each reaction from graphical representations is prone to error. An alternative derivation based on direct access to the proteinomic databases could avoid this error. The use of a spreadsheet to perform the calculation was problematic and could be improved through the use of an approach based on stoichiometric matrices. In the future, a better approach than that used here would be to improve and use the informatic tools available for this type of study. Such tools would use complete databases of known metabolic reactions to construct stoichiometric matrices which include the reactions along all plausible pathways. The tools would then solve the matrices using computer algorithms to automatically calculate maximum yields of products and byproducts. Several long term projects seem to be working on such an informatic platform but face numerous difficulties in the integration of biochemical data, in the development of graph traversal and optimization algorithms, and in the creation of flexible interfaces for researchers to use.
Since the estimates presented in this paper appear plausible as a first approximation, the impact of these yield ratios can be briefly considered by comparing the efficiency of energy generation along the three ATP generating pathways. The pathway involving conversion of the bloodmeal to fat and then to ATP is 83% as efficient as the pathway of direct conversion of the bloodmeal to ATP. The pathway used to provide the metabolic ATP during flight is 70% as efficient as the simple conversion of blood to ATP and 86% as efficient as the pathway of fat catabolism directly through the Krebs cycle. It is interesting to note how little energy is lost by the indirect transformations.
Conclusion
This approach to the analysis of the metabolic efficiency in tsetse succeeds in obtaining the all desired estimates, in improving on the calorimetric approach, and in obtaining estimates which appear reasonable when compared to experimental results. The stoichiometric approach was able to derive yield estimates for each of the major metabolic pathways in tsetse, as described by figure 1, and to include estimates of the uric acid which must be excreted during each of these conversions, of the carbon dioxide released by the conversion, and of the oxygen required. These estimates, presented in table 7, reach a level of accuracy beyond what is possible using a calorimetric approach because they include intrinsically the inefficiencies and costs of biochemical conversion. The estimates derived in this paper, despite the crudeness of the approach and the omission of the biological costs, appear reasonable when compared to laboratory measurements. These estimates therefore provide useful working numbers for future research.
The estimates presented in table 7 can be used to assess the trade-offs made in tsetse physiology throughout the lifecycle between the generation of each of the principal metabolic products. For instance, it becomes possible to analyze the trade-off made by tsetse between the allocation of nutritional resources to the production of metabolic energy as against to the growth of the different components of biomass, muscle or fat. The estimates open up a new, quantitative strategy for reexamining the extensive analysis of digestive physiology in tsetse [20,21,53-56]. The estimates also provide a way to relate the extensive experimental evidence of respiration rates to the experiments of resource use and of changes in body mass. For example, the average respiration rate of mature male tsetse in the field has been measured based on the rate of loss of radioactive caesium isotopes [57,58]; those estimates can now be directly converted into estimates of blood mass use and metabolic energy consumption. The ultimate value of these estimates will become apparent once they are used to assemble a dynamic mass-energy budget which describes the time varying flows of biomass and energy in tsetse populations.
The biochemical approach to the estimation of conversion yields extends the work of ecologists interested in trophic interactions. The approach provides useful estimates which, despite the crude approach of the current study, can readily be refined using a similar analytic strategy but deriving the data from a systematic, experimental approach and using more sophisticated informatic tools. Nutritional ecologists working on digestive metabolism, like their counterparts working on cell metabolism, will be forced to use a biochemical approach in order to distinguish the purely biochemical costs of digestion from the biological costs. Ecologists developing organism specific dynamic mass-energy budgets will benefit from the detail provided by the biochemical approach. Population modelers working on physiologically based population models can use similar biochemical analyses to extend their mechanistic explanations to a more fundamental level and can chain models in multitrophic systems based on the actual currency, mass and energy, which drives ecological systems.
The results presented in this paper have been obtained solely for the purpose of developing an extensive analysis of tsetse and the disease which these flies transmit. The numbers provide the means to develop a dynamic mass-energy budget for tsetse through which to examine their feeding rate. The numbers also provide a quantitative basis from which to estimate the intrinsic rate of growth, in biomass terms, of tsetse populations. Jointly these provide a way to study of the rate of feeding of tsetse populations, which is the critical parameter determining the rate of transmission of trypanosomes and therefore the epidemiology of trypanosomiasis.
Methods
The estimates presented in this paper were calculated by balancing the stoichiometry of all the biochemical reactions involved in the major metabolic pathways in tsetse based on published descriptions of the composition of the reactants, of the pathways used by tsetse metabolism, and of the specific reactions involved. These calculations required several simplifying procedures. Where multiple substrates could be used for the formation of a given end product, the substrate which needed the fewest biochemical steps to produce the desired end product is used. For instance, any amino acid can be used as a precursor to uric acid but here glycine (Gly), serine (Ser), cysteine (Cys) and alanine (Ala) are assigned to the conversion since they require the fewest biochemical steps. Where alternative intermediate pathways are possible, the most energetically efficient path which could be found was chosen. In all pathways, the costs of reactivation of the enzymes and catalysts are explicitly included in the resulting stoichiometric calculations.
A number of stoichiometric equivalences are used in this work. Guanosine triphosphate (GTP) is assumed energetically equivalent to adenosine triphosphate (ATP). The conversion of ATP to adenosine monophosphate (AMP) was assumed equivalent to two conversions of ATP to adenosine diphosphate (ADP). Similarly, reduced nicotinamide-adenine dinucleotide phosphate (NADPH) is assumed equivalent to reduced nicotinamide-adenine dinucleotide (NADH). Each of these is assumed to be energetically equivalent to 2.5 cystolic ATP whereas each reduced flavin-adenine dinucleotide (FADH2) is assumed equivalent to 1.5 cystolic ATP based on Michal [12](see page 44).
All calculations were performed using the free Gnumeric spreadsheet [59] and solved through iterative calculation. Since the proportions of conversion are independent of the actual quantity of substrate converted, each calculation was started, for convenience, with 100 moles of bloodmeal derived amino acids. The quantities of amino acids entering each metabolic pathway were assigned in a stepwise fashion, first to account for the excretion of the excess nitrogen through uric acid formation, next to form the ATP required, and finally to create the desired end products. Where the pathways required proportional formation of several end products, initial estimates for each path were changed iteratively to bring the end products to the correct proportions with the spreadsheet recalculating the metabolic consequences of these changes. The approach used here does not necessarily obtain the optimal solution since the biochemical pathways form a complex graph but this approach serves as a first approximation to the actual net proportion of conversion of substrates to products.
List of abbreviations
AA Amino Acid.
ADP Adenosine diphosphate.
AMP Adenosine monophosphate.
ATP Adenosine triphosphate, the primary carrier of metabolic energy.
CO2 Carbon dioxide, also written CO2.
-CoA CoenzymeA, an enzyme that joins with metabolites in several biochemical reactions, for instance, acetyl-CoA is the complex of an acetyl radical and this enzyme.
FADH2 Flavin-adenine dinucleotide, a molecule used to transfer energy between the Krebs cycle and the ATP generating mitochondria.
Fat The fraction of body mass extracted by alcohol after desiccation.
GTP Guanosine triphosphate, an ATP equivalent carrier of matabolic energy.
mol Mole, Avogadro's number or 6.023 × 1023.
NADH Nicotinamide-adenine dinucleotide, a molecule used to transfer energy between the Krebs cycle and the ATP generating mitochondria.
NADPH Nicotinamide-adenine dinucleotide phospate, a molecule used to transfer energy between the Krebs cycle and the ATP generating mitochondria.
O2 Molecular Oxygen, also written O2.
RDW Residual Dry Weight, the remaining mass after desiccation and extraction of the alcohol soluble fraction. Here 'weight' is used as a synonym of mass, despite the slight inaccuracy.
RNA Ribonucleic acid.
Amino acids
Ala Alanine
Arg Arginine
Asp Asparagine or Aspartic Acid
Cys Cysteine
Glu Glutamine or Glutamic Acid
Gly Glycine
His Histidine
Leu Leucine or Isoleucine
Lyc Lysine
Met Methionine
Phe Phenylalanine
Pro Proline
Ser Serine
Thr Threonine
Try Tryptophan
Tyr Tyrosine
Val Valine
Authors' contributions
A. Custer performed all the research and analysis presented in the article and wrote the text.
Supplementary Material
Additional File 1
The Spreadsheet used for Estimation - Gnumeric Format. This is the spreadsheet used to make the estimates in the text. This version of the file is in the file format used by the Gnumeric spreadsheet [59] which is simply a compressed (GNU zip) XML file. To view this file, download the Gnumeric software. Click here for download:
Click here for file
Additional File 2
The Spreadsheet used for Estimation - Excel Format. This is the spreadsheet used to make the estimates in the text. This version of the file is in the Excel (TM) 97/2000/XP file format. Since this is a binary, unpublished format, Excel may complain on opening the file but the file should work.
Click here for file
Acknowledgements
The advice and encouragement provided by Professors Andrew P. Gutierrez, Nicholas J. Mills, and Gregory S. Biging is greatly appreciated. Three anonymous reviewers provided helpful suggestions for the improvement of the manuscript.
==== Refs
Kooijman S Dynamic Energy and Mass Budgets in Biological Systems 2000 2 Cambridge University Press, Cambridge, UK
Nisbet R Muller EB Lika K Kooijman S From molecules to ecosystems through dynamic energy budget models Journal of Animal Ecology 2000 69 913 926 10.1046/j.1365-2656.2000.00448.x
Kooijman S Quantitative aspects of metabolic organisation: a conceptual introduction Philosophical Transactions of the Royal Society of London (B) 2001 356 331 349 10.1098/rstb.2000.0771
Gutierrez A Baumgärtner J Multitrophic level models of predator-prey energetics: I. age specific energetics models -- pea aphid Acyrthosiphon pisum (harris) (homoptera: Aphididae) as an example Candian Entomologist 1984 116 924 932
Gutierrez A Applied Population Ecology: A Supply-Demand Approach 1996 John Wiley & Sons, Inc., New York, USA
de Roos A Persson L Physiologically structured models from versatile technique to ecological theory Oikos 2001 94 51 71 10.1034/j.1600-0706.2001.11313.x
Habtemariam T Ruppanner R Riemann H Theis J An epidemiologic systems analysis model for African trypanosomiasis Preventative Veterinary Medicine 1983 1 125 136 10.1016/0167-5877(83)90017-X
Gouteux J Artzrouni M Faut-il ou non un contrôle des vecteurs dans la lutte contre la maladie du sommeil? Bulletin de la Société de Pathologie Exotique 1996 89 299 305
Hershey D Digging deeper into helmont's famous willow tree experiment American Biology Teacher 1991 53 458 460
Sterner R Elser J Ecological Stoichiometry: The biology of elements and molecules to the biosphere 2002 Princeton University Press, Princeton, New Jersey, USA
Bursell E Taylor P An energy budget for Glossina (Diptera: Glossinidae) Bulletin of Entomological Research 1980 70 187 196
Michal G Biochemical Pathways: An Atlas of Biochemistry and Molecular Biology 1999 John Wiley & Sons, New York, NY, USA
Secor S Diamond J Effects of meal size on postprandial responses in juvenile burmese pythons (python molurus) American Journal of Physiology: Regulatory, Integrative and Comparative Physiology 1997 272 R902 912
Starck J Moser P Werner R Linke P Pythons metabolize prey to fuel the response to feeding Proccedings of the Royal Society of London (B) 2004 271 903 908 10.1098/rspb.2004.2681
Andersen J Rourke B Caiozzo V Bennett A Hicks J Physiology: Postprandial cardiac hypertrophy in pythons Nature 2005 434 37 38 15744290 10.1038/434037a
Fischer E Sauer U Metabolic flux profiling of escherichia coli mutants in central carbon metabolism using gc-ms European Journal of Biochemistry 2003 270 880 891 12603321 10.1046/j.1432-1033.2003.03448.x
Fuhrer T Fischer E Sauer U Experimental identification and quantification of glucose metabolism in seven bacterial species Journal of Bacteriology 2005 187 1581 1590 15716428 10.1128/JB.187.5.1581-1590.2005
Edwards J Palsson B Metabolic flux balance analysis and the in silico analysis of escherichia coli k-12 gene deletions BMC Bioinformatics 2000 1 1 11001586 10.1186/1471-2105-1-1
Gagneur J Klamt S Computation of elementary modes: a unifying framework and the new binary approach BMC Bioinformatics 2004 5 175 15527509 10.1186/1471-2105-5-175
Hargrove J Packer M Nutritional states of male tsetse flies (Glossina spp.) (Diptera: Glosssinidae) caught in odour-baited traps and artificial refuges: models for feeding and digestion Bulletin of Entomological Research 1993 83 29 46
Loder P Hargrove J Randolph S A model for blood meal digestion and fat metabolism in male tsetse flies (Glossinidae) Physiological Entomology 1998 23 43 52 10.1046/j.1365-3032.1998.2310043.x
Bursell E Nitrogenous waste products of the tsetse fly, Glossina morsitans Journal of Insect Physiology 1965 11 993 1001 5827549 10.1016/0022-1910(65)90202-7
Lehane M Biology of Blood-sucking insects 1991 HarperCollins Academic, London, UK
Bursell E Billing K Hargrove J McCabe C Slack E Metabolism of the bloodmeal in tsetse flies (a review) Acta Tropica 1974 31 297 320 4155596
Leak S Tsetse Biology and Ecology: Their Role in the Epidemiology and Control of Trypanosomiasis 1999 CABI Publishing, New York, NY, USA
Buxton P Lewis D Climate and tsetse flies: Laboratory studies upon Glossina submorsitans and tachinoides Philosophical Transactions of the Royal Entomological Society London Series (B) 1934 224 175 240
Bursell E The water balance of tsetse pupae Philosophical Transactions of the Royal Society of London (B) 1958 241 179 210
Ramos-Elorduy J Moreno J Prado E Perez M Otero J de Guevara O Nutritional value of edible insects from the state of oaxaca, mexico Journal of Food Composition and Analysis 1997 10 142 157 10.1006/jfca.1997.0530
Cmelik S Bursell E Slack E Composition of the gut contents of third-instar tsetse larvae (Glossina morsitans Westwood) Comparative Biochemistry and Physiology 1969 29 447 453 10.1016/0010-406X(69)91764-2
Langley P Pimley R Influence of diet on synthesis and utilization of lipids for reproduction by the tsetse fly Glossina morsitans Journal of Insect Physiology 1979 25 79 85 10.1016/0022-1910(79)90040-4
Langley P Pimley R Storage and mobilization of nutriment for uterine milk synthesis by Glossina morsitans Journal of Insect Physiology 1979 25 193 197 10.1016/0022-1910(79)90099-4
Bursell E Aspects of the metabolism of amino acids in the tsetse fly, Glossina (Diperta) Journal of Insect Physiology 1963 9 439 452 10.1016/0022-1910(63)90054-4
McCabe C Bursell E Interrelationships between amino acid and lipid metabolism in the tsetse fly, Glossina morsitans Insect Biochemistry 1975 5 781 789 10.1016/0020-1790(75)90022-0
Bursell E Slack E Oxidation of proline by sacrosomes of the tsetse fly, Glossina morsitans Insect Biochemistry 1976 6 159 167 10.1016/0020-1790(76)90026-3
Bursell E Synthesis of proline by the fat body of tsetse fly (Glossina morsitans): metabolic pathways Insect Biochemistry 1977 7 427 434 10.1016/S0020-1790(77)90068-3
Bursell E Quantitative aspects of proline utilization during flight in tsetse flies Physiological Entomology 1978 3 265 272
Bursell E Downer R The role of proline in energy metabolism Energy Metabolism in Insects 1981 5 Plenum Press 135 154
Agosin M Rockstein M Functional role of proteins Biochemistry of Insects 1978 5 Academic Press, New York, USA 93 144
Cmelik S Hurrell D Lunat M Lipid content and composition of the tse-tse fly, Glossina morsitans Westwood Comparative Biochemistry and Physiology 1969 31 65 78 4310397 10.1016/0010-406X(69)92169-0
Bursell E The excretion of nitrogen in insects Advances in Insect Physiology 1967 4 33 67
McCabe C Bursell E Metabolism of digestive products in the tsetse fly, Glossina morsitans Insect Biochemistry 1975 5 769 779 10.1016/0020-1790(75)90021-9
Nation J Insect Physiology and Biochemistry 2002 CRC Press, Boca Raton, Florida, USA
Stevens L Avian Biochemistry and Molecular Biology 1996 Cambridge University Press, Cambridge, UK
Taylor P Blood-meal size of Glossina morsitans Westw. and G. pallidipes Austen (Diptera: Glossinidae) under field conditions Transactions of the Rhodesian Scientific Association 1976 57 29 34
Bursell E Post-teneral development of the thoracic musculature in tsetse flies Proccedings of the Royal Entomological Society of London (A) 1961 36 69 74
Rogers D Randolph S Metabolic strategies of male and female tsetse (Diptera: Glossinidae) in the field Bulletin of Entomological Research 1978 68 639 654
Mathews CK van Holde KE Biochemistry 1990 Benjamin/Cummings, Redwood City, CA, USA
Rajagopal P Bursell E The respiratory metabolism of resting tsetse flies Journal of Insect Physiology 1966 12 287 297 10.1016/0022-1910(66)90144-2
Rajagopal P Bursell E The effect of temperature on the oxygen consumption of tsetse pupae Bulletin of Entomological Research 1965 56 219 225 5854752
Bursell E The effect of temperature on the consumption of fat during pupal development in glossina Bulletin of Entomological Research 1960 51 583 598
Alberty R Calculation of biochemical net reactions and pathways by using matrix operations Biophysical Journal 1996 71 507 515 8804633
Shastri A Morgan J Calculation of theoretical yields in metabolic networks Biochemistry and Molecular Biology Education 2004 32 314 318
Langley P Wall R The implication of hunger in the tsetse fly, Glossina pallidipes, in relation to its availability to trapping techniques Journal of Insect Physiology 1990 36 903 908 10.1016/0022-1910(90)90077-S
Randolph S Rogers D Kiilu J The feeding behaviour, activity and trappability of wild female G. pallidipes in relation to the pregnancy cycle Medical and Veterinary Entomology 1991 5 335 350 1768925
Randolph S Williams B Rogers D Connor H Modelling the effect of feeding-related mortality on the feeding strategy of tsetse (Diptera: Glossinidae) Medical and Veterinary Entomology 1992 6 231 240 1421503
Hargrove J Lifetime changes in the nutritional characteristics of female tsetse Glossina pallidipes caught in odour-baited traps Medical and Veterinary Entomology 1999 13 165 176 10484162 10.1046/j.1365-2915.1999.00153.x
Taylor P Radioisotopes as metabolic labels for Glossina (Diptera: Glossinidae). II. The excretion of 137Cs under field conditions as a means of estimating energy utilization, activity and temperature regulation Bulletin of Entomological Research 1978 68 331 340
Hargrove J Coates T Metabolic rates of tsetse flies in the field as measured by the excretion of injected caesium Physiological Entomology 1990 15 157 166
GNOME The Gnumeric Spreadsheet 2004
|
16083496
|
PMC1215481
|
CC BY
|
2021-01-04 16:29:15
|
no
|
BMC Ecol. 2005 Aug 5; 5:6
|
utf-8
|
BMC Ecol
| 2,005 |
10.1186/1472-6785-5-6
|
oa_comm
|
==== Front
BMC Fam PractBMC Family Practice1471-2296BioMed Central London 1471-2296-6-361612487410.1186/1471-2296-6-36Research ArticleSigns and symptoms in children with a serious infection: a qualitative study Van den Bruel Ann [email protected] Rudi [email protected] Etienne [email protected] Peter [email protected] Bert [email protected] Frank [email protected] Department of General Practice, Katholieke Universiteit Leuven, Kapucijnenvoer 33 Blok J, 3000 Leuven, Belgium2 Department of General Practice, Universtaire Instelling Antwerpen, Universiteitsplein 1, 2610 Wilrijk, Belgium3 Department of Pediatrics, Virga Jesseziekenhuis, Stadsomvaart 11, 3500 Hasselt, Belgium4 Department of General Practice, Universiteit Maastricht, Postbus 616, 6200 MD Maastricht, The Netherlands2005 26 8 2005 6 36 36 14 12 2004 26 8 2005 Copyright © 2005 Van den Bruel et al; licensee BioMed Central Ltd.2005Van den Bruel et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Early diagnosis of serious infections in children is difficult in general practice, as incidence is low, patients present themselves at an early stage of the disease and diagnostic tools are limited to signs and symptoms from observation, clinical history and physical examination. Little is known which signs and symptoms are important in general practice. With this qualitative study, we aimed to identify possible new important diagnostic variables.
Methods
Semi-structured interviews with parents and physicians of children with a serious infection. We investigated all signs and symptoms that were related to or preceded the diagnosis. The analysis was done according to the grounded theory approach. Participants were recruited in general practice and at the hospital.
Results
18 children who were hospitalised because of a serious infection were included. On average, parents and paediatricians were interviewed 3 days after admittance of the child to hospital, general practitioners between 5 and 8 days after the initial contact.
The most prominent diagnostic signs in seriously ill children were changed behaviour, crying characteristics and the parents' opinion. Children either behaved drowsy or irritable and cried differently, either moaning or an inconsolable, loud crying. The parents found this illness different from previous illnesses, because of the seriousness or duration of the symptoms, or the occurrence of a critical incident. Classical signs, like high fever, petechiae or abnormalities at auscultation were helpful for the diagnosis when they were present, but not helpful when they were absent.
Conclusion
behavioural signs and symptoms were very prominent in children with a serious infection. They will be further assessed for diagnostic accuracy in a subsequent, quantitative diagnostic study.
==== Body
Background
In general practice, the incidence of acute infections is high. The yearly incidence can be as high as 41%, with acute upper respiratory infections as the most frequent diagnosis and highest incidence rates in children less than 1 year old[1].
In contrast, serious and possibly life-threatening infections are rare, their yearly incidence in children being estimated at 1.5%[2]. The most frequent diagnoses are pneumonia, sepsis, meningitis, pyelonephritis and bacterial gastro-enteritis [3-6]. Other infections, such as osteomyelitis, cellulitis and septic arthritis are even less frequent.
This low incidence and the similarities in the initial presentation make it difficult to distinguish these children from their peers with a non-serious, mostly self-limiting infection.
Still, early diagnosis of a serious infection is important to avoid delay in treatment and improve prognosis [7-11].
Signs and symptoms are the first and most readily available diagnostic tools for the general practitioner. They are the basis of subsequent decisions, such as referral, additional testing or prompt treatment.
However, little is understood about these signs and symptoms in general practice. Prior diagnostic accuracy studies on serious infections in children were predominantly conducted in hospital populations, in very young children or included laboratory or radiology tests results [12-15]. For example, the Yale Observation Scale uses observational information, but was constructed and validated in a hospital setting[16]. The same applies for the Young Infant Observation Scale, which also uses observational information, but which is designed for hospitalised children aged less than 8 weeks old[17]. None of the above was evaluated in general practice, to our knowledge only a few studies were performed in general practice on this subject [18-20]. These studies do not provide quantitative measures of accuracy and cover only meningitis, instead of the entire group of serious infections. Other studies about signs and symptoms of children with a serious infection in primary care were carried out in developing countries in a population that is not comparable to that in Western Europe [21-24].
As the value of a diagnostic test depends on the setting in which it is being used[25,26], the results of these studies can not simply be transferred to general practice and therefore new research is needed.
A number of signs and symptoms have been described in the past as being related to the diagnosis of a serious infection in children. All available information, however, indicates that they are insufficient to reliably diagnose or exclude a serious infection[27,28]. We therefore conducted this qualitative study to generate hypotheses, as part of a large diagnostic study about signs and symptoms of serious infections in children, seen in general practice. With this qualitative study, we intended to identify new signs and symptoms, additional to those found in textbooks or main articles[29] and promising for use in general practice. It is our intention to quantitatively estimate their diagnostic characteristics in a subsequent prospective study.
Methods
We selected a theoretical sample[30] from all children that were admitted with a serious infection to one large regional hospital in the east of Flanders, Belgium. The sample was intended to consist of four children with one of the following infections: pneumonia, sepsis, meningitis, pyelonephritis and complicated gastro-enteritis. Of the latter, two children with a complicated viral gastro-enteritis and two with a bacterial gastro-enteritis were included. Within each subgroup, the children were included consecutively.
This purposive sampling was performed to ensure a wide range of possible signs and symptoms, different perspectives from parents and physicians from both general practice and specialist care[31]. This sampling procedure is characteristic for qualitative studies and should not be mistaken with the classical sampling procedures for a diagnostic accuracy study.
Design
The study was performed between March and October 2003.
The children's parents were invited for a semi-structured interview. We also interviewed the paediatrician who had admitted the child and the general practitioner if he or she had seen the child before admittance. The interviews were carried out as soon as possible after admittance to the hospital or after consultation at the general practice, in order to minimise recall bias.
Our main point of interest was the identification of signs and symptoms that were present in all cases or in all cases of one of the diagnostic categories.
Informed consent was signed by all parents. The study was approved by the Ethical Review Board of the Katholieke Universiteit Leuven.
The interviews were semi-structured[32]: a framework for the interview was set out, covering different areas of possible signs and symptoms. Besides this framework, open ended questions about the history of the disease or any other information the interviewee wanted to communicate, were included in order to retrieve possible new items. Before the start of the pilot study, both the content of the interview and the phrasing of the questions were reviewed by experts in qualitative research and in paediatrics. Changes were made according to their remarks. The interview was piloted twice by the principal investigator with two GPs who had recently seen a child with a serious infection. After these interviews, the GPs were asked to evaluate the content of the questions, the method of performing the interview and the emotions that it possibly evoked. No more substantial changes were needed at this stage.
The interviews were performed by three experienced interviewers and the principal investigator. The interviewers were instructed about the goal of the study and trained during a practice session. All interviews were recorded on tape and transcribed before starting the analysis. During the interview, the atmosphere, non-verbal communication and the possible effect of other people that assisted were described by the interviewers. The framework of the interview is illustrated in table 1.
Table 1 framework of the interviews
1. Opening of the interview: the purpose of the study was repeated and informed consent specifically asked again.
2. Start of the interview with questions on the child's name and the relation of the interviewee to the child.
3. Open questions on the illness episode. Interviewees were asked to tell in their own words what had happened.
4. More directed questions on the start of the illness, the evolution of the illness and possible symptoms.
5. Open questions on the decisions and actions of the interviewee taken during the course of the illness, including the reasons for these decisions.
6. Final open question: whether the interviewee had anything to add to the interview.
Analysis was performed according to the pragmatic variant of the grounded theory approach[33], by which new themes are identified from the data alongside those that were already anticipated from the outset[34]. Every interview was independently analysed by two investigators. The principal investigator analysed all interviews and two other investigators each analysed half of the interviews. Individual codes were assigned to the text by each investigator. The process was iterative, as sampling and data collection were guided by the emerging analysis. During a consensus session, all codes were compared and disagreements were reviewed with the data at hand. This resulted in a set of themes that were striking and seemed important diagnostic features, either across different diagnostic subgroups or in one diagnostic subgroup only. These themes were than translated in hypotheses for subsequent quantitative tests.
Results
In total, 22 children with a serious infection were eligible for the study. Three children were excluded: 1 because a second reading of the X-ray refuted the diagnosis of pneumonia and 2 because the delay between admittance and the first contact with the researcher exceeded 5 days. All parents gave their consent, as well as all paediatricians. In 9 cases, the children were seen by their general practitioner before admittance to the hospital. All GPs agreed to participate. One child was excluded after the interview, because the audiotape was unintelligible. Interviews with the parents and paediatricians of 18 children and with the GPs of 9 children with a serious infection were thus available for analysis. We could include only 2 cases of meningitis within the study period instead of the expected 4 cases. But as no new items were identified in the last two interviews, the data collection was ended.
Demographics
The mean age of the children was 2.5 years (ranging from 14 days to 11 years) and 9 of the 18 children were girls.
As mentioned above, we included 4 children from each diagnostic category, except for meningitis, from which only 2 cases were included.
Average time between admittance to the hospital and the interview with the parents was 2.8 days (range from 1 to 8 days) and to the interviews with the paediatricians was 3.4 days (range from 1 to 8 days). General practitioners were interviewed 7.8 days (range from 1 to 14 days) after seeing the child. The interviewers took notes on the atmosphere of the interview, and of the people present. In most cases the atmosphere was relaxed, especially with the parents; physicians sometimes seemed hurried. The mother of the child was present in the majority of the interviews with the parents. Any non-verbal signs important for the context of the interviews were noted as well. These indicated that most interviewees were motivated and at ease, although one paediatric resident was sceptical on the purpose of the study on three occasions.
Behaviour
Behavioural changes were almost systematically mentioned. They were twofold: on one hand some children were very weak or drowsy; on the other hand there were children who were irritable. Changes in behaviour were mentioned by both parents and physicians, but very meticulously illustrated by the parents as they compared it to the normal behaviour of the child.
Drowsiness, as described by parents
'It seemed as if there was no life in him anymore': this quote was often repeated by different parents, in almost exact the same words. Their children were unusually quiet, did not play nor talk. Some children could not get up, and had to be fed lying in bed. Several children were lying in bed or on the sofa, with their eyes closed without actually sleeping. Many children did not laugh anymore, even when they were being played with.
They were all not in their usual self, on several occasions phrased as 'This was no longer my child.'
[Parent, child with meningococcal sepsis, 1 year old] "It was really different from before. In the morning she was lying in her playpen and she stayed so still, she really wasn't her normal self."
[Parent, child with pneumonia, 5 years old] "He was lying there with his eyes closed, but when I asked him something, he answered me, so he wasn't sleeping, but he felt so miserable, I guess. It was frightening, because normally he is such a lively child, and now he couldn't even watch television."
[Parent, child with sepsis, 8 months old] "He was very passive and normally he is a very cheerful baby, he is always laughing and always content. Now he was not laughing anymore and that's why I knew there was something wrong."
Drowsiness, as described by physicians
Physicians also found it abnormal when a child did not react to strangers, and especially when it was manipulated for a clinical examination or a blood sample. Sometimes, general physicians noticed that the child was not behaving like it did on previous occasions, as they have, more than a paediatrician, a long-term relationship with the family.
[General practitioner, child with pneumonia, 2 years old] "She was so drowsy. Normally she is a very lively child."
[General practitioner, child with meningococcal sepsis, 1 year old] "I went there, and when I arrived I saw immediately that it was serious. The child was so passive!"
[General practitioner, child with pneumonia, 2 years old] "For a two year old, she let herself be examined too easily. Normally we expect some more resistance."
[Paediatrician, child with viral gastro-enteritis with dehydration, 2 years old] "Almost no reactions. While taking the blood sample he did not react either."
Irritability
Some children behaved more irritable, for example they started crying and could not be consoled, slept less than usual or woke up during the night. Other children were irritable when they were examined by a physician.
[Parent, child of 14 days old with pyelonephritis] "He almost did not sleep, very shortly in the afternoon, one hour or so. Normally he sleeps more than four hours; but now he woke after one hour and he did not sleep after since."
[Parent, child with sepsis, 1 year old] "He wanted to go to bed more often, but he could not sleep. And normally, when we put him between us he becomes very quiet, but that did not help this time. Taking him up did not help either, and when he is crying with his dummy in his mouth, then it is not right."
[Paediatrician, child with pyelonephritis, 14 days old] "He was irritable when he was touched."
Crying characteristics
The cry was often found to be different than before, as expressed by their parents, or to be striking, as expressed by the physicians. The way this crying was different, was parallel to the other behavioural changes with children who were drowsy and other who were irritable.
Some children were crying in a nagging, quiet way, with less force than usual. Other children were crying louder than usual, could not be consoled or had a pinching cry.
[General practitioner, child with pneumonia, 2 years old] "Not really crying, but moaning. And she had been doing that at home as well, according to the grandmother, for the last two days."
[Parent, child with meningitis, 2 months old] "Wednesday at midnight, he suddenly started shrieking. Normally he cries a little and then you take him, but now he was really shrieking. He wouldn't be quiet with a bottle or anything."
[Paediatrician, child with sepsis, 8 months old] "The pinching cry was the only thing that struck me. He didn't seem very sick."
Abnormal illness
Apart from behavioural changes, parents found the illness itself different from previous illnesses. Symptoms could be different than usual or the duration of the illness was longer.
[Child, with bacterial gastro-enteritis, 11 years old] "The diarrhoea, I have never had it so seriously. I really panicked."
[Parent, with bacterial gastro-enteritis, 4 years old] "His weight was 17 kg and when I weighed him that morning it was only 15.3 kg. I found he had lost too much weight in 2 days."
[Parent, child with pneumonia, 5 years old] "It went on and on, I thought this can not be right."
[Parent, child with pyelonephritis, 2 years old] "She did not eat anything and she could not hold down her drinks; then you know it is not right."
[Parent, child with pneumonia, 5 years old] "Three days of fever is perfectly normal, even five days; but after these five days I thought it was abnormal."
On several occasions, there was an incident that had never happened before and made the parents anxious or made them decide to seek help immediately. These incidents were variable, a few children suddenly became very pale and limp, and another vomited in her sleep without waking up. One child with viral meningitis became unconscious at school.
[Parent, child with sepsis, 2 years old] "The moment he was sitting on my lap and suddenly collapsed, I was really frightened and came here immediately. At that moment I just knew it was more than just a cold."
[Parent, child with meningitis, 2 months old] "I did not trust it because he wasn't himself. But then he suddenly became very pale after I had cooled him, and it was really strange."
Parents' opinion
Several physicians said that they were guided by the opinion of the parents in their judgement. In addition to other signs, the parents' anxiety or statement that it was an abnormal situation made them cautious. In fact, when we asked them which was the most striking sign or symptom, several physicians answered that it was the parents' opinion.
[General practitioner, child with meningococcal sepsis, 1 year old] "The mother rang back very quickly after a few hours, which is very odd for her. She is very capable of handling a situation like that; normally she doesn't mind a child having a fever for two or three days."
[Paediatrician, child with meningitis, 2 months old] "Especially the mother saying that she did not know the child like that."
Classical signs
Almost all children with a serious infection had a high fever of over 39°C (102 degrees Fahrenheit), except one infant of two weeks old with a pyelonephritis and one toddler with a viral meningitis. The latter two had a body temperature of over 38°C (100.4 degrees Fahrenheit).
Other classical signs as described in textbooks were present in some cases but not in all. For example, neck stiffness, petechiae, crepitations or signs of dehydration, were present in some children and then led towards the diagnosis, but not all children presented with any of these suspected signs. The absence of a sign did not necessarily mean the absence of disease and some classical signs would not appear in all children with a certain disease, as was shown by the management decision of the treating physicians in our sample.
This was especially the case for signs from auscultation and the diagnosis of pneumonia, which were not present in any child with pneumonia. Only one child with pneumonia had a cough at the moment of admittance. Two children with pneumonia started coughing later during the hospital stay.
Other signs such as vomiting or the absence of signs of upper respiratory tract infection could be observed in several of the children with a serious infection. However, in some of the very sick children no classical signs could be found, which was disturbing the physicians on itself.
[General practitioner, child with pneumonia, 2 years old] "Nothing specific, I did not find anything and that child was really sick. I could not find anything, except a sensitive tummy, and then you think: this can not be right."
Disease specific
All children with pneumonia had some sign concerning the breathing. Some children had a higher breathing rhythm, some had superficial breathing, only one child with pneumonia had a dry cough.
Of the children with pyelonephritis, two started wetting their pants again, after they had been toilet-trained long before. The other two children with pyelonephritis in the sample were still in nappies, so it was difficult to observe urination signs.
Loss of weight was prominent in children with gastro-enteritis, but this is to be expected, as it is a criterion for admittance to the hospital.
The children with sepsis all had petechiae but one. In one child, the petechiae were rather large and expanding, in the other two, there were only a few spots that could easily have been missed. In these last two cases, none of the parents had noticed the spots; they were only seen after careful inspection at the emergency department.
For the diagnosis of meningitis, only two cases could be included, both of viral aetiology. The first child was a baby of two months old, who cried in a different way than before and was difficult to console. The second child was a toddler who fell asleep at school and could not be awakened unless with pain stimuli. None of these children had signs of meningeal irritation.
Discussion
Serious infections in children are an important topic in primary care, because of the related mortality and morbidity [35-38]. This was reflected in the high response ratios to our invitation to participate in the study, as all but one parent and one doctor agreed to participate.
The diagnosis of these infections can be difficult, especially in primary care, where the disease is still in an early stage and incidence is low. This, together with the relative inaccessibility of more invasive diagnostic procedures, makes it a difficult challenge for the general practitioner and cases can be missed[39]. In a qualitative study about the diagnosis and management of children suspected with meningitis[40], general practitioners found it difficult to reach a diagnosis and stated they relied upon intuitive rather than systematic methods. In order to be transmissible to younger generations, however, intuition has to be translated in evidence.
In our study, time between admission or consultation and interview was very short (between 2.8 and 7.8 days), especially when this is compared to another qualitative study about the diagnosis of meningitis in primary care, where the mean interval between case and interview was 61 weeks[41]. Even so, in our study, paediatricians and GPs frequently consulted their file, which indicates that although physicians were willing to answer the questions accurately they had difficulties in remembering certain details.
The interviews were mostly taken in a relaxed atmosphere, although GPs and paediatricians often had limited time available. Parents were very motivated for the study in both patient groups. One paediatric resident was sceptical about the study at the beginning, but became more convinced at a later stage. The short time frame, the relaxed and open atmosphere and the motivation and interest for the study strengthen the validity of our study results. Our study is limited by the fact that, although the primary aim was to explore the clinical presentation of children with a serious infection in general practice, we were able to interview only 9 general practitioners. It may be reasonable to assume, however, that the information given by the parents is useful in general practice as well, as physicians should be very sensitive to the information given by the parents during history taking. The information given by parents can be influenced by their educational status, marital status, number of children etcetera. Unfortunately, we do not have any data to check this in our sample.
Our clinical findings are partly in concordance with findings from other studies, but some are different.
Parents can describe their child's behaviour very accurately and can compare this to the normal behaviour of the child and to previous illnesses. The level of detail in which these descriptions were made, was very striking.
Some children were drowsy and weak, more than they were during previous illnesses. Children cried in a different way than they normally did, moaning, nagging, a cry without force. Other children were irritable, cried louder than usual and could not be consoled. A pinching cry was noticed by physicians as well.
Observation variables have been described before [42-44], but these were mainly variables from the doctor's own observations. Our findings suggest that physicians should be very sensitive to what parents are telling them and add this information to their own observations. For this, general practitioners are in a favourable position, as they have a long-term relationship with their patients and can relate this new information to previous contacts.
Parents found this illness different from previous illnesses. Symptoms were more serious, the duration was longer and sometimes there was a critical incident which caused anxiety or warranted further actions. Some physicians were very sensitive to the opinion of the parents, especially if they knew the family before, and saw a difference in reaction of the parents compared to previous occasions.
Serious infections tend to present with high fever. This has been shown before, for example by Hewson[45], Bleeker[46] and Kuppermann [47-49]. Other studies have found no relation between high temperature and serious infections[50,51]. However, it is possible that fever is a valuable sign in an unselected population such as in general practice, but of less value in a selected population, such as children seen at an emergency department. This certainly needs to be addressed in our subsequent, quantitative diagnostic study.
Secondly, signs can be asymmetric, i.e. their presence has more diagnostic power than their absence or vice versa. This has also been demonstrated before [52-55]. In our study, it seems that some 'classical' signs are very informative when they are present, while the absence of these signs provided almost no information to rule out the suspected disease.
Besides these more generic characteristics, we also found disease-specific characteristics that could be important in the diagnosis of these infections. Signs on breathing pattern, urinary symptoms, rashes were present in most of these cases and certainly should be explored further in a general practice setting. Most of these signs have been proven to be of value before [56-58], but hardly ever in general practice [19,59,17]. The diagnostic accuracy of these signs and symptoms could be addressed in a future study.
Overall, this study did not aim to give 'hard' evidence on tests for the diagnosis of serious infections in children. The reason for performing the study was the lack of evidence in general practice; possible new or different signs had to be explored. The results of this study indicate signs that may be apparent in general practice. This qualitative study provides hypotheses, which can be tested in a quantitative study.
Conclusion
This study has revealed several interesting diagnostic signs about serious infections in children in general practice, especially changed behaviour, crying characteristics, parents' opinion and some classical signs.
These hypotheses can be tested in a prospective, quantitative study to determine their diagnostic accuracy, during which possible asymmetries can be evaluated.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
AVDB contributed to the design and protocol of the study, participated in the data collection, was the primary researcher of the analysis and drafted the manuscript.
RB contributed to the protocol and made a substantial contribution to the analysis. EV provided assistance with the analysis and the manuscript. BA helped with the design of the study and the manuscript. PA helped with the design, the protocol and the data collection. FB conceived the study, set up the design, guided and participated in the analysis and revised the manuscript.
All authors read and approved the manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
We are grateful to all the parents and children who so willingly gave their consent to participate in the study, to the general practitioners and paediatricians who donated some of their precious time and to Prof Eeckels (Katholieke Universiteit Leuven), for his advice and support. The study was financed by unconditional grants of the Fonds Wetenschappelijk Onderzoek Vlaanderen (FWO) and Eurogenerics®.
==== Refs
Fleming DM Smith GE Charlton JR Charlton J Nicoll A Impact of infections on primary care–greater than expected Commun Dis Public Health 2002 5 7 12 12070981
Okkes TM Oskam SK Lamberts H Van klacht naar diagnose Episodegegevens uit de huisartspraktijk 1998 Uitgeverij Coutinho Bussum
Bleeker SE Moll HA Steyerberg EW Donders AR Derksen-Lubsen G Grobbee DE Moons KGM External validation is necessary in prediction research: a clinical example J Clin Epidemiol 2003 56 826 832 14505766 10.1016/S0895-4356(03)00207-5
Baker MD Bell LM Unpredictability of serious bacterial illness in febrile infants from birth to 1 month of age Arch Pediatr Adolesc Med 1999 153 508 511 10323632
Bonadio WA The history and physical assessments of the febrile infant Pediatr Clin North Am 1998 45 65 77 9491087 10.1016/S0031-3955(05)70583-6
Bleeker SE Moons KG Derksen-Lubsen G Grobbee DE Moll HA Predicting serious bacterial infection in young children with fever without apparent source Acta Paediatr 2001 90 1226 1232 11808890 10.1080/080352501317130236
Riordan FA Thomson AP Recognition, treatment and complications of meningococcal disease Paediatr Drugs 1999 1 263 282 10935425
Woodward CM Jessop EG Wale MC Early management of meningococcal disease Commun Dis Rep CDR Rev 1995 5 R135 R137 7670577
Cartwright K Strang J Gossain S Begg N Early treatment of meningococcal disease BMJ 1992 305 774 1422350
Strang JR Pugh EJ Meningococcal infections: reducing the case fatality rate by giving penicillin before admission to hospital BMJ 1992 305 141 143 1515826
Orellana P Baquedano P Rangarajan V Zhao JH Eng ND Fettich J Chaiwatanarat T Sonmezoglu K Kumar D Park YH Samuel AM Sixt R Bhatnagar V Padhy AK Relationship between acute pyelonephritis, renal scarring, and vesicoureteral reflux. Results of a coordinated research project Pediatr Nephrol 2004 19 1122 1126 15258842 10.1007/s00467-004-1501-5
Jaskiewicz JA McCarthy CA Richardson AC White KC Fisher DJ Dagan R Powell KR Febrile infants at low risk for serious bacterial infection–an appraisal of the Rochester criteria and implications for management. Febrile Infant Collaborative Study Group Pediatrics 1994 94 390 396 8065869
Bonadio WA McElroy K Jacoby PL Smith D Relationship of fever magnitude to rate of serious bacterial infections in infants aged 4–8 weeks Clin Pediatr (Phila) 1991 30 478 480 1914347
Klassen TP Rowe PC Selecting diagnostic tests to identify febrile infants less than 3 months of age as being at low risk for serious bacterial infection: a scientific overview J Pediatr 1992 121 671 676 1432412
Isaacman DJ Shults J Gross TK Davis PH Harper M Predictors of bacteremia in febrile children 3 to 36 months of age Pediatrics 2000 106 977 982 11061763 10.1542/peds.106.5.977
McCarthy PL Acute infectious illness in children Compr Ther 1988 14 51 57 3282767
Bonadio WA The history and physical assessments of the febrile infant Pediatr Clin North Am 1998 45 65 77 9491087 10.1016/S0031-3955(05)70583-6
Granier S Owen P Pill R Jacobson L Recognising meningococcal disease in primary care: qualitative study of how general practitioners process clinical and contextual information BMJ 1998 316 276 279 9472513
Koorevaar R Bruijnzeels M van der Wouden J van der Does E van der Velden K van Suijlekom-Smit L Patients with suspected meningitis: a study in general practice Eur J Gen Pract 1995 1 21 24
Brennan CA Somerset M Granier SK Fahey TP Heyderman RS Management of diagnostic uncertainty in children with possible meningitis: a qualitative study Br J Gen Pract 2003 53 626 631 14601339
Singhi S Dhawan A Kataria S Walia BN Validity of clinical signs for the identification of pneumonia in children Ann Trop Paediatr 1994 14 53 58 7516135
Haq SA Alam MN Hossain SM Ahmed T Tahir M Value of clinical features in the diagnosis of enteric fever Bangladesh Med Res Counc Bull 1997 23 42 46 9465434
Weber MW Herman J Jaffar S Usen S Oparaugo A Omosigho C Adegbola RA Greenwood BM Mulholland EK Clinical predictors of bacterial meningitis in infants and young children in The Gambia Trop Med Int Health 2002 7 722 731 12225501 10.1046/j.1365-3156.2002.00926.x
Jamuna R Srinivasan S Harish BN Factors predicting occult bacteremia in young children Indian J Pediatr 2000 67 709 711 11105419
Knottnerus JA Leffers P The influence of referral patterns on the characteristics of diagnostic tests J Clin Epidemiol 1992 45 1143 1154 1474411 10.1016/0895-4356(92)90155-G
Buntinx F Knockaert D Bruyninckx R de Blaey N Aerts M Knottnerus JA Delooz H Chest pain in general practice or in the hospital emergency department: is it the same? Fam Pract 2001 18 586 589 11739341 10.1093/fampra/18.6.586
Kuppermann N Malley R Inkelis SH Fleisher GR Clinical and hematologic features do not reliably identify children with unsuspected meningococcal disease Pediatrics 1999 103 E20 9925866 10.1542/peds.103.2.e20
Klassen TP Rowe PC Selecting diagnostic tests to identify febrile infants less than 3 months of age as being at low risk for serious bacterial infection: a scientific overview J Pediatr 1992 121 671 6 1432412
Pope C Mays N Reaching the parts other methods cannot reach: an introduction to qualitative methods in health and health services research BMJ 1995 311 42 45 7613329
Mays N Pope C Qualitative research in health care. Assessing quality in qualitative research BMJ 2000 320 50 52 10617534 10.1136/bmj.320.7226.50
Greenhalgh T Taylor R Papers that go beyond numbers (qualitative research) BMJ 1997 315 740 743 9314762
Britten N Qualitative interviews in medical research BMJ 1995 311 251 253 7627048
Woods L Priest H Roberts P An overview of three different approaches to the interpretation of qualitative data. Part 2: Practical illustrations Nurse Res 2002 10 43 51 12405005
Barbour RS Checklists for improving rigour in qualitative research: a case of the tail wagging the dog? BMJ 2001 322 1115 1117 11337448 10.1136/bmj.322.7294.1115
Bonadio WA Mannenbach M Krippendorf R Bacterial meningitis in older children Am J Dis Child 1990 144 463 465 2321610
McCarthy PL Dolan TF The serious implications of high fever in infants during their first three months. Six years' experience at Yale-New Haven Hospital Emergency Room Clin Pediatr (Phila) 1976 15 794 796 954335
Olesch CA Knight GJ Invasive meningococcal infection in Western Australia J Paediatr Child Health 1999 35 42 48 10234634 10.1046/j.1440-1754.1999.t01-1-00337.x
Bleeker SE Moll HA [Young child with fever of unknown origin in the 'post Haemophilus influenzae era'] Ned Tijdschr Geneeskd 2002 146 3 5 11802335
Riordan FA Thomson AP Sills JA Hart CA Who spots the spots? Diagnosis and treatment of early meningococcal disease in children BMJ 1996 313 1255 1256 8939123
Brennan CA Somerset M Granier SK Fahey TP Heyderman RS Management of diagnostic uncertainty in children with possible meningitis: a qualitative study Br J Gen Pract 2003 53 626 631 14601339
Granier S Owen P Pill R Jacobson L Recognising meningococcal disease in primary care: qualitative study of how general practitioners process clinical and contextual information BMJ 1998 316 276 279 9472513
Brogan PA Raffles A The management of fever and petechiae: making sense of rash decisions Arch Dis Child 2000 83 506 507 11087287 10.1136/adc.83.6.506
Weber MW Herman J Jaffar S Usen S Oparaugo A Omosigho C Adegbola RA Greenwood BM Mulholland EK Clinical predictors of bacterial meningitis in infants and young children in The Gambia Trop Med Int Health 2002 7 722 731 12225501 10.1046/j.1365-3156.2002.00926.x
Olesch CA Knight GJ Invasive meningococcal infection in Western Australia J Paediatr Child Health 1999 35 42 48 10234634 10.1046/j.1440-1754.1999.t01-1-00337.x
Hewson P Poulakis Z Jarman F Kerr J McMaster D Goodge J Silk G Clinical markers of serious illness in young infants: a multicentre follow-up study J Paediatr Child Health 2000 36 221 225 10849220 10.1046/j.1440-1754.2000.00483.x
Bleeker SE Moons KG Derksen-Lubsen G Grobbee DE Moll HA Predicting serious bacterial infection in young children with fever without apparent source Acta Paediatr 2001 90 1226 1232 11808890 10.1080/080352501317130236
Kuppermann N Fleisher GR Jaffe DM Predictors of occult pneumococcal bacteremia in young febrile children Ann Emerg Med 1998 31 679 687 9624306
Bonadio WA McElroy K Jacoby PL Smith D Relationship of fever magnitude to rate of serious bacterial infections in infants aged 4–8 weeks Clin Pediatr (Phila) 1991 30 478 480 1914347
McCarthy PL Grundy GW Spiesel SZ Dolan TF Jr Bacteremia in children: an outpatient clinical review Pediatrics 1976 57 861 868 934743
Haddon RA Barnett PL Grimwood K Hogg GG Bacteraemia in febrile children presenting to a pae3iatric emergency department Med J Aust 1999 170 475 478 10376023
Kuppermann N Malley R Inkelis SH Fleisher GR Clinical and hematologic features do not reliably identify children with unsuspected meningococcal disease Pediatrics 1999 103 E20 9925866 10.1542/peds.103.2.e20
Riordan FA Thomson AP Sills JA Hart CA Who spots the spots? Diagnosis and treatment of early meningococcal disease in children BMJ 1996 313 1255 1256 8939123
Nielsen HE Andersen EA Andersen J Bottiger B Christiansen KM Daugbjerg P Larsen SO Lind I Nir M Olofsson K Diagnostic assessment of haemorrhagic rash and fever Arch Dis Child 2001 85 160 165 11466193 10.1136/adc.85.2.160
Lucero MG Tupasi TE Gomez ML Beltran GL Crisostomo AU Romano VV Rivera LM Respiratory rate greater than 50 per minute as a clinical indicator of pneumonia in Filipino children with cough Rev Infect Dis 1990 12 S1081 S1083 2270407
Olesch CA Knight GJ Invasive meningococcal infection in Western Australia J Paediatr Child Health 1999 35 42 48 10234634 10.1046/j.1440-1754.1999.t01-1-00337.x
Singhi S Dhawan A Kataria S Walia BN Clinical signs of pneumonia in infants under 2 months Arch Dis Child 1994 70 413 417 8017964
Oostenbrink R Moons KG Donders AR Grobbee DE Moll HA Prediction of bacterial meningitis in children with meningeal signs: reduction of lumbar punctures Acta Paediatr 2001 90 611 617 11440091 10.1080/080352501750258649
Brooks D Houston IB Symptomatic urinary infection in childhood: presentation during a four-year study in general practice and significance and outcome at seven years J R Coll Gen Pract 1977 27 678 683 616858
Granier S Owen P Stott NC Recognizing meningococcal disease: the case for further research in primary care Br J Gen Pract 1998 48 1167 1171 9667096
|
16124874
|
PMC1215482
|
CC BY
|
2021-01-04 16:29:13
|
no
|
BMC Fam Pract. 2005 Aug 26; 6:36
|
utf-8
|
BMC Fam Pract
| 2,005 |
10.1186/1471-2296-6-36
|
oa_comm
|
==== Front
BMC GastroenterolBMC Gastroenterology1471-230XBioMed Central London 1471-230X-5-281614455310.1186/1471-230X-5-28Research ArticleRhinosinusitis derived Staphylococcal enterotoxin B possibly associates with pathogenesis of ulcerative colitis Yang Ping-Chang [email protected] Tao [email protected] Bin-Quan [email protected] Tao-Yuan [email protected] Zi-Yuan [email protected] Peng-Yuan [email protected] Dao-Fa [email protected] Department of Pathology and Molecular Medicine, McMaster University, Hamilton, ON, Canada2 Department of Otolaryngology, Shanxi Medical University, the First Hospital, Taiyuan, China3 Department of Gastroenterology, Shanxi Medical University, the First Hospital, Taiyuan, China4 Department of Gastroenterology, Zhengzhou University, the Second Hospital, Zhengzhou, China5 Department of Otolaryngology, Hunan University of Chinese Traditional Medicine, Changsha, China6 an adjunct Professor of Allergy Unit and Department of Otolaryngology, Shanxi Medical University, Taiyuan, China2005 6 9 2005 5 28 28 2 4 2005 6 9 2005 Copyright © 2005 Yang et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
During clinical practice, we noticed that some patients with both ulcerative colitis (UC) and chronic rhinosinusitis (CRS) showed amelioration of UC after treatment of CRS. This study was designed to identify a possible association between CRS and UC.
Methods
Thirty-two patients with both CRS and UC received treatment with functional endoscopic sinus surgery (FESS) for CRS. Clinical symptom scores for CRS and UC, as well as serum levels of anti-Staphylococcal enterotoxin B (SEB) were evaluated at week 0 and week 12. Sinus wash fluid SEB content was measured with enzyme-linked immunosorbent assay (ELISA). The surgically removed tissues were cultured to identify growth of Staphylococcus. aureus (S. aureus). Immunohistochemistry was employed to identify anti-SEB positive cells in the colonic mucosa. Colonic biopsies were obtained and incubated with SEB. Mast cell activation in the colonic mucosa in response to incubation with SEB was observed with electron microscopy and immunoassay.
Results
The clinical symptom scores of CRS and UC severe scores (UCSS) were significantly reduced in the UC-CRS patients after FESS. The number of cultured S. aureus colonies from the surgically removed sinus mucosa significantly correlated with the decrease in UCSS. High levels of SEB were detected in the sinus wash fluids of the patients with UC-CRS. Histamine and tryptase release was significantly higher in the culture supernate in the patients with UC-CRS than the patients with UC-only and normal controls. Anti-SEB positive cells were located in the colonic mucosa.
Conclusion
The pathogenesis of UC in some patients may be associated with their pre-existing CRS by a mechanism of swallowing sinusitis-derived SEB. We speculate that SEB initiates inappropriate immune reactions and inflammation in the colonic mucosa that further progresses to UC.
==== Body
Background
Ulcerative colitis (UC) is a relapsing chronic inflammatory disease of the colon with unknown etiology. The evidence from animal models suggests that an altered immune response towards the commensal gut microbiota plays a key role in the development and perpetuation of this pathological condition [1,2]. Treatment of UC is still a hot topic in clinical research. Although therapeutics has been advanced greatly in the recent years [3-5], generally inflammatory suppression with glucocorticoids and/or sulfasalazine is still the primary therapy for UC [6,7]. However, a large number of patients are resistant or intolerant to medical treatment [8].
The accumulation of evidence from animal studies, cell culture studies, and clinical observations has indicated an important role for microbes and/or their products in the induction or exacerbation of enterocolitis [9]. Microbial antigen Staphylacoccal enterotoxin B (SEB) has been shown to induce enteropathy, characterized by altered villus-crypt in BALB/c and T cell-reconstitued SCID mice [10]. Oral feeding of SEB leads the development of colitis in the absence of regulatory T cells [11]. SEB has been shown to cross the intestinal epithelial cell barrier [12] and is implicated in inducing inappropriate immune reactions in humans [13] and mice [14] leading to activation of T cells to produce proinflammatory cytokines such as IL-1, IL-2, TNF-α and IFN-γ [15,16]. SEB is capable of inducing inflammatory cytokines release from explants of colonic biopsies from patients with IBD [17]. However, our knowledge about the association between SEB and IBD is still limited. SEB is synthesized as a precursor protein of 266 amino acids. This precursor is then activated during excretion by cleavage of the N-terminal portion of the molecule. The active enterotoxin B is a single 239 amino acid chain of molecular weight 28,000 Daltons and isoelectric point of 8.6 [18]. SEB is a superantigen and possesses powerful immune regulatory capabilities that result in increased T cell activation and proliferation. SEB-treated Balb/c mice display a dose-dependent colonic inflammation [19]. SEB can also induce colonic epithelial barrier dysfunction [20] that promotes uptake of exogenous antigens [21], microbial products and other noxious substances into the intestinal tissue to contact immune cells and to initiate immune reactions.
Over the past 15 years, more than 2000 patients with chronic rhinosinusitis (CRS) including some patients with both CRS and UC (UC-CRS) visited our clinic. Their CRS was treated with different remedies, including medical treatment and functional sinus endoscopic surgery (FESS). Apart from the satisfactory efficacy in the treatment of CRS, some of the patients with UC-CRS showed great amelioration of UC as well that couldn't be explained with the primary treatment alone. Therefore, we postulated that there might be an association between CRS and UC in these patients. Microbial infection is the most common cause of CRS. The microbial products produced in the sinuses, such as SEB [22,23] can be discharged into the nasal cavity with secretions through the natural ostia going backward to the pharynx, and then swallowed. After reaching the gastrointestinal tract, SEB is capable of affecting intestinal mucosal physiological functions [24,25]. Therefore, this prospective study was designed to identify a possible association between CRS and UC.
The nasal cavity and sinus are primary sites of colonization by S. aureus [26], and SEB was detected in the nasal cavity of the patients with CRS, nasal polyposis and allergic rhinitis [22,23,27,28]. Based on clinical observations, we hypothesized that sinusitis-derived SEB played a role in the pathogenesis of chronic inflammation in the intestinal mucosa via impairing the epithelial barrier function and inducing inappropriate immune reactions. In this study, we aimed to investigate if (i) colonization of S. aureus can be identified in the sinus cavity of the patients with UC-CRS; (ii) SEB production can be detected in the sinuses; (iii) the improvement of UC associates with the removal of inflammation from the sinuses; (iv) immune cells in the colonic mucosa have been and can be activated by SEB.
Methods
In this study, we evaluated the effect of FESS in the treatment of a group of patients with refractory CRS and moderate UC. The study protocol was approved by the research ethics committee at the First Hospital, Shanxi Medical University. All patients gave written informed consent prior to recruitment. Surgical intervention was chosen for the patients because they had tried different therapies that failed to cure their CRS. The clinical condition of UC was moderate and suitable for sinus surgery. Those patients with severe UC were excluded. Unless otherwise mentioned, all the reagents used in this study were purchased from Sigma Aldrich (USA).
Thirty two patients with refractory CRS and moderate UC were enrolled in this study. Another 32 patients with moderate UC but no CRS were selected as a control group. Twenty-five healthy medical students were enrolled in this study as a healthy control group, their serum and nasal wash fluids were collected as controls. We diagnosed CRS as inflammation of the sinus mucosa with a persistent mucoid or mucopurulent nasal discharge for longer than 3 months that resisted antimicrobial therapy and antral irrigation. Diagnosis was made on the basis of clinical history, rhinoscopic findings, and computed tomographic scan of rhinosinuses. CRS was confirmed by computed tomographic examination that showed diffuse mucosal thickening in the ethmoid or/and maxillary sinuses bilaterally with scores higher than 12 by the Lund-Mackay staging system [29].
The diagnosis and treatment of UC was carried out by one gastroenterologist (Dr An Z). The diagnosis of UC [30] was based on clinical history, colonoscopy and histology. The history included persistent bloody diarrhea, rectal urgency, or tenesmus, stool. Examinations, sigmoidoscopy and biopsy were performed to confirm the presence of colitis and to exclude the presence of infectious etiologies. UC severity scores (UCSS) [31] were recorded for each patient on admission (week 0) and week 12. The UCSS were evaluated from 4 clinical items: stool frequency, rectal bleeding, sigmoidoscopic appearance, and the physician's global assessment.
Sinus and nasal wash fluids collection
Every patient with UC-CRS underwent maxillary sinus puncturing and irrigating with saline. The SWF was collected prior to other procedures. Five ml saline was injected into the sinus cavity and recollected and stored at -70°C for further analysis. Protein content in SWF was evaluated with UV spectrophotometer (Beckman DU-65) at 280 nm. Contents of SEB, SEA and TSST1 in SWF were evaluated with ELISA. Nasal lavage fluids (5 ml saline) instead of sinus irrigation were collected from the healthy controls and UC controls. None of the subjects had acute upper respiratory acute infections within the past month.
Treatment
The procedures of FESS followed the previous reports [32]. Penicillin G (800,000 IU, i.m., Bid) was administered after the surgery for 3 days (no patient in this group showed allergic signs for penicillin G). The surgical side of sinus was irrigated daily until the irrigated fluids became clear. After discharge, the patients were required to visit the hospital to conduct a post-operation examination of nasal cavity and sinuses on a monthly basis for the first year. Any unusual feelings in the operated sinus area were recorded. Physicians would do necessary treatment for relapse of inflammation in the sinuses during the entire follow-up period. The treatment of UC including sulfasalazine and corticosteroids continued for each patient during the 12-week observation period including the patients with UC-CRS and the patients with UC.
Serology of C-reaction protein, IgE and cytokines
Blood samples were collected at week 0 and week 12. The serum was separated and analyzed for contents of C reactive protein and SEB specific IgE respectively. The methods were done as previously reported. [33,34] Blood samples were also collected from the healthy controls and UC-controls twice, 12 weeks apart.
S. aureus identification
Surgically removed tissues were obtained from each patient (a nasal lateral wall swab was obtained from UC-only patients and healthy controls instead) and were sent to the Microbiological Laboratory within 30 minutes. 100 mg mucosa was homogenized in 0.1 ml saline. The same volume of the homogenates from each patient was cultured aerobically on blood agar plates. Colonies of different morphology growing on the agar were separately enumerated, subcultured on blood agar, Gram stained, and tested for catalase production. Isolates with a typical Gram-stained appearance and a positive catalase test were tested for coagulase production, and positive isolates were regarded as S. aureus. The template DNA was extracted from the positive cultured colony of each subject with the procedures reported by Riffon [35]. PCR was performed with the samples from each subject. The nucleotide sequence is available at the GenBank data library (accession number, CP000046). The primers were designed using the software Primer3 and the specificity was determined with Blast. The primers were: 5'-ttgcatatccgcgtcaaata-3' and 5'-cttcatgttctttcgcatcg-3'. Amplification was performed on a Perkin-Elmer (Norwalk, Conn.) thermocycler in 25-μl reaction mixtures. The program consisted of an initial denaturation step at 94°C for 2 min, followed by 22 cycles of 1 s at 94°C, 2 s at 58°C, and 10 s at 72°C, with a final extension step at 72°C for 5 min. Amplification products were separated on a 1% agarose gel and stained with ethidium bromide before being analyzed on a UV bench by using GelDoc2000 (Bio-Rad). The PCR products were routinely sequenced to confirm amplification of the targeted sequence.
Immunohistochemistry
Two pieces of colonic tissue at the edge of ulcers was taken during colonoscopy from each patient. One biopsy was snap frozen with liquid nitrogen in Tissue Tek. Cryosections were air dried overnight and fixed with cold acetone for 15 minutes. The endogenous peroxidase was quenched by incubation in 0.3% H2O2 for 30 minutes and non-specific binding was blocked by incubating the sections in 10% goat serum/phosphate-buffered saline. Specimens were incubated with mouse anti-SEB monoclonal antibody or IgG isotype control immunoglobulin. Subsequently, after incubation with secondary rabbit anti-mouse antibody (HRP conjugated, Vector Laboratories), color was developed applying the Vectastain peroxidase detection kit and diaminobenzidine as a substrate. Counterstaining was performed with hematoxylin. The negative control slides omitted the first antibody or added the isotype control instead.
Tryptase and histamine release from the cultured colonic biopsies in response to the stimulation of SEB
One of the two intestinal biopsy specimens from each patient was subjected to histamine and tryptase release evaluation in response to SEB stimulation. Briefly, upon removal, the specimens were weighed in a sterile vial containing 1 ml pre-warmed (37°C) and oxygenized RPMI 1640 culture medium (no serum). The specimen containing vials were incubated for 30 minutes at 37°C in a humidified 95% and air containing 5% CO2 atmosphere and environment and rocked gently at 20 cycles/min. The supernatants were carefully collected and temporally kept at 4°C. The specimens were washed with pre-warmed no-serum RPMI medium for 3 times, then another RPMI 1640 medium (no serum) with 20 μl SEB/ml was added to the vial and incubated for another 30 minutes. The colonic tissues were carefully taken out and immersed to 2% glutaraldehyde immediately. The supernatants were centrifuged at 17,530 × g for 10 minutes at 4°C. The supernatants were stored at -70°C for further assay. Eight mucosal specimens from surgical removed colonic tissue of the patients with colonic cancer were collected. Samples were selected from the uninvolved part and processed as normal controls.
For tryptase activity assay, triplicate aliquots (10 μL) of the supernatants were added to 200 μL of buffer (50 mmol/L Tris/HCl, pH 7.6, 120 mmol/L NaCl, 20 μg/mL heparin) containing 0.5 mmol/L substrate (tosyl-Glycine-Proline-Arginine-pNitroanilide) and incubated at room temperature for 17 hours (± inhibitors as indicated). Cleavage of the substrate was measured using a microtiter plate reader (absorbance 405 nm) and normalized to the weight of the biopsy specimens used in each specimen.
For histamine assay, triplicate aliquots (50 μL) of the supernatants using a selective enzyme-linked immunoassay (Immunotech, Marseille, France) according to the manufacturer's directions. Histamine was measured using a microtiter plate reader (absorbance 405 nm) and normalized to the weight of the biopsy specimens.
Electron microscopy
After incubation with SEB, the colonic biopsies were fixed with 2% glutaraldehyde for 2 hours at room temperature and postfixed in 1% osmium tetroxide for 1 hour. After dehydration and embedding in EPON, ultra thin section were cut and collected on 200 mesh grids, stained with uranyl acetate and lead citrate, and observed with a JEC 1200 transmission electron microscope. Thirty mast cells were selected randomly from each specimen and photographed. The granules of the selected mast cells were categorized into three types: intact, piecemeal degranulation (PMD) and anaphylactic degranulation (AND) according to reported criteria [36].
Statistical analysis
Data were expressed as means ± standard deviation (SD). Differences between groups were analyzed with the student t test or χ2 test. Correlationship between groups was demonstrated by correlation analyses. The significant criteria was set at p < 0.05.
Results
Clinical symptoms of CRS improved after FESS
All the CRS patients underwent FESS in their maxillary sinusitis. Twelve (37.5%) patients also underwent bilateral ethmoidectomy. Eight patients also underwent unilateral ethmoidectory (25%, right or left side). Four (12.5%) patients also underwent septum orthomorphia. Fourteen (43.75%) patients had nasal polyps that were removed during FESS. Five (15.63%) patients underwent partial middle turbinectomy. Six (18.75%) patients underwent inferior turbinectomy. During the 12-week observation period, all the patients underwent regular follow-up visit (The total follow-up period lasted for one year after FESS). Twelve weeks after FESS, all the patients reported significant improvement in their clinical symptoms of CRS. Clinical symptom scores were recorded at week 0 and week 12 that were defined from 0 to 4. 0 for no symptoms; 4 for severe. Most patients had moderate to severe CRS clinical symptoms before FESS; 29 out of 32 (90.62%, scored 0 and 1) patients showed significant reduction of their clinical scores of sinusitis; 21 out of 32 (65.63%) scored 0 and 1 (Fig 1).
Effect of treatment with FESS on UC clinical symptoms
The demographic characteristics of the patients in the present study are listed in Table 1. In the period of 12 week observation period, treatment for UC of the patients in this study were carried out by one gastroenterologist according to routine therapy. After FESS, the requirement for steroids was gradually reduced in the UC-CRS group and was significantly less than the UC group (Fig 2). From week 0 to week 12, levels of C reactive protein rose in control patients (with UC only) who received routine treatment rather than FESS, from 4 to 20 mg/l, whereas the value reduced significantly in the FESS group. The C reactive protein in the healthy control group was under detectable levels (Fig 3).
Sigmoidoscopy was performed in all subjects during admission and 12 weeks after FESS. The control UC patients were also examined with sigmoidoscopy in the same period. The results showed that the patients with both UC-CRS significantly improved in appearance of the colonic mucosa after FESS, while UC patients showed no change. UCSS assessment also showed significant improvement in the patients with UC-CRS after FESS. The UCSS did not show much difference in the patients with UC-only although they received the routine treatment for their UC (Fig 4).
S. aureus was identified in the sinuses of the patients with CRS
Bacterial culture results showed S. aureus growth in the surgically removed tissue in 24 (75%) patients with UC-CRS. The numbers of S. aureus colonies in the culture ranged from 0 to 76 with an average of 28.5/patient. The reduction of UC clinical symptom scores significantly correlated with the number of the cultured S. aureus colonies from the surgical removed sinus mucosa (Fig 5, r = 0.8399, p < 0.0001). PCR results showed S. aureus DNA amplified products from the surgical removed tissue of 32 (100%) patients that further confirmed the existence of S. aureus infection in the sinuses of patients with UC-CRS (Fig 6). RT-PCR amplified product sequence analysis confirmed that the amplified bands were consistent with the target DNA sequences. Fig 7 depictes serum IgE assay results. Serum specific IgE and total IgE levels were evaluated with ELISA. Before FESS, serum anti-SEB antibody levels were significantly higher in the patients with UC-CRS than the patients with UC only and the healthy controls (p < 0.05). The levels of specific IgE in UC-CRS patients were significantly decreased as measured at week 12 after FESS (p < 0.05). The levels of total IgE were also higher in UC-CRS patients before FESS compared with UC-only patients and healthy controls (p < 0.05). The results showed the total IgE levels were also significantly decreased after FESS as compared with the results before FESS (p < 0.05). A ratio of specific IgE/total IgE was calculated and listed as: control group (28.6% and 15.6%; week 0 and week 12); UC-only group (10.6% and 8.8%) and UC-CRS group (26.3% and 23.5%). Ratio statistical analysis showed that differences between week 0 and week 12 didn't reach significance (p > 0.05) in all three groups indicating the same tendency of change in total IgE and the SEB specific IgE.
SEB specific antibody positive stained cells were located in the colonic mucosa. High concentration of SEB was evaluated in the sinus wash fluids
Immunohistochemistry showed anti-SEB positive staining cells in the colonic mucosa of patients with UC-CRS, but not in patients with UC only (Fig 8). Sinus wash fluids were evaluated for SEB content. A significantly higher SEB level was found in the sinus wash fluids from patients with UC-CRS. SEB content was much lower in the nasal wash fluids of the UC-only patients and healthy controls. Sinus wash fluids were also colleted from 12 patients with CRS, but without UC and 12 patients with nasal polyposis and CRS, but without UC at week 0 and week 12 after FESS. SEB content in the sinus wash fluids was also evaluated (Fig 9). The contents of SEA and TSST1 in the sinus wash fluids were undetectable.
Colonic mucosal mast cell degranulation in response to SEB stimulation ex vivo
The numbers of mast cells in the colonic mucosa were counted with light microscopy. There were significantly more mast cells in patients with UC-only and UC-CRS compared with normal control colonic mucosa. As depicted in Figures 10 and 11, electron microscopy revealed three types of granules in the mast cells of the colonic mucosa, intact, PMD and AND. In the normal colonic mucosa, there was a low ratio of degranulation in the colonic mast cells (PMD, 7.7 ± 4.2%; AND, 2.2 ± 3.1%). A significant increase in the ratio of degranulation of both PMD and AND was observed in colonic specimens of patients with UC-only and UC-CRS before SEB stimulation. A great difference was observed in the mast cell degranulation in the colonic mucosa in response to SEB stimulation between the groups with UC only and UC-CRS ex vivo. AND type degranulation was significantly increased in patients with UC-CRS compared to UC only patients (p < 0.05).
Tryptase and histamine release in response to SEB stimulation ex vivo
Tryptase and histamine was detected in the culture of the colonic biopsies in the first 30-minute. The levels of tryptase and histamine were approximately 3 times higher in the specimens from patients with UC or UC-CRS compared to normal controls. The release of tryptase and histamine further increased after incubation with SEB in the UC-CRS group, but not in the UC group. The tryptase inhibitor BABIM significantly inhibited the effect of tryptase (Fig 12).
Discussion
We have observed the clinical outcome of a group of patients with both CRS and UC after treatment with FESS. Clinical symptom scores of both CRS and UC were significantly reduced 12 weeks after FESS. UC-CRS patients exhibited sinus infection with S. aureus high levels of anti-SEB antibody in the serum, and anti-SEB antibody positive stained cells in the colonic mucosa. In particular, the mast cells in the colonic biopsy specimens from the patients with UC-CRS were activated by incubation with SEB by showing extensive degranulation and release of histamine and tryptase. This is consistent with Dionne's study [17] that reported SEB induced IBD rectal biopsies to release inflammatory mediators ex vivo such as TNF-α and IL-1.
During the 12-week observation period, both groups of UC patients with or without CRS were treated with the routine remedies for UC, mainly sulfasalazine and glucosteroids. Therefore, the marked improvement of UC in patients with UC-CRS should be the removal of inflammation in the sinuses by FESS. Although brief (3 days) penicillin G administration was given to patients after FESS, it did not likely contribute to the amelioration of UC, because penicillin G mainly effects on acute Gram positive bacterial infection which grow fast and have high requirement of cell wall synthesis whereas UC is a chronic inflammatory disease. The prevalence of Crohn's disease, another form of IBD, was elevated in patients with chronic sinonasal disease, occurring in approximately one-half of patients followed at a tertiary IBD center [37].
Chronic rhinosinusitis is a multi-variant disease. Bacterial infection is the accepted cause of acute sinusitis in most circumstances, but the causes of chronic sinusitis are more complicated. S. aureus is a common pathogen that contributes to both acute and chronic rhinosinusitis [38]. The present study has added supportive data to the observation above. We identified S. aureus in the surgically removed tissues from the sinuses in 75% of the patients with UC-CRS and 100% of sinus samples showed positive mRNA for S. aureus. One of the features of CRS is purulent secretions into nasal cavity via the natural ostia. The nasal mucosa is covered with a lining of ciliated epithelium, covers with mucus blanket. One of the functions of this mucus blanket is the trapping of dust from inhaled air and secretions from the sinuses, and removing them from the nasal cavity by rhythmic locomotion. Because of the backward movement of the mucus blanket, some of the trapped substances such as SEB may be swallowed and enter the gastrointestinal tract to alter the mucosal physiological functions. The present data show high levels of SEB in the sinus wash fluids from the patients with UC-CRS. Previous reports also indicate high levels of SEB in nasal/sinus secretions of patients with chronic sinusitis, especially in those with polyposis [22]. The present data also showed nasal/sinus polyposis in 43.75% UC-CRS patients. Thus, it is quite possible that the SEB-containing secretions released from sinuses were swallowed and contribute to the pathophysiological changes in the colonic mucosa. The two groups of patients were treated with approximately the same remedies for their UC during the 12-week observation period. The different outcome between the two groups of patients that supports our hypotheses that CRS derived SEB plays a certain role in the pathogenesis of UC in these patients. However, the present data show that not all patients with S. aureus-infected CRS suffer from UC (Figure 9). Additional factors such as adjuvants may be required to act synergistically with SEB to induce intestinal disorders [27].
Serum IgE antibody level elevation is one of the main features of the type I allergic reactions including acute and chronic allergic diseases [39,40]. The accumulated evidence has emphasized that IgE mediated inflammation plays a role in the pathogenesis of ulcerative colitis [41,42]. We detected high levels of total serum IgE in the UC-CRS patients in this study. This phenomenon stresses that IgE mediated disorders are occurring in these patients. Further evidence we have obtained is that high levels of serum SEB specific IgE antibody were detected in the patients with UC-CRS. Although the antigen SEB that induced the specific IgE production in the present study could be from multiple sources, the fact that a significant reduction of serum SEB specific IgE after FESS indicates that the sinusitis derived SEB may be the obligate antigen. These SEB molecules are also capable of acting as superantigens and reacting with T lymphocytes, inducing massive activation, proliferation, and cytokine production by CD4+ T cells via specific Vbeta elements on the TCR, initiating inappropriate immune reactions in the local tissue and leading to IgE antibody production as well as inflammation [43].
We have located anti-SEB antibody positive stained cells in the colonic biopsies. The positively stained cells were identified in the patients with UC-CRS, but not in the normal colonic mucosa. The phenomenon indicates that there are anti-SEB antibody-bearing cells in the colonic mucosa of these patients. These cells can be the anti-SEB antibody secreting cells (such as plasma cells), or the cells (e.g., mast cells) have been bound by anti-SEB antibodies that are sourced from other cells. We may employ double-antibody staining technique with confocal microscopy in the future to identify the cell types. How have these anti-SEB antibody secreting cells been generated in the colonic mucosa? The mechanism of IgE production induced by SEB is also not clear. SEB from S. aureus is discharged to the nasal cavity from the sinuses, swallowed into the gastrointestinal tract, where the SEB is absorbed by the intestinal epithelial cells via a Toll like receptor (TLR) 2-mediated mechanism. Follicle cells in the Peyer's patch may be one of the paths to transport SEB from the luminal side to the lamina propria. TLR2 has been identified on the M cells of Peyer's patches in intestinal mucosa (44). SEB is one of the ligands for TLR2 (45). Once binding to TLR2 on the M cells, SEB may be internalized and transported across the thin cytoplasm of the M cells. SEB is then released by exocytosis from the basolateral membrane where the SEB can contact immune cells and induce inappropriate immune reactions, such as IgE production and inflammation. However, the detailed mechanism needs to be further understood. SEB then acts as a specific antigen and contacts immune cells in the lamina propria to initiate antibody production. Although this is the first study that reports anti-SEB positive immune cells in the colonic mucosa, there are some studies involved with SEB-colonic mucosa interaction have been published. Lu et al report that SEB compromises mouse colonic epithelial barrier function and induces inflammation in the mucosa [19,20]. McKay et al indicate that SEB induces colonic epithelial barrier deficiency via induction of IFN-γ and TNF-α production that can be partially inhibited by TGF-β [9,24,46]. Dionne et al observed that colonic biopsy specimens from UC patients produce TNF-α significantly more than normal controls in response to SEB stimulation [17]. Taken together, these findings suggest SEB plays an important role in the pathogenesis of chronic inflammatory intestinal disorders. The precise mechanism needs to be further investigated.
We have detected serum C reactive protein levels increasing in the patients with UC of both groups. The C reactive protein levels dropped significantly in the patients with UC-CRS after FESS. C reactive protein is an indicator of body inflammation and although it is non-specific, general elevation of serum C reactive protein in UC patients has been reported [47]. C reactive protein can be a biochemical marker that is useful to stratify patients likely to respond to biologic therapies and to follow response to treatment. The decrease of serum C reactive protein levels may be from the removal of inflammation in the sinuses [48] or from the amelioration of UC or both. Therefore, the decrease of serum C reactive protein can be a useful biological marker of removing inflammation from the sinuses and amelioration of the inflammation in colonic mucosa.
Mast cells play an important role in pathology and pathophysiology of UC [49]. The accumulated evidence shows activated mast cells in the colonic mucosa of the UC patients [49,50]. Our results are in line with previous reports. Furthermore, we provided quantitative data about mast cell activation by showing a significant difference in the ratio of degranulation in the mast cells of the patients with UC-CRS from the patients with UC only and the normal controls. Two types of mast cell degranulation, PMD and AND, were identified in the present study. PMD often presents in mast cells during chronic inflammation whereas AND is more likely in mast cells in acute situation such as in anaphylaxes [51]. In this study, electron microscopy revealed that even in the normal colonic mucosa there are a certain number of degranulated granules in the mast cells including both PMD and AND according to the criteria reported by Crivellato [52]. Most the degranulated granules in the mast cells from the UC patients were of the PMD type. This supports the definition that PMD type degranulation mainly attributes to chronic inflammation [51]. The incubation with SEB further increased mast cell degranulation in the colonic biopsy specimens with AND dominantly in the UC-CRS group. The mechanism of this phenomenon may be that the anti-SEB antibody belongs to the IgE class per se; anti-SEB IgE antibody bound to mast cells through the high affinity IgE receptors (FcåRI); SEB binds to the anti-SEB antibody-IgE receptor complex to activate mast cells.
Mast cells function as effector cells during immune inflammation by releasing chemical mediators. Histamine and tryptase are two of the main mediators and play important roles in both allergic inflammation and chronic inflammation [53,54]. The present study shows mast cell activation in colonic biopsies in response to SEB stimulation ex vivo by showing release of both tryptase and histamine, is consistent with previous reports [17]. The results implicate that when CRS-derived SEB arrives in the colon of these patients with UC-CRS, it will disturb intestinal mucosal function as described below. Tryptase is the main cytosol protein of mast cells, representing approximately 25% of the protein in mast cells [49]. A unique feature of tryptase is its ability to cleave and activate PAR2 receptors on several cell types including intestinal epithelial cells [53]. Intestinal epithelial barrier function may be compromised in response to PAR2 activation [55]. It may contribute partially (if not totally) to the deficiency of the intestinal barrier of patients with UC [56]. The fact that SEB induces tryptase and histamine release from the colonic biopsy specimens of the UC-CRS patients, but not from the UC-only patients or normal controls indicates that tryptase and histamine release is specific. One possible explanation for this is that the mast cells in the colonic mucosa have been primed by pre-binding with anti-SEB IgE antibody. The primed mast cells are then activated when they are re-exposed to SEB. Yet further investigation needs to be done to clarify the mechanism. Supporting evidence observed in the present study is that further increase in AND type degranulation in the mast cells of the colonic mucosa of the patients with UC-CRS after incubation with SEB implicates that the mast cells have had acute release of the granular contents within a short time in contrast of PMD type degranulation, the chronic release of the granular content [36,51]. The amount of tryptase and histamine released from the UC-only biopsies after incubation with SEB didn't show statistical difference from that before the addition of SEB that indicates only a spontaneous mediator released from these mast cells in the colonic tissue.
Conclusion
We have reported a group of patients with both UC and CRS whose UC clinical symptoms were marked improved in addition to the amelioration of CRS after FESS. These data show a possible association between CRS and UC. There may be a subset of UC patients with high levels of IgE some of which is directed towards SEB. Although the present data are not sufficient to draw a conclusion on whether this is the initiating factor in UC or the progression factor, certainly it is a clue for us to further explore pathogenesis of UC about that we have had very limited knowledge yet.
List of abbreviations
UC, ulcerative colitis; UCSS, UC severe score; CRS, chronic rhinosinusitis; SEB, Staphylacoccal enterotoxin B; FESS, functional endoscopic sinus surgery; EM, electron microscopy; PMD, piecemeal degranulation; AND, anaphylactic degranulation; HRP, horseradish peroxidase.
Competing interests
The author(s) declare that they have no competing interests.
Author's contributions
PY was involved in study design, a portion of FESS, histology, EM observation, data analysis and manuscript preparation; TL, BW and TZ were involved CRS treatment, FESS and some experiments; AY was involved in UC management; PZ was involved as a consultant in UC management; TD was involved as a consultant in sinusitis treatment.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
This study was supported by a grant of the National Foundation of Natural Science of China
Figures and Tables
Figure 1 Chronic rhinosinusitis clinical scores of the patients with both chronic rhinosinusitis (CRS) and ulcerative colitis (UC) was recorded before (week 0, A) and after (week 12, B) FESS. 0 stands no symptom; 4 stands severe symptom. The numbers stand for clinical scores (percentage of patient). The results of χ2 analyses showed p < 0.05 as compared the clinical scores of week 0 with that of week 12.
Figure 2 Prednisolone requirement was recorded from the patients with UC-CRS and the patients with UC at week 0 (A) and week 12 (B). The bars represent the number of patients. χ2 analyses were performed with the data between the two groups. The results showed p > 0.05 at week 0 and p < 0.05 at week 12.
Figure 3 Serum C reactive protein. Bars stand for serum levels of C reactive protein. Data were presented mean ± SD. t test was performed with the data before and after FESS. *, p < 0.05, compared with UC-only group. UC-only: patients with UC only. UC-CRS: patients with both UC and CRS.
Figure 4 Ulcerative colitis severe scores of the patients with UC-CRS and the patients with UC-only were recorded at week 0 (A) and week 12 (B). Bars represent summary scores from clinical observation. Data were presented as numbers of patients of each score. χ2 analyses were performed with the data between the two groups. The results showed p > 0.05 at week 0 and p < 0.05 at week 12.
Figure 5 S. aureus was cultured with the surgical removed sinus mucosa of the patients with UC-CRS. The method using to identify S. aureus was described in the text. Bars stand for numbers of patients.
Figure 6 RT-PCR results of S. aureus DNA amplification. DNA templates were extracted from the surgical removed sinus tissue of the patients with UC-CRS. Lane 1 is negative control by adding water instead of DNA template. Lanes 2–11 are representative PCR bands of S. aureus DNA amplification.
Figure 7 Serum IgE antibody. Serum samples were collected from all the subjects at week 0 and week 12. IgE levels were evaluated with ELISA. Data were presented as mean ± SD. *, p < 0.05, compared with the same group at week 0. A, serum total IgE; B, serum anti-SEB IgE.
Figure 8 Anti-SEB positive stained cells were located in the colonic mucosa. Colonic biopsies were got from the patients with UC-only (A) and the patients with UC-CRS (B and C). Cryosections were stained with anti-SEB antibody or IgG isotype control immunoglobulin (B). Sections were visualized by the immunoperoxidase method and counterstained with hematoxylin. SEB specific IgE-bearing cells were stained in brown (C). ×200.
Figure 9 SEB content in sinus wash fluids (nasal lavage fluids for the groups of UC-only and healthy controls). Bars stand for SEB content. Data were presented as mean ± SD. *, p < 0.05, compared with the same group at week 0. CRS-only: a group of 12 patients with CRS but not UC. CRS-Pol: a group of 12 patients with CRS and nasal polyposis.
Figure 10 Mast cells in the colonic mucosa. Representative electron photomicrographs of mast cells were taken from the colonic mucosal specimens of A: normal controls, B: patients with UC-only and C and D: patients with UC-CRS (C, before stimulated with SEB; D, after stimulated with SEB). ×3,000.
Figure 11 The ratio of mast cell degranulation in the colonic mucosa. Under an electron microscope, mast cell degranulation was categorized into piecemeal degranulation (PMD) and anaphylactic degranulation (AND). Percentage of degranulation was calculated with the number of total granules and the number of PMD or AND. Student t test was performed with the data between groups. *, P < 0.05, compared with PMD of normal group (A). #, p < 0.05, compared with PMD of normal group (B). $, p < 0.05, compared with AND of the normal group (A). &, p < 0.05, compared with AND of the normal group (B). @, p < 0.05, compared with PMD of UC only group (A). +, p < 0.05, compared with AND of UC only group (B).
Figure 12 Tryptase and histamine release from the colonic mucosal biopsy specimens in healthy controls, patients with UC-only and patients with UC-CRS. Bars represent contents of tryptase (A) or histamine (B) in the supernatant of the culture of the colonic biopsies. * and #, p < 0.05, compared with the normal control samples; $, p < 0.05, compared with UC-only patients.
Table 1 Demographic and disease features of the patients before FESS
Group Control FESS
Sex Male: 15; female: 17 Male: 18; female: 14
Age 33.23 (20 – 66) 35.82 (26 – 58)
Weight (kg) 62 (55 – 68) 60 (51 – 71)
Duration of UC (months) 28 (15 – 68) 35 (29 – 78)
Prednisolone equivalent (mg/day) 15 (5 – 20) 15 (5 – 20)
Data are shown as medians (interquartile ranges). UC, ulcerative colitis; FESS: functional endoscopic sinus surgery.
==== Refs
Macdonald TT Monteleone G Pender SLF Recent developments in the immunology of inflammatory bowel disease Scand J Immunol 2000 51 2 9 10632969 10.1046/j.1365-3083.2000.00658.x
Campieri M Gionchetti P Bacteria as the cause of ulcerative colitis Gut 2001 48 132 5 11115835 10.1136/gut.48.1.132
Kanai T Watanabe M Clinical application of human CD4+ CD25+ regulatory T cells for the treatment of inflammatory bowel diseases Expert Opin Biol Ther 2005 5 451 62 15934825 10.1517/14712598.5.4.451
Annese V Latiano A Rossi L Lombardi G Dallapiccola B Serafini S Damonte G Andriulli A Magnani M Erythrocytes-Mediated Delivery of Dexamethasone in Steroid-Dependent IBD Patients-A Pilot Uncontrolled Study Am J Gastroentero 2005 100 1370 5 10.1111/j.1572-0241.2005.41412.x
Su C Lichtenstein GR Recent developments in inflammatory bowel disease Med Clin North Am 2002 86 1497 523 12510462 10.1016/S0025-7125(02)00085-8
Cima RR Pemberton JH Medical and surgical management of chronic ulcerative colitis Arch Surg 2005 140 300 10 15781797 10.1001/archsurg.140.3.300
Campieri M New steroids and new salicylates in inflammatory bowel disease: a critical appraisal Gut 2002 50 III43 6 11953332
Baert FJ Rutgeerts PJ Medical therapies for ulcerative colitis and crohn's disease Curr Gastroenterol Rep 2000 2 446 450 11079045
McKay DM Intestinal inflammation and the gut microflora Can J Gastroenterol 1999 13 509 516 10464352
Benjamin MA Lu J Donnely G Dureja P McKay DM Changes in murine jejunal morphology evoked by the bacterial superantigen Staphylococcus aureus enterotoxin B are mediated by CD4+ T cells Infect Immun 1998 66 2193 9 9573107
Zhou P Borojevic R Streutker C Terry NG Croitoru K Oral Administration Of Bacterial Superantigen Activates a Severe Colitis in the Absence Of Regulatory T Cells Gastroenterology 2005 128 170
Hamad AR Marrack P Kappler JW Transcytosis of staphylococcal superantigen toxins J Exp Med 1997 185 1447 54 9126925 10.1084/jem.185.8.1447
Roberts AI Blumberg RS Christ AD Brolin RE Ebert EC Staphylococcal enterotoxin B induces potent cytotoxic activity by intraepithelial lymphocytes Immunology 2000 101 185 90 11012771 10.1046/j.1365-2567.2000.00088.x
Cong Y Brandwein SL McCabe RP Lazenby A Birkenmeier EH Sundberg JP Elson CO CD4+ T cells reactive to enteric bacterial antigens in spontaneously colitic C3H/HeJBir mice: increased T helper cell type 1 response and ability to transfer disease J Exp Med 1998 187 855 64 9500788 10.1084/jem.187.6.855
Siebrecht MS Hsia E Spychalski C Nagler-Anderson C Stimulation of murine intestinal intraepithelial lymphocytes by the bacterial superantigen staphylococcal enterotoxin B Int Immunol 1993 5 717 24 8369237
Spiekermann GM Nagler-Anderson C Oral administration of the bacterial superantigen staphylococcal enterotoxin B induces activation and cytokine production by T cells in murine gut-associated lymphoid tissue J Immunol 1998 161 5825 31 9834060
Dionne S Laberge S Deslandres C Seidman EG Modulation of cytokine release from colonic explants by bacterial antigens in inflammatory bowel disease Clin Exp Immunol 2003 133 108 14 12823284 10.1046/j.1365-2249.2003.02191.x
Baker MD Acharya KR Superantigens: structure-function relationships Int J Med Microbiol 2004 293 529 37 15149028 10.1078/1438-4221-00298
Lu J Wang A Ansari S Hershberg RM McKay DM Colonic bacterial superantigens evoke an inflammatory response and exaggerate disease in mice recovering from colitis Gastroenterology 2003 125 1785 95 14724831 10.1053/j.gastro.2003.09.020
Lu J Philpott DJ Saunders PR Perdue MH Yang PC McKay DM Epithelial ion transport and barrier abnormalities evoked by superantigen-activated immune cells are inhibited by interleukin-10 but not interleukin-4 J Pharmacol Exp Ther 1998 287 128 36 9765331
Yang PC Wang CS An ZY A murine model of ulcerative colitis: induced with sinusitis-derived superantigen and food allergen BMC Gastroenterol 2005 5 6 15745456 10.1186/1471-230X-5-6
Bernstein JM Kansal R Superantigen hypothesis for the early development of chronic hyperplastic sinusitis with massive nasal polyposis Curr Opin Otolaryngol Head Neck Surg 2005 13 39 44 15654214 10.1097/00020840-200502000-00010
Conley DB Tripathi A Ditto AM Reid K Grammer LC Kern RC Chronic sinusitis with nasal polyps: staphylococcal exotoxin immunoglobulin E and cellular inflammation Am J Rhinol 2004 18 273 8 15586797
McKay DM Bacterial superantigens: provocateurs of gut dysfunction and inflammation Trends Immunol 2001 22 497 501 11525940 10.1016/S1471-4906(01)02000-2
Ibbotson JP Lowes JR Potential role of superantigen induced activation of cell mediated immune mechanisms in the pathogenesis of Crohn's disease Gut 1995 36 1 4 7534251
Kluytmans JA Wertheim HF Nasal carriage of Staphylococcus aureus and prevention of nosocomial infections Infection 2005 33 3 8 15750752 10.1007/s15010-005-4012-9
Okano M Hattori H Yoshino T Sugata Y Yamamoto M Fujiwara T Satoskar AA Satoskar AR Nishizaki K Nasal exposure to Staphylococcal enterotoxin enhances the development of allergic rhinitis in mice Clin Exp Allergy 2005 35 506 14 15836761 10.1111/j.1365-2222.2005.02195.x
Rossi RE Monasterolo G Prevalence of serum IgE antibodies to the Staphylococcus aureus enterotoxins (SAE, SEB, SEC, SED, TSST-1) in patients with persistent allergic rhinitis Int Arch Allergy Immunol 2004 133 261 6 14976395 10.1159/000076833
Ashraf N Bhattacharyya N Determination of the "incidental" Lund score for the staging of chronic rhinosinusitis Otolaryngol Head Neck Surg 2001 125 483 6 11700446 10.1067/mhn.2001.119324
Regueiro MD Diagnosis and treatment of ulcerative proctitis J Clin Gastroenterol 2004 38 733 40 15365396 10.1097/01.mcg.0000139178.33502.a3
Schroeder KW Tremaine WJ Ilstrup DM Coated oral 5-aminosalicylic acid therapy for mildly to moderately active ulcerative colitis. A randomized study N Engl J Med 1987 317 1625 9 3317057
Frisch T Arndal H Fons M Outcome for the first 85 patients treated with the functional endoscopic sinus surgery technique Rhinology 1995 33 236 9 8919219
Latkovskis G Licis N Kalnins U C-reactive protein levels and common polymorphisms of the interleukin-1 gene cluster and interleukin-6 gene in patients with coronary heart disease Eur J Immunogenet 2004 31 207 13 15379752 10.1111/j.1365-2370.2004.00476.x
Okano M Takishita T Yamamoto T Hattori H Yamashita Y Nishioka S Ogawa T Nishizaki K Presence and characterization of sensitization to staphylococcal enterotoxins in patients with allergic rhinitis Am J Rhinol 2001 15 417 21 11777251
Riffon R Sayasith K Khalil H Dubreuil P Drolet M Lagace J Development of a rapid and sensitive test for identification of major pathogens in bovine mastitis by PCR J Clin Microbiol 2001 39 2584 9 11427573 10.1128/JCM.39.7.2584-2589.2001
Crivellato E Nico B Mallardi F Beltrami CA Ribatti D Piecemeal degranulation as a general secretory mechanism Anat Rec A Discov Mol Cell Evol Biol 2003 274 778 84 12923888 10.1002/ar.a.10095
Book DT Smith TL McNamar JP Saeian K Binion DG Toohill RJ Chronic sinonasal disease in patients with inflammatory bowel disease Am J Rhinol 2003 17 87 90 12751702
Busaba NY Siegel N Salman SD Bacteriology of nontraumatic maxillary sinus mucoceles versus chronic sinusitis Laryngoscope 2000 110 969 71 10852515 10.1097/00005537-200006000-00016
Gelfand EW Inflammatory mediators in allergic rhinitis J Allergy Clin Immunol 2004 114 S135 8 15536444 10.1016/j.jaci.2004.08.043
Kozel MM Sabroe RA Chronic urticaria: aetiology, management and current and future treatment options Drugs 2004 64 2515 36 15516152
Wilson NW Self TW Hamburger RN Severe cow's milk induced colitis in an exclusively breast-fed neonate. Case report and clinical review of cow's milk allergy Clin Pediatr (Phila) 1990 29 77 80 2406081
Heine RG Pathophysiology, diagnosis and treatment of food protein-induced gastrointestinal diseases Curr Opin Allergy Clin Immunol 2004 4 221 9 15126946 10.1097/00130832-200406000-00015
Tohno M Shimosato T Kitazawa H Katoh S Iliev ID Kimura T Kawai Y Watanabe K Aso H Yamaguchi T Saito T Toll-like receptor 2 is expressed on the intestinal M cells in swine Biochem Biophys Res Commun 2005 330 547 54 15796917 10.1016/j.bbrc.2005.03.017
Torres BA Perrin GQ Mujtaba MG Subramaniam PS Anderson AK Johnson HM Superantigen enhancement of specific immunity: antibody production and signaling pathways J Immunol 2002 169 2907 14 12218104
Fan H Zingarelli B Peck OM Teti G Tempel GE Halushka PV Cook JA Lipopolysaccharide and gram-positive bacteria induced cellular inflammatory responses: Role of heterotrimeric Gαi Proteins Am J Physiol Cell Physiol 2005 289 c293 301 15788486 10.1152/ajpcell.00394.2004
McKay DM Singh PK Superantigen activation of immune cells evokes epithelial (T84) transport and barrier abnormalities via IFN-gamma and TNF alpha: inhibition of increased permeability, but not diminished secretory responses by TGF-beta2 J Immunol 1997 159 2382 90 9278329
Beaven SW Abreu MT Biomarkers in inflammatory bowel disease Curr Opin Gastroenterol 2004 20 318 27 15703659 10.1097/00001574-200407000-00004
Bjerrum L Gahrn-Hansen B Munck AP C-reactive protein measurement in general practice may lead to lower antibiotic prescribing for sinusitis Br J Gen Pract 2004 54 659 62 15353050
He SH Key role of mast cells and their major secretory products in inflammatory bowel disease World J Gastroenterol 2004 10 309 18 14760748
Stoyanova II Gulubova MV Mast cells and inflammatory mediators in chronic ulcerative colitis Acta Histochem 2002 104 185 92 12086339 10.1078/0065-1281-00641
Crivellato E Ribatti D Involvement of mast cells in angiogenesis and chronic inflammation Curr Drug Targets Inflamm Allergy 2005 4 9 11 15720229 10.2174/1568010053622876
Crivellato E Finato N Isola M Ribatti D Beltrami CA Low mast cell density in the human duodenal mucosa from chronic inflammatory duodenal bowel disorders is associated with defective villous architecture Eur J Clin Invest 2003 33 601 10 12814398 10.1046/j.1365-2362.2003.01182.x
Coelho AM Ossovskaya V Bunnett NW Proteinase-activated receptor-2: physiological and pathophysiological roles Curr Med Chem Cardiovasc Hematol Agents 2003 1 61 72 15317291 10.2174/1568016033356715
Schreiber S Raedler A Stenson WF MacDermott RP The role of the mucosal immune system in inflammatory bowel disease Gastroenterol Clin North Am 1992 21 451 502 1512051
Kunzelmann K Schreiber R Konig J Mall M Ion transport induced by proteinase-activated receptors (PAR2) in colon and airways Cell Biochem Biophys 2002 36 209 14 12139406 10.1385/CBB:36:2-3:209
Kim JA Choi SC Yun KJ Kim DK Han MK Seo GS Yeom JJ Kim TH Nah YH Lee YM Expression of protease-activated receptor 2 in ulcerative colitis Inflamm Bowel Dis 2003 9 224 9 12902845 10.1097/00054725-200307000-00002
|
16144553
|
PMC1215483
|
CC BY
|
2021-01-04 16:03:27
|
no
|
BMC Gastroenterol. 2005 Sep 6; 5:28
|
utf-8
|
BMC Gastroenterol
| 2,005 |
10.1186/1471-230X-5-28
|
oa_comm
|
==== Front
BMC Med GenetBMC Medical Genetics1471-2350BioMed Central London 1471-2350-6-261595523710.1186/1471-2350-6-26Case ReportHaploinsufficiency for BRCA1 is associated with normal levels of DNA nucleotide excision repair in breast tissue and blood lymphocytes Latimer Jean J [email protected] Wendy S [email protected] Jennifer M [email protected] Amal [email protected] Victor G [email protected] Stephen G [email protected] Department of Obstetrics, Gynecology and Reproductive Sciences, School of Medicine, University of Pittsburgh, Pittsburgh, PA, USA2 Biochemistry and Molecular Genetics Program, School of Medicine, University of Pittsburgh, Pittsburgh, PA, USA3 Research Institute, Magee-Womens Hospital, Pittsburgh, PA, USA4 Department of Medicine, Northwestern University Feinberg School of Medicine, Chicago, IL, USA5 Evanston Northwestern Healthcare Center for Medical Genetics, Evanston, IL, USA6 Department of Pathology, School of Medicine, University of Pittsburgh, Pittsburgh, PA, USA7 Department of Medicine, School of Medicine, University of Pittsburgh, Pittsburgh, PA, USA8 Department of Environmental and Occupational Health, Graduate School of Public Health, University of Pittsburgh, Pittsburgh, PA, USA2005 14 6 2005 6 26 26 3 2 2005 14 6 2005 Copyright © 2005 Latimer et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Screening mammography has had a positive impact on breast cancer mortality but cannot detect all breast tumors. In a small study, we confirmed that low power magnetic resonance imaging (MRI) could identify mammographically undetectable tumors by applying it to a high risk population. Tumors detected by this new technology could have unique etiologies and/or presentations, and may represent an increasing proportion of clinical practice as new screening methods are validated and applied. A very important aspect of this etiology is genomic instability, which is associated with the loss of activity of the breast cancer-predisposing genes BRCA1 and BRCA2. In sporadic breast cancer, however, there is evidence for the involvement of a different pathway of DNA repair, nucleotide excision repair (NER), which remediates lesions that cause a distortion of the DNA helix, including DNA cross-links.
Case presentation
We describe a breast cancer patient with a mammographically undetectable stage I tumor identified in our MRI screening study. She was originally considered to be at high risk due to the familial occurrence of breast and other types of cancer, and after diagnosis was confirmed as a carrier of a Q1200X mutation in the BRCA1 gene. In vitro analysis of her normal breast tissue showed no differences in growth rate or differentiation potential from disease-free controls. Analysis of cultured blood lymphocyte and breast epithelial cell samples with the unscheduled DNA synthesis (UDS) assay revealed no deficiency in NER.
Conclusion
As new breast cancer screening methods become available and cost effective, patients such as this one will constitute an increasing proportion of the incident population, so it is important to determine whether they differ from current patients in any clinically important ways. Despite her status as a BRCA1 mutation carrier, and her mammographically dense breast tissue, we did not find increased cell proliferation or deficient differentiation potential in breast epithelial cells from this patient which might have contributed to her cancer susceptibility. Although NER deficiency has been demonstrated repeatedly in blood samples from sporadic breast cancer patients, analysis of blood cultured lymphocytes and breast epithelial cells for this patient proves definitively that heterozygosity for inactivation of BRCA1 does not intrinsically confer this type of genetic instability. These data suggest that the mechanism of genomic instability driving the carcinogenic process may be fundamentally different in hereditary and sporadic breast cancer, resulting in different genotoxic susceptibilities, oncogene mutations, and a different molecular pathogenesis.
==== Body
Background
A reduction in breast cancer mortality has been observed in recent years that has been partially attributed to the widespread adoption of screening mammography [1]. Traditional screening mammography, however, fails to detect 15% of incident cancers [2]. New, complementary imaging techniques are therefore under development that may increase the accuracy of primary screening. We performed a small study to validate the use of low power magnetic resonance imaging (MRI) to prospectively detect breast alterations and malignancy and to determine the feasibility of applying this technique to a high-risk population [3]. We present here a subject from that study whose early stage tumor was not detectable by mammography.
This patient was enrolled in the screening study due to her family history of breast and other neoplasias. After tumor diagnosis, she was determined to be heterozygous for a putative inactivating mutation in the BRCA1 gene. In addition, she had dense breast tissue, an impediment to mammography that is in itself a risk factor for breast cancer [4]. Breast development and lactational differentiation also appear to individually modify breast cancer risk, with early term pregnancy conferring a persistent protective effect [5]. Exposure to ionizing radiation, while a lifetime risk factor for breast cancer, appears to be more dangerous when it occurs during alveolar differentiation of the breast at adolescence [6]. Using a novel tissue engineering system [7], we therefore examined the growth and differentiation of normal breast epithelial samples from this patient via live-cell imaging.
The BRCA1 hereditary breast cancer gene has been shown to be involved in DNA double strand break repair [8,9]. DNA repair defects have also been identified in the peripheral blood cells of sporadic breast cancer patients [10-13], but, in this case, it seems to involve a different pathway of DNA repair, nucleotide excision repair (NER) [14-16]. We have extended this observation of NER deficiency to the tumor itself, as well as the adjoining non-diseased normal breast tissue [17]. NER is a complex pathway of DNA repair [18] normally associated with removal of pyrimidine-pyrimidine intrastrand crosslinks (“dimers”) caused by exposure to UV light. NER deficiency is the basis of hereditary xeroderma pigmentosum (XP) [19], a disease with a 1200-fold increase in incidence of skin cancer [20]. The signal for activation of the NER pathway is actually very general; any lesion causing a distortion in the DNA helix, including crosslinks caused by oxidative radicals, certain types of mismatches (purine-purine or pyrimidine-pyrimidine) and so called “bulky” adducts caused by phase I metabolism of polyaromatic hydrocarbons [21]. It has recently been shown that BRCA1 expression can enhance NER activity, although this analysis was not performed in breast cells [22,23]. We therefore applied the functional unscheduled DNA synthesis (UDS) assay for NER capacity to multiple samples of normal tissue from this patient, to determine whether haploinsufficiency for BRCA1 was a mechanism of NER deficiency. We have developed a method to reliably culture non-diseased breast tissue (with a success rate of 100%) and breast tumors (with a success rate of 85%) [7,17].
Case presentation
We describe a breast cancer patient whose tumor was detected by MRI. She was enrolled into a pilot screening study of low power MRI due to her familial risk. She had mammographically dense breasts and her tumor was undetectable mammographically.
Patient description
The patient was a 35.7 year old woman who presented with a very strong family history of breast cancer as depicted in Figure 1, and negative physical and mammographic examination. She had extremely dense breast tissue bilaterally by mammography as well as fibrocystic breast tissue by physical examination. She had no previous personal history of breast biopsy or abnormal mammograms.
Risk profile
The 5 year breast cancer risk for this patient as calculated by the BRCAPRO model was 5.7%, and her probability of being a BRCA1 or BRCA2 carrier was 0.47. The Gail model risk assessment was calculated using the following information: Race-Caucasian; Age-35; Age at first menses-12; Age at first live birth-nulliparous; Number of first-degree relatives with breast cancer-2; Number of previous breast biopsies-0. The calculated 5 year Gail risk was 1.0% and her lifetime risk was 31.3%.
Genetic testing
Following genetic counseling, the patient elected to undergo DNA sequencing of the BRCA1 and BRCA2 genes, which revealed a Q1200X truncation mutation in one of her BRCA1 alleles. The C to T mutation at codon 1200 in exon 11 results in the change of the amino acid glutamine to a stop codon with resulting protein truncation and loss of function. Exon 11 is the largest exon in BRCA1 and has the highest frequency of reported mutations. The Q1200X mutation has been independently observed several times [24].
Imaging
The bilateral screening mammogram was compared to previous films from another hospital. The breast tissue was described as heterogeneously dense, thus lowering the sensitivity. There were no masses, significant calcifications or other findings and the mammogram was interpreted as negative bilaterally. A one-year follow-up was recommended.
The patient was then MRI scanned as previously described [3], with pre- and post-gadolinium enhancement images evaluating both breasts simultaneously in the axial plane. In the upper-outer left breast there was a small (approximately 1 cm), round, well-demarcated enhancing lesion. This lesion was seen on both the initial delay after contrast injection and the delayed contrast enhanced subtraction images. The lesion appeared to accumulate contrast to a greater extent on the delayed subtraction images with an additional lesion adjacent to the first. In the medial aspect of the mid right breast, there were several small punctate areas of enhancement on both the immediate and delayed subtraction views. Also in the right breast just above the nipple level medial and close to the chest wall an additional enhancing lesion was seen. This lesion was approximately 1.5 cm, round, well-demarcated and continued to accumulate contrast on the delayed subtraction images. This lesion appeared to have a small non-enhancing septation.
Core biopsies
Under ultrasound, the lesion of concern in the left breast was identified and biopsied, as well as one lesion in the right breast (Figure 2). The core biopsy of the left breast revealed infiltrating ductal carcinoma in 2 of 5 core fragments; high nuclear grade, with no lymphatic invasion seen. The core biopsy of the right breast demonstrated benign pathology, specifically, fibrosis with focal ductal epithelial hyperplasia.
Final pathology, treatment plan and outcome
Although a surgical candidate for lumpectomy and radiation, the patient chose to undergo left modified radical mastectomy with left axillary lymph node dissection and contralateral prophylactic total mastectomy because of her genetic risk status. The pathology in the left breast was consistent with the imaging and core biopsy in size and description. Tumor size was 8 mm in greatest dimension, nuclear grade III, ER/PR and Her2/neu negative, and the nodal status (0/4) was negative (stage TIaN0M0). The patient underwent 4 cycles of chemotherapy and has been reportedly healthy since. Because of the positive BRCA1 mutation results, she subsequently underwent prophylactic bilateral salpingo-oophorectomy.
Live-cell analysis of tissue explant cultures
A number of life history factors have been associated with breast cancer incidence that are widely interpreted as representing lifetime exposure of the breast tissue to estrogen-induced mitogenesis [25]. An alternative interpretation, based on epithelial cell differentiation, suggests that lactational differentiation, such as occurs during term pregnancy, confers resistance to carcinogenesis [26,27]. We have developed a novel human mammary epithelial (HME) tissue engineering system wherein many aspects of organotypic differentiation are reiterated in vitro [28]. In this system, breast epithelial cells initially retain cell-to-cell contact while they proliferate, then undergo an architectural reorganization, first to form three-dimensional mammospheres, and later vast networks of branching ductal and lobular structures. Tumor and some pre-neoplastic samples fail to form such architecture. Normal tissue from this patient, who is both a BRCA1 mutation carrier and has dense breasts, was evaluated to determine whether either of these factors affected de novo differentiation in this system. Four discrete pieces of fresh tissue were provided for live-cell analysis from each of the patient's ipsilateral and contralateral breasts. In the case of the ipsilateral breast, this tissue was provided at increasing distance from the tumor margin in 1 cm increments. All of these normal samples attached and grew in our culture system and were examined for cell-to-cell interactions and morphology over a period of one month. In the context of breast reduction explant cultures from 22 patients with no breast disease, these patient samples manifested typical mixtures of fibroblastic and epithelial cells. After several days in culture without passaging, the epithelial cells began to self-organize, initially forming three-dimensional mammospheres (Figure 3A), and, after 2 weeks in culture, more complex pre-ductal linear columns of epithelial cells (Figure 3B). The tissue explants from both breasts showed similar patterns of behavior (Figure 3). Tissue cultured from a contemporaneous disease-free control and the contralateral breast of a sporadic breast cancer patient showed similar morphology and architecture (data not shown).
Cell growth kinetics
It has been suggested that the association between breast density and risk of breast cancer is due to increased cell proliferation [29]. One measure of cell growth and viability is the S-phase index (SPI) or the percentage of cells incorporating radiolabeled thymidine over a specific incubation period (in our case, 2 hours). In a previous study with 22 normal breast reduction epithelium [BRE] cultures we observed a wide range of proliferation rates, with SPI ranging from a low of 0.2% to a high of 46.0% (mean of 18.3 ± 2.6%) [30]. The contemporaneous control sample from a disease-free breast reduction patient had an SPI of 30.9%, at the higher end of this normal range. The ipsilateral and contralateral tissue samples from the hereditary breast cancer patient exhibited SPI of 26.6% and 26.2%, respectively, placing them at slightly over the 70th percentile for growth rate. The contralateral sample from the sporadic breast cancer patient had an SPI of 17.0%, placing it slightly under the 50th percentile. Thus, all of these breast cancer patient samples appeared to grow well in our system, with SPI well within the range of our normal samples. The similarity of the SPI values from the two samples from the BRCA1 mutation carrier does not appear to be accidental; the chances of selecting two samples from the normal population with values as close or closer is very small (P = 0.026).
Functional analysis of NER capacity
Peripheral blood lymphocytes and normal breast epithelial tissue from the hereditary cancer patient were then cultured for performance of the functional UDS assay, which requires living cells for radiolabel incorporation during DNA repair synthesis following UV exposure. This assay is diagnostic for the inherited cancer-prone disease XP, where it is usually performed in lymphocytes or skin fibroblasts. Our novel HME tissue engineering system allows us to apply the assay to breast epithelial cells, and we have previously demonstrated tissue-specificity in the NER capacity of these cells in normal samples from patients undergoing breast reduction mammoplasty [30]. Patient data is therefore expressed relative to the average of our breast reduction controls.
Analysis of cultured blood lymphocytes from the patient established that they had normal NER capacity (99.6% of the average of our 33 normal samples) (Figure 4). This is well above the cut-off established in our sporadic breast cancer population, < 70% average normal activity, which when applied to our cases and controls yielded a significant odds ratio of 37.4 [31]. A trend towards age dependence had been noted in the analysis of the UDS data of the normal controls (P = 0.059) [30]; addition of the patient sample supports this trend, but it still fails to reach significance (P = 0.056).
The functional NER assay was then applied to the contemporaneous disease-free breast reduction control sample, one sample each from the ipsilateral and contralateral breasts of the patient, and to a sample from the contralateral breast of an apparently sporadic breast cancer patient. The NER of the BRE non-diseased control was 1.82 times the average of our normal data set for this tissue and within the range of normal. The NER capacity of the ipsilateral breast epithelial sample was 1.05 times the average of our population of BRE controls, clearly exhibiting no overt DNA repair deficiency (Figure 5). The contralateral sample was very similar, with an NER capacity of 1.17 times BRE normal. Although the NER values of these two samples from the same patient are similar, they are not close enough to distinguish themselves as coming from the same individual (P = 0.16). The NER capacity of the contralateral sample from the sporadic breast cancer patient was 1.62 times the average of the BRE controls, also in the normal range.
Our earlier analysis of NER in our normal population revealed no effects of age or cell proliferation (as represented by the SPI). All of these additional patient samples are consistent with those results.
Discussion
At least two types of breast tumors are not accurately detected by traditional screening mammography: "interval" tumors that arise quickly between screenings, and tumors whose density is not sufficient to distinguish them from the surrounding normal tissue. The latter situation is more likely to occur in women with dense normal breast tissue, which, in turn, is more typical of younger women. Thus, mammographically undetectable tumors may have a number of characteristics, such as fast growth, low density, early onset and/or occurrence in dense breasts that might distinguish them from mammographically detectable tumors in terms of molecular etiology and clinical parameters of prognosis and response. The present patient had an early onset breast tumor, but had both hereditary susceptibility due to her BRCA1 mutation and dense breasts, so her presentation is not unusual in this context. It is possible that breast tumors detected by complementary screening methods in the future will demonstrate unique clinical and molecular features, when it becomes feasible to perform such screening in the general population.
Since the BRCA1 gene product is known to play a role in DNA double strand break repair [8,9], it has been suggested that decreased repair capacity is the basis of the breast cancer predisposition observed in mutation carriers [32-35]. Such a cellular phenotype has been difficult to demonstrate, however [36-39]. An alternate possibility is that the mutation affects the growth or differentiation of breast epithelial cells in a manner consistent with cancer susceptibility. It has been suggested that dense breast tissue is indicative of generalized hyperproliferation that might promote oncogenesis [29]. Our findings show that all 8 samples, derived from both the involved and the uninvolved breasts of a hereditary breast cancer patient develop normal epithelial architecture in vitro, implying that the epithelial/stromal (paracrine) interactions necessary for the development of this complex architecture are intact and normal in BRCA1 heterozygotes despite their greater risk of breast cancer. The SPI results also indicate that this non-diseased epithelial tissue falls into the typical range of normal for BRE control cultures and is demonstrating typical growth in our HME tissue engineering system.
NER deficiency is most often associated with XP, sensitivity to UV-induced DNA damage and skin cancer [18-21]. The NER deficiency of XP patients is manifested in other tissues, however, as shown by their high spontaneous frequency of mutation in blood lymphocytes [40] and the occurrence of other types of tumors [41]. The observation that sporadic breast cancer patients have low levels of NER in peripheral lymphocytes suggests that sporadic breast cancer is associated with constitutively low levels of NER [14-16]. Our results from a single patient demonstrate, however, that while overexpression of BRCA1 may enhance NER [22], haploinsufficiency for this gene does not necessarily result in detectable NER deficiency. Since it is clear that genomic instability is a necessary prerequisite for the completion of the complex multi-step carcinogenic pathway(s) involved in breast cancer, a fundamental difference in the mechanisms of genomic instability arising in hereditary and sporadic breast tumors would be likely to translate into fundamentally different patterns of molecular pathogenesis that could impact on clinical management.
The relative NER capacities of tumor and normal tissue may have important practical implications. If breast tumors from hereditary patients exhibit NER deficiency similar to that observed in sporadic patients, while their normal tissues exhibit normal levels of this type of DNA repair, then the tumors would be hypersensitive to a range of chemotherapeutic drugs, including alkylating agents (cyclosphosphamide), cross-linking agents (cis-platinum) and bulky DNA adducting agents (melphalan). Individualization of chemotherapy based on some aspect of NER expression is being pursued in colon [42], testicular [43,44] and ovarian cancer [45].
Conclusion
This patient and her tumor represent the vanguard of a new population of early stage breast cancer patients that will be increasingly diagnosed as new screening technologies complementary to mammography are validated and become practicable. We have shown that low power MRI can detect a stage I tumor in dense breast tissue; the same technology can also impact upon interval tumors by staggering the procedure with mammography rather than applying them coincidently. Although we did not observe obvious differences in the growth rate or differentiation potential of the dense breast tissue from this patient, we cannot rule out the possibility that some or all of the tumors detectable only by complementary screening procedures will differ from the present clinical experience in important ways. Our live-cell analysis takes a step toward defining cellular characteristics that may be useful for cancer risk assessment, but we are only beginning to investigate the possibilities of the system. It may be that different growth conditions, or induction with genotoxic or estrogenic agents, will allow for the greater differentiation of breast tissue and tumor behaviours. This technique also allows for the application of functional assays to patient samples, as exemplified in this report by the UDS assay for NER capacity. Those UDS results, although from a single patient, demonstrate definitively that the constitutively low NER capacities reported in several sporadic breast populations do not arise as a pleiomorphic effect of BRCA1 haploinsufficency. Thus, the basis of genetic instability, a fundamental element in breast carcinogenesis, may differ between sporadic and hereditary breast tumors. This results in different susceptibilities to inducing agents, mutations in different sets of oncogenes and tumor suppressor genes, and, ultimately, tumors of different molecular etiology that express different clinically relevant phenotypes.
Methods
Patients and controls
The patient was a 35.7 year old woman with strong family history of breast cancer recruited into a clinical trial of MRI screening for young woman at high risk for breast cancer with dense breast tissue [3]. Gadolinium enhancement images revealed a small 1 cm lesion in the upper-outer quadrant of the left breast, identified pathologically as an infiltrating ductal carcinoma. The patient underwent a modified radical mastectomy of the left breast and chose to also undergo a contralateral prophylactic total mastectomy. Blood and tissue were obtained for analysis with consent under Magee-Womens Hospital (of the University of Pittsburgh Medical Center) IRB # MWH-94-108.
Data from this hereditary breast cancer patient were compared to that from two additional patients as well as previously published controls. The first new control patient was a 20 year old women undergoing breast reduction mammoplasty. The second contemporaneous control patient was a 36 year old woman undergoing cosmetic surgery on her contralateral breast two years after successful lumpectomy to remove an apparently sporadic stage IIA breast tumor (2.5 cm, negative for estrogen and progesterone receptors, 13 lymph nodes negative). She had undergone standard radiotherapy and chemotherapy with adriamycin and cyclophosphamide. Histopathological analysis confirmed that the breast tissue from both of these control patients was free of cancer and within the acceptable histological range of normal.
Patient tissue culture and analysis
Fresh tissues from the patient were obtained within 5 hours of surgery. After pathological evaluation, excess tissue not needed for diagnosis was placed into DMEM containing 10% fetal calf serum and 3x antibiotic antimycotic (Sigma, St. Louis, MO) at 4°C. This tissue was then processed as described in Latimer et al. [30] and placed into culture on a diluted form of matrigel (1:1 with DMEM) in the novel MWRIα medium [7].
Eight samples of the principal patient's tissue were obtained for culture after bilateral mastectomy surgery. We were not able to obtain a sample of her tumor, because it was utilized entirely for clinical diagnosis. We were able to obtain 4 pieces of histologically normal non-tumor adjacent tissue at increasing 1 cm intervals from the tumor margin from her left (ipsilateral) breast. In addition, we obtained 4 similar pieces of fresh tissue from her contralateral breast. All were placed into primary explant (HME) culture.
For analysis of cell growth and in vitro differentiation, explants were cultured and imaged every second day using a digital Hamamatsu Orca camera for 30–60 days. Images were analyzed on a Macintosh G4 computer using QED imaging software (Media Cybernetics, Inc., Silver Spring, MD).
Control tissue cultures
Breast reduction mammoplasty tissues were obtained from patients ages 20–70 at Magee-Womens Hospital under the above IRB. A neighboring piece of mammoplasty tissue (from the same 0.25 cm2 sample) to that placed into primary culture was fixed and processed in paraffin. These sections were examined by a pathologist to verify the histological features and normality of the tissue. Breast tissue was processed as previously described [30]. Tissue was rinsed three times in PBS containing antibiotics, disaggregated and placed into MWRIα medium [7] on a thin coat of matrigel. Peripheral blood lymphocytes (PBLs) were obtained with consent from normal healthy control subjects ages 20–50 working at Magee-Womens Hospital or students at the University of Pittsburgh. Foreskin fibroblast (FF) tissue was obtained as discarded tissue from newborn infants after circumcision and utilized between passages 7 and 10. These control populations have been previously described in greater detail [30,46]. Breast tissue samples from the two new control patients were processed in the same manner.
Analysis of S-phase indices
Primary cultures of mammary tissue, established 10–14 days, were labeled with 3H-thymidine for a period of 2 hours followed by a chase with cold thymidine for 2 hours and then processed for autoradiography. After a 10–12 day exposure, slides were processed and analyzed by two independent, blinded scorers who evaluated the tissue samples for the percentage of cells in S phase (characterized by complete coverage of the nucleus with silver grains).
Unscheduled DNA synthesis
NER was measured by autoradiography of unscheduled DNA synthesis after UV damage (UDS) [47,48]. After a total of 10–14 days in culture, without passaging, cultures were irradiated with UV light at 254 nm at a mean fluence of 1.2 Joules/m2 for 12 seconds in the absence of culture medium, for a total dose of 14 J/m2. Each sample was represented by at least two chamber slides. One chamber of each 2-chamber slide was shielded from the UV dose to be used as an unirradiated control sample. Primary cultures had not reached confluence and were still actively growing at the time the UDS assay was performed. Control FF were plated subconfluently 2 days before the UDS assay to insure that they also were not in a quiescent state brought on by confluence. After UV exposure, all cultures were incubated in medium supplemented with 10 μCi ml [3H]methyl-thymidine (~80 Ci mmol-1) (PerkinElmer Life Sciences, Boston, MA) for 2 hours at 37°C. Labeling medium was then replaced with unlabeled chasing medium containing 10-3 M non-radioactive thymidine (Sigma) and incubated for a further 2 hours to clear radioactive label from the intracellular nucleotide pools. After incubation in the post-labeling medium, cells were fixed in 1X SSC, 33% acetic acid in ethanol, followed by 70% ethanol and finally rinsed in 4% perchloric acid over night at 4°C. All slides were dried and subsequently dipped in photographic emulsion (Kodak type NTB2) and exposed for 10 to 14 days in complete darkness at 4°C.
The length of exposure of emulsion was determined in each experiment by preparing FF "tester" slides. After 10–12 days these tester slides were developed and grain counting was performed. If the nuclei over the foreskin fibroblasts averaged 50 or more grains per nucleus, then the rest of the experimental slides were developed. If the grain count was below this level, the remaining slides were left to expose 1–3 days longer before being developed.
Grain counting
After photographic development of emulsion, all slides were stained with Giemsa, then examined at a total magnification of 1000X on a Zeiss Axioskop under oil emersion for grains located immediately over the nuclei of non-S phase cells [48]. Local background grain counts were evaluated in each microscopic field over an area the same size as a representative nucleus, and this total was subtracted from the grain count of each nucleus in that field. The average number of grains per nucleus was quantified for each side of the chamber slide, both unirradiated and irradiated. The final NER value for each slide was calculated by subtracting the unirradiated mean (grains per nucleus) from the irradiated mean (grains per nucleus), after the initial subtraction of local background in each field. NER was initially expressed as a percentage of the activity of concurrently analyzed FF. Four FF slides were scored per experiment, by an average of three counters. 200 nuclei were counted per slide, for a total of 800, with an average of 61.6 grains/nucleus. Six slides were evaluated for the patient's PBL sample, two by each of three counters. An average of 195 nuclei were scored per slide (for a total of almost 1200), with an average of 7.5 grains/nucleus. Four slides were counted for the contemporaneous breast reduction control, two each by two counters. There were an average of 200 nuclei per slide and 14.1 grains/nucleus. Six slides were scored from the patient's ipsilateral breast tissue sample, two by each of three independent counters, and five slides were counted from the contralateral sample, again by three independent counters. An average of just over 100 nuclei were evaluated per slide for each sample, for a total of almost 600 nuclei for the ipsilateral sample and over 500 for the contralateral sample. As the NER capacities indicate, these samples had very similar counts; about 35 grains/nucleus for the ipsilateral sample and 28 grains/nucleus for the contralateral sample. Finally, four slides were counted from the contralateral sample of a sporadic breast cancer patient, by three counters. There were an average of 200 nuclei per slide and 29.4 grains per nucleus.
Statistical analysis
To ensure accuracy and guard against transcription errors, raw grain counts from the UDS assay were processed independently in duplicate, once using StatView (version 5.0.1, SAS Institute, Inc., Cary, NC), and once using the Data Analysis Toolpack of the Excel 2001 spreadsheet program (Microsoft Corp., Redmond, WA). The final count from slides of the same cell type within the same experiment and developed the same day were averaged together and expressed as a percentage of concurrently analyzed FF. These results were then normalized by comparison to the average for the tissue type control population [48].
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
JJL conceived of the study, executed it, and drafted the manuscript. WSR recruited and consented the patient, provided clinical samples and information. JMJ evaluated the UDS assay and analyzed the data. AKS performed the histopathological analysis of the tissue. VGV participated in the study design and data interpretation. SGG participated in the design and coordination of the study and helped to draft the manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
This study was supported in part by NIH grant CA 71894, US Army BRCP grants DAMD17-00-1-0681, BC033717, BC991187, DAMD17-00-1-0409, grant BCTR0403329 from the Susan G. Komen Breast Cancer Foundation and grants from the Ruth Estrin Goldberg Foundation and the Pennsylvania Department of Health. We would like to thank our clinical collaborators on this project, Dr. Jules H. Sumkin for his cooperation with this study, and acknowledge the work of our clinical coordinator, Michelle B. Huerbin. We greatly appreciate the technical contributions of Melissa C. Paglia, Shail B. Mehta, Christina M. Cerceo, Crystal M. Kelly, Julie A. Conte, Janiene A. Patterson, Ayodola B. Anise and Lynn R. Janczukiewicz to this study.
Figures and Tables
Figure 1 Pedigree of the patient (indicated by arrow). She, one maternal aunt and one maternal cousin had breast cancer diagnosed at 36, 44 and 41 years old, respectively, as indicated by the half-filled symbols, and her aunt died of the disease. Her cousin underwent lumpectomy followed by chemotherapy, radiotherapy and is presently on tamoxifen. Her mother had breast cancer in both breasts, diagnosed at ages 41 and 42, as indicated by the completely filled symbol. She underwent bilateral mastectomy and hysterectomy followed by chemotherapy and radiotherapy and died of the disease at age 44. A second maternal aunt was diagnosed with colon cancer at age 52 (light half-filled symbol) and breast cancer at age 55 (dark half-filled symbol). Based on this pattern of familial cancer the patient was considered to be at high risk of developing breast cancer and was entered into the low power MRI screening validation and feasibility study. Following her diagnosis, she was confirmed as carrying a Q1200X mutation in the BRCA1 gene.
Figure 2 Ultrasound of the MRI-detected lesion. Following MRI, the patient was scheduled for ultrasound to identify the questionable lesions seen on MRI for possible core biopsy. Under ultrasound the lesion of concern was identified and biopsied at the 1:00 location in the left breast. Additionally, one lesion seen by MRI in the right breast at the 4:00 location was identified and biopsied.
Figure 3 Micrographs of the non-diseased primary human mammary epithelial cultures (HMEC) from the BRCA1 mutation carrier. A) Contralateral breast – A cluster of epithelial cells called a mammosphere is shown on the left center of the image sitting on a field of fibroblasts. B) Ipsilateral breast – The original fresh tissue block from which this culture was derived was located 4 cm from the infiltrating ductal carcinoma. The structure shown is a cluster of rounded epithelial cells manifesting a column configuration called "pre-ductal linearization". Both images were captured under Differential Interference Contrast (DIC) optics on a Zeiss Axiovert 100 microscope at a total of 140x magnification.
Figure 4 Comparison of the NER capacity of a PBL sample from our BRCA1 mutation carrier patient with those of a population of disease-free controls. The dark horizontal line indicates the average for the normal population, while the dotted lines indicate upper limits for residual NER activity in patients with the hereditary NER deficiency disease XP (0.50) and the cut-off established in our breast tissue study that identified tumors with high sensitivity and specificity (0.70).
Figure 5 Comparison of the NER capacities of two samples of normal breast epithelium from our BRCA1 mutation carrier patient with those of a population of disease-free controls who underwent breast reduction mammoplasty. The dark horizontal line indicates the average for the normal population of breast reduction epithelium (BRE), while the dotted lines indicate upper limits for residual NER activity in patients with the hereditary NER deficiency disease XP (0.50) and the cut-off established in our breast tissue study that identified tumors with high sensitivity and specificity (0.70). The patient sample on the left was derived from the ipsilateral (left) breast, while the sample on the right was from the contralateral (right) breast.
==== Refs
Institute of Medicine Mammography and Beyond: Developing Technologies for the Early Detection of Breast Cancer 2001 Washington DC: National Academy Press
Mushlin AI Kouides RW Shapiro DE Estimating the accuracy of screening mammography: a meta-analysis Am J Prev Med 1998 14 143 153 9631167 10.1016/S0749-3797(97)00019-6
Rubinstein W Vogel VG Sumkin JH Huerbin MB Grant SG Latimer JJ Prospective sreening study of 0.5 Tesla dedicated magnetic resonance imaging for the detection of breast cancer in young, high risk women BMC Women's Health
Byng JW Yaffe MJ Jong RA Shumak RS Lockwood GA Tritchler DL Boyd NF Analysis of mammographic density and breast cancer risk from digitized mammograms Radiographics 1998 18 1587 1598 9821201
Lambe M Hsieh C-C Tsaih SW Ekbom A Trichopoulos D Adami HO Parity, age at first birth and the risk of carcinoma in situ of the breast Int J Cancer 1998 77 330 332 9663590 10.1002/(SICI)1097-0215(19980729)77:3<330::AID-IJC3>3.0.CO;2-P
Hancock SL Tucker MA Hoppe RT Breast cancer after treatment of Hodgkin's disease J Natl Cancer Inst 1993 85 25 31 8416252
Latimer JJ Epithelial Cell Cultures Useful For in Vitro Testing US Patent 2000 #6,074,874
Jasin M Homologous repair of DNA damage and tumorigenesis: the BRCA connection Oncogene 2002 21 8981 8993 12483514 10.1038/sj.onc.1206176
Rosen EM Fan S Pestell RG Goldberg ID BRCA1 gene in breast cancer J Cell Physiol 2003 196 19 41 12767038 10.1002/jcp.10257
Jaloszynski P Kujawski M Czub-Swierczek M Markowska J Szyfter K Bleomycin-induced DNA damage and its removal in lymphocytes of breast cancer patients studied by comet assay Mutat Res 1997 85 223 233
Nascimento PA da Silva MA Oliveira EM Suzuki MF Okazaki K Evaluation of radioinduced damage and repair capacity in blood lymphocytes of breast cancer patients Braz J Med Biol Res 2001 34 165 176 11175491 10.1590/S0100-879X2001000200003
Colleu-Durel S Guitton N Nourgalieva K Legue F Leveque J Danic B Chenal C Genomic instability and breast cancer Oncol Rep 2001 8 1001 1005 11496305
Smith TR Miller MS Lohman KK Case LD Hu JJ DNA damage and breast cancer risk Carcinogenesis 2003 24 883 889 12771032 10.1093/carcin/bgg037
Kovacs E Stucki D Weber W Muller H Impaired DNA-repair synthesis in lymphocytes of breast cancer patients Eur J Cancer Clin Oncol 1986 22 863 869 3770043 10.1016/0277-5379(86)90375-5
Xiong P Bondy ML Li D Shen H Wang LE Singletary SE Spitz MR Wei Q Sensitivity to benzo[a]pyrene diol-epoxide associated with risk of breast cancer in young women and modulation by glutathione S-transferase polymorphisms: a case-control study Cancer Res 2001 61 8465 8469 11731429
Kennedy DO Agrawal M Shen J Terry MB Zhang FF Senie RT Motykiewicz G Santella RM DNA repair capacity of lymphoblastoid cell lines from sisters discordant for breast cancer J Natl Cancer Inst 2005 97 127 132 15657342
Latimer JJ Kisin E Zayas-Rivera B Kelley J Johnson R Grant SG Increased somatic mutation and reduced DNA repair in breast cancer patients and their tumors [abstract] Proc Am Assoc Cancer Res 1999 40 440
Mullenders LHF Berneberg M Photoimmunology and nucleotide excision repair: impact of transcription coupled and global genome excision repair J Photochem Photobiol B 2001 56 97 100 10.1016/S1011-1344(01)00244-5
Cleaver JE Defective repair replication of DNA in xeroderma pigmentosum Nature 1968 218 652 656 5655953
Luddy RE Sweren RJ Altman AJ, Schwartz AD Skin cancer Malignant Diseases of Infancy, Childhood and Adolescence 1983 2 Philadelphia: Saunders 552 559
Thompson LH Nickoloff JA, Hoekstra MF. Totowa, NJ Nucleotide excision repair: its relation to human disease DNA Damage and Repair, Volume 2: DNA Repair in Higher Eukaryotes 1998 Humana Press 335 393
Hartman A-R Ford JM BRCA1 induces DNA damage recognition factors and enhances nucleotide excision repair Nat Genet 2002 32 180 184 12195423 10.1038/ng953
Takimoto R MacLachlan TK Dicker DT Niitsu Y Mori T el-Deiry WS BRCA1 transcriptionally regulates damaged DNA binding protein (DDB2) in the DNA repair response following UV-irradiation Cancer Biol Ther 2002 1 177 186 comment 1:187–188 12170778
Breast Cancer Information Core
Gierach G Vogel VG Singletary SE, Robb GL, Hortobagyi GN Epidemiology of breast cancer Advanced Therapy of Breast Disease 2004 2 Hamilton: BC Decker 58 74
Russo IH Koszalka M Gimotty PA Russo J Protective effect of chorionic gonadotropin on DMBA-induced mammary carcinogenesis Br J Cancer 1990 62 243 247 2117463
Russo J Mailo D Hu YF Balogh G Sheriff F Russo IH Breast differentiation and its implication in cancer prevention Clin Cancer Res 2005 11 931s 936s 15701889
Latimer JJ Kanbour-Shakir A Giuliano K Kisin E Kelley J Johnson R Live cell microscopy of long-lived primary human mammary epithelium cultures (HMEC) from breast reduction mammoplasties and breast tumors [abstract] Breast Cancer Res Treat 1998 50 255
Boyd NF Lockwood GA Martin LJ Knight JA Byng JW Yaffe MJ Tritchler DL Mammographic densities and breast cancer risk Breast Dis 1998 10 113 126 15687568
Latimer JJ Nazir T Flowers LC Forlenza MJ Beaudry-Rodgers K Kelly CM Conte JA Shestak K Kanbour-Shakir A Grant SG Unique tissue-specific level of DNA nucleotide excision repair in primary human mammary epithelial cultures Exp Cell Res 2003 291 111 121 14597413 10.1016/S0014-4827(03)00368-9
Kelly CM Johnson JM Wenger SL Vogel V Kelley J Johnson R Amortequi A Mock L Grant SG Latimer JJ Analysis of functional DNA repair in primary cultures of the non-tumor adjacent breast identifies two classes of breast cancer patient [abstract] Proc Am Assoc Cancer Res 2003 44 974 975
Buchholz TA Wu X Hussain A Tucker SL Mills GB Haffty B Bergh S Story M Geara FB Brock WA Evidence of haplotype insufficiency in human cells containing a germline mutation in BRCA1 or BRCA2 Int J Cancer 2002 97 557 561 11807777 10.1002/ijc.10109
Mamon H Dahlberg W Little JB Hemizygous fibroblast cell strains established from patients with BRCA1 or BRCA2 mutations demonstrate an increased rate of spontaneous mutations and increased radiosensitivity Int J Radiat Oncol Biol Phys 2003 57 S346 S347 10.1016/S0360-3016(03)01240-9
Rothfuss A Schutz P Bochum S Volm T Eberhardt E Kreienberg R Vogel W Speit G Induced micronucleus frequencies in peripheral lymphocytes as a screening test for carriers of a BRCA1 mutation in breast cancer families Cancer Res 2000 60 390 394 10667592
Das R Grant SG Unusual pattern of age-related somatic mutation in cell lines derived from carriers of mutations in the BRCA1 and 2 genes [abstract] Am J Hum Genet 2002 71 251
Nieuwenhuis B Van Assen-Bolt AJ Van Waarde-Verhagen MA Sijmons RH Van der Hout AH Bauch T Streffer C Kampinga HH BRCA1 and BRCA2 heterozygosity and repair of X-ray-induced DNA damage Int J Radiat Biol 2002 78 285 295 12020440 10.1080/09553000110097974
Trenz K Rothfuss A Schutz P Speit G Mutagen sensitivity of peripheral blood from women carrying a BRCA1 or BRCA2 mutation Mutat Res 2002 500 89 96 11890937
Baeyens A Thierens H Claes K Poppe B de Ridder L Vral A Chromosomal radiosensitivity in BRCA1 and BRCA2 mutation carriers Int J Radiat Biol 2004 80 745 756 15799620 10.1080/09553000400017937
Trenz K Schutz P Speit G Radiosensitivity of lymphoblastoid cell lines with a heterozygous BRCA1 mutation is not detected by the comet assay and pulsed field gel electrophoresis Mutagenesis 2005 20 131 137 2005, Mar 22; Epub ahead of print 15784691 10.1093/mutage/gei018
Norris PG Limb GA Hamblin AS Lehmann AR Arlett CF Cole J Waugh AP Hawk JL Immune function, mutant frequency, and cancer risk in the DNA repair defective genodermatoses xeroderma pigmentosum, Cockayne's syndrome, and trichothiodystrophy J Invest Dermatol 1990 94 94 100 2295840 10.1111/1523-1747.ep12873952
Kramer KH Lee MM Scotto J DNA repair protects against cutaneous and internal neoplasia: evidence from xeroderma pigmentosum Carcinogenesis 1984 5 511 514 6705149
Arnould S Hennebelle I Canal P Bugat R Guichard S Cellular determinants of oxaliplatin sensitivity in colon cancer cell lines Eur J Cancer 2003 39 112 119 12504667 10.1016/S0959-8049(02)00411-2
Koberle B Grimaldi KA Sunters A Hartley JA Kelland LR Masters JR DNA repair capacity and cisplatin sensitivity of human testis tumor cells Int J Cancer 1997 70 551 555 9052754 10.1002/(SICI)1097-0215(19970304)70:5<551::AID-IJC10>3.0.CO;2-G
Koberle B Masters JR Hartley JA Wood RD Defective repair of cisplatin-induced DNA damage caused by reduced XPA protein in testicular germ cell tumours Curr Biol 1999 9 273 276 10074455 10.1016/S0960-9822(99)80118-3
Selvakumaran M Pisarcik DA Bao R Yeung AT Hamilton TC Enhanced cisplatin cytotoxicity by disturbing the nucleotide excision repair pathway in ovarian cancer cell lines Cancer Res 2003 63 1311 1316 12649192
Forlenza MJ Latimer JJ Baum A The effects of stress on DNA repair capacity Psychol Health 2000 15 881 891
Cleaver JE Thomas GH Friedberg EC, Hanawalt PC. Measurement of unscheduled synthesis by autoradiography DNA Repair: A Laboratory Manual of Research Procedures Volume I New York: Marcel Dekker 1981:277-287
Kelly CM Latimer JJ Unscheduled DNA synthesis: a functional assay for global genomic nucleotide excision repair Meth Mol Biol 2005 291 305 320
|
15955237
|
PMC1215484
|
CC BY
|
2021-01-04 16:03:33
|
no
|
BMC Med Genet. 2005 Jun 14; 6:26
|
utf-8
|
BMC Med Genet
| 2,005 |
10.1186/1471-2350-6-26
|
oa_comm
|
==== Front
BMC Med EducBMC Medical Education1472-6920BioMed Central London 1472-6920-5-311611531210.1186/1472-6920-5-31Research ArticleEvaluation of a task-based community oriented teaching model in family medicine for undergraduate medical students in Iraq Al-Dabbagh Samim A [email protected] Waleed G [email protected] Department of Community Medicine, Mosul College of Medicine, Mosul University, Mosul, Iraq2005 22 8 2005 5 31 31 7 5 2005 22 8 2005 Copyright © 2005 Al-Dabbagh and Al-Taee; licensee BioMed Central Ltd.2005Al-Dabbagh and Al-Taee; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
The inclusion of family medicine in medical school curricula is essential for producing competent general practitioners. The aim of this study is to evaluate a task-based, community oriented teaching model of family medicine for undergraduate students in Iraqi medical schools.
Methods
An innovative training model in family medicine was developed based upon tasks regularly performed by family physicians providing health care services at the Primary Health Care Centre (PHCC) in Mosul, Iraq. Participants were medical students enrolled in their final clinical year. Students were assigned to one of two groups. The implementation group (28 students) was exposed to the experimental model and the control group (56 students) received the standard teaching curriculum. The study took place at the Mosul College of Medicine and at the Al-Hadba PHCC in Mosul, Iraq, during the academic year 1999–2000. Pre- and post-exposure evaluations comparing the intervention group with the control group were conducted using a variety of assessment tools.
Results
The primary endpoints were improvement in knowledge of family medicine and development of essential performance skills. Results showed that the implementation group experienced a significant increase in knowledge and performance skills after exposure to the model and in comparison with the control group. Assessment of the model by participating students revealed a high degree of satisfaction with the planning, organization, and implementation of the intervention activities. Students also highly rated the relevancy of the intervention for future work.
Conclusion
A model on PHCC training in family medicine is essential for all Iraqi medical schools. The model is to be implemented by various relevant departments until Departments of Family medicine are established.
==== Body
Background
More than half of newly graduating physicians will be employed as general practitioners (GP). "Family medicine" is almost synonymous with general practice and constitutes a major component of newly graduated doctors' practices. Characteristics of general practice include accessibility, availability, comprehensiveness, maintaining responsibility and long-term doctor/patient relationship [1,2]. Moreover, general practice plays a vital role as an entry point into the health care system, as a link between self-care and professional medical care and substantially influences overall level of care and use of resources in a community [2,3].
Since the Alma-Ata declaration, the concept of Primary Health Care (PHC) has broadened and encompasses both general and family practice, which refer to the specific medical services provided [4]. The PHC team is now recognized as a crucial element in the delivery of community based medical care. A further challenge is to strengthen relationship between the PHC team and other health networks within the community to achieve the Alma-Ata declaration's vision of PHC [5].
Family physicians play a major role in integrating and coordinating care provided to patients and their families. They are responsible for the implementation of the concept of PHC through their work in general practice. Therefore, a well designed and effective training program in family medicine should be an essential component of medical school curricula.
The first medical school in Iraq, Baghdad Medical College, was established in 1927 with the help of Sir Harry Sinderson, who served as its first Dean. The college adopted the Edinbrough curriculum, which reflected standard teaching curricula of the time. Iraq's second medical school, Mosul Medical College, opened in 1959. Other medical colleges were subsequently established throughout Iraq and all adopted the teaching curriculum of Baghdad Medical College, supported by the Ministry of Higher Education (MOHE), although there were no financial or political reasons to adopt the traditional curriculum. In recent years, suggestions have been put forth to adopt innovative or community based curricula, but such attempts have not received the support of clinicians in medical colleges. In 1989, the MOHE opened a new medical college in Tikrit with an innovative curriculum, supported by the World Health Organization. Despite the WHO imprimatur, a shortage of staff led to incomplete implementation of the innovative teaching model and the result was a combination of the two curricula.
Iraqi medical education is a six year programme. The first three years are devoted to basic sciences and the subsequent three years focus on clinical aspects of medical care. The final year is provides hospital based clinical training in medicine (12 weeks), surgery (12 weeks), pediatrics (10 weeks), gynecology and obstetrics (8 weeks) and community medicine (2 weeks).
There is no specified department for teaching family medicine in any of the Iraqi medical colleges, including that of Tikrit. The curricula of all medical schools in Iraq ignore the concept family medicine as a separate entity. Moreover the clinical teaching component is almost entirely hospital-based; students are not exposed to the presentation of diseases outside of a hospital setting and they are often unprepared for the complexity of general practice [6].
Community-based clinical practice is unique in that it provides real-world scenarios as well as a wide range of learning opportunities [7]. Numerous medical schools have adopted the concept of family medicine to better prepare students for the complex eventualities of general practice. The Arabian Gulf University in Bahrain, for example, has established a separate department for Family and Community Medicine. The Family Medicine Clerkship is mandatory final rotation that students must complete prior to sitting for the final qualifying examination. Another model is the United States, where 95 (out of 127) medical schools have separate departments for family medicine and/or PHC, where students spend an average of 5.7 weeks in clinical clerkships in family/community medicine during their final two academic years [8,9].
Medical educators are now focusing on the relevancy of the medical curriculum to the actual health needs of the community and on the ability of newly graduated doctors to solving common health problems. In another words, medical colleges have made significant efforts to produce competent family physicians [10,11].
Traditional medical colleges have also attempted to apply family medicine to their curriculum by expanding their curricula to include (a) multi-disciplinary approaches (b) the scientific rationale for medicine, (c) problem-solving approaches, (d) strengthening understanding of socio-cultural factors, (e) relating the curriculum to the students' needs, and (f) integrating premedical with pre-clinical medical sciences. Thus, community-oriented medical education seems to be the most progressive and effective way of introducing family medicine into the medical curriculum.
The World Health Organization has encouraged all countries to undertake activities to reform medical education and medical practice with a view to increasing relevance, quality of care, cost-effectiveness and equity in health care. In Iraq, there have been a few studies that have reviewed the teaching curriculum of medical schools during the last few years with the aim of developing a national curriculum for medical colleges with relevance to community needs [12].
The undergraduate teaching of family medicine is not a separate entity in the curriculum of Iraqi Colleges of Medicine and the Ministry of Health programs in the delivery of PHC are rely upon active general family practitioners for success. To achieve this goal, a task-based community oriented model in family medicine for Iraqi medical schools was developed.
The aim of the model was to prepare the future graduates for general practice. The main specific objectives were to train undergraduate medical students to:
• demonstrate familiarity with common health problems and their related health traditions
• provide diagnostic, therapeutic and preventive services
• develop physical, social and psychological relationships with families
• promote the general health status of individuals and their families
• activate and encourage team work in health institutions
• advocate for community participation through regional PHC committees
• use research to assess process of family medicine development
Methods
The study was conducted during the academic year 1999–2000. The methodology of the study was based on a newly proposed job description for general practitioners in Iraq, which combined the activities and responsibilities of GPs with knowledge gained in medical schools and the traditions and customs of the society [13]. The plan was to utilize the tasks and subtasks of the proposed job description to design a task-based health care oriented training model in family medicine for undergraduate medical students. The steps used in the development of the model were the following.
1. Tasks were first grouped under convenience headings and then systematically clustered into four main groups: child health services, maternal health services, common health problems and administration health care activities.
2. A training course programme at a PHCC setting was planned. Learning activities were implemented according to time and place with respect to the tasks and subtasks.
3. Plans for implementing daily activities throughout the course program were outlined. This was done to maximize the effect of the planned training course. The purpose was to create a learning experience that resembled, as much as possible, actual primary care through the introduction of multi-topic health care services. The following activities were conducted daily during the programme:
a. Reviewing the context of learning activities with regards to time, place, resources and students
b. Writing down definite learning objectives
c. Determining critical learning strategies
d. Continuous assessment
4. A sample of medical students was selected for participating in the study. Two groups of students were selected out of eight groups of medical students in their final year at Mosul College of Medicine (28–30 students comprise each administrative group). One group (28 students) was chosen to receive the experimental model (the experiment group). These students had experienced a short gap between their clinical training in the departments of community medicine, pediatrics, and gynecology and obstetric. The comparison group was composed of 2 groups (56 students) who completed their training soon after the experimental group completed their studies.
The duration of implementation of the model was four weeks of block training: two weeks of community medicine, one week of pediatrics and one week of gynecology and obstetrics. Students in the control group following their usual training. The program was to be conducted during the last semester of the academic year, when both groups have finished there usual training in three specialist departments, other than surgery.
Both authors were teaching community medicine in Al-Hadba PHCC. Other tutors were from the staff of community medicine, including psychiatrists and registrars from the departments of pediatrics, gynecology and obstetrics. Tutors were given a one week training course to illustrate the procedures of teaching and evaluation, including seminars and role play. Feed-back from the tutor training was used to modify some of the procedures of implementation of the model.
The setting for the model was the Al-Hadbaa PHCC and the delivery room in the Al-Batool teaching hospital, which is adjacent to the PHCC. Each student completed a log book for all cases and topics of the medical services involved. A pretest evaluation of the intervention group was conducted at the start of the study. A post-training evaluation was conducted for students on the experimental group, as well as for the comparison group using the same roles and criteria.
Different assessment tools were used in this study. For pre- and post-test evaluations, modified essay questions, MCQ's, case management exercises, flowcharts and oral examinations were used. A daily flowchart was given to each student to be completed as homework. The tools were specially designed forms comprising check lists and rating scales to be filled in by the investigators using direct observation. Finally, a 25 item questionnaire was submitted to the 28 trainee students in the intervention group at the end of the training course to solicit their feed-back. Questions referred to the planning of the program, relevance and utility of the working methods, method of running the program and organization attitudes, timing for activities, benefits gained and participants and program evaluation tools used.
A five point rating scale for task analysis was used (0 = not done, 1 = unsatisfactory, 2 = equivocal, 3 = satisfactory, 4 = very satisfactory). Coefficients for the 5 point rating scale were 1, 2, 3, 4 and 5. The score of each item on a question form was expressed as percentage by Guilbert formula [14]:
Where 20 = 100 divided by the maximum coefficient 5
Items related to one variable were grouped together when comparison was needed between variables consisting of many items. The score of each variable is the aggregate score of its components.
Results
Table 1 illustrates the teaching methods, assessment tools, place and duration of the task-based community oriented teaching model in family medicine for medical students. The model is designed to be implemented over four weeks for final year medical students. The setting should be a well equipped and organized PHCC. The tasks and learning objectives of the four principle functions of the model are illustrated in tables 2, 3, 4, 5.
Table 1 Task-based community care oriented teaching model in family medicine
teaching theory Classroom lectures, tutorials, discussion groups, procedure book, problem-based learning and workshops
practical Small groups discussion, observed field work, supervised field work, log book, self study activities
aids Growth chart, handouts, posters, flipcharts, video films, blackboard, overhead projector
assessment toots Log book, problem solving exercises, OSCE, patient management problems, MCQs, Short assays, checklist and rating scale, evaluation of reports and flowcharts
place Primary health care centre
day and time hours Saturday through Thursday 8–12 am
duration one week in each department for each function
Table 2 Function: administrative health care Department: community medicine
principle tasks learning objectives
- Health educational guiding advice propagation. - Prepare a health educational topic, identify opportunities for health education during routine clinical work, initiate health talk and use a suitable communication method to clarify critical health educational subjects.
- Community diagnosis through morbidity and mortality reporting. - Choose simple practical methods for vital events reporting, classify data according to scientific basis, utilize data for diagnosis of community health problems and health needs.
- Surveillance of disease. - Collect, analyze & distribute data to those responsible.
Table 3 Function: child health care department: pediatric
principle tasks learning objectives
- Child Nutrition growth & development - Trace child's growth and monitor feeding habits.
- Breast feeding - Facilitate an exclusive breast feeding practice during the 1st 4–6 months; explain its benefits and when to start giving weaning food.
- ARI control and the rational use of antibiotics - Identify dangerous criteria and the situations of using specific antibodies.
- Diarrhoeal Disease control and the use of Oral rehydration - Diagnose diarrhoeal diseases and assess dehydration. Decide which treatment plan is indicated and explain the value, and method of preparing ORS and its alternatives.
- Child protection from dangerous communicable diseases immunization - Identify type and time of giving vaccines, its contraindications and side effects, its storage, site of giving vaccine and perform properly the national schedule of vaccination to a maximum coverage-rate.
- Management of other common health problems and its related health customs and traditions - Take a brief history and conduct essential clinical examinations, do specific investigations, diagnose and manage the main health problem and communicate well to manipulate the related health customs or traditions.
- Risky child-diagnosis & follow up - Pick dangerous specific criteria, diagnose risky child, prescribe essential method of management, make a follow up schedule, and monitor indicators of improvement
- Health-educational advice propagation - Detect the actual health educational problem through proper communication, determine the essential educational needs, suggest a proper way and timing for introducing the advice.
- Screening for rheumatic heart diseases. - Auscultate pupil's heart carefully, detect any added sounds or murmur, examine the tonsils and relate the findings to past history of recurrent attacks of tonsillitis.
- Essential measures to obtain a healthy eye and vision. - Examine the eyelids and eyeball carefully, identify any disturbed vision, detect critical or risky criteria, diagnose the problem and deal with it accordingly.
- Referral services. - Determine indications for referral, including method and direction and follow up referred cases.
- Prevention and control of any outbreak of epidemic among pupils. - Notify about the case, isolate it by a suitable method, give essential treatment and sick leave needed, examine contacts, put under observation risky ones and provide essential health protective measures.
Table 4 Function: maternal health care department: obstetric and gynecology
principle tasks learning objectives
- Pre-marital & health-care services. - Take proper family, medical history, educate couples about the essentials, order for CXR, blood group, R.H & other investigations for STD.
- Take proper obstetric history, conduct properly clinical examination searching for signs of anemia hypertension, edema, urinary tract disorders, pelvic size and shape, and any fetal abnormality, do a schedule for antenatal visit and weight checking giving antitetanus toxoid at the proper time & calculate the expected date for delivery.
- Management of common gestational health problems. - Diagnose and prescribe treatment for anemia, hypertension, urinary tract infections, toxoplasmosis and diabetes, and follow up patients through out gestational period.
Risky pregnancies – diagnosis and follow up. - Detect and manage risky pregnancies and follow them up.
Delivery-health-care services. - Facilitate process of delivery to be normal uncomplicated vaginal labor and promote uterine contractions for expulsion of a head presented fetus under sterile conditions.
Essential postnatal health care services. - Examine for unevoluted uterus, bleeding & promote healthy purperium, lactation and look for breast complications including over encouragement or crackled nipples.
- Health educational advice for pregnant and lactating women. - Use proper methods of communication to explain essential nutrient materials needed and the value of attending MCH clinics periodically, taking tonic drugs, vaccination and care of breasts.
- Immunization of women during the age of child bearing. - Protect women through proper implementing of vaccination program against German measles and tetanus.
- Referral services. - Put criteria for risky patient's referral and direction, identify proper time and essential measures for referral, and follow up referred patients.
- Screening for the risk of developing breast carcinoma. - Perform a proper clinical examination of breasts and detect any lump and manage accordingly specially for those with positive family history of breast cancer.
- Follow up hydatiform-mole - Monitor women with history of hydatidiform mole, prescribe essential contraceptive, give methotrexate, and follow them up.
- Management of common health problems of lactating breast. - Diagnose and manage common breast problems e.g. crackled nipples, retracted malformed nipples, and breast abscess.
- Infertility and menstrual cycle regulation measures. - Take proper menstrual history, detect any abnormality, relate findings to failure of conception, and suggest proper treatment method.
- Family planning services. - Display a proper action for a better child spacing time, and communicate properly to justify a routine attendance of family planning clinics.
Table 5 Function: management of community health problems department: medicine
principle tasks learning objectives
- Practicing an evidence based clinical medicine - Take proper clinical history, conduct general and systemic examination, select essential investigations and perform diagnosis according to specific criteria
- Making provisional and definitive diagnosis. - Practice putting differential diagnosis and identify the most logical diagnosing criterion
- Prescription and evaluation of treatment. - Practice writing drugs prescription and explain route and method of administration and its main side effects, and evaluate its impact
- Dealing with current health traditions and customs - Choose the negative tradition, plan to alter it, use a suitable & effective method of communication.
- Prevention and control of communicable disease conditions - Display essential measures for preventing and controlling infectious cases (environmental sanitation, active and passive immunization, isolation of cases, tracing of contacts .. etc.
- Health-educational advice - Propagate essential and primitive primary health care educational guidance (specially in regard to safe water, immunization program, and MCH services).
- Building a positive relationship with families. - Practice the initiation and continuity of an active relationship with clients and their family
- Notification and referral services of dangerous diseases - Identify critical cases, practice notification activities, and refer cases according to indication
- Referral of chronic disease - Put criteria of referral, identify its presence, practice referral measures, and follow up cases to ensure continuity of health care
Table 6 shows the pre- and post-test results of the implementation group and those of the comparison group regarding tested knowledge and skills. There was significant improvement in both areas following the application of the training model. The percentage of those who received a total score >75 increased significantly for all types of examinations conducted in comparison to pretest results and those of the control group.
Table 6 Assessment results for knowledge and skills of implementation (pre and post exposure) and control groups.
assessment tools % of students who got the specified degrees
<50 50–75 >75
multiple choice questions a 7.14 42.9 50.0
b 0 10.7 89.3
c 8.9 39.3 51.8
modified assay questions a 3.6 78.6 17.9
b 3.6 39.3 57.1
c 8.9 66.1 25.0
case-management exercises a 25.0 57.1 17.9
b 0 0.0 100.0
c 3.2 58.9 17.9
flow charts a 53.6 46.4 0.0
b 0 0.0 100.0
c 55.4 33.9 10.7
oral examination a 10.7 75.0 14.3
b 0.0 21.4 78.8
c 21.4 60.7 17.9
grand total a 20.0 60.0 20.0
b 0.0 3.6 96.4
c 23.6 51.8 24.6
a. Implementation group (28 students) pre training assessment
b. Implementation group (28 students) post training assessment
c. Control group (56 students)
Table 7 reveals the pre- and post-exposure results for students' attitudes towards clinical training at PHCC. The percentages of students who scored a three or a four on attitudes towards learning, response to advice, initiating and sharing ideas rose from 7.2%, 21.4%, 17.9%, 3.6% to 82.3%, 100.0%, 57.3% and 60.0% respectively after exposure. These rates were also higher than the relevant rates among the control group which were 7.1%, 19.6%, 14.3% and 5.4% respectively.
Table 7 Assessment results of some attitude parameters of implementation (pre and post exposure) and control groups
attitude's parameters % of students who got the specified scores
0–1 1–2 2–3 3–4
attitude toward learning a 91.0 0 1.8 7.2
b 10.7 0 7.0 82.3
c 76.8 12.5 3.6 7.1
response to advice by tutors a 35.7 42.9 0 21.4
b 0 0 0 100.0
c 30.4 41.1 9.0 19.6
initiatives by studies a 67.9 0 35.7 17.9
b 7.0 0 35.7 57.3
c 64.3 5.4 16.1 14.3
sharing ideas within the group and with tutors a 60.7 25.0 21.4 3.6
b 7.0 10.7 21.4 60.9
c 58.9 28.6 7.1 5.4
a. Implementation group (28 students) pre training assessment
b. Implementation group (28 students) post training assessment
c. Control group (56 students)
Table 8 shows that communication skills were poor among students before the application of the training model and increased to a satisfactory level following the intervention training. Table 4 also reveals that there was little improvement in those skills in the control groups after they completed their classical training in the specified departments.
Table 8 Assessment results for communication skills with patients at PHCC among implementation (per and post exposure) and control groups.
Communication parameters % of students show proper communication
a b c
are appropriate visual methods used? 21.4 89.3 23.2
is the communication brief? 43.0 85.7 46.4
is the communication unhurried? 18.0 96.4 16.1
are the facts accurate? 82.0 100.0 71.4
is the argument logical and clearly structured? 43.0 78.6 44.6
is enough detail provided? 29.0 96.4 25.0
are familiar words used? 29.0 75.0 28.6
is the sentence structure simple? 25.0 100.0 28.6
is the patient greeted? 29.0 26.4 30.4
is the patient spoken to by name? 0.0 78.6 10.7
is the patient existing knowledge explored? 0.0 64.3 5.4
are the patient's beliefs respected? 68.0 100.0 69.6
is he credited for appropriate action? 64.8 85.7 60.7
are blame and condemnation avoided? 29.0 100.0 33.9
is concern shown for the patient's problem? 18.0 100.0 25.0
dose any solution offered actually solved the problem as seen by the patient? 32.0 89.3 35.7
is the patient asked to apply information? 3.6 85.7 10.7
are the patient knowledge & understanding tested? 0.0 96.4 8.9
a. Implementation group (28 students) pre training assessment
b. Implementation group (28 students) post training assessment
c. Control group (56 students)
Significant improvement in the skills of measuring arterial blood pressure and preparing a blood film for malaria among the implementation group after the one month training for family medicine at the PHCC were also observed (Tables 9 and 10).
Table 9 Assessment result for student's skills on measuring arterial blood pressure.
component task % of students who got the specified rating scores
0 1 2 3 4
explaining what will be done a 89.3 10.7 0 0 0
b 0 0 3.6 3.6 92.8
explaining the procedure in patient language a 53.6 35.7 10.7 0 0
b 0 0 7.1 21.4 71.4
checking the cuff size a 57.1 21.4 17.9 3.6 0
b 0 0 3.6 3.6 92.8
rolling up sleeve a 35.7 35.7 17.9 0 10.7
b 0 0 0 17.9 82.1
centering the cuff bladder over brachial artery a 0 0 3.6 7.1 89.3
b 0 0 0 0 100.0
positioning & supporting the arm at heart level a 53.6 21.4 7.1 17.8 0
b 0 0 10.7 10.7 78.6
taking palpation a 100.0 0 0 0 0
b 0 0 7.1 10.7 82.0
waiting 30 seconds to allow the arm to rest a 100.0 0 0 0 0
b 0 0 7.1 0 92.9
repositioning of the arm at heart level a 53.6 17.9 10.7 8.9 8.8
b 0 0 0 3.6 96.4
placing the diaphragm over brachial artery a 0 0 0 0 100.0
b 0 0 0 0 100.0
inflating the cuff 20 mm above palpatory artery a 58.9 41.1 0 0 0
b 0 0 10.8 7.1 82.1
recording Bd. P a 0 0 0 0 100.0
b 0 0 0 0 100.0
replacing the arm at rest a 55.4 23.2 10.7 101 0
b 0 0 0 17.9 82.1
offering patient time to ask questions a 100.0 0 0 0 0
b 0 0 25.0 0 75.0
a. Implementation group (28 students) pre training assessment
b. Implementation group (28 students) post training assessment
Table 10 Assessment result for student's skills on making a blood film for malaria.
Procedure/component task % of students got the specified rating score
0 1 2 3 4
clearing the slide a 89.3 3.6 5.4 1.7 0
b 0 0 7.1 10.7 82.2
positioning the drop one cm from slide end a 53.6 10.7 35.7 0 0
b 0 0 0 10.7 89.3
using a cut edge slide as spreader a 35.7 17.9 23.2 5.4 17.8
b 0 3.6 3.6 35.7 57.1
moving hand firmly and steadily a 58.9 23.2 0 10.7 7.2
b 0 0 0 7.1 92.9
making thin blood smear a 35.7 17.9 17.9 26. 8 1.7
b 0 0 0 42.9 57.1
shaking blood Smear in air to dry a 53.6 35.7 0 1.8 8.9
b 0 0 0 32.1 67.9
labeling the patient's name on slide a 73.2 5.4 8.9 12.5 0
b 3.6 0 0 0 96.4
making 2 slides per patient a 82.1 0 0 0 17.9
b 3.6 0 0 0 96.4
putting dry slide upright on slide a 53.6 17.9 10.7 17.8 0
b 0 0 7.1 35.7 57.2
a. Implementation group (28 students) pre training assessment
b. Implementation group (28 students) post training assessment
Finally, evaluation of the training model by the 28 trainee students revealed a satisfaction index ranging from 91.4–100.0% (Table 11).
Table 11 Evaluation of the training model by trainee students using 25 questions items.
evaluation aspects total scores satisfaction index
planning (4 questions) 135.5 96.8%
relevance and utility (5 questions) 136.8 97.7%
program's running & organizer's attitudes (5 questions) 136.8 97.7%
activities according to available time (4 questions) 137.0 97.9%
benefits gained by trainees (4 questions) 135.8 97.0%
model evaluations (3 questions) 140.0 100.0%
Discussion
A task-based community oriented teaching model in family medicine was developed and tested on final year medical students. The study found significant improvement in knowledge and performance skills following exposure to the model.
The study also benefited from the students' enthusiasm. Participating students were eager to learn a greater variety of skills and to examine large number of readily accessible cases. The model is easily applicable at little or no extra cost to standard curricula.
Nevertheless, acceptance of the programme, in the long run, would be improved if tutors from all clinical departments participate in the training model. As the Al-Hadba PHCC is designed for training students, precautions should be taken to reduce the burden on staff if the program is to be implemented in other PHCCs.
The staff evaluation was not blind and was done primarily by community medicine tutors who might be biased towards such a program. This may partially affect the results of the evaluation. Nonetheless, answers to the evaluation questions were pre-determined and were standardized from books and practices by academic staff from different departments to decrease the possibility of such biases.
The teaching of family medicine should be considered for inclusion in the curriculum of Iraqi medical schools. Not only are the majority of newly graduated doctors involved in general practice but there strong reasons for arguing that family medicine should be the core setting for medical training. Most medical schools now have separate departments of family medicine. These departments are responsible for teaching and training students with a curriculum which is frequently updated according to health needs and priorities [15,16].
Medical school students should be exposed to a broad range of educational experiences and gain practical skills. They also should be given time to communicate with their patients to allow them sufficient time to develop an understanding of their problems. Issues relating to accessibility, quality of care and cost can be easily addressed in a PHCC setting where the majority of medical care activities occur. In addition, it has been found that patients at the PHCC both enjoy taking part in undergraduate training and receive greater benefits from the process [17,18].
A task-based community oriented training model was developed. The method integrated PHC services with the tasks of family physicians; this combination helped demonstrate the skills that need to be taught to undergraduate medical students. Analysis of relevant and priority medical services was used to formulated the model. The model is believed to cover all of the required items suggested by similar programs in other studies[19,20]. Community oriented primary health care has consistently showed dramatic positive changes in the health status of the population in several countries. To achieve this goal, obstacles should be overcome so that effective training can be implemented within the traditional medical school curriculum [21].
The principle characteristic that distinguishes the task-based health-care oriented teaching model advocated by the present study is the method by which the skills were identified, using appropriate instructional and assessment methods. The sequence of stages was initiated by obtaining detailed information about the nature of a family physician's job at PHC. This was followed by a break-down of those tasks into subtasks, which were then incorporated into the training model.
The teaching model for undergraduates' clinical training in family medicine relies heavily on the multi-disciplinary nature of instructional strategies. Those which are practiced frequently on a daily basis include: plenary and small group discussions, case management exercises, flow charts, observed and supervised field work, in addition to the teaching and practice of certain essential performance skills. These strategies have enhanced the learning process through student centered approaches which encourage an active learning. This differs from current instructional strategies, which are teacher-centered and often result in passive learning. Students' participation in the learning process of PHC will produce versatile doctors who will be better equipped to cope with constant changes. Moreover, implementation of the training model promoted self-directed teamwork learning, with its significant effect on improving medical education. These structured learning experiences are now believed to constitute a major part in the medical education arena [7,22,23].
Assessment of the trainees and their follow up throughout the course program is similar to that reported in other studies [24,25]. The structured learning experiences are much more likely to change students knowledge, attitude and lead to the development of proper clinical skills.
The model should be further reviewed according to the changing health needs of the community. Reforms and continuous evaluations are also essential for monitoring the model and improving utilization of health services. Finally, better results expected if the model were applied by a department of family medicine, which should be established in each medical school [26,27].
Ironically, in Iraq, emphasis has been placed on family medicine as a specialty for newly graduated doctors although no training program exists for undergraduates. In developed countries, it has been suggested that emphasis on family medicine reversed the decline in general practice and high expectations are now placed on family medicine departments to solve pressing health care system problems [28-30]. In the United States, for example, suggestions have been put forth to design a comprehensive family medicine career development program encompassing elementary through postgraduate education. The aim will be to identify youth who wish to become future family physicians [31,32]. Other models of teaching family medicine include teaching the entire curriculum in the community. Such an approach might be less effective in Iraq than a faculty-based model based on the college's curriculum and supervised by the department of family medicine.
Conclusion
A significant deficiency in knowledge and skills has been observed among students who were not trained for family medicine in a PHCC setting. It is therefore recommended to implement a task-based community oriented model for teaching family medicine in all Iraqi medical schools. Training should be supervised by clinical departments until a separate department of family medicine is established.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
Both authors contributed equally in the designing, implementation of the model, analyzing data, and drafting the manuscript. Both authors read and approved the final manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
==== Refs
Henle DE A trial of increased access to primary care N Eng J Med 1996 335 895 898 10.1056/NEJM199609193351214
Mainous AG Baker R Love MM Grey DP Gill JM Continuity of care and trust in one's physician: evidence from primary care in the United States and United Kingdom Fam Med 2001 33 22 7 11199905
Baskerville NB Hogg W Lemelin J Process of evaluation of a tailored multifaceted approach to changing family physician practice patterns improving preventive care J Fam Pract 2001 50 W242 W249 11252222
Al-Omar BA Bin-Saeed KS Factors influencing patient's utilization of primary health care provider in Saudi Arabia Saudi J Fam and Comm Med 1999 5 23 30
Watt G The long march: the development of academic general practice in the UK and Ireland Eur J Gen Pract 2004 10 98 102 15534574
Al-Ghamdi AH Al-Shehri AM Differences in generalist and specialist problem solving conflict or complimentary purposes Saudi Med J 1996 17 422 7
Al-Shehri AA Al-Faris E Learning objectives of medical students: what is the message? Saudi Med J 1998 19 747 9
Jorden JC Average number of weeks for clinical clerkships at US Medical Schools Curriculum directory 26 Association of American Medical Colleges 12
Skochelak S Thaler S Gierde C The interdisciplinary generalist curriculum project at the university of Wisconsin medical school: the generalist partners program Acad Med 2001 76 S131 3 11299186
Tenore JL Sharp LK Lipsky MS A national survey of procedural skill requirements in family practice residency programs Fam Med 2001 33 28 38 11199906
Long A Implications of the new curriculum for medical trainers Hosp Med 2005 66 135 8 15791870
Al-Awkati NA Al-Dabbagh SA Activities of the programme for development of medical education in collaboration with health services 1996–2002 Health prevention directorate MOH Baghdad Iraq 2002
Al-Dabbagh S Al-Taee W A suggested job description for the general practitioner in Iraq Med J Basrah University 2004 22 56 62
Guilbert JJ Importance of defining professional tasks educational handbook for health personnel 1987 35 Revised WHO publications Geneva
Siddiky A General practitioners for the next millennium: suggestion for medical curriculum reform Br J Gen Pract 2004 54 638 40 15517692
Howe A Education in family medicine-gains and dangers Croat Med J 2004 45 533 6 15495275
Dale J Green J Reid F Glucksman F Primary care in the accident and emergency department: I prospective identification of patients B M J 1995 311 423 6
Mathers J Parry J Lewis S Greenfield S What impact will an increased number of teaching general practices have on patients, doctors and medical students? Med Educ 2004 38 1219 28 15566532 10.1111/j.1365-2929.2004.02014.x
Grant J Ramsay A Bain J Community hospitals and general practice: extended attachments for medical students Med Educ 1997 31 364 8 9488859 10.1046/j.1365-2923.1997.00684.x
Murray E Todd C Modell M Can general internal medicine be taught in general practice? An evaluation of the University College, London Model Med Educ 1997 31 369 74 9488860 10.1046/j.1365-2923.1997.00678.x
Longlett SK Kruse JE Wesley RM Community oriented primary care: historical perspective J Am Board Fam Pract 2001 14 54 63 11206694
Egan AG Schwartz PL Heath CJ Programme components that encourage students in a traditional medicine curriculum to accept innovative teaching methods Teaching and learning in medicine 1994 6 154 60
Azer S Problem-based learning. A critical review of its educational objectives and the rationale for its use Saudi Med J 2001 22 299 305 11331485
Guyatt G Nishikawa J A proposal for enhancing the quality of clinical teaching Med Teach 1993 15 147 161 8246711
Price AD Mitcholl CA A model for clinical teaching and learning Med Educ 1993 27 62 8 8433663
Green LA Graham R Freg JJ Stephens GG Keystone III The role of family practice in a changing health care environment: A dialogue 2001 Washington, DC: American Academy of Family Physicians
Green LA Fryer GE Jr Family practice in the United States: position and prospects Acad Med 2002 77 781 9 12176691
Graham R Roberts RG Ostergaard DJ Kahn NB Pugno PA Green LA Family practice in the United States. A status report JAMA 2002 288 1097 1101 12204078 10.1001/jama.288.9.1097
Edsall RL Recreating family practice Fam Pract Manag 2002 9 10 12375469
Sox HC The future of primary care Ann Intern Med 2003 138 230 2 12558363
Martin JC Avant RF Bowman A Dickinson JC Evans KL Green LA Henley DE Jones WA Matheny SC Nevin JE Panther SL Puffer JC Roberts RG Rodgers DV Sherwood Ra Stange KC Weber CW Future of Family Medicine Project Leadership Committee The future of family medicine: A collaborative project of the family medicine community Annal Family Medicine 2004 2 S1 S32
Bentzen N Family medicine research: implications for Wonca Ann Fam Med 2004 2 S45 S49 15655088 10.1370/afm.190
|
16115312
|
PMC1215485
|
CC BY
|
2021-01-04 16:30:56
|
no
|
BMC Med Educ. 2005 Aug 22; 5:31
|
utf-8
|
BMC Med Educ
| 2,005 |
10.1186/1472-6920-5-31
|
oa_comm
|
==== Front
BMC NeurosciBMC Neuroscience1471-2202BioMed Central London 1471-2202-6-531612239310.1186/1471-2202-6-53Research ArticlePlexin B3 promotes neurite outgrowth, interacts homophilically, and interacts with Rin Hartwig Christine [email protected] Andres [email protected] Sarka [email protected] Georg [email protected] Ulrich [email protected] Institut für Humangenetik, Universitätsklinikum Hamburg-Eppendorf, Universität Hamburg, Hamburg, Germany2 Institute of Gene Technology, Tallinn Technical University, Tallinn, Estonia3 Laboratoriumsmedizin Dortmund, Dortmund, Germany2005 25 8 2005 6 53 53 29 3 2005 25 8 2005 Copyright © 2005 Hartwig et al; licensee BioMed Central Ltd.2005Hartwig et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Plexins, known to date as receptors of semaphorins, are implicated in semaphorin-mediated axon repulsion and growth cone collapse. However, subtype-specific functions of the majority of the nine members of the mammalian plexin family are largely unknown. In order to investigate functional properties of B-plexins, we analyzed the expression of human and murine plexin B3 and expressed full-length human plexins B2 (B2) and B3 (B3) in NIH-3T3 cells.
Results
Unexpectedly, B3 strongly and B2 moderately stimulate neurite outgrowth of primary murine cerebellar neurons. Both plexins mediate Ca2+/Mg2+-dependent cell aggregation due to homophilic trans-interaction, which is strong in the case of B3 and moderate for B2. Using different deletion constructs we show that the sema domain of B3 is essential for homophilic interaction. Using yeast two-hybrid analysis, we identified the neuron-specific and calmodulin-binding Ras-related GTPase Rin as an interaction partner of the intracellular part of B3, but not of B2. Rin, also known for its neurite outgrowth-inducing characteristics, co-localizes and co-immunoprecipitates with B3 in co-transfected COS-7 cells.
Conclusion
Our data suggest an involvement of homophilic interaction of B3 in semaphorin-independent signaling mechanisms positively influencing neuronal morphogenesis or function. Furthermore the neuron-specific small GTPase Rin is involved in downstream signaling of plexin B3.
==== Body
Background
During the development of the nervous system neurons respond to attractive and repulsive guidance cues to navigate to their final targets [1,2]. The nine mammalian plexins, A1–4, B1–3, C1, and D1 [3,4] are characterized by a sema domain, three cysteine-rich repeats (MRS, Met-related sequences, or PSI, plexins, semaphorins, and integrins), three glycine/proline-rich repeats (IPT, immunoglobulin-like fold shared by plexins and transcription factors), a single-pass transmembrane region, and an intracellular SP (sex plexin) domain consisting of two different parts [5]. Plexins are known as semaphorin receptors [6]. Molecules associated with plexins in receptor complexes include cell adhesion molecule L1, the scatter factor receptors Met and Ron, erbB-2, OTK, and VEGFR2 [7-14]. Interactions have been shown between plexin C1 and semaphorin 7A [3,15], plexin D1 and semaphorin 3E [16], plexin B1 and semaphorin 4D [3], and plexin B3 and semaphorin 5A [17]. Semaphorin 5A induces growth cone collapse in retinal ganglion cells, has axon-repelling activity [18], induces cellular collapse, and leads to inhibition of integrin-based adhesion of NIH-3T3 fibroblasts expressing recombinant plexin B3 [17]. The cytoplasmic C-terminus of B plexins activates Rho GTPase through Rho guanine nucleotide exchange factors PDZ-RhoGEF and LARG [19-24]. Based on this C-terminal interaction, plexin B1 mediates semaphorin 4D-induced growth cone collapse in neurons [20]. Independently of this mechanism, a direct down regulation of the activity of neurite outgrowth-promoting GTPase R-Ras by the GTPase activating protein (GAP)-homologous domain of plexin B1 has been shown [25]. Thus, according to published data, plexins appear to be mainly involved in the repulsive activities of semaphorins on neuronal cells.
We found evidence for plexin B3- and B2-dependent stimulation of neurite outgrowth, subtype-specific homophilic interaction of B3 and B2, respectively, and an interaction of B3 with neuron-specific GTPase Rin, the latter one known for its involvement in neurite outgrowth.
Results
Expression and alternative splicing of PLXNB3
Northern blot analysis of 12 different human organs (Figure 1A) revealed a strong band of ~6.2 kb from the brain sample but not the remaining organs, indicating that PLXNB3 is expressed abundantly only in brain. The estimated size of the mRNA corresponds well with that of the mature message predicted from the cloned full-length human cDNA [GenBank:AF149019]. BLASTn screening of human dbEST by AF149019 revealed 56 fully matching entries, all of them representing the 3'-end of the transcript and two variants. EST [GenBank:BF345653] from oligodendroglioma lacks 246 nucleotides of exon 27, corresponding to bp 4,595–4,840 of AF149019. This gap predicts an in-frame loss of 82 codons (aa 1,495–1,575). EST [GenBank:H51489] from adult brain lacks the 67 3'-terminal nucleotides of exon 27 (bp 4,774–4,840 of AF149019). This gap predicts a C-terminally truncated isoform of B3 due to a frame-shift resulting in the inclusion of nine amino acids (aa 1,554–1,563) followed by a premature stop. These findings suggest alternative splicing and the existence of at least three different B3-isoforms due to skipping of various parts of exon 27. Differential expression of the three isoforms in human organs was confirmed by PCR using isoform-specific primers. As shown in Figure 1C the full-length exon 27-isoform was detectable in the majority of the organs analyzed but skeletal muscle and heart. cDNA of the truncated isoform was detectable only in the brain (Figure 1D), whereas the isoform lacking 82 codons was present in skeletal muscle, liver, pancreas, kidney, brain, and heart (Figure 1E). The structures of full length B3 and the two different isoforms are shown in figure 1F–H.
Figure 1 Expression and alternative splicing of PLXNB3 in adult human tissues. (A), expression analysis of PLXNB3 in adult human tissues by poly(A)+ mRNA northern blot. (B), the blot was stripped and reprobed with β-actin probe. Molecular weight in kb is indicated at the left side. (C-E), tissue distribution of three isoforms of PLXNB3 due to alternative splicing of the 3'- part of exon 27. Fragments were amplified by RT-PCR using a common forward primer and isoform-specific reverse primers. Fragment sizes (bp) are given on the right. ΦX174 DNA cleaved by HaeIII was used as size standard. (C), 698 bp fragment containing full length exon 27. (D), 536 bp fragment lacking the 3'-terminal part of exon 27 and coding for a C-terminally truncated B3. (E), 356 bp fragment lacking (in-frame) 246 bp of exon 27. (F-H), possible protein structures of B3 predicted by mRNA isoforms generated through alternative splicing of exon 27. (F), full length isoform; (G), C-terminally truncated B3 predicted by the isoform shown in D; (H), structure of the B3 isoform lacking 246 bp of exon 27 (missing 83 amino acids marked by arrow) as shown in E. SP, signal peptide; Sema, semaphorin domain; MRS, Met-related sequences; IPT, immunoglobulin-like fold shared by plexins and transcription factors; TM, transmembrane domain; SP1/SP2, two different parts of the sex plexin domain.
Analysis of recombinantly expressed and cerebral plexin B3 protein
Conceptual translation of PLXNB3 cDNA [GenBank: AF149019] predicts a protein of 1,909 aa with a molecular mass of 207 kDa and an isoelectric point of 6.21. Western blot (WB) analysis of proteins from COS-7 cells stably overexpressing full-length B3 and from human neocortex (but not from corpus callosum) using antibody pAbB3-B against the third IPT-domain of human B3 revealed bands of ~260 kDa and ~140 kDa that were absent in nontransfected control cells (Figure 2A). Antibody pAbB3-A against the human sema domain of B3 only detected the 260 kDa band (data not shown). Taken together, these data suggest proteolytic processing of the extracellular portion of B3 similar to that described for plexins B1 and B2 [26]. Within the extracellular domains of B3 flanked by the epitopes of pAbB3-A and pAbB3-B there are two RXXR sites corresponding to the minimal consensus motif required by proprotein convertases. Cleavage at one of these sites would remove the sema domain leaving the truncated transmembrane part of B3 not recognizable by pAbB3-A. pAbB3-B did not cross-react with plexins B1, B2, or A1 expressed in COS-7 cells and was therefore also useful for the analysis of co-immunoprecipitation of B3 with these molecules (see figure 12B–D). Mouse B3 was detected by WB analysis of adult murine brain lysate using antibody pAbmB3 against the sema domain of mouse B3 (Figure 2B, first lane). pAbmB3 detected bands of ~260 kDa, ~160 kDa, and ~140 kDa. This suggests posttranslational processing of murine B3 similar or analogous to that of human B3. Minor differences in band patterns between the murine and the human WB may be due to the different epitopes recognized by the antibodies. pABmB3 also recognized C-terminally truncated mouse B3 expressed in pcDNA/mB3V5-transfected COS-7 cells used as positive control (Figure 2B, lane 2). B3 has ten potential N-glycosylation sites (Asn-X-Ser/Thr, X≠Pro). The 260 kDa protein detected by pAbB3-B in human brain (Figure 3A, lane 1) was resistant to EndoH treatment (Figure 3A, lane 2) yielding a distinct band of ~260 kDa in addition to a novel band running at ~200–230 kDa. The latter one suggests the presence of incompletely processed B3 non-resistant to EndoH in the transfected cells. Treatment with tunicamycin resulted in a relatively reduced intensity of the 260 kDa band and the appearance of a novel band of ~200 kDa. These data suggest that the 200 kDa band corresponds to deglycosylated full-length B3, whereas the 260 kDa band represents fully glycosylated, mature, full-length transmembraneous B3 supposed to be located at the cell surface. This was confirmed by exclusive detection of a sharp band of a biotinylated 260 kDa protein after pAbB3-B-immunoprecipitation of surface-biotinylated protein of stably transfected COS-7 cells (Figure 3B, first lane). As shown by immunocytochemistry of living cells stably expressing full-length B3 (Figure 4a), a significant proportion of B3 is detectable at the cell surface.
Figure 2 Detection of plexin B3 protein in human and murine brain. (A) Western blot detection of human plexin B3 (B3) protein in pIRES/B3-tranfected COS-7 cells (lane 1) and human brain (corpus callosum, lane 3; neocortex, lane 4) using antibody pAbB3-B. COS-7 cells transfected with empty pIRES served as negative control (lane 2). (B) Western blot detection of B3 protein in mouse brain (lane 1) and of C-terminally truncated recombinant mouse B3 in pcDNA/mB3V5-transfected COS-7 cells (lane 2) using antibody pAbmB3. COS-7 cells transfected with empty pcDNA3.1D/V5-His served as negative control (lane 3).
Figure 3 Glycosylation analysis and cell surface protein biotinylation of plexin B3. (A) COS-7 cells were transfected with pIRES/B3 and lysed. Untreated lysate served as positive control (B3 pIRES). To analyze cell surface expression and glycosylation of B3 total cell lysates were treated with deglycosylating enzyme Endo H (EndoH). Alternatively, growing cells were incubated by co-translational glycosylation-inhibitor tunicamycin (Tunicamycin). Non-transfected COS-7 cells served as negative control. Lysates were analyzed by Western blot using B3-specific antibody pAbB3-B. (B) After biotinylation of cell surface proteins of cells transfected with pIRES/B3 and immunoprecipitation with pAbB3-B, B3 located on the cell surface was detected in Western blot with alkaline phosphatase-conjugated streptavidin (Surface biotinylation). Non-transfected cells treated as described before served as negative control (COS-7).
Figure 4 Cell surface localization of recombinant plexin B3 protein. (a) Immunocytochemistry using antibody pAbB3-B was performed on living NIH-3T3 cells stably transfected with pIRES/B3 encoding full-length B3. Cy3 conjugated goat anti-rabbit IgG was used as secondary antibody. (b) NIH 3T3 cells transfected with empty pIRES and treated as described for a.
B3 stimulates neurite outgrowth
We analyzed the effects of recombinant B2 and B3 expressed by NIH-3T3 substrate cells on neurite outgrowth of cerebellar neurons of six days old c57BL/6J mice. Polyclonal NIH-3T3 cells stably expressing recombinant human L1 [27,28] (Figure 5A) served as a positive control substrate. Polyclonal NIH-3T3 cells, negative for PLXNB2 and PLXNB3 cDNA in RT-PCR (data not shown), express at comparable levels recombinant human B2 or B3 at the cell surface after stable transfection using the respective full-length expression constructs pFLAG/B2 or pIRES/B3 (Figure 5B–C). Cells expressing B3 displayed increased cell adhesive properties to surfaces and a more flattened cell shape. No such changes were observed in the cells expressing recombinant B2. Non-transfected NIH-3T3 cells, known to stimulate neurite outgrowth only moderately [28], were used as negative substrate control. As suggested by the examples shown in Figure 6A and shown in Figure 6B and 6C, both plexins stimulate neurite outgrowth. Mean outgrowth stimulation through B3-positive substrate cells was significantly higher than that through L1 that in turn showed a higher stimulation than B2. Mean neurite length differed significantly between all groups (ANOVA p ≤ 0.0033). Mean length (± S.D.) of neurites of neurons grown on nontransfected cells was 51 μm ± 30; mean neurite length of neurons grown on cells expressing B2, L1, or B3 was 84 μm ± 37, 93 μm ± 59, and 129 μm ± 58. The neurites of the outgrowth experiments revealed similarly shaped size distribution curves (Figure 6C).
Figure 5 Expression of plexin B2, plexin B3 and L1 protein in stably transfected NIH-3T3 cells. Western blot detection of human L1, plexin B2 and plexin B3 protein in stably transfected NIH-3T3 substrate cells. In all cases same protein amounts were applied and non-transfected NIH-3T3 cells served as negative control. (A) Western blot detection of human L1 protein in stably pIRES/L1-tranfected NIH-3T3 cells using antibody pAbex2. (B) Western blot detection of human plexin B2 (B2) protein in stably pFLAG/B2-tranfected NIH-3T3 cells using anti-FLAG antibody. (C) Western blot detection of human plexin B3 (B3) protein in stably pIRES/B3-tranfected NIH-3T3 cells using antibody pAbB3-B.
Figure 6 Plexins B2 and B3 stimulate neurite outgrowth. (A) Examples of primary murine cerebellar neurons isolated from six days old c57BL/6J mice and grown on L1-, B2-, or B3- expressing transfected (L1, B2, B3) or non-transfected (3T3) NIH-3T3 substrate cells. (B) Mean length of neurites of murine cerebellar neurons grown for 24 h on NIH-3T3 substrate cells. Nontransfected substrate cells (3T3) served as negative control. Cells transfected with pIRES/L1 encoding neuronal cell surface molecule L1 (L1) served as positive control. The strongest stimulation of neurite outgrowth was observed on substrate cells transfected with pIRES/B3 encoding full-length plexin B3 (B3). Plexin B2-expressing substrate cells (B2) were transfected with expression construct pFLAG/B2. Each outgrowth assay was done in three independent experiments. Pooled data of the triple experiments are shown. A minimum of 400 neurons were analyzed for each substrate cell type. Error bars: S.E. of mean × 1. Mean neurite length differed significantly between all groups (ANOVA; **, p < 0.0005; *, p = 0.0033). (C) Culmulative size distribution patterns of the neurites given in B.
Ca2+/Mg2+-dependent B3-homophilic interaction in trans promotes cell adhesion
Primary cerebellar neurons (Figure 7a–f), but not astrocytes (Figure 7g–i) or oligodendrocytes (Figure 7j–l) of six days old mice grown under selective conditions express PlxnB3, suggesting the possibility that neuronally expressed B3 may be involved in mediation of the B3-dependent stimulation of neurite outgrowth. To determine the in vivo expression of PlxnB3, we performed in situ hybridization (ISH) in adult mouse cerebellum using two different probes covering nucleotides 4,647 – 5,936 (Figure 8a) or 3,744 – 5,679 like the probe used by Worzfeld et al. [29] (data not shown). With both probes, strongest labeling was observed in cerebellar neurons, i.e. Purkinje and granular cells. Immunohistochemistry (IHC) of adult human cerebellum using antibody pAbB3-B also revealed staining of Purkinje and granular cells (Figure 8b). Furthermore, murine and human B3 co-immunoprecipitate (see next paragraph). This allowed us to hypothesize that homophilic interaction of B3 and possibly also that of B2 may underlie the B3- and B2-specific stimulation of neurite outgrowth, respectively.
Figure 7 Expression of PlxnB3 in cultivated primary murine cerebellar neurons. Expression of PlxnB3 in cultivated primary cerebellar neurons isolated from six days old c57BL/6J mice (a-f; 400 × magnification) but not in astrocytes (g-i) or oligodendrocytes (j-l) (630 × magnification). The cells were cultivated from a cerebellar homogenate of a six days old mouse and analyzed by combined in situ hybridization (ISH) using a PlxnB3-specific probe (a, d, g, j) and immunocytochemistry (IC) using cell type-specific antibodies (b, e, h, k). Overlays of ISH bright field signals and IC signals (blue) are shown in c, f, i, and l. The cells were grown under the respective selective conditions for neurons, astrocytes or oligodendrocytes. Following ISH slides were used for IC using primary antibodies against Neuromodulin (b) or NeuN (e) for neurons, GFAP for astrocytes (h) or CNPase for Oligodendrocytes (k). The Neuromodulin- and NeuN-positive cells (neurons) shown in b and e are positive for PlxnB3 mRNA (a, d).
Figure 8 Neuronal expression of B3 in adult murine and human cerebellum. (a) Non-radioactive in situ hybridization of PlxnB3 mRNA in adult murine cerebellum using a probe corresponding to nucleotides 4,647 – 5,936 of PlxnB3 cDNA. Purkinje and granular cells show most intense staining (200 × magnification). (b) Detection of plexin B3 protein in adult human cerebellum by immunohistochemistry using antibody pAbB3-B. Positive signal is represented by brown staining and most intense in cerebellar Purkinje cells and neurons of the granular layer (400 × magnification).
In order to investigate the possibility of homophilic interaction, we performed cell aggregation assays using transfected NIH-3T3 cells stably expressing B3 or B2. Transfected cells expressing L1 and non-transfected cells served as positive and negative controls, respectively. B3-, B2-, and L1-expressing cells were stained with DiI (red) and mixed 1:1 with non-transfected cells stained with DiO (green). Aggregates of non-transfected DiI-stained cells mixed 1:1 with DiO-stained ones were composed of approximately equal proportions of DiI- and DiO-labeled cells (Figure 9a). In contrast, there was strong predominance of DiI-stained B3-expressing cells (red) in the aggregates as shown in Figure 9b, indicating an enhanced aggregation due to homophilic interaction of B3 in trans. Similar results were obtained with cells expressing L1 (Figure 9c) and to a smaller extent also for those expressing B2 (Figure 9d). Furthermore, DiI-stained L1-expressing cells (red) were mixed 1:1 with DiO-stained B3-expressing ones (green); the almost monochrome green and red aggregates indicate specific and preferential homotypic interactions of both L1 and B3 (Figure 9e), respectively. As further negative control for the aggregation assay cells expressing B3 (Figure 9f), L1 (Figure 9g) or B2 (Figure 9h) labeled alternatively with DiO or DiI were mixed 1:1, respectively; all aggregates were composed of approximately equal proportions of DiI- and DiO-labeled cells.
Figure 9 Cell aggregation mediated by subtype-specific homophilic interaction in trans of plexins B2 and B3. Aggregation of NIH-3T3 cells expressing full-length recombinant B3, L1, or B2 and non-transfected cells (100 × magnification). (a) Negative control non-transfected cells labeled with DiO (green) were mixed 1:1 with DiI-labeled ones (red) and incubated in DMEM for 45 min prior to fluorescence microscopy. The formed aggregates contain approximately equal proportions of DiO- and DiI-labeled cells. (b)-(d) Non-transfected cells labeled with DiO and DiI-labeled cells stably expressing full-length B3 (b), L1 (c), or B2 (d) were mixed 1:1, respectively and treated as described in a. Predominance of DiI-labeled cells in the formed aggregates indicates cell-cell-adhesion due to homophilic interaction in trans mediated by B3, L1, and B2. (e) L1-transfected cells labeled with DiI and DiO-labeled cells stably expressing B3 were mixed 1:1. The almost pure red or pure green aggregates indicate preferential homotypic interactions of both B3 and L1, respectively. (f-h) As additional negative control cells labeled with DiO and expressing B3 (f), L1 (g) or B2 (h) were mixed 1:1 with DiO labeled cells expressing B3 (f), L1 (g) or B2 (h). The formed aggregates contain approximately equal proportions of DiO- and DiI-labeled cells.
Time dependent aggregation of cells expressing B2 was moderately enhanced, whereas that of cells expressing B3 or L1 was strongly enhanced (Figure 10). After 80 min aggregation time the number of particles decreased by 79%, 76%, 64%, and 52% in L1-, B3-, B2, and non-transfected cells, respectively (Figure 10A). As shown in Figures 10B and 10C, absence of divalent cations abolishes completely both B2- and B3-dependent but not L1-dependent aggregation, the latter one known to be independent of divalent cations [30].
Figure 10 Plexins B2 and B3 promote Ca2+/Mg2+-dependent cell aggregation. Cells stably expressing the neuronal cell adhesion molecule L1 known to promote Ca2+-independent cell aggregation were used as positive control. Time-dependent decrease of aggregation index Nt/N0 indicates decreasing total number of particles due to an increasing proportion of cells aggregated. Given are mean values of three independent experiments with error bars = S.E. of mean × 1. Aggregation assays were performed in DMEM (A), HBSS without Ca2+/Mg2+ (B), and in HBSS with 1 mM Ca2+/0,5 mM Mg2+ (C).
Homophilic interaction of B3 is mediated by the sema domain
Homophilic binding of B3 was further analyzed by co-immunoprecipitation (IP) of full-length B3 and several deletion mutants (Figure 11A). Homophilic binding of full-length B3 was shown by anti-myc IP of lysates of cells co-transfected with expression constructs pEGFP-N1/B3 encoding EGFP-tagged and pSecTag2B/B3 encoding myc-tagged full-length B3, followed by detection with anti-EGFP antibodies (Figure 11B). Full-length B3 co-immunoprecipitates with deletion mutants containing the sema domain alone or lacking the intracytoplasmic part of B3 (B3Δic FLAG) but not with those lacking both the sema domain and the intracytoplasmic part (B3ΔsemaΔic V5; Figure 11C–E). In addition, V5-tagged sema domain (B3sema V5) co-immunoprecipitates with HA-tagged sema domain (B3sema HA; Figure 11F). These data indicate that the sema domain is both essential and sufficient for homophilic binding. Similarly, full-length human B3 (and B3Δic FLAG but not B3ΔsemaΔic V5, data not shown) co-immunoprecipitates with mouse B3 lacking most of its intracellular part (Figure 12A). This supports the assumption that a quasi homophilic interaction of human and murine B3 in trans may underlie the observed B3-dependent stimulation of neurite outgrowth. Full-length B3 co-immunoprecipitated with its known [17] ligand semaphorin 5A (data not shown); IP of lysates of cells co-expressing B3 and plexins A1, B1, or B2 revealed no evidence for heterophilic interaction of B3 with these molecules (Figure 12B–D).
Figure 11 Co-immunoprecipitation experiments showing homophilic interaction of B3 mediated by the sema domain. COS-7 cells were co-transfected with various full-length and deletion constructs of B3. Cells transfected with a putative interaction partner and the corresponding/respective vector lacking an insert served as negative controls. Cells were lysed and immunoprecipitations (IP) were performed by various antibodies. Total lysates and IPs were analyzed by Western blot (WB) using antibodies as indicated in the figures. Bands marked with * represent antibodies precipitated by protein-A-agarose. (A) Human B3 constructs used for the IP experiments. (B) Cells were co-transfected with pSecTag2B/B3 encoding myc-tagged full-length B3 and pEGFP-N1/B3 encoding EGFP-tagged full-length B3. IP was performed with anti-myc antibody and shows homophilic interaction of B3. (C-E) Co-IP of three different B3 deletion mutants with anti-myc antibody against myc-tagged full-length B3, demonstrating homophilic interaction mediated through the sema domain. Cells were co-transfected with pSecTag2B/B3 and pFLAG/B3Δic encoding B3 missing the intracellular part (C), pcDNA3.1/B3ΔsemaΔic encoding a V5-tagged fragment of B3 missing the intracellular part and the sema domain (D), or pcDNA3.1/B3sema encoding the V5-tagged sema domain of B3 (E). (F) Cells were co-transfected with pcDNA3.1/B3sema encoding V5-tagged sema domain of B3 and with pcDNA3.1/B3semaHA encoding HA-tagged sema domain of B3. Co-IP was performed using anti-HA antibody and suggests that the sema domain is essential and sufficient for homophilic binding.
Figure 12 Co-immunoprecipitation (IP) experiments showing homophilic interaction of human and murine B3 and no interaction of B3 with plexins A1, B1 or B2. Total lysates and IPs were analyzed by Western blot (WB) using antibodies as indicated in the figures. IP was performed with pAbB3-B against human B3 and shows interaction between mouse and human B3 and no interaction between B3 and human plexins A1, B1 and B2. (A) Cells were co-transfected with pSecTag2B/B3 encoding myc-tagged human full-length B3 and pcDNA3.1/mB3 encoding V5-tagged mouse B3 lacking most of its intracellular part. Cells transfected with pcDNA3.1/mB3 and pSecTag2B vector without insert served as negative control. (B) COS-7 cells were co-transfected with pIRES/B3 encoding non-tagged full-length human B3 and pcDNAVSV/A1 encoding VSV-tagged full-length human plexin A1. Cells co-transfected with pcDNAVSV/A1 and pIRES vector without insert served as negative control. (C) COS-7 cells were co-transfected with pIRES/B3 and pcDNAVSV/B1 encoding VSV-tagged full-length human plexin B1. Cells co-transfected with pcDNAVSV/B1 and pIRES vector without insert served as negative control. (D) COS-7 cells were co-transfected with pIRES/B3 and pFLAG/B2 encoding FLAG-tagged full-length human plexin B2. Cells co-transfected with pFLAG/B2 and pIRES vector without insert served as negative control.
Rin, an intracellular interaction partner of B3
In order to determine which intracellular signaling pathways may be involved in B3 dependent neurite outgrowth, we performed yeast two-hybrid screens using the Sos-recruitment system (Figure 13). pSos fusion constructs of the intracellular parts of B3 (pSos/B3IC) and B2 (pSos/B2IC) expressed in cdc25H cells served as bait and pMyr fusion constructs of human fetal brain library cDNAs as prey molecules. Putative positive clones were used for re-transformation of cdc25H cells together with pSos/B3IC or pSos/B2IC, or pSos vector without insert. Only those clones were defined positive in which co-expression of the bait was essential in order to restore cell growth at 37°C on galactose. These experiments suggested an interaction between B3 and Rin (Ras-like protein expressed in neurons) (Figure 13, line 8), a small GTP-binding protein belonging to the Ras superfamily of GTPases. For B2, no interaction partner could be identified with this system and no interaction with Rin could be demonstrated (Figure 13, line 9). Three independent cDNAs containing the complete coding region of Rin were obtained by repeated yeast two-hybrid screens. Rin co-transfected with pSos-vector without insert was not able to induce growth on galactose at 37°C (Figure 13, line 10), showing that Rin does not activate the system unspecifically. Correspondingly, B3 and Rin could be co-immunoprecipitated from COS-7 cells transiently co-transfected with pFLAG/Rin and pIRES/B3 (Figure 14A). Rin did not co-immunprecipitate with B2 (Figure 14B).
Figure 13 Plexin B3 interacts with Rin in the Sos recruitment system. Various pMyr and pSos plasmid combinations (as indicated on the left) were used for co-transformation of cdc25H yeast cells representing a positive control (line 1) and negative controls (lines 2–7, 10). Six independent transformants were spotted and grown for six days on glucose medium at 22°C (left panel) or 37°C (middle panel), and on galactose medium at 37°C (right panel). Cdc25H cells re-transformed with pSos-B3IC and pMyr-Rin (line 8) grow on galactose medium at 37°C for six days (right panel), showing interaction between full length human Rin and the intracellular part of human B3 in this system. Cells re-transformed with pSos-B2IC and pMyr-Rin (line 9) show no growth on galactose medium at 37°C after six days (right panel), demonstrating that the intracellular part of human B2 and human Rin do not interact in this system. Rin co-transfected with pSos-vector without insert is not able to induce growth on galactose at 37°C (line 10).
Figure 14 Co-immunoprecipitation (IP) experiments showing interaction of B3 with Rin. Total lysates and IPs were analyzed by Western blot (WB) using antibodies as indicated in the figures. (A) COS-7 cells were co-transfected with pIRES/B3 encoding non-tagged full-length human B3 and pFLAG/Rin encoding FLAG-tagged full-length human Rin. Cells co-transfected with pIRES/B3 and pFLAG vector without insert served as negative control. IP was performed using anti-FLAG antibodies. (B) COS-7 cells were co-transfected with pFLAG/B2 encoding FLAG-tagged full-length human plexin B2 and pFLAG/Rin encoding FLAG-tagged full-length human Rin. Cells co-transfected with pFLAG/B2 and pFLAG vector without insert served as negative control. IP was performed using anti-Rin antibodies. Bands marked with * represent antibodies precipitated by protein-A-agarose.
COS-7 cells transiently overexpressing B3 and Rin were analyzed by confocal laser scanning microscopy (Figure 15a, b). In line with previous findings [31], Rin-specific immunofluorescence was enhanced at the plasma membrane and the nucleus of both transfected COS-7 cells (Figure 15b). Co-localization of B3 and Rin could be demonstrated at plasma membrane-associated sites (Figure 15c, d).
Figure 15 Subcellular localization of Rin and of co-localization of Rin and B3 at the plasma membrane. COS-7 cells were transiently transfected with pIRES/B3 encoding non-tagged full-length plexin B3 and pFLAG/Rin encoding FLAG-tagged full-length human Rin and analyzed by confocal laser scanning microscopy. (a) B3 was labeled using pAbB3-B and an Alexa-Fluor®568 (red)-conjugated secondary antibody. (b) Rin was labeled using anti-Flag and an Alexa-Fluor®488 (green)-conjugated secondary antibody. (c) Overlay of the images shown in a and b; yellow signals represent co-localization of B3 and Rin at the plasma membrane (arrowheads). (d) Magnification of the region marked in c. (e) Phase-contrast image of the cell shown in a-c.
Discussion
Using a neurite outgrowth assay with murine cerebellar neurons and substrate cells expressing recombinant human plexins B2 or B3, we found evidence of neurite outgrowth-promoting activity of both plexins. Up to now, in the nervous system plexins were known to act as receptors involved in repulsion and growth cone collapse. The observed stimulation of neurite outgrowth by human plexins B2 and B3 could not be explained by the mechanisms known so far for B-plexins. For Xenopus plexin, homophilic interaction in trans has been described [32]. Homophilic interaction of various neuronal transmembrane proteins, including molecules involved primarily in repulsion, has been implicated in stimulation of neurite outgrowth. Therefore, we hypothesized that homophilic interaction of plexins B2 and B3 may underlie the neurite outgrowth-promoting activity of both plexins. Using cell aggregation assays and immunoprecipitation we found evidence of human B2- and B3-specific homophilic interaction in trans. Furthermore, human and murine B3 co-immunoprecipitated, and we could show expression of PlxnB3 in cultured murine cerebellar neurons. There seemed to be a correlation between plexin-dependent cell aggregation, adhesion, and neurite outgrowth stimulation, as B3, compared to B2, had a stronger effect on all features. These data, along with the fact that homophilic interaction of various neuronal CAMs has been implicated in stimulation of neurite outgrowth [33], support the hypothesis that homophilic interaction of B3 and also possibly that of B2 is involved in stimulation of neurite outgrowth and that in the assay presented both human plexins may stimulate neurite outgrowth of murine cerebellar neurons via a quasi homophilic interaction of the human and murine homologs. However, a reverse signaling mechanism in which B3 would act as a ligand for a yet unknown receptor cannot be ruled out as an explanation for the observed B3-dependent stimulation of neurite outgrowth. Such a reverse signaling mechanism has been described for semaphorin 6D / plexin A1 in cardiac development [34].
For plxnB3, both glial and neuronal expression have been demonstrated [17,29,35]. Using combined in situ hybridization and immunocytochemistry, we found plxnB3 mRNA in cultured primary cerebellar neurons of six days old mice. PlxnB3 mRNA was also detected in adult murine cerebellum and we observed prominent neuronal B3-specific immunostaining in adult human cerebellum. Therefore, homophilic interaction of B3 is supposed to be a possible mechanism underlying the stimulation of neurite outgrowth of cerebellar neurons in the assay presented. The new ISH-probe used in this work and the probe used by Cheng et al. [35], both cover major parts of the 3'-UTR of PlxnB3. Probes hybridizing more upstream appear to detect a lower level of neuronal and a more pronounced non-neuronal expression of PlxnB3 [29,36]. We also found in addition to neuronal staining a non-neuronal staining pattern using the same probe as Worzfeld et al. [29]. These data suggest the existence of cell-type specific isoforms of B3 with different 3'-ends of the mRNA. In human organs we found evidence for the expression of such isoforms. However, in mouse EST database (NCBI dbEST) the 3'-end of PlxnB3 transcripts is strongly overrepresented with more upstream sequences represented very scarcely thus not allowing the rapid detailed analysis of tissue-specific expression of B3 isoforms.
B3 is a known receptor of semaphorin 5A that induces cellular collapse, growth cone collapse, has axon-repelling activity, and leads to inhibition of integrin-based adhesion of NIH-3T3 fibroblasts expressing transfected plexin B3 [17,18]. Our findings of both B3- and B2-associated cell aggregation and stimulation of neurite outgrowth suggest the possibility that plexins B2 and B3, via homophilic interaction, respectively, are also involved in signaling pathways independent of semaphorins. The sema domain of semaphorins contains a plexin interaction site [37]. By immunoprecipitation of various deletion constructs we showed that the sema domain of B3 was necessary and sufficient for the homophilic interaction. Since B3 does not co-immunoprecipitate with plexins A1, B1, or B2, and since B3-positive cells do not aggregate with L1-positive cells, the homophilic interaction seems to be highly specific. Therefore, the sema domain of B3 may be involved in both homophilic interaction and heterophilic interaction with semaphorin 5A. Under the experimental conditions presented for homophilic co-IP of recombinant B3, B3 also co-immunoprecipitates with semaphorin 5A. Since recombinant B3 and semaphorin 5A were co-expressed in the cells used for these experiments and semaphorin 5A could be co-immunoprecipitated by B3 despite its strong homophilic trans-interaction, both homophilic and heterophilic interactions of B3 may co-exist in vivo. Although the sema domain of B3 seems to be involved in both types of interaction, it is not clear whether semaphorin 5A and plexin B3 compete directly for B3 binding, whether different co-receptors are involved, and which signal transduction pathways are triggered by the different types of interaction.
B3 may be a multifunctional player in cell adhesion and both neurite outgrowth and repulsion, possibly due to competing ligands inducing "opposite" effects on neuronal morphology. There is a growing number of CAMs showing bifunctional characteristics with respect to involvement of homophilic interaction in stimulation of neurite outgrowth, neuronal attraction, migration, or axonal fasciculation, and heterophilic interactions in various functions including repulsion and growth cone collapse. A prominent example is represented by L1, known for its strong stimulation of neurite outgrowth in association with its homophilic binding [38] and also involved in semaphorin 3A mediated repulsion of cortical axons as part of a heteromultimeric receptor complex including plexin A1 and neuropilin 1 [12]. The Roundabout (Robo) receptor, a transmembrane glycoprotein sharing structural homology with a number of neuronal CAMs of the immunoglobulin (Ig) superfamily, is receptor for Slit, an extracellular matrix protein. Slit controls midline crossing of axons by inducing growth cone repulsion upon interaction with Robo [39]. On the other hand, homophilic trans-interaction of Robo promotes cell adhesion and neurite outgrowth [40], most likely reflecting Robo's known role in selective axon fasciculation. The majority of the known CAMs which show involvement of homophilic interaction in stimulation of neurite outgrowth or neuronal migration contain Ig (± fibronectin type III) or cadherin domains; these CAMs include L1, NCAM, Robo1, Robo2, fasciclin II, LAMP, DM-GRASP, N-cadherin, and Celsr2 [38,40-48]. Currently, the number of known CAMs with homophilic binding and neurite outgrowth stimulating characteristics is rapidly growing, with an increasing variety of molecular features not shared by Igs, fibronectins, or cadherins, as e.g. the AMIGOs or ninjurins [49-51]. Our work shows that B3 and suggests that also B2 belong to this latter group, adding further molecular heterogeneity to the group of homophilic CAMs with neurite outgrowth stimulating capacity.
Searching for intracellular pathways involved in the B2- and B3-dependent neurite outgrowth using the intracellular parts of B2 and B3 as bait in a yeast two-hybrid screen of human fetal brain cDNA we identified Rin, a neuron-specific and calmodulin-binding Ras-related GTPase, as interaction partner for B3 [52]. An interaction between B3 and Rin in mammalian cells (but not between B2 and Rin) could be shown by co-immunoprecipitation of recombinant B3 and Rin expressed in COS-7 cells in which co-localization of these proteins at plasma membrane-associated sites could be shown by confocal laser scanning microscopy. Therefore, one may assume physiological interaction of B3 and Rin as part of a neuronal receptor complex involved in B3-dependent signaling. Expression of recombinant Rin induces neurite outgrowth in rat pheochromocytoma PC12 cells and Rin interacts with the transcription factor Brn-3a, which is known to regulate different genes involved in neuronal differentiation and survival [31,53]. If the homophilic interaction of B3 is responsible for B3-dependent stimulation of neurite outgrowth one may speculate therefore that Rin may be involved in this process. Since Rin and B2 do not interact in yeast or mammalian cells, this interaction seems to be specific for B3 and other mechanisms seem to be involved in B2-dependent stimulation of neurite outgrowth. The Rin-interacting intracellular subdomain of B3 remains to be identified. This may help to elucidate how plexins may be involved in common and subtype-specific intracellular signaling pathways [8].
Further experiments are required to investigate whether homophilic interaction of B2 or B3 is responsible for the observed stimulation of neurite outgrowth, or whether these plexins function as heterophilic ligands of yet unknown neuronal receptors in a reverse signaling mechanism.
Conclusion
Our data suggest an involvement of the homophilic interaction of plexin B3 in stimulation of neurite outgrowth. The neuron-specific small GTPase Rin, known for its neurotrophic characteristics, was identified as intracellular interaction partner of B3. Therefore, both, neuronally and non-neuronally expressed plexin B3 may be involved in semaphorin-independent signaling positively influencing neuritogenesis.
Methods
Analysis of the expression of full length B3 and its isoforms
For probe preparation, a PLXNB3-specific fragment including nucleotides 833–1,814 of AF149019 was amplified by RT-PCR and cloned into pCRII-TOPO (Invitrogen, Karlsruhe, Germany). The insert was labeled with [α32P]-dCTP and hybridized to normalized Multiple Tissue northern Blot and poly (A)+ Multiple Tissue Expression Array (BD Biosciences). The northern blot was stripped and rehybridized with a β-actin control probe.
The distribution of three isoforms of PLXNB3 due to alternative splicing of the 3'- part of exon 27 was analyzed by RT-PCR form various human organs using forward primer TTCCTCCTCACG|CTCATCCACAC (splice junction marked by bar) and isoform-specific reverse primers. A 698 bp fragment containing full length exon 27 was amplified using reverse primer TCTGGGAC|CTTGTAGTGTTG. A 536 bp fragment lacking the 3'-terminal part of exon 27 and coding for a C-terminally truncated B3 was amplified using reverse primer CAGGCCTGAGCGCCACT|CTTCTC. A 356 bp fragment lacking (in-frame) 246 bp of exon 27 was amplified using reverse primer AGGCCTGAGCGCCACT|CTGTCAC.
Plasmid constructs
pcDNA3 expression constructs encoding VSV-tagged human plexins B1 and A1 were kindly provided by Dr. L. Tamagnone. Human PLXNB2 cDNA (KIAA 0315) was kindly provided by Dr. T. Nagase. PLXNB3 cDNA [GenBank:AF149019] identified by screning fetal brain 5' Stretch Plus cDNA λ phage library (BD Biosciences) and 5'RACE and PLXNB2 cDNA were cloned into pIRES (BD Biosciences) and pFLAG (Sigma, Taufkirchen, Germany) vectors. The expression constructs were named correspondingly pFLAG/B2, pIRES/B3, and pFLAG/B3. Stop codon of PLXNB3 was changed to codon for Ser by PCR-based site directed mutagenesis. The mutant product was cloned in-frame to pEGFP-N1 (BD Biosciences) or myc-tagged pSecTag2B (Invitrogen, Karlsruhe, Germany). The constructs were named pEGFP-N1/B3 and pSecTag2B/B3. pFLAG/B3Δic, lacking intracellular part (aa 1,328–1,909) of B3, was generated by deleting a BamHI/EcoRI fragment of pFLAG/B3. B3 deletion constructs pcDNA3.1/B3sema and HA-tagged pcDNA3.1/B3semaHA, encoding sema domain (aa 1–468) and pcDNA3.1/B3ΔsemaΔic lacking sema and intracellular (aa 1,337–1,909) domains were generated by PCR and cloned directly to pcDNA3.1D/V5-His (Invitrogen). Mouse B3 cDNA was PCR-amplified from mouse brain 5'STRETCH PLUS cDNA (BD Biosciences) and cloned to pcDNA3.1D/V5-His in order to generate pcDNA/mB3V5 encoding C-terminally V5-tagged truncated mouse plexin B3 (aa 1–1,252), lacking most of its intracellular part. Human full-length semaphorin 5A cDNA was PCR amplified from human brain QUICK-Clone cDNA and cloned into pFLAG. Human full-length Rin was PCR amplified from the pMyr construct identified in the CytoTrap® system and cloned into pFLAG (pFLAG/Rin). The expression construct for L1 (pIRES/L1) was described before [28].
Antibodies
Rabbit polyclonal antibodies (pAb) against extracellular domains of B3 were produced by immunization with peptides TSRCVTLPLDSPESYP (human sema domain, aa 354–369) for pAbB3-A, VQASRAQPQDPQPRRSC (third IPT domain of human B3, aa 1,058–1,074) for pAbB3-B, and VFRRRGARAQTEYRS (mouse B3 sema domain aa 227–241) for pAbmB3. Affinity-purified antisera were diluted 1:50 for immunocytochemistry (IC) or immunoprecipitation (IP), and 1:3,000 for Western blotting (WB). pAbex2 against human L1 has been described before [28]. Mouse monoclonal antibody (mAb) anti-Sos1 (BD Biosciences) for detection of bait constructs was used at 1:4,000 dilutions for WB. Anti-hRin mAb (ICN; Eschwege, Germany) was diluted 1:5,000 for WB and 1:50 for IC. EGFP-tagged recombinant proteins were detected by WB using mAb diluted 1:100 (Living Colors® A.v. peptide anti-EGFP; BD Biosciences). Anti-c-myc mAb (Sigma) was diluted 1:500 for WB and 1:50 for IP. Anti-HA-tag rabbit pAb (BD Biosciences) was diluted 1:1,000 for WB and 1:50 for IP. Anti-FLAG mAb (Sigma), Anti-VSV-Glycoprotein mAb (Sigma) and Anti-V5 mAb were diluted for WB 1:1000, 1:10000 and 1:5000, respectively.
Protein expression, Western blotting (WB), and immunoprecipitation (IP)
Complete mouse brain, human corpus callosum, and human neocortex samples were lysed in tissue lysis buffer (20 mM tris-HCl pH7.6, 140 mM NaCl, 5 mM EDTA, 5% glycerol, 1% NP40, 0.1% SDS, complete protease inhibitor cocktail [Roche, Mannheim, Germany]), separated under reducing conditions by SDS-PAGE, blotted onto PVDF membrane (Millipore, Eschborn, Germany), followed by detection using the antibodies described above and Immun-Star Chemiluminescent system (Bio-Rad, Munich, Germany). Lipofectamine 2000 (Invitrogen) was used for all transfections. Polyclonal cell lines stably overexpressing B3 or B2 were generated as described previously for L1 [28]. Homogeneous cell surface expression and similar total levels of expression of B3, B2, and L1 was verified by living-cell IC and WB. For IP cells were lysed 30 h post-transfection in 500 μl cell lysis buffer (20 mM tris-HCl pH7.5, 150 mM NaCl, 1% TritonX-100, complete protease inhibitor cocktail) and centrifuged (10 min, 12,000 rpm). Supernatants were incubated (2 h, 4°C) with pAbB3-B, anti-myc, anti-FLAG (Sigma), or anti-HA (BD Biosciences) and precipitated (12 h, 4°C) by protein-A agarose (Roche). After washing once with cell lysis buffer and twice with washing buffer (20 mM tris-HCl pH 7.5, 500 mM NaCl and 0,1% TritonX-100), bound protein was eluted by boiling in SDS-PAGE gel loading buffer, and immunodetected as described for WB.
Glycosylation analysis and surface protein biotinylation
One day after transfection COS-7 cells were treated with 10 μg/ml glycosylation inhibitor tunicamycin (Sigma) for 24 h prior to lysis. Alternatively, lysates of untreated cells were incubated (12 h, 37°C) with 750 U endo-β-N-acetylglucosaminidase H (Endo H; New England Biolabs, Frankfurt, Germany). For surface biotinylation adherent cells were washed three times with PBS at 4°C and incubated in 0.5 mg EZ-Link® Sulfo-NHS-LC-Biotin (Pierce, Bonn, Germany) /ml PBS (30 min) on ice. After washing three times in PBS (4°C) cells were lysed in 1 ml ice-cold lysis buffer and centrifuged (10 min, 4°C, 12,000 rpm). The supernatant was incubated (2 h, 4°C) with antibodies pAbB3-B or anti-FLAG, and precipitated (12 h, 4°C) with Protein A-agarose. After washing twice with washing buffer bound protein was eluted. PAGE, blotting, and detection were done as described above. Biotinylated protein was detected by alkaline phosphatase-conjugated streptavidin (Sigma).
Immunocytochemistry (IC), in situ hybridization (ISH) and immunohistochemistry (IHC)
To detect cell surface expression of B3, IC was performed on living COS-7 and NIH-3T3 cells stably transfected with pIRES/B3 using pAbB3-B and Cy3-conjugated secondary antibodies. For detection of Rin expression and co-localization experiments cells transfected with pFLAG/Rin or co-transfected with pFLAG/Rin and pIRES/B3 were fixed with 4% paraformaldehyde (PFA) in PBS before incubation with anti-Rin and pAbB3-B and Alexa-Fluor®568- or -488-conjugated secondary antibodies and analyzed by confocal laser scanning microscopy.
For ISH, two different digoxigenin-labeled riboprobes were synthesized. The probes correspond to nucleotides 4,647 – 5,936 and 3,744 – 5,679 [29] of PlxnB3 cDNA [GenBank:NM_019587]. After fixation (4% PFA) and acetylation slides were hybridized (12 h, 55°C), followed by washing twice with 0.2 × SSC (20 min, 55°C), three times with 0.2 × SSC in 50% formamide (60 min, 55°C), once with 0.2 × SSC (10 min, RT), and final equilibration in TBS (10 min). Colorimetric immunodetection (NBT/BCIP) was performed according to the manufacturer's instructions (Roche).
To analyze the expression of PlxnB3 in various brain cell types, astrocytes, oligodendrocytes, and neurons of mechanically dissociated cerebellum of six days old mice were grown under cell type specific selective conditions according to the protocols described by [54] and used for combined ISH and IC. ISH was done as described before using a digoxigenin-labeled riboprobe corresponding to nucleotides 4,647–5,936 of PlxnB3 cDNA. After hybridization slides were washed twice with 0.2 × SSC (20 min, 55°C) and three times with 0.2 × SSC in 50% formamide (20 min, 55°C). For IC slides were equilibrated in PBS, incubated (30 min) in blocking solution (Roche), followed by incubation (12 h, 4°C) with mAb against Neuromodulin (Transduction Laboratories), NeuN (1:25; Chemicon), GFAP (1:100; Transduction Laboratories, Heidelberg, Germany) or CNPase (Sigma), washed three times in PBS, followed by detection (45 min, RT) using Cy3-labeled rabbit-anti-mouse antibody (1:200; Dianova, Hamburg, Germany), washing with PBS /0.1% Tween, and mounting in Mowiol (Calbiochem).
For IHC formaldehyde-fixed and paraffin-embedded slices (7 μm) of post mortem human brain (post mortem time 47 h) were dewaxed, hydrated and further fixed in 4% PFA. Endogenous peroxidase was quenched by 3% H2O2 (30 min). The slides were treated with 6 M urea /0.1 M glycin (pH3.5), blocked (2% BSA, 3% goat serum, 0.2% TritonX-100 in PBS), and incubated with pAbB3-B (1:50; 12 h, 4°C) detected using Vectastain ABC Kit (Vector Laboratories), DAB chromogen, and immunoperoxidase reaction. The slides were counterstained with hematoxylin.
All animal experiments have been approved by the local government body regulating animal research. The human brain tissue used in the experiments was obtained from the officially approved local research brain bank.
Neurite outgrowth and cell aggregation assays
NIH-3T3 cells were stably transfected in order to express recombinant L1CAM, PLXNB2, and PLXNB3 and used for cell aggregation assays and as substrate cells for primary murine cerebellar neurons in neurite outgrowth assay performed as described previously [28]. Cell aggregation assays were performed essentially according to the method of [30]. Monolayers of substrate cells were incubated with 2 mM EDTA in PBS (15 min, RT), dispersed by pipetting, diluted to 1 × 106 cells/ml (N0) in DMEM or Hanks' balanced salt solution (HBSS) ± 1 mM Ca2+ and 0.5 mM Mg2+, and incubated at 37°C. Nt/N0, representing aggregation-dependent decrease of total particle number at incubation time t was determined in aliquots taken every 20 min out of the suspension immediately after mixing the suspension by gentle inversion. To determine the molecular specificity of cell aggregation, control and transfected cells were stained by lipophilic dyes DiO and DiI, respectively (Molecular Probes, Leiden, The Netherlands) using 5 μl dye solution /ml (2 h, 37°C). DiO- and DiI-stained cells were diluted in DMEM to 5 × 105 cells/ml, mixed 1:1, incubated (45 min, 37°C), spotted on coverslips, fixed (4% PFA; 10 min), and washed in PBS (10 min) prior to fluorescence microscopy.
Yeast two-hybrid analysis
To identify intracellular interaction partners of plexins B2 and B3 we used the CytoTrap® yeast two-hybrid system (Stratagene) also known as Sos Recruitment System [55]. This system is based on the recruitment of human Sos (hSos) to the plasma membrane in the mutant temperature-sensitive yeast strain cdc25H. This strain is unable to grow at the restrictive temperature of 37°C unless activation occurs through translocation of hSos to the plasma membrane via interaction between two-hybrid proteins. The intracellular parts of B2 (bp 3,960–5,840) and B3 (bp 3,840–5,680) were amplified by PCR using primers with SalI and NotI restriction sites and cloned into SalI/NotI-cleaved pSos vector in order to create hSos fusion constructs as bait for the CytoTrap system (pSos/B2IC and pSos/B3IC). We screened a human fetal brain plasmid cDNA library with an average insert size of 1.3 kb in the pMyr expression vector (Stratagene) according to the manufacturer's instructions. Conventional yeast transformation by the lithium acetate method was used. Cdc25H cells were co-transformed with pSos/B2IC or pSos/B3IC and pYES/mGAP in order to reduce isolation of Ras GTPase false positive clones [56]. Expression of bait constructs was confirmed by immunoblotting of yeast lysates with an anti-hSos1 antibody. These pre-transformed cdc25H strains were transformed with pMyr-cDNA library plasmids and incubated at 22°C for 5 days on selective minimal glucose plates lacking leucin, uracil, and tryptophan. A total of ~2.5 × 106 transformants were replica-plated onto selective minimal galactose plates and grown up (7 days, 37°C). From a first selection of 350 clones 39 putative positive clones were isolated after a second round of selection on galactose plates at 37°C. Bait-prey protein interactions of putative positive clones were analyzed by re-transformation of the cdc25H yeast strain with both the respective prey-encoding pMyr-plasmid together with pSos/B2IC, pSos/B3IC, or pSos vector without insert.
Authors' contributions
AV and UF conceived of the study and participated in its design and coordination; UF wrote the grant application. AV participated in cloning and supervised some of the experiments. UF, AV, and CH drafted the manuscript. CH participated in cloning and carried out all experiments except in situ hybridization, immunohistochemistry, northern blot and RT-PCR that were done by SK. GR participated in yeast two hybrid screening and confocal laser microscopy. All authors read and approved the final manuscript.
Acknowledgements
Human plexin B1 and A1 expression vectors were generous gift from Dr. L. Tamagnone (University of Torino, IRCC, Italy). Human PLXNB2 cDNA (KIAA 0315) was kindly provided by Dr. T. Nagase (Kazusa DNA Research Institute, Kisarazu, Japan). This work was supported by Deutsche Forschungsgemeinschaft (grant number SFB 444, C3, D6) and BMBF (NGFN-FKZ 01GS0119), and Estonian Science Foundation (A.V.; grant number ETF 5915).
==== Refs
Dickson BJ Molecular mechanisms of axon guidance Science 2002 298 1959 1964 12471249 10.1126/science.1072165
Tessier-Lavigne M Goodman CS The molecular biology of axon guidance Science 1996 274 1123 1133 8895455 10.1126/science.274.5290.1123
Tamagnone L Artigiani S Chen H He Z Ming GI Song H Chedotal A Winberg ML Goodman CS Poo M Tessier-Lavigne M Comoglio PM Plexins are a large family of receptors for transmembrane, secreted, and GPI-anchored semaphorins in vertebrates Cell 1999 99 71 80 10520995 10.1016/S0092-8674(00)80063-X
van der Zwaag B Hellemons AJ Leenders WP Burbach JP Brunner HG Padberg GW Van Bokhoven H PLEXIN-D1, a novel plexin family member, is expressed in vascular endothelium and the central nervous system during mouse embryogenesis Dev Dyn 2002 225 336 343 12412018 10.1002/dvdy.10159
Maestrini E Tamagnone L Longati P Cremona O Gulisano M Bione S Tamanini F Neel BG Toniolo D Comoglio PM A family of transmembrane proteins with homology to the MET-hepatocyte growth factor receptor Proc Natl Acad Sci U S A 1996 93 674 678 8570614 10.1073/pnas.93.2.674
Pasterkamp RJ Verhaagen J Emerging roles for semaphorins in neural regeneration Brain Res Brain Res Rev 2001 35 36 54 11245885 10.1016/S0165-0173(00)00050-3
Toyofuku T Zhang H Kumanogoh A Takegahara N Suto F Kamei J Aoki K Yabuki M Hori M Fujisawa H Kikutani H Dual roles of Sema6D in cardiac morphogenesis through region-specific association of its receptor, Plexin-A1, with off-track and vascular endothelial growth factor receptor type 2 Genes Dev 2004 18 435 447 14977921 10.1101/gad.1167304
Pasterkamp RJ Kolodkin AL Semaphorin junction: making tracks toward neural connectivity Curr Opin Neurobiol 2003 13 79 89 12593985 10.1016/S0959-4388(03)00003-5
Takahashi T Fournier A Nakamura F Wang LH Murakami Y Kalb RG Fujisawa H Strittmatter SM Plexin-neuropilin-1 complexes form functional semaphorin-3A receptors Cell 1999 99 59 69 10520994 10.1016/S0092-8674(00)80062-8
Giordano S Corso S Conrotto P Artigiani S Gilestro G Barberis D Tamagnone L Comoglio PM The Semaphorin 4D receptor controls invasive growth by coupling with Met Nat Cell Biol 2002 4 720 724 12198496 10.1038/ncb843
Castellani V De Angelis E Kenwrick S Rougon G Cis and trans interactions of L1 with neuropilin-1 control axonal responses to semaphorin 3A Embo J 2002 21 6348 6357 12456642 10.1093/emboj/cdf645
Castellani V Chedotal A Schachner M Faivre-Sarrailh C Rougon G Analysis of the L1-deficient mouse phenotype reveals cross-talk between Sema3A and L1 signaling pathways in axonal guidance Neuron 2000 27 237 249 10985345 10.1016/S0896-6273(00)00033-7
Conrotto P Corso S Gamberini S Comoglio PM Giordano S Interplay between scatter factor receptors and B plexins controls invasive growth Oncogene 2004 23 5131 5137 15184888 10.1038/sj.onc.1207650
Swiercz JM Kuner R Offermanns S Plexin-B1/RhoGEF-mediated RhoA activation involves the receptor tyrosine kinase ErbB-2 J Cell Biol 2004 165 869 880 15210733 10.1083/jcb.200312094
Comeau MR Johnson R DuBose RF Petersen M Gearing P VandenBos T Park L Farrah T Buller RM Cohen JI Strockbine LD Rauch C Spriggs MK A poxvirus-encoded semaphorin induces cytokine production from monocytes and binds to a novel cellular semaphorin receptor, VESPR Immunity 1998 8 473 482 9586637 10.1016/S1074-7613(00)80552-X
Gu C Yoshida Y Livet J Reimert DV Mann F Merte J Henderson CE Jessell TM Kolodkin AL Ginty DD Semaphorin 3E and plexin-D1 control vascular pattern independently of neuropilins Science 2005 307 265 268 15550623 10.1126/science.1105416
Artigiani S Conrotto P Fazzari P Gilestro GF Barberis D Giordano S Comoglio PM Tamagnone L Plexin-B3 is a functional receptor for semaphorin 5A EMBO Rep 2004 5 710 714 15218527 10.1038/sj.embor.7400189
Oster SF Bodeker MO He F Sretavan DW Invariant Sema5A inhibition serves an ensheathing function during optic nerve development Development 2003 130 775 784 12506007 10.1242/dev.00299
Oinuma I Katoh H Harada A Negishi M Direct interaction of Rnd1 with Plexin-B1 regulates PDZ-RhoGEF-mediated Rho activation by Plexin-B1 and induces cell contraction in COS-7 cells J Biol Chem 2003 278 25671 25677 12730235 10.1074/jbc.M303047200
Swiercz JM Kuner R Behrens J Offermanns S Plexin-B1 Directly Interacts with PDZ-RhoGEF/LARG to Regulate RhoA and Growth Cone Morphology Neuron 2002 35 51 63 12123608 10.1016/S0896-6273(02)00750-X
Hirotani M Ohoka Y Yamamoto T Nirasawa H Furuyama T Kogo M Matsuya T Inagaki S Interaction of plexin-B1 with PDZ domain-containing Rho guanine nucleotide exchange factors Biochem Biophys Res Commun 2002 297 32 37 12220504 10.1016/S0006-291X(02)02122-8
Perrot V Vazquez-Prado J Gutkind JS Plexin B regulates Rho through the guanine nucleotide exchange factors Leukemia-associated RhoGEF (LARG) and PDZ-RhoGEF J Biol Chem 2002 277 43115 43120 12183458 10.1074/jbc.M206005200
Driessens MH Olivo C Nagata K Inagaki M Collard JG B plexins activate Rho through PDZ-RhoGEF FEBS Lett 2002 529 168 172 12372594 10.1016/S0014-5793(02)03323-9
Aurandt J Vikis HG Gutkind JS Ahn N Guan KL The semaphorin receptor plexin-B1 signals through a direct interaction with the Rho-specific nucleotide exchange factor, LARG Proc Natl Acad Sci U S A 2002 99 12085 12090 12196628 10.1073/pnas.142433199
Oinuma I Ishikawa Y Katoh H Negishi M The Semaphorin 4D receptor Plexin-B1 is a GTPase activating protein for R-Ras Science 2004 305 862 865 15297673 10.1126/science.1097545
Artigiani S Barberis D Fazzari P Longati P Angelini P Van De Loo JW Comoglio PM Tamagnone L Functional regulation of semaphorin receptors by proprotein convertases J Biol Chem 2003 278 10094 10101 12533544 10.1074/jbc.M210156200
Williams EJ Doherty P Turner G Reid RA Hemperly JJ Walsh FS Calcium influx into neurons can solely account for cell contact-dependent neurite outgrowth stimulated by transfected L1 J Cell Biol 1992 119 883 892 1429842 10.1083/jcb.119.4.883
Michelson P Hartwig C Schachner M Gal A Veske A Finckh U Missense mutations in the extracellular domain of the human neural cell adhesion molecule L1 reduce neurite outgrowth of murine cerebellar neurons Hum Mutat 2002 20 481 482 12442287 10.1002/humu.9096
Worzfeld T Puschel AW Offermanns S Kuner R Plexin-B family members demonstrate non-redundant expression patterns in the developing mouse nervous system: an anatomical basis for morphogenetic effects of Sema4D during development Eur J Neurosci 2004 19 2622 2632 15147296 10.1111/j.0953-816X.2004.03401.x
Miura M Asou H Kobayashi M Uyemura K Functional expression of a full-length cDNA coding for rat neural cell adhesion molecule L1 mediates homophilic intercellular adhesion and migration of cerebellar neurons J Biol Chem 1992 267 10752 10758 1587850
Calissano M Latchman DS Functional interaction between the small GTP-binding protein Rin and the N-terminal of Brn-3a transcription factor Oncogene 2003 22 5408 5414 12934100 10.1038/sj.onc.1206635
Ohta K Mizutani A Kawakami A Murakami Y Kasuya Y Takagi S Tanaka H Fujisawa H Plexin: a novel neuronal cell surface molecule that mediates cell adhesion via a homophilic binding mechanism in the presence of calcium ions Neuron 1995 14 1189 1199 7605632 10.1016/0896-6273(95)90266-X
Kiryushko D Berezin V Bock E Regulators of neurite outgrowth: role of cell adhesion molecules Ann N Y Acad Sci 2004 1014 140 154 15153429 10.1196/annals.1294.015
Toyofuku T Zhang H Kumanogoh A Takegahara N Yabuki M Harada K Hori M Kikutani H Guidance of myocardial patterning in cardiac development by Sema6D reverse signalling Nat Cell Biol 2004 6 1204 1211 15543137 10.1038/ncb1193
Cheng HJ Bagri A Yaron A Stein E Pleasure SJ Tessier-Lavigne M Plexin-A3 mediates semaphorin signaling and regulates the development of hippocampal axonal projections Neuron 2001 32 249 263 11683995 10.1016/S0896-6273(01)00478-0
Perala NM Immonen T Sariola H The expression of plexins during mouse embryogenesis Gene Expr Patterns 2005 5 355 362 15661641 10.1016/j.modgep.2004.10.001
Antipenko A Himanen JP van Leyen K Nardi-Dei V Lesniak J Barton WA Rajashankar KR Lu M Hoemme C Puschel AW Nikolov DB Structure of the Semaphorin-3A Receptor Binding Module Neuron 2003 39 589 598 12925274 10.1016/S0896-6273(03)00502-6
Lemmon V Farr KL Lagenaur C L1-mediated axon outgrowth occurs via a homophilic binding mechanism Neuron 1989 2 1597 1603 2627381 10.1016/0896-6273(89)90048-2
Long H Sabatier C Ma L Plump A Yuan W Ornitz DM Tamada A Murakami F Goodman CS Tessier-Lavigne M Conserved roles for Slit and Robo proteins in midline commissural axon guidance Neuron 2004 42 213 223 15091338 10.1016/S0896-6273(04)00179-5
Hivert B Liu Z Chuang CY Doherty P Sundaresan V Robo1 and Robo2 are homophilic binding molecules that promote axonal growth Mol Cell Neurosci 2002 21 534 545 12504588 10.1006/mcne.2002.1193
Williams G Williams EJ Doherty P Dimeric versions of two short N-cadherin binding motifs (HAVDI and INPISG) function as N-cadherin agonists J Biol Chem 2002 277 4361 4367 11726665 10.1074/jbc.M109185200
Doherty P Cohen J Walsh FS Neurite outgrowth in response to transfected N-CAM changes during development and is modulated by polysialic acid Neuron 1990 5 209 219 2200449 10.1016/0896-6273(90)90310-C
Sandig M Rao Y Siu CH The homophilic binding site of the neural cell adhesion molecule NCAM is directly involved in promoting neurite outgrowth from cultured neural retinal cells J Biol Chem 1994 269 14841 14848 8182092
Wright JW Snyder MA Schwinof KM Combes S Copenhaver PF A role for fasciclin II in the guidance of neuronal migration Development 1999 126 3217 3228 10375511
Pimenta AF Zhukareva V Barbe MF Reinoso BS Grimley C Henzel W Fischer I Levitt P The limbic system-associated membrane protein is an Ig superfamily member that mediates selective neuronal growth and axon targeting Neuron 1995 15 287 297 7646886 10.1016/0896-6273(95)90034-9
Zhukareva V Levitt P The limbic system-associated membrane protein (LAMP) selectively mediates interactions with specific central neuron populations Development 1995 121 1161 1172 7743928
DeBernardo AP Chang S Native and recombinant DM-GRASP selectively support neurite extension from neurons that express GRASP Dev Biol 1995 169 65 75 7750658 10.1006/dbio.1995.1127
Shima Y Kengaku M Hirano T Takeichi M Uemura T Regulation of dendritic maintenance and growth by a mammalian 7-pass transmembrane cadherin Dev Cell 2004 7 205 216 15296717 10.1016/j.devcel.2004.07.007
Kuja-Panula J Kiiltomaki M Yamashiro T Rouhiainen A Rauvala H AMIGO, a transmembrane protein implicated in axon tract development, defines a novel protein family with leucine-rich repeats J Cell Biol 2003 160 963 973 12629050 10.1083/jcb.200209074
Araki T Zimonjic DB Popescu NC Milbrandt J Mechanism of homophilic binding mediated by ninjurin, a novel widely expressed adhesion molecule J Biol Chem 1997 272 21373 21380 9261151 10.1074/jbc.272.34.21373
Araki T Milbrandt J Ninjurin2, a novel homophilic adhesion molecule, is expressed in mature sensory and enteric neurons and promotes neurite outgrowth J Neurosci 2000 20 187 195 10627596
Lee CH Della NG Chew CE Zack DJ Rin, a neuron-specific and calmodulin-binding small G-protein, and Rit define a novel subfamily of ras proteins J Neurosci 1996 16 6784 6794 8824319
Hoshino M Nakamura S Small GTPase Rin induces neurite outgrowth through Rac/Cdc42 and calmodulin in PC12 cells J Cell Biol 2003 163 1067 1076 14662747 10.1083/jcb.200308070
Itoh K Sakurai Y Asou H Umeda M Differential expression of alternatively spliced neural cell adhesion molecule L1 isoforms during oligodendrocyte maturation J Neurosci Res 2000 60 579 586 10820428 10.1002/(SICI)1097-4547(20000601)60:5<579::AID-JNR2>3.0.CO;2-#
Aronheim A Zandi E Hennemann H Elledge SJ Karin M Isolation of an AP-1 repressor by a novel method for detecting protein-protein interactions Mol Cell Biol 1997 17 3094 3102 9154808
Aronheim A Improved efficiency sos recruitment system: expression of the mammalian GAP reduces isolation of Ras GTPase false positives Nucleic Acids Res 1997 25 3373 3374 9241255 10.1093/nar/25.16.3373
|
16122393
|
PMC1215486
|
CC BY
|
2021-01-04 16:39:09
|
no
|
BMC Neurosci. 2005 Aug 25; 6:53
|
utf-8
|
BMC Neurosci
| 2,005 |
10.1186/1471-2202-6-53
|
oa_comm
|
==== Front
BMC NeurosciBMC Neuroscience1471-2202BioMed Central London 1471-2202-6-541612239410.1186/1471-2202-6-54Research ArticleSoy isoflavone glycitein protects against beta amyloid-induced toxicity and oxidative stress in transgenic Caenorhabditis elegans Gutierrez-Zepeda Astrid [email protected] Ross [email protected] Zhixin [email protected] Marishka [email protected] YanJue [email protected] Ikhlas [email protected] Christopher D [email protected] Baolu [email protected] Yuan [email protected] Department of Biological Sciences, University of Southern Mississippi, Hattiesburg, MS 39406, USA2 Department of Human Sciences, Alcorn State University, Alcorn, MS 39096, USA3 National Center for Natural Products Research, School of Pharmacy, Oxford, MS 38655, USA4 Institute for Behavioral Genetic, University of Colorado, Boulder CO 80309, USA5 Laboratory of Visual Information Processing, Center of Brain & Cognitive Science, Institute of Biophysics, Academia Sinica, Beijing 100101, P.R. China6 Department of Pharmaceutical Science, School of Pharmacy, University of Maryland, Baltimore, MD 21201-1180, USA2005 25 8 2005 6 54 54 2 2 2005 25 8 2005 Copyright © 2005 Gutierrez-Zepeda et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Epidemiological studies have associated estrogen replacement therapy with a lower risk of developing Alzheimer's disease, but a higher risk of developing breast cancer and certain cardiovascular disorders. The neuroprotective effect of estrogen prompted us to determine potential therapeutic impact of soy-derived estrogenic compounds. Transgenic C. elegans, that express human beta amyloid (Aβ), were fed with soy derived isoflavones genistein, daidzein and glycitein (100 μg/ml) and then examined for Aβ-induced paralysis and the levels of reactive oxygen species.
Results
Among the three compounds tested, only glycitein alleviated Aβ expression-induced paralysis in the transgenic C. elegans. This activity of glycitein correlated with a reduced level of hydrogen peroxide in the transgenic C. elegans. In vitro scavenging effects of glycitein on three types of reactive oxygen species confirmed its antioxidant properties. Furthermore, the transgenic C. elegans fed with glycitein exhibited reduced formation of β amyloid.
Conclusion
These findings suggest that a specific soy isoflavone glycitein may suppress Aβ toxicity through combined antioxidative activity and inhibition of Aβ deposition, thus may have therapeutic potential for prevention of Aβ associated neurodegenerative disorders.
==== Body
Background
Estrogen, a natural steroid long associated with effects on the female reproductive system, also plays a role in the central nervous system (CNS) through binding estrogen receptors located in the brain [1,2]. It has been demonstrated that estrogen has neuroprotective and neurotrophic properties [1-9]. Epidemiological studies suggest that post-menopausal women using Estrogen Replacement Therapy (ERT) have a decreased risk of developing dementia [10-12]. However, the beneficial effect of ERT on dementia associated with Alzheimer's disease (AD) is yet inconclusive [13-15]. Although ERT alleviates the symptoms associated with menopause and has positive effects on bones, ERT in post-menopausal women has been linked to a higher incidence of uterine and breast cancer. Consequently, the Selective Estrogen Receptor Modulators (SERMs) compounds that exert tissue specific estrogenic effects may provide the benefits of ERT without the risks. A group of natural SERMs are the soy-derived phytoestrogens, which are structurally similar to estrogen [16], and may serve as an alternative to ERT [17-19].
Soybeans contain a large amount of isoflavones, including genistein (4', 5'7-trihydroxyisoflavone), daidzein (4', 7-dihydroxyisoflavone), glycitein (6-methoxydaidzein) and their glycosides [20]. Experimental evidence suggests that soy isoflavones possess many properties including estrogenic [16], antioxidant [21] hypocholesterolemic [22], and inhibition of cell proliferation and DNA synthesis [23,24]. Phytoestrogens exert estrogen agonist and antagonist characteristics [17], in part because of differential binding affinities for the estrogen receptor (ER) isoforms; with higher affinity for ERβ than for ERα. Areas of the brain responsible for cognitive function and susceptible to AD (basal forebrain, hippocampus, cerebral cortex), express higher levels of ERβ compared to ERα [25]. Thus, interest in these compounds has grown because they could be used as SERMs, to delay or prevent the cognitive decline associated with AD [3,26] without increasing the risk of developing cancer [27].
AD is widely recognized as a serious public health problem [28]. The clinical symptoms of AD begin with memory impairment that eventually progresses to dementia, a process postulated to be the consequence of selective degeneration of nerve cells in those brain regions critical for memory, cognitive performance and personality [29]. AD is characterized by the presence of amyloid beta peptide (Aβ1–42) aggregation and increased oxidative stress, both causing neuronal injury and death [30]. An "amyloid cascade" hypothesis states that accumulation of Aβ deposition initiates a series of downstream neurotoxic events, which result in neuronal dysfunction and death [31]. The strongest evidence supporting this hypothesis comes from molecular genetic studies. Patients with Down's Syndrome, a disease related to an extra copy of chromosome 21 containing the APP gene, develop AD with the formation of Aβ deposits, an early sign of brain lesion [32]. All familiar forms of AD (FAD)-linked mutations, in the APP gene or two presenilin genes (PS1 and PS2), result in increased production of Aβ42, which is the more amyloidogenic form [33]. Transgenic mice overexpressing the mutant APP develop Aβ-containing amyloid plaques similar to those found in AD. Furthermore, inducing toxicity and cognitive dysfunction by introducing Aβ into organisms that do not have endogenous Aβ [34, 56] provided "gain of function" evidence for the "amyloid hypothesis". In addition, other structure lesions including neurofibrillary tangles and AproE might contribute to an imbalance between Aβ production and clearance [31]. Therefore, modulation of Aβ production and clearance in the brain is one approach for treatment of AD.
In order to understand the neuroprotective mechanism of phytoestrogens, we performed several experiments using a transgenic Caenorhabditis elegans model expressing the human amyloid-beta peptide (Aβ1–42). The transgenic C. elegans exhibits β amyloid fluorescence staining similar to those observed in the human brain [34], along with a concomitant progressive paralysis phenotype [35]. Results of these experiments suggest that the neuroprotective effect of phytoestrogens is, at least in part, due to its antioxidative activity.
Results
1. Glycitein alleviates Aβ-induced paralysis in the transgenic C. elegans
A relationship between the onset of Aβ expression and paralysis behavior has been established in the temperature-inducible transgenic C. elegans strain CL4176 [35]. We first conducted the paralysis assay using this strain to determine the effects of the isoflavones on Aβ-induced toxicity in the organism. We have observed in an independent study that the same transgenic C. elegans fed with Ginkgo biloba extract EGb 761, known for its antioxidant properties and beneficial effect for dementia, exhibited a delayed paralysis at the concentration ranging from 10 to 500 μg/ml, and this effect was not dose-dependent (data not shown). Age-synchronized C. elegans (CL4176, 100 worms/group) were fed with daidzein, glycitein, genistein or vehicle for 48 h prior to temperature up shift and then scored for paralysis. Figure 1A is a time course of a paralysis assay comparing a transgenic control strain CL1175, which does not express Aβ, with the Aβ-expressing strain CL4176 to demonstrate the specificity of Aβ-expression induced paralysis. Figure 1B and 1C represent paralysis in four groups of C. elegans CL4176 fed with one of the three different isoflavones (100 μg/ml) or vehicle. Apparently, Aβ-induced paralysis was delayed in worms fed with glycitein (Fig. 1B, filled circle compared with open squares, n = 3 assays, 100 worms/assay). Genistein, known to have more estrogenic activity than diadzein or glycitein [16], did not affect Aβ-induced paralysis in the nematode CL 4176 at the concentration applied (Fig. 1B filed squares, n = 3 assays, 100 worms/assay). The Aβ-induced paralysis was moderately accelerated at the end of the assay in the CL4176 worms fed with daidzein (Fig. 1B filled triangles, n = 3 assays, 100 worms/assay). Figure 1C shows a statistical analysis of the paralysis assays displayed in Fig. 1B. We define PT50 as time duration at which 50% worms were paralyzed from 30 hrs after up shift of temperature to 23°C. Statistically, a significant delay of Aβ-induced paralysis was only observed in the worms fed with glycitein (Fig. 1C, Control, PT50 = 2.6 ± 0.08 h vs. Glycitein, PT50 = 3.3 ± 0.25 h. p = 0.036; Daidzein, PT50 = 2.5 ± 0.10 h, p = 0.46; Genistein, PT50 = 2.6 ± 0.15 h. p = 0.76; n = 3 assays each drug, 40 worms in each assay group). Although Daidzein accelerated paralysis at the end point, PT50 did not indicate significant difference (Fig 1C) compared with that of the controls. It is known that the effective concentration for genistein to activate the estrogen receptor and tyrosine kinases is much lower (nM). Differential concentration effects of genistein might contribute to protection against Aβ toxicity/paralysis. Thus, we conducted experiments using genistein at two lower doses (10 μg/ml and 0.1 μg/ml). Aβ-induced paralysis was not affected in the worms fed with either of the two concentrations (data not shown), supporting the view that the effect of glycitein is specific.
To determine the overall effect of the isoflavones on the behavioral of the C. elegans, we conducted oxidative stress sensitivity assay and life span assay. We found that the C. elegans fed with glycitein were more resistant toward an oxidative stressor Juglone than the worms fed with daidzein and genistein (data not shown). However, the maximum life span was not affected in the C. elegans CL2006 fed with glycitein compared with untreated control worms (data not shown).
2. Glycitein attenuates levels of H2O2 in the Aβ-expressing C. elegans & in vitro
Given that soy isoflavones are potent antioxidants, we determined whether the antioxidative properties of the isoflovones might contribute to protection against Aβ-toxicity. Previously, we established an in vivo assay for the measurement of intracellular H2O2-associated ROS in C. elegans [36]. The transgenic C. elegans were fed with or without the isoflavones, prior to induction of Aβ-expression, followed by measurement of the levels of H2O2 in the organism. Figure 2A demonstrates that the levels of ROS in the C. elegans CL2006 fed with glycitein for 36 h were reduced (control 100 ± 23%, glycitein 68.9 ± 7 %, n = 3, p = 0.05). Although genistein increased the levels of ROS compared with the untreated controls (Ctrl 100 ± 23 %, genistein 126.1 ± 18 %, n = 3, p = 0.28 total 300 worms in each group), it is not statistically significant. Daidzein did not affect Aβ-induced elevation of ROS (Ctrl 100 ± 23%, daidzein 104.4 ± 6%, n = 6, p = 0.74). These results suggest the decreased Aβ toxicity by glycitein might be, in part, a consequence of its antioxidative action.
To confirm the scavenging effect of glycitein on different species of oxidative free radicals in vitro, we first measured its effect on hydroxyl radicals. The hydroxyl free radicals were generated from Fenton reaction (H2O2 3%, FeSO4 0.1 mM and tapped by DMPO (0.1 mol/l). A spectrum with 4 lines and 1:2:2:1 intensity (g = 2.0045, aN = aH = 14.9 G) were obtained (Fig. 2B). Figure 2B demonstrates the signal intensity decrease with different concentrations of the soy isoflavone glycitein added into this system. The soy isoflavone glycitein appears to have very strong scavenging effects on hydroxyl radical generated from Fenton reaction (IC50 = 0.035 mg/ml).
We then determined the scavenging effect of glycitein on superoxide free radicals. The superoxide free radicals were generated from xanthine/xanthine oxidase and trapped by DMPO. A signal with 12 lines (aN = 14.2 G, aHβ = 11.2 G, aHγ = 1.3 G) was obtained (Fig. 2C), and it was decreased with addition of glycitein as shown in Fig. 2C. Apparently, soy isoflavone glycitein has moderate scavenging effect on superoxide free radicals generated from the reaction of xanthine/xanthine oxidase (IC50 = 2 mg/ml).
The reaction of NO with superoxide free radicals is very fast (6.4 × 109mol/L-1s-1) and forms peroxynitrite (ONOO-). In alkaline solution, it is stable but has a pKa of 6.6 at 0°C and decays rapidly once protonated, to hydroxyl radical-like species and NO2, which can oxidize sulfhydryls and membrane lipid causing cell toxicity and some diseases. To determine the scavenging effects of the soy isoflavone glycitein on ONOO-, the methyl free radical was generated from the oxidation of DMSO by ONOO- and trapped by tNB and a spectrum with 12 lines (aN = 17.2 G, aH = 14.2 G) (Zhao et al. 1996) was obtained (Fig. 2D). A strong scavenging effect of glycitein on ONOO- (IC50 = 0.13 mg/ml) was found as shown in Fig. 2D.
3. β amyloid were significantly reduced in transgenic C. elegans fed with glycitein
The modified "amyloid hypothesis" states that Aβ-induced oxidative stress may speed up β amyloid formation and lead to neuronal cell death in AD [37]. To determine whether soy isoflavones affect β amyloid formation in vivo, we measured β amyloid in the transgenic C. elegans CL2006 by thioflavin S staining. β amyloid was stained and the fluorescent images were quantified. Quantitatively (Fig. 3), the mean numbers of β amyloid staining per head area of the nematode are significantly reduced only in the transgenic C. elegans (CL2006) fed with glycitein (4.1 ± 0.4) compared with unfed controls (6.9 ± 0.5). A moderate reduction, although not significant, was observed in the C. elegans fed with genistein (6.1 ± 0.5). No change of Aβ deposits was observed in the worms fed with daidzein (6.9 ± 0.6). None of the three soy isoflavones inhibited Aβ aggregation in vitro (data not shown), suggesting that the decreased β amyloid by glycitein in the transgenic C. elegans (Fig. 1) is not due to its direct binding to Aβ, but might be a consequence of its antioxidative action (Fig. 2).
Discussion
In this study, we employed a transgenic C. elegans model to evaluate the pharmacological effect of the soy-derived isoflavones genistein, glycitein and daidzein, on Aβ-initiated toxicity and oxidative stress. Results of these assays indicate that among the three isoflavones tested, glycitein delayed Aβ induced paralysis and attenuated the levels of amyloid formation in the transgenic C. elegans. In addition, glycitein significantly scavenged hydroxyl free radicals and inhibited the oxidation of peroxynitrite in vitro.
There has been strong evidence for the neuroprotective role of estrogen in aging animal studies and human studies [8,11,26,38-42]. Evidence for estrogens effect on cognition in women with AD is controversial [10,14]. However, it was reported that ovariectomized guinea pigs had a pronounced accumulation of β-amyloid plaques compared to intact controls and that estrogen replacement reversed the accumulation [3]. A proposed mechanism for estrogen inhibition of plaque formation is that estrogen induces the cleavage of membranous amyloid precursor protein (APP) generating a soluble proteolytic fragment that precludes the development of β-amyloid plaque formation [5,6]. The possible link between estrogen and Aβ prompted us to determine the effect of phytoestrogens on Aβ-induced toxicity in a model organism. Knowing that apl-1, the member of APP family in C. elegans, lacks a recognizable Aβ sequence [57, 58], the effect of phytoestrogens may have different mechanisms of action. Phytoestrogens have received increasing attention due to their potential protective effects against age-related diseases and hormone-dependent cancers. Phytoestrogens have the ability to selectively activate estrogen receptors, thus affecting many of the biological responses that are caused by endogenous levels of estrogen without concurrent and undesired side effects. Phytoestrogens may act both as an agonist and antagonist in a tissue specific manner [4]. It was suggested that phytoestrogenes can significantly influence sexually dimorphic cognitive behavior by enhancing spatial memory in young adult female animals but inhibit this ability in male [4].
Our observation that glycitein, with weaker estrogenic activities than genistein and daidzein, inhibited Aβ-induced paralysis and deposition, suggests that neuroprotection by phytoestrogens may not be mediated through the estrogenic activity of the compounds. Compared to other soy isoflavones, the estrogenic activity of glycitein is 20 times lower than genistein and daidzein and 200 times lower than 17β estradiol [16]. Soybeans contain large amounts of glycitein and its glycosides, which have been reported to inhibit growth and DNA synthesis of smooth muscle cells [23].
Apparently, it is the antioxidant activity that contributed to the protective effect by glycitein against Aβ-toxicity (Fig. 1) since glycitein is the only soy isoflavone which significantly attenuated the levels of ROS in the C. elegans (Fig. 2). Oxidative free radicals have been postulated as a cause of aging and of some degenerative diseases [45,46]. The formation of free radicals by Aβ in vitro [46] and profound induction of protein carbonyl in the transgenic C. elegans suggests that Aβ-induced oxidative stress triggers Aβ-induced paralysis in the C. elegans [47]. Although Aβ aggregations have been identified as neurotoxic to the brain, oxidative stress is predicted to occur before these aggregations [47] leading to cell apoptosis. Thus the observed reduction in amyloid formation might also be due to the anti-oxidant activities of glycitein. These observations go along with the free radical hypothesis of aging, which states that there is an imbalance of free radicals and reactive oxygen species (ROS) in the brain causing significant damage to key cellular components [45]. This imbalance may be the causative agent for the pathology of neurodegenerative disorders (such as AD) since most of these disorders are associated with age [48]. The toxicity of free radicals depends on the kinetics of their production, as well as on their stability and transfer efficiency to lipids and proteins. These radicals may interact with other radicals to produce Aβ aggregates [49], and promote the cleavage of the Aβ precursor (APP) supporting the idea that AD can be attributed to continuous oxidative stress, along with a weakened antioxidant status [49].
The causal relationship between ROS and Aβ has been long debated in the field. The transgenic C. elegans would allow us to address the issue. We have conducted a paralysis assay in the C. elegans fed with vitamin C and EGb 761, a Ginkgo biloba leaf extract. Surprisingly, vitamin C alone did not delay Aβ-induced paralysis, but it did when combined with EGb 761, which also inhibits Aβ oligomerization (data not shown), suggesting that it is the combined actions of antioxidants and other protection against Aβ toxicity that is necessary for alleviating Aβ-induced paralysis. Thus, we consider that the antioxidant action is only partially contributing to the protection against Aβ toxicity. Same argument may apply to the discrepancy of the genistein's effect between Fig 2A and Fig 3B; the increased levels of ROS by genistein did not correlate with a decreased Aβ deposition. Defining a functional relationship between Aβ deposition and toxicity, and ROS level is certainly one of our future directions.
The assumption that the protective effect by glycitein against Aβ toxicity might not be mediated by its action on the estrogen receptor is supported by our observation that genistein, with strongest estrogenic activity among soy isoflavones, did not offer protection again Aβ-toxicity. Genistein is a known tyrosine kinase inhibitor. The effective concentration for genistein to activate the estrogen receptor and inhibit tyrosine kinases is much lower (nM-μM) than the concentration we applied to the worms [4] and [5]. These differential concentration effects of genitein might offer protection against Aβ toxicity/paralysis. However, our additional experiments using much lower dosage of genistein did not provide evidence to support this notion. Aβ-induced paralysis was not affected in the worms fed with either of the two lower concentrations. Since at the given concentration (10 μg/ml, i.e. 37 μM), we observed effects with glycitein but not genistein, we assume that they have differential effects on Aβ-induced paralysis. It has been shown that high dose of genistein (μM) could cause apoptosis in rat primary cortical neurons in vitro via a calcium dependent pathway [43].
We demonstrated a consistent, correlative effect by glycitein against Aβ-induced toxicities using different assays, which suggests that C. elegans is a valid model for mechanistic examination of the transgene products as well as for pharmacological analysis of time course and kinetics of drug effect [50,51]. A relationship between Aβ amino-acid sequence, amyloid formation and oxidative damage was established using this model. Yatin et al. [46] showed both in vitro and in the C. elegans model that methionine (Met35) is critical for free radical production by Aβ1–42, and it is also critical for β-sheet formation in the transgenic C. elegans lines [52]. A correlation between a progressed paralysis phenotype with increased levels of protein carbonyls in CL4176 [47] supports the advanced "amyloid hypothesis" [37]. Mammalian αB-crystallin (CRYAB) a stress-inducible chaperone protein, which inhibits fibril formation of Aβ-(1–42) [53], has a protein homologue HSP-16 in the C. elegans. This protein has been reported to be colocalized with intracellular Aβ and up regulated in the transgenic Aβ-expression strain of C. elegans [35]. We previously demonstrated that a neuroprotectant, EGb 761, an extract from the ginkgo biloba tree leave, suppressed HSP-16 expression [54]. Although many protein molecules including estrogenic receptors are conserved in the nematode [55], the lack of correlation between isoflavone estrogenic activity and suppression of Aβ toxicity in this model system may not exclude the neuroprotection estrogen in AD patients. Nevertheless, it is likely that the temporal sequence of events manifested in the transgenic worms is the same as the one demonstrated in a Drosophila model of AD [56] in that accumulation of Aβ42 in the brain is sufficient to cause cognitive impairment and neurodegeneration.
Conclusion
We used a transgenic C. elegans model to evaluate the pharmacological effect of the soy-derived isoflavones genistein, glycitein and daidzein, on Aβ-initiated toxicity and oxidative stress. Among the three compounds tested, only glycitein alleviated Aβ expression-induced paralysis in the transgenic C. elegans, which correlated with a reduced level of hydrogen peroxide and β amyloid. These findings suggest that the neuroprotective effect of phytoestrogens is probably due, at least in part, to its antioxidative activities.
Methods
Soy isoflavones were obtained from the National Natural Products Research Center (Oxford, MS). Stock solutions of the soy isoflavones (1 mg/ml, 1000× stock solution) were made in 100% ethanol. The final concentration of ethanol in the food did not exceed 0.01%. DMPO (5,5-dimethyl-1-pyroline-1-oxide, tNB(3,3,5,5-tetramethyl-pyrroline N-oxide) were purchased from Sigma Chem Co. DMPO was purified by active charcoal.
C. elegans strains
The construction and characterization of the transgenic nematode strains CL2006 and CL4176 have been described previously [34,35]. The CL2006 strain constitutively produces a muscle-specific Aβ1–42, whereas the expression of Aβ1–42 in CL4176 depends on a temperature up-shift from 16 to 23°C. Age-synchronized wild type (N2) and the transgenic CL2006 were propagated at 20°C in a temperature-controlled incubator (Sheldon Manufacturing, Model 2005, Cornelius, OR), CL4176 at 16°C, on solid nematode growth medium (NGM) seeded with E. coli (OP50) for food. All chemicals for treatment of experimental animals were added directly to the OP50 food source and began when larvae were 2 days old (for CL2006). In most cases, the nematodes were treated for 4 days (after hatching) with their respective drug. In the life span assay, the C. elegans were treated with the drug for the duration of their lifetime.
Paralysis assays
C. elegans strain CL4176 [35,47] was maintained at 16°C and egg-synchronized onto 35 × 10 mm culture plates containing vehicle or drug. The worms (100 worms on each plate) were allowed to grow for 38 h at 16°C. After 38 hours the temperature was up shifted to 23°C to induce Aβ expression. Paralysis was scored at 1 h intervals until all worms were paralyzed
H2O2 assay in C. elegans
Intracellular levels of H2O2-related reactive oxidative species (ROS) were measured in C. elegans using 2,7-dichlorofluorescein diacetate (DCF-DA; Molecular Probes). At the end of the specified treatment times, the C. elegans were collected into 100 μl phosphate buffered saline (PBS) (molarity) with 1% Tween-20 (PBST) in eppendorf tubes. The worms were then sonicated (Branson Sonifier 250, VWR Scientific, Suwanee, GA) and pipetted into wells of 96-well plates containing DCF-DA (final concentration 50 μM in PBS). Samples were read every 10 min for 2.5 h. in an FLx800 Microplate Fluorescent Reader (Bio-Tek Instruments, Winooski, VT) at 37°C at excitation 485 nm and emission 530 nm.
ESR assay of free radicals
In order to measure the effect of glycitein on free radicals, the spin trap and the system-generated free radicals were mixed and measured with ESR spectrometer and the signal intensity was taken as Ho. Then the system with addition of glycitein was measured again. Hydroxyl radicals (H2O2 3%), Fe2SO4 (0.1 mM) and DMPO (0.1 mol/l) were mixed and sucked into a quartz capillary for ESR measurement, and the signal intensity was taken as Hx. The scavenging effect was calculated by [(Ho-Hx)/Ho] × 100%. The ESR spectra were recorded with Brucker ER200 D-SRC ESR spectrometer. Parameters were employed as follows: X-band, 100 kHz modulation with amplitude 1 G, microwave power 10 mW, central magnetic field 3,250 G, sweep width 200 G, temperature 20°C.
Fluorescent staining and quantitation of β amyloid
Individual CL2006 transgenic nematodes were fixed in 4% paraformaldehyde/PBS, pH 7.4, for 24 h at 4°C, and then permeabilized in 5% fresh β-mercaptoethanol, 1% Triton X-100, 125 mM Tris pH 7.4, in a 37°C incubator for 24 h. The nematodes were transferred, stained with 0.125% thioflavin S (Sigma) in 50% ethanol for 2 min, destained for 2 min in 50% ethanol, washed with PBS and mounted on slides for microscopy. Fluorescence images were acquired using a 40× objective of a fluorescence microscope (BX 60, Olympus, Tokyo, Japan) equipped with a digital camera (Micropublisher 5.0, QIMAGING, Burnaby BC, Canada). The Thioflavin S-reactive deposits anterior of the pharyngeal bulb in individual animals were scored.
Statistical analyses
All statistical tests were performed using a PC-based version of the statistical program Origin 6.0 software (Microcal Software, Inc., Northampton, MA). Statistical comparisons between treatments were done with unpaired student t-test. All figures indicate means and standard error of the mean. Differences with a p value less than 0.05 were defined as statistically significant.
List of abbreviations used
AD, Alzheimer's disease
ROS, reactive oxygen species
H2O2, hydrogen peroxide
Aβ, beta amyloid peptide
ERT, Estrogen Replacement Therapy
SERMs, Selective Estrogen Receptor Modulators
Authors' contributions
AGZ carried out the paralysis assay, the oxidative stress assay and measurement of levels of ROS. MB conducted some of the oxidative stress assays. ZW performed the fluorescence staining for Aβ deposits and the qantitation; JW performed additional experiments for the revision; IK provided the soy isoflavones; CL generated the transgenic C. elegans; BZ participated in the design and analysis of the experiments regarding in vitro assay of ROS; YL participated in the general design of the study, organized collaboration as well as finalizing the manuscript. All authors have read and approved the final manuscript.
Acknowledgements
This study was supported by The Mississippi Functional Genomic Network from a NIH-NCRR grant P20RR64176 (YL and RS), by NIH grant R01AT001928-01A1 from NIH National Center for Complimentary and Alternative Medicine (YL), and by a grant from the National Natural Science Foundation of P. R. China (BZ).
Figures and Tables
Figure 1 Paralysis assays in the transgnic C. elegans. A. Time course of paralysis in the transgenic strain CL4176 and the control strain CL1175. B. Paralysis assay in C. elegans CL4176 fed with different isoflavones. Synchronized eggs were maintained at 16°C for 38 h, on the 35 × 10 mm culture plates (~100 eggs/plate) containing vehicle (control), daidzein, glycitein, or genistein (100 μg/ml), followed by up-shifting the temperature to 23°C to induce the transgene expression. The paralysis was scored at 60 min intervals. Data are expressed as percentage of non-paralyzed worms from three independent assays of 100 worms in each experiment. C. The paralysis assays were quantitated for mean time duration at which 50% worms were paralyzed from 30 hrs after up shift temperature to 23°C (PT50). P values were obtained from 3 independent assays for the worms fed with different drugs each paired with untreated controls. Total 100 worms were used in each assay.
Figure 2 Scavenging effect of glycitein in the transgenic C. elegans and in vitro. A. H2O2 level in transgenic C. elegans CL4176 fed with different isoflavones. C. elegans strain CL4176 was maintained at 16°C for 38 h and then temperature up-shifted to 23°C for 48 h, followed by measurement of H2O2 (DCF assay described in methods). CL4176 worms were fed vehicle (Ctrl), 100 μg/ml daidzein, genistein or glycitein from 1 day of age until 3 days of age. At least 60 worms from each group were analyzed for levels of H2O2. Results are expressed as a percentage of fluorescence (%DCF) relative to control. B. Scavenging effect of glycitein on hydroxyl radicals in vitro. The ESR conditions: X-band, 100 kHz modulation with amplitude 1 G, microwave power 10 mW, central magnetic field 3,250 G, sweep width 200 G, temperature 20°C. Inset: ESR spectrum of DMPO-OH generated from Fenton reaction and trapped by DMPO. C. Scavenging effect of glycitein on superoxide radicals in the system. The ESR conditions are the same as in Fig. 3B. Inset: ESR spectrum of DMPO-OOH generated from Xanthine/xanthine oxidase and trapped by DMPO; D. Scavenging effect of Soy isoflavone glycitein on .CH3 free radicals in the in vitro system. Inset: ESR spectrum of CH3-tNB generated from the oxidation of DMSO by ONOO-and trapped by tNB.
Figure 3 Effect of soy isoflavones on Aβ deposits in transgenic C. elegans. A. Representative images of thioflavin S staining in the transgenic (left) or wild type (right) worms. B. Quantitative Aβ formation. β amyloid were stained with thioflavin S in C. elegans CL2006 fed with or without isoflavones (100 μg/ml) for 4 days starting at the second day of age. β amyloid were examined using a fluorescence microscope. The quantity of β amyloid is expressed as mean number of Aβ deposits/worm head area (n = 24).
==== Refs
Behl C Skutella T Lezoualc'h F Post A Widmann M Newton CJ Holsboer F Neuroprotection against oxidative stress by estrogens: structure-activity relationship Mol Pharmacol 1997 51 535 541 9106616
Li R Shen Y Estrogen and brain: synthesis, function and diseases Front Biosci 2005 10 257 267 15574366
Gandy S Estrogen and neurodegeneration Neurochem Res 2003 28 1003 1008 12737524 10.1023/A:1023246921127
Lund TD West TW Tian LY Bu LH Simmons DL Setchell KD Adlercreutz H Lephart ED Visual spatial memory is enhanced in female rats (but inhibited in males) by dietary soy phytoestrogens BMC Neurosci 2001 2 20 11801187 10.1186/1471-2202-2-20
Goodman Y Bruce AJ Cheng B Mattson MP Estrogens attenuate and corticosterone exacerbates excitotoxicity, oxidative injury, and amyloid beta-peptide toxicity in hippocampal neurons J Neurochem 1996 66 1836 1844 8780008
Kim H Bang OY Jung MW Ha SD Hong HS Huh K Kim SU Mook-Jung I Neuroprotective effects of estrogen against beta-amyloid toxicity are mediated by estrogen receptors in cultured neuronal cells Neurosci Lett 2001 302 58 62 11278111 10.1016/S0304-3940(01)01659-7
Mook-Jung I Joo I Sohn S Kwon HJ Huh K Jung MW Estrogen blocks neurotoxic effects of beta-amyloid (1-42) and induces neurite extension on B103 cells Neurosci Lett 1997 235 101 104 9406879 10.1016/S0304-3940(97)00632-0
Green PS Simpkins JW Neuroprotective effects of estrogens: potential mechanisms of action Int J Dev Neurosci 2000 18 347 358 10817919 10.1016/S0736-5748(00)00017-4
Cordey M Gundimeda U Gopalakrishna R Pike CJ Estrogen activates protein kinase C in neurons: role in neuroprotection J Neurochem 2003 84 1340 1348 12614334 10.1046/j.1471-4159.2003.01631.x
Yoon BK Kim DK Kang Y Kim JW Shin MH Na DL Hormone replacement therapy in postmenopausal women with Alzheimer's disease: a randomized, prospective study Fertil Steril 2003 79 274 280 12568834 10.1016/S0015-0282(02)04666-6
Zec RF Trivedi MA The effects of estrogen replacement therapy on neuropsychological functioning in postmenopausal women with and without dementia: a critical and theoretical review Neuropsychol Rev 2002 12 65 109 12371603 10.1023/A:1016880127635
LeBlanc ES Janowsky J Chan BK Nelson HD Hormone replacement therapy and cognition: systematic review and meta-analysis Jama 2001 285 1489 1499 11255426 10.1001/jama.285.11.1489
Mulnard RA Cotman CW Kawas C van Dyck CH Sano M Doody R Koss E Pfeiffer E Jin S Gamst A Grundman M Thomas R Thal LJ Estrogen replacement therapy for treatment of mild to moderate Alzheimer disease: a randomized controlled trial. Alzheimer's Disease Cooperative Study Jama 2000 283 1007 1015 10697060 10.1001/jama.283.8.1007
Thal LJ Thomas RG Mulnard R Sano M Grundman M Schneider L Estrogen levels do not correlate with improvement in cognition Arch Neurol 2003 60 209 212 12580705 10.1001/archneur.60.2.209
Zec RF Trivedi MA Effects of hormone replacement therapy on cognitive aging and dementia risk in postmenopausal women: a review of ongoing large-scale, long-term clinical trials Climacteric 2002 5 122 134 12051107
Song TT Hendrich S Murphy PA Estrogenic activity of glycitein, a soy isoflavone J Agric Food Chem 1999 47 1607 1610 10564025 10.1021/jf981054j
Zhao L Chen Q Diaz Brinton R Neuroprotective and neurotrophic efficacy of phytoestrogens in cultured hippocampal neurons Exp Biol Med (Maywood) 2002 227 509 519 12094016
Tikkanen MJ Wahala K Ojala S Vihma V Adlercreutz H Effect of soybean phytoestrogen intake on low density lipoprotein oxidation resistance Proc Natl Acad Sci U S A 1998 95 3106 3110 9501223 10.1073/pnas.95.6.3106
Zeng H Chen Q Zhao B Genistein ameliorates beta-amyloid peptide (25-35)-induced hippocampal neuronal apoptosis Free Radic Biol Med 2004 36 180 188 14744630 10.1016/j.freeradbiomed.2003.10.018
Murphy PA Farmakalidis E Johnson LD Isoflavone content of soya-based laboratory animal diets Food Chem Toxicol 1982 20 315 317 7201958 10.1016/S0015-6264(82)80185-5
Carroll KK Review of clinical studies on cholesterol-lowering response to soy protein J Am Diet Assoc 1991 91 820 827 2071797
Anthony MS Clarkson TB Williams JK Effects of soy isoflavones on atherosclerosis: potential mechanisms Am J Clin Nutr 1998 68 1390S 1393S 9848505
Pan W Ikeda K Takebe M Yamori Y Genistein, daidzein and glycitein inhibit growth and DNA synthesis of aortic smooth muscle cells from stroke-prone spontaneously hypertensive rats J Nutr 2001 131 1154 1158 11285318
Santell RC Kieu N Helferich WG Genistein inhibits growth of estrogen-independent human breast cancer cells in culture but not in athymic mice J Nutr 2000 130 1665 1669 10867033
Kuiper GG Lemmen JG Carlsson B Corton JC Safe SH van der Saag PT van der Burg B Gustafsson JA Interaction of estrogenic chemicals and phytoestrogens with estrogen receptor beta Endocrinology 1998 139 4252 4263 9751507 10.1210/en.139.10.4252
Fillit HM The role of hormone replacement therapy in the prevention of Alzheimer disease Arch Intern Med 2002 162 1934 1942 12230415 10.1001/archinte.162.17.1934
Miltyk W Craciunescu CN Fischer L Jeffcoat RA Koch MA Lopaczynski W Mahoney C Crowell J Paglieri J Zeisel SH Lack of significant genotoxicity of purified soy isoflavones (genistein, daidzein, and glycitein) in 20 patients with prostate cancer Am J Clin Nutr 2003 77 875 882 12663286
Brookmeyer R Gray S Kawas C Projections of Alzheimer's disease in the United States and the public health impact of delaying disease onset Am J Public Health 1998 88 1337 1342 9736873
Price DL Sisodia SS Borchelt DR Alzheimer disease--when and why? Nat Genet 1998 19 314 316 9697686 10.1038/1196
Selkoe DJ Alzheimer's disease: genotypes, phenotypes, and treatments Science 1997 275 630 631 9019820 10.1126/science.275.5300.630
Hardy J Selkoe DJ The amyloid hypothesis of Alzheimer's disease: progress and problems on the road to therapeutics Science 2002 297 353 356 12130773 10.1126/science.1072994
Mann DM Cerebral amyloidosis, ageing and Alzheimer's disease; a contribution from studies on Down's syndrome Neurobiol Aging 1989 10 397 9; discussion 412-4 2530459 10.1016/0197-4580(89)90073-0
Sherrington R Rogaev EI Liang Y Rogaeva EA Levesque G Ikeda M Chi H Lin C Li G Holman K Cloning of a gene bearing missense mutations in early-onset familial Alzheimer's disease Nature 1995 375 754 760 7596406 10.1038/375754a0
Link CD Expression of human beta-amyloid peptide in transgenic Caenorhabditis elegans Proc Natl Acad Sci U S A 1995 92 9368 9372 7568134
Link CD Taft A Kapulkin V Duke K Kim S Fei Q Wood DE Sahagan BG Gene expression analysis in a transgenic Caenorhabditis elegans Alzheimer's disease model Neurobiol Aging 2003 24 397 413 12600716 10.1016/S0197-4580(02)00224-5
Smith JV Luo Y Elevation of Oxidative Free Radicals in Alzheimer's Disease Models Can Be Attenuated by Ginkgo biloba Extract EGb 761 J Alzheimers Disease 2003 5 287 300 14624024
Butterfield DA Amyloid beta-peptide [1-42]-associated free radical-induced oxidative stress and neurodegeneration in Alzheimer's disease brain: mechanisms and consequences Curr Med Chem 2003 10 2651 2659 14529455 10.2174/0929867033456422
Lee SJ McEwen BS Neurotrophic and neuroprotective actions of estrogens and their therapeutic implications Annu Rev Pharmacol Toxicol 2001 41 569 591 11264469 10.1146/annurev.pharmtox.41.1.569
Wise PM Dubal DB Wilson ME Rau SW Bottner M Rosewell KL Estradiol is a protective factor in the adult and aging brain: understanding of mechanisms derived from in vivo and in vitro studies Brain Res Brain Res Rev 2001 37 313 319 11744096 10.1016/S0165-0173(01)00136-9
Wise PM Dubal DB Wilson ME Rau SW Bottner M Minireview: neuroprotective effects of estrogen-new insights into mechanisms of action Endocrinology 2001 142 969 973 11181507 10.1210/en.142.3.969
Roof RL Hall ED Gender differences in acute CNS trauma and stroke: neuroprotective effects of estrogen and progesterone J Neurotrauma 2000 17 367 388 10833057
Roof RL Hall ED Estrogen-related gender difference in survival rate and cortical blood flow after impact-acceleration head injury in rats J Neurotrauma 2000 17 1155 1169 11186229
Harman D Ageing: a theory based on free radical and radiation chemistry J Gerontol 1957 2 298 300
Yatin SM Varadarajan S Link CD Butterfield DA In vitro and in vivo oxidative stress associated with Alzheimer's amyloid beta-peptide (1-42) Neurobiol Aging 1999 20 325 30; discussion 339-42. 10588580 10.1016/S0197-4580(99)00056-1
Drake J Link CD Butterfield DA Oxidative stress precedes fibrillar deposition of Alzheimer's disease amyloid beta-peptide (1-42) in a transgenic Caenorhabditis elegans model Neurobiol Aging 2003 24 415 420 12600717 10.1016/S0197-4580(02)00225-7
Christen Y Maixent JM What is Ginkgo biloba extract EGb 761? An overview--from molecular biology to clinical medicine Cell Mol Biol 2002 48 601 611 12396070
Hensley K Carney JM Mattson MP Aksenova M Harris M Wu JF Floyd RA Butterfield DA A model for beta-amyloid aggregation and neurotoxicity based on free radical generation by the peptide: relevance to Alzheimer disease Proc Natl Acad Sci U S A 1994 91 3270 3274 8159737
Linford NJ Yang Y Cook DG Dorsa DM Neuronal apoptosis resulting from high doses of the isoflavone genistein: role for calcium and p42/44 mitogen-activated protein kinase J Pharmacol Exp Ther 2001 299 67 75 11561064
Gutierrez-Zepeda A Luo Y Testing the amyloid toxicity hypothesis of alzheimer s disease in transgenic Caenorhabditis elegans model Front Biosci 2004 9 3333 3338 15353361
Driscoll M Gerstbrein B Dying for a cause: invertebrate genetics takes on human neurodegeneration Nat Rev Genet 2003 4 181 194 12610523 10.1038/nrg1018
Fay DS Fluet A Johnson CJ Link CD In vivo aggregation of beta-amyloid peptide variants J Neurochem 1998 71 1616 1625 9751195
Fonte V Kapulkin V Taft A Fluet A Friedman D Link CD Interaction of intracellular beta amyloid peptide with chaperone proteins Proc Natl Acad Sci U S A 2002 99 9439 9444 12089340 10.1073/pnas.152313999
Strayer A Wu Z Christen Y Link CD Luo Y Expression of small heat-shock protein Hsp16-2 in Caenorhabditis elegans is suppressed by ginkgo biloba extract EGb 761 FASEB J 2003 17 2305 2307 14525938
Shao W Halachmi S Brown M ERAP140, a conserved tissue-specific nuclear receptor coactivator Mol Cell Biol 2002 22 3358 3372 11971969 10.1128/MCB.22.10.3358-3372.2002
Iijima K Liu HP Chiang AS Hearn SA Konsolaki M Zhong Y Dissecting the pathological effects of human Abeta40 and Abeta42 in Drosophila: a potential model for Alzheimer's disease Proc Natl Acad Sci U S A 2004 101 6623 6628 15069204 10.1073/pnas.0400895101
|
16122394
|
PMC1215487
|
CC BY
|
2021-01-04 16:03:48
|
no
|
BMC Neurosci. 2005 Aug 25; 6:54
|
utf-8
|
BMC Neurosci
| 2,005 |
10.1186/1471-2202-6-54
|
oa_comm
|
==== Front
BMC PediatrBMC Pediatrics1471-2431BioMed Central London 1471-2431-5-321612022910.1186/1471-2431-5-32Research ArticleParents' psychological adjustment in families of children with Spina Bifida: a meta-analysis Vermaes Ignace PR [email protected] Jan MAM [email protected] Anna MT [email protected] Jan RM [email protected] Institute of Family and Child Care Studies, Radboud University Nijmegen, Montessorilaan 3, P.O. Box 9104, 6500 HE Nijmegen, The Netherlands2 Department of Educational Sciences, Radboud University Nijmegen, Montessorilaan 3, P.O. Box 9104, 6500 HE Nijmegen, The Netherlands2005 25 8 2005 5 32 32 5 4 2005 25 8 2005 Copyright © 2005 Vermaes et al; licensee BioMed Central Ltd.2005Vermaes et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Spina Bifida (SB) is the second most common birth defect worldwide. Since the chances of survival in children with severe SB-forms have increased, medical care has shifted its emphasis from life-saving interventions to fostering the quality of life for these children and their families. Little is known, however, about the impact of SB on family adjustment. Reviewers have struggled to synthesize the few contradictory studies available. In this systematic review a new attempt was made to summarize the findings by using meta-analysis and by delimiting the scope of review to one concept of family adjustment: Parents' psychological adjustment. The questions addressed were: (a) do parents of children with SB have more psychological distress than controls? (b) do mothers and fathers differ? and (c) which factors correlate with variations in psychological adjustment?
Methods
PsycInfo, Medline, and reference lists were scanned. Thirty-three relevant studies were identified of which 15 were eligible for meta-analysis.
Results
SB had a negative medium-large effect on parents' psychological adjustment. The effect was more heterogeneous for mothers than for fathers. In the reviewed studies child factors (age, conduct problems, emotional problems, and mental retardation), parent factors (SES, hope, appraised stress, coping, and parenting competence), family factors (family income, partner relationship, and family climate), and environmental factors (social support) were found to be associated with variations in parents' psychological adjustment.
Conclusion
Meta-analysis proved to be helpful in organizing studies. Clinical implications indicate a need to be especially alert to psychological suffering in mothers of children with SB. Future research should increase sample sizes through multi-center collaborations.
==== Body
Background
Worldwide, SB is the second most common congenital birth defect [1]. Its prevalence varies per geographic region, depending on genetic and environmental factors [2]. In developing countries the occurrence of SB tends to be higher than in Western countries. For example, in Tanzania the incidence of SB live births is estimated at 1.35‰ [3] whereas in the US, the incidence of SB pregnancies is estimated at 0.41‰ [4] and the number of SB live births at 0.21‰ [5]. Thus, despite primary prevention programs, such as the fortification of cereal grain products with folic acid in the US [5], and despite estimates that at least 40% of the early detected SB pregnancies in Europe are terminated [6], the number of children who are born with SB remains substantial.
Children with spina bifida (SB) live with a range of disabilities, depending on where in the spinal column formation the defect is located and whether it is closed or open. Since the mid 1960s, early surgical treatment of SB has increased the survival rates of children with severe forms of SB and in more recent years, the development of prenatal surgery around the 20th week of pregnancy has further improved children's chances of survival [7]. Consequently, medical teams face the task of fostering the quality of life for these children and their families. On the one hand, enhancing the quality of life depends on medical advances (e.g. urological, orthopedic, and hydrocephalus research). On the other hand, it depends on the development of a scientifically based understanding of the psychosocial aspects involved with chronic illness in general and SB in particular [8].
To date, a limited number of studies have investigated psychosocial aspects of SB. Typically, these studies have focused on two broad topics: (1) the impact of SB on the child and (2) the impact of SB on the family [8]. Although attempts have been made to integrate findings [8-12], most reviewers have struggled to draw conclusions on family adjustment to SB. One problem is the dearth of empirically sound studies. Another problem is the small number of studies with theoretically driven research questions and hypotheses [8]. Both problems have led to a fragmented picture of mixed findings, because the few studies available have investigated outcome variables reflecting different levels of family functioning (e.g. marital adjustment, parenting stress, and family atmosphere) as indicators of family adjustment. Based on family-systems theory and family-resilience theory it can be argued that SB will have a differential effect on different levels of the family structure [13].
Therefore, in this review a new attempt was made to synthesize findings by concentrating on one level of family adjustment only: parental adjustment. Moreover, the traditional narrative methods used by earlier reviews were replaced with statistical meta-analysis to summarize findings more systematically. The goal of this approach was to exhaust the limited studies available to maximize the information concerning parents' adjustment to having a child with SB.
Conceptualization of adjustment: psychological adjustment
A preliminary inventory of the literature uncovered that we could divide the research on parental adjustment to SB into three areas: (1) psychological adjustment, (2) interpersonal adjustment in the dyadic partner and parenting relationships, and (3) parents' perceptions of the family atmosphere. The inspiration to discern these areas of adjustment stemmed from Wallander's model of maternal adjustment with chronic illness [14]. In this model the areas mental health, physical health and social functioning are distinguished as indicators of maternal adjustment.
Parents' psychological adjustment can be defined as the adaptive task of managing upsetting feelings aroused by the illness of the child and preserving a reasonable emotional balance [15]. Pless and Pinkerton [16] have postulated that adjustment to chronic illness changes over time and that at any given moment psychological adjustment will reflect the cumulative product of earlier transactions. Thus, on the one hand parents' psychological adjustment reflects the outcome of parents' ability to maintain a balance between the demands of stressful situations and the availability of personal (e.g. optimism) and social resources (e.g. partner support), whereas on the other hand, parents' psychological adjustment enhances the accomplishment of other general adaptive tasks, such as: preserving a satisfactory self-image, keeping the family together, and preparing for an uncertain future, as well as the accomplishment of illness-related tasks, for example: dealing with the symptoms of the illness, dealing with treatment related stressors, and establishing functional relationships with health caregivers [15]. Positive experiences in achieving such tasks will in turn enforce parents' emotional balance through so called positive-feedback loops [16].
Based on these ideas, we opted to delimit the concept of parental adjustment to parents' psychological adjustment. Only studies using psychological outcomes as indicators of parents' adjustment to SB were included in the review.
Hypotheses and research questions
A considerable number of studies have been devoted to children with severe disabilities and their families [17]. Two approaches have emerged: A categorical and a non-categorical. Categorical studies aim at investigating the unique effects of a specific disease on family life, for example SB, whereas non-categorical studies aim at examining shared effects of different chronic diseases on family life [18].
From non-categorical accounts a few broad hypotheses regarding parents' psychological adjustment with SB were derived. Parents of children with physical impairments have been found to report higher levels of stress, anxiety, and depression than parents of able-bodied children [19], however parents' adjustment to chronic illness has also been found to be marked by great individual variation [14,16]. Studies have yielded several conceptual models based on stress-coping theories and socio-ecological views on family functioning to explain the differential effects of chronic illness on parents' adjustment [19]. In short, most of these models view the child's chronic illness as a potential stressor. The severity of the illness and associated delays in the child's development are expected to determine the functional care strain on the family as a whole and on parents in particular. Besides illness-related stressors, other major life events and daily hassles may add to the demands on parents. Stress-coping theories maintain that the extent to which parents are negatively affected by these demands, will depend on how they appraise, or give meaning, to them. In the process of appraisal parents estimate how their personal capacities and their resources of social support meet the demands of stressful situations. The personal capacities to interact with stressful situations are determined by parents' personality characteristics, coping styles, and strategies. The social resources are determined by the extent to which parents have access to emotional as well as instrumental support from their relationships with others, for example, marital support, family support, informal support from extended family and friends, and formal support from professional caregivers. Depending on how parents estimate the balance between the illness-related stressors, their personal capacities, and their social resources, they can be expected to have more or less difficulties to adjust to having a child with SB. Thus, variability in parents' psychological adjustment can be expected to be associated with multiple factors concerning: characteristics of the child (e.g. severity of illness and developmental delays), characteristics of the parent (e.g. personality characteristics and coping styles), characteristics of the family (e.g. marital quality and family climate), and characteristics of parents' environment outside the family (e.g. social support from extended family and friends).
Although most studies have focused on maternal adjustment to chronic illness, individual differences may be expected between mothers and fathers because of role differentiations in care and work [19]. Mothers are often their child's main caregiver. Consequently, they are more exposed to illness-related situations than fathers and may therefore experience more psychological stress than fathers.
In this review the above premises were studied, guided by three research questions identified in the literature on parents' adjustment with SB: (a) do parents of children with SB have higher levels of psychological distress than do parents of able-bodied children? (b) do mothers and fathers differ in psychological adjustment? and (c) which factors are correlated with parents' psychological adjustment? Four categories of factors were discerned: (a) child factors, (b) parent factors, (c) family factors, and (d) other environmental factors.
Methods
Identification of studies
For the meta-analysis, primary research reports were located and coded in four steps:
Step 1: Identification of studies on parents' adjustment
The PsycInfo and Medline databases from 1966 to January 2005 were scanned using the key terms "spina bifida" or "neural tube defect" (NTD) or "myelomeningocele" (MMC) and "family" or "parenting" or "parents" and "adjustment" or "adaptation". This resulted in 925 abstracts. Two reviewers (IV and JJ) selected 65 abstracts based on the following inclusion criteria: (1) available in English, (2) reported primary research, and (3) studied parents' adjustment with SB. Agreement between raters was 96.6% (Cohen's Kappa = .92). Differences between reviewers were resolved through discussion.
The reference lists of the 65 reports were scanned to check whether other studies had been missed in the first scan of PsycInfo and Medline. Despite this check, one report was overlooked because at first glance its appearance was similar to another report of the same authors published in the same year [20,21].
Step 2: Selection of studies on parents' psychological adjustment
The two reviewers coded each publication with regard to the area of parents' adjustment. Three areas were distinguished: (1) individual psychological adjustment, (2) interpersonal adjustment in dyadic partner and parent-child relationships, and (3) perceptions of family functioning. The coders found that 33 out of 66 studies reported findings on psychological adjustment. Their interrater reliability was 90.8% (Cohen's Kappa = .82). Total agreement was achieved through discussion.
Step 3: Coding of research reports
The 33 studies were classified by study and sample characteristics (see Table 1). The study characteristics were: number of participants, design, presence of comparison group, and outcome measure. The sample characteristics were: parent gender, child impairment, child age, and treatment timing of SB. The two coders agreed between 87% and 100% (Cohen's Kappa = .84 to 1.00). Discussion led to total agreement.
Table 1 Study and sample characteristics of reports on parents' psychological adjustment
Reports included in meta-analysis N Design Comparison Group Parent Gender Child Impairment1 Child Age Child Treatment Outcome Measure2
Barakat & Linney, 1992 [31] 29 Prospective Control Mothers SB
(MMC-non retarded) 6–11 Early BSI
Barakat & Linney, 1995 [32] 29 Prospective Control Mothers SB
(MMC-non retarded) 6–11 Early BSI
Evans, Tew, & Laurence, 1986 [48] 124 Longitudinal Control Fathers Combined: NTD 18 Late General Health Questionnaire
Fagan & Schor, 1993 [61] 50 Prospective Norm scores Mothers SB M = 8.1 Early Malaise Inventory
Holmbeck, Gorey-Ferguson, Hudson, Seefeldt, Shapera, Turner, & Uhler, 1997 [43] 55 Prospective Control Mothers & fathers SB 8–9 Early SCL-90R
Horton & Wallander, 2001 [23] 33 Prospective Norm scores Mothers SB M = 10.6 Early BSI
Kazak & Marvin, 1984 [62] 56 Prospective Control Mothers & fathers SB (MMC) 1–16 Early Langner Symptom Checklist
King, King, Rosenbaum, & Goffin, 1999 [22] 164 Prospective Norm scores Mothers & fathers Combined: CP, SB, NOS 3–6 Early SCL-90R
Kronenberger & Thompson, 1992 [20] 66 Prospective Norm scores Mothers SB (MMC) 0–18 Early SCL-90R
Kronenberger & Thompson, 1992 [21] 66 Prospective Norm scores Mothers SB (MMC) 0–18 Early SCL-90R
Lemanek, Jones, & Lieberman, 2000 [56] 59 Prospective Norm scores Mothers SB-non retarded 3–16 Early SCL-90R
Tew & Laurence, 1973 [33] 51 Longitudinal Norm scores Mothers SB M = 11.6 Late Malaise Inventory
Tew & Laurence, 1975 [34] 51 Longitudinal None Mothers SB M = 11.6 Late Malaise Inventory
Wallander, Varni, Babani, DeHaan, Thompson, Wilcox, & Tweddle Banis, 1989 [14] 50 Prospective Norm scores Mothers Combined: SB, CP 6–11 Early Malaise Inventory
Wiegner & Donders, 2000 [45] 34 Prospective Norm scores Mothers SB 3–12 Early BSI
Reports excluded from meta-analysis N Design Comparison Group Parent Gender Child Impairment1 Child Age Child Treatment Outcome Measure2
Dorner, 1973 [63] 63 Prospective None Mothers SB 13–19 Late Malaise Inventory
Dorner, 1974 [44] 63 Prospective None Mothers SB 13–19 Late Malaise Inventory
Dorner, 1975 [64] 63 Prospective None Mothers SB 13–19 Late Malaise Inventory
Dorner & Atwell, 1985 [65] 25 Prospective None Mothers & fathers Non-surviving SB - - Malaise Inventory
Downey, 1981 [66] Cohorts None - Combined: SB, Down syndrome 0–2 - Standardized questionnaire
Eden-Piercy, Blacher, & Eyman, 1986 [67] 77 Prospective None Mothers Combined: SB, autistism, mentally retarded 1–10 Early Questionnaire on emotions
Hare, Laurence, Payne, & Rawnsley, 1966 [52] 120 Longitudinal None Mothers & fathers Combined: SB, ANC, HYDRO Late Semi-structured interview
Kazak, 1987 [46] 125 Prospective Control Mothers & fathers Combined: SB, PKU, mentally retarded 1–16 Early Langner Symptom Checklist
Kolin, Scherzer, New, & Garfield, 1971 [68] 13 Prospective None Mothers SB (MMC) 7–11 Late Psychiatric observation
Kronenberger, 1991(abstract) [69] 66 Prospective None Mothers SB (MMC) 0–18 Early SCL-90R
Loebig, 1990 [70] 10 Prospective None Mothers SB (MMC) 5–11 Early Semi-structured interview
McAndrew, 1976 [47] 116 Retrospective None Mothers & fathers Combined: MMC, CP, limb deficit 5–10 - Semi-structured interview
Murdoch, 1984 [53] 109 Retrospective None Mothers SB 2–10 Early Semi-structured interview
Nielsen, 1980 [54] 30 Longitudinal None Mothers SB (MMC) 0–6 Early Semi-structured interview
Richards & McIntosh, 1973 [71] 86 Prospective None Mothers & fathers SB (SBA) 2–6 Late Semi-structured interview
Rolle, Niemeyer, & Grafe, 2000 [72] 80 Retrospective None Mothers & fathers Combined: SB, HYDRO 0–18 Early Coping Skills
Spaulding & Morgan, 1986 [59] 19 Prospective Control Mothers & fathers SB-non retarded 5–15 Early Social Readjustment Rating Scale
Walker, Thomas, & Russell, 1971 [55] 108 Retrospective None Mothers & fathers SB 0–3 Early Standardized questionnaire
1ANC = anencephaly, CP = cerebral palsy, HYDRO = hydrocephalus, MMC = myelomeningocele, NOS = not otherwise specified, NTD = neural tube defect, PKU = phenylketonuria, SB = spina bifida, SBA = spina bifida aperta.
2BSI = Brief Symptom Inventory, SCL-90R = Symptom CheckList-90 items Revised.
As shown in Table 1, most studies lacked a comparison group in their design. Only seven studies compared SB-parents with matched control groups, an additional eight studies used standardized outcome measures enabling the comparison of SB-parents with non-clinical norm groups. Most studies included mothers only. Five studies included fathers too, but two of these studies did not specify gender in their analyses [22,23]. Furthermore, some studies included parents of children of all ages whereas others focused on parents of children in a specific developmental period.
Twenty-four reports studied parents of children with SB exclusively. A few studies included late-treated children, that is, children who were born before the time that early surgical treatment had come into practice. Ten studies explicitly confined their samples to the severer forms of SB, namely myelomeningocele (MMC) and spina bifida aperta (SBA). Other studies included a combination of SB with other neural tube defects (NTDs) or with other disabilities. From those non-categorical studies the findings on SB-parents were abstracted for this review. Only one study examined parents' adjustment with the loss of a baby with SB.
Through the years, studies evolved from qualitative to quantitative data collection. Qualitative studies mostly used semi-structured interviews. Quantitative studies used questionnaires to assess symptoms of psychological distress. Three of these measures were adaptations of the Cornell Medical Index, namely: the Malaise Inventory [24], the Symptom Check List-90R (SCL-90R) [25], and the Brief Symptom Inventory (BSI) [26,27]. Other similar questionnaires were the General Health Questionnaire (GHQ) [28] and the Langner Symptom Checklist [29].
Step 4: Allocation of studies eligible for meta-analysis
The reviewers selected studies for meta-analysis guided by the following criteria: (1) quantitatively measuring psychological adjustment in samples that include parents of children with SB, (2) including control group scores or using standardized measures for which norm scores are available, (3) reporting sufficient statistics to estimate effect sizes of SB on parents' psychological adjustment and/or to estimate effect sizes of relationships between other factors and parents' psychological adjustment.
Fifteen research reports were eligible for meta-analysis and 18 were not. The reviewers' agreement was 89.8% (Cohen's Kappa = .88). Differences were resolved through discussion.
Meta-analytic procedures
Weighted average effect size d+
To estimate the effect of SB on parents' psychological adjustment the weighted average effect size d+ was calculated [30]. First, one effect size per sample was obtained through combining multiple reports on the same sample to avoid overrepresentation [31-34]. Second, for studies using standardized outcome measures without matched control groups, Malaise Inventory scores of SB-mothers were compared with norm scores of 33-year-old women (N = 5678, M = 2.81, SD = 3.18; physical health M = .89, SD = 1.17; mental health M = 1.89, SD = 2.37) of the National Child Development Study [35]; scores on the Symptom Check List-90 Revised Global Severity Index were compared with the adult non-patient scores of women (N = 480, M = .36, SD = .35 or T = 50, SD = 10) and of men (N = 494, M = .25, SD = .24 or T = 50, SD = 10) [25]; and T-scores on the Brief Symptom Inventory Global Severity Index were compared with the norms for women (N = 480, T = 50, SD = 10) and men (N = 494, T = 50, SD = 10). Third, the statistical program SISA Binomial [36] was used to estimate a corrected number of degrees of freedom in cases where experimental and comparison groups had different variances. Fourth, effect sizes g were calculated based on means and standard deviations or based on t-test scores [30,37]. Fifth, g's were converted into d's correcting for bias because the reports in this review had relatively small samples. Finally, the weighted average d+ was calculated [30]. For all d+'sStouffer's combined probability effect sizes Zc were reported as indicators of significance.
To check whether d+ encompassed zero, a 95% confidence interval (CI 95%) was estimated. The actual magnitude of d+ was interpreted through use of Cohen's [38] guidelines: d+ ≤ .2 (small effect), d+ ≤ .5 (medium effect), and d+ ≤ .8 (large effect). Furthermore, d+'s were transformed into percentiles of the normal distribution (U3) using Cohen's [38] table to study the amount of non-overlap between experimental and comparison groups. Finally, the homogeneity statistic Q [30] was calculated to determine whether the set of d's on which d+ was based shared a common effect.
Moderating effects of study or sample characteristics on d+ were not tested because of the small number of studies (k = 15).
Weighted average effect size r
To estimate associations between parents' psychological adjustment and various factors the weighted average effect size r [39] was computed. First, t-test and F-test estimates were converted into Pearson's correlations. Second, raw correlation coefficients r were transformed into Fisher's Zr allowing the sampling distribution of r to approximate a Gauss curve. Third, each Zr was weighted by the reciprocal of its estimated within-study variance [30]. Combined probability levels Zc were obtained through dividing the average effect sizes Zr by their standard errors.
Regarding the interpretation of r, most authors recognize that a minimum of three studies is needed for r to be a valid estimate of the population effect size Rho [37]. However, since the objective of this review was to exhaust the limited studies available as much as possible, r's were also calculated on two correlation coefficients. Our justification is that any significant correlation expresses a representative estimate of an association in a certain population. Thus, although two combined correlations do not sufficiently approximate effect size Rho of the universal population, they do at least indicate a valid association in two independent populations.
The r's based on three or more correlation coefficients were interpreted as follows. The magnitude of r was interpreted using Cohen's [38] guidelines: r = .1 (small effect), r = .3 (medium effect), and r = .5 (large effect). Furthermore, the Q statistic was computed to test the homogeneity of studies underlying r.
File drawer analysis Fail Safe N
Reviews based on published studies only, may be at risk for Type I errors. The underlying assumption is that studies revealing nonsignificant results (confirming the null-hypothesis) are less likely to be published than studies reporting significant results. One way to correct for such bias is to calculate the number of studies confirming the null-hypothesis that would be necessary to reverse a conclusion that a significant relationship exists [37]. Because unpublished manuscripts were beyond the scope of this review, both meta-analyses were followed by File Drawer Analysis [40,41]. In this review, the Fail Safe N [42] was calculated.
Results
Weighted average effect size d+
The first question was whether parents of children with SB showed higher levels of psychological distress than comparison groups. The group means, standard deviations, t-tests, raw group differences, and Hedges' standardized effect sizes g and d of SB-parents and comparison groups are displayed in an additional file [see Additional file 1]. Based on these data, d+'s were computed. Table 2 presents the statistics d+, C.I. 95%, Cohen's U3, homogeneity test Q, Stouffer's combined Zc, and Fail Safe .05 N for mothers, fathers, and parents.
Table 2 Weighted average effect sizes of SB on parents' psychological adjustment
k nexp Weighted Mean
Effect Size d+ 95% Confidence Interval U3 Homogeneity Test Q Stouffer's Combined Test Zc Fail Safe .05 N
Mothers 10 500 .73 .38 – .97 76.7% 66.21*** 9.15*** 299.1
Fathers 3 134 .54 .35 – .76 70.5% 0.24 3.93*** 14.1
Parents 15 831 .76 .48 – .86 77.6% 73.54*** 11.25*** 686.7
*** p < .0001
For mothers of children with SB the average amount of psychological distress was .73 standard deviations higher than for controls (see Table 2). This effect size was between medium and large. The C.I. 95% did not include zero, indicating that the chance of not finding a negative effect of SB was less than 5%. Furthermore, there was 76.7% of non-overlap between SB-mothers and comparison groups. The summary index of statistical significance (Zc) further underscored the probability of the effect. Finally, the Fail Safe N revealed that more than 299 studies confirming a null-hypothesis would be needed to overturn the effect.
For fathers of children with SB more moderate though similar findings were obtained based on three studies. Their levels of psychological distress were approximately half a standard deviation higher than the comparison groups, indicating a medium to large effect size. The corresponding non-overlap between the groups was 70.5%. The C.I. 95% and Zc indicated that the effect was consistent and significant. Moreover, 14 nonsignificant studies would be needed to reverse the effect.
For all parents taken together a medium to large negative effect (d+ = .76) of SB psychological adjustment was found. There was 77.6% non-overlap between SB-parents and comparison groups. The C.I. 95% did not include zero and Zc confirmed the overall significance of the effect. What is more, 687 studies confirming a null-hypothesis would be required to undermine the effect size.
Notwithstanding the above results, the significance of Q indicated that the effects of SB on mothers' psychological functioning varied greatly. For fathers a homogeneous underlying effect size was confirmed by a nonsignificant Q but the number of studies (k = 3) was rather limited.
Weighted average effect size r
Possible explanations for the heterogeneity of the SB effect on parents' psychological adjustment were studied by examining associated factors. The variables studied in relationship with parents' psychological functioning had been categorized as: child factors, parent factors, family factors and environmental factors. A summary of the converted effect sizes (Pearson's r, p-value, and Fisher Zr) found in the literature is displayed in an additional file [see Additional file 2]. The weighted average effect sizes r are depicted in Table 3. In the following, only the results for the average effect sizes r based on three or more studies will be briefly described.
Table 3 Weighted average effect sizes of categories associated with parents' psychological symptoms
Category k n Weighted Zr Effect Size r Zc Homogeneity Q Fail Safe .05 N
Child factors
Disability parameters 4 385 .14 .14 2.75** 2.93 10.1
Behavior problems 3 273 .38 .37 6.22*** 2.60 30.9
Emotional problems 2 193 .50 .47 6.90*** 1.03 30.4
Social competence 2 109 -.12 -.12 -1.26 .02 0.0
Parent factors
Socio-economic characteristics 3 264 -.13 -.13 -2.13* 1.45 .36
Appraised stress 2 177 .56 .63 7.32*** 5.90* 30.8
Coping 2 76 .40 .38 3.31*** 8.55** 10.9
Parenting satisfaction/competence 2 109 -.44 -.41 -4.44*** .09 12.1
Family factors
Partner presence 3 211 -.16 -.16 -2.22* .69 1.6
Marital adjustment 2 97 -.43 -.40 -4.12*** .23 10.4
Family income 2 214 -.22 -.22 -3.15** 1.05 6.2
Positive family environment 5 340 -.45 -.42 -8.14*** 1.17 108.6
Environment factors
Quantity social support 4 240 -.29 -.28 -4.35*** 3.16 22.9
Satisfaction social support 4 351 -.29 -.28 -5.37*** 6.68 37.9
Formal support 2 214 -.07 -.07 -1.07 .01 .0
* p < .05, ** p < .01, *** p < .001
Seven child variables were reported in association with parents' psychological adjustment: gender, age, cognitive capacities, disability parameters, behavior problems, emotional problems, and social competence [see Additional file 2]. Disability parameters had a small positive and behavior problems had a medium positive association with parents' psychological symptoms (see Table 3). Both effects were homogeneous.
Five parent variables were studied in relation to parental adjustment: socio-economic characteristics, appraised stress, hope, coping, and parenting satisfaction-competence [see Additional file 2]. Socio-economic characteristics correlated inversely and very minimally to parents' psychological complaints (see Table 3). The significance level that was reached mainly reflected correlations found by Kronenberger and Thompson [20].
Eight family variables were studied in association with parental adjustment: partner presence, marital adjustment, family income, family size, family coping style, impact on family, negative family environment, and positive family environment [see Additional file 2]. The presence of a partner was correlated with fewer psychological symptoms, albeit minimally (see Table 3). Moreover, two nonsignificant studies would be enough to nullify the association. Positive family environment was moderately but consistently related with less psychological complaints. Both r's were homogeneous.
Three environmental factors were reported in connection with parents' adjustment: Quantity of social support, social support satisfaction and formal support [see Additional file 2]. For both the amount of social support and satisfaction with social support medium effects were found on psychological distress (see Table 3). The Fail Safe N indicated that the effects would not be easily overturned. Both r's were homogeneous.
Discussion
Overall results
In this section the meaning of the above findings will be addressed and specific gaps in our understanding of parental psychological adjustment with SB will be identified.
Levels of psychological distress in parents of children with SB
The results confirmed our hypothesis that the presence of SB in families predicts higher levels of psychological strain in parents. The heterogeneity of the effect for mothers however also indicated that SB does not necessarily provoke psychopathology in all parents. Reports on the proportions of SB-parents, scoring within clinical ranges of psychopathology, further illustrate this. Within samples of SB-mothers varying proportions of psychopathology were found: 19.2% [43], 31.9% [44], 41% [45], 50% [46] and 56% [47]. Less variability was found for SB-fathers: 25.6% [43], and 28% [48].
Gender differences in parents' psychological adjustment
It was hypothesized that differences in the effect of SB on adjustment could be expected between mothers and fathers because of role differentiations in care and work. The effect for mothers seemed somewhat higher than for fathers, but the difference could not be tested reliably because of the few studies on fathers. There was some indication that the effect of SB was more homogeneous for fathers than for mothers. Hypothetically, the division of care and work between partners may provide a theoretical explanation for this difference. Work outside the home can be an opportunity to release some of the stress around SB [49]. While at the same time, full-time working schedules may impede contacts with health professionals and therefore diminish opportunities to discuss worries concerning SB [48]. Fathers tend to work fulltime schedules while mothers' occupational lives are more likely to vary [48,49]. In addition, the nursing burden for children with SB varies greatly. Thus SB-related stress on mothers may be much more variable than on fathers. Further enquiries on father's psychological adjustment with SB will be needed to determine whether this hypothesis can be empirically underscored.
Factors correlated with parents' psychological adjustment
Variability in parents' psychological adjustment was expected to be associated with child, parent, family, and environment factors. In terms of models explaining adjustment to chronic illness, parents' psychological adjustment was regarded as the outcome of transactions among multiple factors representing demands and resources. Theoretically, such transactions may involve interactions as well as main effects; however in this meta-analysis direct associations were estimated only.
This review yielded correlation coefficients based on one study only (Additional file 2; representative of one population), average effect sizes based on two studies (Table 3; representative of two populations), and average effect sizes based on three or more studies (Table 3; representative of all populations). Figure 1 displays a summary of these associations.
Figure 1 Factors found to be associated with parents' psychological adjustment.
All the associations were cross-sectional. Hence, inferences about causalities, whether uni- or bi-directional, could not be made on an empirical basis. In the light of this situation, it is feasible that future longitudinal studies will reveal that not all of the associations found in this review will be of decisive importance to explain parents' adjustment with SB. Therefore, we labeled factors associated with reductions in psychological distress as "positive associations" and factors associated with increases in parents' psychological distress as "negative associations".
Child factors
Associations of the child's cognitive capacities with parents' psychological adjustment were hardly reported, despite indications from non-categorical studies that cognitive limitations are likely to put extra strains on parents [17] and despite indications from neuropsychological research that children with SB have specific profiles of cognitive strengths and weaknesses [50]. More research will be needed to understand the impact of children's cognitive profiles on parents' adjustment.
Most studies did not find associations between the degree of the physical disability and parents' psychological adjustment, except one study [34]. Kronenberger and Thompson [21] have noted that this particular study included children with milder forms of SB. Another explanation may be that indexes of the severity of SB have not been conceptualized in a consistent way. Some studies used indicators of physical impairments only (e.g., lesion level of the defect), others added functional limitations (e.g., the degree of mobility), and/or indicators of treatment intensity (e.g., number of shunt revisions). Conceptual refinement of SB-parameters and treatment will be needed to more effectively investigate which factors cause stress in parents and which do not.
Theoretically, the marginality hypothesis [16] may further explain why a linear relationship between SB parameters and parents' psychological functioning was barely found. This theory holds that children with minor disabilities tend to exhibit more psychosocial problems than severely impaired children because they have difficulties identifying themselves with either able-bodied or disabled peers. Similar identification problems could arise for parents of marginally disabled children with SB.
The associations of behavior and emotional problems with parents' psychological symptoms may signify that such problems put additional strain on parents. Once children have developed conduct and/or emotional disorders, this line of reasoning is plausible. However, it is well known from the parenting literature that parent-child relationships are bidirectional, meaning that parents and children mutually influence each other through long-term transactions. For example, attachment theorists have emphasized that during the early years of the child's life, parents' sudden mood changes, depressive symptoms, and grief are potential risks to the development of affective attunement between parent and child [51]. Parents descriptions in open interviews of their struggle with emotions during the first year after the birth of their child with SB seem to support the hypothesis that these children might be at risk of insecure attachment [47,52-55]. In the long term, the insecure parent-child relationship may contribute to the development of behavioral and emotional problems. It may be well worth studying the early development of parent-child relations in families with SB to uncover how children's behavioral and emotional problems interplay with parents' psychological adjustment over time.
Children's lack of social competence was not found to be related to parents' psychological health even though Lemanek et al. [56] reported that children with SB had significantly fewer social skills than children in norm groups. These findings provide indirect support for the hypothesis that parents do not expect equal proficiency in social skills, such as cooperation, assertion, responsibility, and self-control, from a child with SB as from an able-bodied child. This may explain why a limitation in social skills of children with SB does not affect parents' psychological adjustment. More studies investigating parents expectations are required to affirm this assumption.
Parent factors
Very few studies investigated the role of parents' appraisals and coping styles. This is remarkable, since the role of appraisal and coping are of central importance to understanding how stressful events affect people [57]. The scarce findings suggest that parents' appraisals (e.g., appraised stress and hope) and coping styles are highly predictive of positive as well as negative adjustment. Besides appraisal and coping, hardly any intra-personal resources of parents were studied in relation with psychological adjustment to SB. In the light of current theories on affect-processing, the absence of studying personality characteristics, such as ego-resilience, could be regarded a major gap in our knowledge of parents' adjustment to SB. For example, J. Block has pointed out that some individuals are characteristically maladaptive while others are characteristically resourceful in responding to environmental stressors [58]. This characteristic ability to dynamically and progressively adapt to stress appears to be more person-related than situation-related. Thus, studies on the associations between personality characteristics, affect-regulation, and psychological adjustment may prove to be fruitful.
Family factors
As expected, parents' psychological health was consistently associated with a supportive family climate. The quality of parents' partner relationship also appeared to be a promising correlate of psychological bonadjustment. Future research may need to study more closely though, whether the measure of marital satisfaction reflects satisfaction with the joint care for the child with SB or satisfaction with a relationship that meets parents' personal needs of intimacy and companionship.
Environmental factors
In line with expectations, there appears to be fair evidence that a large informal social network of family and friends that matches parents' needs, enhances parents' psychological adjustment to SB. Unexpectedly, formal types of support were not related to parents' psychological adjustment. Apparently, dissatisfaction with formal support does not necessarily imply increased risks of psychological maladjustment.
Strengths and limitations of studies and future research
The chronology of studies was in line with contemporary trends in behavioral sciences. Qualitative descriptive research was followed by quantitative analytical designs. Standardized measures of psychological symptoms came into use and were updated, passing from the Malaise Inventory to the Brief Symptom Inventory. Statistical procedures moved from correlational analyses to multiple regression equations and structural equation models.
Inevitably, studies also had methodological limitations. In the first place, studies had sampling problems. Samples tended to be small, risking Type II errors (i.e. not detecting a relationship which in fact exists). For example, one study (n = 19) did not find a significant relationship between SB and parents' psychological adjustment [59]. Furthermore, the recruitment of participants via hospitals and/or SB associations may have led to unbalanced sampling. Members of SB associations may not be representative of all SB-parents. Moreover, parents with psychiatric problems may have refused to participate in studies. And finally, fathers were underrepresented.
A second area of concern is the quality of the associations reported in studies. Most associations were cross-sectional. Hence, the causal interpretations were based on theoretical assumptions only. Furthermore, correlations may have been inflated because studies relied on parents' self-reports. Especially studies examining depression and anxiety are at risk of common method variance, because the respondents' affective states may influence their ratings of other concepts [60].
Future studies will need to increase their sample sizes through merging datasets from different studies and establishing long-term multi-centered collaborations. Special efforts, such as home visits after office hours, must be made to include more fathers. Longitudinal designs are needed to empirically validate assumed directions of associations. And finally, studies need to collect data from multiple informants and/or observational data to avoid common method variance.
Conclusion
Our study confirms that SB represents a considerable challenge to parents' psychological well-being. Especially mothers are at risk of psychological suffering, although there is great variety among mothers in their psychological adjustment to having a child with SB. Studies indicate that the extent to which SB affects parents depends on the quality of parents' partner relationship, family climate, and support from informal social networks.
Clinical implications
Bearing these results in mind, it is important to monitor parents' psychological well-being on a regular basis, that is, to ask parents at different stages of their child's life how they cope, how they keep the care strains manageable, how they support one another, and how they reserve time to balance the care for their child with SB and other primary tasks with their personal needs. Alertness to the quality and amount of social support around the family may prevent parents from becoming overburdened.
It may be important to advise parents to think strategically about how their relationships with others can support them emotionally as well as instrumentally at times when the care for their child intensifies due to acute medical situations or at times when chronic burdens pile up. At the same time, it may be equally important to advise parents to think about how much attention these relationships need in order to be maintained.
In conclusion, the medium-large effect of SB on parents' psychological health indicates that spina bifida health care should include psychological support to parents of children with this condition to ensure the well-being of the whole family.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
IV conceived of the study, carried out the literature search, the meta-analyses, and drafted the manuscript. JJ helped with the literature search (coding and categorizing abstracts), supervised the meta-analyses procedures, and helped to draft the manuscript. AB participated in the interpretation of the findings and critically revised the content of the manuscript. JG conceived of the study, supervised its design and coordination, and critically revised the content of the manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Supplementary Material
Additional File 1
Effect sizes of SB on parents' psychological adjustment. This file contains a table with the statistics that we gathered from primary reports. Based on these statistics the weighted average effect sizes of spina bifida on parents' psychological adjustment were estimated.
Click here for file
Additional File 2
Reported correlations of factors associated with parents' psychological adjustment. This file contains a summary of the converted effect sizes (Pearson's r, p-value, and Fisher Zr) found in the literature. Based on these data the weighted average effect sizes of categories associated with parents' psychological symptoms were estimated.
Click here for file
Acknowledgements
Completion of this manuscript was funded by a Research Grant from the University Board of the Radboud University to the Nijmegen Interdisciplinary Spina Bifida (NISB) research program. This research program is directed by Jan Gerris, PhD (Institute of Family and Child Care Studies, Radboud University Nijmegen) and Jan Rotteveel, MD PhD (Interdisciplinary Paediatric Neurological Centre, University Medical Centre St Radboud Nijmegen).
We would like to thank the reviewers for their helpful comments on earlier drafts of this manuscript and Judith Semon Dubas for revising the English.
==== Refs
Detrait ER George TM Etchevers HC Gilbert JR Vekemans M Speer MC Human neural tube defects: Developmental biology, epidemiology, and genetics Neurotoxicology and Teratology 2005 27 515 524 15939212 10.1016/j.ntt.2004.12.007
EUROCAT World atlas of birth defects 1998 Geneva, CH, World Health Organisation 20 31
Kinasha A Manji K Neural Tube Defects in Dar es Salaam
Mersereau P Kilker K Carter H Fasett E Williams J Flores A Prue C Williams L Mai C Mulinare J Spina bifida and Anencephaly before and after Folic Acid Mandate - United States, 1995-1996 and 1999-2000 Morbidity and Mortality Weekly Report 2004 53 362 365 15129193
Mathews TJ Honein MA Erikson JD Spina bifida and Anencephaly prevalence - United States, 1991 - 2001 Morbidity and Mortality Weekly Report 2002 51 9 11
EUROCAT Prevalence of neural tube defects in 20 regions of Europe and the impact of prenatal diagnosis, 1980-1986 Journal of Epidemiology and Community Health 1991 45 52 58 2045746
Tulipan N Sutton LN Bruner JP Cohen BM Johnson M Adzick NS The effect of intrauterine myelomeningocele repair on the incidence of shunt-dependent hydrocephalus Pediatric Neurosurgery 2004 38 27 33 12476024 10.1159/000067560
Thompson RJJ Kronenberger WG A heuristics review of the psychosocial aspects of myelodysplasia Advances in Developmental and Behavioral Pediatrics 1992 10 89 108
de Tychey C Spina-bifida et adaptation "psycho socio-affective." [Spina bifida and "psycho-socioaffective" adjustment] Annales Medico Psycholiques (Paris) 1983 141 731 756
Kalnins IV Cross-illness comparison of separation and divorce among parents having a child with a life-threatening illness Children's Health Care 1984 12 72 77
Singh DK Families of children with spina bifida: A review Journal of Developmental and Physical Disabilities 2003 15 37 55 10.1023/A:1021452220291
Tizard J Children with myelomeningocele; Social and educational problems Developmental Medicine and Child Neurology 1968 15 1 5
Nevin RS Easton JK McCubbin HI Birkebak RR Parental coping in raising children who have spina bifida cystica Z Kinderchir Grenzgeb 1979 28 417 425 399416
Wallander JL Varni JW Babani L DeHaan CB Wilcox KT Banis HT The social environment and the adaptation of mothers of physically handicapped children Journal of Pediatric Psychology 1989 14 371 387 2677297
Moos RH Tsu VD Moos RH The crisis of physical illness: An overview Coping with physical illness Current topics in mental health Ahmed P I, Plog S C 1977 London, U.K., Plenum Medical Book Company 3 21
Pless IB Pinkerton P Chronic childhood disorder: Promoting patterns of adjustment 1975 Chicago, IL, Year Book Medical Publishers
Sloper P Turner S Risk and resistance factors in the adaptation of parents of children with severe physical disability Journal of Child Psychology and Psychiatry 1993 34 167 188 8444991
Wallander JL Varni JW Effects of pediatric chronic physical disorders on child and family adjustment Journal of Child Psychology and Psychiatry 1998 39 29 46 9534085 10.1017/S0021963097001741
Thompson RJGKE Adaptation to chronic childhood illness 1996 Washington, DC, American Psychological Association
Kronenberger WG Thompson RJJ Psychological adaptation of mothers of children with spina bifida: Association with dimensions of social relationships Journal of Pediatric Psychology 1992 17 1 14 1545317
Kronenberger WG Thompson RJJ Medical stress, appraised stress, and the psychological adjustment of mothers of children with myelomeningocele Developmental & Behavioral Pediatrics 1992 13 405 411
King G King S Rosenbaum P Goffin R Family-centered caregiving and well-being of parents of children with disabilities: Linking process with outcome Journal of Pediatric Psychology 1999 24 41 53 10.1093/jpepsy/24.1.41
Horton TV Wallander JL Hope and social support as resilience factors against psychological distress of mothers who care for children with chronic physical conditions Rehabilitation Psychology 2001 46 382 399 10.1037//0090-5550.46.4.382
Rutter M Tizard J Whitmore K Education, health, and behavior 1970 London, U.K., Longmans
Derogatis LR SCL-90-R: Administration, scoring, and procedures manual (3rd Edition) 1994 3rd Edition Mineapolis, MN, NCSD Pearson Inc.
Derogatis LR Brief Symptom Inventory 1993 Mineapolis, MN, NCS Pearson Inc.
Derogatis LR Melisaratos N The Brief Symptom Inventory: An introductory report. Psychological Medicine 1983 13 595 605 6622612
Goldberg D Manual of General Health Questionnaire 1978 Slough, U.K., NFER-Nelson
Langner TA A 22 item screening score of psychiatric symptoms indicating impairment Journal of Health and Human Behavior 1962 3 269 276 13928667
Hedges L Olkin I Statistical Methods for Meta-Analysis 1985 New York, NY, Academic Press
Barakat LP Linney JA Children with physical handicaps and their mothers: The interrelation of social support, maternal adjustment, and child adjustment Journal of Pediatric Psychology 1992 17 725 739 1484335
Barakat LP Linney JA Optimism, appraisals, and coping in the adjustment of mothers and their children with spina bifida Journal of Child and Family Studies 1995 4 303 320
Tew B Laurence KM Mothers, brothers and sisters of patients with spina bifida Developmental Medicine and Child Neurology 1973 15 69 76 4580876
Tew B Laurence KM Some sources of stress found in mothers of spina bifida children British Journal of Preventive and Social Medicine 1975 29 27 30 124188
Rodgers B Pickles A Power C Collishaw S Maughan B Validity of the Malaise Inventory in general population samples Social Psychiatry and Psychiatric Epidemiology 1999 34 233 241 10.1007/s001270050153
Uitenbroek DG SISA Binomial [Computer software]
Wolf FM Meta-analysis: Quantitative methods for research synthesis Series: Quantitative applications in social sciences L. Sullivan J., G. Niemi R. 1986 Beverly Hills, CA, Sage publications
Cohen J Statistical power analysis for the behavioral sciences (Revised edition) 1977 New York, Academic Press
Hedges L Vevea JL Fixed- and random-effects models in meta-analysis. Psychological Methods 1998 3 486 504 10.1037//1082-989X.3.4.486
Rosenthal R Comparing significance levels of independent studies Psychological Bulletin 1979 86 1165 1168 10.1037//0033-2909.86.5.1165
Rosenthal R Bickman L and Rog D Meta-analytic procedures for social research Applied social research methods series (Vol 6) 1984 Beverly Hills, CA, Sage publications
Cooper HM Statistically combining independent studies: A meta-analysis of sex differences in conformity research Journal of Personality and Social Psychology 1979 37 131 146 10.1037//0022-3514.37.1.131
Holmbeck GN Gorey Ferguson L Hudson T Seefeldt T Shapera W Turner T Uhler J Maternal, paternal, and marital functioning in families of preadolescents with spina bifida Journal of Pediatric Psychology 1997 22 167 181 9114641
Dorner S Some problems of families with an adolescent spina bifida patient Developmental Medicine and Child Neurology 1974 16 153 154
Wiegner S Donders J Predictors of parental distress after congenital disabilities Journal of Developmental and Behavioral Pediatrics 2000 21 271 277 10972250
Kazak AE Families with disabled children: stress and social networks in three samples Journal of Abnormal Child Psychology 1987 15 137 146 3553273 10.1007/BF00916471
McAndrew I Children with a handicap and their families Child: Care, Health and Development 1976 2 213 218 134850
Evans O Tew B Laurence KM The fathers of children with spina bifida Zeitschrift für Kinderchirurgie [Surgery in Infancy and Childhood] 1986 41 Suppl 1 42 44
Joosten J Accounting for changes in family life of families with spina bifida children Zeitschrift für Kinderchirurgie [Surgery in Infancy and Childhood] 1979 28 412 417
Fletcher JM Northrup H Landry SH Kramer LA Brandt ME Dennis M Barnes MA Blaser SE Hannay HJ Copeland K Francis DJ Spina bifida: Genes, brain, and development International Review of Research in Mental Retardation 2004 29 63 117 10.1016/S0074-7750(04)29003-6
Field T Bornstein MH Psychologically depressed parents Handbook of parenting: Applied and practical parenting (Vol 4) 1995 London, U.K., Lawrence Erlbaum Associates, Publishers
Hare EH Laurence KM Payne H Rawnsley K Spina bifida cystica and family stress British Medical Journal 1966 2 757 760 5331185
Murdoch JC Immediate post-natal management of the mothers of Down's syndrome and spina bifida children in Scotland 1971-1981 Journal of Mental Deficiency Research 1984 28 67 72 6232389
Nielsen HH A longitudinal study of the psychological aspects of myelomeningocele Scandinavian Journal of Psychology 1980 21 45 54 7361075
Walker JH Thomas M Russell IT Spina bifida--and the parents Developmental Medicine and Child Neurology 1971 13 462 476 4934305
Lemanek KL Jones ML Lieberman B Mothers of children with spina bifida: Adaptational and stress processing Children's Health Care 2000 29 19 35 10.1207/S15326888CHC2901_2
Skinner EA Edge K Altman J Sherwood H Searching for the structure of coping: A review and critique of category systems for classifying ways of coping Psychological Bulletin 2003 129 216 269 12696840 10.1037/0033-2909.129.2.216
Block J Personality as an affect-processing system: Toward an integrative theory 2002 London, U.K., Lawrence Erlbaum Associates, Publishers
Spaulding BR Morgan SB Spina bifida children and their parents: A population prone to family dysfunction? Journal of Pediatric Psychology 1986 11 359 374 3534207
Atkinson M Characterizing quality of life among patients with chronic mental illness: A critical examination of the self-report methodology: Reply American Journal of Psychiatry 1997 154 1479 9326848
Fagan J Schor D Mothers of children with spina bifida: factors related to maternal psychosocial functioning American Journal of Orthopsychiatry 1993 63 146 152 8427306
Kazak AE Marvin RS Differences, difficulties and adaptation: Stress and social networks in families with a handicapped child Family Relations: Journal of Applied Family and Child Studies 1984 33 67 77
Dorner S Psychological and social problems of families of adolescent spina bifida patients: a preliminary report Developmental Medicine and Child Neurology 1973 15 24 26
Dorner S The relationship of physical handicap to stress in families with an adolescent with spina bifida Developmental Medicine and Child Neurology 1975 17 765 776 1107105
Dorner S Atwell JD Family adjustment to the early loss of a baby born with spina bifida Developmental Medicine and Child Neurology 1985 27 461 466 3896902
Downey JB The relationship between age of child and parental adjustment to congenitally defective infants Dissertation Abstracts International Section B: Sciences and Engineering 1981 42 2029 2030
Eden-Piercy GVS Blacher JB Eyman RK Exploring parents' reactions to their young child with severe handicaps Mental Retardation 1986 24 285 291 2945980
Kolin IS Scherzer AL New B Garfield M Studies of the school-age child with meningomyelocele: Social and emotional adaptation Journal of Pediatrics 1971 78 1013 1019 4930795
Kronenberger WG The effects of perceived social, marital, and familial environments on the psychological adjustment of mothers of children with spina bifida Dissertation Abstracts International Section B: Sciences and Engineering 1991 52 3299
Loebig M Mothers' assessments of the impact of children with spina bifida on the family Maternal Child Nursing Journal 1990 19 251 264 2136578
Richards ID McIntosh HT Spina bifida survivors and their parents: a study of problems and services Developmental Medicine and Child Neurology 1973 15 292 304 4578076
Rolle U Niemeyer C Grafe G Coping strategies of parents from patients with spina bifida and hydrocephalus of various aetiologies European Journal of Pediatric Surgery 2000 10 62 63 10770251
|
16120229
|
PMC1215488
|
CC BY
|
2021-01-04 16:31:07
|
no
|
BMC Pediatr. 2005 Aug 25; 5:32
|
utf-8
|
BMC Pediatr
| 2,005 |
10.1186/1471-2431-5-32
|
oa_comm
|
==== Front
BMC PediatrBMC Pediatrics1471-2431BioMed Central London 1471-2431-5-351614304810.1186/1471-2431-5-35Research ArticleEqual antipyretic effectiveness of oral and rectal acetaminophen: a randomized controlled trial [ISRCTN11886401] Nabulsi Mona [email protected] Hala [email protected] Ramzi [email protected] Ziyad [email protected] Shadi [email protected] Hadi [email protected] Mohammad [email protected] Department of Pediatrics, American University of Beirut Medical Center, Beirut, Lebanon2 Department of Epidemiology and Population Health, Faculty of Health Sciences, American University of Beirut, Beirut, Lebanon3 Department of Pharmacology and Therapeutics, Faculty of Medicine, American University of Beirut, Beirut, Lebanon4 Department of Pediatrics, Middle East Hospital, Beirut, Lebanon2005 6 9 2005 5 35 35 24 2 2005 6 9 2005 Copyright © 2005 Nabulsi et al; licensee BioMed Central Ltd.2005Nabulsi et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
The antipyretic effectiveness of rectal versus oral acetaminophen is not well established. This study is designed to compare the antipyretic effectiveness of two rectal acetaminophen doses (15 mg/kg) and (35 mg/kg), to the standard oral dose of 15 mg/kg.
Methods
This is a randomized, double-dummy, double-blind study of 51 febrile children, receiving one of three regimens of a single acetaminophen dose: 15 mg/kg orally, 15 mg/kg rectally, or 35 mg/kg rectally. Rectal temperature was monitored at baseline and hourly for a total of six hours. The primary outcome of the study, time to maximum antipyresis, and the secondary outcome of time to temperature reduction by at least 1°C were analyzed by one-way ANOVA. Two-way ANOVA with repeated measures over time was used to compare the secondary outcome: change in temperature from baseline at times1, 2, 3, 4, 5, and 6 hours among the three groups. Intent-to-treat analysis was planned.
Results
No significant differences were found among the three groups in the time to maximum antipyresis (overall mean = 3.6 hours; 95% CI: 3.2–4.0), time to fever reduction by 1°C or the mean hourly temperature from baseline to 6 hours following dose administration. Hypothermia (temperature < 36.5°C) occurred in 11(21.6%) subjects, with the highest proportion being in the rectal high-dose group.
Conclusion
Standard (15 mg/kg) oral, (15 mg/kg) rectal, and high-dose (35 mg/kg) rectal acetaminophen have similar antipyretic effectiveness.
==== Body
Background
Parents of febrile children often conceive fever as a disease that requires treatment, rather than being a symptom or a sign of illness. In their anxious quest to treat fever, parents suffering from "fever phobia" may end up unintentionally overdosing their children with different antipyretics, or with different preparations of the same antipyretic [1-3]. Acetaminophen, in its various preparations, is a widely used drug because of its established analgesic and antipyretic effects. Whereas the analgesic efficacy achieved with standard (10–20 mg/kg) oral, standard rectal (10–20 mg/kg), and high rectal (40–45 mg/kg) acetaminophen doses have been well investigated [4-7], the comparative antipyretic effects of oral and rectal acetaminophen is not well studied. Parents, as well as physicians, use the standard dose (10–20 mg/kg) of oral and rectal acetaminophen preparations interchangeably to treat fever in children, assuming they have equal antipyretic effects. However, although the evidence for rapid absorption (within 30–60 minutes) and the pharmacokinetics of a single acetaminophen oral dose is well-established [8,9], the pharmacokinetics of a single rectal dose reveal the absorption to be erratic and prolonged, varying with the suppository size, composition of its base, rate of dissolution, position in the rectum, and the rectal contents [5]. Moreover, an increasing body of evidence indicates that the rectal acetaminophen dose of 10–15 mg/kg fails to achieve antipyretic serum levels of 10–20 μg/ml. Indeed, a rectal acetaminophen dose of 30–45 mg/kg is needed to achieve antipyretic serum levels in that range [5,7,10-12].
To our knowledge, only three randomized controlled trials had previously investigated the antipyretic effects of rectal acetaminophen in comparison to the oral one, with contradictory results [13-15]. Whereas Leary et al. found oral paracetamol to be superior to the rectal preparation in reducing the temperature of febrile children [13], both Vernon et al [15] and Scolnik et al [14] found no difference in the antipyretic responses of oral and rectal acetaminophen. However, Vernon's study was unblinded and lacked placebo control, and compared the standard doses of 15–20 mg/kg of oral and rectal acetaminophen only. Scolnik's study, also lacking blinding and placebo control, was further limited by the fact that it assessed antipyresis during the first three hours after drug administration, a time during which maximum antipyresis of rectal acetaminophen may not have occurred.
In view of the conflicting results in the literature, we conducted this study to compare the antipyretic effectiveness of two different rectal doses of acetaminophen: 15 mg/kg and 35 mg/kg to that of a standard oral dose of 15 mg/kg, over a six-hour period to allow detection of late antipyresis that may occur with rectal acetaminophen. The results of this study will provide further evidence on the comparative antipyretic effects of different doses of rectal acetaminophen versus the standard oral one.
Methods
Setting
This study was conducted between November 2000 and September 2002, in the paediatric inpatient services of two hospitals in Beirut: the American University of Beirut Medical Center (AUBMC), which is a tertiary care facility, and the Middle East Hospital (MEH), a secondary care facility. The Institutional Review Board and the Ethics Committee at the American University of Beirut, as well as the Board of the Middle East Hospital, approved this study. Written informed consent was obtained from all parents as well as the oral consent of children aged 10 years or more.
Subjects
Subjects approached for enrolment in the study were febrile inpatients whose ages were between 6 months and 13 years, and whose rectal temperature was ≥ 38.5°C. A wide age range was permitted to enhance recruitment, since the antipyretic effect of acetaminophen does not vary with age [9]. Exclusion criteria included any of the following conditions: acute or chronic gastroenteritis, vomiting, any medical or surgical condition that precluded oral or rectal drug administration, acute or chronic hepatic disease, rectal bleeding, malabsorption syndromes, acute or chronic renal disease with the exception of urinary tract infection, chronic metabolic disease, bleeding disorders, chronic neurological disease that may affect central thermoregulation, cancer, immune suppression, sepsis, critical medical status, or known allergy to acetaminophen. Children with concurrent or previous intake of antibiotics were not excluded. All antipyretics were stopped for 8 hours prior to the initiation of the study.
Study design
This is a randomized, double blind, and double dummy design clinical trial. Subjects were randomized according to a computer-generated, random-number list that was kept with the hospital pharmacist until the end of the study, into one of three treatment groups: standard oral acetaminophen dose (15 mg/kg) and rectal placebo suppositories; standard rectal acetaminophen dose (15 mg/kg) and oral placebo; high-dose rectal acetaminophen (35 mg/kg) and oral placebo. The allocation sequence was generated by one of the co-investigators (HT) who was not involved in subject enrolment or outcome assessment. The pharmacist who prepared the study medications was aware of subjects' treatment allocation, whereas subjects, parents, nurses, treating physicians, research assistant responsible for subject enrolment, data analyst (co-investigator ZM) and investigators were all blinded to the assignment of the patients.
Study medications
The drugs used in this study: acetaminophen and its placebo were supplied by Julphar (Gulf Pharmaceutical Industries, United Arab Emirates). The oral acetaminophen used was a 250 mg acetaminophen/5 ml suspension (Adol, Julphar), while the placebo was a suspension with a similar colour and exipient to Adol. The suppositories (Adol) came in three sizes: 125 mg, 250 mg, and 500 mg. The suppository base is lipophyllic and consists of semi-synthetic glycerides (1140 mg of saturated fatty acids from C8 to C18). Placebo suppositories consisted of the same base, and came in similar colour, shape, and sizes.
Study procedure
After obtaining the approval of the treating physician, the parent(s) of the eligible child was approached for interview and enrolment. During the interview, a trained research assistant administered a structured questionnaire designed to collect information on the following variables: diagnosis, previous or concurrent antibiotics, antipyretic intake, fever duration, gender, and date of birth. The purpose and procedure of the trial were fully explained to the family, and written parental consent, as well as oral consent of the subject when older than ten years of age were obtained. Children enrolled in the study were then assigned a random number as mentioned previously.
Baseline rectal temperature was recorded using a portable thermistor with single-use disposable probe covers (Sure Temp 679, Welch Allyn). One thermometer was used for the whole duration of the study in each hospital. Investigational drugs were prepared by the pharmacist who was aware of subjects' treatment allocation as follows: the oral group would receive oral acetaminophen at a dose of 15 mg/kg through a syringe, followed by placebo rectal suppositories which were estimated assuming a rectal acetaminophen dose of 35 mg/kg; the second group would receive oral placebo at a volume similar to the volume obtained if oral acetaminophen were to be given at a dose of 15 mg/kg, and rectal suppositories consisting of 15 mg/kg acetaminophen and 20 mg/kg placebo; the third group would receive oral placebo, and rectal acetaminophen suppositories at 35 mg/kg dose. In order to avoid cutting suppositories, rectal acetaminophen dose was rounded to the suppository size nearest to the calculated dose. More than one suppository could be used to achieve the desired rectal dose. The patient's nurse, who was blinded to the treatment allocation, administered all drugs and checked suppository retention for 30 minutes following administration and at hourly intervals for the duration of the study. Rectal temperatures were subsequently recorded at 1, 2, 3, 4, 5, and 6 hours from baseline.
Statistical analyses
The primary outcome was time to maximum antipyresis following administration of a single dose of acetaminophen. Secondary outcomes: included time to fever reduction by at least 1°C, the temperatures at one, two, three, four, five, and six hours from administration and possible side effects such as hypothermia defined as a rectal temperature < 36.5°C.
For sample size calculation, we considered a one-hour difference in the average time to reach maximum antipyresis between any of two treatment groups to be a clinically significant outcome. Using this one-hour difference to maximum antipyresis, a standard deviation of one hour, 80% power, and alpha of 0.05, the calculated sample size was 48 subjects, 16 in each group.
We used the Chi square test to study the association between categorical variables and treatment groups, and one-way ANOVA to investigate the relationship between continuous variables and treatment groups. Two-way ANOVA with repeated measures over time was used to compare the changes in temperature from baseline at each time (t = 1, 2, ..., 6), among the three groups. Intent-to-treat analysis was planned with statistical significance set at P < 0.05.
Results
Baseline characteristics
Between November 2000 and September 2002, 125 parents were approached for interview and questionnaire administration. The progress of these subjects through the study is shown in Figure 1. There were no differences in the baseline characteristics of the patients who completed the study and those who did not. The study was completed with 51 subjects: 16 in the oral group, 18 in the rectal standard-dose group, and 17 in the rectal high-dose group. Their mean (SD) age was 3.9(3.0) years, with an age range of 6 months-13.1 years. The median duration of fever was 3.0 days, with a range of 0.5–101.0 days. Three patients had prolonged fever ranging from 3 weeks to 3 months that were later diagnosed to be due to juvenile rheumatoid arthritis, central fever, and viral etiology. These patients were analyzed in their groups since intent to treat analysis was planned. There were 29 (56.9%) males, and 35 (68.6%) subjects were receiving at least one antibiotic when entered into the study.
Figure 1 Flow diagram of the subjects' progress through the study.
The characteristics of the three groups are shown in Table 1. There were no significant differences with respect to sex, age, underlying basic disease causing fever, duration of fever, previous antipyretic use, or concurrent antibiotic administration. In addition, there were no differences in the characteristics or treatment allocation of the patients recruited from AUBMC (39), and those recruited from MEH (12). The baseline temperature was similar in the three groups with a mean (SD) of 39.2 (0.7)°C. As for the rectal acetaminophen doses, the means (SD) and ranges were: 14.1 (2.3) mg/kg and 10.7–18.5 mg/kg for the standard rectal dose; and 31.7 (6.7) mg/kg, and 12.5–43.4 mg/kg for the rectal high-dose respectively. One subject who was allocated to the rectal high-dose group received a low rectal dose by mistake. Excluding the dose received by this patient, the mean (SD) and range for the high-dose rectal group becomes 33.0(4.7) mg/kg and 27.2–43.4 mg/kg respectively. Since intent-to-treat analysis is planned, this subject was analyzed in the high-dose group.
Table 1 Subject characteristics and main outcome.
TOTAL RECTAL (15 MG/KG) ORAL (15 MG/KG) RECTAL (35 MG/KG)
Total 51 18 16 17
Male Gender
N (%) 29 (56.9) 11 8 10
Age (years):
Mean (SD) 4.1 (3.5) 3.8 (2.8) 4.0 (3.6) 3.5 (3.0)
Range 0.5–13.1 0.6–13.1 0.5–10.2 0.5–12.4
Diagnosis:¶ N (%)
Pneumonia 10 (19.6) 3 2 5
UTI 7 (13.7) 2 4 1
Virus 17 (33.3) 6 4 7
Bacteremia 5 (9.8) 2 2 1
Others 36 (70.5) 15 12 9
Previous antipyretic
N (%)
Acetaminophen 40(78.4) 15 14 11
Ibuprofen 8(15.6) 2 2 4
Antibiotic intake
N (%) 35 (68.6) 16 8 11
Acetaminophen Dose (mg/kg):
Mean (SD) § 14.1(2.3) 15.0(0.0) 31.7(6.7)
Range 10.7–18.5 15.0–15.0 12.5–43.4
Duration of fever (days):
Median 3.0 3 1.3 5
Range 0.5–101.0 1.0–90.0 0.5–15.0 1.0–101.0
Baseline temperature
Mean (SD) 39.2 (0.7) 39.1 (0.9) 39.3 (0.6) 39.1 (0.6)
Time to max AP in hr:
Mean 3.6 3.3 3.6 3.9
95% CI 3.2–4.0 2.4–4.2 2.8–4.3 3.3–4.6
§ P = 0.000 (ANOVA); ¶ Percentages do not add up to 100 since more than one diagnosis is entered for some subjects; max AP: maximum antipyresis; hr: hour.
Primary outcome
Intent-to-treat analysis of the time to maximum antipyresis, the primary outcome measure of the study, revealed no significant differences among the three groups (P = 0.5). The overall mean (95% CI) was 3.3 (2.4–4.2) hours for the rectal low-dose group, 3.6 (2.8–4.3) hours for the oral group, and 3.9 (3.3–4.6) hours for the rectal high-dose group [Table 1]. Repeat analysis excluding the three patients with long duration of fever, and the only patient who received a low-dose rectal acetaminophen instead of his allocated high-rectal dose, revealed similar results to the intent-to-treat analysis. Therefore, analyses of both primary and secondary outcomes were kept as intent-to-treat.
Secondary outcomes
The mean (95%CI) maximum decline in temperature was 1.6(1.3–2.0)°C in the rectal low-dose group, 1.7(1.2–2.2)°C in the oral group, and 2.0(1.4–2.5)°C in the rectal high-dose group (P = 0.5). The time to fever reduction by at least 1°C was similar among the three groups: mean (95% CI) of 2.4 (1.8–3.1) hours in the rectal low-dose, 3.5 (2.6–4.4) hours in the oral and 2.8 (2.1–3.6) hours in the rectal high-dose groups (P = 0.13). Two-way ANOVA with repeated measures over time did not reveal statistically significant differences in the changes in temperature from baseline at times 1, 2, ..., 6 hours among the three groups (P = 0.25).
As for side effects of the medications, hypothermia, defined as a body temperature below 36.5°C rectally, occurred in 11 (21.6%) subjects: 2(11.1%) with the rectal standard dose, 3(18.8%) in the oral group and 6(35.3%) with the rectal high-dose. These proportions however were not statistically significant (P = 0.2). The temperature range of hypothermic episodes was between 35.5°C and 36.4°C (mean 36.1°C).
There were three mortalities among our subjects, which were judged to be unrelated to the investigational drugs. The first one was a one-year-old male infant who died 10 days after enrolment from systemic Epstein-Barr viral infection. The second mortality occurred in a 12-year-old child who succumbed to bacterial endocarditis and myocardial abscesses two weeks following enrolment in the study, and the third patient was a six month old boy whose clinical status deteriorated four hours after enrolment, the time at which he was withdrawn from the study. This patient died 14 hours later from complicated respiratory infection, sepsis, and respiratory failure.
Discussion
Acetaminophen is the most widely used antipyretic in paediatric medicine. Despite the well-established antipyretic effects of oral and rectal acetaminophen, controversy regarding the comparative antipyretic effectiveness of the two types of acetaminophen preparations is yet unresolved. Whereas some investigators have reported better antipyresis with oral acetaminophen [13], others have reported equal antipyretic effects [14,15]. Faced with this uncertainty, the use of either preparation is often influenced by the child's acceptance of the oral medication, his medical condition (presence of vomiting for example), and parental or physician preferences.
Our results reveal no difference in the antipyretic effectiveness among oral, rectal standard-dose, or rectal high-dose acetaminophen. The time to maximum antipyresis was not significantly different among the three doses or preparations of the drug, with an overall mean (SD) of 3.3(95%CI: 2.4–4.2) hours. In addition, the three regimens behaved similarly with respect to the maximum decline in temperature at any time during the six hours and the time to fever reduction by at least one degree. Our findings are in agreement with those of Vernon, et al [15] and Scolnik, et al [14], but different from those of Leary et al [13]. These differences may be attributed to the fact that in Leary's study, all outcome measurements were based on axillary temperatures, the reliability of which is uncertain [16].
This study is the fourth randomized controlled trial that compares the antipyretic effectiveness of oral and rectal acetaminophen, and the second one to investigate the differences in antipyresis between standard oral, standard rectal, and high-dose rectal acetaminophen. The strengths of this study, as compared to the previous ones, include the fact that the six-hour study duration permitted detection of any delayed antipyretic responses, if present. In addition, it was double-blinded with double-dummy technique. In contrast, the study of Scolnik et al [14] assessed antipyresis for the first three hours only, the time at which maximum antipyresis may not have occurred. In addition, it was neither blinded, nor placebo-controlled. Similarly, Vernon et al's study [15] lacked both blinding and placebo control, and compared the standard doses of 15–20 mg/kg of oral and rectal acetaminophen only. Finally, the main drawback of Leary et al's study [13] was their use of axillary temperature instead of the gold standard rectal measurements, which undermines the reliability of their results.
Our study is limited by the fact that it included inpatients only, the majority of which were on antibiotics and had previously received antipyretics. However, our findings can be generalized to febrile children who are treated as outpatients, since the antipyretic response to acetaminophen is not known to vary between inpatients and outpatients. In addition, and since our outcome of interest is "effectiveness" rather than "efficacy", we did not exclude subjects receiving antibiotics from enrolment nor subjects with prior intake of antipyretics. Antipyretics however were stopped for 8 hours prior to enrolment, the time at which a febrile subject may receive antipyretic treatment in "real clinical life". It may be argued that the antipyretic effects of the investigational drugs are confounded by antibiotic administration and previous antipyretic intake. However since this is a randomized clinical trial, we anticipated that the randomization process will dilute these effects by distributing these subjects equally among the three treatment groups. Indeed, the proportions of subjects receiving antibiotics and those with prior antipyretic intake were not significantly different among the three groups, suggesting adequate randomization. A weakness of this study is the inter-individual variability of the acetaminophen dose in the rectal high-dose group which ranged between 27.2 and 43.4 mg/kg, after exclusion of the subject who received a low dose by mistake. It is possible that the lower acetaminophen doses in this range may have attenuated the mean antipyretic effect of this group resulting in similar antipyretic responses among the three different groups. This problem however is difficult to avoid with rectal administration and is frequently encountered in real clinical life. We cannot therefore eliminate the possibility of some imprecision of the results in the rectal high-dose group due to dosage variability.
It is interesting to note that one fifth of our patients developed hypothermia during the study interval, a finding that has not been previously reported. Though the differences in the proportions of patients with hypothermia among the three groups did not achieve statistical significance, the rectal high-dose group tended to have a higher proportion with hypothermia. This observation needs to be further investigated with a larger sample size, since our study was not powered to detect whether hypothermia is more common in one group as compared to the others.
Conclusion
In conclusion, oral and rectal acetaminophen preparations seem to have equal antipyretic effectiveness which is in line with earlier studies. There is no evidence to support the belief that rectal suppositories, whether prescribed in the standard dose of 15 mg/kg, or in the high dose of 30–40 mg/kg, are superior to oral acetaminophen in terms of rapidity of action, or in the extent of temperature reduction. Though the oral route may be preferred because of its predictable rapid absorption, the rectal route seems to be a good and equally effective alternative in special circumstances like vomiting, or conditions preventing oral administration. High-dose rectal acetaminophen should be used with caution, since it may result in hypothermia, a finding that deserves further exploration in the future. Physicians should educate parents about fever being a benign symptom of illness, rather than a disease in itself. While it is desirable to treat fever in children, parents need to be aware that fever per se is not a usual cause of mortality in a child, while acetaminophen overdose can be [1-3].
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
MN prepared grant submission in relation to this study, contributed to the design, data acquisition, analysis and interpretation, drafting, revision and final approval of the manuscript. HT contributed to the design, data analysis and interpretation, drafting, revision and final approval of the manuscript. RS participated in grant submission, design, revision and final approval of the manuscript. ZM contributed to statistical analysis, revision and final approval of the manuscript. SM and HD contributed substantially to data acquisition, drafting and final approval, MM participated in grant submission, drafting, revision and final approval of the manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
This study was funded by the Medical Practice Plan of the Faculty of Medicine, at the American University of Beirut, Grant number 686056. We are thankful to all the pediatric nurses at the American University of Beirut Medical Center and the Middle East Hospital for their enthusiastic and dedicated work that made this study possible. We also thank the following physicians for approving the enrolment of their patients in the study: from AUBMC, Drs. Majd Ariss, Bishara Atiyeh, Fadi Bitar, Ibrahim Dabbous, Ghassan Dbaibo, Salim Firzli, Nabil Fuleihan, George Haddad, Malek Hubballah, Salman Mroueh, Salim Musallam, Mounir Obeid, Sami Sanjad, Nabil Shararah, Samir Shehadeh, Jinan Usta, and Khaled Yunis; from MEH, Drs. Bassem Abou Merhi, Mohamad-Bilal Arab, Imad Chokor, Mohammad Itani, Imad Mofti, Muhieddine Mohebb, Zafer Shehadeh, and Ikram Tannir. We are also grateful to Dr. Samer Jabbour for his help in the design of the study, to Julphar (Gulf Pharmaceutical Industries, United Arab Emirates) for their generous donation of all the investigational drugs, and to Miss Huda Dagher, our research assistant, whose dedication and enthusiasm made this study possible.
==== Refs
American academy of Pediatrics Committee on Drugs: Acetaminophen toxicity in Children Pediatrics 2001 108 1020 1024 11581462 10.1542/peds.108.4.1020
Heubi JE Barbacci MB Zimmerman HJ Therapeutic misadventures with acetaminophen: hepatotoxicity after multiple doses in children J Pediatr 1998 132 22 27 9469995
Rivera-Penera T Gugig R Davis J McDiarmid S Vargas J Rosenthal P Berquist W Heyman M Ament M Outcome of acetaminophen overdose in pediatric patients and factors contributing to hepatotoxicity J Pediatr 1997 130 300 304 9042136
Anderson B Kanagasundarum S Woollard G Analgesic efficacy of paracetamol in children using tonsillectomy as a pain model Anaesth Intensive Care 1996 24 669 673 8971314
Birmingham PK Tobin MJ Henthorn TK Fisher DM Berkelhamer MC Smith FA Fanta KB Cote CJ Twenty-four-hour pharmacokinetics of rectal acetaminophen in children. An old drug with new recommendations Anesthesiology 1997 87 244 252 9286887 10.1097/00000542-199708000-00010
Bremerich DH Neidhart G Heimann K Kessler P Behne M Prophylactically-administered rectal acetaminophen does not reduce postoperative opioid requirements in infants and small children undergoing elective cleft palate repair Anesth Analg 2001 92 907 912 11273923 10.1097/00000539-200104000-00020
van der Marel CD van Lingen RA Pluim MA Scoones G van Dijk M Vaandrager JM Tibboel D Analgesic efficacy of rectal versus oral acetaminophen in children after major craniofacial surgery Clin Pharmacol Ther 2001 70 82 90 11452248 10.1067/mcp.2001.116794
Don Brown R Wilson JT Kearns GL Eichler VF Johnson VA Bertrand KM Single-dose pharmacokinetics of Ibuprofen and Acetaminophen in febrile children J Clin Pharmacol 1992 32 231 241 1564127
Wilson JT Don Brown R Kearns GL Eichler V Johnson V Bertrand KM Lowe BA Single-dose, placebo-controlled comparative study of ibuprofen and acetaminophen antipyresis in children J Pediatr 1991 119 803 811 1941390
Birmingham PK Tobin MJ Fisher DM Henthorn TK Hall SC Cote CJ Initial and subsequent dosing of rectal acetaminophen in children: a 24-hour pharmacokinetic study of new dose recommendations Anesthesiology 2001 94 385 389 11374595 10.1097/00000542-200103000-00005
Hahn TW Henneberg SW Holm-Knudsen RJ Eriksen K Rasmussen SN Rasmussen M Pharmacokinetics of rectal paracetamol after repeated dosing in children Br J Anaesth 2000 85 512 519 11064607 10.1093/bja/85.4.512
Montgomery CJ McCormack JP Reichert CC Marsland CP Plasma concentrations after high-dose (45 mg.kg-1) rectal acetaminophen in children Can J Anaesth 1995 42 982 986 8590508
Leary PM Walker KG van der Meulen W Antipyretic effect of oral versus rectal paracetamol [letter] S Alr Med J 1997 87 708
Scolnik D Kozer E Jacobson S Diamond S Young NL Comparison of oral versus normal and high-dose rectal acetaminophen in the treatment of febrile children Pediatrics 2002 110 553 556 12205259 10.1542/peds.110.3.553
Vernon S Bacon C Weightman D Rectal paracetamol in small children with fever Arch Dis Child 1979 54 469 479 475432
Anagnostakis D Matsaniotis N Grafakos S Sarafidou E Rectal-axillary temperature difference in febrile infants and children Clin Pediatr (Phila) 1993 32 268 272 8324970
|
16143048
|
PMC1215489
|
CC BY
|
2021-01-04 16:31:07
|
no
|
BMC Pediatr. 2005 Sep 6; 5:35
|
utf-8
|
BMC Pediatr
| 2,005 |
10.1186/1471-2431-5-35
|
oa_comm
|
==== Front
BMC PediatrBMC Pediatrics1471-2431BioMed Central London 1471-2431-5-361615014610.1186/1471-2431-5-36Research ArticleDifferences in tidal breathing between infants with chronic lung diseases and healthy controls Schmalisch G [email protected] S [email protected] RR [email protected] Clinic of Neonatology (Charité), Humboldt-University of Berlin, Germany2005 8 9 2005 5 36 36 11 5 2005 8 9 2005 Copyright © 2005 Schmalisch et al; licensee BioMed Central Ltd.2005Schmalisch et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
The diagnostic value of tidal breathing (TB) measurements in infants is controversially discussed. The aim of this study was to investigate to what extent the breathing pattern of sleeping infants with chronic lung diseases (CLD) differ from healthy controls with the same postconceptional age and to assess the predictive value of TB parameters.
Methods
In the age of 36–42 postconceptional weeks TB measurements were performed in 48 healthy newborns (median age and weight 7d, 3100 g) and 48 infants with CLD (80d, 2465 g)) using the deadspace-free flow-through technique. Once the infants had adapted to the mask and were sleeping quietly and breathing regularly, 20–60 breathing cycles were evaluated. Beside the shape of the tidal breathing flow-volume loop (TBFVL) 18 TB parameters were analyzed using ANOVA with Bonferroni correction. Receiver-operator characteristic (ROC) curves were calculated to investigate the discriminative ability of TB parameters.
Results
The incidence of concave expiratory limbs in CLD infants was 31% and significantly higher compared to controls (2%) (p < 0.001). Significant differences between CLD infants and controls were found in 11/18 TB parameters. The largest differences were seen in the mean (SD) inspiratory time 0.45(0.11)s vs. 0.65(0.14)s (p < 0.0001) and respiratory rate (RR) 55.4(14.2)/min vs. 39.2(8.6)/min (p < 0.0001) without statistically significant difference in the discriminative power between both time parameters. Most flow parameters were strongly correlated with RR so that there is no additional diagnostic value. No significant differences were found in the tidal volume and commonly used TB parameters describing the expiratory flow profile.
Conclusion
The breathing pattern of CLD infants differs significantly from that of healthy controls. Concave TBFVL and an increased RR measured during quiet sleep and under standardized conditions may indicate diminished respiratory functions in CLD infants whereas most of the commonly used TB parameters are poorly predictive.
==== Body
Background
With increasing numbers of infants born preterm, respiratory diseases associated with immature lungs and the need for mechanical ventilation or supplemental oxygen is becoming very common [1]. However, in the post surfactant era the "classic" or severe form of chronic lung disease (CLD) has been replaced by less severe forms ("mild" CLD) which are observed in very small premature infants who survive after prolonged mechanical ventilation [2,3]. Increasing awareness that both inflammation and disturbed lung development may cause respiratory problems in the middle and old age has emphasized the need for simple methods to assess lung function during the early age [4].
Several studies have shown that the degree of impaired respiratory function can be assessed by respiratory function testing [5-7]. However these complex techniques are limited to specialized centers. In contrast, tidal breathing (TB) measurements can be performed relatively easily in healthy and sick neonates at the bedside [8], and therefore they are frequently used for clinical and research purposes [9-11].
Air flow V'(t) and volume V(t) determined by numerical integration of V'(t) are the basic signals of a TB analysis. Both signals plotted together in an x-y diagram represent the tidal breathing flow-volume loop (TBFVL). In addition to assessing the shape of the TBFVL, several parameters were measured or derived from the time signals or the TBFVL to describe the breathing pattern.
Tidal breathing is commonly measured at the airway opening using a pneumotachograph (PNT) connected to a face mask [8]. In preterm and sick neonates, the total apparatus deadspace (VD, app) may exceed the infant's own deadspace and limits the time of measurement, even when a small face mask and a very low deadspace flow meter are used. Therefore for TB measurements in small infants, the deadspace-free flow-through technique (FTT) was developed [12] which virtually eliminates VD, app by a background flow. In contrast to other techniques, the elimination of VD, app by the FTT enables pneumotachographic long-term measurements which are an essential prerequisite for reliable TB measurements in infancy [12].
In the past tidal breathing measurements in CLD infants were commonly performed in infants with the severe form of CLD using the conventional technique (face mask and PNT) [13-15]. In these infants a significantly higher respiratory rate, a longer expiratory time, an earlier peak expiratory flow and concave expiratory limbs of the TBFVL were seen. Can we see such alterations also in the current milder forms of CLD and how many infants show such breathing pattern? Furthermore, Emralino and Steele [16] have shown that TB parameters measured with the conventional technique are significantly affected by the measuring technique itself and they cannot be compared with parameters measured during quiet breathing. Therefore the aim of this study was to investigate to what extent the breathing pattern of sleeping infants with the current mild forms of CLD differ from healthy controls with the same postconceptional age using the deadspace-free FTT and to assess the predictive value of the different TB parameters.
Methods
Subjects
In a prospective clinical study over 3 years TB measurements were performed in 96 infants (37 female and 59 male) with a postconceptional age between 36 and 42 weeks. All measurements were performed in the respiratory function laboratory of the Clinic of Neonatology at the Humboldt University (Charité). Inclusion criteria for this study were: spontaneous ventilation, quiet sleep according to Prechtl [17], and written parental consent.
Exclusion criteria were: upper respiratory airway disease and acute upper respiratory airway infection, congenital heart disease with exception of persistent ductus arteriosus, congenital diaphragmatic hernia and other pulmonary malformations, central nervous respiratory dysregulation.
In this study all CLD infants of the study period and the same number of healthy controls matched for postconceptional age were enrolled.
Group 1: 48 healthy newborns born in the Charité Hospital without any signs of respiratory disease or need for extra oxygen were recruited as controls.
Group 2: 48 infants with CLD came from different departments of neonatology. In our laboratory we measured our patients as well as patients admitted from other hospitals for pulmonary follow-up. All infants had a history of a severe postnatal RDS and required mechanical ventilation for >72 hours. The diagnosis of CLD was given by oxygen requirement at day 28 of age. Oxygen dependency was defined as the inability to keep oxygen saturation >92% in room air for at least 12 hours a day. 45/48 CLD infants were of a gestational age <32 weeks. 46/48 CLD infants were breathing room air at 36 weeks post-conception. According to the recently proposed diagnostic criteria of CLD by Jobe and Bancalari [3] the most CLD infants of this study had a mild form of CLD.
The patient characteristics are shown in Table 1. Although the controls were of the same postconceptional age, the controls were more mature at birth and heavier at the time of measurement. Because it is rare for infants born at 24 to 32 weeks gestational age not to develop any form of lung disease [5] a term control group of healthy newborns was used. Unfortunately, from the most of the outborn patients, no information was available about antenatal corticosteroids, postnatal surfactant treatment and the exact duration of mechanical ventilation and oxygen support, so that these important predictors could not be evaluated.
Table 1 Patient characteristics (median and range in brackets)
Healthy neonates (n = 48) CLD infants (n = 48)
Birth weight (g) 3280 (1610 – 4670) 890*** (450 – 3860)
Gestational age (weeks) 39 (34 – 41) 27*** (24 – 34)
Age (days) 7 (3 – 12) 86*** (33 – 125)
Postconceptional age (weeks) 39.6 (35.6 – 42.3) 38.7 (36.0 – 42.3)
Body weight at time of measurement (g) 3120 (1590 – 4580) 2400*** (1950 – 3800)
Comparison with healthy controls: *** p < 0.001
This study was approved by the ethical committee of the medical faculty (Charité) of the Humboldt University (protocol 54/92). Parents were given a full explanation of the tests and equipment used before their written consent was obtained.
Equipment for TB measurements
Tidal breathing was measured as described previously [12] using custom made equipment based on the flow-through technique. Briefly, the face mask is continuously rinsed thoroughly by a constant background flow higher than the infant's peak tidal inspiratory flow. The flow in and out of a modified transparent face mask (Vital Signs Inc., Totowa, USA) were measured by two screen PNTs (Baby PNT Jaeger, Wuerzburg, Germany) with a low flow resistance (0.2kPa·L-1·s). The infant's tidal flow was measured by the difference between the two flow signals. Both PNTs were calibrated simultaneously with room air at the beginning of each measurements using a 100 mL calibration syringe (Hans Rudolph, Kansas City, USA). The continuous background flow, which was generated by an air mixer and flow regulator, did not have a significant effect on the calibration and measurement accuracy. For background flows up to 7L/min the in-vitro volume error was <3% [18]. Changes of temperature, humidity, and gas viscosity between calibration and measurement were numerically corrected by the software.
The flow signal was filtered by an analogue Bessel filter of 4th order with 48Hz cutoff frequency to avoid aliasing, sampled with a 16 bit analogue/digital converter and recorded at 200 Hz.
Protocol of lung function testing
All infants were tested when well and clinically free from an upper airway infection since ≥3weeks. Most infants were studied during natural, quiet sleep, assessed by behavioural criteria [17], but 12 infants (12%) (1 healthy neonate and 11 CLD infants) were sedated with chloral hydrate (50 mg·kg-1) given orally 15–30 minutes before testing.
Sleeping infants were in a supine position with the neck in a neutral position supported with a roll. The background flow was adjusted (about the six fold of the expected minute ventilation of 220 mL/kg [18]) before the face mask was placed. Only in few infants an increase of the background flow was necessary to prevent rebreathing. After a period of accommodation (5–20 min), TB was measured while the airtight seal of the mask on the infant's face was checked by continuous leak monitoring [19]. The end of the accommodation period is commonly characterized by a more regular breathing pattern without any visible drift in TB parameters [20]. The graphical display of the instantaneous respiratory rate over the last 60 breathing cycles was used to assess the stability of the TB parameters. The duration of the TB measurements was normally 20–30 minutes depending on the period of accommodation to the face mask. All infants were continuously monitored by pulse oxymetry to prevent any adverse event particularly in sedated infants [21]. Parents were usually present during the respiratory function testing.
Depending on the variability of the breathing pattern, an interval of 20–60 consecutive artifact-free breaths with the same basic pattern of the TBFVLs were selected by the investigator and evaluated by the software at bedside as shown in Fig. 1. A subjective influence of the investigator on the calculated TB parameters can be excluded. From the recorded breathing cycles an averaged breathing loop was calculated as described previously [22]. After finishing the study the averaged loops were blinded and classified by three investigators (GS, SW, RW) according to typical patterns shown in Fig 2. A majority vote was accepted in only a small number of cases where f the loop pattern could not be clearly identified by all three investigators.
Figure 1 Tidal breathing parameters used in this study of a) the volume and b)the flow signals and c) the flow-volume loop. Abbreviations: tI-inspiratory time, tE-expiratory time, VT-tidal volume, PTIF, PTEF-peak tidal inspiratory and expiratory flow, tPTEF-time to peak tidal expiratory flow, TIF 50-tidal inspiratory flow when 50% of VT is inspired, VPTEF exhaled volume to peak tidal expiratory flow, TEF75, TEF50, TEF25- expiratory flow when 75%, 50% and 25% of tidal volume remains in the lung.
Figure 2 Typical shapes of tidal-breathing flow volume loops in newborns. To reduce the breath-to-breath variability a zeroing of volume at the begin of each inspiration was performed. In accordance with the common presentation of flow-volume loops, the inspiration started on the right side and continues in the lower quadrant, whereas the expiration follows in the upper quadrant.
For the quantitative evaluation, eleven basic parameters (Fig. 1) were measured from the flow and volume signals [inspiratory time (tI), expiratory time (tE), tidal volume (VT), tidal inspiratory flow when 50% of VT is inspired (TIF 50), peak tidal inspiratory and expiratory flow (PTIF, PTEF), time to peak tidal expiratory flow (tPTEF), exhaled volume to peak tidal expiratory flow (VPTEF) and expiratory flow when 75%, 50% and 25% of tidal volume remains in the lung (TEF75, TEF50, TEF25)]. From these parameters seven characteristic TB parameters were derived [respiratory rate (RR), minute ventilation (V'E), mean inspiratory flow VT/tI, mean initial expiratory gas acceleration PTEF/tPTEF, the ratios tPTEF/tE, VPTEF/VT, and the axis ratio of the TBFVL given by (PTIF+PTEF)/VT)].
Statistical methods
Patient characteristics are recorded as the median and range and compared using the Wilcoxon, Mann Whitney test. Differences in the pattern of the TBFVL between the groups were tested by means of the Chi2-test. Mean and standard deviations (SD) were calculated for all TB parameters, and analysis of variance (ANOVA) with Bonferroni correction for multiple comparisons was used to quantify differences between the patient groups. Birth weigh and gestational age were taken as covariates to investigate the effect of prematurity on the difference in TB parameters. Receiver operating characteristic curves (ROC) were calculated to investigate the discriminative ability of tidal breathing parameters in order to distinguish breathing patterns of both patient groups. The 95% confidence interval of the area under the normalized ROC curve (AUC) was calculated as described by Hanley and McNeil [23]. A level of statistical significance of p < 0.05 was accepted.
Results
The qualitative evaluation of the TBFVLs showed that the shape of the inspiratory limb was commonly convex and there were no statistically significant differences in the inspiratory shapes between the patient groups. In contrast to the inspiratory limb the shape of the expiratory limb after PTEF varied widely. Table 2 shows the distribution of typical patterns in both patient groups with significant differences (p < 0.001) in the distribution. As shown in this table, concave expiratory limbs were rarely seen in healthy infants (2%) but in about one third of all CLD infants.
Table 2 Distribution of typical shapes of the expiratory limb of the tidal breathing flow-volume loop after the peak tidal expiratory flow (PTEF) (absolute number and percentages in brackets)
Shape Healthy neonates (n = 48) CLD infants (n = 48)
Convex 10 (21%) 9 (25%)
Linear 31 (65%) 13 (27%)
Concave 1 (2%) 15 (31%)
Flow limitation 5 (10%) 5(11%)
Other shapes 1 (2%) 3 (6%)
The comparison of the TB parameters between CLD infants and healthy controls is shown in Table 3. Because the controls were significantly heavier at the day of measurement (Table 1), all flow and volume parameters were related to the body weight to reduce the inter-subject variability.
Table 3 Comparison of TB parameters between both patient groups ordered according to the p-value of the ANOVA (Presented are group means ± SD, statistically significant p-values after Bonferroni correction (p < 0.0028) are printed in bold)
Parameter Healthy neonates (n = 48) CLD infants (n = 48) p-value CLD
tI(s) 0.65 ± 0.14 0.45 ± 0.11 p < 0.0001
RR (min-1) 39.2 ± 8.6 55.4 ± 14.2 p < 0.0001
(PTIF+PTEF)/VT (s-1) 0.27 ± 0.06 0.37 ± 0.09 p < 0.0001
tE (s) 0.98 ± 0.24 0.72 ± 0.22 p < 0.0001
VT/tI(mL·s-1·kg-1) 8.9 ± 2.2 11.6 ± 2.8 p < 0.0001
V'E(mL·min-1·kg-1) 215 ± 51.7 276 ± 76.8 p < 0.0001
TIF 50 (L·min-1·kg-1) 0.75 ± 0.21 0.98 ± 0.26 p < 0.0001
PTIF (L·min-1·kg-1) 0.83 ± 0.20 1.05 ± 0.28 p < 0.0001
PTEF/tPTEF (L·s-2·kg-1) 2.90 ± 1.68 5.25 ± 3.66 p = 0.0001
PTEF (L·min-1·kg-1) 0.64 ± 0.19 0.82 ± 0.29 p = 0.0006
tptef (s)*) 0.24 ± 0.09 0.18 ± 0.11 p = 0.001
TEF75 (L·min-1·kg-1) 0.60 ± 0.20 0.76 ± 0.29 p = 0.003
TEF50 (L·min-1·kg-1) 0.52 ± 0.17 0.66 ± 0.25 p = 0.003
TEF25 (L·min-1·kg-1) 0.38 ± 0.11 0.46 ± 0.16 p = 0.006
Vptef (mL/kg)*) 1.68 ± 0.52 1.37 ± 0.44 p = 0.006
VT (mL·kg-1) 5.57 ± 1.06 5.15 ± 1.35 p = 0.09
VPTEF/VT(%)*) 29.4 ± 6.6% 27.2 ± 6.1% p = 0.13
tPTEF/tE (%)*) 25.8 ± 9.7% 23.2 ± 7.8% p = 0.20
*)Loops with flow limitations, grunting or other deformations were excluded from the evaluation (6 controls, 8 CLD infants)
Abbreviations: tI,E-inspiratory, expiratory time, RR-respiratory rate, PTIF, PTEF-peak tidal inspiratory and expiratory flow, VT-tidal volume, TIF 50-tidal inspiratory flow when 50% of VT is inspired, TEF 75, TEF 50, TEF 25-expiratory flow when 75%, 50% and 25% of tidal volume remains in the lung, V'E-minute ventilation, tPTEF,VPTEF-time and volume to peak tidal expiratory flow
The parameters in Table 3 were ordered according to the p-value of the ANOVA. The covariates birth weight and gestational age didn't have any statistically significant effect on the differences in TB parameters between the patient groups. The lowest p-values between the patient groups were found in the time parameters of the breathing cycle. The best discriminating parameter was tI whereas the differences in tE were distinctly lower. Nevertheless, the respiratory rate which is the reciprocal value of the sum tI+tE is one of the most important parameters distinguishing CLD infants from healthy controls.
The high differences in the axis ratio of the TBFVL (given by (PTIF+PTEF)/VT) between the groups can be explained by the strong correlation with RR. For a sinusoidal flow signal is RR = π·(PTIF+PTEF)/VT. Large differences were also found in the mean inspiratory flow VT/tI, minute ventilation V'E, and in the initial expiratory gas acceleration denoted as PTEF/tPTEF.
The differences between the patient groups in the flow parameters (PTIF, PTEF) are statistically significant, but considerably lower than the differences in RR. Furthermore, most flow parameters are strongly correlated with RR so that there is no additional diagnostic value. A surprising result is that the differences between the patient groups decreased in the course of expiration. In contrast to PTEF, the differences in the end-expiratory flow (TEF50, TEF25) are not statistically significant. No statistically significant differences were found in tidal volume related to the body weight and in the widely used TB-parameters tPTEF/tE and VPTEF/VT.
The discriminative power of TB parameters between both patients groups was investigated by using the ROC analysis (Fig. 3). In Table 4 the area under the ROC curve (AUC), the optimal cut-off value for each significantly different TB parameter and the resulting sensitivity and specificity were presented. For the best discriminating parameters tI and RR the area under the curve was not statistically different. This means that there will not be a large difference in the diagnostic value of tI compared to the more commonly used parameter RR. The sensitivity of both parameters using the optimal cutoff value was the same (70.8%). The sensitivity of the other TB parameters was distinctly lower and likely too low for the most clinical applications.
Figure 3 ROC curves of inspiratory time tI, respiratory rate (RR), tidal volume (VT) and the ratio tPTEF/tEbetween CLD infants and healthy controls.
Table 4 ROC analysis of commonly used TB parameters between CLD infants and healthy controls. If the 95% confidence interval (95% CI) of the area under the normalized ROC curve (AUC) include the 0.5 value (no discrimination) than there is no evidence that the TB parameters has the ability to distinguish between the two groups
Parameter AUC with 95%CI Optimal cut-off point Sensitivity Specificity
tI 0.879 (0.808 to 0.950) 0.48 s 70.8% 91.7%
RR 0.842 (0.754 to 0.909) 49.1 min-1 70.8% 89.6%
VT/tI 0.809 (0.721 to 0.896) 11.1 mL·s-1·kg-1 58.3% 89.6%
V'E 0.776 (0.682 to 0.869) 250 mL·min-1·kg-1 64.6% 85.4%
PTIF 0.747 (0.649 to 0.845) 0.95 L·min-1·kg-1 62.5% 79.2%
PTEF/tPTEF 0.688 (0.582 to 0.794) 4.57 L·s-2·kg-1 52.8% 87.5%
PTEF 0.673 (0.566 to 0.780) 0.79 L·min-1·kg-1 45.8% 83.3%
TEF25 0.653 (0.544 to 0.762) 0.67 L·min-1·kg-1 43.7% 87.5%
VT 0.610 (0.497 to 0.722) - - -
VPTEF/VT 0.565 (0.450 – 0.620) - - -
tPTEF/tE (%) .552 (0.437 – 0.670) - - -
(Abbreviation see Table 3)
Discussion
The main goal of this study was to investigate to what extent the tidal breathing pattern of CLD infants differ from healthy controls. We found in CLD infants a high incidence (31%) of concave TBVFL and significant differences mainly in the time parameters of the breathing cycle. tI was decreased and RR was increased in about 71% of all CLD infants whereas the most TB parameters were poorly predictive
The fact that time parameters show the largest differences between CLD infants and controls indicate the main problem of TB measurements. It is well recognized by numerous studies that the respiratory rate of an infant is affected by the measurement equipment itself [12,16,24], the time of measurement [20], behavioural states [17] or non-pulmonary diseases (e.g. infections). Therefore, the standardization of equipment and measuring conditions is an urgently necessary to obtain reliable results.
Techniques of tidal breathing measurements
For monitoring purposes tidal breathing in infants is commonly measured by indirect methods like breathing belts or measurement of transthoracic impedance changes. These techniques do not affect the air flow, however, accurate air flow measurements are only possible after circumstantial calibration and if measurements conditions are very stable [25]. Reliable TBFVL can not be obtained by indirect methods. Dead space free ventilatory measurements without any facial attachment are possible by "face out" body plethysmography. However, this technique is too expensive and cumbersome for routine bedside application [26]. Thus, the use of a face mask connected to a pneumotach is commonly used for precise ventilatory measurements [8].
The main problem of this conventional technique is the relatively high apparatus dead space which limits the duration of measurement due to CO2 rebreathing [27] so that a sufficient adaptation time after application of the face mask [28] can not be realized. Some of these influencing factors are eliminated by the FTT. The virtual elimination of the apparatus dead space by the background flow permits long-term measurements, so that the duration of measurements can readily be adapted to the prevailing measuring conditions (e.g., time required to reach a steady state) as well as to the variability of the respiratory signals. Furthermore, the airtight placement of the face mask can be monitored [19].
TB-Parameters
In the present study, the differences in the TB parameters between the patient groups are in good agreement with published results. Hjalmarson and Sandberg [5] recently showed in a prospective clinical study that preterm infants with mild or moderate CLD had significant higher RR compared to controls but no statistically significant changes in VT related to body weight. Ranganathan et al. [29] found in young infants with cystic fibrosis an elevated RR but no significant changes in the commonly used TB parameters. Tepper et al. [14] reported a significantly higher RR in CLD infants but no significant changes in VT. In the present study the unchanged tidal volume related to body weight and the much higher RR in CLD infants compared to healthy neonates explain the significantly higher flow parameters. In contrast to tidal breathing, it is well recognized that at forced expiration CLD infants show a significant flow limitation due to poor growth of the airways and the resulting higher peripheral airway resistance [14]. During tidal breathing we have never seen a reduced end-expiratory flow (TEF25) in CLD infants probably due to their high RR and the resulting higher expiratory flow rates.
The parameters tPTEF, tPTEF/tE as well as VPTEF, VPTEF/VT (which are strongly correlated), describe the site of PTEF in the flow and in the TBFVL, respectively. These parameters were frequently used in the past to detect airway obstructions [10,30,31]. However, the association of these parameters with small airway caliber remain speculative and could not be demonstrated in previously published studies [32,33]. In the present study we found only for tPTEF a statistically significant reduction in CLD infants The higher PTEF and the shorter tPTEF in CLD infants explain the significant differences between the patient groups in the mean initial expiratory gas acceleration given by PTEF/tPTEF.
In a recent study we have shown [34] that TB parameters of newborns describing the flow profile or the shape of the TBFVL describe rather the breathing strategy than an impaired lung function. Neonates have a highly compliant chest wall which may cause several problems during breathing e.g., small end-expiratory lung volume, low oxygen stores, and a high risk for airway occlusion and atelectasis [35]. Therefore, infants compensate for this mechanical disadvantage by actively maintaining lung volume above the resting volume which affects significantly the measured TB parameters. This may explain the decreasing significant differences in the TB parameters between the patient groups in the course of expiration.
An unexpected finding of our recent study [34] was that conventionally used TB parameters (e.g., RR, VT, V'E, mean V'I) are relatively robust against changes in dynamically elevated lung volume. Furthermore, these parameters are clearly defined and easy to derive from the measured respiratory signals. Therefore, TB measurements in neonates should be focused much more on the evaluation of these conventional parameters measured under standardized conditions which had also shown in the present study the highest discriminative potential.
The interpretation of TB measurements remains difficult because they reflect both the control of breathing and respiratory mechanics. Thus the breathing pattern can be influenced by factors other than impaired respiratory mechanics (e.g. changes in glottic aperture) and a TB measurement can never reveal impaired respiratory functions with complete reliability, because there is always the chance that the breathing pattern has been affected by an abnormality in the neural control of breathing [36]. Nevertheless, concave expiratory limbs (Table 2) were nearly exclusively seen in CLD infants (31%) and in about 70% of the CLD infants, RR was notably increased. Thus, despite its methodological limitations TB measurements can give valuable information about impaired respiratory function which should be investigated more in detail by further methods.
There are several limitations to the interpretation of our results. First, the control group was more mature at birth than the CLD infants so that the difference in TB parameters could also be affected by the immaturity of the CLD infants. However, this could not be confirmed in the ANOVA using birth weight and gestational age as covariates. Second, with exception of one healthy infant sedation was only used in CLD infants (23%). However, we did not find any statistically significant differences in the TB parameters between sedated and unsedated infants which is in well agreement with earlier investigations [12]. Third, from the majority of the admitted CLD infants important predictors (e.g. administration of steroids, duration of mechanical ventilation and oxygen therapy) were not known so that their influence on the changes in tidal breathing could not be investigated.
Reference ranges
The main problem of current TB measurements are the missing reference ranges. Despite repeated efforts during the last 50 years, the published reference values [11] are highly specific to the equipment used and the behavioral state of specific populations. They are unlikely to be of relevance when using other measurement techniques or equipment [12]. The fact that some TB parameters in neonates depend on the infant's breathing strategy makes it difficult to establish reference ranges for these parameters. In contrast, TB parameters describing breathing rate and depth are relatively independent from the breathing strategy and new efforts should be undertaken to determine reference ranges considering the biological development.
Conclusion
Diminished respiratory functions in infants after neonatal intensive care may be derived easily and non-invasively by TB measurements. Beside the shape of the TBFVL, time parameters of the breathing cycle showed the highest sensitivity. However, reliable measurements are only possible during quiet sleep and under standardized long-term measurements. Although the breathing pattern is affected by both the neural control of breathing and respiratory mechanics, TB-measurements can be used as a first-line tool in the respiratory function testing of infants after neonatal intensive care. However, the causes of the diminished respiratory function have to be investigated by more specialized methods.
Abbreviations
AUC – Area under the ROC curve
CLD – Chronic lung disease
FTT – Flow-through technique
PNT – Pneumotachograph
PTIF, PTEF – Peak tidal inspiratory and expiratory flow
RDS – Respiratory distress syndrome
ROC – Receiver operatic characteristic
RR – Respiratory rate
TB – Tidal breathing
TBFVL – Tidal breathing flow-volume loop
TEF 75, 50, 25 – Expiratory flow when 75%, 50% and 25% of VT remains in the lung,
tI, E – Inspiratory, expiratory time
TIF 50 – Tidal inspiratory flow when 50% of VT is inspired
tPTEF, VPTEF – Time and volume to peak tidal expiratory flow
V(t) – Volume
V'(t) – Air flow
V'E – Minute ventilation
VD – Dead space
VT – Tidal volume
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
GS and RW had primary responsibility for study design, protocol development, data analysis and writing of the manuscript. SW carried out all lung function measurements and GS performed statistical analysis. All authors read and approved the final manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
This work was supported by the German Ministry for Education and Research, project "Perinatal Lung" (grant 01 ZZ 9511). The authors thank Dr. Bertram Foitzik and Dr. Mario Schmidt for development of the hardware and software, Mrs. Silke Schmidt for her assistance in respiratory function testing and Dr. Noga Rogalla for the linguistic revision of the manuscript.
==== Refs
Bancalari E del Moral T Bronchopulmonary dysplasia and surfactant Biol Neonate 2001 80 7 13 11359038 10.1159/000047170
Bancalari E Changes in the pathogenesis and prevention of chronic lung disease of prematurity Am J Perinatol 2001 18 1 9 11321240 10.1055/s-2001-12940
Jobe AH Bancalari E Bronchopulmonary dysplasia Am J Respir Crit Care Med 2001 163 1723 1729 11401896
Hislop AA Lung growth and computed tomography Eur Respir J 2003 22 195 196 12952246 10.1183/09031936.03.00040603
Hjalmarson O Sandberg KL Lung function at term reflects severity of bronchopulmonary dysplasia J Pediatr 2005 146 86 90 15644829 10.1016/j.jpeds.2004.08.044
Lui K Lloyd J Ang E Rynn M Gupta JM Early changes in respiratory compliance and resistance during the development of bronchopulmonary dysplasia in the era of surfactant therapy Pediatr Pulmonol 2000 30 282 290 11015127 10.1002/1099-0496(200010)30:4<282::AID-PPUL2>3.0.CO;2-D
Schmalisch G Wauer RR Bohme B Effect of early ambroxol treatment on lung functions in mechanically ventilated preterm newborns who subsequently developed a bronchopulmonary dysplasia (BPD) Respir Med 2000 94 378 384 10845438 10.1053/rmed.1999.0751
Bates JH Schmalisch G Filbrun D Stocks J Tidal breath analysis for infant pulmonary function testing. ERS/ATS Task Force on Standards for Infant Respiratory Function Testing. European Respiratory Society/American Thoracic Society Eur Respir J 2000 16 1180 1192 11292125 10.1034/j.1399-3003.2000.16f26.x
Clarke JR Aston H Silverman M Evaluation of a tidal expiratory flow index in healthy and diseased infants Pediatr Pulmonol 1994 17 285 290 8058421
Greenough A Zhang YX Yuksel B Dimitriou G Assessment of prematurely born children at follow-up using a tidal breathing parameter Physiol Meas 1998 19 111 116 9522392 10.1088/0967-3334/19/1/010
Lodrup Carlsen KC Magnus P Carlsen KH Lung function by tidal breathing in awake healthy newborn infants Eur Respir J 1994 7 1660 1668 7995397 10.1183/09031936.94.07091660
Schmalisch G Foitzik B Wauer RR Stocks J Effect of apparatus dead space on breathing parameters in newborns: "flow-through" versus conventional techniques Eur Respir J 2001 17 108 114 11307739 10.1183/09031936.01.17101080
Durang M Rigatto H Tidal volume and respiratory frequency in infants with bronchopulmonary dysplasia (BPD) Early Hum Dev 1981 5 55 62 6781856 10.1016/0378-3782(81)90070-0
Tepper RS Morgan WJ Cota K Taussig LM Expiratory flow limitation in infants with bronchopulmonary dysplasia J Pediatr 1986 109 1040 1046 3783328
Gerhardt T Hehre D Feller R Reifenberg L Bancalari E Serial determination of pulmonary function in infants with chronic lung disease J Pediatr 1987 110 448 456 3819948
Emralino F Steele AM Effects of technique and analytic conditions on tidal breathing flow volume loops in term neonates Pediatr Pulmonol 1997 24 86 92 9292899 10.1002/(SICI)1099-0496(199708)24:2<86::AID-PPUL3>3.0.CO;2-G
Prechtl HF The behavioural states of the newborn infant (a review) Brain Res 1974 76 185 212 4602352 10.1016/0006-8993(74)90454-5
Schmalisch G Wauer RR Adjustment of the background flow for measurement of ventilation in newborn using the flow-through technique Pneumologie 1995 49 461 465 7479640
Foitzik B Schmidt M Windstetter D Wauer RR Schmalisch G Leak measurements in spontaneously breathing premature newborns by using the flow-through technique- Special Communication J Appl Physiol 1998 85 1187 1193 9729598
Patzak A Foitzik B Mrowka R Schmalisch G Time of measurement influences the variability of tidal breathing parameters in healthy and sick infants Respir Physiol 2001 128 187 194 11812383 10.1016/S0034-5687(01)00277-8
Malviya S Voepel-Lewis T Tait AR Adverse events and risk factors associated with the sedation of children by nonanesthesiologists Anesth Analg 1997 85 1207 1213 9390581 10.1097/00000539-199712000-00005
Schmalisch G Schmidt M Foitzik B Novel technique to average breathing loops for infant respiratory function testing Med Biol Eng Comput 2001 39 688 693 11804177
Hanley JA McNeil BJ The meaning and use of the area under a receiver operating characteristic (ROC) curve Radiology 1982 143 29 36 7063747
Dolfin T Duffty P Wilkes D England S Bryan H Effects of a face mask and pneumotachograph on breathing in sleeping infants Am Rev Respir Dis 1983 128 977 979 6650989
Brown K Aun C Jackson E Meckersie A Hatch D Stocks J Validation of respiratory inductive plethysmography using the Qualitative Diagnostic Calibration method in anaesthetized infants Eur Respir J 1998 12 935 943 9817172 10.1183/09031936.98.12040935
Edberg KE Sandberg K Silberberg A Sjoqvist BA Ekstrom Jodal B Hjalmarson O A plethysmographic method for assessment of lung function in mechanically ventilated very low birth weight infants Pediatr Res 1991 30 501 504 1754309
Katz SM Hertzberg T Lagercrantz H CO2-sensitivity in newborn and young infants tested by the rebreathing method. Methodological aspects Biol Neonate 1991 59 126 132 1905160
Fleming PJ Levine MR Goncalves A Changes in respiratory pattern resulting from the use of a facemask to record respiration in newborn infants Pediatr Res 1982 16 1031 1034 7155672
Ranganathan SC Goetz I Hoo AF Lum S Castle R Stocks J Assessment of tidal breathing parameters in infants with cystic fibrosis Eur Respir J 2003 22 761 766 14621082 10.1183/09031936.03.00024703
Martinez FD Morgan WJ Wright AL Holberg CJ Taussig LM Diminished lung function as a predisposing factor for wheezing respiratory illness in infants N Engl J Med 1988 319 1112 1117 3173442
Lodrup CKC Carlsen KH Nafstad P Bakketeig L Perinatal risk factors for recurrent wheeze in early life Pediatr Allergy Immunol 1999 10 89 95 10478609 10.1034/j.1399-3038.1999.00028.x
Aston H Clarke J Silverman M Are tidal breathing indices useful in infant bronchial challenge tests? Pediatr Pulmonol 1994 17 225 230 8208592
Williams EM Madgwick RG Thomson AH Morris MJ Expiratory airflow patterns in children and adults with cystic fibrosis Chest 2000 117 1078 1084 10767243 10.1378/chest.117.4.1078
Schmalisch G Wauer RR Foitzik B Patzak A Influence of preterm onset of inspiration on tidal breathing parameters in infants with and without CLD Respir Physiolo Neurobiol 2003 135 39 46 10.1016/S1569-9048(03)00029-6
Stark AR Cohlan BA Waggener TB Frantz ID Kosch PC Regulation of end-expiratory lung volume during sleep in premature infants J Appl Physiol 1987 62 1117 1123 3571069 10.1063/1.339720
Bates JHT Detecting airway obstruction from the tidal flow profile Eur Respir J 1998 12 1008 1009 9863988 10.1183/09031936.98.12051008
|
16150146
|
PMC1215490
|
CC BY
|
2021-01-04 16:31:06
|
no
|
BMC Pediatr. 2005 Sep 8; 5:36
|
utf-8
|
BMC Pediatr
| 2,005 |
10.1186/1471-2431-5-36
|
oa_comm
|
==== Front
BMC Plant BiolBMC Plant Biology1471-2229BioMed Central London 1471-2229-5-161610721210.1186/1471-2229-5-16Research ArticleDevelopment of ESTs from chickpea roots and their use in diversity analysis of the Cicer genus Buhariwalla Hutokshi K [email protected] Jayashree [email protected] K [email protected] Jonathan H [email protected] c/o JH. Crouch, Centro International de Mejoramiento de Maiz Y trigo (CIMMYT), Apdo. postal 6-641, 06600 Mexico, D.F., Mexico2 Centro International de Mejoramiento de Maiz Y trigo (CIMMYT), Apdo. postal 6-641, 06600 Mexico, D.F., Mexico3 International Crop Research Institute for the Semi-Arid Tropics (ICRISAT), Patancheru, Andhra Pradesh 502 324, India2005 17 8 2005 5 16 16 29 11 2004 17 8 2005 Copyright © 2005 Buhariwalla et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Chickpea is a major crop in many drier regions of the world where it is an important protein-rich food and an increasingly valuable traded commodity. The wild annual Cicer species are known to possess unique sources of resistance to pests and diseases, and tolerance to environmental stresses. However, there has been limited utilization of these wild species by chickpea breeding programs due to interspecific crossing barriers and deleterious linkage drag. Molecular genetic diversity analysis may help predict which accessions are most likely to produce fertile progeny when crossed with chickpea cultivars. While, trait-markers may provide an effective tool for breaking linkage drag. Although SSR markers are the assay of choice for marker-assisted selection of specific traits in conventional breeding populations, they may not provide reliable estimates of interspecific diversity, and may lose selective power in backcross programs based on interspecific introgressions. Thus, we have pursued the development of gene-based markers to resolve these problems and to provide candidate gene markers for QTL mapping of important agronomic traits.
Results
An EST library was constructed after subtractive suppressive hybridization (SSH) of root tissue from two very closely related chickpea genotypes (Cicer arietinum). A total of 106 EST-based markers were designed from 477 sequences with functional annotations and these were tested on C. arietinum. Forty-four EST markers were polymorphic when screened across nine Cicer species (including the cultigen). Parsimony and PCoA analysis of the resultant EST-marker dataset indicated that most accessions cluster in accordance with the previously defined classification of primary (C. arietinum, C. echinospermum and C. reticulatum), secondary (C. pinnatifidum, C. bijugum and C. judaicum), and tertiary (C. yamashitae, C. chrossanicum and C. cuneatum) gene-pools. A large proportion of EST alleles (45%) were only present in one or two of the accessions tested whilst the others were represented in up to twelve of the accessions tested.
Conclusion
Gene-based markers have proven to be effective tools for diversity analysis in Cicer and EST diversity analysis may be useful in identifying promising candidates for interspecific hybridization programs. The EST markers generated in this study have detected high levels of polymorphism amongst both common and rare alleles. This suggests that they would be useful for allele-mining of germplasm collections for identification of candidate accessions in the search for new sources of resistance to pests / diseases, and tolerance to abiotic stresses.
==== Body
Background
Chickpea (Cicer arietium L.) is one of the world's more important but less studied leguminous food crop with over 10 M ha grown across the Americas, the Mediterranean basin, East Africa, the Middle East, Asia and Australia [1]. As a grain legume it plays a significant role in the nutrition of the rural and urban poor in the developing world, as it provides a protein-rich supplement to cereal-based diets particularly of vegetarians and subsistence farmers who cannot afford meat. Despite its economic importance, chickpea productivity has been low because of yield losses to foliar and soil-borne fungal diseases (ascochyta blight, fusarium wilt and botrytis grey mould), insect pests (helicoverpapod borer) and abiotic stresses such as drought, cold and salinity. Sources of resistance and tolerance to these constraints exist in the wild Cicer germplasm yet remain largely unused by conventional breeding programs [2-5]. Three annual Cicer gene-pools have been defined based on the available hybridization reports, biochemical and molecular diversity analysis [6]: species within the primary gene-pool (C. arietinum, C. reticulatum and C. echinospermum) can be readily crossed usually generating fully fertile progeny; while species within the secondary gene-pool (C. bijugum, C. pinnatifidum and C. judaicum) can be successfully crossed with the cultigen C. arietinum, providing hybrid embryos are rescued. However, the progeny of crosses between primary and second gene-pools are frequently sterile; finally, species within the tertiary gene-pool (C. cuneatum, C. chorassanicum, C. yamashitae and others) have not yet been successfully crossed with the cultigen Cicer arietinum.
Cultivated chickpea has a relatively low level of diversity, due to a series of bottlenecks caused by restricted distribution of the wild progenitors and the founder effect associated with domestication [7]. In addition, chickpea breeding programs have limited themselves to a small number of cultivated genotypes having sources of biotic stress resistance and abiotic stress tolerance with little or no use of wild species [5]. This has resulted in limited durability of resistances to many of the major pests and diseases, and limited progress in abiotic stress tolerance breeding. Wild germplasm accessions held at the International Centre for Agricultural Research in Dry Areas (ICARDA) in Syria, the International Institute for the Semi-Arid Tropics (ICRISAT) in India and other genebanks in USA, Europe and Australia contain valuable sources of novel genetic variation for improvement of these traits [6]. The wild Cicer species from the secondary gene-pool (particularly C. bijugum, C. pinnatifidum and C. judaicum) are known to possess multiple sources of pest and disease resistance and tolerance to abiotic stresses including drought and cold [7,8].
Where species barriers can be overcome, molecular markers will facilitate rapid and efficient transfer of economically important traits into cultivated breeding pools. This will have the added advantage of broadening the narrow genetic base of this crop and thereby reducing its vulnerability to evolving pest and disease pressures, while simultaneously providing stable increases in yield far beyond the current 0.6% annual improvement [1].
Most types of molecular markers have been tested in chickpea including isozymes [9-12], restriction fragment length polymorphism (RFLP) [13,14] random amplified polymorphic DNA markers (RAPDs) [15,16] amplified fragment length polymorphisms (AFLPs) [17], sequence characterized amplified regions (SCARs) [18], inter-simple sequence repeat (ISSRs) [15], simple sequence repeat (STMS or SSR) [19-22], resistance gene analogs (RGAs) [23] DNA amplification fingerprinting (DAF) [24], and expressed sequence tags (ESTs) [25]. However, there is a low level of polymorphism detected in cultivated chickpea using isozyme/allozyme markers [9,10] and RFLP analysis [13,14]. In contrast, SSR markers have been shown to be highly polymorphic within the cultigen C. arietinum [26] and have been routinely utilized for creating genetic linkage maps [27-29]. These SSR markers have also been used as reference points for integration of the different chickpea linkage groups derived from inter- and intra- specific crosses [28,29]. To date there are over 500 SSR markers developed from chickpea genomic libraries [19,20,22]. In general, 30–50% of the chickpea SSR markers are polymorphic in any given breeding or intra-specific mapping population (HK Buhariwalla, unpublished data).
SSR markers remain the marker of choice for marker assisted selection in many breeding programs. However, SSR motifs may evolve too rapidly to be valuable as the sole assay for interspecific diversity analysis. In addition, markers shown to have tight genetic linkage to target genes in interspecific mapping populations may lose their selective power when used in backcross programs based on interspecific derivatives. Thus, there is a need for the development and utilization of gene-based markers that better serve molecular breeding applications and diversity analysis of germplasm.
In this context, the most important criteria for new markers are high reproducibility, detection of co-dominance polymorphism and suitability for rapid large-scale low cost screening. EST-based markers fulfill these criteria and since they are associated with the coding regions of the genome they also enhance molecular germplasm evaluation by capturing variation across transcribed regions and in genes of known function. This potentially resolves the problem of limited genomic coverage suffered by traditional single gene phylogenetic studies [30,31]. Targeted EST development has already begun in chickpea, focusing on ABA-related mechanisms of water-deficit tolerance in epicotyl tissue [32].
Our objective in this study was to increase the public EST resource for chickpea with a particular focus on root tissue, as constitutive drought tolerance mechanisms (such as root system development) are a common target for the crop physiology community [33] and are increasingly identified by genomics studies as important components of stress tolerance mechanisms [32,34]. Thus, an EST library was constructed after SSH of root tissue from two very closely related chickpea (C. arietinum) genotypes the landrace ICC 4958 and the popular local variety Annigeri, both considered to possess important sources of drought tolerance [35,36]. Following sequencing and functional annotation, we have generated over 100 EST markers which we have used in preliminary diversity analysis of wild and cultivated Cicer germplasm.
Results and discussion
The SSH process resulted in over 4000 clones expected to be largely associated with ICC 4958 (SSH tester) but could conceivably also include some highly constitutively expressed genes from Annigeri (SSH driver). Typically the cDNA inserts ranged from 0.5–1 kb. All clones were partially or fully sequenced through a single pass read from the 3' end and trimmed of vector and low quality sequences. A length threshold of 170 bp was set in view of the minimum expected size of a functional gene-encoding region (exon) [37]. Removal of short sequences (below 170 bp: largely comprising of problematic repeats and sequences predominantly composed of poly-A) resulted in 2858 high quality EST sequences with an average length of 480 bp all available at the ICRISAT Chickpea EST Database [38].
Analysis of ESTs and assembly of consensus sequences
Gene annotations were based on similarities to either known or putative ESTs in the public databases. All annotations were based on Blast searches, with a score threshold of ≥ 200 for BLASTn. For tBLASTx a score threshold of >100 was set, as these generally had e-values <10-5 with a minimum of 50% identity over at least 30% of the length of the protein, which are the commonly used thresholds for reliable sequence annotation [39,40]. Tentative functional annotations of EST sequences based on tBLASTx were grouped under 12 general categories based on the biochemical functions of the predicted proteins (Figure 1). The 2858 sequences were assembled into 210 contigs with 267 ESTs remaining unassociated (singletons). Putative identifications could be assigned to 2179 of the 2858 sequences (76%), representing 73% of the unigenes (178 singletons and 162 contigs), and all of these proteins were represented in other plant species. The 2858 sequences were assembled into 210 contigs with 267 ESTs remaining unassociated (singletons). Assuming that in most cases each contig represents one gene [41] a maximum number of 477 genes are represented in this dataset (available at NCBI: accession numbers CK148643–CK149150). The total length of contigs ranged from 240 to 1291 bases (with an average of 648 bases) while the largest number of ESTs assembled into one contig was 255. EST assembly leading to the generation of tentative consensus (TC) sequences has dramatically reduced the redundancy in this database, as reported elsewhere [42] (see Additional file 1).
Over three quarters of the 210 consensus sequences (77% equivalent to 164 contigs) were found to have significant similarity (tBLASTx >100) to sequences in public databases. Many of these annotations (47 consensus sequences) were validated through comparison with the CDD database using RPS-BLAST which compares a protein query sequence to a position-specific score matrix prepared from the underlying conserved protein domain alignment [43]. Conserved motifs consistent with tBLASTx annotations were identified in 47 consensus sequences providing an additional level of confidence for these assigned gene functions (see Additional file 2).
As expected, we observed a substantial number of the EST contigs and singletons to have no significant homology (NSH: 48 contigs and 89 singletons; 29%) to either nucleotide or protein sequences in public databases at the time of analysis. This compares well to the higher proportion of sequences (44%) reported with little or no homology from chickpea leaf tissue [44]. In addition, a substantial number of ESTs (approximately 19% of the unigenes) were found to have similarity with unannotated hypothetical, putative or unknown proteins contained in the public databases and were, therefore, placed in the unidentified function (UF) class.
PCR optimisation and unit costs
Optimization of PCR conditions is a critical precursor for accurate and robust large-scale marker screening and offers additional benefits for reducing the unit cost of genotyping [45]. As optimization relies on the sequential investigation of each reaction variable, the process can take a long time, thus the conditions of optimization are rarely identified empirically and individual components are rarely all tested simultaneously. However, Cobb and Clarkson [46] devised a strategy in which the components of a PCR reaction could be tested together using the least number ofexperiments based on the Taguchi principle [47]. We adapted the Cobb and Clarkson [46], 9 reactions of 4 PCR components to a 5 reactions of 5 PCR components, the additional component being enzyme concentration. This further reduced the time and cost of optimization, whilst also minimizing the cost of the PCR screening by also considering enzyme concentration. Final primer (0.15 - 0.5 pmol) and Taq DNA polymerase (0.2 – 0.5 U) concentrations resulted in a reagent unit cost of US$0.08 per PCR sample. This compares very favorable with other reports of PCR unit costs [48,49]. Clearly DNA extraction costs are an equally important component for molecular breeding programs and can also be greatly reduced as has been reported elsewhere [50]. The primer sequences and optimization parameters of all the polymorphic markers are given in Table 2. Optimization also resulted in fewer spurious amplification products and clearer polymorphic bands amongst the accessions studied (Figure 2).
EST marker diversity analysis
A total of 106 EST primers were designed, optimized and screened, of which 48 gave good amplification products in ICC 4958 (the SSH tester accession in the EST generation process) while 58 were deemed unsuitable as they either did not produce an amplification product or generated a complex pattern of bands which were difficult to evaluate. Of the 48 primers pairs screened across 1 cultivated and 8 wild Cicer species, 44 were polymorphic (Table 2). A total of 167 polymorphic fragments ranging in size from 200 to >1000 bp was scored on polyacrylamide or agarose electrophoresis gels depending on product size (Figure 2). The number of alleles detected per primer ranged from 2 – 8 with the polymorphic information content (PIC) ranging 0.03 – 0.89 (Table 5).
Fourteen of the 106 EST's contained SSR motifs, of which 10 contained perfect trinucleotide repeats. A relatively large proportion of these EST-SSR markers (8) were derived from the no significant homology (NSH) and unknown protein functional (UF) classes, based on sequence comparisons (BLASTn tBLASTx and PSI-BLAST). Markers from these EST classes have readily detected polymorphism in the diverse germplasm tested (PIC 0.20–0.82). Since over 48% of the EST's generated from the chickpea root library fall into NSH and UF classes, these are clearly useful markers with potentially important, yet unknown functions.
Eleven of the polymorphic EST's were from stress annotated transcripts (AGLC2, AGLC16, AGLC 20, AGLC29, AGLC 34, AGLC45, AGLC52, AGLC53, AGLC55, AGLC66, AGLC68 and AGLC93); of these four contained SSR motifs (AGLC 34, AGLC55, AGLC67 and AGLC84). These stress related EST markers may be useful for allele-mining of germplasm collections for the identification of candidate accessions with new sources of agronomically important traits and for candidate gene mapping [51].
The EST data-set generated in this study was used to compute pair-wise genetic distances between species according to Band similarity coefficient and UPGMA clustering. In order to display, with minimal distortion, the genetic relationships between species, a Principle Coordinate Analysis (PCoA) was carried out. The first two dimensions of the PCoA plot indicate the presence of 5 clusters, which account for 20% and 17% of the total variation (Figure 3). Dollo and polymorphism parsimony analysis (DOLLOP) conducted on this data-set found one most parsimonious tree with 301 steps (Figure 4) corresponding well with the PCoA analysis.
Both PCoA and parsimony analysis provide clear separation of most species. Both analysis grouped C. echinospermum, C. reticulatum and C. arietinum, together. These three species are classified together as the primary gene pool based on crossability studies. Similarly, all previous diversity studies based on allozymes [52], RAPD [16,53], microsatellite [54], ISSR [15,55] and AFLP [17] markers have also grouped C. echinospermum, C. reticulatum and C. arietinum together. The EST-based diversity analysis presented here shows C. arietinum (believed to be the wild annual progenitor [56] of cultivated chickpea) to be slightly distant to C. echinospermum and C. reticulatum but still within the same cluster (best reflected by the PCoA analysis Figure 3). All previous molecular diversity studies have clustered C. bijugum, C. pinnatifidum and C. judaicum together, although sometimes not well differentiated from the primary gene-pool species based on molecular diversity analysis. These three species are commonly considered to represent the secondary gene-pool as they require embryo rescue to recover viable progeny when crossed with the cultigen [6]. Both parsimony and PCoA analyses of the EST data-set clustered C. judaicum, C. pinnatifidum and C. bijugum together. This is in agreement with previous studies which demonstrated the close association of these three species. The current analysis suggests that C. pinnatifidum and C. judaicum are more closely related. This is in agreement with previous reports based on morphology analysis [56], seed storage protein analysis [57] and isozyme analysis [12]. In contrast, ecogeographical studies of the wild Cicer germplasm have suggested that C. judaicum to be quite diverged [7], which may explain the outlying position of one C. judaicum accession studied here (IG 69986) (Figure 4). Most other species generally fall into one of two clusters generally collectively referred to as the tertiary gene pool (Figure 3 and Figure 4) from which no viable hybrids have been reported from crosses with the cultigen.
Parsimony analysis grouped C. chorassanicum with C. cuneatum, but placed the C. yamashitae accessions as outliers (Figure 4). This data supports the tertiary group proposed by Croser et al. [6] to consist of; C. chorassanicum which has never been successfully crossed with cultivated chickpea and C. yamashitae and C. cuneatum which have been crossed with the cultigen but the resultant progeny have failed to flower or proven sterile. However, this is not reflected in the PCoA (Figure 3).
The distribution of accessions from the same species within secondary and tertiary clusters suggests that these two gene-pools may not be as distinct as previously expected. This may indicate that there are some tertiary gene-pool species that might be more readily crossed with the cultigen than their taxonomic classification would otherwise indicate. However, since the current study is based on a limited numbers of accessions for both C. cuneatum and C. yamashitae further research is required to fully define the relationships between these species and the scope of new opportunities for plant breeders.
Conclusion
This is the first report of the use of EST-based markers for estimating genetic distances between annual Cicer species. We report on the development and characterization of 106 EST markers from chickpea sequences with functional annotations or unknown functions, some of which contain SSR motifs. These markers have detected high levels of polymorphism amongst the wild species studied here and initial results indicate that around 20% are polymorphic in intraspecific (C. arietinum) mapping populations, either directly or as cleaved amplified polymorphic sequences (CAPS) based markers [58].
The EST marker-based diversity analysis reported here broadly supports the Cicer taxonomy based on allozymic data and conclusions by Croser et al. [6]. The concurrence of our EST data with that of allozymic studies in particular suggests that many gene-based loci do have a common ancestory across Cicer species, as proposed but not substantiated by Choumane et al. [54]. Clustering patterns based on RAPD, AFLP and ISSR are generally very similar [15,17,55] but are substantially different to those based on the EST marker analysis reported here and the previously reported allozyme-based analyses [59]. This may be a consequence of RAPD, AFLP and ISSR markers sampling fundamentally different regions of the genome (under different evolutionary pressures) as compared with markers based on expressed genes. Unfortunately a detailed comparison of different assays is confounded by the virtual absence of common accessions and the adoption of different statistical analyses in each study [6].
The EST database developed in this study provides a preliminary profile of some differentially expressed genes that may be associated with constitutive mechanisms important for stress tolerance and root development in chickpea roots. These include transcripts with putative annotations for proteases, T6P synthase, non-specific lipid transfer proteins, MRP-like ABC transporters, chaperones- HSP70, TCP-1-alpha, bZIP transcription factor, calcium ATPases, protein kinases, MRP4 glutathione-conjugate transporter, glutathione S-transferase, phosphoenol pyruvate carboxylase, S-adenosyl methionine synthetase (see Additional files 1 and 2 for full details). It is envisaged that these differentially expressed genes can be validated by the chickpea community and the most promising ESTs used in candidate gene mapping and allele mining. The resultant annotated EST markers will then be of substantial value for marker-assisted introgression programs based on interspecific crosses using wild Cicer species that harbour agronomically valuable genes. One of the aims of chickpea breeding is to address the continuing need for cultivars adapted to particular geographical regions or with specific new ideotypes to ensure sustainability and profitability of production. An important means of continuing to achieve such breeding objectives is through the use of novel germplasm [60]. Many breeding programs, in particular those involving rice and wheat, have successfully utilized molecular and statistical approaches in accelerating the introduction of novel genes from wild species through marker accelerated backcross breeding.
The availability of increasing amounts of sequence data in many legume species now offers the potential for routine development of gene-based markers. These provide the ultimate assay (or so-called 'perfect marker') for indirect trait selection and map-based cloning. Combining gene-based markers together with highly polymorphic flanking SSR markers will greatly assist in reducing linkage drag and increasing the speed and efficiency of subsequent introgression programs. Finally, molecular genetic diversity analysis based on EST markers may also provide a means of predicting which accessions are most likely to produce fertile progeny when crossed with chickpea cultivars.
Methods
Plant material and growth conditions for SSH
Fifteen seeds from each chickpea accession (ICC 4958 and Annigeri) were sterilized with 25% chlorex (v/v) for 10 mins, rinsed twice with sterile distilled water, before placing in pre-sterilized plant pots containing sterilized vertisol. The soil was collected from the field, passed through a 2 mm sieve and autoclaved in plastic bags using a 60 min cycle at 120°C and repeated three times on consecutive days. Seeds were sown immediately after the last autoclave cycle.
Three seeds were sown per pot; the plants were maintained in a Conviron growth chamber (Conviron, Winnipeg, Man.) under optimal physiological conditions for chickpea growth: 25°C day (11 h) and 12°C night (13 h), 455 μmoles m-2 sec-1 illumination intensity, 80% day and 40% night relative humidity, and irrigation as required.
Roots were harvested 28 days after germination (with flowering expected after 30 days). The soil was washed off the roots in running water, the intact whole plant was lifted out of the soil and soil particles were gently rubbed off in running water with gloved hands. The root tissue was subsequently treated with (0.1% v/v) DEPC solution and immersed in sterile Milli Q water, excess water from the roots was blotted onto filter paper and roots were weighed (approx. 2 g), flash frozen in liquid nitrogen and stored at -80°C.
RNA isolation and suppressed subtractive hybridization
The subtraction of the two genomes was achieved through a process of PCR-based suppression of genes common to both genotypes, and genes that are differentially expressed are enriched. As the process is based on PCR, low copy number cDNA can be detected from the tester (ICC 4958). RNA isolation and library construction was carried out by Avestha Gengraine Technologies (Bangalore, India). Total RNA was isolated from the root tissue of ICC 4958 and Annigeri using Trizol reagents (GIBCO-BRL) for RNA isolation, mRNA was purified using Oligotex (Qiagen GmbH, Hilden, Germany).
The suppression subtractive hybridization (SSH) process was carried out using a Clontech PCR-Select cDNA subtraction kit (Clontech INC. USA). Tester (ICC 4958) and driver (Annigeri) cDNA was restricted with RsaI for 90 mins and cDNA from the ICC 4958 was ligated with the required adaptors. Adaptor ligation was confirmed by PCR, followed by one round of over-night hybridisation with Annigeri as the driver and ICC 4958 as the tester. A second round of hybridisation was followed by primary PCR (27 cycles) and secondary PCR (12 cycles) according to the manufacturer's instructions for the PCR-Select. The PCR products of the subtraction were analyzed by gel electrophoresis and cloned using pCR4-TOPO cloning kit (Invitrogen Inc., USA). The cDNA clones were amplified in E. coli DH10B cells by electroporation and insert containing clones were selected by plating on LB agar containing 100 ug/ml ampicillin. Large colony forms were picked and used to regenerate single clone cultures in 96-well microtitre plates. After growth over-night at 37°C, glycerol was added to a final concentration of 15% and cultures were stored at -80°C.
EST sequencing
Cultures of transformed E. coli were grown over-night in 100 ul LB media containing 100 ug/ml ampicillin, the cells were centrifuged and suspended in sterile distilled water, heat denatured for 10 mins at 95°C, 1 ul of supernatant was used for insert amplification with M13 forward and reverse primers. Inserts (5 ul) were separated by electrophoresis on 1.2% agarose gel containing ethidium bromide to confirm amplification of a single insert and for quantification. After SAP/exonuclease I digest, 4 ul of the insert was sequenced using T7 primer and 1/8 fraction of the recommended volume of dye terminator reagent (Big Dye terminator cycle sequencing kit v2, ABI-Foster City) for cycle sequencing. After removal of dye blobs by ethanol precipitation, DNA sequences were analyzed by capillary-electrophoresis using an Applied Biosystems genetic analyzer model ABI 3700 or ABI 3100.
Sequence analysis for predicting gene function
Base calling was performed using the ABI DNA Sequence Analysis Software (v3.7) or by Chromas v2.2 (Technelysium Pty Ltd. Australia). All sequences were scanned visually for quality of the peak shapes and corresponding base call. Vector sequences were trimmed using Sequencher v.4 (Gene Codes, Ann Arbor, MI) software and low quality bases (quality score <20) were trimmed from both ends of sequences. A total of 4000 sequences were analyzed, only sequences of more than 170 bp with less than 5% ambiguity were processed. All sequences containing interspersed or simple repeats were masked with the RepeatMasker software [61] using Arabidoposis as reference. The masked sequences were screened against the following DNA databases: Arabidopsis thaliana, Medicago truncatula, Glycine max, Zea mays and Oryza sativa, in the TIGR Gene Indices using default settings for BLASTn and tBLASTx searches [62]. The same sequences were also used to search the NCBI EST database (dbEST) using Viridiplantae as a limiter [63].
The comparison of chickpea sequences with those in public databases was carried out during 2002 – 2003, and blast searches with unique sequences were repeated in November 2004. An in-house script was used to retrieve best matches from the outputs of BLASTn and tBLASTx text files, retaining the following information: gene identifier, description, BLAST score, percent identity, alignment length, reading frame, and database name from the top-scoring alignments. Alignments with scores of ≥200 for BLASTn and ≥100 for tBLASTx were extracted into an MS-Access table. The individual ESTs were assembled into tentative consensus sequences (TCs) using Sequencher, the assembly parameters used for 'dirty data' were (minimum percentage match of 90 and minimum overlap 40 bp). A sequence was retained in a contig if it matched at least one sequence already in the contig based on pair-wise BLAST [63], with a percentage identity threshold of 90%. A consensus sequence was then derived for each cluster. Consensus sequences were translated into six possible reading frames with Transeq software [64], using the standard genetic code. The translated sequences were then submitted to Reverse Position Specific (RPS)-BLAST [63] to search for matches to the conserved domain database (CDD; all 4540-PSSMs) using default parameters. The RPS-BLAST outputs were similarly processed using an in-house script, retaining the following information from the top-scoring alignments: gene identifier, description, score and e-value. Each result was manually interpreted and classified according to the biochemical function of the predicted protein.
Plant material and DNA isolation
Fifteen chickpea accessions representing 9 annual Cicer species, including the cultivated species C. arietinum, were used in this study (Table 1). Seeds of these accessions were germinated in Jiffy pots in a growth room and DNA was extracted from 4–5 pinnules of 15–20 day old individual plants with 2% CTAB extraction buffer containing 0.03% mercaptoethanol as described in [50]. DNA concentrations were determined by comparing the sample intensity (1 μl) with that of known amounts of uncut lambda DNA by electrophoresis in 1.2% 'Ready-to-run' agarose gels (Amersham) containing ethidium bromide. DNA was diluted to a working stock concentration of 10 ng/μl and checked before use.
Primer design of EST markers
A total of 106 EST sequences were chosen that represented diverse functional/stress annotations, a selection of which contained SSR motifs. In addition, primers were also designed from a random selection of sequences categorized as having 'no significant homology' (NSH) and 'unknown protein function' (UF) when compared with either nucleotide or protein sequences in public databases.
Primer sequences were designed using Genefisher 1.1 software [65] using the following criteria 57 – 63°C melting temperature (Tm), 40–50% GC content, 17–24 bp primer length and the difference in Tm between forward and reverse primer pairs was limited to 2°C. For those EST sequences containing repeat motifs, primers were designed to flank the microsatellite region. Optimal primer pairs were selected using Netprimer, the primer design algorithm which analyzes all possible secondary structures including hairpins, self-dimers and cross-dimers [66]. The Tm and product length of 8 forward and reverse pairs per sequence from Genefisher software were evaluated by Netprimer software for minimal or no secondary structure. All primer and EST sequences are accessible through the ICRISAT Chickpea EST Database [38].
PCR optimization and evaluation
The chickpea accession ICC 4958 (from which the EST sequences were derived) was used as the reference accession for primer optimization. A total of 106 EST primers (synthesized by MWG, Germany) were optimized simultaneously for the following PCR components using a modified (5×5) grid [46]. On this basis initial optimization of the five major components in a PCR (concentrations of primer, template DNA, Mg++, dNTP and enzyme) were empirically determined for each primer used in this study (Table 3). Optimal touch-down temperature and number of amplification cycle profiles were also determined for each primer pair (Table 4). This procedure is important to minimize non-specific amplification, maximize data accuracy and minimizes cost.
PCRs were performed in a total volume of 5 μl containing 5–10 ng of DNA and optimized concentration of the following: primer (0.15–0.3 pmoles), dNTP (0.1–0.2 mM), MgCl2 (1.5–2 mM), Qiagen Taq (0.2–0.5 u) and 1 x buffer. PCR amplifications were carried out using a GeneAmp model 9700 thermocycler (Perkin Elmer-Applied Biosystems). As the annealing temperatures of the markers ranged from 47–65°C we opted to use 3 catagories of "touchdown" temperature cycles (see Table 4).
PCR amplified products with a size range of 200–600 bp were resolved by non-denaturing polyacrylamide gels electrophoresis (6–10% acrylamide/bisacrylamide, 20:1 in TBE). Bands were visualized through a modified silver staining protocol [67] as follows: gels were immersed in water for 3 mins, followed by 20 mins in 0.1% CTAB solution and 0.3% ammonia solution for 15 mins. Silver staining solution was freshly prepared each day, consisting of 0.1 % (w/v) AgNO3 in 4 mM NaOH solution, 0.5 to 0.6 ml of 25% ammonia was titrated until the cloudy suspension became clear. Gels were gently agitated in the silver nitrate solution for 30 mins, and developed in 1.5% (w/v) sodium carbonate and 0.02% (v/v) formamide solution until bands appeared. The gels were rinsed in water, fixed in 1.5% glycerol solution and documented by scanning. PCR products above 600 bp were resolved by agarose gels electrophoresis (0.8–1%) containing ethidium bromide.
EST marker data collection and analysis
The amplification profile from each EST marker was checked for reproducibility and subsequently scored visually (and independently rescored) for the presence (1) or absence (0) of polymorphic bands. The degree of polymorphism was quantified using the polymorphic information content (PIC) calculator [68] based on the following formula:
PIC = 1 - ∑ Pi2 - ∑ ∑ Pi2Pj2
i = 1 i = 1 j = i+1
where Pi is the frequency of an individual genotype.
Pair-wise genetic distances were calculated with NTSYS-pc software version 2.0 [69] as Sxy = 1 - (2nxy/nx + ny) as Sxy = 1 - (2nxy/nx + ny) based on band coefficient [70], where nxy are shared bands amongst the individuals, nx and ny are the number of fragments exhibited by each individual. Distance based on Band coefficient is similar to the genetic distance (GD) formula by Nei and Li [71]: GD = 1 - Sxy. The 15 × 15 similarity matrix (based on Band coefficient) was subjected to sequential agglomerative hierarchical nested (SAHN) clustering using UPGMA (unweighted pair-group method analysis). Principle coordinate analysis based on these distance estimates was performed using NTSYS, using consecutive commands of 'DCENTRE' and 'EIGEN' in order to generate a PCoA plot which is more informative in highlighting differences between major taxonomic groups than dendrogram representations. Maximum parsimony analysis was conducted using the PHYLIP program Dollo and polymorphism parsimony analysis version 3.63 [72] and the consensus tree viewed in tree view.
Authors' contributions
HKB coordinated library construction and sequencing of a large proportion of the ESTs, experimental design, marker analysis and manuscript development. JB coordinated and/or carried out all bioinformatic analyses and participated in all aspects of manuscript development. KE carried out all aspects of marker optimization and screening, and contributed to various aspects of experimental design and data analysis. JHC participated in all aspects of experimental design, data interpretation and manuscript development and provided overall coordination of the project. All authors read and approved all versions of the manuscript.
Supplementary Material
Additional File 1
List of the most abundant EST from the chickpea root EST database. Number of chickpea ESTs forming tentative consensus sequences (TC) with corresponding TC number and where known, putative annotation*.
Click here for file
Additional File 2
Table of chickpea unigene sequences classified through RPS-Blast. Functional categories, with corresponding genbank ID numbers, Blastx descriptors, domains identified through RPS-Blast and e-values of chickpea unigene sequences.
Click here for file
Acknowledgements
This work was supported by ICRISAT core funds including earmarked unrestricted grants from the governments of UK, Japan and the European Union. The authors would like to thank Dr HC Sharma for providing access to chickpea material, Pratibha R. for technical assistance with initial EST marker screening, Dr HD Upadhyaya for provision of gene-bank passport data for accessions used in this study, Seetha K for lab management, PVNS Prasad and KDV Prasad for assistance with figures and to Drs R Serraj and R Folkertsma for helpful suggestions during manuscript development. We greatly appreciate suggestions from an anonymous reviewer concerning this manuscript.
Figures and Tables
Figure 1 A summary of the number of EST clones and respective gene contigs classified in various functional categories based on alignments with public databases. Putative identifications were assigned to 2130 of the 2858 sequences generated from a chickpea root subtractive hybridization library. Percentages indicate the proportion of unigenes from the total number of unigenes identified in whole dataset.
Figure 2 Polymorphic profiles of EST primers AGLC34, AGLC45, AGLC51 and AGLC52 screened on representatives of 8 wild species (IG69947, IG69960, IG69961, IG69974, IG69976, IG69986, IG69992, IG70029, ICC17116, ICC17121, ICC17122, ICC17126, ICC17141, ICC17148) and one cultivated genotype (ICC4958), separated by 6% non-denatured polyacrylamide gel electrophoresis and visualized by silver staining.
Figure 3 Principle Coordinate Analysis (PCoA) plot from diversity analysis of 44 EST markers screened across 14 accessions representing 8 wild Cicer species and one cultivated genotype (ICC4958).
Figure 4 Dendogram of 14 Cicer accessions (representing 8 wild species and one cultivated genotype, ICC4958) based on parsimony analysis of 103 EST alleles.
Table 1 Passport data of Cicer accessions used in this study.
Accessions Species Alternate accession identifier Biological status Origin and location Latitude Longitude
ICC4958 C. arietinum JGC 1 Cultivar India no data no data
IG69947 C. bijugum Wild Turkey (Diyarbakir) 40.0 33.1
ICC17122 C. bijugum ICCW 7 Wild Turkey (Savur) 37.5 40.9
IG70029 C. chorassanicum Wild Afghanistan (Bamian) 67.5 34.5
ICC17141 C. chorassanicum ICCW 26 Wild Afghanistan (Shahidan) 34.2 69.7
IG69976 C. cuneatum Wild Ethiopia (Tigray) 38.5 14.1
IG69974 C. echinospermum Wild Turkey (Sanli Urfa) 30.4 37.4
IG69986 C. judaicum Wild Syria (Tartous) 35.5 35.0
ICC17148 C. judaicum ICCW 33 Wild Lebanon no data no data
IG69961 C. pinnatifidum Wild Turkey (Elazig) 39.2 38.4
ICC17126 C. pinnatifidum ICCW 11 Wild Turkey 38.7 39.3
IG69960 C. reticulatum Wild Turkey (Mardin) 40.6 37.3
ICC17121 C. reticulatum ICCW 6 Wild Turkey (Savur) 37.5 40.9
IG69992 C. yamashitae Wild Afghanistan (Kabul) 69.4 34.4
ICC17116 C. yamashitae ICCW 1 Wild Afghanistan (Shezghan) 34.7 69.7
Table 2 Characteristics of chickpea EST markers developed from EST root library with primer sequence, GenBank accession numbers, PCR optimized parameters including touch down amplification profile.
Marker Primer Sequence (5'--3') Forward Primer Sequence (5'--3') Reverse GenBank accession numbers Touch -down profile (°C) PCR optimised reagents
Primer (pmol) Mg++ (mM) dNTP (mM) Taq (u)
AGLC2 TGTCAGACTGAGCTGTGTATGAGA TTGCCCGTATGGTTATGTTAGGAA CK148953* 55–45 0.3 1.5 0.15 0.3
AGLC7 GACCCCCAAAAATGAAAAAGCA TTGCCCATACATTCTTCACCCAA CK148966 60–55 0.3 1.5 0.15 0.3
AGLC8 CAAACTCCTCAATAGCAGGCACA GCTGTATCGGAGAGTGGTCAGA CK149041* 55–45 0.3 1.5 0.15 0.3
AGLC9 ACTCCTGTAGTGGCATATCTTCGA TGGTCCATTTATGCCGCTGGTA CK148709 60–55 0.3 1.5 0.15 0.3
AGLC14 GCAGCAACTATTTACACTGGTA CTCTCTGGGAGAAAGCTCGGAA CK149086* 60-55 0.45 1.5 0.15 0.3
AGLC15 ACTGATCAAGGTCTCTTCTAGACA CCCAACAAACTGGACAAAGCAGA CK149086* 60-55 0.45 1.5 0.15 0.3
AGLC16 GAGTACTTGCCAACTAGCTTAGGA TTGGATATAACAGATGACGGGGAA CK149087* 60-55 0.3 1.5 0.15 0.3
AGLC19 GCATCCTTCCCACTTCTTTGCA GAATGGACTCGGATGTCTTAAGCA CK148924* 60-55 0.3 1.5 0.15 0.3
AGLC20 AATGGTGATTCGTCAGTCGCCTA CTGTCTGAAGAAAGTGAACGAA CK148978* 60-55 0.3 1.5 0.15 0.3
AGLC27 CAAATTTCTGTTCTTCCACCCCAA GGCGATCTTCGAGTCCATCGA CK148933 60-55 0.3 1.5 0.15 0.3
AGLC28 GCTAAACCTTAGAGCAATGACTCA CCTTGCTTGTGCCTTATCTTCCA CK148934 60-55 0.3 1.5 0.15 0.3
AGLC29 TCTTCAACACCTCCATCTAACCTA GACATGAAACCAAAGCATCACA CK148945* 60-55 0.3 1.5 0.15 0.3
AGLC34 CTTTACCAAAACCACCTTCACCAA TCTCTCTCTCTCTCTTCTGTTCCA CK149140* 60-55 0.5 1.5 0.15 0.3
AGLC39 GCCGAGGTACACTTTACCAA TCCTCACACTTCAGGTTCAACGA CK149136 55-45 0.5 1.5 0.15 0.3
AGLC45 CTCCTCTTCTCCGTCGTAGCA CTGGTCCTTGACGGGAGTGA CK149070* 65-60 0.5 1.5 0.15 0.3
AGLC47 GTTTACATCATGACCCGCCCTA TCACCAAGACCAGAACGTTCCA CK148819 60-55 0.5 1.5 0.15 0.3
AGLC48 TGCCCAACGGTTTCTTTTACCA TCAGAGATACTCGCCCACCAA CK148960 55-45 0.5 1.5 0.15 0.3
AGLC51 TCTTTGAGCAGCATTCATTCCACA GAGTGCTACCTTCAAAGACTGCA CK148677 55-45 0.5 1.5 0.15 0.3
AGLC52 CGATCAAGAACCCAGTTTTGCAA AAGATCGACAGGCGATCTGGTA CK148718 55-45 0.5 1.5 0.15 0.3
AGLC53 CACTCTCCGTTCCGGTTCCA CTGTCCATGCCCTTGTCCA CK148806 60-55 0.3 1.5 0.2 0.5
AGLC55 CAGGTCGCGTTGTTGCA GGCCGAGGTACACTTTTCCA CK149133* 60-55 0.5 2 0.1 0.3
AGLC57 TTCATCTGGCACTAGCATATCTGA CGACAATTCTTGCTTCAACAACCA CK148900* 60-55 0.3 2 0.1 0.3
AGLC60 CATGTTTTCTACCCTCACAATGCA TACTCACTTGTTGTTCCAGACA CK149124* 55-45 0.2 1 0.1 0.2
AGLC61 TTCGATCCTCCGACCCCGAA TTCGCTAGATCTGGATACTTCTCA CK148802 60-55 0.3 2 0.1 0.3
AGLC64 TCTTCTTCTTCTTCTTCAGCCACA GTGGATTGGGAAATGTGAATGTCA CK149128* 55-45 0.2 1 0.1 0.2
AGLC66 CCACAAAGGACGACAACAACGA CCCAACACGAACCACACGA CK149070 60-55 0.2 2 0.2 0.5
AGLC67 ATCCATCACAACCCTCAACTCA CTCCGTCAACCTTTCCGCAA CK149102* 55-45 0.3 2 0.1 0.3
AGLC68 TGTTGTCTCGCCAATTCAAAGCA CGTTTGGTGGCATTCCTGCA CK148906* 55-45 0.3 2 0.1 0.3
AGLC72 TTTAATTACGCGGTTTCCACGA GAAGACTTGAGACATGGGCACA CK148928* 55-45 0.2 1 0.1 0.2
AGLC74 CGTGGGATTGAAAAAGTTGCTA CACTACCAGCCAAAGCACTCA CK149006* 55-45 0.2 1 0.1 0.2
AGLC75 CAACAACAACCTATCCGAACCTCA ACTATCCCTAACCTTCCATCACCA CK148853 60-55 0.5 1 0.2 0.5
AGLC76 CATGAGTGGTAGTGGGAGTGGA GTTCGTTTGAGTCGTTTACTGGAA CK148862 60-55 0.2 2 0.2 0.5
AGLC78 TCAACAACGCTACCCGATCCAA TTCTCAAGAGCACCACAAAAGAGA CK149132 60-55 0.3 2 0.1 0.3
AGLC79 CGGCGGCTATATTGGTTTTGCA TCCTAAACCCCACTTATCTCCCTA CK149080* 60-55 0.3 2 0.1 0.3
AGLC82 TTTGTGATGGTCCTGCTCTCTCA ACCGCTTCAGGATCAACTCGA CK148871 60-55 0.3 2 0.1 0.3
AGLC83 TCTTCCGATCCTAAGAAAGAGCAA ACCAATATGGAGAGCACCAGTCA CK148894 60-55 0.3 2 0.1 0.3
AGLC84 CCACCTTCCATCTCCAATTCCAA GACTGAATCGGAGAAGGTTTCTCA CK149089 60-55 0.15 2 0.1 0.2
AGLC85 CCAGCTTCTAATGTAGGTCTGCA CAGCAGCAGCAGAGAGAGCA CK148761 60-55 0.15 1 0.1 0.2
AGLC88 ACTTGGGCGTTCAAAAATCTCA CCATTACGATCAAAGAGCTCAGGA CK148768 55-45 0.3 2 0.1 0.3
AGLC93 GTCCGAGCTGTGGATAGGGAA GTTCCGCCTTCAATCCATGGAA CK148904* 60-55 0.2 2 0.2 0.5
AGLC94 CCAACTTCCCTCATTCTTATTCCA ACCAATTCCAAATTTCCAGCTCGA CK149093 55-45 0.2 1 0.1 0.2
AGLC96 TCCATATGGCTGAAGAACCCCAA TTCTGAGGTTCAGGTAGTTCGGAA CK149016* 60-55 0.2 2 0.2 0.5
AGLC98 CTCTTTCTTTCCCTCTAGTTTCCA CGGCGAACTCGTGTTTGCTA CK149016 55-45 0.2 1 0.1 0.2
AGLC101 TGTCCAAAATTGGGATCAGAGA AGAACGACTTCAGCAGCAGCA CK148993* 55-45 0.3 2 0.1 0.3
*Genbank accession numbers of the longest EST sequence of the contig, used in primer design.
Table 3 Permutations of PCR component concentrations used for optimization of PCR conditions for each marker.
Reactions Primer (pmol) Template (ng) MgCl2 (mM) dNTP (mM) Enzyme (u)
1 0.2 5 1 0.1 0.2
2 0.2 15 2 1.5 0.5
3 0.3 5 1.5 0.2 0.5
4 0.3 10 2 0.1 0.3
5 0.5 10 1 0.2 0.3
Table 4 PCR amplification profiles used for optimization of PCR conditions for each marker.
Name Profile Temperature°C Time min. sec Number of cycles
65-60 Denaturation 95 3.00 1
Touch down 94 0.20 5
65-60 0.20
72 0.30
Normal 94 0.20 30
59 0.20
72 0.30
Extension 72 20.0 1
60-55 Denaturation 95 3.00 1
Touch down 94 0.20 5
60-55 0.20
72 0.30
Normal 94 0.20 30
56 0.20
72 0.30
Extension 72 20.0 1
55-45 Denaturation 95 3.00 1
Touch down 94 0.20 10
55-45 0.20
72 0.30
Normal 94 0.20 30
48 0.20
72 0.30
Extension 72 20.0 1
Table 5 Characteristics of chickpea EST markers including SSR motif where present, annotation information (NSH indicates no significant homology), and, size range of amplification products, number of alleles and PIC (polymorphism information content) values based on screening of germplasm in Table 1.
Marker SSR motif Annotation Product size (bp) No. of alleles PIC
AGLC2 Valyl tRNA synthetase 620 2 0.50
AGLC7 D-3-phosphoglycerate dehydrogenase 330–335 2 0.44
AGLC8 RNA helicase 525–550 5 0.78
AGLC9 Unknown protein 680 4 0.62
AGLC14 Hypothetical protein 520–600 2 0.45
AGLC15 Hypothetical protein 520–600 4 0.65
AGLC16 WD-repeat protein like 365–370 2 0.32
AGLC19 Formin-like protein 350–355 4 0.50
AGLC20 Type IIB calcium ATPase 230–235 3 0.61
AGLC27 Putative Scarecrow gene regulator 330–335 3 0.43
AGLC28 Type 2A protein phosphatase 335–350 3 0.86
AGLC29 Non-specific lipid transfer protein 325–400 6 0.64
AGLC34 AG(20) Putative protein kinase 350–600 5 0.75
AGLC39 AG(17) No significant homology 275–290 3 0.74
AGLC45 Probable cysteine proteinase (EC 3.4.22.-) 410–420 2 0.29
AGLC47 (NM_127184) putative beta-ketoacyl-CoA synthase 200–210 2 0.59
AGLC48 Putative peroxisome assembly factor-2 (NM_100181) 400–410 3 0.89
AGLC51 RNA binding protein 170–180 4 0.55
AGLC52 Protein kinase 390–410 6 0.89
AGLC53 Acyl- [acyl-carrier protein] desaturase 390–400 5 0.87
AGLC55 CT(17) Putative protein kinase 225–275 4 0.89
AGLC57 GGA(3) Unknown protein; protein >1000 3 0.56
AGLC60 TTC(3) Hypothetical protein 310–350 7 0.82
AGLC61 GAA(3) Protein of unknown function 450–470 5 0.55
AGLC64 TTC(6) No significant homolgy 270–360 4 0.20
AGLC66 Probable cysteine protenase (EC 3.4.22.-) 500–650 8 0.63
AGLC67 TCT(4) Tryptophan synthase beta-subunit (TSB2) >1000 4 0.56
AGLC68 Lon protease homolog 1 precursor 220–230 3 0.39
AGLC72 Putative Grr1 protein 230–250 4 0.62
AGLC74 TTTA(3) Glycosyl hydrolase family 17 400–450 5 0.75
AGLC75 AAC(5) No significant homology 200–230 3 0.48
AGLC76 AAG(3) RGA2 protein Arabidopsis thaliana 440–480 3 0.71
AGLC78 CAA(5) Translation initiation factor eIF3 – like protein 430–450 2 0.44
AGLC79 No significant homology 330–410 5 0.03
AGLC82 Unknown protein; protein id: At1g80280.1 300–320 3 0.50
AGLC83 Putative valyl tRNA synthetase [Oryza sativa-japonica cultivar-group] >1000 3 0.74
AGLC84 AAG(3) DEAD/DEAH box RNA helicase 430–450 3 0.42
AGLC85 Emb trypsin inhibitor cme precursor 410–450 3 0.47
AGLC88 Chromomethylase CMT2 Arabidopsis thaliana >1000 5 0.78
AGLC93 BZIP family transcription factor >1000 2 0.51
AGLC94 T48 ankyrin-like protein 480–500 4 0.67
AGLC96 Hypothetical protein CAB95829.1 310–360 5 0.59
AGLC98 AG(19) Hypothetical protein 405–450 5 0.78
AGLC101 No Significant homology 370–400 4 0.70
==== Refs
FAOSTAT data 2004
Erskine W Siddique K Khan T Cowling W Maxted N, Bennett SJ Utilization of grain legume diversity Plant genetic resourses of legumes in the Mediterranean 2001 Kluwer Academic, The Netherlands 311 326
Knight E Brinsmead R Fordyce M Wood J Kelly A Harden S Use of wild relative Cicer echinospermum in chickpea improvement Proceedings of the 12 th Australian Plant Breeding Conference Perth Australia 2002 150 154
Muehlbauer FJ Kaiser WJ Simon CJ Potential for wild species in cool season food legume breeding Euphytica 1994 73 109 114 10.1007/BF00027187
Singh KB Malhotra RS Halila H Knight EJ Verma MM Current status and future strategy in breeding chickpea for resistance to abiotic and biotic stresses Euphytica 1994 73 137 149 10.1007/BF00027190
Croser JS Ahmad F Clarke HJ Siddique KHM Utilisation of wild Cicer in chickpea improvement- progress, constraints and prospects Aust J Agric Res 2003 54 429 444 10.1071/AR02157
Berger J Abbo S Turner NC Ecogeography of annual wild Cicer species: the poor state of the world collection Crop Sci 2003 43 1076 1090
Singh KB Ocampo B Robertson LD Diversity for abiotic and biotic stress resistance in the wild annual Cicer species Genet Res Crop Evol 1998 45 9 17 10.1023/A:1008620002136
Gaur PM Slinkard AE Inheritance and linkage of isozyme codominant genes in chickpea J Hered 1990 81 455 461
Gaur PM Slinkard AE Genetic control and linkage relations of additional isozyme markers in chickpea Theor Appl Genet 1990 80 648 656
Kazan K Muehlbauer FJ Weeden NF Ladizinsky G Inheritance and linkage relationships of morphological and isozyme loci in chickpea Theor Appl Genet 1993 86 417 426 10.1007/BF00838556
Labdi M Robertson LD Singh KB Charrier A Genetic diversity and phylogenetic relationships among the annual Cicer species as revealed by isozyme polymorphism Euphytica 1996 88 181 188 10.1007/BF00023889
Udupa SM Sharma A Sharma RP Pai RA Narrow genetic variability in Cicer arietinum L. as revealed by RFLP analysis J Plant Biochem Biotech 1993 2 84 86
van Rheenen HA Reddy MV Kumar J Haware MP Singh KB, Saxena MC Breeding for resistance to soil-borne diseases in chickpea Disease Resistance Breeding in Chickpea 1992 ICARDA, Aleppo, Syria 55 70
Iruela M Rubio J Cubero JI Gil J Millán T Phylogenetic analysis in the genus Cicer and cultivated chickpea using RAPD and ISSR markers Theor Appl Genet 2002 104 643 651 12582669 10.1007/s001220100751
Sudupak M Akkaya M Kence A Analysis of genetic relationships among perennial and annual Cicer species growing in Turkey using RAPD markers Theor Appl Genet 2002 105 1220 1228 12582902 10.1007/s00122-002-1060-8
Nguyen TT Taylor PWJ Reeden RJ Ford R Genetic diversity estimates in Cicer using AFLP analysis Plant Breeding 2004 123 173 179 10.1046/j.1439-0523.2003.00942.x
Mayer MS Tullu A Simon CJ Kumar J Kraft JM Muehlbauer FJ Development of a DNA marker for fusarium wilt resistance in chickpea Crop Sci 1997 37 1625 1629
Hüttel B Winter P Weising K Choumane W Weigand F Sequence-tagged microsatellite site markers for chickpea (Cicer arientinum L.) Genome 1999 42 210 217 10231957 10.1139/gen-42-2-210
Lichtenzveig J Scheuring C Dodge J Abbo S Zhang H-B Construction of BAC and BIBAC libraries and their applications for generation of SSR markers for genome analysis of chickpea, Cicer arietinum L Theor Appl Genet 2005 110 492 510 15712010 10.1007/s00122-004-1857-8
Niroj SK Bhumika S Bhatia S Isolation and characterization of a sequence-tagged microsatellite site markers in chickpea (Cicer arietinum L.) Mol Ecol Notes 2003 3 428 430 10.1046/j.1471-8286.2003.00472.x
Winter P Pfaff T Udupa SM Hüttel B Sharma PC Sahi S Arreguin-Espinoza R Weigand F Muehlbauer FJ Kahl G Characterisation and mapping of sequence-tagged microsatellite sites in the chickpea (Cicer arietinum L.) genome Mol Gen Genet 1999 262 90 101 10503540 10.1007/s004380051063
Hüttel B Santra D Muehlbauer F Kahl G Resistance gene analogues of chickpea (Cicer arietinum L.): isolation, genetic mapping and association with a Fusarium resistance gene cluster Theor Appl Genet 2002 105 479 490 12582554 10.1007/s00122-002-0933-1
Benko-Iseppon AM Winter P Huettel B Staginnus CM Muehlbauer FJ Kahl G Molecular markers closely linked to fusarium resistance genes in chickpea show significant alignment to pathogenesis-related genes located on Arabidopsis chromosomes 1 and 5 Theor Appl Genet 2003 107 379 386 12709786 10.1007/s00122-003-1260-x
Pfaff T Kahl G Mapping of gene-specific markers on the genetic map of chickpea (Cicer arietinum L.) Mol Genet Genomics 2003 269 243 251 12756536
Udupa SM Robertson LD Weigand F Baum M Kahl G Allelic variation at (TAA)n microsatellite loci in a world collection of chickpea (Cicer arietinum L.) germplasm Mol Gen Genet 1999 261 354 63 10102371 10.1007/s004380050976
Tekeoglu M Rajesh N Muehlbauer FJ Integration of sequence tagged microsatellite sites to the chickpea genetic map Theor Appl Genet 2002 105 847 854 12582909 10.1007/s00122-002-0993-2
Udupa SM Baum M Genetic dissection of pathotype-specific resistance to ascochyta blight disease in chickpea (Cicer arietinum L.) using microsatellite markers Theor Appl Genet 2003 106 1196 1202 12748770
Winter P Benko-Iseppon AM Hüttel B Ratnaparkhe M Tullu A Sonnante G Pfaff T Tekeoglu M Santra D Sant VJ Rajesh PN Kahl G Muehlbauer FJ A linkage map of the chickpea (Cicer arietinum L.) genome based on recombinant inbred lines from a C. arietinum × C. reticulatum cross: localization of resistance genes for fusarium wilt races 4 and 5 Theor Appl Genet 2000 101 1155 1163 10.1007/s001220051592
Doyle JJ Doyle JL Harbison C Chloroplast-expressed glutamine synthease in Glycine and related Leguminosae: Phylogeny, gene duplication and ancient polyploidy Systematic Botany 2003 28 567 577
Kajita T Ohashi H Tateishi Y Bailey CD Doyle JJ rbcL and legume phylogeny, with particular reference to Phaseoleae, Millettieae, and Allies Systematic Botany 2001 26 515 536
Romo S Labrador E Dopico B Water stress-regulated gene expression in Cicer arietinum seedlings and plants Plant Physiol Biochem 2004 39 1017 1026 10.1016/S0981-9428(01)01318-3
Sinclair TR Muchow RC System analysis of plant traits to increase grain yield on limited water supplies Agron J 2001 93 263 270
Colmenero-Flores JM Moreno LP Smith CE Covarrubias AA Pvlea-18, a member of a new late-embryogenesis-abundant protein family that accumulates during water stress and in the growing regions of well-irrigated bean seedlings Plant Physiol 1999 120 93 103 10318687 10.1104/pp.120.1.93
Saxena NP Krishnamurthy L Johansen C Registration of a drought resistant chickpea germplasm Crop Sci 1993 33 1424
Saxena NP Saxena NP Management of drought in chickpea – a holistic approach Management of Agricultural drought – Agronomic and Genetic Options 2003 Oxford & IBH Publishers Co Pvt. Ltd, New Delhi, India 103 122
Glick B Pasternak J Molecular biotechnology: Principles and applications of recombinant DNA 1998 American Society for Microbiology; Washington DC, USA
ICRISAT chickpea EST database:
Quackenbush J Liang F Holt I Pertea G Upton J The TIGR Gene Indices: reconstruction and representation of expressed gene sequences Nucl Acids Res 2000 28 141 145 10592205 10.1093/nar/28.1.141
Ronning CM Stegalkina SS Ascenzi RA Bougri O Hart AL Utterbach TR Vanaken SE Riedmuller SB White JA Cho J Pertea GM Lee Y Karamycheva S Sultana R Tsai J Quackenbush J Griffiths HM Restrepo S Smart CD Fry WE Van Der Hoeven R Tanksley S Zhang P Jin H Yamamoto ML Baker BJ Buell CR Comparitive analysis of potato expressed sequence tag libraries Plant Physiol 2003 131 419 429 12586867 10.1104/pp.013581
White JA Todd J Newman T Focks N Girke T Ilarduya OM Jaworski JG Ohlrogge JB Benning C A new set of Arabidopsis Expressed Sequence Tags from developing seeds. The metabolic pathway from carbohydrates to seed oil Plant Physiol 2000 124 1582 1594 11115876 10.1104/pp.124.4.1582
Rounsley SD Glodek A Sutton G Adams MD Somerville CR Venter JC Kerlavage AR The construction of Arabidopsis expressed sequence tag assemblies. A new resource to facilitate gene identification Plant Physiol 1996 112 1177 1183 8938416 10.1104/pp.112.3.1177
Altschul SF Madden TL Schaffer AA Zhang J Zhang Z Miller W Lipman DJ Gapped BLAST and PSI-BLAST: A new generation of protein database search programs Nucl Acids Res 1997 25 3389 3402 9254694 10.1093/nar/25.17.3389
Coram T Pang E Ascochyta blight of chickpea – searching for resistance gene analogues and characterizing an enriched EST library Proceedings of the 12 th Australian Plant Breeding Conference, 15–20 September, Perth, Australia 2002 17 22
Beaulieu M Larson GP Geller L Flanagan SD Krontiris TG PCR candidate region mismatch scanning: adaption to quantitive, high-throughput genotyping Nucleic Acid Res 2001 29 1114 1124 11222761 10.1093/nar/29.5.1114
Cobb BD Clarkson JM A simple procedure for optimising the polymerase chain reaction (PCR) using modified Taguchi methods Nucl Acids Res 1994 22 3801 3805 7937094
Taguchi G Introduction to Quality Engineering, Asian Productivity Organization 1986 American Supplier Institute Inc., Dearborn, MI
Dreher K Khairallah M Ribaut JM Morris M Money matters (I): costs of field and laboratory procedures associated with conventional and marker-assisted maize breeding at CIMMYT Mol Breeding 2003 11 221 234 10.1023/A:1022820520673
Ragot M Hoisington DA Molecular markers for plant breeding: comparisons of RFLP and RAPD genotyping costs Theor Appl Genet 1993 86 975 984 10.1007/BF00211050
Mace ES Buhariwalla HK Crouch JH A high throughput DNA extraction protocol for tropical molecular breeding programs Plant Mol Biol Rep 2004 21 459a 459h
Kresovich S Luongo AJ Schloss SJ Engels JMM, Ramanatha Rao V, Brown AHD, Jackson MT Mining the Gold: Finding Allelic variants for improved crop conservation and use Managing Plant Genetic Diversity 2002 IPGRI 379 386
Kazan K Muehlbauer FJ Allozyme variation and phylogeny in annual species of Cicer (Leguminosae) Plant Syst Evol 1991 175 11 21 10.1007/BF00942142
Ahmad F Random amplified polymorphic DNA (RAPD) analysis reveals genetic relationships among the annual Cicer species Theor Appl Genet 1999 98 657 663 10.1007/s001220051117
Choumane W Winter P Weigand F Kahl G Conservation and variability of sequence-tagged microsatellite sites (STMSs) from chickpea (Cicer aerietinum L.) within the genus Cicer Theor Appl Genet 2000 101 269 278 10.1007/s001220051479
Sudupak M Inter and Intra-species Inter Simple sequence Repeat (ISSR) variation in the genus Cicer Euphytica 2004 135 229 238 10.1023/B:EUPH.0000014938.02019.f3
Ladizinsky G Adler A Genetic relationships among the annual species of Cicer L Theor Appl Genet 1976 48 197 203 10.1007/BF00527371
Ahmad F Slinkard AE Genetic relationships in the genus Cicer L. as revealed by polyacrylamide gel electrophoresis of seed storage proteins Theor Appl Genet 1992 84 688 692 10.1007/BF00224169
Buhariwalla HK Gaur PM Pande S Crouch JH Asian Development Bank Half-yearly progress report in chickpea for project 'Rapid crop improvement for poor farmers in the semi-arid tropics' 2002 ICRISAT, India
Gargav S Gaur PM Phylogenetic relationships of perennial chickpea species Cicer microphylhum (Benth.) with its annual relatives as revealed by allozyme polymorphism Indian Journal of Genetics 2001 61 218 222
Heslop-Harrison JS Exploiting novel germplasm Aust J Agric Res 2002 53 873 879 10.1071/AR02078
RepeatMasker software
TIGR BLASTn and tBLASTx analysis
NCBI BLASTn analysis:
Transeq software:
Giegerich R Meyer F Schleiermacher C GeneFisher – Software Support for the Detection of Postulated Genes Proceedings of the Fourth International Conference on Intelligent Systems for Molecular Biology 1996 AAAI Press, USA 68 77
Netprimer analysis
Tegelstrom H Hoelzel AR Detection of mitochondrial DNA fragments in molecular genetic analysis of populations A Practical Approach 1992 IRL Press, Oxford 89 114
Polymorphic information content calculator
Rohlf FJ NTSYS-pc Numerical taxonomy and multivariate analysis system, Version 20 1998 Applied Biostatistics, New York
Lynch M The similarity index and DNA fingerprinting Mol Biol Evol 1990 7 478 484 2263197
Nei M Li WH Mathematical model for studying genetic variation in terms of restriction endonucleases Proc Natl Acad Sci 1979 74 5267 5273
Felsenstein J PHYLIP (Phylogeny Inference Package), version 363 1995 the author, Univ. of Washington, Seattle, Washington
|
16107212
|
PMC1215491
|
CC BY
|
2021-01-04 16:03:52
|
no
|
BMC Plant Biol. 2005 Aug 17; 5:16
|
utf-8
|
BMC Plant Biol
| 2,005 |
10.1186/1471-2229-5-16
|
oa_comm
|
==== Front
BMC PsychiatryBMC Psychiatry1471-244XBioMed Central London 1471-244X-5-321612022810.1186/1471-244X-5-32Research ArticleThe characteristics of suicides within a week of discharge after psychiatric hospitalisation – a nationwide register study Pirkola Sami [email protected] Britta [email protected] Kristian [email protected] Mental Health Group, Health and Social Services Division, National Research and Development Centre for Welfare and Health (STAKES), Lintulahdenkuja 4, FIN-00530, Helsinki Finland2 Department of Mental Health and Alcohol Research, National Public Health Institute (KTL), Helsinki Finland3 Vaasa Central Hospital, Vaasa, Finland2005 25 8 2005 5 32 32 29 4 2005 25 8 2005 Copyright © 2005 Pirkola et al; licensee BioMed Central Ltd.2005Pirkola et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
The characteristics of victims of immediate post-discharge suicides are not well known. We explored these characteristics for the purposes of better recognition and preventive efforts of potential immediate post-discharge suicides.
Methods
Suicides from a Finnish nationwide register were linked with preceding periods of psychiatric inpatient treatment. Characteristics of suicides within a week of discharge were compared to those occurring later after discharge.
Results
Compared to other previously hospitalised suicide victims, those committing suicide within a week of discharge were more often female, unmarried, had a higher grade of education and a diagnosis of schizophrenia spectrum or affective disorder, tended to use more drowning and jumping from heights as the methods for suicide and had gained a smaller improvement in psychological functioning during hospitalization.
Conclusion
These characteristics indicate a more severe psychopathology, relatively poorer level of functioning, less global response to hospitalisation, and a more frequent choice of lethal and easily available method for suicide. Potentially suicidal psychiatric patients should be better recognized and an immediate follow-up arranged if it is decided they be discharged.
==== Body
Background
A psychiatric illness that necessitates hospitalisation is one of the strongest risk factors for suicide [1]. Although the specific risk factors of psychiatric patients for completed suicide are not well established, the time period after discharge from a psychiatric hospital is known to be a high risk time for attempted and completed suicide for up to one year [2-8]. A clustering of suicides within the first month and especially the first week after discharge seems evident [4,9]. The factors associated with immediate post-discharge suicides are not well known, although it has been suggested that young adults, females, and those with affective disorders or short hospitalizations are at higher risk [9,10]. Identifying subjects who carry a risk for immediate post-discharge suicide is particularly important given the current tendency to further shorten psychiatric hospitalizations. In terms of suicide prevention it is possible that identification of an obvious suicide risk and consequent actions may lead to prevention of suicide at least in individual cases. Postponing a potentially lethal suicide attempt ("winning time") may offer opportunities for effective treatment and suicide prevention. In addition to the relatively unspecific risk factors reported for immediate post-discharge suicide [10], more specific characteristic can be distinguished by exploring these suicides within the population of all previously hospitalized suicide completers.
Within the MERTTU project on the effectiveness of mental health services, we set out to investigate the nationwide pattern of post-discharge suicides and factors associated with suicides in the period subsequent to inpatient care. In a comprehensive, register-based database we had data on all completed suicides between 1980 and 2001 and any psychiatric hospitalisations that preceded their death. We aimed at characterising the victims of immediate (within one week) post-discharge suicides compared to victims who committed suicide later than the first week who had also had some previous psychiatric hospitalisation.
Methods
Registers
We collected all suicides (N = 22717) during the years 1980–2001 in Finland from the National Cause of Death Register maintained by Statistics Finland. The personal identification codes of these subjects were linked to the Finnish Hospital Discharge Register (FHDR) and the Finnish Health Care Register (FHCR). We collected data on the psychiatric hospitalisations preceding suicide and discharge diagnosis. Furthermore, we collected details on involuntary treatment acts during the last hospitalisation, which were available from 1995 onwards. Sociodemographic variables were recorded from the registers of Statistics Finland. These included data on total years of education categorised into three groups (primary- or lower secondary; upper- or post secondary but non-tertiary; and tertiary or higher education) and) and other classified variables about marital status and occupation -based socioeconomical status. Level of functioning was assessed by the Global Assessment Scale (GAS), which has been registered from 1995 onwards both at hospital intake and at discharge.
Discharge diagnoses
During the follow-up period, the official diagnostic classification has changed twice: from ICD-8 to ICD-9 in 1982 and from ICD-9 to ICD-10 in 1996. The primary discharge diagnoses from psychiatric treatment periods were converted to current ICD-10 codes for metacategories of substance use disorders (F1*), schizophrenia spectrum disorders (F2*), affective disorders (F3*), stress- and anxiety-related and somatoform disorders (F4*), disorders related to physiology and bodily functions – for instance eating disorders (F5*), personality disorders (F6*) and other disorders – for instance developmental disorders or syndromes of organic origin (F0*, F7*–F9*).
Statistical methods
In analyzing the data, basic statistical tests were used for the bivariate analyses: chi-a square test and a two-tailed t-test. An age- and sex-adjusted logistic regression model was used to estimate the significance of individual factors in predicting a suicide within a week of discharge from psychiatric inpatient care. The SPSS (version 11.5) statistical package was used for the analyses.
Results
A proportion of 6% (1407/22717) of all suicide victims had died within a week of being discharged after a psychiatric hospitalisation. Compared to other previously hospitalised suicide victims, those committing suicide within a week of discharge were more often female, unmarried, and more likely to have used drowning, jumping or hanging as suicide methods (Table 1). They suffered more often from schizophrenia spectrum or affective disorders, and less often from substance-related disorders. They had also more often and for longer periods been in involuntary care according to the Mental Health Act during the last hospital period (mean of sum 25.7 days vs. 11.9 days, independent samples t-test, F = 40,59, p < 0.001). No differences were found in the frequency of individual coercive treatment acts, including injected medication, restrictions or constraint.
Table 1 The characteristics of suicides carried out within a week of discharge compared to other suicides with previous hospitalisations
Suicide later % Suicide within a week from discharge % All % Age- and sex adjusted logistic model
N = 6689 N = 1407 N = 8096 Exp(B) 95,0% C.I.
Age, y * 41.5 43.6
Male 70.4 63.3 69.2 ref
Female * 29.6 36.7 30.9 2.30 2.10 – 2.58
Marital status *
Unmarried 43.2 47.1 43.9 1.29 1.12–1.49
Widow 5.4 5.8 5.5 0.90 0.70 – 1.17
Separated 1.0 0.7 0.9 1.08 0.56–2.06
Divorced 23.6 16.3 22.4 1.00 0.85 – 1.18
Married 26.8 30.1 27.3 ref
Education *
High grade 12.5 16.9 13.3 1.57 1.34–1.83
Middle grade 36.5 34.5 36.2 1.07 0.94 – 1.21
Low grade 50.9 48.6 50.5 ref
Socioeconomic status *
Upper employee 4.1 5.8 4.4 1.78 1.37 – 2.33
Lower employee 8.8 11.9 9.3 1.51 1.22 – 1.86
Entrepreneur 17.0 15.4 16.7 1.25 0.95 – 1.63
Worker 4.9 5.5 5.0 0.91 0.75 – 1.10
Student 5.1 7.4 5.5 1.24 0.97 – 1.59
Retired 37.5 38.6 37.7 2.54 2.11 – 3.06
Other or undetermined 22.6 15.4 21.4 ref
Suicide method *
Intoxication, any substance 33.7 17.3 30.9 0.89 0.73 – 1.08
Hanging or other suffocation 27.5 34.3 28.7 1.44 1.21 – 1.71
Drowning 6.6 16.4 8.3 3.38 2.75 – 4.16
Shooting or exploding 10.2 6.6 9.6 0.40 0.31 – 0.51
Jumping from heights 5.4 11.2 6.4 3.29 2.63 – 4.10
Other 16.6 14.2 16.2 ref
Discharge diagnosis *
Substance-related disorders 19.2 3.7 16.5 0.3 0.22 – 0.47
Schizophrenia and similar psychoses 25.4 37.2 27.4 2.3 1.67 – 3.07
Affective disorders 33.1 45.6 35.2 2.3 1.68 – 3.08
Neurotic-, stress-related and somatoform disorders 7.3 4.1 6.7 0.9 0.58 – 1.28
Personality disorders 9.2 5.7 8.6 0.9 0.63 – 1.33
Other 6.0 .7 5.6 ref
* = in univariate testing, all significant at level p < 0.001
Typical for immediate post-discharge suicides was a more modest improvement between arrival and discharge in functional status as measured by GAS scores (3 vs. 16, t-test for means, t = 16.63, two-tailed p < 0.001), as well as a worse functional status (42 vs. 57, t-test for means, t = 18.97, two-tailed p < 0.001), though this information is only available among the more recent cases.
Discussion
In this register-based study we collected a comprehensive dataset covering all suicides in Finland during 1980–2001. On the basis of the available information on previously hospitalised victims, we found that subjects committing suicide soon after discharge from hospital treatment for psychiatric disorders differed from later suicides of previously hospitalised patients in more often being female, having more often received treatment for a schizophrenia spectrum or affective disorder and less often for a substance-related disorder. They had more often used suicide methods of easier availability (particularly drowning and jumping from heights), had more often been an employee, and had more often had a higher grade of education. Their psychological functioning improved less during the last hospital period than the functioning of subjects who committed suicide later. They were more often involuntarily treated, and they also had worse GAS scores at discharge. These findings from a comprehensive nationwide suicide population help in efforts to characterise psychiatric inpatients at risk for immediate suicide after discharge, and in adding to our understanding of the role of their hospitalisation and post-hospital follow-up. These individuals may represent patients whose discharges should be particularly well-planned and monitored.
The distinctive characteristics we found are not specific for suicides in general and they do not represent suicide risk factors. They rather help to identify a special population comprising a total of 6% of all suicides, a part of which we believe, could be prevented by alertness in mental health in-patient services. A better recognition of risk and prevention of immediate post-discharge suicides may act towards winning time for appropriate management of effective care. It may be that a final set of risk factors at the time of immediate post-discharge suicide are no longer valid when sufficient time has passed. In this regard, a successful recognition of risk among this special population offers a means for effective suicide prevention in a portion of potential suicide attempters. For instance, a portion of the immediate post-discharge suicide victims may have suffered from a relatively fast decline in their psychiatric and psychosocial condition. This disruption may have gone unnoticed and a relatively premature discharge has occurred. In these cases, a longer treatment period and the management of proper aftercare, including family support, might have been preventive for suicide [10].
Discharge diagnoses
Our finding that in a nationwide sample, schizophrenia spectrum- and affective disorders carried an elevated risk for suicide soon after discharge is somewhat discordant with Ho [9], who in a record-linkage follow-up study found that among psychiatric patients, no particular diagnosis seems to carry a specific risk for immediate (1–28 days) post-discharge suicide. The lack of statistically significant differences in suicide risk between diagnostic groups may be explained by the fact that the analysis by Ho (2003) included only 280 suicides, which is considerably less than the 1407 suicides in the current study.
King et al.[7] reported within a selected case-control setting study that affective- and schizophrenia-like disorders are the most frequent diagnoses among inpatient- and discharged patient suicides. In line with our findings, the majority of in-patient suicides are reportedly diagnosed with a current or previous affective disorder or schizophrenia [11-13]. Particular alertness and a focusing on immediate follow-up when discharging patients in these diagnostic groups seems justified. An interesting diagnostic finding was also the relative infrequency of substance-related discharge diagnoses. It seems that the triggers and timing for suicide manifest differently among the victims with primarily substance-related disorders. Particular challenges in their treatment may include the assessment of an appropriate outpatient setting in the long run.
The suicide methods
The overrepresented methods in suicides within the first week of discharge (drowning and jumping from heights) are of a more serious lethality and relatively easier availability than the other methods (shooting or intoxication by any substance). Drowning (6.9%) and jumping from heights appear relatively uncommon suicide methods in general [14], suggesting that victims of post-discharge suicides have suffered from a particular impulsivity or lability. It may be that some of the immediate post-discharge suicides have occurred without preceding preparations or planning, but rather in a state of impulsive mood, anxiety or psychotic disturbance. In these cases, discharge may have been premature and follow-up arrangements in community care insufficient. The continuity of treatment contacts has been suggested as of importance in efforts to reduce post-discharge suicides [10,15]. Our findings regarding the lower level of functioning at discharge and poorer functional improvement during hospitalisation indicate that victims of immediate post-discharge suicides may have been discharged earlier than their clinical status would have allowed.
Sociodemographic factors
Victims of suicides soon after discharge had certain sociodemographic characteristics. In addition to being more often female, they tended to have a relatively better sociodemographic status in terms of profession and education, and they were slightly more often married (in addition to being unmarried) rather than divorced. It remains speculative as to whether their suicidal process included a more recent clinical change and concomitant psychosocial disadvantage or disruption, similar to what has been reported among alcohol-misusing suicide victims [15]. If so, this again should alert us to the possibility of a post-discharge suicide.
Methodological considerations
Our unselected population-based suicide victims do not result in selection bias and are totally representative of the hospitalised psychiatric patients in this respect. However, certain limitations arise from the fact that the Finnish Health Care Register includes data from all hospital treatments in Finland, but the data collection is limited to details of the treatment period. Therefore personal history, as well as any outpatient treatment data, is beyond the reach of this study. Evaluating the effectiveness of clinical practices, including psychosocial management and medication, needs to be studied more in clinical settings.
In the current study we were not able to use a control group consisting of post-discharge survivors. Therefore, we are basically describing the characteristics of possibly prematurely discharged psychiatric patients who have died by suicide. We do assume that suicides occurring later after discharge are affected more by a variety of other risk factors that may be more effectively identified and prevented in outpatient settings.
Conclusion
Our findings indicate that in retrospective, suicides soon after discharge after a psychiatric hospitalisation have some typical characteristics that indicate a more severe psychopathology, a lower level of functioning, and a preferential choice of more lethal and easily available methods for suicide. These suggest the possibility of better recognition during treatment, and for preventive efforts in selected populations. With regard to suicide prevention, there is a need for a better recognition of suicidal risk among psychiatric patients during a period of decreased total use of psychiatric hospital treatment. Most likely, immediate follow-treatment for discharged patients is needed.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
All authors have made a substantive intellectual contribution to this study and participated in all stages of this work, including the design of the study. In addition, SP drafted the manuscript and performed the statistical analyses. BS participated the statistical designing and interpretation of the data, and revised the text. KW participated in conceiving the study, participated in its coordination and critically revised the text. All authors have read and approved the final manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
This study was supported by the Academy of Finland grant No. 203742 (the MERTTU Project)
==== Refs
Mortensen PB Agerbo E Erikson T Qin P Westergaard-Nielsen N Psychiatric illness and risk factors for suicide in Denmark Lancet 2000 355 9 12 10615884 10.1016/S0140-6736(99)06376-X
Pirkis J Burgess P Jolley D Suicide among psychiatric patients: a case-control study Aust N Z J Psychiatry 2002 36 86 91 11929443 10.1046/j.1440-1614.2002.00993.x
Goldacre M Hawton K Repetition of self-poisoning and subsequent death in adolescents who take overdoses Br J Psychiatry 1985 146 395 398 4016442
Goldacre M Seagroatt V Hawton K Suicide after discharge from psychiatric inpatient care Lancet 1993 342 283 286 8101307 10.1016/0140-6736(93)91822-4
Geddes JR Juszczak E O'Brien F Kendrick S Suicide in the 12 months after discharge from psychiatric inpatient care, Scotland 1968-92 J Epidemiol Community Health 1997 51 430 434 9328552
Geddes JR Juszczak E Period trends in rate of suicide in first 28 days after discharge from psychiatric hospital in Scotland, 1968-92.[see comment] BMJ 1995 311 357 360 7640540
King EA Baldwin DS Sinclair JM Campbell MJ The Wessex Recent In-Patient Suicide Study, 2. Case-control study of 59 in-patient suicides Br J Psychiatry 2001 178 537 542 11388970 10.1192/bjp.178.6.537
King EA Baldwin DS Sinclair JM Baker NG Campbell MJ Thompson C The Wessex Recent In-Patient Suicide Study, 1. Case-control study of 234 recently discharged psychiatric patient suicides Br J Psychiatry 2001 178 531 536 11388969 10.1192/bjp.178.6.531
Ho TP The suicide risk of discharged psychiatric patients J Clin Psychiatry 2003 64 702 707 12823086
Qin P Nordentoft M Suicide risk in relation to psychiatric hospitalization: evidence based on longitudinal registers Arch Gen Psychiatry 2005 62 427 432 15809410 10.1001/archpsyc.62.4.427
Roy A Draper R Suicide among psychiatric hospital in-patients Psychol Med 1995 25 199 202 7792356
Deisenhammer EA DeCol C Honeder M Hinterhuber H Fleischhacker WW In-patient suicide in psychiatric hospitals Acta Psychiatr Scand 2000 102 290 294 11089729 10.1034/j.1600-0447.2000.102004290.x
Powell J Geddes J Deeks J Goldacre M Hawton K Suicide in psychiatric hospital in-patients. Risk factors and their predictive power Br J Psychiatry 2000 176 266 272 10755075 10.1192/bjp.176.3.266
Pirkola S Isometsa E Lonnqvist J Do means matter?: differences in characteristics of finnish suicide completers using different methods J Nerv Ment Dis 2003 191 745 750 14614342
Pirkola SP Isometsa ET Heikkinen ME Lonnqvist JK Suicides of alcohol misusers and non-misusers in a nationwide population Alcohol Alcohol 2000 35 70 75 10684781
|
16120228
|
PMC1215492
|
CC BY
|
2021-01-04 16:33:03
|
no
|
BMC Psychiatry. 2005 Aug 25; 5:32
|
utf-8
|
BMC Psychiatry
| 2,005 |
10.1186/1471-244X-5-32
|
oa_comm
|
==== Front
Biomed Eng OnlineBioMedical Engineering OnLine1475-925XBioMed Central London 1475-925X-4-481609553210.1186/1475-925X-4-48ResearchAlgorithm for identifying and separating beats from arterial pulse records Treo Ernesto F [email protected] Myriam C [email protected] Max E [email protected] Departamento de Bioingeniería, Instituto Superior de Investigaciones Biológicas (INSIBIO), Consejo Nacional de Investigaciones Científicas y Técnicas (CONICET), Universidad Nacional de Tucumán (UNT), CC327, 4000, Tucumán, Argentina2 Departamento de Bioingeniería, Instituto Superior de Investigaciones Biológicas (INSIBIO), Consejo Nacional de Investigaciones Científicas y Técnicas (CONICET), Universidad Nacional de Tucumán (UNT), CC327, 4000, Tucumán, Argentina. Also with Facultad de Ciencias Exactas y Tecnología (FACET), Universidad Nacional de Tucumán (UNT), Av. Independencia 1800, 4000, Tucumán, Argentina3 Departamento de Bioingeniería, Instituto Superior de Investigaciones Biológicas (INSIBIO), Consejo Nacional de Investigaciones Científicas y Técnicas (CONICET), Universidad Nacional de Tucumán (UNT), CC327, 4000, Tucumán, Argentina2005 11 8 2005 4 48 48 2 3 2005 11 8 2005 Copyright © 2005 Treo et al; licensee BioMed Central Ltd.2005Treo et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
This project was designed as an epidemiological aid-selecting tool for a small country health center with the general objective of screening out possible coronary patients. Peripheral artery function can be non-invasively evaluated by impedance plethysmography. Changes in these vessels appear as good predictors of future coronary behavior. Impedance plethysmography detects volume variations after simple occlusive maneuvers that may show indicative modifications in arterial/venous responses. Averaging of a series of pulses is needed and this, in turn, requires proper determination of the beginning and end of each beat. Thus, the objective here is to describe an algorithm to identify and separate out beats from a plethysmographic record. A secondary objective was to compare the output given by human operators against the algorithm.
Methods
The identification algorithm detected the beat's onset and end on the basis of the maximum rising phase, the choice of possible ventricular systolic starting points considering cardiac frequency, and the adjustment of some tolerance values to optimize the behavior. Out of 800 patients in the study, 40 occlusive records (supradiastolic- subsystolic) were randomly selected without any preliminary diagnosis. Radial impedance plethysmographic pulse and standard ECG were recorded digitizing and storing the data. Cardiac frequency was estimated with the Power Density Function and, thereafter, the signal was derived twice, followed by binarization of the first derivative and rectification of the second derivative. The product of the two latter results led to a weighing signal from which the cycles' onsets and ends were established. Weighed and frequency filters are needed along with the pre-establishment of their respective tolerances. Out of the 40 records, 30 seconds strands were randomly chosen to be analyzed by the algorithm and by two operators. Sensitivity and accuracy were calculated by means of the true/false and positive/negative criteria. Synchronization ability was measured through the coefficient of variation and the median value of correlation for each patient. These parameters were assessed by means of Friedman's ANOVA and Kendall Concordance test.
Results
Sensitivity was 97% and 91% for the two operators, respectively, while accuracy was cero for both of them. The synchronism variability analysis was significant (p < 0.01) for the two statistics, showing that the algorithm produced the best result.
Conclusion
The proposed algorithm showed good performance as expressed by its high sensitivity. The correlation analysis demonstrated that, from the synchronism point of view, the algorithm performed the best detection. Patients with marked arrhythmic processes are not good candidates for this kind of analysis. At most, they would be singled out by the algorithm and, thereafter, to be checked by an operator.
second derivative beat detectionlimb impedance plethysmographypatient screeningpreventive medicine
==== Body
Background
Outpatients coming daily for consultation to a general public hospital are often preventively checked for signs suggestive of infectious, cardiovascular and/or any other endemic disease. The positive detected fraction is derived for further confirmatory study, which may lead to eventual treatment. Within such concept, this project was specifically designed as an epidemiological aid-selecting tool for a small country health center serving a large rural area (see Acknowledgments). Essential requirements were low cost and simplicity. The general objective was to screen out possible coronary patients.
Peripheral artery function can be non-invasively evaluated by impedance plethysmography, either in lower or upper limbs [1]. Changes in these vessels appear as good predictors of future coronary behavior [2,3]. Basically, impedance plethysmography detects volume variations due to the pulsating blood flow that, after simple mechanical occlusive maneuvers, may show indicative modifications in arterial/venous responses [4,5].
Pulse plethysmographic analysis, based on variations of its amplitude or waveform [6], requires the averaging of several beats.
There are specific algorithms for the detection of the dicrotic notch [7]; some papers make a beat-to-beat analysis of the arterial pressure [8-11]. Commercial equipment (like Complior ®SP, Artech Medical, y SphygmoCor ® Vx, Atcor Medical, ) carry out the above mentioned type of plethysmographic signal analysis. Schroeder et al [12], by means of MATLAB, developed a cardiovascular package (named HEART), which permits beat identification using two sequential processes (one of coarse approximation and a second one of fine adjustment). Unfortunately, none of these procedures offer detailed descriptions.
Besides, several algorithms have been developed to detect electrocardiographic beats, by and large based on the recognition of the QRS complex [13]. However, our design has been thought to operate independently of the ECG signal; for these reasons they are not applicable in this case.
Thus, the objective here is to describe an algorithm to identify and separate out beats from a plethysmographic record during an occlusive maneuver. As a secondary objective, we intended to compare the output given by human operators (trained and not trained) against the algorithm. The method herein proposed is potentially applicable to other cardiac signals.
Methods
Outline
An occluding cuff produces a limb short ischemia. Basal and post-occlusion plethysmographic arterial pulse records are compared searching for either amplitude and/or waveform modifications or both. Since the possible change in a single beat record does not supply enough information, valid results call for averaging of a series of pulses and this, in turn, requires proper determination of the beginning and end of each beat. In other words, good beats must be identified and singled out discarding abnormal pulses.
Each applied maneuver consists of a supradiastolic and subsystolic occlusion; thus, it permits inflow of blood while stops outflow, which leads to limb volume increase. A typical record is showed in Figure 1. The pulse identification algorithm detected the beat's onset and end on the basis of the maximum rising phase, the choice of possible ventricular systolic starting points considering cardiac frequency, and the adjustment of some tolerance values to optimize the behavior.
Figure 1 Typical dc plethysmographic record of an occlusive maneuver. Insets show expanded pulses during each stage; differences are evident. An arbitrary zero time is also showed from which the pre-inflation time was measured (PRE); thereafter, the intra-occlusion (INTRA) period followed (bounded by the beginning and end of inflation), until post-occlusion was reached (POST).
Patient population
From the daily inflow of hospital outpatients, we obtained 800 records out of which 40 supradiastolic-subsystolic occlusive ones were randomly selected without any preliminary diagnosis. All patients accepted and signed the informed consent. The attending physician, including the measurements leading to the quantitative data mentioned above, carried out routine clinical interrogation. Blood pressure was obtained with the oscillometric method using the contra-lateral arm to that where the test was to be performed. Patients rested for at least 5 min in the supine position prior to the test.
Impedance Plethysmography and Recording System
Radial pulse was picked up with two metallic electrodes (ECG standard type) placed over the forearm artery line, 2 cm below the ante-cubital fold, and 5 to 10 cm apart. The forearm was always at the left atrial level. Besides, a simultaneous standard ECG was obtained. Impedance was obtained with a custom-made laboratory apparatus [14].
Digital acquisition (sampling frequency sf = 200 Hz, at 16 bits) was carried out using a commercial system (BIOPAC System Inc, AcqKnowledge II for MP100WSW). Each occlusive maneuver record included basal plethysmographic pulses (PRE), a period of 2 to 3 minutes of occlusive cuff inflation (INTRA), and a post-occlusive (POST) after release; the overall duration was always in the order of 5–6 min (Figure 1).
Algorithm
Figure 2 summarizes the sequential steps the recorded signal S went through. The first step is the detection of the average cardiac frequency fc, which is expected to be within the 0.5 Hz – 2.5 Hz range, and is divided in two stages: First, the signal goes through a band-pass FIR (Finite Impulsive Response) filter, with cut-off frequencies of 0.8 Hz and 2.8 Hz, and attenuation at least of -50dB at 0.25 Hz and 3.25 Hz. Thereafter, the Power Spectral Density function (PSD, an averaging variant of the Fast Fourier Transform), is applied to find fc. The first minute of data acquisition corresponds to the basal or pre-occlusion stage (Fig. 1). This period is divided into overlapping sections; each is linearly detrended, then windowed with a Hanning function (4,096 samples) and, thereafter, zero-padded to a length of 8,192 samples. The magnitude squared of the Fast Fourier Transform of each section is averaged to form Pxx, the Spectral Density Function. Each section overlaps with the previous 2000 samples. The maximum value for the first harmonic component corresponds to the cardiac frequency fc (see Signal Processing Toolbox for Use with MATLAB, The MathWorks, Inc, Natick, MA).
Figure 2 Flow Diagram of the Algorithm.
The original signal is derived twice applying the well-known iterative series of subtractions [15], that is,
S'(n) = S(n + 1) - S(n)
S"(n) = S'(n + 1) - S'(n) [1]
where S'(n) and S"(n) stand, respectively, for the first and second derivative and n represents the sample number. Figure 3, upper trace, shows 6 beats of a typical plethysmographic record. Thereafter, both signals are processed binarizing the first derivative and rectifying the second derivative. Binarization means to replace 1's for positive values and 0's for negative ones. The process of rectification leaves only the positive excursions. Clearly, the product of the binarized and rectified signals is a trace showing large peaks and some significantly lower ones in between, which is called the weighing signal W1. When the latter product is compared with the original S signal, one can easily see that the large peaks are obviously coincident with true good beats while the small spikes correspond to other changes (not pulses) in that signal. A spike detection routine based on the first derivative sign change identifies all maxima, including both the large and the small spikes; thus, filtering is required to remove the small spikes and preserve the large ones.
Figure 3 Different stages while processing a typical signal. S is the original signal; S'b is the binarized first derivative; S''r is the rectified second derivative; W1 is the weighing signal showing all spikes detected by marks; W2 is the same weighing signal, with only those spikes preserved by the weighed filter. W3 is the same weighing signal with only those spikes preserved by the frequency filter. The bottom trace represents again S, with all onsets and ends well identified. Tolerance values were Tol1 = 0.4 and Tol2 = 0.2.
Weighed Filter
The sequence of maximum spikes after multiplication (Figs. 2 and 3) can be used as a filtering criterion to separate out the true beats. Whenever the interval between two pulses is much smaller than the cardiac period (for example, less than one half), it can be assumed that the two spikes are too close together and cannot represent a beginning (and ending) of a whole cycle. Consequently, one must be removed.
Since, by and large, the beginning of a cycle corresponds to a steep rise time, the second derivative has more weight, and the multiplication result clearly indicates to retain that particular spike (Fig. 3, trace W2).
Mathematically, this is treated as
t(Si+1) - t(Si) <Tol1 × Tc [2]
where t(Si+1) and t(Si) are, respectively, the time of appearance of spikes i+1 and i in signal W1, Tol1 is a preset tolerance value and Tc = 1/fc stands for the cardiac period expressed in seconds. Each pair of adjacent spikes is analyzed and, if the comparison result is true, the smallest is removed and the new pair of contiguous spikes is now chosen. If the result is false, i is incremented. The process repeats until no true result is obtained. Adjusting the tolerance value Tol1, the number of removed spikes can be increased or decreased.
Frequency Filter
Once all small spikes have been removed, the remaining spikes must be analyzed to check if they correspond to the beginning and end of a cycle. Figure 3 (fifth trace W2) shows the peaks remaining after the weighed filter. Spike 4 is a misdetection that must be removed. A second filter matches pairs of peaks (not necessarily consecutive) checking whether they correspond to a beat limits or not; the distance between them should be fixed between Tc ± Tol2, where the latter is a second tolerance value, generally chosen close to 20%.
Starting from the first detected spike 1 at time t (Fig. 3, fifth trace), and assuming it corresponds to the beginning of a cycle, a second spike should be located within the interval t + Tc ± 20%. If this second spike exists, the time corresponding to both spikes is stored as the limits of a cardiac cycle. In fact, this second spike exists in Figure 3, marked as 2. The process is repeated starting now from 2 and so on. Now, let us consider that the first spike detected was 4. When searching its partner spike ahead, the algorithm will not find it because 5 and 6 are, respectively, too close and too far from 4. In this case, 4 is discarded and the process continues to the next one.
Cardiac period has been assumed constant up to now, however, it is known to be modulated by the respiratory heart rate response. To have a better estimation of Tc, each time two spikes are found to be (Tc ± Tol2) seconds apart, their difference is used to update a new value of Tc to be applied in the following calculations.
Statistics
For each of the 40 patients, a 30 s trace was chosen at random, which was analyzed by two operators. One of them (operator 1) was trained and familiar with the procedure and another (operator 2) without any previous training. Both operators received the same instructions regarding the analysis to be performed. Each operator marked manually the beginning and end of each beat as the 30 s sample was presented on the monitor. The selection criterion was to identify that point previous to the rapid rising of the ejective period, not necessarily coincident with the previous minimum. In this way, there were two marks that clearly bounded each positive cycle. When the beat limits were not clearly defined or the signal was lost due to circuit saturation, the portion between the last observed beat and the following beginning was classified as negative (i.e., rejected). Thereafter, the algorithm was applied and coincidences with the operators' results were searched.
Since an exact coincidence is almost nil, we adopted a threshold level to specify the maximum difference to be accepted between the limits marked by the operators and the algorithm. A program was developed to determine the beginning and ending points closest to those selected by the operator (Fig. 4). When the addition of the differences (Δt1 + Δt2) was lower than 10% of the cardiac period, the beat was considered as true positive (tp). If the difference was larger, the beat was classified as false negative (fn). However, when the operator could not bound the beat there is an undetermined signal length, which must not be processed. If coincidently the algorithm did not classify any beat within that same length, the selection is marked as true negative (tn). Instead, if the algorithm picked up at least one beat, the selection is considered as false positive (fp).
Figure 4 A single beat indicating the maximum error admitted for the statistical analysis.
Sensitivity of the procedure was defined as the percentage of beats correctly selected by the algorithm with respect to the total number of beats marked as true by the operator, that is,
s [%] = tp /(tp + fn) × 100 [3]
Accuracy, instead, was defined as the total number of sections correctly rejected by the algorithm with respect to the total number of sections discarded by the operator,
a [%] = tn /(tn + fp) × 100 [4]
Bounding of the beats is also important for the correct synchronization of the averaging procedure. Thus, those beats correctly classified by all three methods (operators 1 and 2 and the algorithm) were selected to compare the synchronization ability. For that matter, the time between the beginning and the first maximum coincident with ventricular ejection was measured for each beat. This time was, of course, different for the operators and the algorithm, each with a specific coefficient of variation. The latter was taken as the statistical estimator.
Moreover, for each patient a correlation analysis was carried out between all possible combinations of the tp beats. For each pair of beats a correlation factor was obtained, thus producing a non-normal distribution when all combinations are considered, which is usually characterized by the median value. In the end, we obtained three of these values for each patient according to the classification methods (two operators and algorithm).
The coefficient of variation (also with a non-Gaussian distribution) and the median should be analyzed by non-parametric techniques. In our case, we used Friedman's ANOVA and Kendall Concordance.
Results
Figure 5 shows three examples of signals and their bounding obtained by the algorithm. Its last section D belongs to an arrhythmic patient.
Figure 5 Several signals and the results of the algorithmic process. Vertical lines mark the beginning (dotted) and end (dashed) of each cycle. (A) Patient with Parkinson disease; (B) Deflation of the cuff; (C) Inflation of the cuff. The two latter in the same patient. Section D shows the ECG and the plethysmographic signal from a patient with cardiac arrhythmia.
The sensitivity was 97% and 91% for operators 1 and 2, respectively. Patients with low sensitivity were retrospectively analyzed. Only one patient produced a low sensitivity according to the criteria of both operators; after careful analysis, we found large heart rate variability mainly due to ectopic beats (Fig. 5D, Table 1).
Table 1 Results of the statistical analysis for both operators.
Operator 1 Operator 2
POSITIVE NEGATIVE POSITIVE NEGATIVE
Algorithm POSITIVE 1403 (tp) 5 (fp) POSITIVE 1321 (tp) 7 (fp)
NEGATIVE 44 (fn) 0 (tn) NEGATIVE 124 (fn) 0 (tn)
Accuracy for both operators was 0 because traces marked as negative were somehow classified by the algorithm, thus, producing fp beats.
The synchronism variability analysis gave off significant values p (p < 0.01) for the two statistics. However, Kendall coefficients were 0.62 and 0.17, respectively, for the median correlation and the coefficient of variation. Such results suggest that synchronism is not the same for the three separation criteria. Figure 6 shows the typical overlapping of several beats along with their averaged result. This represents a good way of visualizing what could be the pattern of, say, a normal beat, which could serve as a comparison reference. The same beats were separated according to the criteria of the two operators and the algorithm. The right lower graph shows a histogram with the correlation coefficients of all beats. Very few produced low values while most of them are grouped rather close to 1; as expected, the algorithm produced the maximum of the three.
Figure 6 Superimposed separated beats (thin colored lines) and one averaged beat (thick black line), from a particular patient, as they were singled out by (A) Operator 1, (B) operator 2 and (C) the algorithm. Section (D) shows the histogram of correlation values for all beats.
The detection routine for the average cardiac frequency was a critical factor in the analysis of the algorithm; any failure in it can produce an error that would propagate to any subsequent processing. Thus, this parameter was checked by visual inspection of the 40 signals and their frequency spectra.
Discussion
This algorithm allows the averaging of non-invasively obtained arterial pulses for the evaluation of the vascular response to peripheral occlusive maneuvers employing only the plethysmographic signal. The ECG served as monitor of cardiac activity and was used to help the operators in their task. Since the algorithm was designed thinking of a possible commercial equipment based only on the plethysmographic signal, the ECG, cannot be included in the analysis.
The algorithm sensitivity depends on the operator and it was always higher than 90%. However, accuracy was always cero. None of the sections marked as negative by any operator was correctly rejected by the algorithm. Analysis of the fp beats showed that the portions discarded by the operators are, by and large, sections with large drifts mainly due to cuff inflation or by movement artifacts. An important problem lies on the fact that the algorithm, many times, picks up beats that are placed in the course of a drift. An operator would never select such a beat, as exemplified in Figure 7. The possible utility of such fp beats perhaps ought to be studied because they might still contain clinical information.
Figure 7 Typical case of a beat falsely classified as positive by the algorithm. Δ are the points selected by the operator and ▼ are those detected by the algorithm.
The algorithm is based on the analysis of the maximum systolic slope discarding pulses not consistently separated out from their respective previous beats or when the derivative value is too low. Usually, beat separation in blood pressure records is obtained by the minimum value previous to the dicrotic notch. Noise, however, may perturb this kind of determination. When pulses selected by this criterion are overlapped for their averaging, the systolic maximum does not temporally coincide in all beats and a small shift, unpredictable and unknown, shows up. Low amplitude noise, when present, tends to interfere with the temporal location of the valve opening point. The maximum second derivative criterion does not really represent valve opening but rather represents maximum rise during systole. In most of the patients, in our experience, the latter reference showed better periodicity and seems to be a better time reference when averaging is required. The second event seems to be less sensitive to interferences because signal growth during systole is larger than noise changes. This observation was supported by the variability analysis. Kendall coefficient indicates that correlation is a reliable statistic and its variability is similar in the three methods.
The algorithm, due to its philosophy of design, does not have to identify all beats, so conferring to it a practical characteristic, i.e., rather frequently, due to patient's movements, the system's electronics may saturate. In such case, the algorithm disregards the piece resuming the search after signal recovery. However, the two filtering criteria are based on the cardiac frequency and, when the latter is too variable (for example, due to arrhythmias) the sensitivity falls drastically (figure 5D).
The average cardiac frequency is not affected during the occlusive maneuvers [16,17], and for patients without rhythm alterations, a mean cardiac frequency assumptions appears as reasonable.
Tolerances, in turn, are useful to modify the algorithm's performance according with the prevalent conditions (noise, drift, saturation). However, sensitivity higher than 90% is enough when the recording time is long (say, 5–6 min or more). The tolerance values suggested here produced in our opinion the best results.
Interested investigators are encouraged to request the algorithm in order to test it using signals obtained from other sources. These authors would be happy to make it available.
Conclusion
The proposed algorithm showed good performance as expressed by its high sensitivity. The correlation analysis demonstrated that, from the synchronism point of view, the algorithm performed the best detection. Patients with marked arrhythmic processes are not good candidates for this kind of analysis. At most, these patients would be singled out by the algorithm to be checked by an operator.
Authors' contributions
These authors contributed equally to this work; the former developed the basic idea of the algorithm while the latter actually obtained the records in the hospital environment. The corresponding last author gave orientation, revised the data, and wrote the paper in its different stages. Design of the protocol was a team collaborative task.
Acknowledgements
We deeply thank the medical and paramedical staff of the General Lamadrid Hospital (City of Monteros, Province of Tucumán, Argentina) for their collaboration during the clinical tests reported here. We also thank Dr. Elena Bru for her useful contributions and discussions, especially in aspects regarding to the statistical analysis. Preliminary partial results were communicated to the III Latin American Congress of Biomedical Engineering and XIX Congresso Brasileiro de Engenharia Biomédica, João Pessoa (Brazil), September 22–25, 2004. This work was partially supported by the Consejo Nacional de Investigaciones Científicas y Técnicas (CONICET, PID#3134-700-88) and by the Consejo de Investigaciones de la Universidad Nacional de Tucumán (CIUNT, E349/2005).
==== Refs
Nyboer J Electrical Impedance Plethysmography 1970 2 Springfield, IL: Charles C Thomas
Franklyn SS Khan SA Wong ND Larson MG Levy D Is Pulse pressure useful in predicting risk for coronary heart disease? The Framingham Heart Study Circulation 1999 100 354 360 10421594
Pepine CL Drexler H Dzau VJ Endothelial Function in Cardiovascular Health and Disease 1997 New York: Landmark Programs, Inc
Chin-Dusting JP Cameron JD Dart AM Jennings GL Human forearm venous occlusion plethysmography. Methodology, presentation and analysis Clin Science (London) 1999 96 439 440
Race C Ceravolo R Notarangelo L Crescenzo A Ventura G Tamburrini O Perticone F Gnasso A Comparison of endothelial function evaluated by strain gauge plethysmography and brachial artery ultrasound Atherosclerosis 2001 158 53 59 11500174 10.1016/S0021-9150(01)00406-3
O'Rourke MF Pauca A Jiang X-J Pulse wave analysis Br J Clin Pharmacol 2001 51 507 522 11422010 10.1046/j.0306-5251.2001.01400.x
Hoeksel SAAP Jansen JRC Blom JA Schreuder JJ Detection of dicrotic notch in arterial pressure signals J Clin Monit 1997 13 309 316 9338845 10.1023/A:1007414906294
van Lieshout JJ Toska K van Lieshout EJ Eriksen M Walløe L Wesseling KH Beat-to-Beat Noninvasive Stroke Volume from Arterial Pressure and Doppler Ultrasound Eur J Appl Physiol 2003 90 131 137 12851826 10.1007/s00421-003-0901-8
Romano SM Lazzeri C Chiostri M Gensini GF Franchi F Beat-to-Beat Analysis of Pressure Wave Morphology for Pre-Symptomatic Detection of Orthostatic Intolerance During Head-Up Tilt J Am College Cardiol 2004 44 1891 1897 10.1016/j.jacc.2004.07.046
Papaioannou TG Stamatelopoulos KS Gialafos E Vlachopoulos C Karatzis E Nanas J Lekakis J Monitoring of Arterial Stiffness Indices by Applanation Tonometry and Pulse Wave Analysis: Reproducibility at Low Blood Pressures J Clin Monit 2004 18 137 144
Teng XF Zhang YT The effect of contacting force on photoplethysmographic signals Physiol Meas 2004 25 1323 1335 15535195 10.1088/0967-3334/25/5/020
Schroeder MJ Perreault B Ewert DL Koenig SC HEART: an Automated Beat-to-Beat Cardiovascular Analysis Package Using MATLAB Comp Biol & Med 2004 34 371 388 10.1016/S0010-4825(03)00087-8
Tsipouras MG Fotiadis DI Sideris D An Arrhythmia Classification System Based on the RR-Interval Signal Artif Intelligence Med 2005 33 237 250 10.1016/j.artmed.2004.03.007
Feldman G Herrera MC Técnica impedancimétrica: Variabilidad de la respuesta vascular ante apremio suprasistólico (Impedance plethysmography: Variability of vascular response to suprasystolic injury) Revista de la Federación Argentina de Cardiología 2003 32 254 258 in Spanish
Ambardar A Analog and Digital Signal Processing 1995 PWS Publishing Company
Baldassare D Amato M Palombo C Morizzo C Pustina L Sirtori CR Time Course of Forearm Arterial Compliance Changes During Reactive Hyperemia Am J Physiol (Heart Circ Physiol) 2001 281 H1093 H1103 11514275
Higashi Y Sasaki S Nakagawa K Matsuura H Kajyama G Oshima T A Noninvasive Measurement of Reactive Hyperemia that can be Used to Assess Resistance Artery Endothelial Function in Humans Am J Cardiol 2001 87 121 125 11137850 10.1016/S0002-9149(00)01288-1
|
16095532
|
PMC1215493
|
CC BY
|
2021-01-04 16:37:34
|
no
|
Biomed Eng Online. 2005 Aug 11; 4:48
|
utf-8
|
Biomed Eng Online
| 2,005 |
10.1186/1475-925X-4-48
|
oa_comm
|
==== Front
Biomed Eng OnlineBioMedical Engineering OnLine1475-925XBioMed Central London 1475-925X-4-521614304710.1186/1475-925X-4-52ResearchContribution of non-extensor muscles of the leg to maximal-effort countermovement jumping Nagano Akinori [email protected] Taku [email protected] Shinsuke [email protected] Senshi [email protected] Computational Biomechanics Unit, RIKEN; Hirosawa 2-1, Wako, Saitama, 351-0198, Japan2 Department of Computer Engineering and Information Technology, City University of Hong Kong; 83 Tat Chee Avenue, Kowloon, Hong Kong3 Dpartment of Life Sciences (Sports Sciences), the University of Tokyo; Komaba 3-8-1, Meguro, Tokyo, 153-8902, Japan2005 6 9 2005 4 52 52 29 6 2005 6 9 2005 Copyright © 2005 Nagano et al; licensee BioMed Central Ltd.2005Nagano et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
The purpose of this study was to determine the effects of non-extensor muscles of the leg (i.e., muscles whose primary function is not leg extension) on the kinematics and kinetics of human maximal-effort countermovement jumping. Although it is difficult to address this type of question through experimental procedures, the methodology of computer simulation can be a powerful tool.
Methods
A skeletal model that has nine rigid body segments and twenty degrees of freedom was developed. Two sets of muscle models were attached to this skeletal model: all (most of) major muscles in the leg ("All Muscles" model) and major extensor muscles in the leg (i.e., muscles whose primary function is leg extension; "Extensors Only" model). Neural activation input signal was represented by a series of step functions with a step duration of 0.05 s. Simulations were started from an identical upright standing posture. The optimal pattern of the activation input signal was searched through extensive random-search numerical optimization with a goal of maximizing the height reached by the mass centre of the body after jumping up.
Results
The simulated kinematics was almost two-dimensional, suggesting the validity of two-dimensional analyses when evaluating net mechanical outputs around the joints using inverse dynamics. A greater jumping height was obtained for the "All Muscles" model (0.386 m) than for the "Extensors Only" model (0.301 m). For the "All Muscles" model, flexor muscles developed force in the beginning of the countermovement. For the "All Muscles" model, the sum of the work outputs from non-extensor muscles was 47.0 J, which was 13% of the total amount (359.9 J). The quantitative distribution of the work outputs from individual muscles was markedly different between these two models.
Conclusion
It was suggested that the contribution of non-extensor muscles in maximal-effort countermovement jumping is substantial. The use of a computer simulation model that includes non-extensor muscles seems to be more desirable for the assessment of muscular outputs during jumping.
==== Body
Background
Jumping motions have been investigated by many researchers in the field of biomechanics in an effort to understand the coordination of the human body during explosive activities. A maximal-effort jumping is a suitable subject for this purpose, as the objective of a maximal-effort jumping can be defined in a very straightforward manner: "jump up as high as possible". Therefore less inter-subjects and intra-subject variability of body coordination is expected. In addition, jumping motions play important roles in many athletic activities such as track and field, basketball and volleyball. Therefore it is practically valuable to understand the biomechanics of the human body during jumping.
Researchers have reported many valuable insights regarding maximal-effort jumping motions using two-dimensional computer modelling and simulation [1-5]. Typically in these studies, leg muscles that have a primary function of leg extension (e.g., the m. gluteus maximus, m. rectus femoris, hamstrings, mm. vasti, m. gastrocnemius, m. soleus) were included in the model. In other words, other leg muscles that have a different primary function (e.g., joint flexion, abduction/adduction, rotation etc.; "non-extensor muscles") often were not explicitly implemented. Although it is true that the motion of the leg is mostly extension during jumping, there is a possibility that these non-extensor muscles do contribute to a jumping performance because of their three-dimensional anatomical configuration. Especially, when looking at the location of the origin, insertion and via-points of most muscles, it is observed that three-dimensional vectors instead of two-dimensional vectors better represent their line of action [6,7]. Therefore it is valuable to investigate whether or not these non-extensor muscles of the leg make a substantial contribution to jumping performance. (Note that muscles whose primary function is not leg extension are called "non-extensor muscles" in this paper. This nomenclature does not imply that these muscles do not contribute to leg extension at all. In fact, as the human body is a linked segmental system, the activity of a muscle can affect the actions of multiple joints/degrees of freedom in the system [8]. This paper utilized this nomenclature for the sake of simplicity.)
For that purpose, it is likely that the use of a three-dimensional neuromusculoskeletal model instead of a two-dimensional model is more straightforward. Anderson and Pandy [9] expanded their research on jumping motions using a three-dimensional model. Nagano et al. [10] also constructed a three-dimensional musculoskeletal model of a human ancestor's body that can be scaled up to represent the musculoskeletal system of modern humans [11]. The purpose of this study was to evaluate the contribution of non-extensor muscles of the leg to maximal-effort countermovement jumping using a three-dimensional neuromusculoskeletal model.
Methods
A 3D neuromusculoskeletal model of the human body was constructed using DADS-3D (LMS CADSI, Coralville, Iowa, USA) with the FORTRAN-based USER.FORCE option. Detailed properties of this model have been reported in preceding studies [10,11]. The musculoskeletal model consisted of nine rigid body segments (the head-arms-trunk (HAT), right and left upper legs, right and left lower legs, right and left feet and right and left toes) connected with frictionless joints (Figure 1). Body segmental parameter values were derived from [12] (body mass = 73.1 kg). The hip joints were modelled as ball-and-socket joints that have three degrees of freedom. The knee joints were modelled as hinge joints. The ankle joints were modelled as universal joints [13]. The metatarsophalangeal joints were modelled as hinge joints with a tilted axis [7]. The total number of degrees of freedom of the model was 20.
Figure 1 The musculoskeletal model developed for this study. The simulation model contained 9 rigid body segments, and the degrees of freedom of the model was 20.
The body was actuated by two different sets of muscles to construct the "All Muscles" model and the "Extensors Only" model. Thirty-two Hill-type lower limb muscles (16 muscles in each leg) were implemented in the "All Muscles" model (Table 1). These include all (more precisely, most of) major muscles found in a human leg. Fourteen muscles (7 muscles in each leg) were implemented in the "Extensors Only" model (Table 1). These include major leg extensor muscles only (i.e., muscles whose primary function is leg extension). Note that such biarticular muscles as the hamstrings, m. rectus femoris and m. gastrocnemius were regarded as extensor muscles. Under the joint configurations assumed during jumping, these muscles do develop more joint extension moments than joint flexion moments.
Table 1 The muscle parameter values used in this study. The values for each muscle are shown. Fmax: maximal isometric force of the contractile element. LCEopt: optimal length of the contractile element. αpen: pennation angle. Lslack: slack (unloaded) length of the series elastic element. ILIOP: m. iliopsoas. GMAXI: m. gluteus maximus. GMEDI: m. gluteus medius. GMIN: m. gluteus minimus. ADDLO: m. adductor longus. ADDMA: m. adductor magnus. ADDBR: m. adductor brevis. HEXRO: merged hip external rotator muscles. RECTF: m. rectus femoris. HAMST: merged hamstrings. VASTI: mm. vasti. BFESH: m. biceps femoris short head. GASTR: m. gastrocnemius. TIBAN: m. tibialis anterior. SOLEU: m. soleus. OPFLE: merged monoarticular planter flexor muscles other than m. soleus. All of these muscles were implemented in the "All Muscles" model. Muscles whose primary function is leg extension are noted as "Extensor". Only these muscles were implemented in the "Extensors Only" model.
Fmax (N) LCEopt (m) αpen (deg) Lslack (m)
ILIOP 1544 0.104 8 0.130
GMAXI 1883 0.142 5 0.125 Extensor
GMEDI 1966 0.054 8 0.078
GMINI 849 0.038 1 0.051
ADDLO 716 0.138 6 0.110
ADDMA 1916 0.087 5 0.060
ADDBR 531 0.133 0 0.020
HEXRO 1512 0.054 0 0.024
RECTF 1353 0.084 5 0.432 Extensor
HAMST 3054 0.080 15 0.359 Extensor
VASTI 6718 0.087 3 0.315 Extensor
BFESH 256 0.173 23 0.100
GASTR 2044 0.045 17 0.408 Extensor
TIBAN 532 0.098 5 0.223
SOLEU 5881 0.030 25 0.268 Extensor
OPFLE 3137 0.031 12 0.310 Extensor
Muscles investigated in [7] and [14] were considered for implementation. In order to perform computer simulation and numerical optimization within feasible computation time, it was necessary to limit the complexity of the model. Therefore muscles that have similar biomechanical function were merged to compose a single muscle group. For example, the m. vastus medialis, m. vastus intermedialis and m. vastus lateralis were merged as mm. vasti. Muscles or muscle groups whose maximal isometric force is greater than 500 N were selected. The m. biceps femoris short head, whose Fmax is smaller than 500 N, was also selected as the only mono-articular knee flexor muscle (Table 1). Coordinates of the origin, insertion and via-points of these muscles were derived from [7]. Muscle parameter values, i.e., the optimal contractile element length (LCEopt), maximal isometric force of the contractile element (Fmax), pennation angle (αpen) and unloaded length of the series elastic element (Lslack), were derived from [7] and [14]. A specific tension value of 31.5 N/cm2 [15] was utilized. A bilateral symmetry was assumed between the right side and the left side of the body.
A muscle-tendon complex was composed of a contractile element (CE) and a series elastic element (SEE) serially connected with a pennation angle (αpen) (Figure 2). The mathematical model of the contractile element represented the force-length-velocity relations. Passive stress-strain property of the series elastic element was modelled with a quadratic function. A detailed mathematical representation of these components can be found in [16]. Neural activation input to individual muscles was represented by a series of step functions with duration of 0.050 s [17]. Excitation dynamics of the contractile element was modelled with a first-order ordinary differential equation as described in [18].
Figure 2 The musculotendon model utilized in this study. The musculotendon model was composed of a contractile element (CE) and a series elastic element (SEE). The effect of pennation angle (αpen) was also taken into consideration. The contractile element had the force-length-velocity relation, and the series elastic element had a non-linear force-length relation.
The interaction between the foot segments and the ground was modelled using the same form of equation as was reported in [9]. Passive joint properties that function to limit the joint range of motion were adopted from [9].
Maximal-effort countermovement jump motions were generated through computer simulation with the "All Muscles" model and the "Extensors Only" model. A simulation was initiated from an upright posture with the hip, knee and ankle joints slightly flexed (5 degrees: dorsiflexed for the ankle joint) to facilitate the generation of countermovement. Simulations were performed from exactly the same initial posture for these two models. Muscle activation input profiles were modified through Bremermann's numerical optimization [19] in which the jumping height was maximized. The optimal combination of the activation input profiles for the muscles was searched. The optimization process was terminated when the objective function value had not improved for 10,000 successive iterations, which corresponds to approximately 60,000 function evaluations without any improvement [11].
The instantaneous power output value of the contractile element (PCE) was calculated as the product of the force development (FCE) and the shortening speed (VCE; positive value for shortening and negative value for lengthening) of the contractile element:
PCE = FCE·VCE (Eq. 1)
The work output of the contractile element (WCE) was calculated as the time integration of PCE from the start of simulation through the instant of take off:
Results
The maximal height reached by the mass centre of the body measured from the floor was 1.317 m for the "All Muscles" model and 1.233 m for the "Extensors Only" model (Table 2). The jumping height measured from the starting posture was 0.386 m and 0.301 m, respectively. With the body mass of 73.1 kg and the gravitational acceleration of 9.81 m/s2, the energy gain of the mass centre of the body throughout the jumping motion was 277 J and 216 J, respectively.
Table 2 The results of the numerical optimization obtained in this study. Hinit: initial height of the mass centre of the body. Hmax: maximal height reached by the mass centre of the body. Egain: energy gain of the mass centre of the body through the jumping motion calculated from the jump height, body mass and acceleration due to gravity. Δ: difference between the values for the "All Muscles" model and for the "Extensors Only" model.
Hinit (m) Hmax (m) Jump Height (m) Egain (J)
All Muscles 0.931 1.317 0.386 277
Extensors Only 0.931 1.233 0.301 216
Δ - 0.084 0.084 61
Realistic kinematics of jumping was generated both for the "All Muscles" model and for the "Extensors Only" model. Sagittal views of the kinematics are presented as Figure 3. The motions of the segments/joints outside of the sagittal plane were small (~10 deg; not shown), suggesting that the motion of the skeletal system was mostly two-dimensional. Take-off occurred at 0.65 s and 0.61 s after the start of simulation for the "All Muscles" model and for the "Extensors Only" model, respectively. Ground reaction force profiles are shown in Figure 4.
Figure 3 The countermovement jumping kinematics generated in this study (sagittal view). The take-off occurred 0.65 s and 0.61 s after the start of simulation for the "All Muscles" model and for the "Extensors Only" model, respectively.
Figure 4 The profile of ground reaction force. The dashed curve represents the value for the "All Muscles" model, whereas the solid curve represents the value for the "Extensors Only" model. The dashed vertical lines represent the instant of take-off. The dashed horizontal line represents the body weight.
In the "All Muscles" model, joint flexor muscles such as the m. iliopsoas, m. biceps femoris short head and m. tibialis anterior were activated in the beginning of the countermovement phase (Figure 5). For the hamstrings, mm. vasti and other plantar flexor muscles, force output was greater for the "Extensors Only" model than for the "All Muscles" model. For the m. rectus femoris, m. gastrocnemius and m. soleus, force output was greater for the "All Muscles" model than for the "Extensors Only" model.
Figure 5 The profile of muscle force output. The dashed curve represents the value for the "All Muscles" model, whereas the solid curve represents the value for the "Extensors Only" model. The dashed vertical lines represent the instant of take-off. The added values for two contralateral muscles are shown. The muscles whose primary function is leg extension (Table 1) are noted by (E).
For the "All Muscles" model, non-extensor muscles such as hip adductors and external rotators performed relatively little work (Table 3, Figure 6), although the sum of the work outputs was substantial (47.0 J). The behaviour of the hamstrings was markedly different between the "All Muscles" model and the "Extensors Only" model. Specifically, for the "All Muscles" model, the hamstrings exerted relatively small magnitude of positive work suggesting that the action of this muscle was mostly isometric. On the other hand, for the "Extensors Only" model, this muscle had relatively large negative work suggesting that the action of this muscle was mostly eccentric (Table 3, Figure 6).
Figure 6 The profile of the power output of the contractile element. The dashed curve represents the value for the "All Muscles" model, whereas the solid curve represents the value for the "Extensors Only" model. Positive is concentric and negative is eccentric. The dashed vertical lines represent the instant of take-off. The added values for two contralateral muscles are shown. The muscles whose primary function is leg extension (Table 1) are noted by (E).
Table 3 The amount of the mechanical work performed by the contractile element. The integrated values from the start of simulation through the instant of take-off. The added values for two legs (two contralateral muscles) are shown. Δ: difference between the values for the "All Muscles" model and for the "Extensors Only" model.
CE Work (two legs) (J)
All Muscles Extensors Only Δ
ILIOP 2.0 0.0 2.0
GMAXI 47.8 32.6 15.2
GMEDI 4.1 0.0 4.1
GMINI 4.5 0.0 4.5
ADDLO 10.3 0.0 10.3
ADDMA 9.1 0.0 9.1
ADDBR 2.4 0.0 2.4
HEXRO 14.8 0.0 14.8
RECTF 15.0 14.7 0.3
HAMST 10.1 -36.1 46.2
VASTI 133.1 172.2 -39.1
BFESH 0.4 0.0 0.4
GASTR 27.1 24.7 2.4
TIBAN -0.6 0.0 -0.6
SOLEU 28.3 40.3 -12.0
OPFLE 51.5 39.4 12.1
SUM 359.9 287.8 72.0
Discussion
The purpose of this study was to evaluate quantitatively the contribution of non-extensor muscles (muscles whose primary function is not leg extension) of the leg to maximal-effort countermovement jumping. Details of the simulation model utilized in this study have been described in [11]. In that study, a countermovement jumping motion simulated with the "All Muscles" mode have been analyzed and compared with the experimental data reported in preceding studies, and the validity of the modelling and simulation has been discussed. The optimized jumping height was smaller for the "Extensors Only" model than for the "All Muscles" model by 0.084 m (Table 2; 28.0%). In this study, the decrement in performance was caused by the absence of non-extensor muscles. This implies that non-extensor muscles do have substantial contributions to a maximal-effort countermovement jumping performance.
In both cases, the general characteristics of the jumping kinematics obtained through the numerical optimization process (Figure 3) were similar to the ones obtained through experimental data collection of human countermovement jumping [20], although the motion of the body and joint excursions were greater for the "All Muscles" model than for the "Extensors Only" model (Figure 3). Only limited motions of the skeletal system were observed outside of the sagittal plane in this study. It should be noted that the computer simulation model utilized in this study has a capability to perform fully three-dimensional motions (e.g., hip joint abduction/adduction etc.). Nonetheless, the simulation model chose to perform almost two-dimensional motions. This finding supports that the two-dimensional inverse dynamic analyses on jumping performed in numerous preceding studies are mostly valid. Especially when calculating such mechanical variables as net joint reaction forces, net joint moments and power outputs of joints, reliable calculations can be assumed.
When performing computer simulation of jumping, it is assumed to be acceptable to construct a two-dimensional skeletal model of the human body for the same reason. However, when attaching muscle models to the skeletal model, it will be more appropriate to explicitly consider the contribution of non-extensor muscles of the leg. Implementing three-dimensional configuration of these muscles will be the most straightforward solution. Calculating the projection of the line of action of these muscles to the sagittal plane will be another option to accomplish this.
Regarding the profiles of ground reaction force (Figure 4), two peaks were observed during the push-off phase. The first peak was mostly caused through the interaction between the heel and the floor in the beginning of the push-off phase, whereas the second peak was mostly caused through the interaction between the toe and the floor in the last part of the push-off phase. This profile of ground reaction force with two peaks is often observed in ground reaction force data collected from human subjects during a maximal-effort countermovement jumping [1,21]. The profile of ground reaction force was bumpy because each foot was modelled with only five contact points [11]. The profile will become smoother with more contact points in a foot, although this modification will greatly increase the computation time.
In the beginning of countermovement, joint flexor muscles, such as the m. iliopsoas, m. biceps femoris short head and m. tibialis anterior, developed force (Figure 5). This resulted in a greater countermovement for the "All Muscles" model than for the "Extensors Only" model. This seems to suggest that the contribution of these flexor muscles in the beginning of countermovement should be considered when investigating the mechanism of maximal-effort countermovement jumping motion. For the "All Muscles" model, non-extensor muscles such as the m. gluteus medius, m. gluteus minimus, adductors and hip external rotators had relatively minor individual contributions in terms of mechanical work and power outputs (Table 3). However, when the work outputs of these muscles were added together, the amount was substantial (47.0 J in 359.9 J; 13%), suggesting that the contribution of these muscles in jumping motion is not negligible.
For the m. gluteus maximus, m. rectus femoris, hamstrings, m. gastrocnemius and other monoarticular plantar flexor muscles (OPFLE), the work output was greater for the "All Muscles" model than for the "Extensors Only" model (Table 3). This result is very reasonable considering that the jumping height was greater for the "All Muscles" model than for the "Extensors Only" model (Table 2). Generally speaking, to achieve a higher jumping performance in an optimally-coordinated movement, muscles need to perform more work. As the jumping height was greater for the "All Muscles" model, greater mechanical outputs of muscles are reasonably expected for this model than for the "Extensors Only" model.
However, there were two exceptions; for the mm. vasti and m. soleus, the work output was greater for the "Extensors Only" model than for the "All Muscles" model. This result came from the fact that the "Extensors Only" model underwent a smaller countermovement compared to the "All Muscles" model (Figure 3, Figure 4). In this study, the amount of work output was calculated as a net (positive and negative) value from the start of a motion through the instant of take-off (Eq. 1 and 2). As the magnitude of countermovement (negative phase) was smaller for the "Extensors Only" model, the net amount of work output of the mm. vasti and m. soleus was calculated to be greater for this model. This phenomenon can be observed in Figure 6, where a smaller negative power output of these muscles during the countermovement is exhibited for the "Extensors Only" model. As the mm. vasti is a major knee extensor and the m. soleus is a major ankle plantarflexor, these muscles had to function to brake the downward momentum generated during the countermovement. This had the effect of reducing the net work output of these muscles.
For the hamstrings, a positive work output (10.1 J) was calculated for the "All Muscles" model, whereas a negative (-36.1 J) value was calculated for the "Extensors Only" model. It is observed that the contractile element of the hamstrings was mostly stretched in an eccentric manner in the "Extensors Only" model in the latter phase of the countermovement, resulting in a negative power (Figure 6) and work (Table 3) outputs. This is because only a few muscles that can act as extensors were available to brake the countermovement of the trunk segment in this model. Specifically, the m. adductor longus, m. adductor brevis and m. adductor magnus had been removed from the model. Therefore the inertial load of the trunk segment (moving downwards) imposed on the hamstrings in the latter phase of the countermovement was so great as to stretch this muscle in an eccentric manner, although this muscle was vigorously activated during this period (Figure 5). In other words, the hamstrings was not strong enough to brake the downward momentum of the trunk segment in a concentric manner. This discussion is consistent with the muscle force development profile shown in Figure 5. The force development of the hamstrings was greater for the "Extensors Only" model than for the "All Muscles" model (Figure 5), which is reasonable considering that the action of this muscle was mostly eccentric for the "Extensors Only" model (eccentric part of the force-velocity relation; Figure 2). On the other hand, as the "All Muscles" model had more muscles to function to brake the countermovement, the hamstrings could shorten itself and produce positive (concentric) work and power outputs. For example, the m. adductor magnus did function to brake the countermovement in this model (Figure 6). This result is consistent with the observation that there was a greater countermovement for the "All Muscles" model than for the "Extensors Only" model. As there were not enough muscles to brake the downward momentum generated during the countermovement, the optimal magnitude of countermovement for the "Extensors Only" model was smaller than that for the "All Muscles" model.
Conclusion
As a result of this computer simulation study, it was found that the dynamics of the body motion is altered by the effects of non-extensor muscles. This finding is noteworthy considering that the overall kinematics of the body (Figure 3) and the ground reaction force profiles (Figure 4) were similar between the "All Muscles" model and the "Extensors Only" model. This result implies that it is desirable to consider explicitly the mechanical contribution of non-extensor muscles of the leg when investigating human jumping motions in terms of mechanical outputs of muscles. The use of a three-dimensional neuromusculoskeletal model seems to be more suitable for this purpose. However, results of this computer simulation study also supported that the nature of an optimally-coordinated countermovement jumping motion is mostly two-dimensional, which suggests the validity of the two-dimensional inverse dynamic analyses of net mechanical outputs around joints performed in many preceding studies.
Authors' contributions
AN constructed the simulation model, performed computer simulation and other calculations, and drafted the manuscript. TK and SY participated in the process of model construction and numerical optimization. SF contributed valuable discussions and suggestions throughout this project, including the stage of manuscript writing. All authors read and approved the final manuscript.
Acknowledgements
AN would like to thank the special post-doctoral program of RIKEN. SY would like to thank the Junior Research Associate program of RIKEN. AN and SY would like to thank Professor Ryutaro Himeno at RIKEN for his supports. This study was partly supported by the Ministry of Education, Culture, Sports, Science and Technology in Japan (No: 16300205).
==== Refs
Anderson FC Pandy MG Storage and Utilization of Elastic Strain-Energy during Jumping J Biomech 1993 26 1413 1427 8308046 10.1016/0021-9290(93)90092-S
Bobbert MF Van Soest AJ Effects of muscle strengthening on vertical jump height: a simulation study Med Sci Sports Exerc 1994 26 1012 1020 7968418
Pandy MG Zajac FE Optimal Muscular Coordination Strategies for Jumping J Biomech 1991 24 1 10 2026629 10.1016/0021-9290(91)90321-D
Pandy MG Zajac FE Sim E Levine WS An Optimal-Control Model for Maximum-Height Human Jumping J Biomech 1990 23 1185 1198 2292598 10.1016/0021-9290(90)90376-E
van Soest AJ Schwab AL Bobbert MF van Ingen Schenau GJ The influence of the biarticularity of the gastrocnemius muscle on vertical-jumping achievement J Biomech 1993 26 1 8 8423165 10.1016/0021-9290(93)90608-H
Brand RA Crowninshield RD Wittstock CE Pedersen DR Clark CR van Krieken FM A model of lower extremity muscular anatomy J Biomech Eng 1982 104 304 310 7154650
Delp SL Loan JP Hoy MG Zajac FE Topp EL Rosen JM An interactive graphics-based model of the lower extremity to study orthopaedic surgical procedures IEEE Trans Biomed Eng 1990 37 757 767 2210784 10.1109/10.102791
Yamaguchi GT Norwell MA Dynamic modeling of musculoskeletal motion: a vectorized approach for biomechanical analysis in three dymensions 2001 USA: Kluwer Academic Publishers
Anderson FC Pandy MG A Dynamic Optimization Solution for Vertical Jumping in Three Dimensions Comput Methods Biomech Biomed Engin 1999 2 201 231 11264828
Nagano A Umberger BR Marzke MW Gerritsen KG Neuromusculoskeletal computer modeling and simulation of upright, straight-legged, bipedal locomotion of Australopithecus afarensis (A.L. 288-1) Am J Phys Anthropol 2005 126 2 13 15386246 10.1002/ajpa.10408
Nagano A Komura T Fukashiro S Himeno R Force, work and power output of lower limb muscles during human maximal-effort countermovement jumping J Electromyogr Kinesiol 2005 15 367 376 15811607 10.1016/j.jelekin.2004.12.006
de Leva P Adjustments to Zatsiorsky-Seluyanov's segment inertia parameters J Biomech 1996 29 1223 1230 8872282 10.1016/0021-9290(95)00178-6
Inman VT Baltimore MD The joints of the ankle 1976 USA: Williams and Wilkins
Friederich JA Brand RA Muscle fiber architecture in the human lower limb J Biomech 1990 23 91 95 2307696 10.1016/0021-9290(90)90373-B
Brown IE Satoda T Richmond FJR Loeb GE Feline caudofemoralis muscle – Muscle fibre properties, architecture, and motor innervation Exp Brain Res 1998 121 76 91 9698193 10.1007/s002210050439
Nagano A Gerritsen KGM Effects of neuromuscular strength training on vertical jumping performance – A computer simulation study J Appl Biomech 2001 17 113 128
Nagano A Fukashiro S Komura T Contribution of series elasticity in human cyclic heel-raise exercise J Appl Biomech 2003 19 340 352
He JP Levine WS Loeb GE Feedback Gains for Correcting Small Perturbations to Standing Posture Ieee Transactions on Automatic Control 1991 36 322 332 10.1109/9.73565
Bremermann H A method of unconstrained optimization Math Biosci 1970 9 1 15 10.1016/0025-5564(70)90087-8
Bobbert MF Gerritsen KG Litjens MC Van Soest AJ Why is countermovement jump height greater than squat jump height? Med Sci Sports Exerc 1996 28 1402 1412 8933491
Fukashiro S Komi PV Jarvinen M Miyashita M In vivo Achilles tendon loading during jumping in humans Eur J Appl Physiol Occup Physiol 1995 71 453 458 8565978 10.1007/BF00635880
|
16143047
|
PMC1215494
|
CC BY
|
2021-01-04 16:37:34
|
no
|
Biomed Eng Online. 2005 Sep 6; 4:52
|
utf-8
|
Biomed Eng Online
| 2,005 |
10.1186/1475-925X-4-52
|
oa_comm
|
==== Front
J CarcinogJournal of Carcinogenesis1477-3163BioMed Central London 1477-3163-4-131610918010.1186/1477-3163-4-13ResearchGenetic polymorphisms in DPF3 associated with risk of breast cancer and lymph node metastases Hoyal Carolyn R [email protected] Stefan [email protected] Richard B [email protected] Richard [email protected] George [email protected] Marion [email protected] Ulrike [email protected] Lyn R [email protected] Florian [email protected] Joachim [email protected] Matthew R [email protected] Andreas [email protected] Sequenom, Inc., San Diego, California, USA2 Department of Obstetrics & Gynecology, Technical University of Munich, Germany3 Genomics Research Centre, School of Health Science, Griffith University Gold Coast, Australia4 I. Frauenklinik, Klinikum Innenstadt, University of Munich, Germany2005 19 8 2005 4 13 13 5 5 2005 19 8 2005 Copyright © 2005 Hoyal et al; licensee BioMed Central Ltd.2005Hoyal et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Several studies have identified rare genetic variations responsible for many cases of familial breast cancer but their contribution to total breast cancer incidence is relatively small. More common genetic variations with low penetrance have been postulated to account for a higher proportion of the population risk of breast cancer.
Methods and Results
In an effort to identify genes that influence non-familial breast cancer risk, we tested over 25,000 single nucleotide polymorphisms (SNPs) located within approximately 14,000 genes in a large-scale case-control study in 254 German women with breast cancer and 268 age-matched women without malignant disease. We identified a marker on chromosome 14q24.3-q31.1 that was marginally associated with breast cancer status (OR = 1.5, P = 0.07). Genotypes for this SNP were also significantly associated with indicators of breast cancer severity, including presence of lymph node metastases (P = 0.006) and earlier age of onset (P = 0.01). The association with breast cancer status was replicated in two independent samples (OR = 1.35, P = 0.05). High-density association fine mapping showed that the association spanned about 80 kb of the zinc-finger gene DPF3 (also known as CERD4). One SNP in intron 1 was found to be more strongly associated with breast cancer status in all three sample collections (OR = 1.6, P = 0.003) as well as with increased lymph node metastases (P = 0.01) and tumor size (P = 0.01).
Conclusion
Polymorphisms in the 5' region of DPF3 were associated with increased risk of breast cancer development, lymph node metastases, age of onset, and tumor size in women of European ancestry. This large-scale association study suggests that genetic variation in DPF3 contributes to breast cancer susceptibility and severity.
==== Body
Background
Breast cancer etiology is a complex process, involving genes in the multiple stages of carcinogenesis, from initial cell cycle dysregulation to metastatic potential [1,2]. Approximately ten percent of breast cancer cases occur within families in which the disease segregates in a Mendelian fashion. BRCA1 and BRCA2 have been identified to be responsible for a substantial proportion of familial breast cancer [3,4]. Other genes involved in the same DNA double-strand break repair pathway, TP53 [5], ATM [6], and PTEN [7], are also known to contribute to familial cases, but are more rare. Such high penetrance germ line mutations are responsible for less than 10% of all breast cancer cases. However, genetic variation is estimated to contribute approximately 25% to the population risk of breast cancer, likely accounted for by a large number of yet undiscovered common, low penetrance alleles [8,9]. It is possible that these low penetrance markers may be useful in the development of practical prognostic and diagnostic indicators with greater utility in the general population.
Many candidate gene studies have been performed to identify the genes that contribute to risk for sporadic breast cancer [10]. Unfortunately, these efforts have been largely unsuccessful. Some of the more consistently reported candidates include variations in metabolizing enzymes, such as the cytochrome P-450 family [11], N-acetyltransferases [12], and glutathion-S-transferases [13]. The candidate susceptibility allele CHEK2*1110delC was shown to confer an increased breast cancer risk [14,15], which was more recently supported by results obtained in a large case-control study [16]. In an effort to identify novel genes involved in breast cancer susceptibility, we have conducted a large-scale, case-control study using more than 25,000 SNPs located within approximately 14,000 genes. We previously reported the findings on two breast cancer candidates identified in this study [17,18]. Herein, we describe variations in intron 1 of DPF3 on chromosome 14q24.3-q31.1 that are associated with increased risk of breast cancer, lymph node metastases, earlier age of diagnosis, and tumor size.
Methods
Subjects and Study Design
The participants in the large-scale association study (referred to as the discovery sample) were recruited among patients attending the Frauenklinik Innenstadt, University of Munich, Germany, and comprised 254 breast cancer cases. At the time of assessment, 94 cases (37%) displayed positive lymph node status, and 18 cases (7%) had known distant metastases. Twenty-seven cases (11%) reported having at least one first- or second-degree relative with breast cancer. The median age of diagnosis was 56 yr (range = 23–87 yr). During the same period, 268 controls with a median age of 57 yr (range = 17–88 yr) were recruited from patients with benign disease being seen at the clinic. Controls with a reported family history of breast or ovarian cancer were excluded from the current study. Both parents of each study participant were reported to be of German descent.
The participants in the German replication sample were recruited from the Department of Obstetrics and Gynecology, Technical University of Munich, and consisted of 188 cases and 150 controls. Most breast cancer cases were recruited at pre-operative visits, and the female controls were recruited from healthy individuals or patients with non-malignant diagnoses. Median age of diagnosis for cases was 59 yr (range = 22–87 yr) and median age of controls was 50 yr (range = 19–91 yr). Two participants reported one parent of non-German, Eastern European origin; otherwise both parents were of German descent.
The participants in the Australian replication sample were recruited from the Pathology Department of Gold Coast Hospital or by the Genomics Research Centre, Southport. The collection included 180 breast cancer cases with a median age of diagnosis of 50 yr (range = 24–74 yr). Controls consisted of 180 healthy volunteers without family history of cancer recruited through the Genomics Research Centre. Controls were individually age-matched to cases (± 5 yr) with the median age for controls at 60 yr (range = 28–94 yr).
All subjects involved in our studies signed a written informed consent and the institutional ethics committees of participating institutions approved the experimental protocols.
SNP Markers, Genotyping, and Resequencing
A set of 25,494 SNPs covering the human genome was selected from a larger collection of 125,799 experimentally validated polymorphisms [19]. This set includes SNPs that are located in gene coding regions (within 10 kb of 13,735 genes annotated in Entrez Gene), have a minor allele frequencies greater than 0.02 (95% have frequencies greater than 0.1), and a median inter-marker spacing of 40 kb. SNP annotation was based upon the NCBI dbSNP database, refSNP build 118 [20]. Genomic annotation was based on NCBI Genome Build 34. Gene annotation was based upon Entrez Gene genes for which NCBI was providing positions on the Mapview FTP site [21].
DNA pools were formed by combining equimolar amounts of each sample as described elsewhere [22,23]. For SNP assays carried out on pooled DNA, 25 ng of DNA was used. All PCR and MassEXTEND™ reactions were conducted using standard conditions [23]. Relative allele frequency estimates were derived from calculations based on the area under the peak of mass spectrometry measurements from four analyte aliquots as described elsewhere [23]. The same procedure was used for individual genotyping except 2.5 ng DNA was used and only one mass spectrometry measurement was taken. Primers used for genotyping are presented in Table 1. Sequencing was performed under standard conditions for MassCLEAVE™ [24] using 5 ng of DNA. For Exon 1, the amplification primers used were 5'-AACGGCAGAGCACATGTAGTAA-3' and 5'-ATATTGAAACCACGCGGAATA-3'. For Exon 2, the amplification primers used were 5'-CTGGGTGTGTTTCAGTCTTCC-3' and 5'-CTGGTTTCCCAGACAAGCTG-3'.
Table 1 Primer sequences used in genotyping.
rs8010957 Forward primer 5'-TGC TGG GAT TAT GAG CCA CT-3'
Reverse primer 5'-GTG TGT CTC CAG TAA AGG GC-3'
Extend primer 5'-AAA ACT CTG CTA CTG GC-3'
rs4307892 Forward primer 5'-GCA AAA TGC TAG TAA ATG GTG-3'
Reverse primer 5'-GAA AAA TGG CAA GCC TTC TG-3'
Extend primer 5'-AGA GCA ATG AAC ACC AAT ATC C-3'
rs1990440 Forward primer 5'-AAG TCA CTA ACC CCA CAC AC-3'
Reverse primer 5'-CCA GGG TGT GTT CTA ATA CG-3'
Extend primer 5'-CGT CAG CAA ATG TGT ACC GA-3'
rs4899445 Forward primer 5'-AGG AGA GTC TGC CCA TTT GA-3'
Reverse primer 5'-AGA AAA CTC ACC TCC CTG AC-3'
Extend primer 5'-AGC CCT CTC CAG GGC CAT GC-3'
rs4378563 Forward primer 5'-GCC GTG TGC ATA TCC TGA TC-3'
Reverse primer 5'-TTA TGG CTT CCT CTC CCT AC-3'
Extend primer 5'-CCT CCA TGC CCT GCT TA-3'
rs12232220 Forward primer 5'-TTA AAA ATA CAA TGA TGG CC-3'
Reverse primer 5'-TCC CGA CCT CAG GTG ATG TG-3'
Extend primer 5'-GAT TAC AGG TAT GAG CCA C-3'
Statistical Methods
Tests of association between disease status and each SNP using pooled DNA were carried out in a similar fashion as explained elsewhere [25]. Sources of measurement variation included pool formation, PCR/mass extension, and chip measurement. When three or more replicate measurements of an allele frequency were available within a model level, the corresponding variance component was estimated from the data. Otherwise, the following historical laboratory averages were used: pool formation = 5.0 × 10-5, PCR/mass extension = 1.7 × 10-4, and chip measurement = 1.0 × 10-4. Tests of association using individual genotypes were carried out using a chi-square test of heterogeneity based on allele and genotype frequencies. Selected tests of association involving contingency tables with rare or missing cells were carried out using Fisher's exact test. The DerSimonian-Laird random effects meta-analysis method [26] was used for the analysis of replication samples to test for the consistency of association while permitting allele frequencies to differ among collections. All tests of allele frequencies involving only replication samples are one-sided, confirming the effect observed in the discovery sample. P-values were derived using the log odds of each contrast and their standard errors. Multiple approaches were explored in an effort to identify haplotypes demonstrating a stronger association with disease status than single sites. These included analyses of six SNP haplotypes and subsets thereof using the coalescent theory-based PHASE v2.0 [27] and the score method that relies on the EM algorithm [28]. No attempt was made to correct P-values for multiple testing. Rather, P-values are provided to compare the relative strength of association from multiple dependent (e.g. SNPs within samples) and independent (e.g. SNPs between samples) sources of information. P-values less than 0.05 are referred to as statistically significant.
Results
SNP markers associated with breast cancer status were identified using a three-phased approach. In the first phase, pools of case and control samples were subjected to a single PCR reaction and primer extension for each of the 25,494 SNP assays. Four aliquots of the extension products were measured. The relative allele frequencies were compared, and 1,619 SNPs (~5%) with the most statistically significant associations were selected to be tested in the second phase. In this phase, allele frequencies were measured in three separate PCR and primer extension reactions using case and control pools, and compared as in the first phase. The 74 most significant SNPs (~5%) from the second phase were selected for individual genotyping in the samples that comprised the case and control pools (Figure 1).
Figure 1 Univariate distributions of summary statistics of the 74 genotyped SNP assays. Empirical densities (histograms) are provided for A) the minor allele frequency in the controls, B) the odds ratios (ORs) on a log2 scale presenting the fold-change comparing allele frequencies between cases and controls, and C) the P-values from the tests of association.
Case-control studies employing tens of thousands of SNPs in a genome-wide approach using liberal selection criteria are expected to yield a high proportion of false positive associations. To determine if the observed association was a true genetic effect, the 74 SNPs were subsequently genotyped in two additional breast cancer case-control collections. After reviewing the results of all three samples, one significant result was observed for a C-to-G SNP, rs1990440, in intron 1 of the DPF3 gene on chromosome 14q24.3-q31.1. The frequency of the G allele in discovery control subjects was 0.08, similar to the NCBI reported average allele frequency [29]. The frequency was increased by 4% in the cases. Table 2 shows the association of rs1990440 with breast cancer in the discovery and two replication collections. Even though this SNP was only marginally associated in the German discovery sample (OR = 1.49, P = 0.069), German replication sample (OR = 1.33, P = 0.29), and Australian replication sample (OR = 1.36, P = 0.22), the estimated effects were consistent and the analysis of all three samples resulted in a combined significance of P = 0.016 (OR = 1.40) and a significance of P = 0.054 (OR = 1.35) within the replication samples only.
Table 2 Distribution of genotype counts and relative allele frequencies of the DPF3 polymorphisms1 in breast cancer and control groups for discovery and replication samples.
Study Na Genotype count (%) MAFb ORc P-valued
rs8010957 (Intron 1 G>A) AA AG GG A
German2 Case 235 0 (0) 8 (3) 227 (97) 2% 0.34 0.006
Control 256 0 (0) 23 (9) 233 (91) 5%
German3 Case 182 0 (0) 9 (5) 173 (95) 2% 0.94 0.912
Control 134 0 (0) 7 (5) 127 (95) 3%
Australian Case 163 0 (0) 5 (3) 158 (97) 2% 0.62 0.407
Control 164 0 (0) 8 (5) 156 (95) 2%
Replication only 0.79 0.270
Total (all of centres) 0.55 0.058
rs4307892 (Intron 1 G>A) AA AG GG A
German2 Case 238 2 (1) 59 (25) 177 (74) 13% 1.56 0.030
Control 252 1 (0) 44 (17) 213 (83) 9%
German3 Case 185 1 (1) 40 (22) 144 (78) 11% 1.44 0.177
Control 141 2 (1) 19 (13) 120 (85) 8%
Australian Case 173 0 (0) 33 (19) 140 (81) 10% 1.23 0.445
Control 171 1 (1) 25 (15) 145 (85) 8%
Replication only 1.33 0.068
Total (all of centres) 1.43 0.010
rs19904404 (Intron 1 C>G) GG GC CC G
German2 Case 206 2 (1) 46 (22) 158 (77) 12% 1.49 0.069
Control 253 2 (1) 39 (15) 212 (84) 8%
German3 Case 191 1 (1) 39 (20) 151 (79) 11% 1.33 0.286
Control 145 1 (1) 22 (15) 122 (84) 8%
Australian Case 179 1 (2) 34 (19) 144 (80) 11% 1.36 0.224
Control 171 1 (1) 25 (15) 145 (85) 8%
Replication only 1.35 0.054
Total (all of centres) 1.40 0.016
rs4899445 (Intron 1 A>G) GG GA AA G
German2 Case 240 2 (1) 48 (20) 190 (79) 11% 1.56 0.045
Control 257 1 (0) 35 (14) 221 (86) 7%
German3 Case 185 2 (1) 28 (15) 155 (84) 9% 1.72 0.094
Control 143 0 (0) 15 (10) 128 (90) 5%
Australian Case 176 1 (1) 33 (19) 142 (81) 10% 1.47 0.160
Control 172 1 (1) 22 (13) 149 (87) 7%
Replication only 1.56 0.016
Total (all of centres) 1.56 0.003
rs4378563 (Intron 1 C>T) TT CT CC T
German2 Case 235 2 (1) 58 (25) 175 (74) 13% 1.86 0.004
Control 252 2 (1) 34 (13) 216 (86) 8%
German3 Case 175 1 (1) 30 (17) 144 (82) 9% 1.15 0.624
Control 131 1 (1) 19 (15) 111 (85) 8%
Australian Case 162 1 (1) 30 (19) 131 (81) 10% 1.20 0.506
Control 167 1 (1) 26 (16) 140 (84) 8%
Replication only 1.18 0.210
Total (all of centres) 1.44 0.027
rs12232220 (Intron 2 G>A) AA AG GG A
German2 Case 237 0 (0) 16 (7) 221 (93) 3% 0.55 0.059
Control 253 1 (0) 28 (11) 224 (89) 6%
German3 Case 186 0 (0) 25 (13) 161 (87) 7% 2.23 0.038
Control 144 0 (0) 9 (6) 135 (94) 3%
Australian Case 166 0 (0) 16 (10) 150 (90) 5% 1.27 0.536
Control 169 0 (0) 13 (8) 156 (92) 4%
Replication only 1.66 0.960
Total (all of centres) 1.13 0.770
1Presented in order respective to position in DPF3; 2German discovery sample; 3German replication sample; 4Marker SNP identified in large-scale association study.
aNumber of subjects with genotypes; bMinor relative allele frequency; cOdds ratio for test comparing allele frequencies between cases and controls; dP-value for test comparing allele frequencies between cases and controls.
To fine map the region of association, we tested an additional 394 SNPs located within the DPF3 gene using the discovery case and control pools (Figure 2). We observed that the contiguous region of highest significance extended approximately 65 kb, spanning the 3' region of intron 1 with additional evidence for a 15 kb region that includes exon 2 and a part of intron 2. Using a cleavage assay and mass spectrometry [24], we re-sequenced exons 1 and 2 with their flanking intron sequences in six breast cancer cases and five controls to determine if any additional SNPs with stronger disease association or apparent functional relevance could be discovered. We identified only one SNP in intron 1 that was not publicly annotated (data not shown) and found to have an average allele frequency that did not significantly differ between the case and control pools. No previously described SNPs reside in exon 1 or 2, and no novel SNPs were discovered by our efforts. We selected five SNPs with allele frequencies that differed significantly between case and control pools, roughly distanced 20 kb apart, for genotyping in the discovery and replication samples and for further analysis (Table 1, 2). The SNPs most strongly associated in the discovery and replication samples, rs4307892, rs4899445 and rs4378563, were flanking the original marker SNP and were in strong linkage disequilibrium (all |D'| > 0.9, r2 > 0.7). Of the additional SNPs genotyped, rs4899445 demonstrated the most consistent differences between cases and controls, with a slightly larger effect in the discovery sample (OR = 1.56, P = 0.045) and a substantially more consistent effect in the German (OR = 1.72, P = 0.094) and Australian (OR = 1.47, P = 0.16) replication samples (Table 2). The effect of the combined replication sample was significant at the 0.05 level (OR = 1.56, P = 0.016) and equal to the estimate from all three samples (OR = 1.56, P = 0.003). Analyses of haplotypes consisting of subsets of the six genotyped SNPs did not reveal any haplotype with stronger association than individual SNPs (data not shown).
Figure 2 Fine mapping of breast cancer susceptibility on chromosome 14q24.3-q31.1. 395 SNPs in a 250-kb window were compared between pooled cases and controls. The x-axis corresponds to the chromosomal position and the y-axis to the test P-value (shown on the -log10 scale). The continuous dark line represents a goodness – of – fit test for excess of significance (compared to the 0.05 proportion expected by chance alone) in a 10-kb sliding window assessed at 1-kb increments. The continuous light grey line is the result of a non-linear smoothing function showing a weighted average of the P-values across the region. The shade of each point corresponds to the minor allele frequency of the corresponding SNP in the control sample (see legend below graph). The Entrez Gene annotation for NCBI genome build 34 is included.
The data collected on the patients in the German discovery collection included information on family history of breast cancer, age of onset, and disease severity. Further analysis revealed associations between the initial marker SNP, rs1990440, and multiple traits indicative of cancer aggressiveness (Table 3), including lower mean age of diagnosis of breast cancer (P = 0.01) and lymph node metastases (P = 0.006). Associations with organ metastases (P = 0.35) and tumor size (P = 0.17) were not statistically significant. The SNP most strongly associated with breast cancer risk across all three samples, rs4899445, was also found to be significantly associated with lymph node metastases (P = 0.008), and increased tumor size (P = 0.007). Though not statistically significant, the risk allele carriers tended to be younger at age of diagnosis (P = 0.35) and to have a higher proportion of breast cancer family history (P = 0.13).
Table 3 Association of the DPF3 polymorphisms with traits of interest in discovery breast cancer cases.
Trait Genotype Estimates P-value
AA AG GG
rs4307892 G>A N N = 2 N = 59 N = 177
Age of diagnosis (Years) 235 49.654.5 59.4 45.1 54.6 60.3 51.2 56.9 63.5 0.0981
Years since diagnosis 233 4.6 8.3 12.1 0.6 2.8 5.7 0.6 2.2 5.9 0.7931
Familial breast cancer history 254 1 (50) 6 (10) 17 (10) 0.2552
Tumor size3 236 0 (0) 22 (37) 47 (28) 0.3182
Lymph node metastases (Yes) 254 2 (100) 23 (39) 46 (26) 0.0142
Organ metastases (Yes) 254 0 (0) 6 (10) 10 (6) 0.3392
GG GC CC
rs1990440 C>G N N = 2 N = 46 N = 158
Age of diagnosis (Years) 235 49.654.5 59.4 45.0 51.7 60.1 51.9 57.4 65.0 0.0121
Years since diagnosis 233 4.6 8.3 12.1 0.4 3.0 6.8 0.6 2.2 5.4 0.7621
Familial breast cancer history 254 1 (50) 6 (13) 15 (9) 0.1652
Tumor size3 236 0 (0) 20 (43) 44 (30) 0.1692
Lymph node metastases (Yes) 254 2 (100) 19 (41) 39 (25) 0.0062
Organ metastases (Yes) 254 0 (0) 6 (13) 11 (7) 0.3472
GG GA AA
rs4899445 A>G N N = 2 N = 48 N = 190
Age of diagnosis (Years) 235 49.654.5 59.4 46.4 53.8 60.6 50.6 56.5 63.3 0.3501
Years since diagnosis 233 4.6 8.3 12.1 0.7 2.8 5.7 0.6 2.3 5.9 0.7431
Familial breast cancer history 254 1 (50) 6 (12) 17 (9) 0.1322
Tumor size3 236 0 (0) 23 (48) 47 (26) 0.0072
Lymph node metastases (Yes) 254 2 (100) 20 (42) 50 (26) 0.0082
Organ metastases (Yes) 254 0 (0) 5 (10) 11 (6) 0.4152
TT TC CC
rs4378563 C>T N N = 2 N = 58 N = 175
Age of diagnosis (Years) 235 49.654.5 59.4 46.0 53.8 60.6 51.0 56.9 63.4 0.2121
Years since diagnosis 233 4.6 8.3 12.1 0.4 2.5 6.0 0.7 2.3 6.0 0.7231
Familial breast cancer history 254 1 (50) 5 (9) 16 (9) 0.2742
Tumor size3 236 0 (0) 26 (46) 43 (26) 0.0112
Lymph node metastases (Yes) 254 2 (100) 23 (40) 46 (26) 0.0132
Organ metastases (Yes) 254 0 (0) 6 (10) 10 (6) 0.3382
a ab crepresent the lower quartile a, the median b, and the upper quartile c for continuous variables. 1Kruskal-Wallis test; 2Fisher's exact test; 3T ≥ 2 cm.
Discussion
Here, we report variants in DPF3, identified through a large-scale, genome-wide association study, that are associated with increased breast cancer risk, lymph node metastases, decreased age of onset, and increased tumor size. Our study suggests that individuals that carry one or more G alleles of the C-to-G variant rs1990440 have a nominally significant increase in breast cancer risk in comparison with the CC homozygotes. This association was substantiated in two independent collections from Germany and Australia. Fine mapping narrowed down the region of association to approximately 80 kb, spanning the majority of intron 1, exon 2 and a portion of intron 2. Subsequent genotyping of additional SNPs identified an intron 1 SNP, rs4899445, that was more consistently associated with breast cancer status across the three samples (OR = 1.56, P = 0.003).
The initial marker in DPF3 was one of 74 SNPs identified from a large-scale association study. The estimated effect of this marker was relatively small and would have been discounted had similar effects not been observed in the two replication samples. Even so, the statistical significance observed in the replication samples alone for the marker SNP (rs1990440; P = 0.054) or the more significant SNP identified nearby (rs4899445; P = 0.016) would not hold up to an experiment-wide type I error rate of 0.05 after correcting for multiple testing. Given that that 74 regions followed up in replication were largely independent on a population level, a conservative Bonferroni correction would require P-values to be less than 0.0006 to achieve the stated experiment-wide false positive rate. Indeed, validation of these results will require a larger sample collection. If we were to assume that the true effect is as estimated by the replication samples (OR = 1.5) with a population allele frequency of 6%, aggregate sample sizes on the order of 1,000 cases and controls will be necessary to have adequate power to substantiate the effect of this region on breast cancer risk.
DPF3 encodes a zinc finger protein on chromosome 14q24.3-q31.1. Zinc-fingers consist of clusters of cysteines or cysteines and histidines that coordinately bind zinc ions. DPF3 is a highly conserved gene homologous to members of the d4 family of zinc-fingers. Previous characterization of the DPF3 gene demonstrated that while it is similar to DPF1 (neuro-d4) and DPF2 (ubi-d4/REQ), the introns of DPF3 are much larger [30]. This fact, in conjunction with the close similarity in the amino acid sequences of the encoded proteins, has led to the suggestion that the increase in intron size of DPF3 may result in changes in the regulation of DPF3 gene expression. DPF3 has a C2H2 domain and a PHD domain, suggesting a role in the direct binding of DNA and in the assembly of large protein complexes. Functional studies of the d4 gene family have suggested that its members participate in regulation of myeloid programmed cell death through the induction of apoptosis [31]. More recently, DPF3 has been identified by microarray analysis as a transcription factor that may play a role in the pathogenesis of incipient Alzheimer's disease [32]. Publicly available information on DPF3 gene expression is limited; however it has been shown among others to be expressed in both normal and cancerous breast tissue and cell lines [33].
While the current study suggests that variants in DPF3 intron 1 are associated with increased breast cancer risk, the possible mechanism by which these variants predispose to breast cancer are purely speculative. The susceptibility allele might be associated with decreased DPF3 activity through the down-regulation of transcription levels or by negatively impacting RNA splicing. This, in turn, may result in a reduction in the ability of DPF3 to induce apoptosis at the cellular level. Apoptosis is a physiological mechanism of cell death that plays an important role in many disease states, including cancer [34]. Imbalance of pro-apoptotic and anti-apoptotic proteins resulting in altered apoptosis may result in tumor development or poor response to adjuvant therapy. Apoptosis requires de-novo synthesis of mRNA and protein, and alterations of DPF3 may lead to reduced response to apoptotic signaling. Additional experimental studies will be required to precisely elucidate the role of DPF3 in breast cancer etiology and progression.
Conclusion
Our study in women of European ancestry identified significant associations between polymorphisms in DPF3 and breast cancer susceptibility, lymph node metastases, earlier age of onset, and tumor size. While three independent samples from the current study support the observed associations, additional studies are needed to verify the results and to further characterize the gene in order to fully understand the role of DPF3 in the etiology and progression of breast cancer. These and similar still undiscovered variations of small effect may be useful in the assessment of individual breast cancer risks and in the decisions surrounding patient care.
Authors' contributions
CRH drafted the manuscript, participated in the development of the SNP marker set and sequencing, as well as in the study design and data analysis. SK participated in the development of the SNP marker set, supervised the operational aspects of the study, and helped to draft the manuscript. RBR was study project leader and participated in data analysis. RR participated in the study design. GM participated in the development of the SNP marker set. MK and USB designed and collected the German replication sample. LRG designed and collected the Australian replication sample. FE and JR designed and collected the German discovery sample. MRN participated in study design and performed the statistical analyses. AB participated in the study design and had the overall scientific responsibility for the study. All authors read and approved the final manuscript.
Acknowledgements
The authors would like to thank the laboratory teams for their continuous efforts, and the patients and control individuals for their participation in the study.
==== Refs
McPherson K Steel CM Dixon JM ABC of breast diseases. Breast cancer-epidemiology, risk factors, and genetics Bmj 2000 321 624 628 10977847 10.1136/bmj.321.7261.624
Rebbeck TR The contribution of inherited genotype to breast cancer Breast Cancer Res 2002 4 85 89 12052249 10.1186/bcr430
Lancaster JM Wooster R Mangion J Phelan CM Cochran C Gumbs C Seal S Barfoot R Collins N Bignell G Patel S Hamoudi R Larsson C Wiseman RW Berchuck A Iglehart JD Marks JR Ashworth A Stratton MR Futreal PA BRCA2 mutations in primary breast and ovarian cancers Nat Genet 1996 13 238 240 8640235 10.1038/ng0696-238
Futreal PA Liu Q Shattuck-Eidens D Cochran C Harshman K Tavtigian S Bennett LM Haugen-Strano A Swensen J Miki Y BRCA1 mutations in primary breast and ovarian carcinomas Science 1994 266 120 122 7939630
Malkin D Li FP Strong LC Fraumeni JFJ Nelson CE Kim DH Kassel J Gryka MA Bischoff FZ Tainsky MA Germ line p53 mutations in a familial syndrome of breast cancer, sarcomas, and other neoplasms Science 1990 250 1233 1238 1978757
Thorstenson YR Roxas A Kroiss R Jenkins MA Yu KM Bachrich T Muhr D Wayne TL Chu G Davis RW Wagner TM Oefner PJ Contributions of ATM mutations to familial breast and ovarian cancer Cancer Res 2003 63 3325 3333 12810666
Liaw D Marsh DJ Li J Dahia PL Wang SI Zheng Z Bose S Call KM Tsou HC Peacocke M Eng C Parsons R Germline mutations of the PTEN gene in Cowden disease, an inherited breast and thyroid cancer syndrome Nat Genet 1997 16 64 67 9140396 10.1038/ng0597-64
Weber BL Nathanson KL Low penetrance genes associated with increased risk for breast cancer Eur J Cancer 2000 36 1193 1199 10882856 10.1016/S0959-8049(00)00082-4
Pharoah PD Antoniou A Bobrow M Zimmern RL Easton DF Ponder BA Polygenic susceptibility to breast cancer and implications for prevention Nat Genet 2002 31 33 36 11984562 10.1038/ng853
Hodgson SV Morrison PJ Irving M Breast cancer genetics: unsolved questions and open perspectives in an expanding clinical practice Am J Med Genet 2004 129C 56 64 10.1002/ajmg.c.30019
Feigelson HS Coetzee GA Kolonel LN Ross RK Henderson BE A polymorphism in the CYP17 gene increases the risk of breast cancer Cancer Res 1997 57 1063 1065 9067272
Ambrosone CB Freudenheim JL Marshall JR Graham S Vena JE Brasure JR Michalek AM Laughlin R Nemoto T Shields PG The association of polymorphic N-acetyltransferase (NAT2) with breast cancer risk Ann N Y Acad Sci 1995 768 250 252 8526358
Helzlsouer KJ Selmin O Huang HY Strickland PT Hoffman S Alberg AJ Watson M Comstock GW Bell D Association between glutathione S-transferase M1, P1, and T1 genetic polymorphisms and development of breast cancer J Natl Cancer Inst 1998 90 512 518 9539246 10.1093/jnci/90.7.512
Vahteristo P Bartkova J Eerola H Syrjakoski K Ojala S Kilpivaara O Tamminen A Kononen J Aittomaki K Heikkila P Holli K Blomqvist C Bartek J Kallioniemi OP Nevanlinna H A CHEK2 genetic variant contributing to a substantial fraction of familial breast cancer Am J Hum Genet 2002 71 432 438 12094328 10.1086/341943
Meijers-Heijboer H van den Ouweland A Klijn J Wasielewski M de Snoo A Oldenburg R Hollestelle A Houben M Crepin E van Veghel-Plandsoen M Elstrodt F van Duijn C Bartels C Meijers C Schutte M McGuffog L Thompson D Easton D Sodha N Seal S Barfoot R Mangion J Chang-Claude J Eccles D Eeles R Evans DG Houlston R Murday V Narod S Peretz T Peto J Phelan C Zhang HX Szabo C Devilee P Goldgar D Futreal PA Nathanson KL Weber B Rahman N Stratton MR Low-penetrance susceptibility to breast cancer due to CHEK2(*)1100delC in noncarriers of BRCA1 or BRCA2 mutations Nat Genet 2002 31 55 59 11967536 10.1038/ng879
The CHEK2 Breast Cancer Case-Control Consortium CHEK2*1100delC and susceptibility to breast cancer: a collaborative analysis involving 10,860 breast cancer cases and 9,065 controls from 10 studies Am J Hum Genet 2004 74 1175 1182 15122511 10.1086/421251
Kammerer S Roth RB Reneland R Marnellos G Hoyal CR Markward NJ Ebner F Kiechle M Schwarz-Boeger U Griffiths LR Ulbrich C Chrobok K Forster G Praetorius GM Meyer P Rehbock J Cantor CR Nelson MR Braun A Large-scale association study identifies ICAM gene region as breast and prostate cancer susceptibility locus Cancer Res 2004 64 8906 10 15604251
Kammerer S Roth RB Hoyal CR Reneland R Marnellos G Kiechle M Schwarz-Boeger U Griffiths LR Ebner F Rehbock J Cantor CR Nelson MR Braun A Association of the NuMA region on chromosome 11q13 with breast cancer susceptibility Proc Natl Acad Sci U S A 2005 102 2004 2009 15684076 10.1073/pnas.0409806102
Nelson MR Marnellos G Kammerer S Hoyal CR Shi MM Cantor CR Braun A Large-scale validation of single nucleotide polymorphisms in gene regions Genome Res 2004 14 1664 1668 15289484 10.1101/gr.2421604
The NCBI dbSNP database
NCBI Mapview
Buetow KH Edmonson M MacDonald R Clifford R Yip P Kelley J Little DP Strausberg R Koester H Cantor CR Braun A High-throughput development and characterization of a genomewide collection of gene-based single nucleotide polymorphism markers by chip-based matrix-assisted laser desorption/ionization time-of-flight mass spectrometry Proc Natl Acad Sci U S A 2001 98 581 584 11136232 10.1073/pnas.021506298
Bansal A van den Boom D Kammerer S Honisch C Adam G Cantor CR Kleyn P Braun A Association testing by DNA pooling: an effective initial screen Proc Natl Acad Sci U S A 2002 99 16871 16874 12475937 10.1073/pnas.262671399
Stanssens P Zabeau M Meersseman G Remes G Gansemans Y Storm N Hartmer R Honisch C Rodi CP Bocker S van den Boom D High-throughput MALDI-TOF discovery of genomic sequence polymorphisms Genome Res 2004 14 126 133 14707174 10.1101/gr.1692304
Barratt BJ Payne F Rance HE Nutland S Todd JA Clayton DG Identification of the sources of error in allele frequency estimations from pooled DNA indicates an optimal experimental design Ann Hum Genet 2002 66 393 405 12485472 10.1046/j.1469-1809.2002.00125.x
DerSimonian R Laird N Meta-analysis in clinical trials Control Clin Trials 1986 7 177 188 3802833 10.1016/0197-2456(86)90046-2
Stephens M Donnelly P A comparison of bayesian methods for haplotype reconstruction from population genotype data Am J Hum Genet 2003 73 1162 1169 14574645 10.1086/379378
Schaid DJ Rowland CM Tines DE Jacobson RM Poland GA Score tests for association between traits and haplotypes when linkage phase is ambiguous Am J Hum Genet 2002 70 425 434 11791212 10.1086/338688
NCBI dbSNP = rs1990440
Ninkina NN Mertsalov IB Kulikova DA Alimova-Kost MV Simonova OB Korochkin LI Kiselev SL Buchman VL Cerd4, third member of the d4 gene family: expression and organization of genomic locus Mamm Genome 2001 12 862 866 11845289 10.1007/s00335-001-3039-1
Gabig TG Mantel PL Rosli R Crean CD Requiem: a novel zinc finger gene essential for apoptosis in myeloid cells J Biol Chem 1994 269 29515 29519 7961935
Blalock EM Geddes JW Chen KC Porter NM Markesbery WR Landfield PW Incipient Alzheimer's disease: microarray correlation analyses reveal major transcriptional and tumor suppressor responses Proc Natl Acad Sci U S A 2004 101 2173 2178 14769913 10.1073/pnas.0308512100
The Cancer Genome Anatomy Project
Thompson CB Apoptosis in the pathogenesis and treatment of disease Science 1995 267 1456 1462 7878464
|
16109180
|
PMC1215495
|
CC BY
|
2021-01-04 16:39:20
|
no
|
J Carcinog. 2005 Aug 19; 4:13
|
utf-8
|
J Carcinog
| 2,005 |
10.1186/1477-3163-4-13
|
oa_comm
|
==== Front
Cardiovasc DiabetolCardiovascular Diabetology1475-2840BioMed Central London 1475-2840-4-111603364710.1186/1475-2840-4-11MethodologyHypertension control: results from the Diabetes Care Program of Nova Scotia registry and impact of changing clinical practice guidelines Russell Cory [email protected] Peggy [email protected] Sonia [email protected] Ingrid [email protected] George [email protected] Department of Community Health and Epidemiology, Dalhousie University, Halifax, Canada2 Diabetes Care Program of Nova Scotia, Halifax, Canada3 College of Pharmacy, Dalhousie University, Halifax, Canada2005 20 7 2005 4 11 11 24 4 2005 20 7 2005 Copyright © 2005 Russell et al; licensee BioMed Central Ltd.2005Russell et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
The objective of this study was to determine the rate of blood pressure control according to 4 sets of Canadian guidelines published over a decade in patients with diabetes mellitus attending Diabetes Centres in the province of Nova Scotia.
Methods
One hundred randomly selected charts from each of 13 Diabetes Centres audited between 1997 and 2001 were extracted from the Diabetes Care Program of Nova Scotia Registry. Multivariate logistic regression analyses examined the relationship between individual characteristics and self-reported antihypertensive use. Included were 1132 adults, mean age 63 years (48% male), with 9 years mean time since diagnosis of diabetes.
Results
According to the 1992 guidelines, 63% of the patients and according to the 2003 guidelines, 84% of patients were above target blood pressure or receiving antihypertensive medications. Forty-seven percent of patients are considered to be hypertensive and not on treatment according to 2003 guidelines. The results of the multivariate analyses showed that the only factors independently associated with anti-hypertensive use was oral anti-hyperglycemic use.
Conclusion
Hypertension is an additional risk factor in those with diabetes mellitus for macrovascular and microvascular complications. The health and budgetary impacts of addressing the treatment gap need to be further explored.
==== Body
Background
Achievement of target blood pressures in hypertensive patients is often difficult. In Halifax County, Nova Scotia, 57% of men and 42% of women with hypertension were not well controlled[1]. Adequate blood pressure control is of particular concern in patients with diabetes as hypertension increases morbidity and mortality associated with stroke and cardiovascular disease[2,3], as well as microvascular complications such as retinopathy and nephropathy[4]. Cardiovascular disease rates have been shown to be 2–4 times higher in diabetes than in matched non-diabetic populations[5,6].
A Canadian study reported 43% of people (age 18–74 years) had an optimal blood pressure (<120/80 mmHg), and of those with a diagnosis of hypertension, only 13% were below target (defined as 140/90 mm Hg). In this study, about 50% of patients with diabetes were hypertensive, and of these only 9% were under control[7]. An internal review at the Diabetes Care Program of Nova Scotia (DCPNS) from 1997–2001 showed that only 27.5% of a random selection of patients attending Diabetes Centres fell within the recommended target blood pressure for people with a diagnosis of diabetes (< 130/85 mm Hg) [8].
The United Kingdom Prospective Diabetes Study (UKPDS) emphasized the need for adequate blood pressure control in type 2 diabetes. The evidence suggested that good blood pressure control may be as important if not more important than blood glucose control in reduction of the cardiovascular complications[3,9,10]. Further, adequate blood pressure control in the UKPDS decreased risk for multiple diabetes end-points: 32% in deaths related to diabetes; 44% decreased risk of stroke; and a 34% reduction in risk for all macrovascular diseases, as well as a significantly decreased risk for other complications. [4] Clinical trials and epidemiologic studies have suggested the target blood pressure goal of <130/80 mmHg[11-14].
The treatment of hypertension in patients with diabetes has changed over the last decade. Studies and Clinical Practice Guidelines for the management of hypertension in patients with diabetes suggest lower blood pressure targets for diagnosis and control than for the general population[11,15-17].
This study determined the degree of blood pressure control in patients with diabetes according to four sets of Canadian Clinical Practice Guidelines published between 1992 and 2003[2,18-20] and described demographic and treatment variables associated with antihypertensive treatment.
Methods
The cohort was selected as part of a DCPNS (Diabetes Care Program of Nova Scotia) internal audit of approximately 100 records from each of 13 Diabetes Centres between 1997 and 2001. All patients were referred to the Diabetes Centre following a diagnosis of diabetes by a physician. Information gathered included: age, gender, weight, blood pressure, duration of diabetes, serum creatinine, urinary protein and specific antihypertensive treatment regimens. The mercury sphygmomanometer was used for blood pressure measurement and recorded by nurses and averaged over all visits for all individuals. The nurses were aware of the correct procedure for obtaining a blood pressure measurement, and performed the procedure regularly. Eligibility criteria for the cohort included being a non-pregnant adult over the age of 19; a diagnosis of type 1 or 2 diabetes; a visit to the centre within 12 months of the audit date; and at least 15 months of followup.
The final cohort included 1132 subjects. The population consists of both genders (48% male), with an average age of 63 (SD 12). Over 95% of patients had type 2 diabetes, and the average length of time since diagnosis was 9.3 (SD 8) years previous. Average scores were obtained for most tests and attributes. Kidney function was estimated using both Couchoud cutpoints and the Cockroft-Gault formula [21,22] Drug information was reclassified using the WHO Anatomical Therapeutic Chemical (ATC) categories[23]. Prevalence of hypertension and trends in Clinical Practice Guidelines over time were determined.
Guidelines used for analysis included the following: 1992 Clinical Practice Guidelines for Treatment of Diabetes Mellitus – hypertension subcategory [18]; 1998 Clinical Practice Guidelines for the Management of Diabetes in Canada – hypertension subcategory [2]; 1999 Canadian Hypertension Society Recommendations for the Management of Hypertension – diabetes subcategory [19]; 2003 Canadian Hypertension Society Recommendations for the Management of Hypertension – diabetes subcategory[20].
Hypertension was defined using anti-hypertensive drug use and blood pressure records. Patients with any antihypertensive drug use and/or average blood pressure above the guideline cutpoints (systolic, diastolic, or both) were designated to be hypertensive. Rates and risk factors for hypertension were calculated for each specific guideline. Logistic regression was performed to determine predictors of antihypertensive treatment.
SAS version 8.2 (SAS Institute, Cary, NC, USA, 2001) was used for analysis.
Results
The use of the 2003 guidelines (target blood pressure: systolic < 130 mmHg; diastolic < 80 mmHg) increased the percentage of patients not meeting target to 84% from 63% using 1992 guidelines (target blood pressure: systolic < 140 mmHg; diastolic <90 mmHg). Those considered to be hypertensive and not on treatment increased to 47% using the 2003 guidelines from 26% with the 1992 guidelines. (Figure 1); Clinical Practice Guidelines Effects on Nova Scotia Patients with Diabetes Classified as Hypertensive; Blank cells indicate that the category was not applicable for that guideline. The "Isolated Systolic" category in the 1999 and 2003 guidelines is used synonymously with the "Elderly" category used in the 1998 guidelines for data display purposes) Among all potential predictors of antihypertensive drug treatment related to the 2003 guidelines included in our database, only the patients receiving oral antihyperglycemics with or without insulin were more likely to be treated. (Table 1: Predictors of Treatment among Patients with Hypertension, 2003 Guidelines)
Figure 1 The effect of clinical practice guidlines changes on the percentage of Nova Scotia patients with diabetes classified as hypertensive.
Table 1 Predictors of Treatment among Patients with Hypertension, 2003 Guidelines
Variable Crude Odds Ratio (95% CI) Adjusted Odds Ratio (95% CI)†
Age
≤50 1.00
51–60 1.17 (0.70–1.96)
61–70 1.47 (0.91–2.37)
≥71 1.42 (0.88–2.28)
Gender
Male 1.00
Female 1.26 (0.97–1.63)
Diabetes
Type I 1.55 (0.75–3.18)
Type II 1.00
Treatment *
Diet 1.00 1.00
Oral meds 1.89 (1.31–2.74) 1.89 (1.31–2.74)
Insulin 1.04 (0.77–1.41) 1.04 (0.77–1.41)
Insulin + Oral 2.48 (1.13–5.43) 2.48 (1.13–5.43)
Kidney Disease **
None 1.00
Mild 1.32 (0.90–1.94)
Moderate 0.97 (0.65–1.45)
Severe 1.23 (0.34–4.50)
Failure not valid
Years Since Diagnosis
≤5 1.00
6–10 0.86 (0.62–1.19)
≥11 1.32 (0.98–1.78)
Years Since Referral
≤3 1.00
4–7 0.84 (0.61–1.17)
≥8 0.94 (0.67–1.31)
Weight (KG)
≤55 1.00
56–67 0.78 (0.35–1.71)
68–90 0.87 (0.42–1.80)
≥91 0.99 (0.47–2.06)
* Oral meds mean taking only oral antihyperglycemic medication; Insulin + Oral means patients who are taking insulin plus oral antihyperglycemic medication.
** Presence and stage of kidney disease as measured using the Cockroft-Gault formula.
† Treatment was the only significant predictor, and therefore is not adjusted.
Discussion
Many Nova Scotia patients with diabetes mellitus had uncontrolled blood pressure and were not receiving antihypertensive medication. Achieving control of high blood pressure may be more important for long-term outcomes than glycemic control[3,10]. The rates were similar to other studies where 54–58% were above target blood pressure and 22–28% were not receiving antihypertensive treatment [24,25]. These populations have a decreased prevalence of hypertension, yet a higher rate of treatment in those affected.
Changing Clinical Practice Guidelines affect the criteria for diagnosis, the treatment targets, the population to be treated and the type of treatment. Many patients with diabetes mellitus previously considered to be normotensive are now above the defined cutpoints. Adherence to the newer guidelines would result in more patients being treated and increased drug expenditures, but may lead to decreased overall health service utilization and improved patient outcomes. Further work will be needed to determine the rate of adoption of the newer guidelines and the facilitators and barriers to adoption. For example, it is unclear how well guidelines apply to patients above age 85 or the frail elderly.
This study is a population-based study in the real world. The study included cardiovascular risk factors, and documentation of kidney disease unlike many survey reports[26]. Drug data was recorded by patient self-report at each visit by Diabetes Centre personnel. The quality of the DCPNS Registry evolved over time, particularly the details related to antihypertensive drug therapy. Self-report has had good concordance with pharmacy claims data[27,28]. We were unable to determine how patients used the medications, if antihypertensive medications were used for hypertension or for another disease, any contraindications to therapy, the comorbid conditions, target organ damage, or response to previous antihypertensive therapy. Blood pressure measurements were part of routine care. Family history of cardiovascular disease, smoking, and lifestyle factors and the level of blood pressure at which treatment was started were not determined.
Conclusion
Many patients with diabetes mellitus and hypertension were not treated according to guidelines, with 47% of the patients meeting the 2003 guidelines definitions of hypertension not being treated with antihypertensive medications. By reducing the cutpoints for defining hypertension, the proportion of people affected increased substantially. Specific risk factors determined may aid in identifying patients at high-risk for inadequate treatment. Patient and provider education, public health approaches, and health system changes are needed to address these issues. Further work is needed to determine the reasons for lack of control, approaches to improve control and long-term patient outcomes, and the budget impact and cost effectiveness of using the 2003 guidelines.
Competing Interests
The author(s) declare that they have no competing interests.
Authors' Contributions
Cory Russell was involved with the design of the study, analysis and interpretation of the data, drafting and editing the manuscript, and gave final approval of the manuscript. All other authors were involved with the design of the study, interpretation of the data, editing and revising the manuscript, and gave final approval of the manuscript to be published.
Financial Support
Cory Russell was funded by the Drug Use Management and Policy Residency, a summer studentship that aims to build student and faculty understanding about how the creation of knowledge and dissemination of evidence is used by decision makers for drug used management and policy analysis. The Canadian Health Services Research Foundation, Canadian Institutes of Health Research and the Nova Scotia Health Research Foundation support this residency.
Dr. Ingrid Sketris holds a Chair funded by the Canadian Health Services Research Foundation(CHSRF)/Canadian Institutes for Health Research (CIHR) Chair in Health Services Research, co-sponsored by the Nova Scotia Health Research Foundation (NSHRF).
Declaration of Business Interests/Disclaimer
The data used in this research was made available by the Diabetes Care Program of Nova Scotia. Any opinions expressed by the authors do not necessarily reflect the opinion of DCPNS.
Presented in part at the 2004 Canadian Association for Population Therapeutics conference, June 6–8 2004 in Winnipeg, Mb
Note
The Effect of Clinical Practice Guidelines Changes on the Percentage of Nova Scotia Patients with Diabetes Classified as Hypertensive
Acknowledgements
The authors wish to thank Peggy Dunbar and Sonia Salisbury for participating as a preceptor in the Drug Use Management and Policy Residency. We are grateful to Charmaine Cooke, Jocelyn Leclerc, and Amanda Hayden, for their useful insights, support and help in the preparation of this manuscript.
==== Refs
Wolf HK Andreou P Bata IR Comeau DG Gregor RD Kephart G Trends in the prevalence and treatment of hypertension in Halifax County from 1985 to 1995 CMAJ 1999 161 699 704 10513276
Meltzer S Leiter L Daneman D Gerstein HC Lau D Ludwig S 1998 clinical practice guidelines for the management of diabetes in Canada CMAJ 1998 159 Canadian Diabetes Association S1 S29 9834731
Johansen OE Birkeland KI Preventing Macrovascular Disease in patient with Type 2 Diabetes Mellitus Am J Cardiovasc Drugs 2003 3 283 297 14728081
The UK Prospective Diabetes Study Group Tight blood pressure control and risk of macrovascular and microvascular complications in type 2 diabetes BMJ 1998 317 703 13 9732337
Haffner SM The importance of hyperglycemia in the nonfasting state to the development of cardiovascular disease Endocr Rev 1998 19 583 92 9793758 10.1210/er.19.5.583
Kannel WB Mcgee DL Diabetes and cardiovascular risk factors: the Framingham study Circulation 1979 59 8 13 758126
Joffres MR Hamet P MacLean DR L'italien GJ Fodor G Distribution of blood pressure and hypertension in Canada and the United States AJH 2001 14 1099 1105 11724207 10.1016/S0895-7061(01)02211-7
Dunbar, Peggy MEd, PDT, Coordinator for the Diabetes Care Program of Nova Scotia Personal communication 2003
Hansson L Zanchetti A Carruthers SG Effects of intensive blood-pressure lowering and low-dose aspirin in patients with hypertension: principal results of the Hypertension Optimal Treatment (HOT) randomised trial Lancet 1998 351 1755 1762 9635947 10.1016/S0140-6736(98)04311-6
Mogensen CE Combined high blood pressure and glucose in type 2 diabetes: double jeopardy BMJ 1998 317 693 4 9732334
Chobanian AV Bakris GL Black HR Cushman WC Green LA Izzo JL JrJones DW Materson BJ Oparil S Wright JT JrRoccella EJ the National Heart, Lung, and Blood Institute Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure, the National High Blood Pressure Coordinating Committee The Seventh Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure: the JNC 7 Report JAMA 2003 289 2560 2572 12748199 10.1001/jama.289.19.2560
Adler AI Stratton IM Neil HA Yudkin JS Mathhew DR Cull CA Wright AD Turner RC Holman RR Association of systolic blood pressure with macrovascular and microvascular complications of type 2 diabetes (UKPDS 36): prospective observational study BMJ 2000 321 413 419 10.1136/bmj.321.7258.412
Lewington S Clarke R Qizilbash N Peto R Collins R Age-specific relevance of usual blood pressure to vascular mortality: a meta-analysis of individual data for one million adults in 61 prospective studies Lancet 2002 360 1903 1913 12493255 10.1016/S0140-6736(02)11911-8
Stamler J Vaccaro O Neaton JD Wentworth D Diabetes, other risk factors, and 12-yr cardiovascular mortality for men screened in the multiple risk factor intervention trial Diabetes Care 1993 16 434 444 8432214
Carruthers SG Hypertension in Gray J Therapeutic Choices Canadian Pharmacists Association 2000 180 190
American Diabetes Association Standards of Medical Care in Diabetes Diabetes Care 2004 S15 35 14693923
Canadian Diabetes Association Clinical Practice Guidelines ExpertCommittee Canadian Diabetes Association Clinical Practice Guidelines for the Prevention and Management of Diabetes in Canada Can J Diabetes 2003 27 S1 S152
Expert Committee of the Canadian Diabetes Advisory Board Clinical practice guidelines for treatment of diabetes mellitus CMAJ 1992 147 697 712 1521215
Feldman RD Canadian Recommendations for the Management of Hypertension CMAJ 1999 161 S1 17 10624417
Canadian Hypertension Education Program,
Cockroft DW Gault MH Prediction of creatinine clearance from serum creatinine Nephron 1976 16 31 41 1244564
Couchoud C Pozet N Labeeuw M Pouteil-Noble C Screening early renal failure: Cut-off values for serum creatinine as an indicator of renal impairment Kidney Int 1999 55 1878 84 10231450 10.1046/j.1523-1755.1999.00411.x
World Health Organization Collaborating Centre for Drug Statistics Methodology ATC Index with DDDs, 2002 2002 Oslo, Norway: World Health Organization
Toth EL Majumdar SR Guirguis LM Lewanczuk RZ Lee TK Johnson JA Compliance with Clinical Practice Guidelines for type 2 diabetes in rural patients: treatment gaps and opportunities for improvement Pharmacotherapy 2003 23 659 65 12741441 10.1592/phco.23.5.659.32203
Grant RW Cagliero E Murphy-Sheehy P Singer DE Nathan DM Meigs JB Comparison of hyperglycemia, hypertension, and hypercholesterolemia management in patients with Type 2 diabetes Am J Med 2002 112 603 9 12034408 10.1016/S0002-9343(02)01103-8
Wolf-Maier K Cooper RS Kramer H Banegas JR Giampaoli S Joffres MR Poulter N Primatesta P Stegmayr B Thamm M Hypertension treatment and control in five European countries, Canada, and the United States Hypertension 2004 43 10 17 14638619 10.1161/01.HYP.0000103630.72812.10
Kwon A Bungay KM Pei Y Rogers WH Wilson IB Zhou Q Adler DA Antidepressant use: concordance between self-report and claims records Med Care 2003 41 368 74 12618640 10.1097/00005650-200303000-00005
Saunders K Simon G Bush T Grothaus L Assessing the feasibility of using computerized pharmacy refill data to monitor antidepressant treatment on a population basis: a comparison of automated and self-report data J Clin Epidemiol 1998 51 883 90 9762882 10.1016/S0895-4356(98)00053-5
|
16033647
|
PMC1215496
|
CC BY
|
2021-01-04 16:25:04
|
no
|
Cardiovasc Diabetol. 2005 Jul 20; 4:11
|
utf-8
|
Cardiovasc Diabetol
| 2,005 |
10.1186/1475-2840-4-11
|
oa_comm
|
==== Front
Cardiovasc UltrasoundCardiovascular Ultrasound1476-7120BioMed Central London 1476-7120-3-231612902810.1186/1476-7120-3-23ResearchRight ventricular dyssynchrony in patients with pulmonary hypertension is associated with disease severity and functional class López-Candales Angel [email protected] Kaoru [email protected] Navin [email protected] Matthew [email protected] Srinivas [email protected] John [email protected] Kathy [email protected] Cardiovascular Institute at the University of Pittsburgh Medical Center, Pittsburgh, PA, USA2005 29 8 2005 3 23 23 1 8 2005 29 8 2005 Copyright © 2005 López-Candales et al; licensee BioMed Central Ltd.2005López-Candales et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Abnormalities in right ventricular function are known to occur in patients with pulmonary arterial hypertension.
Objective
Test the hypothesis that chronic elevation in pulmonary artery systolic pressure delays mechanical activation of the right ventricle, termed dyssynchrony, and is associated with both symptoms and right ventricular dysfunction.
Methods
Fifty-two patients (mean age 46 ± 15 years, 24 patients with chronic pulmonary hypertension) were prospectively evaluated using several echocardiographic parameters to assess right ventricular size and function. In addition, tissue Doppler imaging was also obtained to assess longitudinal strain of the right ventricular wall, interventricular septum, and lateral wall of the left ventricle and examined with regards to right ventricular size and function as well as clinical variables.
Results
In this study, patients with chronic pulmonary hypertension had statistically different right ventricular fractional area change (35 ± 13 percent), right ventricular end-systolic area (21 ± 10 cm2), right ventricular Myocardial Performance Index (0.72 ± 0.34), and Eccentricity Index (1.34 ± 0.37) than individuals without pulmonary hypertension (51 ± 5 percent, 9 ± 2 cm2, 0.27 ± 0.09, and 0.97 ± 0.06, p < 0.005, respectively). Furthermore, peak longitudinal right ventricular wall strain in chronic pulmonary hypertension was also different -20.8 ± 9.0 percent versus -28.0 ± 4.1 percent, p < 0.01). Right ventricular dyssynchrony correlated very well with right ventricular end-systolic area (r = 0.79, p < 0.001) and Eccentricity Index (r = 0.83, p < 0.001). Furthermore, right ventricular dyssynchrony correlates with pulmonary hypertension severity index (p < 0.0001), World Health Organization class (p < 0.0001), and number of hospitalizations (p < 0.0001).
Conclusion
Lower peak longitudinal right ventricular wall strain and significantly delayed time-to-peak strain values, consistent with right ventricular dyssynchrony, were found in a small heterogeneous group of patients with chronic pulmonary hypertension when compared to individuals without pulmonary hypertension. Furthermore, right ventricular dyssynchrony was associated with disease severity and compromised functional class.
Dyssynchronyright ventricleoutcomespulmonary hypertensionstrain imagingtissue Doppler imaging
==== Body
Background
Right ventricular systolic dysfunction has been identified as a key element in determining prognosis of patients afflicted with chronic pulmonary hypertension [1-7]. Although echocardiography has proved invaluable to noninvasively assess pulmonary artery pressures; evaluation of right ventricular size and systolic function by echocardiography is somewhat more difficult, largely because of the complex RV anatomy that limits its evaluation [8-14]. Therefore, identification of early right ventricular dysfunction is of outmost clinical importance since as many as two-thirds of the deaths in patients with chronic pulmonary hypertension may be attributed to right ventricular failure [15-19].
The recent introduction of strain and strain rate echocardiography using tissue Doppler imaging (TDI) has provided an objective means to quantify global and regional left ventricular function with improved accuracy and greater reproducibility than conventional echocardiography [20-23]. We have recently reported that TDI is also useful in identifying right ventricular free wall mechanical delay in patients with chronic pulmonary hypertension [24]. However, its clinical significance and potential relevance remains to be determined. We therefore designed this study to answer two critical questions. First, determine if right ventricular dyssynchrony in patients with chronic pulmonary hypertension is associated with indices of disease severity and impaired functional class when compared to individuals without pulmonary hypertension. Second, determine if TDI can identify right ventricular dyssynchrony in patients with chronic pulmonary hypertension before any visible abnormalities of right ventricular size or function are apparent by using routine transthoracic echocardiography.
Methods
Study population
Fifty-two patients (mean age 46 ± 15 years, 22 males) who were referred to our echocardiographic laboratory underwent a complete echocardiographic examination. In the population studied, 24 patients had chronic pulmonary hypertension, as determined by echocardiography [25], including 8 patients with parenchymal lung disease, 4 patients with idiopathic pulmonary hypertension, 5 patients with chronic thrombo-embolic pulmonary hypertension, 3 patients with porto-pulmonary hypertension, and 4 patients with connective tissue disorder. In these patients with chronic pulmonary hypertension, all electronic hospital records were retrospectively reviewed to assess how often these patients were hospitalized, seen in the emergency room visits, or had evidence of clinical deterioration for a period of 6-months prior to the echocardiogram.
Patients with an irregular heart rhythm such as atrial fibrillation, history of significant coronary artery disease, previous myocardial infarction, resting wall motion abnormalities, cardiomyopathy, abnormal left ventricular systolic function, valvular heart disease or the presence of a pacer or defibrillator wire in the right ventricle were all excluded.
The Institutional Review Board of the University of Pittsburgh Medical Center approved the study and all patients gave informed consent.
Standard echocardiography
All patients underwent a complete transthoracic echocardiographic study including two-dimensional, color flow and spectral Doppler as well as tissue strain imaging using a GE-Vingmed Vivid 7 system (GE Vingmed Ultrasound, Horten, Norway). Standard two-dimensional echocardiographic evaluation of RV size and function was performed as routinely [26]. In addition, right ventricular end-diastolic and end-systolic areas were measured from the apical 4-chamber view to calculate right ventricular fractional area change. Eccentricity Index using the mid-ventricular short axis image at the level of the papillary muscles in both systole and diastole and right ventricular Myocardial Performance Index were calculated as previously reported [19,26,27].
Pulmonary artery systolic pressures were estimated using the approach of calculating the systolic pressure gradient between right ventricle and right atrium by the maximum velocity of the tricuspid regurgitant jet using the modified Bernoulli equation and then adding to this value an estimated right atrial pressures based on both the size of the inferior vena cava and the change in caliber of this vessel with respiration [25,28].
To verify that the severity of tricuspid regurgitation was reliable, we also determined the vena contracta width as previously documented by Tribouilloy et al [29]; specifically the position of the transducer was modified to optimize visualization of the flow convergence region and the regurgitant flow proximal and distal to the tricuspid valve, the aliasing velocity ranged from 46 to 96 cm/s and the narrowest neck of the regurgitant flow just distal to the flow convergence region was measured in mid systole by an observer unaware of the clinical examination.
Tissue Doppler
Color-coded tissue Doppler cine loops were obtained as routinely performed in our laboratory from 3 beats obtained from apical 4-chamber views at the depths of 14 ± 2 cm with pulse repetition frequency set at 1 kHz, Nyquist velocity range ± 16 cm/sec and frame rates 99 ± 9 Hz [20,30,31]. Initial length for longitudinal strain measurement was set at 12 mm and the regions of interest with a length of 20 to 24 mm and width of 6 to 8 mm were placed in the basal and mid segments of the right ventricular lateral free wall (RVw), inter-ventricular septum (IS), and left ventricular lateral (LVL) wall to measure the peak longitudinal systolic strain and time-to-peak strain from the onset of Q-wave on the electrocardiogram and shown as mean values for both basal and mid segments for the corresponding right and left ventricular walls. Right and left ventricular dyssynchrony was determined as the difference in time-to-peak strain from IS to RVw and from IS to LVL. We also determine difference in right to left ventricular synchrony as the differences in time-to-peak strain from RVF to LVL. Finally, longitudinal right ventricular annular displacement was also measured by placing transducer region of interest with a 7 by 7 mm sample volume in the junction of the right ventricular free wall and the tricuspid valve.
Statistical analysis
All echocardiographic parameters were calculated using the commercially available software EchoPAC PC version 3.00 (GE Vingmed Ultrasound) and determined by a single observer. All intervals were corrected for heart rate (corrected interval = measured interval/ (RRinterval)1/2) [30,31]. Group data (mean ± SD) were compared using the 2-tailed Student's t-test for paired and unpaired data, respectively. Linear regression analysis was used to examine relations between various dependent variables. Univariate analysis of right ventricular dyssynchrony to clinical and echocardiographic variables was also performed. To assess if there was any correlation between right ventricular dyssynchrony and clinical variables, we gathered information regarding World Health Organization (WHO) symptom class, systolic pulmonary artery pressure (SPAP) severity index, and hospitalizations due to pulmonary hypertension or heart failure symptoms over a 12-month period prior to the echocardiographic examination. We determined SPAP severity index as follows: A pulmonary hypertension severity of 1 corresponded to a SPAP 36–50; 2 as a SPAP 51–75; and 3 as a SPAP > 75. P-values of less than .05 were considered to be statistically significant.
Results
Echocardiographic results obtained in all 52 patients are summarized in Table 1. Although the mean values for all standard echocardiographic parameters to assess right ventricular size and function were found within normal limits; there was a wide spread of values. In the case of right ventricular end-diastolic areas the mean value obtained for the entire population studied was 26 ± 10 cm2, with a minimum value of 12 and a maximum value of 59 cm2. Similarly, the mean value for right ventricular end-systolic areas was 16 ± 10 cm2 but ranged from 5 to 45 cm2; consequently the calculated mean right ventricular fractional area change was also found within normal (43 ± 13), with a minimum value of 13 and a maximum value of 65. With regards to Eccentricity Index, the mean value was 1.16 ± 0.32 and included values from 0.84 to 3.28. Similarly, the Myocardial Performance Index was 0.53 ± 0.35 but ranged from 0.13 to 1.26. For the whole population studied, the mean pulmonary artery systolic pressure was 55 ± 33 mmHg, ranging from 15 to 116 mmHg. Since most healthy subjects only have a trivial amount of tricuspid regurgitation, estimation of peak pulmonary artery systolic pressures, in these individuals was not possible. However, the mean vena contracta width in patients with chronic pulmonary hypertension was 5.7 ± 2.2 mm.
Table 1 Standard two-dimensional echocardiographic and Doppler data.
Parameters Population Studied
Standard Echocardiography
RV end diastolic area (cm2) 26 ± 10
RV end systolic area (cm2) 16 ± 10
RV fractional area change (%) 43 ± 13
Eccentricity index 1.16 ± 0.32
Myocardial performance index 0.53 ± 0.35
Pulmonary artery systolic pressure (mmHg) 53 ± 33
Left ventricular systolic function (%) 57 ± 5
It is only when we analyze this echocardiographic data according to the presence or absence of chronic pulmonary hypertension that certain correlations become apparent. As seen in Figure 1A, right ventricular fractional area change is significantly higher in individuals without pulmonary hypertension when compared to patients with chronic pulmonary hypertension. A significantly lower Eccentricity is noted among individuals without pulmonary hypertension when compared to patients with chronic pulmonary hypertension as seen in Figure 1B. Similarly, a lower RV Myocardial Performance Index (Figure 1C) is also noted among normal individuals when compared to patients with chronic pulmonary hypertension.
Figure 1 (A) Right ventricular fractional area change, expressed as a percentage, is significantly higher in individuals with no pulmonary hypertension (no PHTN), as demonstrated by the black columns, than in patients with chronic pulmonary hypertension (PHTN), as seen in the grey columns. (B) A significantly lower Eccentricity Index, as expected, is noted in individuals with no pulmonary hypertension when compared to patients with chronic pulmonary hypertension. (C) Similarly, a lower right ventricular Myocardial Performance Index is noted among normal individuals when compared to patients with chronic pulmonary hypertension.
A complete TDI examination was also obtained in all 52 patients and the results are summarized in Table 2. To better understand the echocardiographic findings based on TDI data we compared time differences between interventricular septum and the RVw, expressed in milliseconds (ms), against the right ventricular end-diastolic area and noted a very strong correlation (r = 0.70, p < 0.001) as shown in Figure 2A. An even better correlation was noted in the time difference between interventricular septum and the RVw activation when compared to right ventricular end-systolic area (r = 0.79, p < 0.001) as shown in Figure 2B. On Figure 3, a very strong correlation is noted between Eccentricity Index and the time difference between interventricular septum and the RVw (r = 0.83, p < 0.001). Finally, a better representation of the statistically significant differences between individuals without pulmonary hypertension and patients with chronic pulmonary hypertension is seen in both Figure 4 and 5. In Figure 4, we demonstrate the differences in peak strain, expressed in percent, for the RVw, IS, and LVL. It is important to note the statistically significant difference in peak strain generation between individuals without pulmonary hypertension and patients with chronic pulmonary hypertension with regards to both IS (Figure 4A) and RVw Figure 4C). In Figure 5, we demonstrate the time differences in mechanical activation, expressed in ms, between IS and LVL, IS and RVw, and LVL and RVw. A statistically significant difference in time difference of mechanical activation between individuals without pulmonary hypertension and patients with chronic pulmonary hypertension was only seen for both IS and RVw (Figure 5B) and LVL and RVw (Figure 5C), suggestive of the presence of right ventricular dyssynchrony. It is also demonstrated in Figure 5A, the presence of normal synchrony in the left ventricle.
Table 2 Tissue Doppler imaging data.
Parameters Population Studied
Tissue Doppler Imaging
Peak strain (%)
RVw -24 ± 8
IS -16 ± 5
LVL -14 ± 5
Time to peak strain (ms)
RVw 417 ± 62
IS 364 ± 39
LVL 380 ± 47
Time difference (ms)
IS – RVw 53 ± 66
IS – LVL 16 ± 37
RVw – LVL 37 ± 65
Figure 2 (A) A very strong correlation is shown between time differences between interventricular septum and the RVw, measured in ms, against the right ventricular end-diastolic area (R = 0.70, p < 0.001). (B) This graph shows the correlation between the time difference between interventricular septum and RVw activation and right ventricular end-systolic area (R = 0.79, p < 0.001).
Figure 3 A very strong correlation is seen between Eccentricity Index (EI) and the time difference between interventricular septum and the RVw (R = 0.83, p < 0.001).
Figure 4 (A) Bar graph showing the mean and standard deviation values for individuals without pulmonary hypertension (black bars) and for patients with chronic pulmonary hypertension (grey bars) with regards to peak strain measured by TDI expressed in percent for the interventricular septum (IS). (B) Peak strain for the left ventricular lateral (LVL) wall and (C) and peak strain for the right ventricular wall (RVw). The p value for each one is represented.
Figure 5 (A) Bar graph showing the mean and standard deviation values for individuals without pulmonary hypertension (black bars) and for patients with chronic pulmonary hypertension (grey bars) with regards to time difference in mechanical activation expressed in milliseconds (ms) between the interventricular septum (IS) and left ventricular lateral (LVL). (B) Time difference in mechanical activation between interventricular septum (IS) and right ventricular wall (RVw). (C) Time difference in mechanical activation between the left ventricular lateral (LVL) wall and the right ventricular wall (RVw). The p value for each one is represented.
A representative tissue Doppler image is shown in Figure 6 demonstrating (A) Almost synchronized time-to-peak longitudinal peak systolic strain generation of both right ventricular free wall (yellow curve) and ventricular septum (green curve) in an individual without pulmonary hypertension. (B) A slightly delayed time-to-peak longitudinal right ventricular free wall peak systolic strain generation (yellow curve) is already noticeable when compared to the ventricular septum peak systolic strain (green curve) generation in a patient with mild chronic pulmonary hypertension. Please note that in both (A) and (B) the peak systolic strain (yellow curve) generated by the right ventricular wall when is higher than the peak systolic strain generated by the inter-ventricular septum. (C) A more noticeable delayed time-to-peak longitudinal right ventricular free wall peak systolic strain (yellow curve) generation when compared to ventricular septum peak systolic strain (green curve) generation is now evident in a patient with moderate chronic pulmonary hypertension as well as in (D) in a patient with severe chronic pulmonary hypertension and severe right ventricular dysfunction. In both (C) and (D), we can appreciate a significant reduction in peak systolic strain (yellow curve) generated by the right ventricular wall when compared to the peak systolic strain generated by the inter-ventricular septum (IS) as seen in the green curve in these patients with moderately severe pulmonary hypertension.
Figure 6 A representative tissue Doppler image is shown demonstrating (A) Almost synchronized time-to-peak longitudinal peak strain of both right ventricular free wall (yellow curve) and ventricular septum (green curve) in an individual without PAH. (B) Slightly delayed time-to-peak longitudinal right ventricular free wall peak strain (yellow curve) when compared to ventricular septum peak strain (green curve) in a patient with mild PAH. Note that there is no significant reduction in RVw strain generation. (C) A more noticeable delayed time-to-peak longitudinal right ventricular free wall peak strain (yellow curve) when compared to ventricular septum peak strain (green curve) is now evident in a patient with moderate PAH as well as in (D) in a patient with severe PAH and severe RV dysfunction. Note that in both (C) and (D) there is a significant reduction in RVw strain generation.
It is important to note that the electrocardiographic QRS duration between patients with chronic pulmonary hypertension was no different than the electrocardiographic QRS duration of individuals without pulmonary hypertension (91 ± 13 vs. 86 ± 9 ms, p = NS).
Univariate analysis between right ventricular dyssynchrony and extent of disease severity markers including World Health Organization (WHO) functional class, systolic pulmonary artery pressure (SPAP), hospitalizations, and deaths between individuals without pulmonary hypertension and patients with chronic pulmonary hypertension is seen in Table 3. Despite being a small sample we found that right ventricular dyssynchrony correlated significantly with WHO class, with SPAP severity index, and with the number of hospitalizations. Although there was also a correlation seen in terms of the number of deaths among patients with chronic pulmonary hypertension, the number of deaths is too small to be confident.
Table 3 shows data analysis regarding right ventricular dyssynchrony and extent of disease severity markers including World Health Organization (WHO) functional class, systolic pulmonary artery pressure (SPAP), hospitalizations, and deaths in individuals without pulmonary hypertension and patients with chronic pulmonary hypertension.
Variables No Pulmonary Hypertension Chronic Pulmonary Hypertension P value
WHO class 1 ± 0 2.9 ± 0.9 0.0001
SPAP severity 0 2.5 ± 0.8 0.0001
Hospitalizations 0 33 0.0001
Deaths 0 3 0.005
Discussion
The results of this study suggest that right ventricular dyssynchrony, represented as the time difference from interventricular septal to RVw activation [24], occurs in patients with chronic pulmonary hypertension and is strongly correlated with markers of right ventricular size and Eccentricity Index, a well-known parameter of right ventricular pressure overload [14]. In addition, the presence of right ventricular dyssynchrony correlates with markers disease severity including pulmonary hypertension severity index, World Health Organization class, and number of hospitalizations. Finally, right ventricular dyssynchrony is clearly evident with mild elevations in the pulmonary artery systolic pressure even when standard echocardiographic indices of right ventricular size and function are still within normal limits.
It is important to emphasize that in this study, right ventricular dyssynchrony was present even with a normal electrocardiographic QRS interval duration. A finding that is in agreement with previous data stating that an abnormal electrical conduction is not necessarily needed to produce left ventricular mechanical dyssynchrony; since left ventricular dyssynchrony has been identified in the failing myocardium with a normal QRS duration [32-34]. Therefore, it appears that not all contributing mechanisms resulting in mechanical dyssynchrony have been identified and are probably complex.
The results of this study can be quite useful in the evaluation of chronic pulmonary hypertension patients given the well-know limitations of standard echocardiography in the assessment of right ventricular size and function due to the complex structure and asymmetrical shape of this cardiac chamber [8-14]. We used changes in right ventricular area obtained from the apical 4-chamber views as indices of right ventricular size and global systolic function rather than ventricular volumes and ejection fraction. In addition we also used Myocardial Performance Index, a well-recognized measure of global systolic and diastolic function that is independent on any geometric assumptions and heart rate [26,27]. Finally, the use of the Eccentricity Index that indicates the degree of ventricular septal displacement is also a well-recognized marker of right ventricular deformation by either pressure or volume loads [14]. The variability in our measurements of RV size, morphology, and functional performance in patients with variable degree of chronic pulmonary hypertension is probably due to the right ventricular geometric remodeling that occurs in these patients with pulmonary hypertension as recently described by Sukmawan and associates [35].
The value of TDI has been widely applied to quantify regional left ventricular myocardial function under different clinical scenarios and all the available evidence suggests that it is quite useful to assess left ventricular mechanical dyssynchrony [36,37]. However, despite all this knowledge on left ventricular mechanical activation, no attempts have been made to quantify right ventricular dyssynchrony using strain imaging. In our study, we evaluated longitudinal shortening and its time sequence using strain imaging from apical 4-chamber views for two reasons; ease of accessibility from this window and functional dominancy of the longitudinal shortening over short-axis shortening [38].
Potential limitations of this present study obviously include the small number of patients. However, even with this small number of patients we were able to reach our primary goal of identifying the presence of a statistically significant right ventricular mechanical dyssynchrony in patients with chronic pulmonary hypertension. Second, the presence of a heterogeneous population of patients with regards to the etiology of their pulmonary hypertension. At first hand, this might be a crucial disadvantage; since the time sequence of events might be different with regards to different etiologies; however, what appears evident is that regardless of the initiating mechanism the same final pathway abnormality in right ventricular mechanics might be not that dissimilar between different etiologies. Third, an invasive pressure measurement was not used in this study; therefore, assessments of right ventricular time-pressure plots, dp/dt, and pulmonary vascular resistance were not available to compare with our echo and TDI data. Similarly, peak systolic pulmonary arterial pressures were estimated simply based on tricuspid regurgitation measurements. However, this widely used Doppler-derived pressure estimation is well recognized and has been documented to have a good correlation with simultaneously obtained catheter-derived measurements; particularly in patients with elevated systolic pulmonary artery pressures [39]. Last, the use of fractional area change as an index of global right ventricular systolic function has a limitation of being highly afterload dependent particularly in patients with pulmonary hypertension [40,41]. This effect might be compounded by tricuspid regurgitation that by reducing systolic afterload augments right ventricular systolic function. However, in this study we also assessed the severity of tricuspid regurgitation by measuring the width of the vena contracta and found the width of the vena contracta correlated with decreasing right ventricular fractional area change rather than playing role in augmenting right ventricular function.
It is important to clarify that right ventricular dyssynchrony was due to delayed RVw peak strain rather than to a septal motion abnormality. In addition, we found no significant dyssynchrony between septal to LV lateral activation in these patients. Although a clear mechanism to explain the delayed RVw contractility is beyond the scope of this study, we speculate that either ischemia with consequent tethering (post-systolic shortening) of the RVw or differences in afterload-dependency of the right ventricular free wall when compared to the interventricular septum might be considered possible mechanisms. In fact, post-systolic shortening of the RVw was observed in 50% of patients with chronic pulmonary hypertension in this study.
Conclusion
We conclude that right ventricular dyssynchrony, represented as the time difference from interventricular septal to RVw activation, occurs in patients with chronic pulmonary hypertension and is strongly correlated with markers of right ventricular size and Eccentricity Index. In addition, right ventricular dyssynchrony is associated with disease severity as it correlates with pulmonary hypertension severity index, World Health Organization class, and number of hospitalizations. Finally, right ventricular dyssynchrony is clearly evident with mild elevations in the pulmonary artery systolic pressure even when standard echocardiographic indices of right ventricular size and function are still within normal limits. Therefore, it is tempting to suggest that TDI might be useful in the early identification of patients with subclinical evidence of right ventricular dysfunction but further studies are required. In addition, the long-term effects of right ventricular dyssynchrony on morbidity and mortality as well as whether right ventricular resynchronization therapy that might correct right ventricular dyssynchrony and restore right ventricular function with resultant improvement of markers of disease severity and functional capacity also require investigation.
List of Abbreviations
TDI = Tissue Doppler imaging
RVw = Right ventricular lateral free wall
IS = Inter-ventricular septum
LVL = Left ventricular lateral wall
WHO = World Health Organization
SPAP = Systolic pulmonary artery pressure
dp/dt = Delta pressure / delta time
ms = milliseconds
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
All the listed authors contributed to gather and interpret data as well as to write the full length of this manuscript.
Acknowledgements
None
==== Refs
Brieke A DeNofrio D Right ventricular dysfunction in chronic dilated cardiomyopathy and heart failure Coron Artery Dis 2005 16 5 11 15654193 10.1097/00019501-200502000-00002
Di Salvo TG Mathier M Semigran MJ Dec GW Preserved right ventricular ejection fraction predicts exercise capacity and survival in advanced heart failure J Am Coll Cardiol 1995 25 1143 1153 7897128 10.1016/0735-1097(94)00511-N
Meluzin J Spinarova L Dusek L Toman J Hude P Krejci J Prognostic importance of the right ventricular function assessed by Doppler tissue imaging Eur J Echocardiogr 2003 4 262 271 14611821 10.1016/S1525-2167(02)00171-3
Peacock A Naeije R Galie N Reeves JT End points in pulmonary arterial hypertension: the way forward Eur Respir J 2004 23 947 953 15219011
Hoeper MM Oudiz RJ Peacock A Tapson VF Haworth SG Frost AE Torbicki A End points and clinical trial designs in pulmonary arterial hypertension: clinical and regulatory perspectives J Am Coll Cardiol 2004 43 48 55 10.1016/j.jacc.2004.02.010
McLaughlin VV Presberg KW Doyle RL Abman SH McCrory DC Fortin T Ahearn G American College of Chest Physicians Prognosis of pulmonary arterial hypertension: ACCP evidence-based clinical practice guidelines Chest 2004 126 78 92 10.1378/chest.126.1_suppl.78S
Chin KM Kim NH Rubin LJ The right ventricle in pulmonary hypertension Coron Artery Dis 2005 16 13 18 15654194 10.1097/00019501-200502000-00003
Berger HJ Matthay RA Loke J Marshall RC Gotschalk A Zaret BL Assessment of cardiac performance with quantitative radionuclide angiocardiography: right ventricular ejection fraction with reference to findings in chronic obstructive pulmonary disease Am J Cardiol 1978 41 897 905 645599 10.1016/0002-9149(78)90731-2
Incalzi RA Fuso L De Rosa M Di Napoli A Basso S Pagliari G Pistelli R Electrocardiographic signs of chronic cor pulmonale: a negative prognostic finding in chronic obstructive pulmonary disease Circulation 1999 99 1600 1605 10096937
Burrows B Kettel LJ Niden AH Rabinowitz M Diener CF Patterns of cardiovascular dysfunction in chronic obstructive lung disease N Engl J Med 1972 286 912 918 5013974
Barnard D Alpert JS Right ventricular function in health and disease Curr Probl Cardiol 1987 12 417 449 2959452
Armour JA Randall WC Structural basis for cardiac function Am J Physiol 1970 218 1517 1523 4986555
Weyman AE Wann S Feigenbaum H Dillon JC Mechanism of abnormal septal motion in patients with right ventricular volume overload: a cross-sectional echocardiographic study Circulation 1976 54 179 186 939018
Ryan T Petrovic O Dillon JC Feigenbaum H Conley MJ Armstrong WF An echocardiographic index for separation of right ventricular volume and pressure overload J Am Coll Cardiol 1985 5 918 927 3973294
Nagel E Stuber M Hess OM Importance of the right ventricle in valvular heart disease Eur Heart J 1996 17 829 836 8781821
Bommer W Weinert L Neumann A Neef J Mason DT DeMaria A Determination of right atrial and right ventricular size by two-dimensional echocardiography Circulation 1979 60 91 100 445737
Suzuki J Caputo GR Masui T Chang J O'Sullivan M Higgins CB Assessment of right ventricular diastolic and systolic function in patients with dilated cardiomyopathy using cinemagnetic resonance imaging Am Heart J 1991 122 1035 1040 1927854 10.1016/0002-8703(91)90469-X
Marangoni S Scalvini S Schena M Vitacca M Quadri A Levi G Right ventricular diastolic function in chronic obstructive lung disease Eur Respir J 1992 5 438 443 1563502
Mizushige K Morita H Senda S Matsuo H Influence of right ventricular pressure overload on left and right ventricular filling in cor pulmonale assessed with Doppler echocardiography Jpn Circ J 1989 53 1287 1296 2533278
Gorcsan J 3rd Tissue Doppler Echocardiography Curr Opin Cardiol 2000 15 323 329 11128184 10.1097/00001573-200009000-00003
Miyatake K Yamagishi M Tanaka N Uematsu M Yamazaki N Mine Y Sano A Hirama M New method for evaluating left ventricular wall motion by color-coded tissue Doppler imaging: in vitro and in vivo studies J Am Coll Cardiol 1995 25 717 724 7860919 10.1016/0735-1097(94)00421-L
Heimdal A Stoylen A Torp H Skjaerpe T Real-time strain rate imaging of the left ventricle by ultrasound J Am Soc Echocardiogr 1998 11 1013 1019 9812093
Donovan CL Armstrong WF Bach DS Quantitative Doppler tissue imaging of the left ventricular myocardium: validation in normal subjects Am Heart J 1995 130 100 104 7611097 10.1016/0002-8703(95)90242-2
López-Candales A Dohi K Bazaz R Edelman K Relation of Right Ventricular Free Wall Mechanical Delay to Right Ventricular Dysfunction as Determined by Tissue Doppler Imaging Am J Cardiol 2005 96 602 606 16098321 10.1016/j.amjcard.2005.04.028
Schiller NB Pulmonary artery pressure estimation by Doppler and two-dimensional echocardiography Cardiol Clin 1990 8 277 287 2189562
Miller D Farah MG Liner A Fox K Schluchter M Hoit BD The relation between quantitative right ventricular ejection fraction and indices of tricuspid annular motion and myocardial performance J Am Soc Echocardiogr 2004 17 443 447 15122184
Thohan V Prognostic implications of echocardiography in advanced heart failure Curr Opin Cardiol 2004 19 238 249 15096957 10.1097/00001573-200405000-00009
Yock PG Popp RL Noninvasive estimation of right ventricular systolic pressure by Doppler ultrasound in patients with tricuspid regurgitation Circulation 1984 70 657 662 6478568
Tribouilloy CM Enriquez-Sarano M Bailey KR Tajik AJ Seward JB Quantification of tricuspid regurgitation by measuring the width of the vena contracta with Doppler color flow imaging: a clinical study J Am Coll of Cardiol 2000 36 472 478 10933360 10.1016/S0735-1097(00)00762-2
Gorcsan J 3rdGulati VK Mandarino WA Katz WE Color-coded measures of myocardial velocity throughout the cardiac cycle by tissue Doppler imaging to quantify regional left ventricular function Am Heart J 1996 131 1203 1213 8644601 10.1016/S0002-8703(96)90097-6
Sade LE Kanzaki H Severyn D Dohi K Gorcsan J 3rd Quantification of radial mechanical dyssynchrony in patients with left bundle branch block and idiopathic dilated cardiomyopathy without conduction delay by tissue displacement imaging Am J Cardiol 2004 94 514 518 15325944 10.1016/j.amjcard.2004.04.071
Rouleau F Merheb M Geffroy S Berthelot J Chaleil D Dupuis JM Victor J Geslin P Echocardiographic assessment of the interventricular delay of activation and correlation to the QRS width in dilated cardiomyopathy Pacing Clin Electrophysiol 2001 24 1500 1506 11707043 10.1046/j.1460-9592.2001.01500.x
Kass DA Predicting cardiac resynchronization response by QRS duration – The long and short of it J Am Coll Cardiol 2003 42 2125 2127 14680738 10.1016/j.jacc.2003.09.021
Yu CM Lin H Zhang Q Sanderson JE High prevalence of left ventricular systolic and diastolic asynchrony in patients with congestive heart failure and normal QRS duration Heart 2003 89 54 60 12482792 10.1136/heart.89.1.54
Sukmawan R Akasaka T Watanabe N Akiyama M Wada N Taniguchi M Kawamoto T Yoshida K Quantitative assessment of right ventricular geometric remodeling in pulmonary hypertension secondary to left-sided heart disease using real-time three-dimensional echocardiography Am J Cardiol 2004 94 1096 1099 15476638 10.1016/j.amjcard.2004.06.079
Breithardt OA Sinha AM Schwammenthal E Bidaoui N Markus KU Franke A Stellbrink C Acute effects of cardiac resynchronization therapy on functional mitral regurgitation in advanced systolic heart failure Am Coll Cardiol 2003 41 765 770 10.1016/S0735-1097(02)02937-6
Kanzaki H Bazaz R Schwartzman D Dohi K Sade LE Gorcsan J 3rd A mechanism for immediate reduction in mitral regurgitation after cardiac resynchronization therapy: insights from mechanical activation strain mapping J Am Coll Cardiol 2004 44 1619 1625 15489094 10.1016/j.jacc.2004.07.036
Kukulski T Hubbert L Arnold M Wranne B Hatle L Sutherland GR Normal regional right ventricular function and its change with age: a Doppler myocardial imaging study J Am Soc Echocardiogr 2000 13 194 204 10708468 10.1067/mje.2000.103106
Caso P Galderisi M Cicala S Cioppa C D'Andrea A Lagioia G Liccardo B Martiniello AR Mininni N Association between myocardial right ventricular relaxation time and pulmonary arterial pressure in chronic obstructive lung disease: analysis by pulsed Doppler tissue imaging J Am Soc Echocardiogr 2001 14 970 977 11593201 10.1067/mje.2001.115033
Morita S Kormos RL Mandarino WA Eishi K Kawai A Gasior TA Deneault LG Armitage JM Hardesty RL Griffith BP Right ventricular/arterial coupling in the patient with left ventricular assistance Circulation 1992 86 316 325
Katz WE Gasior TA Quinlan JJ Lazar JM Firestone L Griffith BP Gorcsan J 3rd Immediate effects of lung transplantation on right ventricular morphology and function in patients with variable degrees of pulmonary hypertension J Am Coll Cardiol 1996 27 384 391 8557910 10.1016/0735-1097(95)00502-1
|
16129028
|
PMC1215497
|
CC BY
|
2021-01-04 16:38:31
|
no
|
Cardiovasc Ultrasound. 2005 Aug 29; 3:23
|
utf-8
|
Cardiovasc Ultrasound
| 2,005 |
10.1186/1476-7120-3-23
|
oa_comm
|
==== Front
Cost Eff Resour AllocCost effectiveness and resource allocation : C/E1478-7547BioMed Central London 1478-7547-3-81609113910.1186/1478-7547-3-8ResearchThe cost-effectiveness of neonatal screening for Cystic Fibrosis: an analysis of alternative scenarios using a decision model Simpson Neil [email protected] Rob [email protected] Franco [email protected] Alexandra [email protected] Peter [email protected] Karen [email protected] Heather [email protected] Department of Child Health, Newbridge Hill, Bath, BA1 3QE, UK2 Peninsula Technology Assessment Group (PenTAG) & Institute for Health & Social Care Research, Peninsula Medical School, Universities of Exeter and Plymouth, Plymouth, PL6 8BU, UK 3 Department of Social Policy, The London School of Economics and Political Science, Houghton Street, London, WC2A 2AE, UK4 LSE Health and Social Care, The London School of Economics and Political Science, Houghton Street, London, WC2A 2AE, UK5 Department of Medical Sciences, University of Bath, Bath, BA2 2 BB, UK6 University of Toronto, Canada. Associate Scientist, Institute of Clinical Evaluative Sciences (ICES), G-214, 2075 Bayview Avenue, Toronto, Ontario, M4N 3M5, Canada2005 9 8 2005 3 8 8 20 8 2004 9 8 2005 Copyright © 2005 Simpson et al; licensee BioMed Central Ltd.2005Simpson et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
The use of neonatal screening for cystic fibrosis is widely debated in the United Kingdom and elsewhere, but the evidence available to inform policy is limited. This paper explores the cost-effectiveness of adding screening for cystic fibrosis to an existing routine neonatal screening programme for congenital hypothyroidism and phenylketonuria, under alternative scenarios and assumptions.
Methods
The study is based on a decision model comparing screening to no screening in terms of a number of outcome measures, including diagnosis of cystic fibrosis, life-time treatment costs, life years and QALYs gained. The setting is a hypothetical UK health region without an existing neonatal screening programme for cystic fibrosis.
Results
Under initial assumptions, neonatal screening (using an immunoreactive trypsin/DNA two stage screening protocol) costs £5,387 per infant diagnosed, or £1.83 per infant screened (1998 costs). Neonatal screening for cystic fibrosis produces an incremental cost-effectiveness of £6,864 per QALY gained, in our base case scenario (an assumed benefit of a 6 month delay in the emergence of symptoms). A difference of 11 months or more in the emergence of symptoms (and mean survival) means neonatal screening is both less costly and produces better outcomes than no screening.
Conclusion
Neonatal screening is expensive as a method of diagnosis. Neonatal screening may be a cost-effective intervention if the hypothesised delays in the onset of symptoms are confirmed. Implementing both antenatal and neonatal screening would undermine potential economic benefits, since a reduction in the birth incidence of cystic fibrosis would reduce the cost-effectiveness of neonatal screening.
==== Body
Background
Cystic fibrosis is an inherited disorder associated with considerable morbidity and reduced life expectancy. The UK birth prevalence is about 0.4 per 1,000, or 300 new cases each year [1,2]. A recent report from the UK NHS Health Technology Assessment Programme recommended that antenatal screening for cystic fibrosis should be offered routinely, and that "Health Authorities could consider introducing neonatal screening" [3].
Antenatal screening aims to prevent affected births and neonatal screening aims to improve prognosis by early intervention. Studies of the effectiveness of neonatal screening have measured short-term outcomes or are subject to statistical biases, including selection and lead-time bias, and the use of historical controls [4-8]. Thus the ability of neonatal screening to alter long-term prognosis is not proven, although there is limited circumstantial evidence favouring a benefit [3].
All four existing economic evaluations of neonatal screening for cystic fibrosis are from the USA, and are ten or more years old [9-12]. None used any measure of health effect or performed extensive sensitivity analysis, and only two of the studies compared screening to a do-nothing alternative. Exclusion of the no-screening alternative seems inappropriate, as the type and scale of the benefits of cystic fibrosis screening remain uncertain.
Given the short-comings of these studies, and the lack of full economic evaluations based on UK service and survival data, we aimed to compare the lifetime cost-effectiveness of neonatal screening with no screening, under different possible scenarios for survival with cystic fibrosis. We undertook this from the perspective of a hypothetical UK Health Authority that has an existing routine neonatal screening programme for congenital hypothyroidism and phenylketonuria but not for cystic fibrosis.
Methods
To allow a full exploration of the uncertainty surrounding our cost-effectiveness estimates, they were created using a decision tree, incorporating Markov processes to model lifetime costs and quality of life [13]. We modelled a screening programme using two-stage immunoreactive trypsin combined with genetic testing strategy (Figure 1). This model has replaced two stage immuno-reactive trypsin tests in a number of programmes internationally, and is the commonest protocol used in programmes starting after 1990 [3].
Figure 1 Decision tree for no screening and neonatal screening strategies for cystic fibrosis.
The disease progression was modelled as a Markov process. All cases are 'born' into the pre-symptomatic health-state. Then, each year, there is a given probability of moving into the symptomatic disease state, then into the severe irreversible lung disease-state, and finally death (with the probabilities of moving into the last two states changing over time). These states broadly correspond with the stages of disease described in paediatrics, and also reflect thresholds between different therapy regimes [14].
The model excluded those cases (15% in the base case scenario) diagnosed at or shortly after birth, for example by meconium ileus or family history, since these infants would have received the same prognosis and treatment under both strategies. Under the "no screening" strategy infants would be diagnosed with cystic fibrosis symptomatically (late diagnosis). Under the screening strategy most cystic fibrosis cases would be detected by screening (early diagnosis), with the remainder – the false negatives – experiencing the disease under late diagnosis assumptions.
The putative benefit of early diagnosis through neonatal screening was modelled as a difference in the annual transition probability of remaining pre-symptomatic. In the initial model this probability was 69% for those diagnosed through screening (compared with 59% for those diagnosed symptomatically) resulting in a delay of the emergence of symptoms of 6 months.
The decision analysis model uses three types of data; probability data, cost data, and quality of life estimates for the three health states: these data were obtained from a variety of different sources as referenced in Tables 1, 2 and 3.
Table 1 Model parameters, data sources and values used in the model.
Probabilities
Variable Base case value Range used in sensitivity analysis
Incidence of cystic fibrosis [2] 0.0004 0.00067 – 0.00029
% diagnosed at birth (MI & family history) [39] 0.15 0.10 – 0.40
IRT test sensitivity [40] 0.9 0.99
IRT test specificity [40] 0.995 0.999
DNA test: % of mutations detected [40] 0.88 0.85 – 0.95
DNA test sensitivity [40] 0.9856 0.9975
DNA test specificitya 1.0
Increased annual transition probability of remaining without symptoms (in early-diagnosed cases)b 10% 10 – 40%
a It is assumed that there are no false positives (from the combined DNA and sweat tests), because all test results that are either homozygous or heterozygous for cystic fibrosis mutations, are confirmed using sweat tests.
b Assumed effect of early diagnosis
Table 2 Model parameters, data sources and values used in the model.
Costs (all inflated to reference year 1998)
Variable Base case value Range used in sensitivity analysis
Costs of screening
Additional time to explain test (survey by NS) £0.40a (2.1 mins) £0 – £1.44 (0 – 7.6 mins)
Obtaining and transport of blood sampleb £0
IRT test (Bradley DM – pers. comm.) £0.97 £0.50 – £1.50
DNA test (Bradley DM – pers. comm.) £79.48 £40.00 – £120.00
Sweat test (Walker S – pers. comm) £29.40 £15.00 – £45.00
Administration and feedback of resultsc £0
Cost of pre-diagnosis care in unscreened group (audit by NS)
Presumed GP visits (mean number of visits)d £14.77 (1.27) £11.63 – £46.52 (1 – 4)
Outpatient attendances (mean number)e £129.07 (1.47) £0 – £263.40 (0 – 3)
Inpatient admissions (mean number of admissions and days per admission)f £792.55 (0.87) (3.0 days) £0 – £1821.96 (0–2)
Costs of treatment per year in health state by age group g14
Presymptomatic 0–5 £2,950
6–10 £3,995
11–15 £4,570
> 16 £4,275
Symptomatic 0–5 £15,241
6–10 £15,704
11–15 £19,247
>16 £19,291
Severe irreversible symptoms 0–5 £28,722
6–10 £30,692
11–15 £37,224
>16 £37,388
a Whitley Councils for the Health Services Pay Scales – 1/4/98 (Nursing and Midwifery) + 16% on-costs = £11.33/hour
bc Assumed to be a shared cost with existing screening programmes
d The Government's Expenditures Plans 1996–97, Cost/GP consultation = £11.63
e Annual financial returns for NHS Trusts 97/98, Cost//paediatric outpatient attendance = £87.80
f Annual financial returns for NHS Trusts 97/98, Cost//patient-day in paediatric ward) = £303.66
g Hospital staff, overhead and capital costs are included in the average cost per day for inpatient stays and the average cost of outpatient attendance. Drug costs were derived from the monthly index of medical specialities (MIMMS) and do not include individual hospital discount arrangements (1996 costs inflated to 1998).
Table 3 Model parameters, data sources and values used in the model.
Utility values of symptom states
Variable Base case value Range used in sensitivity analysis
Asymptomatic – latea 0.95 0.90
Asymptomatic – earlyb 0.95 0.90
Symptoms (FEV1 – 60%, range 40%–80%) [17,41] 0.75 0.65 – 0.90
Severe irreversible symptoms (FEV1 – 30%, range 20–40%) [17,41] 0.68 0.58 – 0.78
a utility value to reflect "taking regular medication or staying on a prescribed health diet for health reasons" derived from community surveys.
b arbitrary utility weight to reflect the probable repeated visits to GP with undiagnosed CF
Probability data
The transition probabilities in the Markov model were estimated to achieve age-specific survival rates and other estimated parameters, in three alternative scenarios; based on conservative, balanced and optimistic assumptions of recent UK age-specific survival data supplied by one of the authors, PL (Figure 2 and Table 3). The annual transition probabilities that best predicted these calibration data were (for the balanced scenario, 'late diagnosis'): from asymptomatic to symptomatic, 0.491 per year (with the remainder all staying asymptomatic); symptomatic to severe irreversible lung disease, 0.0064 increasing exponentially according the accumulated years with symptoms (with hazard rates derived from Dodge et al. 1997); severe irreversible lung disease to death, 0.09 increasing according to the number of years spent in the severe irreversible disease stage. (Excel spreadsheets available on request from the first author, NS). These allowable transitions effectively make the simplifying assumptions that all people with CF ultimately die of CF-related respiratory symptoms, and that all pass through both the symptomatic and severe irreversible lung disease stages before they die.
Figure 2 Survival curves under conservative, balanced and optimistic assumptions.
Cost data
The following screening costs were the included in the model: counselling time required by midwives to obtain consent for testing, immunoreactive trypsin test, DNA analysis and sweat chloride test. Other costs related to obtaining the blood spot and feedback of results by health visitors, were assumed to be sunk in the existing neonatal screening programmes for phenylketonuria and congenital hypothyroidism.
We assume the addition of cystic fibrosis screening does not increase refusals or insufficient blood samples and that 100% of the neonatal population would be covered by the programme [15]. Time for genetic counselling for carriers identified by the screening programme were excluded from the model.
The pre-diagnosis healthcare costs for children with late diagnosis (no-screening) was estimated via an audit of the clinical notes of 25 children with cystic fibrosis, attaching unit costs to derive the mean cost of pre-diagnosis care (Table 2).
Disease state-specific costs of treatment were derived from the cost of care of 161 patients at a large UK cystic fibrosis unit during 1996 which were based on annual medical costs for patients at different age groups and at different disease stages (Table 2) [14]. Only CF-related health care costs are included, including any costs incurred during additional years of life added by early diagnosis. All cost data was adjusted for inflation at 5% to the reference year of 1998. Future treatment costs in the model were discounted at 6% per year [16].
Effectiveness data
It has been shown [17,18] that forced expiratory volume in one second (FEV1) is associated with quality of life, as measured by the Quality of Well-Being Scale (a preference-based quality of life instrument used in economic evaluation) [19], as well as with morbidity and mortality [20,21]. These Quality of Well-Being scores remain the only preference-based estimates of health-related quality of life in people with CF. On the basis of these findings, a quality of life value was assigned to each Markov state (Table 3) and multiplied by survival time in the same state to produce quality-adjusted life expectancy. Future QALYs were discounted in the model (2% in base case analysis) [16].
A wide range of one-way, and selected multi-way sensitivity analyses were undertaken. The sensitivity analyses reported here are those that either (a) had a significant impact on the incremental cost-effectiveness ratio, or (b) related to parameters for which reliable published estimates were not available (e.g. QALYs per year of life lived with CF symptoms or severe irreversible symptoms, and the cost of pre-diagnosis care for those diagnosed symptomatically). Microsoft Excel© (version 5.0) spreadsheet software was used to develop the model and conduct the analyses – the models are available on request from the lead author.
Results
Our base case assumptions gave an estimated cost per diagnosed infant of £5,387 (or £1.83 per infant screened), compared with an estimated cost per case diagnosed clinically of £936. £4,020 (75%) of this cost is due to those components of the screening process that are carried out for every infant screened i.e. the immunoreactive trypsin test, and the explanation of the test by midwives (Table 4).
Table 4 Components of the average cost of diagnosis by screening
Cost component Mean cost of diagnosis per infant screened (£) Mean cost of diagnosis per infant diagnosed (£) Cost to average Health Authoritya per year (£)
Midwife time explaining test 0.40 1,167 2,380
IRT tests (kits, & overheads) 0.97 2,853 5,820
DNA tests (kits, labour & overheads) 0.42 1,240 2,530
Confirmatory sweat tests 0.01 22 44
Pre-diagnosis care of false negatives 0.04 106 216
Total average cost 1.83 5,387 10,990
a Health Authority with a population of 500,000 and assumed crude birth rate of 12 per 1000 population per year.
If the explanation to parents of cystic fibrosis screening could be incorporated within existing screening arrangements, the average cost per diagnosed infant would fall by 22% to £4,220. Conversely, if explaining cystic fibrosis screening cannot be incorporated in the existing process, but instead commits midwives to seven and a half minutes more time [22], then the cost per diagnosed infant rises to £8,443. Even with highly optimistic assumptions regarding the specificity and sensitivity of both screening tests the mean cost of diagnosis by screening falls by only a fifth, to £4,351. The cost of diagnosis is more sensitive to the incidence of the disease, and the proportion of cases detected at birth (Figure 3).
Figure 3 Cost of diagnosis (£) by screening assuming different disease incidence and different proportions diagnosed at birth by other means.
Neonatal screening for cystic fibrosis produced (under base case assumptions) an average of 0.36 additional QALYs per life with cystic fibrosis at an additional cost of £2,895 (or a cost per QALY gained of £6,864) (Figure 4).
Figure 4 Incremental cost-effectiveness (£/QALY) of neonatal screening for cystic fibrosis compared with no screening, assuming different cost discount rates and assumed effects of preventive therapy.
A delay in the emergence of symptoms, or an increase in survival of 11 months or more (i.e. an increase in the probability of remaining without symptoms of 15 percentage points or more), compared to unscreened infants, would produce lower costs and better outcomes than no screening (Figure 4). A wide range of different quality of life estimates for the two disease states produce negligible changes in the cost-effectiveness ratio. This is because in this model the main effect of early diagnosis is to delay the emergence of symptoms, but the progression of disease thereafter is the same under the late and early diagnosis assumptions.
If pre-diagnosis care (of those diagnosed symptomatically) does not involve admission to hospital then the incremental cost-effectiveness ratio falls by 32% to £4,640 per QALY. Under any of the survival models if annual treatment costs are increased by as much as 20% the cost per QALY gained only increases by 11%. Also, as life expectancy increases so does the incremental cost-effectiveness of neonatal screening (Table 5).
Table 5 Cost-effectiveness of screening compared to no screening under various survival scenarios
Scenario: Conservative survivala Balanced survivala Optimistic survivala
Model parameter:
% surviving to age 16 90% 95% 97%
% surviving to age 30 72% 84% 90%
% surviving to age 50 30% 45% 55%
Median survival (years) 41 48 52
Mean survival (years) 39.4 45.8 49.9
Mean years spent:
before symptoms emerge 1.0 1.0 1.0
with symptoms 32.3 38.7 42.8
in severe irreversible stage 6.1 6.1 6.1
Incremental C/E ratio (£/QALY) £7,474 £6,864 £6,532
a Under late diagnosis assumptions
Discussion
As a method of diagnosis neonatal screening is relatively expensive. At £5,387 per diagnosed infant, our estimate of the cost of diagnosis is similar to a previous (1997) estimate of £6,400 [23]. Compared to no screening, neonatal screening for cystic fibrosis (under base case assumptions) produces an incremental cost-effectiveness of £6,864 per QALY gained.
There are a number of key assumptions and potential limitations to this study. Firstly, in relation to the model structure, we have simplified the representation of cystic fibrosis. However our representation of the disease – as three health states within a Markov model – makes best use of available knowledge. Also, we have assumed that the effect of early diagnosis is a delay in the emergence of respiratory symptoms after prophylactic treatment (but, with the subsequent treatment cost and quality of life in those states being the same for screened and unscreened infants). This choice was made in the absence of evidence to suggest other ways of modelling the effect of early diagnosis. Unfortunately, while recent trials or observational studies have shown that early (neonatal) diagnosis and treatment results in improvements in nutritional status, height and weight, and cognitive functioning, evidence about the long-term impact of early diagnosis on lung function (and therefore mortality) remains uncertain [24-28].
The model assumes that cystic fibrosis in infants is a relatively homogeneous condition. However, the spectrum of cases ranges from neonates that are severely affected, to cases who live a normal life undiagnosed until adulthood. It is possible that the more severe but asymptomatic cases would benefit most from early diagnosis. Those with milder forms of the disease would be diagnosed later under the no-screening strategy and would have their age at diagnosis advanced most under screening. The data in existing population-based data sets are insufficient to investigate this issue (PL – author), so this point has been ignored.
Our survival estimates are based on the most recent and reliable age-specific survival data. As yet there is a limited understanding of the interactions of cystic fibrosis with the normal ageing process. Recent clinical experience indicates that potentially life-shortening complications such as diabetes mellitis and liver disease may become more frequent with advancing age, so projections based on experience of younger people may be inaccurate [2,29]. Although it is possible that recent improvements in survival may be confounded by the introduction of neonatal screening in some areas in the 1970s and 1980s, regional variations in mortality do not show this (PL – author). Further, using Quality of Well-Being (QWB) scores as a basis for weighting the quality of extra years survived may not provide a true indicator of relative preferences for being in these different health states. However, it remains the only health-related quality of life instrument that has been used widely in people with cystic fibrosis [30].
Overall, as far as presently available data allow, the model structure and data inputs would satisfy most of the criteria that are recommended in current guidelines for good practice in decision analytic modelling [31]. External validation of the model is more problematic since valid data about either the long-term health effects of screening or the future survival of people with cystic fibrosis does not yet exist. Due to the absence of published data on the distributions underlying the means of most parameters our sensitivity analysis is restricted to one-way and two-way sensitivity analysis. Future modelling of these policy choices should attempt to use data that allow more probabilistic sensitivity analysis to be conducted, but with the proviso that model structure (or methodological) uncertainty can only be explored using traditional 'non-probabilistic' methods.
A number of costs, which may in theory are important have been omitted, either because of the complexity of deriving estimates, or their probable minimal effect on the main findings. The omitted costs are; the potential effects of distress or reassurance related to screening [32], self- and lay-care costs (for example, therapy provided by parents or by the patients themselves), unrelated health care costs and savings resulting from increasing life expectancy (for example, additional years of economically productive life); the costs of genetic counselling (generated by the identification of carriers); and the treatment option of heart-lung transplantation.
The omission of genetic counselling costs was partly because the benefits of such counselling would be difficult to quantify. Existing evidence also shows that, of the small number of carriers that will be identified, only a minority take up counselling and the cost of providing this counselling is small in relation to total screening costs [33,34]. With regard to heart-lung transplantation, this treatment option is currently only available to a relatively small proportion of people with cystic fibrosis. Even if this changes there is no reason to assume that the costs, benefits or availability would be different for those diagnosed at birth by screening and those diagnosed later symptomatically [35-39].
With regard to the generalisability of the findings, this study has most relevance to the UK context: it employs cost estimates based on NHS care and the survival estimates are derived from the UK National Cystic Fibrosis Survey [1]. Although comparisons with other studies are difficult; the costs of care at the Leeds unit (average annual cost of £10,567 [14]) from which our data are derived were comparable with another UK study [40].
Although a number of alternative screening protocols are being used in the UK and world-wide, they all employ an IRT test as the initial screening stage for all neonates. Our analysis shows that it is the cost of this initial stage, carried out on all infants, which most affects the cost-effectiveness ratio, and also that differences in the performance of the screening protocol produce only minor changes. Therefore, it is unlikely that substantially different results would be obtained with alternative protocols.
We have shown that, in the absence of antenatal screening, neonatal screening costs of £6,864 per QALY gained, based on an assumed benefit of six months average delay in the onset of symptoms, and that it would be less costly and more beneficial if the benefit were shown to be 11 months or more. As further evidence becomes available it will be clearer if this threshold can be realised. The model used here could be adapted to reflect new effectiveness data, associations between genotype and phenotype, new treatments for cystic fibrosis as they become available; and local information concerning populations or services.
Comparisons of cost per affected pregnancy identified by antenatal screening with averted treatment costs are generally favourable [3]. However, these studies give no value to a life lived with cystic fibrosis. They also assume around 90% uptake of prenatal diagnosis and effectively universal termination of affected pregnancies. This contrasts with surveys of affected individuals and close family members which suggest that only about half find termination of an affected pregnancy acceptable [41-43]. Changes in public attitudes about prenatal diagnosis and termination might further affect the economic value of antenatal screening.
In the UK policy context Murray et al. recommended both that antenatal screening should be offered routinely and that health authorities 'could consider' introducing neonatal screening. They also suggest that routine antenatal screening would reduce the birth prevalence of cystic fibrosis by between 43% and 49% [3]. According to our analysis, by halving the birth prevalence of the disease, the cost per QALY gained for neonatal screening would increase to £19,543. With this lower prevalence at birth, early diagnosis would have to delay the emergence of symptoms on average by 20 months or more for neonatal screening to be less costly and more beneficial.
In 1998 in the UK, up to a quarter of the population were covered by six regional programmes of neonatal screening for cystic fibrosis and we are aware of only one scheme which provides routine antenatal screening (in Edinburgh) [3]. Any economic evaluation of antenatal screening compared with neonatal screening is impossible as the strategies have such different aims. Fundamentally, the benefits are not comparable: in antenatal screening the aim is to allow reproductive choice, including the option to terminate affected pregnancies, whereas neonatal screening primarily aims to improve the length and quality of life of sufferers. Economic evaluations of antenatal screening have therefore tended to give no value to a life with cystic fibrosis, and instead attribute financial savings to lives with cystic fibrosis avoided.
In conclusion, according to the scenarios explored here, as long as the birth incidence of cystic fibrosis remains stable, there is no reason for existing neonatal screening programmes to be discontinued on cost-effectiveness grounds. Although UK Health Authorities wanting to introduce neonatal screening may want to see more reliable evidence of the health benefits of early diagnosis before making a decision, this evidence is now beginning to emerge, especially from the Wisconsin trial [44]. It has prompted the (UK) NHS National Screening Committee to implement a national neonatal screening programme for England (Scotland introduced theirs in 2003). As this programme is rolled out to different regions from April 2005 (see ) more reliable data on how early diagnosis alters lung function and long-term survival will become available, and could be used to update this cost-effectiveness analysis.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
NS, RA, FS, AP, KT, HS all contributed to conception, design of study and editing, RA developed spreadsheet model, NS, RA and FS undertook analysis and wrote the paper, and PL contributed to analysis and editing.
==== Refs
Dodge JA Morison S Lewis PA Coles E Geddes D Russel G Cystic Fibrosis in the United Kingdom, 1968–1988: incidence, population and survival Pediatr & Perinat Epidemiol 1993 2 157 166
Dodge JA Morison S Lewis PA Coles E Geddes D Russel G Incidence, population and survival of cystic fibrosis in the UK, 1968–95 Arch Dis Child 1997 77 493 496 9496181
Murray J Cuckle H Taylor G Littlewood J Hewison J Screening for cystic fibrosis Health Technology Assessment 1999 3
Farrell PM Kosorok MR Laxova A Guanghong S Koscik R Bruns T Nutritional benefits of neonatal screening for cystic fibrosis N Engl J Med 1997 337 963 969 9395429 10.1056/NEJM199710023371403
Chatfield S Owen G Ryley H Williams J Alfaham M Goodchild M Neonatal screening for cystic fibrosis in Wales and the West Midlands: clinical assessment after five years of screening Arch Dis Child 1991 66 29 33 1996888
Kuzemko JA Neonatal screening for cystic fibrosis Lancet 1986 i 327 328
Dankert-Roelse J Meeman GJ Martijin A Kate LP Knol K Survival and clinical outcome in patients with cystic fibrosis, with or without neonatal screening J Pediatr 1989 114 262 267
Waters D Wilken B Irwig L Van Asperen P Mellis C Simpson J Clinical outcomes of newborn screening for cystic fibrosis Arch Dis Child Fetal Neonatal Ed 1999 80 F1 F7 10325803
Gregg RG Wilfond BS Farrell P Laxova A Hassemer D Mischler E Application of DNA analysis in a population-screening program for neonatal diagnosis of cystic fibrosis (CF): comparison of screening protocols American Journal of Human Genetics 1993 52 616 626 7680526
Dauphinais RM A cost-analysis of blood spot screening of newborns for cystic fibrosis J Clin Immunoassay 1992 15 121 125
Farrell PM Mischler EH Newborn screening for cystic fibrosis Advances in Pediatrics 1992 39 35 70 1442316
Pauly MV The Economics of Cystic Fibrosis Boston, John Wright, PSG 1983
Sonnenberg F Beck R Markov Models in medical decision making: a practical guide Medical Decision Making 1993 13 322 338 8246705
Littlewood JM Cross E Present day treatment of cystic fibrosis: its content and cost Clinical Economics in Gastroenterology 1999 220 249
Simpson N Randall R Lenton S Walker S Audit of neonatal screening programme for phenylketonuria and congenital hypothyroidism Arch Dis Child 1997 47 228 234
Torgerson DJ Raftery J Economics notes – Discounting BMJ 1999 319 914 915 10506056
Orenstein D Nixon PA Ross EA Kaplan RM The quality of well-being in cystic fibrosis Chest 1989 92 344 347 2914486
Orenstein D Pattishall EN Nixon PA Ross EA Kaplan RM Quality of well-being before and after antibiotic treatment of pulmonary exacerbation on patients with cystic fibrosis Chest 1990 98 1081 1084 2225948
Bowling A Measuring Health – a review of Quality of Life measurement scales Milton Keynes 1991 Open Clinical Press
Congleton J Hodson M Duncan-Skingle F Quality of life in adults with cystic fibrosis Thorax 1996 51 936 940 8984707
Corey M Farewell V Determinants of mortality from cystic fibrosis in Canada, 1970–1989 Am J Epidemiol 1996 143 1007 1017 8629607
Simpson N Audit of time for midwives to explain cystic fibrosis screening Unpublished 1999
Pollitt R Green A McCabe C Booth A Cooper N Leonard J Neonatal screening for inborn errors of metabolism: cost, yield and outcome (Review) Health Technology Assessment 1997 1
Lai HJ Cheng Y Cho H Kosorok MR Farrel P Association between initial disease presentation, lung disease outcomes, and survival in patients with cystic fibrosis American J Epidemiol 2004 159 537 546 10.1093/aje/kwh083
Koscik RL Farrel PM Kosorok MR Zaremba KM Laxova A Lai HC Cognitive function of children with cystic fibrosis: deleterious effect of early malnutrition Pediatrics 2004 113 1549 1558 15173473 10.1542/peds.113.6.1549
Assael BM Castellani C Ocampo MB Iansa P Callegaro A Valsecchi MG Epidemiology and survival analysis of cystic fibrosis in an area of intense neonatal screening over 30 years American Journal of Epidemiology 2002 156 397 401 12196308 10.1093/aje/kwf064
Siret D Bretaudeau G Branger B Dabadie A Dagorne M David V de Braekeleer M Comparing the clinical evolution of cystic fibrosis screened neonatally to that of cystic fibrosis diagnosed from clinical symptoms: a 10-year retrospective study in a French region (Brittany) Pediatric Pulmonology 2003 35 342 349 12687590 10.1002/ppul.10259
Wang SS FitzSimmons SC O'Leary LA Rock MJ Gwinn ML Khoury MJ Early diagnosis of cystic fibrosis in the new-born period and risk of Pseudomonas aeruginosa acquisition in the first 10 years of life: a registry-based longitudinal study Pediatrics 2001 107 274 279 11158458 10.1542/peds.107.2.274
Lewis P Morison S Dodge J Geddes D Coles E Russell G Survival estimates for adults with cystic fibrosis born in the United Kingdom between 1947 and 1967 Thorax 1999 54 420 422 10212106
Kotwicki RJ Condra L Vereulen L Wolf T Douglas J Farrel PM Assessing the quality of life in children with cystic fibrosis Western Medical Journal 2001 100 50 54
Weinstein M O'Brien B Hornberger J Jackson J Johannesson M McCabe C Principles of good practice for decision analytic modeling in health-care evaluation: Report of the ISPOR Task Force on Good Research Practices – Modelling Studies Value in Health 2003 6 9 17 12535234 10.1046/j.1524-4733.2003.00234.x
Sassi F McKee M Roberts J Economic evaluation of diagnostic technology: methodological challenges and viable solutions Int J Technolol assessment in Health Care 1997 13 613 630
Shannon N Evans S Pollitt R Quarrell O Follow up of Delta F508 detected as a result of neonatal screening in the Trent Region J Med Genet 1997 34 P6.05
Ranieri E Lewis B Morris C Wilcken B Dodge JBD, Widdicombe JH Neonatal screening using combined biochemical and DNA based techniques Cystic Fibrosis – current topics 1996
Wiebe K Wahlers T Harringer W vd Hardt H Fabel H Haverich A Lung transplantation for cystic fibrosis – a single centre experience over 8 years European Journal of Cardio-Thoracic Surgery 1998 14 191 196 9755006 10.1016/S1010-7940(98)00163-8
Coueitil J Soubrane O Houssin D Dousset B Chevalier P Combined heart-lung-liver, double lung-liver, and isolated liver transplantation for cystic fibrosis in children Transplant International 1997 10 33 39 9002149
Whitehead B Helms P Goodwin M Martin I Scott J Smyth R Heart-lung transplantation for cystic fibrosis. 2: Outcome Archives of Disease in Childhood 1991 66 1022 1026 1929506
Balfour-Lynn I Martin I Whitehead B Rees P Elliot M de Leval M Heart-lung transplantation for patients under 10 with cystic fibrosis Archives of Disease in Childhood 1997 76 38 40 9059159
Busschbach J Horikx P van den Bosch J Brutal de la Riviere A de Charro F Measuring the quality of life before and after bilateral lung transplantation in patients with cystic fibrosis Chest 1994 105 911 917 8131563
Robson M Abbott J Webb K Dodd M Walsworth-Bell J A cost description of an adult cystic fibrosis unit and cost analysis of different categories of patients Thorax 1992 47 684 689 1440461
Watson EK Williamson R Chapple J Attitudes to carrier screening for cystic fibrosis: a survey of health care professionals, relatives of sufferers and other members of the public Br J Gen Pract 1991 41 237 240 1931202
Watson EK Marchant J Bush A Williamson B Attitudes towards prenatal diagnosis and carrier screening for cystic fibrosis among the parents of patients in a paediatric cystic fibrosis clinic J Med Genet 1992 29 490 491 1640429
Conway SP Allenby K Pond MN Patient and parental attitudes toward genetic screening and its implications at an adult cystic fibrosis clinic Clin Gen 1994 45 308 312
Grosse SD Boyle CA Botkin JR Comeau AM Kharrazi M Rosefeld M Newborn screening for cystic fibrosis: evaluation of benefits and risks and recommendations for state newborn screening programs Morbidity & mortality weekly report 2004 53 Centre for Disease Control. Atlanta, Georgia, USA 1 36
|
16091139
|
PMC1215498
|
CC BY
|
2021-01-04 16:39:14
|
no
|
Cost Eff Resour Alloc. 2005 Aug 9; 3:8
|
utf-8
|
Cost Eff Resour Alloc
| 2,005 |
10.1186/1478-7547-3-8
|
oa_comm
|
==== Front
Curr Control Trials Cardiovasc MedCurrent Controlled Trials in Cardiovascular Medicine1468-67081468-6694BioMed Central 1468-6708-6-141614656610.1186/1468-6708-6-14ReviewPredictors of post-operative mortality following treatment for non-ruptured abdominal aortic aneurysm Urbonavicius Sigitas [email protected] Henrik [email protected] Grazina [email protected] Mindaugas [email protected] Dainius [email protected]é Bent [email protected] Kaunas University of Medicine, Institute of Cardiology, Lithuania2 University of Aarhus, Institute of Medical Biochemistry, Denmark3 Kaunas University of Medicine, Clinic of Surgery, Lithuania4 Kaunas University of Medicine, Clinic of Cardiovascular Surgery, Department of Vascular Surgery, Lithuania2005 7 9 2005 6 1 14 14 24 10 2004 7 9 2005 Copyright © 2005 Urbonavicius et al; licensee BioMed Central Ltd.2005Urbonavicius et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
The aim of this prospective study of patients undergoing repair of non-ruptured abdominal aortic aneurysm between 1999 and 2003 was to evaluate and compare risk factors for mortality after surgery, to determine a complex of informative factors for lethal outcome, and to define patient risk groups. Logistic regression analysis revealed a complex of informative factors, including female gender, previous myocardial infarction, age greater than 75 years, and clinical course of abdominal aortic aneurysm as important indicators for lethal outcome. A risk score model identified low-, moderate- and high-risk groups with mortality rates of 2.9%, 8.0% and 44.4%, respectively.
abdominal aortic aneurysmsurgical treatmentmortality rate
==== Body
Introduction
Care of patients with abdominal aortic aneurysm (AAA) has been a benchmark of progress in vascular surgery for more than 50 years. By the early 1990s, elective repair of AAA was regularly carried out with a mortality rate of less than 5% [1]. Risk of rupture increases rapidly after the AAA reaches 5 cm in maximal diameter. Most surgeons therefore recommend surgical repair of AAAs 5 cm in diameter or larger and conservative treatment of smaller AAAs. However, rupture sometimes occurs with AAAs less than 5 cm in diameter, whereas some AAAs that are 5 cm or larger in diameter never rupture [2].
The aim of this study was threefold: (1) to evaluate and compare risk factors that influence the outcome of treatment in patients undergoing surgery for non-ruptured AAA, (2) to identify a complex of informative factors for lethal outcome, and (3) to define various risk groups.
Materials and methods
Patients
From January 1999 to August 2003 100 patients with a diagnosis of AAA were treated surgically at the Department of Vascular Surgery at the Kaunas University of Medicine. Thirty-one patients (31%) had surgery for a ruptured AAA and were excluded from our study. The 69 remaining patients (69%) underwent surgery for a non-ruptured AAA; these cases were analyzed in our study. There were 56 bifurcated (37.7% biiliacal, 29% bifemoral, and 14.5% femoral-iliac) and 13 (18.8%) straight grafts. All patients analyzed in this study had a midline laparotomy incision, and during AAA resection, the abdominal aorta was clamped below the renal arteries. Sixty-one patients survived while eight patients (11.6%) died within 30 days after surgery. Variables recorded for each patient were categorized as preoperative, intraoperative or postoperative.
Preoperative variables
Cardiac function
All patients had consultations with an internist and a cardiologist. Cardiac ultrasonography was performed, and left ventricular ejection fraction (LVEF) was estimated. Patients without coronary artery disease (CAD) or with minor CAD included asymptomatic patients, patients with normal electrocardiogram (ECG) results, and patients in whom myocardial infarction (MI) was diagnosed with electrocardiography more than 6 months earlier.
Renal system
Patients were considered as having normal or only minor renal dysfunction when the serum creatinine level was less than 110 μmol/L.
Respiratory system
Patients with normal pulmonary function included asymptomatic patients with normal results on chest radiographs and normal results on respiratory function assessments.
Abdominal aorta
AAA diameter was measured during examination of the patient with computerized tomography (CT) and ultrasonography. Forty-two patients (61%) were examined with aortography.
Intraopeartive variables
Operative risk was assessed according to the American Society of Anesthesiologists (ASA) classification. The following variables were recorded: type of operation, abdominal aorta "cross-clamping" time, declamping hypotension, blood loss during operation, urine output during operation (ml per hour) and total operation time.
Postoperative variables
We recorded the following variables: durability of artificial pulmonary ventilation, length of stay in intensive care, length of hospital stay after operation, and total length of hospital stay. Complications during the postoperative period were not analyzed in this study.
Statistical analysis
All variables defined above were subjected to univariate analysis using Microsoft Excel 2000 to test their influence on the mortality. Fisher's exact test was performed with SPSS 10.0 for Windows and Statistics v.5 for Windows. The complex influence of several predictors on the probability of lethal outcome (LO) was evaluated using the multidimensional logistic regression model [3]. The significance of variables in the logistic model was evaluated using the likelihood ratio (G2) and Wald statistics. When using the multidimensional logistic model coefficients, we developed a risk score model for the evaluation of complexes of risk factors. The accuracy of significant variables for prediction of LO was estimated by means of Receiver Operating Characteristic (ROC) curves. Variables that were statistically significant (p < 0.05) between groups of survivors and deceased were entered into a maximum probability logistic regression program. For each significant variable in the logistic regression analysis, the odds ratio (OR) was calculated (95% confidence interval). In order to determine low-, medium-, and high risk groups, we used LO probability estimation in the multinomial regression model, or logit function of the probability estimation log (p/(1-p)).
Results
Sixty-nine patients underwent surgery for non-ruptured AAA. For each patient a set of variables were compared between the surviving group of patients and the group of deceased. The variables that differed significantly between the groups of survivors and deceased were age, blood loss, days of hospitalization, days of hospitalization after operation and LVEF as given in Table 1. Variables like aneurysm size, diuresis, duration of operation and aortic cross clamping time did not differ significantly between the groups (not shown).
Table 1 Characteristics of patients undergoing surgical repair of non-ruptured abdominal aortic aneurysm (95% confidence interval of mean)
Variable Survivors Deceased p
n = 61 n = 8
Age (years) [69.0; 73.2] [73.2; 85.7] 0.02
Blood loss (ml) [1442; 1908] [310; 6364] 0.01
Length of hospitalization (days) [17.5; 23.1] [2.7; 7.27] 0.006
Length of hospitalization after operation (days) [11.3; 15.9] [0.18; 3.2] 0.007
LVEF (%) [43.4; 47.7] [30.9; 45.0] 0.019
LVEF-left ventricular ejection fraction
Table 2 shows additional groups of binominal variables that differed significantly with respect to mortality rate. We found significantly higher mortality rate among patients above 75 years, among females, as well as among patients with previous MI, with symptomatic course of AAA, with insufficient respiratory function or with insufficient renal function. In our material the mortality rate was not significantly higher among patients with ischemic heart disease, arterial hypertension, aorto coronary bypass operation, diabetes mellitus or LVEF below 40% (not shown).
Table 2 Binominal variables among 69 patients operated for non-ruptured abdominal aortic aneurysm.
Variable Number of operated patients Number of deceased patients (%) P (Fisher)
Age
≤75 yrs. 51 3 (5.9%)
>75 yrs. 18 5 (38.5%) 0.024
Distribution by sex:
Males 52 5 (9.6%)
Females 17 3 (17.6%) 0.008
MI in anamnesis
Yes 34 7 (20.6%)
No 35 1 (2.9%) 0.025
Clinical course of AAA
Asymptomatic 48 3 (6.3%)
Symptomatic 21 5 (23.8%) 0.04
Insufficient respiratory function
Yes 18 5 (27.8%)
No 51 3 (5.9%) 0.024
Insufficient renal function
Yes 11 3 (27.2%)
No 58 5 (8.6%) 0.01
MI-myocardial infarction
The ability of significant variables to predict LO was estimated by means of ROC curves. The area under the ROC curve is a measure of how well the groups are separated. The ROC curve's position above the mean line demonstrates the capability of the method to predict lethal outcome with some degree of precision. An area of 1 represents a perfect test. In our study, the area under the age curve was 0.778; CI [0.616; 0.94]; p = 0.011; the area under the sex distribution curve was 0.714; CI [0.5; 0.92]; p = 0.050; and the area under previous MI was 0.716; CI [0.55; 0.883]; p = 0.048 (not shown). Other variables were not found to be significant as estimated by the ROC curve.
To evaluate the informative parameters for LO, we used logistic regression. Age, previous MI, insufficiency of respiratory function, blood loss, LVEF and clinical course of AAA were confirmed as informative factors for LO (Table 3). Previous MI signified the highest risk; the value of this parameter was 8.8. The risk values associated with insufficiency of respiratory function and clinical course of AAA were somewhat lower (6.15 and 4.69, respectively). Each incremental blood loss of 100 ml during surgery was calculated to possess a risk value of 1.05. In our material the following variables were not considered as informative, size of diuresis, insufficiency of renal function and arterial hypertension.
Table 3 Prognostic values of the informative parameters for lethal outcome
Parameter χ2 p OR CI
Age, years 7.36 0.006 1.14 [1.03–1.26]
Previous MI 5.85 0.016 8.8 [1.02–76.06]
Clinical course of AAA 4.01 0.045 4.69 [1.004–21.87]
Blood loss, ml 5.48 0.019 1.05 [0.99–1.11]
Insufficient respiratory function 5.42 0.02 6.15 [1.3–29.17]
LVEF, % 5.8 0.016 0.89 [0.81–0.99]
AAA – abdominal aortic aneurysm, OR – odds ratio, LVEF – left ventricular ejection fraction, CI – confidence intervals, MI – myocardial infarction.
A logistic model for LO was established by using an informative parameter method, taking into consideration the strong correlation between some of the variables. For this model, the optimal criterion was a p-value that was adequate for the logistical model criterion G2. The logistic regression model included the variables as long as the p-value was significantly decreasing. Two models turned out to be able to predict LO. One model included the variables patient gender, previous MI and age above 75 years:
p(LO) = 1/1+exp{1.83 × gender + 2 × previous MI + 1.44 × (age>75 years) - 6.5}
χ2 = 14.5 p = 0.023
A second model included the variables patient gender, previous MI and clinical course of AAA:
p(LO) = 1/1+exp{1.8 × gender + 2.3 × previous MI + 1.12 × (clinical course of AAA) - 7.2}
χ2 = 15.1 p = 0.0017
The risk score model allows evaluation of the groups for risk stratification. Score evaluation was performed by eβ. The value of female gender (β = 1.83) was 6 (6.21); the value of previous MI (β = 2) was 7 (7.4); and the value of age greater than 75 years (β = 1.44) was 4 (4.23). The maximum sum of risk scores was 17. In 54 patients (78%), the score was less than 10, and mortality in this group was 3.7% (two patients). In 15 patients (22%), the score was higher than 10, and mortality in this group was 40% (six patients).
Using the second logistic model, the value of female gender (β = 1.8) was 6 (6.1), the value of previous MI (β = 2.3) was 10 (9.94), and the value of clinical course of AAA (β = 1.12) was 3 (3.06). The maximal sum of scores was 19. In 35 patients (51%), the sum was less than 13, and mortality in this group was 2.9% (1 patient). Twenty-five patients (36%) had scores of 13 to 16, and two patients (8.0%) in this group died postoperatively. Nine patients (13%) had scores higher than 16; four patients (44.4%) from this group died postoperatively. These findings were used for establishing low-, moderate-, and high-risk groups (Table 4). The prognostic values of both models is given by means of ROC curves in Fig. 1. The area under the 1st logistic model curve was 0.78 (CI [0.6; 0.97]; p = 0.009) and the area under the 2nd logistic model curve was 0.83; (CI [0.63; 1.02]; p = 0.003). These data suggest that both models predict LO fair enough but the 2nd logistic model demonstrates a slightly better sensitivity.
Table 4 Risk groups according to the sums of scores in model 2
Risk groups Score Mortality (%)
Low-risk group <13 2.9
Moderate-risk group 13–16 8.0
High-risk group >16 44.4
Figure 1 ROC curves of the logistic models: the 1st model (---) included gender, previous myocardial infarction and age above 75 years; the 2nd model (_ _ _ _) included gender, previous myocardial infarction and clinical course of abdominal aortic aneurysm.
Discussion
The 30-day postoperative mortality rate reported in the literature ranges from 0 to 10.5%. In our material it is slightly higher, 11.6%, probably due to the small number of patients. In a previous report based on a large number of patients collected from studies published between 1985 and 1997, the mortality rate was 5.8% [4].
In the elective situation, advanced age and cardiac, pulmonary, and renal disease increase the risk of postoperative mortality [5]. As in previous studies, age, renal dysfunction, and previous MI were found to be strong independent predictors of postoperative death [6,7]. Recent data suggest that female gender may increase mortality associated with repair of AAA [7,8]. We found that age greater than 75 years, previous MI, clinical course of AAA, and female gender were informative factors for lethal outcome. Steyerberg and colleagues developed a clinical prediction model to estimate the operative mortality risk for individual patients undergoing elective AAA repair [9]. With respect to informative parameters, our logistic model was similar to that of Steyerberg. However, in our material we found that size of the aneurysm did not influence perioperative mortality.
Blood loss during the operation was a statistically significant factor for increased mortality. Other investigators have also concluded that blood loss during surgery is an informative factor for lethal outcome [10].
All postoperative complications increase length of stay in intensive care. However, postoperative complications were not analyzed in our study.
Conclusion
Logistic regression analysis revealed a combination of informative factors for mortality after surgical repair of non-ruptured AAA. In one model these factors were female gender, previous MI, and age greater than 75 years. In a second model female gender, previous MI, and clinical course of AAA were important predictors of lethal outcome. The latter risk score model allows identification of groups of patients running a high risk for postoperative death. Mortality was 2.9% in the low-risk group, 8.0% in the moderate-risk group, and 44.4% in the high-risk group.
==== Refs
Thompson RW Reflections on the pathogenesis of abdominal aortic aneurysms Cardiovascular Surgery 2002 10 4 389 394 12359414 10.1016/S0967-2109(02)00042-X
Lindholt JS Considerations and experiences of screening for abdominal aortic aneurysm Ph D thesis 1998 University of Aarhus 1 109
Hosmer DW Lemeshow S Applied logistic regression 1989 New York: John Wiley
Becquemin JP Chemla E Chatellier G Allaire E Melliere D Desgranges P Peroperative factors influencing the outcome of elective abdominal aorta aneurysm repair Eur J Vasc Endovasc Surg 2000 20 1 84 89 10906304 10.1053/ejvs.2000.1102
UK Small Aneurysm Trial Participants Mortality results for randomised controlled trial of early elective surgery or ultrasonographic surveillance for small abdominal aortic aneurysms Lancet 1998 352 1649 1655 9853436 10.1016/S0140-6736(98)10137-X
Hollier L Taylor L Ochsner J Recommended indications for operative treatment of abdominal aortic aneurysms. Report of a subcommittee of the Joint Council of the Society for Vascular Surgery and the North American Chapter of the International Society for Cardiovascular Surgery J Vasc Surg 1992 15 1046 1056 1597887 10.1067/mva.1992.37163
D'Angelo F Vaghi M Zorzoli C Gatti S Tacconi A Is age an important risk factor for the outcome of elective abdominal aneurysm surgery? J Cardiovasc Surg (Torino) 1993 34 153 155 8320250
Katz DJ Stanley JC Zelenock GB Operative mortality rates for intact and ruptured abdominal aortic aneurysms in Michigan: an eleven-year statewide experience J Vasc Surg 1994 19 804 817 8170034
Steyerberg EW Kievit J de Mol Van Otterloo JCA van Bockel JH Eijkemans MJ Habema JD Perioperative mortality of elective abdominal aortic aneurysm surgery: A clinical prediction rule based on literature and individual patient data Arch Intern Med 1995 155 1998 2004 7575054 10.1001/archinte.155.18.1998
Amundsen S Skjaerven R Trippestad A Soreide O Abdominal aortic aneurysms- a study of factors influencing postoperative mortality. Norwegian Aortic Aneurysm Trial Eur J Vasc Surg 1989 3 5 405 409 2680610 10.1016/S0950-821X(89)80046-5
|
16146566
|
PMC1215499
|
CC BY
|
2021-01-04 16:47:32
|
no
|
Curr Control Trials Cardiovasc Med. 2005 Sep 7; 6(1):14
|
utf-8
|
Curr Control Trials Cardiovasc Med
| 2,005 |
10.1186/1468-6708-6-14
|
oa_comm
|
==== Front
Epidemiol Perspect InnovEpidemiologic perspectives & innovations : EP+I1742-5573BioMed Central London 1742-5573-2-71603365210.1186/1742-5573-2-7MethodologyReporting incidence from a surveillance system with an operational case definition of unknown predictive value positive Kegler Scott R [email protected] Office of Statistics and Programming, National Center for Injury Prevention and Control, Centers for Disease Control and Prevention, 4770 Buford Highway NE, Mailstop K59, Atlanta GA 30341-3724, USA2005 20 7 2005 2 7 7 2 5 2005 20 7 2005 Copyright © 2005 Kegler; licensee BioMed Central Ltd.2005Kegler; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
When reporting incidence rate estimates for relatively rare health conditions, associated case counts are often assumed to follow a Poisson distribution. Case counts obtained from large-scale electronic surveillance systems are often inflated by the presence of false positives, however, and adjusted case counts based on the results of a validation sample will have variances which are hyper-Poisson. This paper presents a simple method for constructing interval estimates for incidence rates based on case counts that are adjusted downward using an estimate of the predictive value positive of the surveillance case definition.
==== Body
Introduction
Large-scale surveillance for selected medical or health conditions often relies on electronic data sources which provide comprehensive coverage of a given population. For example, the Centers for Disease Control and Prevention conduct surveillance of brain injuries involving hospitalization or death, based on electronic hospital discharge and vital statistics data received from twelve to fifteen states each year [1]. To identify cases, electronic records are scanned for specified diagnosis codes which collectively form the operational case definition. The resulting case counts are subsequently combined with population data to estimate incidence rates.
As with most surveillance methods, an operational case definition as described above may admit some records that do not represent true cases under a strict clinical definition ("false positives") and may also fail to capture some records representing true cases ("false negatives"). The customary terms reflecting these aspects of an operational case definition are predictive value positive (PVP) and sensitivity, defined in the present context as the conditional probabilities [2]:
PVP = Pr{case meets clinical definition | case detected under operational definition};
sensitivity = Pr{case detected under operational definition | case meets clinical definition}.
Depending on the extent to which false positives and/or false negatives are believed to influence the surveillance process, it may be appropriate to use estimates of PVP and/or sensitivity to adjust incidence rate estimates accordingly. It is not generally possible to assess PVP or sensitivity using electronic surveillance data alone. The most direct approach to obtaining the additional data required for estimation of PVP involves manual review of medical records for a random sample of provisional cases identified via the operational case definition. Obtaining the additional data necessary for estimation of sensitivity may be more labor-intensive, particularly when considering an uncommon condition. Without additional "markers" (apart from the operational case definition) to narrow the scope of review, it may be necessary to select a very large sample of general medical records in order to identify enough true cases to support a stable estimate of sensitivity.
The methodology described in this paper is oriented to surveillance of relatively rare health conditions. Because validation data quantifying the influence of false positives will typically be easier to obtain than data quantifying the influence of false negatives in this setting, the development concentrates on incidence rate estimates reflecting adjustments for PVP. This emphasis is not intended to diminish the potential influence of false negatives; rather, it reflects the logistical difficulties associated with obtaining data on false negatives as part of ongoing surveillance. If there is sufficient doubt surrounding the sensitivity of case ascertainment for any particular surveillance process, the proposed methodology should be applied with due caution.
Analysis
For a given surveillance period, it is assumed that case confirmation data are available for a random sample (selected without replacement) of provisional cases. Data obtained through such validation efforts allow estimation of PVP as well as adjustments to case counts to eliminate the bias due to false positives. To illustrate, suppose that for a set period (e.g., one year) of observation:
N = size of the at-risk population covered by the surveillance system;
M = count of provisional cases detected under the operational case definition;
MT = count of true cases (unknown) among the provisional cases;
MF = count of false positive cases (unknown) among the provisional cases = M - MT;
S = number of provisional cases sampled for case confirmation;
CT = count of confirmed true cases among those sampled;
CF = count of cases determined to be false positives among those sampled = S - CT.
The usual estimate of PVP is given by [3]:
= CT/S = CT/(CT + CF).
Noting that is definable only when M > 0 (assuming also that S > 0) a reasonable estimate of the population of true cases which eliminates the false positive bias is:
Case counts obtained through comprehensive surveillance may be considered inherently variable even though they are essentially census-level quantities, in the sense that a case count can be viewed as representing one observation from a hypothetically repeatable process [4-7]. For relatively rare conditions such case counts are often assumed to follow a Poisson distribution [6,7]. For example, suppose that all M provisional cases were to be reviewed so that the count of true cases MT could be determined. When reporting the corresponding incidence rate R = MT/N one might also make use of the variance estimate , based on the assumption that MT represents one observation from a Poisson process [6,7]. Due to the estimation of PVP, however, the adjusted case count cannot be treated in a similar fashion. Depending on the validation sample and the underlying PVP, for example, Var() can be well in excess of the variance that would be estimated under the assumption that simply follows a Poisson distribution.
The remainder of this paper addresses three aspects of the problem outlined above: (i) a simple model for the true and false positive case counts within the defined framework, (ii) selected properties of under a broadly applicable validation sample plan, and (iii) the relative frequency of coverage for interval estimates formulated using these properties.
A Case Count Model
To evaluate the proposed estimator , a working model characterizing the process underlying the case counts M, MT, and MF is needed. For a given at-risk population and surveillance period it will be assumed that the provisional case count M is generated according to a Poisson process with parameter λ. Each provisional case, independent of other provisional cases, will be assumed to be a true case with probability equal to the underlying PVP. These assumptions are reflected in the following mixture model [8]:
M ~ POI(λ);
MT|M ~ BIN(M, PVP)
where POI denotes the Poisson distribution and BIN denotes the binomial distribution. The count of false positive cases is implicitly given by MF = M - MT. It is well-established that under this type of decomposition MT and MF are independent Poisson random variables such that MT ~ POI(τ) and MF ~ POI(φ), where τ = λ·PVP and φ = λ·(1-PVP) [9,10]. In this model, the parameter λ represents the average size of the recurring count of provisional cases and τ represents the average size of the recurring count of true cases among the provisional cases. The quantity 1/PVP can be viewed as the factor by which the count of true cases is inflated (on average) under the operational case definition. Finally, the parameters λ, τ and φ are implicitly dependent on the size of the at-risk population N; however, the functional form of this dependency is not important in the present development.
A Validation Sample Plan
This section examines several important properties of the estimator when a fixed fraction of provisional cases are sampled for confirmation. The properties presented are derived in Appendix A. Letting 0 < f < 1 denote the fixed sampling fraction, assume that the sample size S = where the quantity f·M is rounded up. Under this procedure:
E[] = τ (2)
and when f·λ is sufficiently large:
Equality (2) indicates that is an unbiased estimator for the mean recurring count of true cases. The first component τ on the right-hand side of (3) represents the variance of the true case count MT. The second component approximates the addition to variance that results from the case count adjustment based on . Note that for any given PVP the variance inflation factor is essentially constant as a result of holding the sampling fraction fixed.
It is noted in passing that when case populations are typically small, it may be feasible to adopt the practice of confirming all provisional cases. Under this approach will be equivalent to the true case count MT and it follows that ~ POI(τ). Based on familiar properties of the Poisson distribution [8] it follows that E[] = Var() = τ and customary analysis methods are applicable.
Application
The remaining objective is the formulation of a simple method for constructing interval estimates for τ and the corresponding incidence rate. From (2) it is already known that is an unbiased estimator of τ. In Appendix B it is shown that the following estimator is nearly unbiased for the right-hand side of (3):
Based on (4) an approximate (1-α)·100% confidence interval (adjusted for the false positive bias) for the recurring case count τ is given by:
where zα/2 represents the appropriate quantile of the standard normal distribution. The corresponding interval estimate for the population-based incidence rate is:
where it will be recalled that N is the size of the at-risk population under surveillance. As an example, suppose that an interval estimate providing 95% relative frequency of coverage is desired for the population-based incidence rate. Table 1 shows the relative frequency with which interval (5) covers the underlying incidence rate in repeated Monte Carlo simulations involving various underlying values of PVP, λ, and f. For several cells f·λ is small and the coverage is below the nominal (95%) level, providing an illustration of where the interval estimation procedure begins to break down. In the remaining table cells coverage is close to the nominal level.
Table 1 Estimated Relative Coverage Frequencies of a Nominal 95% Interval with Variance Correction.
PVP = 0.70 PVP = 0.80 PVP = 0.90
f λ = 100 λ = 500 λ = 1000 λ = 100 λ = 500 λ = 1000 λ = 100 λ = 500 λ = 1000
0.10 0.92 0.94 0.95 0.92 0.95 0.95 0.94 0.95 0.95
0.25 0.94 0.95 0.95 0.95 0.95 0.95 0.95 0.95 0.95
0.50 0.95 0.95 0.95 0.95 0.95 0.95 0.94 0.95 0.95
To illustrate the importance of the correction to the variance, Table 2 shows the relative coverage frequencies (again based on repeated simulations) if the adjusted case counts are simply assumed to follow a Poisson distribution. It is apparent that for smaller sampling fractions, coverage is well below the nominal level even with the larger case populations.
Table 2 Estimated Relative Coverage Frequencies of a Nominal 95% Interval w/o Variance Correction.
PVP = 0.70 PVP = 0.80 PVP = 0.90
f λ = 100 λ = 500 λ = 1000 λ = 100 λ = 500 λ = 1000 λ = 100 λ = 500 λ = 1000
0.10 0.73 0.70 0.68 0.78 0.77 0.76 0.86 0.84 0.84
0.25 0.84 0.85 0.85 0.87 0.88 0.88 0.92 0.91 0.91
0.50 0.91 0.92 0.91 0.92 0.93 0.93 0.93 0.94 0.94
Extensions to independent subgroups (e.g., age groups) and aggregates (e.g., age-adjusted rates) are straightforward. Provided that subgroup boundaries do not divide the surveillance population too finely, the error associated with the interval estimation method described above should remain minimal.
Conclusion
This paper was motivated by considerations related to analysis of data from the brain injury surveillance system mentioned in the introduction. Beginning with surveillance year 2000, a number of participating states identified provisional cases which were subsequently determined to be false positives upon in-depth review. Preliminary estimates of PVP were observed to fall close to 0.9 for some states, suggesting the need for adjusted incidence rate estimates. This issue is also relevant in a broader context, as a wide range of PVP estimates have been reported for other surveillance systems [11].
Adjustments to incidence rate estimates to eliminate the false positive bias are straightforward. However, since the PVP estimates used to make such downward adjustments are subject to random variation, the adjusted rates have an additional source of variation beyond what is usually assumed. Interval estimates failing to account for this fact may have coverage frequencies well below the nominal level. This paper presents a simple method of interval estimation for rates that have been adjusted to remove the bias due to false positives, applicable in large-scale surveillance settings.
The methodology presented does not address the potential bias associated with false negatives. In situations where validation data also support estimation of sensitivity, surveillance case counts could be further adjusted to reduce or eliminate such bias. This in turn would introduce another source of variation in the adjusted case counts and associated rates. Other types of sampling plans might also be considered. For example, a fixed sample size s* might be preferred, in which case S = min(s*, M) and an alternate expression for Var() would result. Technical details aside, the essential point is that data available from validation samples can have a nontrivial influence on point and interval estimates, and should be factored into surveillance statistics whenever feasible.
Appendix A. Moments of the Estimator
In the sampling procedure considered, the size of the validation sample depends on the provisional case count M. To make the analysis generic, the sample size will be denoted by s(M) where s(·) depends on the particular sampling procedure but is assumed positive whenever M > 0. The PVP-adjusted case count (1) can then be defined more precisely as:
where implicitly = CT/s(M). When M > 0 the distribution of CT conditional on M and MT is hypergeometric [12], that is, CT|M, MT ~ HYP(s(M), MT, M). It is not difficult to show that when M > 0 the distribution of CT conditional on M only is binomial, that is, CT|M ~ BIN(s(M), PVP). It follows that E[|M] = M·PVP for M ≥ 0. Applying principles of conditional expectation [8] it is readily established that is an unbiased estimator of τ = λ·PVP:
E[] = E[E[|M]] = E[M·PVP] = λ·PVP.
To determine Var() it is convenient to employ the following variance decomposition [8]:
Var() = E[Var(|M)] + Var(E[|M]).
Since E[|M] = M·PVP it follows that Var(E[|M]) = λ·PVP2. Evaluation of the first component of variance is more complicated. Defining:
it follows from (A.1) and the fact that CT|M ~ BIN(s(M), PVP) when M > 0 that:
Var(|M) = PVP·(1-PVP)·g(M).
The task is thus reduced to determining E[g(M)]. When s(M) = it holds that g(M) ≤ M/f and hence that E[g(M)] ≤ E[M/f] = λ/f. Given fixed f the upper bound is a good approximation provided that λ is sufficiently large, so that E[g(M)] ≌ λ/f and E[Var(|M)] ≌ PVP·(1-PVP)·λ/f. Combining variance components and simplifying results in:
Numerical calculation of Var() across a range of values for PVP, λ, and f shows that for f ≥ 0.01 and f·λ ≥ 50, the relative error of (A.2) is less than 0.01.
Appendix B. An Estimate of Var()
The following is proposed as an estimator of the right-hand side of (A.2):
Defining:
it follows from the treatment in Appendix A that the expected value of the variance estimator (B.1) conditioned on M is:
Then, since it follows that:
When s(M) = it holds that h(M) ≤ 1/f and hence that E[h(M)] ≤ 1/f. Given fixed f the upper bound is a good approximation provided that λ is sufficiently large. Substituting 1/f in place of E[h(M)] results in:
Algebraic simplification results in:
As f·λ becomes large, approximation (A.2) results.
Competing interests
The author(s) declare that they have no competing interests.
==== Refs
Thurman DJ Sniezek JE Johnson D Greenspan A Smith SM Guidelines for Surveillance of Central Nervous System Injury 1995 Atlanta: Centers for Disease Control and Prevention
Greenland S Rothman KJ, Greenland S Basic methods for sensitivity analysis and external adjustment Modern Epidemiology 1998 2 Philadelphia: Lippincott-Raven Publishers 343 357
Romaguera RA German RR Klaucke DN Teutsch SM, Churchill RE Evaluating public health surveillance Principles and Practice of Public Health Surveillance 2000 New York: Oxford University Press 176 193
Keyfitz N Sampling variance of standardized mortality rates Human Biology 1966 38 309 317
Brillinger DR The natural variability of vital rates and associated statistics Biometrics 1986 42 693 734 3814721
Greenland S Rothman KJ Rothman KJ, Greenland S Introduction to categorical statistics Modern Epidemiology 1998 2 Philadelphia: Lippincott-Raven Publishers 231 252
Anderson RN Minino AM Fingerhut LA Warner M Heinen MA Deaths: Injuries, 2001 National Vital Statistics Reports 2004 52 Hyattsville MD: National Center for Health Statistics
Casella GC Berger RL Statistical Inference 1990 Belmont CA: Duxbury Press
Ross SM Introduction to Probability Models 1989 4 San Diego: Academic Press
Taylor HM Karlin S An Introduction to Stochastic Modeling 1984 Orlando: Academic Press
German RR Sensitivity and predictive value positive measurements for public health surveillance systems Epidemiology 2000 11 720 727 11055638 10.1097/00001648-200011000-00020
Cochran WG Sampling Techniques 1977 3 New York: Wiley
|
16033652
|
PMC1215500
|
CC BY
|
2021-01-04 16:36:38
|
no
|
Epidemiol Perspect Innov. 2005 Jul 20; 2:7
|
utf-8
|
Epidemiol Perspect Innov
| 2,005 |
10.1186/1742-5573-2-7
|
oa_comm
|
==== Front
Emerg Themes EpidemiolEmerging Themes in Epidemiology1742-7622BioMed Central London 1742-7622-2-81613732710.1186/1742-7622-2-8Analytic PerspectiveAssessing observational studies of medical treatments Hartz Arthur [email protected] Suzanne [email protected] Mary [email protected] Douglas [email protected] Yogita [email protected] G Mustafa [email protected] Kjell [email protected] University of Iowa, College of Medicine, Department of Family Medicine, Iowa City, IA 52242 USA2 University of Iowa, College of Medicine, Department of Family Medicine, Iowa City, IA 52242 USA3 University of Iowa, College of Medicine, Department of Family Medicine, Iowa City, IA 52242 USA4 VA Medical Center, 500 East Veterans Street, Tomah, WI 54660 USA5 University of Iowa, College of Medicine, Department of Family Medicine, Iowa City, IA 52242 USA6 Section of Community Psychiatry, St. George's Hospital Medical School, London, UK7 Family Practice Clinic, North Colorado Medical Center, Greeley, Colorado USA2005 1 9 2005 2 8 8 15 3 2005 1 9 2005 Copyright © 2005 Hartz et al; licensee BioMed Central Ltd.2005Hartz et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Previous studies have assessed the validity of the observational study design by comparing results of studies using this design to results from randomized controlled trials. The present study examined design features of observational studies that could have influenced these comparisons.
Methods
To find at least 4 observational studies that evaluated the same treatment, we reviewed meta-analyses comparing observational studies and randomized controlled trials for the assessment of medical treatments. Details critical for interpretation of these studies were abstracted and analyzed qualitatively.
Results
Individual articles reviewed included 61 observational studies that assessed 10 treatment comparisons evaluated in two studies comparing randomized controlled trials and observational studies. The majority of studies did not report the following information: details of primary and ancillary treatments, outcome definitions, length of follow-up, inclusion/exclusion criteria, patient characteristics relevant to prognosis or treatment response, or assessment of possible confounding. When information was reported, variations in treatment specifics, outcome definition or confounding were identified as possible causes of differences between observational studies and randomized controlled trials, and of heterogeneity in observational studies.
Conclusion
Reporting of observational studies of medical treatments was often inadequate to compare study designs or allow other meaningful interpretation of results. All observational studies should report details of treatment, outcome assessment, patient characteristics, and confounding assessment.
comparative studiesepidemiologic research designmeta-analysisreproducibility of resultsreporting criteria for observational studies
==== Body
Introduction
As compared to randomized controlled trials of medical interventions, observational studies (OSs) are likely to be more timely, less expensive, and include patients more representative of usual clinical practice. In addition, OSs avoid ethical issues caused by compromising the patients' or physicians' therapeutic choices. However, their validity is vigorously debated [1,2]. Concerns about validity were heightened by a well-publicized randomized controlled trial (RCT) that found an increased risk of heart disease for women on hormone replacement therapy [3]. This study contradicted results from several previous high profile OSs [4-6].
It is possible that OSs are particularly ill-suited to evaluate hormone replacement therapy. The choice of this therapy is greatly influenced by patient ideas about youth and femininity, which may be highly confounded by unmeasured factors affecting health. Not all OSs found a decreased risk, however [7,8], and one study found evidence that adjusting results for socioeconomic status yielded similar results to randomized controlled trials [9]. Other factors contributing to differences between the types of studies may include design features that do not necessarily influence validity, such as the older age of women in the randomized trials.
Even if OSs of hormone replacement therapy are shown to be invalid, the observational study design may still have a role in the assessment of other medical treatments. For example, these studies may give good results for evaluation of a surgical procedure that is determined primarily by physician familiarity with a specific treatment or by treatment availability. Support for the validity of some OSs comes from reviews that found that observational and randomized studies often give similar results [10-13].
The present study investigated the comparisons of OSs and RCTs in more depth. In addition to the comparisons, we examined design features of the OSs that could have influenced these comparisons. In the process of this examination, we assessed how well OSs of medical treatments were reported.
Materials and methods
Each of the reviewed studies compared outcomes of a given medical treatment to outcomes of a comparison group, which was most often standard therapy. Studies were selected from articles that compared results from observational and randomized studies previously assessed in meta-analyses or systematic reviews [13,14]. The reason for including only studies previously included in systematic reviews or meta-analyses was to increase the likelihood that all articles on a given topic were reviewed. Meta-analyses of fewer than four OSs were excluded from our review, because they provided limited ability to evaluate influences of study design characteristics on results of OSs.
Study characteristics abstracted
To determine characteristics that should be abstracted, we reviewed the literature as to how RCTs should be reported [15,16] and evaluated [17], how OSs should be reported and evaluated [18], and how patient and treatment characteristics influence RCTs [19]. Some information from RCTs was not relevant for OSs (e.g., blinding of the randomization process), some was relevant but not available in OSs (e.g., protocols for administering the primary treatment and managing intermediate outcomes), and some was not relevant to comparing studies (e.g., power of the study when data have already been collected).
Based on previous literature and our own experience with OSs, we developed the schema in Table 1 for study characteristics that could influence results by influencing either the study's applicability or validity. Factors that could influence applicability include specific characteristics of the treatments, outcomes, or subjects; results that apply to studies using specific types of treatments, outcome measurements or subjects may be valid, but may not be reproduced by other studies using different types of treatments, outcome measurements or subjects.
Table 1 Study Characteristics That Influence Results
Characteristics That Influence Applicability
Treatment specifics
Details of procedure
Details of ancillary treatments or management of intermediate outcomes
Outcome
Definition
Method of patient contact and assessment at follow-up
Length of follow-up
Patient characteristics
Study setting
Inclusion/exclusion criteria
Reported characteristics
Characteristics That Influence Validity
Information bias
Incorrectly ascertaining treatment or outcome
Selection bias
Pretreatment – subjects on different treatments have different risks
Post treatment – lost to follow-up depends on outcome and treatment
Confounding
Caused by pretreatment selection bias
Demonstrated by differences in risk factors between treatment groups
Possibly reduced by risk-adjustment
Factors that could influence validity include those that could contribute to confounding, selection, or information (also called measurement) bias [20]. Confounding arises when subjects who receiving one treatment differ in risk from subjects receiving another, independent of the effect of treatment. Selection bias occurs when the association between exposure and disease differs between those who complete a study and those in the target population. In cohort studies of medical treatments, such as those reviewed below, pre-treatment selection bias leads to confounding and post-treatment selection bias results from incomplete follow-up that differs according to both outcome and treatment. Information bias occurs when errors are made in assessing which treatment or outcome a patient had. Although post-treatment selection bias and information bias distort estimates of effect size, they were difficult to assess in the papers reviewed and were not recorded in our analysis.
This schema guided the type of information abstracted from the reviewed articles. Although it does not include all 27 items considered important for measuring the quality of OSs in one schema [18], it is conceptually simple and should include most study aspects that influence interpretation of results. For each article reviewed, we noted critical data elements omitted from the article.
We deemed that the likelihood of confounding would be increased if treatment choice were related to time, so that recent patients generally received one treatment, whereas patients from several years previously received another. Confounding could also be more likely if treatment was allocated on the basis of patient characteristics that contribute to treatment failure, either by the physician or through patient self-selection. Confounding was considered less likely if the physicians treating the patients used only one procedure. An implicit assumption in this criterion is that patient risk and quality of care are similar across physicians; this assumption may not be valid in all cases, but we wanted to judge the studies as generously as possible so that reports of deficiencies in these studies would be conservative. Another criterion for decreased likelihood of confounding was an abrupt change in patient care, so that all patients received one treatment before the change and all patients received another treatment after the change.
All data abstraction was from the original articles. Although the majority of articles were assessed independently by two different reviewers, some articles were only assessed by the same reviewer several months apart. Disagreements between reviewers, between reviews at different times, or between our reviewers and the published meta-analysis were resolved through discussion.
Statistical methods
Results were reported as statistically significant if p < 0.05, although p-values were often much lower. We used a 2-by-2 χ2-test for contingency tables to compare OS and RCT subjects for pooled failure rates on the same treatment. Significantly different failure rates for OS and RCT studies of one treatment in a comparison but not the other is a sensitive indication of possible confounding in the OSs. Significantly different failure rates for both treatments suggest that the two types of studies may differ with respect to features that influence failure rates (e.g., patients, outcome measures, specifics of the treatment, or uses of ancillary treatments). We also evaluated whether it might be worthwhile to search for important study factors that caused heterogeneity by examining variation among OSs for failure rates of a given treatment. The p-value for the statistical significance of this variation was determined using a 2-by-k χ2-test for contingency tables, where k was the number of studies that evaluated a given treatment. Pearson's correlation coefficient, denoted r, was used to compute the p-value for the association between the failure rates in the treatment group and the failure rates in the control group at the 0.05 significance level.
Statistical methods were used to combine odds ratios from several studies and to test the difference between the summary odds ratios from the observational and randomized studies. To combine odds ratios from several studies and to find the standard error of the combined odds ratio, we used a fixed-effects calculation [21]. By using fixed- rather than random-effects calculations [22], we obtained smaller standard errors and decreased the chances of missing true differences. However, this method may increase the likelihood of finding spurious differences.
We tested the difference between two odds ratios using the equation
Z = (Ln1 - Ln2) / √(SE12 + SE22)
where Z has a normal distribution with mean zero and variance 1, Ln1 and Ln2 are the logarithms of the two odds ratios, and SE1 and SE2 are the standard errors of these logarithms. Heterogeneity in odds ratios was tested with the Breslow-Day test for homogeneity at the 0.05 significance level.
Results
Meta-analyses selected for review
The selected analyses are shown in Table 2. These analyses addressed 10 topics: anticoagulants for treatment of myocardial infarction, quinidine for atrial fibrillation, trial of labor for patients with a breech delivery, colposuspension compared to anterior colporrhaphy for urinary incontinence, colposuspension compared to needle suspension for urinary incontinence, transcutaneous electrical nerve stimulation (TENS) for treatment of postsurgical pain, early discharge following childbirth, hip screws for hip fracture, local anesthesia for patients with carotid endarterectomy, and hysterosalpingography (HSG) media on pregnancy.
Table 2 Meta-analyses Selected for Review
Brief Title Year Medical Condition Treatment 1 v Treatment 2 Failure Outcome No. of studies (RCT, OS) Reasons for excluding OSs previously compared to RCTs [13]
Anticoagulants 1977 Myocardial Infarction Control v Anticoagulants Mortality Ioannidis: (6, 12) Kunz: (6, 12) Three used alternately assigned controls [23-25].
Quinidine 1992 Atrial fibrillation Control v Quinidine Relapse into atrial fibrillation Ioannidis: (6, 5) Kunz: (6, 6) One did not have 3-month follow-up [53].
Trial of Labor 1995 Breech delivery No trial v Trial of Labor 5 minute Apgar Ioannidis: (2, 6) None excluded.
Colposuspension 1 1996 Incontinence Anterior colporrhaphy v Colposuspension No cure of incontinence Ioannidis: (4, 11) Five were in a foreign language [54-58].
Colposuspension 2 1996 Incontinence Needle suspension v Colposuspension No cure of incontinence Ioannidis: (3, 9) One used failed surgery instead of incontinence as an outcome [59]. One was a controlled trial [26]. Three were in a foreign language [55, 56, 60].
Transcutaneous Electrical Nerve Stimulation (TENS) 1996 Post-operative pain Control v TENS No pain relief Ioannidis: (2, 4) Kunz: (17, 19) None excluded.
Early Discharge 1997 Childbirth Conventional v Early Maternal Morbidity Ioannidis: (1, 3) Added one study from original meta-analysis [36].
Hip screw 1999 Hip fracture Fixed nail plates v Sliding hip screw Total complications Ioannidis: (1, 6) One "OS" was an RCT [27].
Local Anesthesia 1996 Carotid Endarterectomy General v Local Anesthesia Stroke or death Ioannidis: (3, 14) One from a non-peer reviewed abstract [61], 1 from unpublished data [62], one in a foreign language [63].
Hysterosalpingo-graphy (HSG) 1999 Infertility Water v Oil in Hysterosalpingography No Pregnancy Ioannidis: (5, 6) None excluded.
RCT = Randomized Controlled Trial
OS = Observational Study
With five exceptions, we considered as observational all studies that were considered observational in the reviewed meta-analyses [13]: three of these were excluded because they used alternately assigned controls [23-25], and two were RCTs [26,27]. We did not exclude studies that used historical controls. Seven additional studies that were not in English were excluded because we were not able to accurately abstract detailed information about them.
Some studies assessed more than one outcome. With one exception, we reported results for the same outcomes that were assessed in the study by Ioannidis et al [13]. The exception was the meta-analysis of quinidine [28]. The outcome used by Ioannidis et al. from that analysis was mortality, which was zero or near zero for most studies. We used relapse of atrial fibrillation following cardioversion, which was used by our other source of meta-analyses [14]. Failure rates were used to compute odds ratios not computed by the original studies. For some studies the success rates and odds ratios in the primary studies [29-32] differed from those reported by Ioannidis et al. [13] or the meta-analysis [28]. When there was a discrepancy, we used the rates reported in the primary studies. Rates in primary studies for positive endpoints (e.g. pregnancy) were converted to failure rates (e.g. no pregnancy).
Comparisons of observational and randomized studies
The comparison of the combined odds ratios for the two types of studies are shown in Figure 1. In general, the confidence intervals were wider for the RCTs than for the OSs, reflecting the larger sample sizes for the OSs. Wide confidence intervals for randomized controlled studies of trial of labor, transcutaneous electrical nerve stimulation (TENS), early discharge, and local anesthesia prevented meaningful comparisons for these treatment areas. The only treatment area for which the odds ratios differed significantly was studies of anticoagulants following an acute myocardial infarction.
Figure 1 Comparison of Confidence Intervals for Combined Odds Ratios from Observational Studies and Randomized Controlled Trials.
The observational and randomized studies differed with respect to several failure rates (see Table 3). For some treatment comparisons there were dissimilar failure rates between the types of studies for both treatment and control groups (TENS for postoperative pain and early discharge following childbirth), and for other treatment comparisons there were significant differences between study designs with respect to the rates for patients with the new treatments, but not for patients on the older treatments (quinidine for the treatment of atrial fibrillation and colposuspension versus two older treatments for urinary incontinence).
Table 3 Outcome Differences Between RCTs and OSs
Brief Title (outcome) Number of Studies Average Failure Rate (Number of Patients)
Control** Treatment
Anticoagulants (MI) [13, 14, 64]
RCT 6 17% (1748) 14% (2106)
OS 9 31%† (3615) 16%† (2598)
Quinidine (Afib) [13, 14, 28] 3 months
RCT 6 54% (390) 36% (413)
OS 5 61%* (200) 53% (342)
Trial of labor (Breech) [13, 65]
RCT 2 2% (128) 3% (182)
OS 6 5% (1043) 4%† (1552)
Colposuspension 1 (Incontinence) [13, 66]
RCT 2 33% (134) 12% (139)
OS 6 37%† (508) 26% (374)
Colposuspension 2 (Incontinence) [13, 66]
RCT 2 31% (132) 12% (139)
OS 4 32%† (190) 23%† (349)
TENS (Pain) [13, 14, 67]
RCT 2 18% (34) 3% (34)
OS 4 76%† (172) 56%† (136)
Early Discharge (Childbirth) [13, 68]
RCT 1 8% (38) 5% (93)
OS 4 21%† (379) 19%† (402)
Hip Screw (Hip Fx) [13, 69]
RCT 1 50% (26) 12% (33)
OS 5 35%† (290) 8% (560)
Local Anesthesia (CEA) [11, 13, 70]
RCT 3 5% (79) 5% (75)
OS 11 5%* (1509) 2% (1713)
HSG (Infertility) [11, 13, 71]
RCT 5 81% (527) 69% (302)
OS 6 74%† (734) 58%† (1072)
RCT = Randomized Controlled Trial; OS = Observational Study
MI = Myocardial Infarction
Afib = Atrial fibrillation
TENS = Transcutaneous electrical nerve stimulation
Hip Fx = Hip Fracture
CEA = Carotid Endarterectomy
HSG = Hysterosalpingography
** The treatment group is listed in the row title. As seen in Table 1, the control (i.e. comparison) groups are the negative of the listed treatment except for the following: Control group for colposuspension 1 is colporrhaphy, colposuspension 2 is needle suspension, hip fracture is fixed nail plates, CEA is general anesthesia, infertility is water soluble medium.
Significance testing was only done to test heterogeneity among the failure rates for the observational studies
* P < .05 for test of heterogeneity of failure rates combined to create the average
† P < .001 for test of heterogeneity of failure rates combined to create the average
As indicated in Table 3, several studies show considerable heterogeneity in results among OSs. For each treatment comparison there was statistically significant variation in failure rates for at least one of the treatments. There was statistically significant heterogeneity in the odds ratios for studies of anticoagulants, colporrhaphy, needle suspension, and hysterosalpingography. Despite small numbers of studies in each treatment area, failure rates were significantly correlated for studies of anticoagulants (r = 0.79, p = 0.01), trial of labor (r = 0.75, p = 0.08), early discharge (r = 0.99, p = 0.01), hip screws (r = 0.93, p = 0.02), and local anesthesia (r = 0.66, p = 0.03). This correlation might be explained by substantial differences among study features that influence failure rates.
Reporting of treatments and outcomes in OSs
Reporting details of the primary treatment, ancillary treatments, and management of intermediate outcomes was uniformly poor. Most aspects of outcome were also poorly reported. However, outcome definitions were generally well reported. Even here there were exceptions: one study of surgical treatment for incontinence defined the subjective outcome only as "cured" [33], and another defined it as "symptom free" [34].
Length of follow-up, which may substantially influence outcome and comparisons of treatments, was usually not well reported. Of the studies reviewed, only two studies of hysterosalpingography and five of surgical treatment for stress incontinence provided both the mean (or median) and range (or other measure of spread) of follow-up times. Eleven studies provided no follow-up information, and the remainder provided only one number (median, minimum, or undefined).
Considerations of patient selection in OSs
Even though the majority of studies were based on chart abstraction, none described methods for reducing selection or information bias.
Results from studies were sometimes combined, even though they differed with respect to potentially important patient characteristics. For example, studies of surgical treatment for incontinence varied with respect to exclusions due to previous surgery for incontinence, detrussor instability, and other pathologic findings. Another example is that criteria for studies of local anesthesia for carotid endarterectomy varied on the basis of whether patients were included who were simultaneously undergoing a coronary artery bypass grafting procedure or who had an acute stroke. Among studies of early discharge, one unique inclusion criteria was caesarian delivery [35] and another was primiparity [36]. Of the two studies of HSG that provided detail on inclusion and exclusion criteria, one required infertility for at least two years [37] and a second required infertility for only one year [38].
In Table 4 articles are rated for their reporting of patient characteristics in a descriptive table. Articles were rated as 'A' if they reported at least one item in each of the categories of medical history, demographics, and clinical assessment. Even with these minimal criteria a minority of studies were categorized as 'A'; the only treatment areas that had primarily 'A's were local anesthesia for carotid endarterectomy and colporrhaphy or needle suspension for incontinence. For one treatment area, early versus conventional discharge, none of the OSs provided information on maternal comorbidities or other relevant aspects of medical history.
Table 4 Reporting of Patient Characteristics and Efforts to Assess and Control Confounding
Quinidine (Afib) (n = 6) Trial of Labor (Breech) (n = 6) Colposuspension 1 (Incontinence) (n = 6) Colposuspension 2 (Incontinence) (n = 4) Early Discharge (Childbirth) (n = 4) Hip Screw (Hip Fx) (n = 5) Local Anesthesia (CEA) (n = 11) HSG (Infertility) (n = 6)
Patient Characterization*
A = 4 A = 3 A = 5 A = 3 A = 0 A = 3 A = 7 A = 1
B = 2 B = 3 B = 1 B = 1 B = 3 B = 1 B = 4 B = 3
C = 0 C = 0 C = 0 C = 0 C = 1 C = 1 C = 0 C = 0
D = 0 D = 0 D = 0 D = 0 D = 0 D = 0 D = 0 D = 2
Treatment Selection↖
A = 0 A = 0 A = 0 A = 0 A = 1 A = 1 A = 5 A = 3
B = 0 B = 0 B = 0 B = 0 B = 1 B = 0 B = 3 B = 0
C = 0 C = 6 C = 5 C = 4 C = 2 C = 0 C = 1 C = 1
D = 6 D = 0 D = 1 D = 0 D = 0 D = 4 D = 2 D = 2
Comparison of risk factors*
A = 0 A = 2 A = 5 A = 3 A = 0 A = 1 A = 5 A = 1
B = 3 B = 3 B = 1 B = 0 B = 3 B = 2 B = 6 B = 2
C = 0 C = 1 C = 0 C = 0 C = 1 C = 0 C = 0 C = 0
D = 3 D = 0 D = 0 D = 1 D = 0 D = 2 D = 0 D = 3
Statistical Adjustment‡
A A = 0 A = 0 A = 0 A = 0 A = 1 A = 0 A = 1 A = 0
B B = 1 B = 3 B = 1 B = 2 B = 0 B = 2 B = 0 B = 0
C C = 5 C = 3 C = 5 C = 2 C = 3 C = 3 C = 10 C = 6
Afib = Atrial fibrillation
Hip Fx = Hip Fracture
CEA = Carotid Endarterectomy
HSG = Hysterosalpingography
* Characterizations or comparisons were given an 'A' if they included least one element in each of the following categories: demographics, medical history, and clinical assessment. They were given a 'B' if they included one medical history or clinical assessment variable, a 'C' if they included one demographic variable, and a 'D' if there were no characterization or comparisons.
↖ Treatment selection methods were given an 'A' if they probably reduced confounding, a 'B' if the effect on confounding was uncertain, a 'C' if confounding was probably increased, and a 'D' if they were not described.
† Statistical adjustment was given an 'A' for using multiple regression, 'B' for using stratification, and 'C' for no adjustment.
Factors that influence confounding
Table 4 also describes how study characteristics were reported that could influence confounding. Confounding was more likely in two studies because subjects on one treatment were treated several years previously compared with subjects on another. Confounding was also more likely in other studies (the majority of trials of labor and surgery for incontinence and half the studies of early discharge [32,33]) because treatment was allocated on the basis of patient characteristics likely to influence the possibility of treatment failure. Confounding may have been less likely if the physicians treating the patients used only one procedure. This occurred in a few studies of local anesthesia for carotid endarterectomy, hip screws for hip fracture, and contrast media for HSG. Confounding was considered less likely in another study because of an abrupt change in patient care [35]. In several studies it was not possible to assess how patient preferences may have influenced confounding [39-42].
Table 4 shows whether studies assessed the possibility of confounding by comparing patients on the two treatments with respect to at least one variable from the categories of medical history, demographics, and clinical assessment. Most studies did not make these comparisons; the few that did should have evaluated additional potential confounders. In addition, once potential confounders were identified, the studies made only minimal use of statistical methods to control for confounding. A few studies attempted to control for confounding by stratifying on the basis of some risk factors, but only one study performed a regression analysis that adjusted for multiple risk factors [43].
Reasons for OS heterogeneity
We found evidence that variation in outcome definition and length of follow-up caused heterogeneity in results. For example, in studies of trial of labor, the study with the lowest failure rate [44] was also the study that defined poor outcome in the newborn as a five-minute Apgar score less than five, instead of less than seven as used in other studies (the lower the score the more likely the newborn is to require resuscitation). For studies of early discharge, the lowest failure rate came from a study that examined post-operative complications of C-section patients, and the highest rates came from a study that included many common symptoms in the definition of maternal morbidity (e.g., cold, flu, and constipation). In the studies comparing colposuspension to either needle suspension or colporrhaphy, the lowest failure rates in the colposuspension groups came from studies in which follow-up was less than one-year, and those low failure rates were very similar to the RCTs, both of which had follow-up of one year. For the study of HSG with the lowest odds ratio (0.98) [37], the duration of follow-up was two years, as compared to the other studies which had follow-up of one year or less. It is possible that infertility problems that improved with oil-contrast media may have resolved in any case over a two-year period.
Reasons for differences between observational and randomized studies
Differences for the studies of TENS for postoperative pain and early discharge following childbirth may have been due in part to dissimilar definitions of failure. The OSs of TENS defined this as whether or not a patient received post-operative medications. The randomized controlled trials used verbal ratings of pain that were dichotomized into "satisfactory" or "unsatisfactory". For studies of early discharge, the randomized controlled trial defined failure as maternal problems requiring physician referral. These problems were primarily infections: urinary tract infections, episiotomy infection, mastitis, subinvolution, and endometritis. Most OSs defined failure as maternal problems determined from physical assessment or self-report. Since these problems included constipation, flu-like symptoms, and lethargy as well as infections, failure rates were generally higher for observational than for randomized studies. The one exception was an observational study that examined outcomes post-caesarian section and defined failure as fever, wound infection etc [35]. The failure rates for this study were 6% for early discharge and 7% for the conventional group, which are similar to the rates from randomized studies. Without this study of C-section patients, the overall failure rates for the observational and randomized studies would have differed even more.
The primary concern about OSs is confounding. There was evidence of obvious confounding that was not taken into account in three treatment comparisons: 1) influence of anticoagulants on survival of myocardial infarction (historical controls treated several years earlier [45-47] and anticoagulants preferentially given to younger patients and patients at lower risk for other reasons [48,49]) 2) quinidine for the treatment of arrhythmias (significantly higher [50,51] rates of valvular heart disease in the quinidine group), and 3) colposuspension versus anterior colporrhaphy (substantially and significantly higher rates in the colposuspension group of severe pre-surgery incontinence [52]). [32,34] In no study showing obvious confounding did the authors assess or adjust for confounding, or even raise it as a concern.
Discussion
Previous studies have compared results of OSs and RCTs. The present investigation was the first to evaluate what design features could have influenced results of OSs and, therefore, the comparisons of results from OSs and RCTs. We found evidence that some factors unrelated to validity (treatment specifics, patient characteristics, and methods of measuring outcomes) could have influenced results in some of the studies. However, the comparisons of RCTs and OSs (and in many cases the original meta-analyses that combined the studies) did not take these study features into account. Therefore, it is possible that some differences between some RCTs and OSs may be due to factors other than lack of validity of OSs.
Clearly, however, a critical validity issue (confounding) influenced the results of some OSs. Patients on some treatments differed substantially from patients on another with respect to risk factors or ancillary treatments that probably influenced outcomes and altered the observed relative effectiveness of the two treatments. Unfortunately, few studies assessed the possibility of confounding, and almost none made a sophisticated effort to control for it. Because of the potential for confounding to invalidate the results of OSs, the lack of concern with confounding was surprising and disturbing.
The primary finding of this investigation was that few OSs of medical treatments provided sufficient information for their results to be adequately interpreted. The poor reporting impaired the ability of the systematic reviews and meta-analyses that included these articles to explain differences in results or decide how results should be combined. It may also have contributed to our inability to account for most of the variation in results among OSs and the causes of discrepancies between OSs and RCTs. Differences may have occurred because the OSs and RCTs evaluated different treatments, defined outcomes differently, or had obvious confounding. The OSs reviewed did not provide sufficient evidence to assess whether they were invalid because of undetectable and unavoidable confounding. This type of confounding is of greatest concern in OSs and may have been responsible for differences between OSs and RCTs of hormone replacement therapy. Undetectable confounding may be less likely when patients have little influence on treatment choice, such as decisions about a specific surgical procedure.
In summary, our study provided little evidence either for or against the validity of OSs. However, it suggested that causes of differences previously found between OSs and RCTs are difficult to determine. The OSs we examined may not be representative of all OSs that evaluated medical treatments. However, the severe reporting problems in the 61 studies reviewed suggest that many other published studies provide inadequate information. Without improved standards for reporting, it will be difficult to assess how OSs on a given topic should be interpreted or, more generally, the appropriate role for OSs in the evaluation of medical treatments. Standards can be improved by developing criteria for studies and involving more researchers with a strong epidemiological background in the design, reporting, and review of OSs of medical treatments.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
AH was responsible for much of the study design and writing. SB, MC, YB, and KB reviewed the articles, helped develop the format for abstraction, and examined relationships between study characteristics and results. DL and MS helped with the conceptualization and writing of the article.
Acknowledgements
Supported in part by grant 1 RO1 HS10739-01 from the DHS-Agency for Healthcare Research and Quality. The authors appreciate the assistance of Laurie Wallace with manuscript preparation.
==== Refs
D'Agostino R Kwan H Measuring effectiveness. What to expect without a randomized control group Med Care 1995 33 AS95 105 7723466
Pocock S Randomized Trials or Observational Tribulations? N Eng J Med 2000 342 1907 1908 10.1056/NEJM200006223422511
Writing Group for the Women's Health Initiative I. Risks and benefits of estrogen plus progestin in healthy postmenopausal women: principal results From the Women's Health Initiative randomized controlled trial JAMA-J Am Med Assoc 2002 288 321 333 10.1001/jama.288.3.321
Grodstein F Manson JE Colditz GA Willett WC Speizer FE Stampfer MJ A prospective, observational study of postmenopausal hormone therapy and primary prevention of cardiovascular disease Ann Intern Med 2000 133 933 941 11119394
Heckbert SR Weiss NS Koepsell TD Lemaitre RN Smith NL Siscovick DS Lin D Psaty BM Duration of estrogen replacement therapy in relation to the risk of incident myocardial infarction in postmenopausal women Arch Intern Med 1997 157 1330 1336 9201007 10.1001/archinte.157.12.1330
Bush TL Barrett-Connor E Cowan LD Criqui MH Wallace RB Suchindran CM Tyroler HA Rifkind BM Cardiovascular mortality and noncontraceptive use of estrogen in women: results from the Lipid Research Clinics Program Follow-up Study Circulation 1987 75 1102 1109 3568321
Wilson PW Garrison RJ Castelli WP Postmenopausal estrogen use, cigarette smoking, and cardiovascular morbidity in women over 50. The Framingham Study N Engl J Med 1985 313 1038 1043 2995808
Sidney S Petitti DB Quesenberry CP Jr Myocardial infarction and the use of estrogen and estrogen-progestogen in postmenopausal women.[see comment] Ann Intern Med 1997 127 501 508 9313017
Humphrey L Chan BK Sox HC Postmenopausal hormone replacement therapy and the primary prevention of cardiovascular disease Ann Intern Med 2002 137 273 284 12186518
Britton A McKee M Black N McPherson K Sanderson C Bain C Choosing between randomised and non-randomised studies: a systematic review Health Technol Assess 1998 2
Benson K Hartz A Observational studies versus randomized controlled trials N Engl J Med 2000 342 1878 1886 10861324 10.1056/NEJM200006223422506
Concato J Shah N Horwitz RI Randomized, controlled trials, observational studies, and the hierarchy of research designs N Eng J Med 2000 342 1887 1892 10.1056/NEJM200006223422507
Ioannidis JP Haidich AB Pappa M Pantazis N Kokori SI Tektonidou MG Contopoulos-Ioannidis DG Lau J Comparison of evidence of treatment effects in randomized and nonrandomized studies JAMA-J Am Med Assoc 2001 286 821 830 10.1001/jama.286.7.821
Kunz R Oxman A The unpredictability paradox: review of empirical comparisons of randomised and non-randomised clinical trials Br Med J 1998 317 1185 1190 9794851
Begg C Cho M Eastwood S Horton R Moher D Olkin I Pitkin R Rennie D Schulz KF Simel D Stroup DF Improving the quality of reporting of randomized controlled trials. The CONSORT statement [see comments] JAMA-J Am Med Assoc 1996 276 637 639 10.1001/jama.276.8.637
Moher D Schulz KF Altman DG The CONSORT statement: Revised recommendations for improving the quality of reports of parallel-group randomized trials Ann Intern Med 2001 134 657 662 11304106
Moher D Jadad A Nichol G Penman M Tugwell P Walsh S Assessing the quality of randomized controlled trials: an annotated bibliography of scales and checklists Control Clin Trials 1995 16 62 73 7743790 10.1016/0197-2456(94)00031-W
Downs S Black N The feasibility of creating a checklist for the assessment of the methodological quality both of randomised and non-randomised studies of health care interventions J Epidemiol Community Health 1998 52 377 384 9764259
Horwitz RI Complexity and contradiction in clinical trial research Am J Med 1987 82 498 510 3548349 10.1016/0002-9343(87)90450-5
West S King V Carey TS Lohr KN McKoy N Sutton SF Lux L Systems to rate the strength of scientific evidence. Evidence report/technology assessment no. 47. Prepared by the Research Triangle Institute – University of North Carolina Evidence-based Practice Center under Contract No. 290-97-0011 AHRQ Publication No 02-E016 2002 Rockville, MD: Agency for Healthcare Research and Quality
Mantel N Haenszel WH Statistical aspects of the analysis of data from retrospective studies of disease J Natl Cancer Inst 1959 22 719 748 13655060
DerSimonian R Laird N Meta-analysis in clinical trials Control Clin Trials 1986 7 177 188 3802833 10.1016/0197-2456(86)90046-2
Wright I Marple CD Beck DF Report of the committee for the evaluation of anticoagulants in the treatment of coronary thrombosis with myocardial infarction: a progress report on the statistical analysis of the first 800 cases studied by this committee Am Heart J 1948 36 810 815 10.1016/0002-8703(48)90278-6
Feldman L O'Connor WR Friedman IA Anticoagulant therapy in acute myocardial infarction Am Heart J 1952 44 112 118 14933314 10.1016/0002-8703(52)90178-6
Rashkoff I Schaefer LE Magida MGM Evaluation of dicumarol therapy in 287 cases of acute myocardial infarction J Mt Sinai Hosp 1952 18 350 357
Mundy A A trial comparing the Stamey bladder neck suspension procedure with colposuspension for the treatment of stress incontinence Br J Urol 1983 55 687 690 6360297
Bannister G Gibson AG Ackroyd CE Newman JH The fixation and prognosis of trochanteric fractures. A randomized prospective controlled trial Clin Orthop 1990 254 242 246 2182255
Reimold S Chalmers TC Berlin JA Antman EM Assessment of the efficacy and safety of antiarrhythmic therapy for chronic atrial fibrilliation: observations on the role of trial design and implications of drug-related mortality Am Heart J 1992 124 924 932 1388328 10.1016/0002-8703(92)90974-Z
Clawson D Trochanteric fractures treated by the sliding screw plate fixation method J Trauma 1964 4 737 752 14225321
Songane FF Thobani S Malik H Bingham P Lilford RJ Balancing the risks of planned cesarean section and trial of vaginal delivery for the mature, selected, singleton breech presentation J Perinat Med 1987 15 531 543 3452635
Weil A Reyes H Bischoff PP Rottenberg RD Krauer F Modifications of the urethral rest and stress profiles after different types of surgery for urinary stress incontinence Br J Obstet Gynaecol 1984 91 46 55 6537885
Thunedborg P Rasmussen WF Jensen SB Stress urinary incontinence and posterior bladder suspension defects. Results of vaginal repair versus Burch colposuspension Acta Obstet Gynecol Scand 1990 69 55 59 2346081
Fowler J Experience with suprapubic vesicourethral suspension and endoscopic suspension of the vesical neck for stress urinary incontinence in females Surg Gynecol Obstet 1986 163 437 441 3775620
Van Geelen JM Theeuwes AG Eskes TK Martin CB The clinical and urodynamic effects of anterior vaginal repair and Burch colposuspension Am J Obstet Gynecol 1988 159 137 144 3394733
Strong TJ Brown WL JrBrown WL Curry CM Experience with early postcesarean hospital dismissal Am J Obstet Gynecol 1993 169 116 119 8333434
Lemmer C Early discharge: outcomes of primiparas and their infants J Obstet Gynecol Neonatal Nurs 1987 16 230 236 3650321
Barwin N Hysterosalpingography in infertility Ulster Med J 1971 41 61 65 5150066
Mackey R Glass RH Olson LE Vaidya R Pregnancy following hysterosalpingography with oil and water soluble dye Fertil Steril 1971 22 504 507 4329273
Shah D Darling RC Chang BB Bock DE Paty PS Carotid endarterectomy in awake patients: its safety, acceptability, and outcome J Vasc Surg 1994 19 1015 1020 8201702
Becquemin J Paris E Valverde A Pluskwa F Melliere D Is regional anesthesia always appropriate? J Cardiovasc Surg 1991 32 592 598 1939322
Corson J Chang BB Shah DM Leather RP DeLeo BM Karmody AM The influence of anesthetic choice on carotid endarterectomy outcome Arch Surg 1987 122 807 812 3592971
Norr K Nacion KW Abramson R Early discharge with home follow-up: impact on low-income mothers and infants J Obstet Gynecol Neonatal Nurs 1989 18 133 141 2540302
Allen B Anderson CB Rubin BG Thompson RW Flye MW Young-Beyer P Frisella P Sicard GA The influence of anesthetic technique on perioperative complications after carotid endarterectomy J Vasc Surg 1994 19 834 843 8170037
Bingham P Hird V Lilford RJ Management of the mature selected breech presentation: an analysis based on the intended method of delivery Br J Obstet Gynaecol 1987 94 746 752 3311132
Loudon I Pease JC Cooke AM Anticoagulants in myocardial infarction Br Med J 1953 1 911 913 13032552
Eastman G Cook ET Shinn ET A clinical study of anticoagulants in acute myocardial infarction with particular reference to early heparin therapy Am J Med Sci 1957 233 647 653 13424531
Honey G Truelove SC Prognostic factors in myocardial infarction Lancet 1957 1 1155 1161 1209-1212 13439991 10.1016/S0140-6736(57)91735-X
Modan B Shani M Schor S Modan M Reduction of hospital mortality from acute myocardial infarction by anticoagulant therapy N Engl J Med 1975 292 1359 1362 1138162
Tonascia J Gordis L Schmerler H Retrospective evidence favoring use of anticoagulants for myocardial infarctions N Engl J Med 1975 292 1362 1366 1138163
Radford M Evans DW Long-term results of DC reversion of atrial fibrillation Br Heart J 1968 30 91 96 5637563
Szekely P Sideris DA Batson GA Maintenance of sinus rhythm after atrial defibrillation Br Heart J 1970 32 741 746 5212345
Park G Miller EJ Surgical treatment of stress urinary incontinence: a comparison of the Kelly plication. Marshall-Marchetti-Krantz, and Pereyra procedures Obstet Gynecol 1988 71 575 579 3353049
Coelho E Serra Pinto L Sales Luiz A Macieira Coelho E Leite Pereira A Barreiros R Long-term results of conversion of atrial fibrillation by direct current countershock Cardiologia 1967 50 147 155 6077090
De Souza A Stress incontinence of urine Int Surg 1976 8 396 402
Ralph G Lichtenegger W Tamussino K Urodynamic results after incontinence operations. Uroynamische ergebnisse nack inkontinenzoperationen Zentralbl Gynakol 1989 111 235 242 2728666
Ralph G Effect of various operations for incontinence on the dynamics and topography of the bladder and bladder neck Wien Klin Wochenschr 1991 185 3 14
Halaska M Voigt R Michels W Martan A Voigt P Abdominal versus vaginal stress incontinence surgery. Results of the university women's clinic in jena abdominale versus vaginale stressinkontinenzoperation der frau – ergebnisse an der universitats-frauenklinik jena Arch Gynecol Obstet 1993 254 491 493
Casado J Virseda Chamorro M Ramirez Fernandez JC Rapariz Gonzalez M Esteban Fuertes M Resel Estevez L Comparative results of different types of urethropexy in the treatment of stress urinary incontinence in women. Resultados comparativos de los diferentes tipos de uretropexia en el tratamiento de la incontinencia urinaria de esfuerzo fmenina Arch Esp Urol 1994 8 761 768
Carey M Dwyer PL Position and mobility of the urethrovesical junction in continent and in stress incontinent women before and after successful surgery Aust NZ J Obstet Gynaecol 1991 31 279 284
Penttinen J Kaar K Kauppila A Colposuspension and transvaginal bladder neck suspension in the treatment of stress incontinence Gynecol Obstet Invest 1989 29 101 105 2792913
Jopling M deSanctis CA McDowell DE Graf DF Mertinez OA Savrin RA Anesthesia for carotid endarterectomy: a comparison of regional and general techniques (abstract) Anesthesiology 1983 59 A217
European Carotid Surgery Trialists Collaborative Group The morbidity and mortality of carotid endarterectomy in the European Carotid Surgery Trial Unpublished data
Bartoloni A Savron F Rigo V Pasetto A Spezie C Perbellini A Effectiveness of regional anesthesia for carotid endarterectomy. Retrospective review of 147 endarterectomies (author's translation) Minerva Anestesiol 1991 57 75 82 1870730
Chalmers T Matta RJ Smith H JrKunzler AM Evidence favoring the use of anticoagulants in the hospital phase of acute myocardial infarction N Engl J Med 1977 297 1091 1096 909566
Gifford D Morton SC Fiske M Kahn K A meta-analysis of infant outcomes after breech delivery Obstet Gynecol 1995 85 1047 1054 7770253 10.1016/0029-7844(95)00053-T
Black N Downs SH The effectiveness of surgery for stress incontinence in women: a systematic review Br J Urol 1996 78 497 510 8944504
Carroll D Tramer M McQuay H Nye B Moore A Randomization is important in studies with pain outcomes: systematic review of transcutaneous electrical nerve stimulation in acute postoperative pain Br J Anaesth 1996 77 798 803 9014639
Grullon K Grimes DA The safety of early postpartum discharge: a review and critique Obstet Gynecol 1997 90 860 865 9351780 10.1016/S0029-7844(97)00405-5
Chinoy M Parker MJ Fixed nail plates versus sliding hip systems for the treatment of trochanteric femoral fractures: a meta analysis of 14 studies Injury 1999 30 157 163 10476259 10.1016/S0020-1383(99)00074-1
Tangkanakul C Counsell C Warlow C Local versus general anaesthesia for carotid endarterectomy. Cochrane Stroke Group Cochrane Database of Systematic Reviews 2002
Vandekerckhove P Watson A Lilford R Harada T Hughes E Oil-soluble versus water-soluble media for assessing tubal patency with hysterosalpingography or laparoscopy in subfertile women Cochrane Database of Systematic Reviews 1999
|
16137327
|
PMC1215501
|
CC BY
|
2021-01-04 16:38:27
|
no
|
Emerg Themes Epidemiol. 2005 Sep 1; 2:8
|
utf-8
|
Emerg Themes Epidemiol
| 2,005 |
10.1186/1742-7622-2-8
|
oa_comm
|
==== Front
Genet Vaccines TherGenetic Vaccines and Therapy1479-0556BioMed Central London 1479-0556-3-61611531610.1186/1479-0556-3-6ResearchProduction and characterization of amplified tumor-derived cRNA libraries to be used as vaccines against metastatic melanomas Carralot Jean-Philippe [email protected] Benjamin [email protected] Oliver [email protected] Jochen [email protected] Birgit [email protected] Regina [email protected] Ingmar [email protected] Claus [email protected] Hans-Georg [email protected] Steve [email protected] CureVac, Paul Ehrlich Strasse 15, 72076 Tübingen, Germany2 University of Tübingen, Institute for Cell Biology, Department of Immunology; Auf der Morgenstelle 15; 72076 Tübingen, Germany3 Section for Dermatological Oncology, Tübingen University Hospital, Liebermeisterstraße 25, 72076 Tübingen, Germany2005 22 8 2005 3 6 6 22 6 2005 22 8 2005 Copyright © 2005 Carralot et al; licensee BioMed Central Ltd.2005Carralot et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Anti-tumor vaccines targeting the entire tumor antigen repertoire represent an attractive immunotherapeutic approach. In the context of a phase I/II clinical trial, we vaccinated metastatic melanoma patients with autologous amplified tumor mRNA. In order to provide the large quantities of mRNA needed for each patient, the Stratagene Creator™ SMART™ cDNA library construction method was modified and applied to produce libraries derived from the tumors of 15 patients. The quality of those mRNA library vaccines was evaluated through sequencing and microarray analysis.
Results
Random analysis of bacterial clones of the library showed a rate of 95% of recombinant plasmids among which a minimum of 51% of the clones contained a full-Open Reading Frame. In addition, despite a biased amplification toward small abundant transcripts compared to large rare fragments, we could document a relatively conserved gene expression profile between the total RNA of the tumor of origin and the corresponding in vitro transcribed complementary RNA (cRNA). Finally, listing the 30 most abundant transcripts of patient MEL02's library, a large number of tumor associated antigens (TAAs) either patient specific or shared by several melanomas were found.
Conclusion
Our results show that unlimited amounts of cRNA representing tumor's transcriptome could be obtained and that this cRNA was a reliable source of a large variety of tumor antigens.
==== Body
Background
The identification by van der Bruggen et al. [1] of the first tumor associated (TAA) antigen recognized by specific cytotoxic T lymphocytes (CTLs) in melanoma patients boosted the development of anti-cancer immunotherapy strategies. During the last years, vaccination protocols targeting differentiation antigens (MART-1/Melan-A [2,3], gp100 [4], Tyrosinase [5,6]) or cancer-testis antigens (MAGE [1,7], NY-ESO1 [8]) were tested and showed encouraging results [9-11].
However, a growing body of evidence suggests that, instead of using defined antigens, targeting the whole spectrum of tumor antigens would represent an alternative, potentially more efficacious method [12-14]. Indeed, the use of total tumor material for vaccination allows the development of B and T cells directed against a large variety of known but also unknown TAAs [15]. In addition, stimulating such a large spectrum of specific effectors directed against multiple epitopes restricted by diverse HLA class I and II types would reduce the risk of tumor escape through antigen loss or MHC downregulation [16-19]. Finally, another advantage of the whole tumor approach is that, in an autologous setting, patient's TAAs eventually stemming from tumor-specific somatic mutations could be targeted [20,21].
In order to vaccinate patients with the whole spectrum of TAAs, several methods were developed. In 1998, Soiffer et al. [22] disclosed the results obtained by vaccinating patients with autologous irradiated tumor cells engineered to produce GM-CSF. The same year, Nestle et al. [23] showed partial or complete tumor remissions in six melanoma patients vaccinated with dendritic cells (DC) loaded with autologous tumor lysate. Alternatively, Boczkowski et al. [24] reported that mouse DCs pulsed in vitro with tumor RNA could trigger an anti-tumor immunity in vivo. Several groups further developed and optimized those different strategies [25-27] but faced the limitation imposed by the requirement of large amounts of tumor tissue for lysate preparation or for sufficient RNA yields extraction. In order to overcome this drawback, Boczkowski et al. [28] modified the SMART method (BD Biosciences Clontech, Palo Alto, CA) in order to in vitro transcribe tumor cDNA and performed therefore a one-step amplification of tumor mRNA. Transfected into antigen presenting cells (APCs), this amplified cRNA was shown in vitro to induce anti tumor immunity [29,30]. As an alternative vaccination method, Hoerr et al. [31] demonstrated the capacity of mRNA coding for defined antigens or of total cRNA to trigger an antigen-specific immune response after direct intra-dermal injections of the ribonucleic acid. Similarly, Granstein et al. [15] showed protection against S1509 tumor cells in mice that received three intradermal injections of total RNA extracted from S1509 cells. Although still marginally studied compared to mRNA-loaded DC vaccines, the direct injection of mRNA represents a technology that offers the important advantage to circumvent the time and money consuming steps of generation of DCs.
In 2003, we initiated the first phase I/II clinical study to test the feasibility, safety, and efficacy of a vaccine composed of autologous amplified tumor mRNA in stage III/IV patients with metastatic melanoma (The detailed evaluation of the toxicity, clinical and immunological efficacy of this treatment will be reported in a following manuscript). Fifteen patients received from 3 to 16 intradermal injections of 200 μg of amplified autologous tumor cRNA. The amount of injected RNA was limited by the maximal intradermal injection volume (100 μl) and set according to the preclinical results which indicated that a concentration of ca. 0.8 μg/μl was leading to a good gene transfer. The injection's schedule consisted in applications every two weeks of four injections and then once every month. It was decided empirically since no previous data on the toxicity and efficacy of this immunization method are available in humans. In mice one injection was shown to be sufficient to trigger an immune response [31]. However, in cancer patients, a sustained stimulation of the immunity is probably requested in order to get an efficient anti-tumor immune response. According to this protocol, the required amount of cRNA for a complete therapy was between 0.6 and 3.2 mg per patient (table 1). In order to get unlimited amounts of product, a new method for the amplification of the tumor mRNA was developed. Briefly, a cDNA library was generated from tumor RNA using the SMART (Switch Mechanism At the 5'end of RNA Templates) system (BD Clontech) and then was cloned in our RNActive™ vector (CureVac GmbH), amplified in Escherichia coli, and finally transcribed in vitro. As opposed to the protocol described by Boczkowski et al. [28] in which the PCR-amplified tumor cDNA library was directly used as template for the in vitro transcription (resulting in limited cRNA amounts), the method applied in our laboratory provided us with unlimited amounts of the tumor-derived cRNA.
Table 1 Summary of mRNA libraries and clone analysis. In the case of MEL14, total RNA was extracted from ~5 × 104 pleural tumor cells (NA: Not applicable)
Patients Weight of tumor sample (mg) Quantity of extracted total RNA (μg) Number of clones (cfu) Size range of analyzed clones (nt) Quantity of mRNA library prepared (mg) Number of injection performed
1 MEL01 32.4 9.4 1 × 105 500 – 4000 2.8 10
2 MEL02 33 10.5 1 × 105 200 – 8000 5.0 12
3 MEL03 33.6 26.3 5 × 105 250 – 1000 1.9 6
4 MEL04 38 36.7 2 × 105 400 – 3500 3.6 8
5 MEL05 85 14.2 2 × 105 500 – 1000 5.0 13
6 MEL06 60 115.2 3 × 105 300 – 1200 4.0 16
7 MEL07 76 70.5 5 × 105 500 – 1200 2.8 7
8 MEL08 34.1 60.9 3 × 105 500 – 1200 5.3 16
9 MEL09 95 84 4 × 104 350 – 800 4.5 10
10 MEL10 78.7 62.5 2 × 105 600 – 1200 4.2 16
11 MEL11 77.3 9.56 3 × 105 400 – 1000 2.7 3
12 MEL12 34.3 15.4 6 × 104 400 – 1200 3.9 10
13 MEL13 72 9.14 2 × 105 750 – 2000 4.4 16
14 MEL14 NA 13.2 1 × 105 400 – 10000 1.8 4
15 MEL15 60 41.5 3 × 105 500 – 4000 4.1 8
Average 57.8 38.6 2 × 105 450 – 3250 3.7 10
Whereas the SMART method was reported to maintain the relative levels of RNAs contained in the original transcriptome regardless of their size or their baseline expression [32], the cloning step in E. coli was on the contrary described to introduce a bias favoring short fragments [33]. We thus analyzed the quality of the produced amplified-mRNA libraries to be used as a vaccine in melanoma patients. Several clones randomly picked-up within the produced libraries were analyzed by PCR and sequenced. In addition, the gene expression profiles of two metastases were compared to their corresponding cRNA-libraries.
Results and discussion
Tumor-derived mRNA library quality
Total RNA was extracted form 15 fresh melanoma tissues. It was then used to generate cDNA libraries according to the SMART protocol (BD Clontech, Palo Alto, CA). The obtained cDNA libraries were ligated into the RNActive™ vector (CureVac GmbH, Tübingen, Germany) containing the mRNA stabilizing sequences of 5' and 3' UnTranslated Regions (UTR) from β- and α-globin respectively [34,35]. Moreover, this vector introduced a 70 A poly(A) tail further enhancing mRNA translation potential (data not shown). RNActive™ libraries were transformed into ultracompetent E. coli. The total primary transformant numbers were ranging from 6 × 104 to 5 × 105 clones with an average number of 2 × 105 (table 1). In order to determine the ligation efficiency, 24 clones per library were randomly picked up and submitted to 35 PCR cycles using primers flanking the cDNA insertion sites. Ninety-eight percent of the 312 analyzed clones had an insert with sizes ranging from 200 bp to 10 kbp (data not shown). In order to further test the quality of the libraries, plasmid DNA of 9 clones randomly picked up were extracted for 5'end sequencing (Figure 1). Out of the 112 readable sequences, 2 clones had no insert confirming the 2% rate of self-ligation found by PCR analysis. Among the remaining 110 clones, 3 (3%) showed sequences classified as aberrant with insert sizes inferior to 50 bp probably corresponding to recombination events. The other 107 sequences were analyzed BLAST [36] (Basic Local Alignment Search Tool) using the nr database. About half of the clones (51%) corresponded to full Open Reading Frames (ORFs) of annotated sequences. The other sequences were homologous to ESTs (Expressed Sequence Tags) coding for potential proteins with unknown functions. Full-length clone sizes ranged from 344 to 5925 bp with an average size of 1395 bp correlating with the average insert size of 1.4 kb in cDNA libraries described by Draper et al. [37]. Interestingly, 28 % of the clones which aligned to annotated ORFs (14% of all sequenced clones) were tumor related genes (for instance S100 Calcium binding protein A4-metastasin [38]), or genes reported to be overexpressed in tumors (for instance Laminin receptor [39]). This observation fitted with the objective of using the cRNA libraries as anti-tumor vaccines.
Figure 1 BLAST analysis of sequenced clones. Nine clones per library were randomly picked up and their plasmid DNA was sequenced using a T7 promoter primer. Readable sequences (n = 112) were submitted to a BLAST analysis and their relative distribution plotted.
Relative representation of transcripts
In order to determine whether a bias was introduced by the amplification protocol, the relative gene expression in extracted total RNA from two metastases was compared to the relative gene expression in the corresponding amplified cRNA libraries. Biotin-labeled complementary RNA of tumor total RNA and of amplified cRNA libraries from patients MEL02 and MEL10 were generated using the Affymetrix eukaryotic sample and array processing standard procedure and hybridized on HG-U133A microarrays. According to the Microarray Analysis Suite 5.0 software (MAS 5.0; Affymetrix), 34% and 36% of the genes that were reported as "present" in the tumor total RNA were also detected as "present" in the amplified libraries of patients MEL02 and MEL10 respectively (Theses transcripts are qualified as "recovered" in the following). The other transcripts reported by the Microarray Analysis Suite 5.0 software as "present" in tumor's mRNA but as "absent" in the corresponding cRNA library were termed as "lost". In order to determine the factors influencing the biased amplification of genes, the size distributions of "lost" and "recovered" transcripts were compared and plotted in figure 2A. For both patients MEL02 and MEL10, the average size distribution of "recovered" genes (2218 and 1965 nt respectively) was significantly lower (t test, P < 0.0001) than the average size of transcripts lost during the amplification process (2894 and 3222 nt respectively). This suggests a biased amplification disfavoring large fragments as observed by Wellenreuther et al. [33]. In addition, the fluorescence values in the original tumor of genes reported as "recovered" and "lost" in the cRNA were compared as shown in figure 2B. Average signals of 2711 and 3709 for MEL02 and MEL10 respectively were observed for the genes present in the cRNA library and thus preserved during the process. In contrast, the genes "lost" during amplification had an average signal of only of 708 and 1174 for patients MEL02 and MEL10 respectively. These values were significantly lower (t test, P < 0.0001) than those found for the group of "recovered" genes. Thus, transcripts of higher abundance in the original tumor were preferentially preserved during the amplification process whereas mRNAs of lower abundance were eventually lost.
Figure 2 Comparison of sizes and fluorescence signal intensities in the original tumor sample for "lost" and "recovered" transcripts. A. The sizes of transcripts Present in tumor's total RNA and reported as Present (PP) or Absent (PA) in the cRNA library of MEL02 and MEL10 patients were plotted. The average size of "recovered" genes was significantly lower (t test, P < 0.0001) than the average size of "lost" genes. B. The fluorescence signals of genes "Present" in the original tumor were compared for the transcripts reported as Present (PP) or Absent (PA) in the corresponding mRNA libraries. The group of "recovered" genes showed a significantly higher (t test, P < 0.0001) mean signal than the group of genes "lost" during the library production.
The fluorescence signal intensities of the genes reported as "present" in both the tumor and the corresponding cRNA libraries were plotted in figure 3. The correlation factors to a linear regression were 0.55 and 0.42 for patients MEL02 and MEL10 libraries respectively, confirming that the mRNA amplification was quite heterogeneous. However, in the group of "recovered" transcripts no significant correlation was evidenced between the transcript size or their signal intensity in the tumor of origin and their amplification factor during the cloning (data not shown).
Figure 3 Correlation of signal intensities in tumor and corresponding mRNA libraries for patients MEL02 and MEL10. Fluorescence signals in original metastases and amplified tumor cRNA libraries for patients MEL02 and MEL10 were compared for all genes reported as "Present" in the library by MAS 5.0 software.
Patient-specific gene expression
In order to evaluate the relevance of the autologous approach, signal intensities for the genes present in metastases of MEL02 and MEL10 were compared in figure 4. As expected, the two melanoma samples were quite similar with a correlation coefficient to a linear regression of 0.75 for the 6693 genes shared between the two tumors. However, the two tissues showed specific profiles with 3222 and 483 mRNA transcripts expressed only in patient's MEL02 and MEL10 metastasis respectively. These patient-specific antigens might represent particularly interesting immunological targets [13] and argue for the injection of autologous tumor cRNA rather than the use of cRNA library derived from tumor cell lines as a vaccine.
Figure 4 Correlation of fluorescence of genes present in tumors of patients MEL02 and MEL10. Fluorescence signals of patients MEL02 and MEL10 tumor transcripts reported by MAS 5.0 software as present in both samples (6993 genes, ◆), only in MEL02 metastasis (3222 genes, ●) or only in MEL10 melanoma (483 genes, ▲).
Several tumor antigens are present in the cRNA libraries
The transcripts showing the highest fluorescence signals in the injected cRNA library were listed in Table 2 for patient MEL02. Among the 30 genes displaying the highest signals likely representing most abundant transcripts, 13 have already been described as being involved in tumor genesis or observed as being overexpressed in different cancer types. The well-defined and widely used tumor antigen Melan-A was found as one of the most abundant transcripts showing the relevance of cRNA libraries as melanoma vaccines. In addition, the presence in the injected library of many other "tumor-related" antigens rarely used in vaccines targets highlights the potential of such a product to vaccinate patients against a large panel of TAAs.
Table 2 List of the thirty transcripts showing the highest fluorescence signals in MEL02 amplified cRNA library
Title Symbol Observations
1 Ribosomal protein L23a RPL23a
2 Eukaryotic translation elongation factor 1 alpha 1 EEF1A1
3 Ribosomal protein S3A RPS3A
4 RNase A family, 1 (pancreatic) RNASE1
5 Peptidylprolyl isomerase A, cyclophilin A PPIA Overexpressed in several cancers [42]
6 Ribosomal protein S23 RPS23
7 Ribosomal protein L39 RPL39
8 Melan-A MLANA Melanoma differentiation antigen [2]
9 Ribosomal protein L31 RPL31
10 Cytochrome c oxidase subunit VIc COX6C Overexpressed in carcinomas [43]
11 Ribosomal protein L7 RPL7 Overexpressed in gliomas [44]
12 Ribosomal protein L37a RPL37A
13 Ribosomal protein S29 RPS29
14 Secreted phosphoprotein 1, osteopontin SPP1 Important for tumorgenesis [45]
15 Calmodulin 2 CALM2 Overexpressed in several cancers [42]
16 Ribosomal protein S11 RPS11
17 "Ribosomal protein S4, X-linked" RPS4X
18 Nascent-polypeptide-associated complex alpha NACA Overexpressed in gliomas [44]
19 Ribosomal protein L23a RPL23A Involved in tumor proliferation [46]
20 ATP synthase, mitochondrial F0 complex, subunit g ATP5L
21 Tubulin, alpha, ubiquitous K-ALPHA-1 Overexpressed in breast cancers [47]
22 NADH dehydrogenase (ubiquinone) 1 alpha subcomplex NDUFA4
23 Ribosomal protein S27a RPS27A Overexpressed in breast cancers [48]
24 Beta-2-microglobulin B2M
25 H2A histone family, member Z H2AFZ Overexpressed in several cancers [42]
26 SRY (sex determining region Y)-box 4 SOX4 Overexpressed in lung cancers [49]
27 ATP synthase, mitochondrial F1 complex, epsilon subunit ATP5E
28 Tumor protein, translationally-controlled 1 TPT1 Involved in malignant transformation [50]
29 Cytochrome c oxidase subunit VIIa polypeptide 2 COX7A2
30 Ubiquitin B UBB Related to sustained proliferation [51]
Conclusion
In order to vaccinate metastatic melanoma patients with autologous amplified tumor-derived cRNA, fifteen libraries were produced using a modified SMART method. Despite a heterogeneous amplification of tumor genes, this method provided us with an unlimited source of tumor and patient specific TAAs. Indeed, the microarray analysis of the libraries indicated the presence of high copy numbers of well-known tumor associated antigens such as Melan-A but also of abundant tumor-related antigens scarcely targeted in immunotherapy. Although not addressed in the present work, this method might also allow the targeting of tumor-specific mutations. These features makes of the amplification of tumor mRNA the method of choice to easily obtain unlimited amounts of RNA coding for patient's specific TAAs that can be applied as anti-tumor immunotherapy.
Materials and methods
Tumor samples
Immediately after surgery, metastatic tissues from fully informed patients (Ethic committee approval Nr.: 269/2002) were chopped in ~0,1 cm3 pieces, and submerged in RNAlater solution from Ambion (Hungtingdon, UK), and stored at 4°C until histological identification as melanoma by an experienced pathologist.
Tumor total RNA extraction
Total RNA was extracted from tumors using the RNeasy mini kit from Qiagen (Hilden, Germany) following the instructions of the provider. Briefly, 15 to 30 mg samples placed in a 2 ml eppendorf tube were snap-frozen in a liquid nitrogen bath and disrupted with micropistils from Eppendorf (Hamburg, Germany). Tumor powder was resuspended in RLT buffer, homogenized through a 20-gauge needle and digested with 200 μg of proteinase K (Qiagen) at 55°C during 10 min. Samples were then clarified, loaded on RNeasy mini columns, washed and finally eluted in 50 μl of RNAse-free water. RNA was quantified by U.V spectrophotometry (O.D260/O.D280 ratio was over 1.8 in all cases) and analyzed on a 1,2% formaldehyde/agarose gel.
cDNA library generation
cDNA libraries of tumor total RNA were generated using the slightly modified Creator™ SMART™ PCR cDNA library construction kit from BD Biosciences Clontech (Heidelberg, Germany). Briefly, 1 μg of total RNA was reverse transcribed using SMART IV™ and CDS III/3' oligo-dT primers provided by the manufacturer. After termination of the reaction, 2 μl of cDNA were amplified using the Advantage 2 PCR kit (BD Biosciences Clontech). DNA polymerase was then inactivated with proteinase K and the cDNA library was digested with 200 U of Sfi I enzyme. cDNA libraries were then gel-purified on an 1% agarose gel and fragments from 300 bp to 10 kbp were extracted using E.Z.N.A.™ Gel Extraction kit from Peqlab GmbH (Erlangen, Germany). After precipitation, the cDNA library was ligated to dephosphorylated Sfi I-digested RNactive™ vector provided by CureVac GmbH (Tübingen, Germany) in three separated reactions to optimize vector/insert ratios.
Cloning
The three ligation products were used to transform XL10-Gold ultracompetent cells from Stratagene (Heidelberg, Germany). For analysis, 1 and 10 μl of transformation broth were plated on 2 LB-ampicillin agar plates and, after overnight culture at 37°C, the number of clones was counted. The inserts of 8 clones per transformation were amplified by PCR using primers flanking the insertion sites and amplicons were analyzed on a 1% agarose gel. Libraries having more than 104 clones/ml and less than 20% of non recombinant clones, were amplified in three 300 ml maxicultures in 2X LB-ampicillin medium during 20 h at 33°C in order to limit uneven amplification of clones.
DNA preparation and linearization
Maxicultures were pooled, centrifuged down at 5 000 rpm for 10 min and plasmid DNA was extracted using EndoFree Plasmid Maxi (Qiagen). After precipitation, 100 μg of cDNA library were digested with 100 U of Not I enzyme. After phenol/chloroform extraction and ammoniumacetate precipitation, linearized cDNA libraries were resuspended in RNAse-free water, quantified by U.V. spectrophotometry (O.D260/O.D280 ratio was over 1.8 in all cases) and analyzed on 1% agarose gel.
cRNA In vitro transcription
Twenty to hundred micrograms of linear cDNA library were in vitro transcribed using T7 mRNA Optikit from CureVac GmbH. After mRNA synthesis, DNA template was digested with 40 to 100 U of recombinant DNAse I purchased from Ambion. mRNA was then LiCl precipitated, phenol/chloroform purified, NaCl precipitated, and finally resuspended in PBS. cRNA was filter sterilized (0,2 μm), heat denatured at 80°C for 10 min before final sterile aliquoting. cRNA was quantified by U.V spectrophotometry (O.D260/O.D280 ratio was over 1.8 in all cases) and analyzed on 1.2% formaldehyde/agarose gel. Sterility of cRNA was checked by inoculating LB medium (in all cases, no bacterial growth was observed after 4 days at 37°C) and endotoxin content was determined using Bio-Whittaker (Verviers, Belgium) LAL assay kit (endotoxin content was always below 7 EU/ml).
Clone sequencing
For each library, 3 colonies per transformation were randomly picked-up with a pipette tip and used to inoculate 3 ml of LB-ampicillin medium. After overnight culture at 37°C, plasmid DNA was extracted using E.Z.N.A miniprep kit (Peqlab). Clone sequencing was performed using the ABI Big Dye and a T7 promoter primer. Sequences were purified on Autoseq. G-50 columns (Amersham Pharmacia Biotech, Freiburg, Germany), run on a 310 Genetic Analyzer from ABI PRISM™ (Applied Biosystems, Darmstadt, Germany) and analyzed with the Sequencing Analyzing 3.4.1 software (ABI PRISM). Finally, the BLAST algorithm [36] was used to identify matches to known genes.
Microarray analysis
Expression analysis of total tumor RNA and amplified tumor cRNA was performed on HG-U133A microarrays from Affymetrix (High Wycombe, UK) according to the manufacturer's eukaryotic sample and array processing standard procedure [40], which is based on the IVT method originally described by Van Gelder et al. [41]. Briefly, 1st-strand cDNA synthesis was performed using an oligo(dT)24 primer containing a T7 promoter sequence. After RNA template degradation and cDNA's second strand cDNA synthesis, complementary RNA (cRNA) was transcribed in vitro using biotinylated NTPs and T7 RNA polymerase. After purification using RNeasy columns (Qiagen), 18 μg of biotin-labeled cRNAs were fragmented by metal-induced hydrolysis. Hybridization, staining, and scanning of microarrays were performed by the Microarray Facility Tübingen. Scanned images were processed using the Microarray Analysis Suite 5.0 (MAS 5.0; Affymetrix) and expression differences between tumor and library samples were determined by baseline comparison algorithms provided by the software. Data were further processed using Microsoft Access™ and Excel™.
Authors' contributions
JPC carried out the total tumor RNA extraction, the production of mRNA libraries, the analysis of clones, participated in the microarray analysis of RNA samples, evaluated the microarray data and drafted the manuscript. OS carried out the microarray analysis of RNA samples with the help of the Microarray Facility Tübingen. BW and CG carried out the patient's recruitment and the tumor sample preparation. JP, BS and RT participated in the design of the study, the development of the work and helped to draft the manuscript. CG, IH, HGR and SP conceived, designed and coordinated this study. All authors read and approved the final manuscript.
Acknowledgements
JPC is supported by a "Fortüne" grant from the University of Tübingen and JP is supported by the DFG Graduiertenkollegue "Infektionsbiologie" of Tübingen.
==== Refs
Van der Bruggen P Traversari C Chomez P Lurquin C De Plaen E Van den Eynde B Knuth A Boon T A gene encoding an antigen recognized by cytolitic T lymphocytes on a human melanoma. Science 1991 254 1643 1647 1840703
Kawakami Y Eliyahu S Delgado CH Robbins PF Rivoltini L Topalian SL Miki T Rosenberg SA Cloning of the gene coding for a shared human melanoma antigen recognized by autologous T cells infiltrating into tumor. Proc Natl Acad Sci USA 1994 91 3515 3519 8170938
Coulie PG Brichard V Van Pel A Wolfel T Schneider J Traversari C Mattei S De Plaen E Lurquin C Szikora JP A new gene coding for a differentiation antigen recognized by autologous cytolytic T lymphocytes on HLA-A2 melanomas J Exp Med 1994 180 35 42 8006593 10.1084/jem.180.1.35
Kawakami Y Eliyahu S Delgado CH Robbins PF Sakaguchi K Appell E Yannelli JR Adema GJ Miki T Rosenberg SA Identification of a Human Melanoma Antigen Recognized by Tumor-Infiltrating Lymphocytes Associated with in vivo Tumor Rejection PNAS 1994 91 6458 6462 8022805
Brichard V Van Pel A Wolfel T Wolfel C De Plaen E Lethe B Coulie P Boon T The tyrosinase gene codes for an antigen recognized by autologous cytolytic T lymphocytes on HLA-A2 melanomas J Exp Med 1993 178 489 495 8340755 10.1084/jem.178.2.489
Robbins PF el-Gamil M Kawakami Y Stevens E Yannelli JR Rosenberg SA Recognition of tyrosinase by tumor-infiltrating lymphocytes from a patient responding to immunotherapy Cancer Res 1994 54 3124 3126 8205528
Gaugler B Van den Eynde B van der Bruggen P Romero P Gaforio JJ De Plaen E Lethe B Brasseur F Boon T Human gene MAGE-3 codes for an antigen recognized on a melanoma by autologous cytolytic T lymphocytes J Exp Med 1994 179 921 930 8113684 10.1084/jem.179.3.921
Chen YT Scanlan MJ Sahin U Tureci O Gure AO Tsang S Williamson B Stockert E Pfreundschuh M Old LJ A testicular antigen aberrantly expressed in human cancers detected by autologous antibody screening PNAS 1997 94 1914 1918 9050879 10.1073/pnas.94.5.1914
Marchand M van Baren N Weynants P Brichard V Dreno B Tessier MH Rankin E Parmiani G Arienti F Humblet Y Bourlond A Vanwijck R Lienard D Beauduin M Dietrich PY Russo V Kerger J Masucci G Jaeger E De Greve J Atzpodien J Brasseur F Coulie PG van der Bruggen P T. B Tumor regressions observed in patients with metastatic melanoma treated with an antigenic peptide encoded by gene MAGE-3 and presented by HLA-A1 Int J Cancer 1999 80 219 230 9935203 10.1002/(SICI)1097-0215(19990118)80:2<219::AID-IJC10>3.0.CO;2-S
Cebon J Jaeger E Shackleton MJ Gibbs P Davis ID Hopkins W Gibbs S Chen Q Karbach J Jackson H MacGregor DP Sturrock1 S Vaughan H Maraskovsky E Neumann A Hoffman E Sherman ML Knuth A Two phase I studies of low dose recombinant human IL-12 with Melan-A and influenza peptides in subjects with advanced malignant melanoma Cancer Immunity 2003 3 7 25 12862418
Chen Q Jackson H Shackleton M Parente P Hopkins W Sturrock S MacGregor D Maraskovsky E Tai TY Dimopoulos N Masterman KA Luke T Davis ID Chen W Cebon J Characterization of antigen-specific CD8+ T lymphocyte responses in skin and peripheral blood following intradermal peptide vaccination Cancer Immunity 2005 5 5 12 15755075
Ferrone S Finerty JF Jaffee EM Nabel GJ How much longer will tumour cells fool the immune system? Immunology Today 2000 21 70 72 10712001 10.1016/S0167-5699(99)01569-8
Gilboa E The makings of a tumor rejection antigen Immunity 1999 11 263 270 10514004 10.1016/S1074-7613(00)80101-6
Heiser A Coleman D Dannull J Yancey D Maurice MA Lallas CD Dahm P Niedzwiecki D Gilboa E Vieweg J Autologous dendritic cells transfected with prostate-specific antigen RNA stimulate CTL responses against metastatic prostate tumors. The Journal of Clinical Investigation 2002 109 409 417 11828001 10.1172/JCI200214364
Granstein RD Ding W Ozawa H Induction of anti-tumor immunity with epidermal cells pulsed with tumor-derived RNA or intradermal administration of RNA. J Invest Dermatol 2000 114 632 636 10733665 10.1046/j.1523-1747.2000.00929.x
Jaeger E Ringhoffer M Karbach J Arand M Oesch F Knuth A Inverse relationship of melanocyte differentiation antigen expression in melanoma tissues and CD8+ cytotoxic-T-cell responses: evidence for immunoselection of antigen-loss variants in vivo Int J Cancer 1996 66 470 476 8635862 10.1002/(SICI)1097-0215(19960516)66:4<470::AID-IJC10>3.0.CO;2-C
Jaeger E Ringhoffer M Altmannsberger M Arand M Karbach J Jager D Oesch F Knuth A Immunoselection in vivo: independent loss of MHC class I and melanocyte differentiation antigen expression in metastatic melanoma. Int J Cancer 1997 71 142 147 9139833 10.1002/(SICI)1097-0215(19970410)71:2<142::AID-IJC3>3.0.CO;2-0
Kerkmann-Tucek A Banat GA Cochlovius B Zoller M Antigen loss variants of a murine renal cell carcinoma: implications for tumor vaccination. Int J Cancer 1998 77 114 122 9639402 10.1002/(SICI)1097-0215(19980703)77:1<114::AID-IJC18>3.0.CO;2-E
Singluff CLJ Colella TA Thompson L Graham DD Skipper JC Caldwell J Brinckerhoff L Kittlesen DJ Deacon DH Oei C Melanomas with concordant loss of multiple melanocytic differentiation proteins: immune escape that may be overcome by targeting unique or undefined antigens. Cancer Immunol Immunother 2000 48 661 672 10752474 10.1007/s002620050015
Boon T van der Bruggen P Human tumor antigens recognized by T lymphocytes. The Journal of Experimental Medicine 1996 183 725 729 8642276 10.1084/jem.183.3.725
Jaeger E Bernhard H Romero P Ringhoffer M Arand M Karbach J Ilsemann C Hagedorn M Knuth A Generation of cytotoxic T-cell responses with synthetic melanoma-associated peptides in vivo, implications for tumor vaccines with melanoma-associated antigens. Int J Cancer 1996 66 162 169 8603805 10.1002/(SICI)1097-0215(19960410)66:2<162::AID-IJC4>3.0.CO;2-0
Soiffer R Lynch T Mihm M Jung K Rhuda C Schmollinger JC Hodi FS Liebster L Lam P Mentzer S Singer S Tanabe KK Cosimi AB Duda R Sober A Bahn A Daley J Neuberg D Parry G Rokowich J Richards L Drayer J Berns A Clift S Dranoff G Vaccination with irradiated autologous melanoma cells engineered to secrete human granolucyte-macrophage colony-stimulating factor generates potent antitumor immunity in patients with metastatic melanoma. Proc Natl Acad Sci USA 1998 95 13141 13146 9789055 10.1073/pnas.95.22.13141
Nestle FO Alijagic S Gilliet M Sun Y Grabbe S Dummer R Burg G Schadendorf S Vaccination of melanoma patients with peptide- or tumor lysate-pulsed dendritic cells. Nature 1998 4 328 332 10.1038/nm0398-328
Boczkowski D Nair SK Snyder D Gilboa E Dendritic cells pulsed with RNA are potent antigen presenting in vitro and in vivo. Journal of Experimental Medicine 1996 184 465 472 8760800 10.1084/jem.184.2.465
Simons JW Mikhak B Chang JF DeMarzo AM Carducci MA Lim M Weber CE Baccala AA Goemann MA Clift SM Ando DG Levitsky HI Cohen LK Sanda MG Mulligan RC Partin AW Carter HB Piantadosi S Marshall FF Nelson WG Induction of immunity to prostate cancer antigens: results of a clinical trial of vaccinated with the irradiated autologous prostate tumor cells engineered to secreted granulocyte-macrophage colony-stimulating factor using ex vivo gene transfer. Cancer Research 1999 59 5160 5168 10537292
Kobayashi T Yamanaka R Homma J Tsuchiya N Yajima N Yoshida S Tanaka R Tumor mRNA-loaded dendritic cells elicit tumor-specific CD8+ cytotoxic T cells in patients with malignant glioma. Cancer Immunol Immunother 2003 52 632 637 12827308 10.1007/s00262-003-0408-5
Griffioen M Borghi M Schrier PI Osanto S Schadendorf D Analysis of T-cell responses in metastatic melanoma patients vaccinated with dendritic cells pulsed with tumor lysates. Cancer Immunol Immunother 2004 53 715 722 14997347 10.1007/s00262-004-0514-z
Boczkowski D Nair SK Nam JH Lyerly HK Gilboa E Induction of tumor immunity and cytotoxic T lymphocytes responses using dendritic cells transfected with messenger RNA amplified from tumor cells. Cancer Research 2000 60 1028 1034 10706120
Heiser A Maurice MA Yancey DR Wu NZ Dahm P Pruitt SK Boczkowski D Nair SK Ballo MS Gilboa E Vieweg J Induction of polyclonal prostate cancer-specific CTL using Dendritic cells transfected with amplified tumor RNA. The Journal of Immunology 2001 166 2953 2960 11207244
Grünebach F Müller MR Nencioni A Brossart P Delivery of tumor-derived RNA for the induction of cytotoxic T-lymphocytes. Gene Therapy 2003 10 367 374 12601391 10.1038/sj.gt.3301901
Hoerr I Obst R Rammensee HG Jung G In vivo application of RNA leads to induction of specific cytotoxic T lymphocytes and antibodies. European Journal of Immunology 2000 30 1 7 10602021 10.1002/1521-4141(200001)30:1<1::AID-IMMU1>3.3.CO;2-R
Seth D Gorrell MD McGuinness PH Leo MA Lieber CS McCaugan GW Haber PS SMART amplification maintains representation of relative gene expression: quantitative validation by real time PCR and application to studies of alcoholic liver disease in primates. J Biochem Biophzs Methods 2003 55 53 66 10.1016/S0165-022X(02)00177-X
Wellenreuther R Schupp I Consortium TGDNA Poutska A Wiemann S SMART amplification combined with cDNA size fractionation in order to obtain large full-length clones. BMC Genomics 2004 5 36 44 15198809 10.1186/1471-2164-5-36
Yu J Russell E Structural and functional analysis of an mRNP complex that mediates the high stability of human ß-globin mRNA. Molecular and Cellular Biology 2001 21 5879 5888 11486027 10.1128/MCB.21.17.5879-5888.2001
Wang Z Day N Trifillis P Kiledjian M An mRNA stability complex functions with poly(A)-binding protein to stabilize mRNA in vitro. Molecular and Cellular Biology 1999 19 4552 4560 10373504
Altschul SF Gish W Miller W Meyers EW Lipman DJ Basic local alignment search tool. J Mol Biol 1990 215 403 410 2231712 10.1006/jmbi.1990.9999
Draper MP August PR Connolly T Packard B Call KM Efficient cloning of full-length cDNAs based on cDNA size fractionation. Genomics 2002 79 603 607 11944994 10.1006/geno.2002.6738
Wibe E Paus E Aamdal S Neuron specific enolase (NSE) in serum of patients with malignant melanoma. Cancer Letters 1990 52 29 31 2354416 10.1016/0304-3835(90)90073-7
Kondoh N Schweinfest CW Henderson KW Papas TS Differential Expression of S19 Ribosomal-Protein, Laminin-Binding Protein, and Human Lymphocyte Antigen Class-I Messenger-Rnas Associated with Colon-Carcinoma Progression and Differentiation Cancer Res 1992 52 791 796 1339304
Affymetrix´s Eukaryotic Sample and Array Processing [http://www.affymetrix.com/support//downloads/manuls/expression_s2_manual.pdf]
Van Gelder RN von Zastrow ME Yool A Dement WC Barchas JD Eberwine JH Amplified RNA synthesized from limited quantities of heterogeneous cDNA. Proc Natl Acad Sci USA 1990 87 1663 1667 1689846
GNF Gene Expression Atlas [http://expression.gnf.org/cgi bin/index.cgi]
Wang FL Wang YB Wong WK Liu Y Addivinola FJ Liang P Chen LB Kantoff PW Pardee AB Two differentially expressed genes in normal human prostate tissue and in carcinoma Cancer Res 1996 56 3634 3637 8705997
Kroes RA Jastrow A McLone MG Yamamoto H Colley P Kersey DS Yong VW Mkrdichian E Cerullo L Leestma J Moskal JR The identification of novel therapeutic targets for the treatment of malignant brain tumors Cancer Lett 2000 156 191 198 10880769 10.1016/S0304-3835(00)00462-6
Standal T Borset M Sundan A Role of osteopontin in adhesion, migration, cell survival and bone remodeling Exp Oncol 2004 26 179 184 15494684
Jiang HP Lin JJ Tao J Fisher PB Suppression of human ribosomal protein L23A expression during cell growth inhibition by interferon-beta Oncogene 1997 14 473 480 9053844 10.1038/sj.onc.1200858
Banerjee A Increased levels of tyrosinated alpha-, beta(III)-, and beta(IV)-tubulin isotypes in paclitaxel-resistant MCF-7 breast cancer cells Biochem Biophys Res Commun 2002 293 598 601 12054644 10.1016/S0006-291X(02)00269-3
Adams SM Sharp MGF Walker RA Brammar WJ Varley JM Differential Expression of Translation-Associated Genes in Benign and Malignant Human Breast-Tumors Br J Cancer 1992 65 65 71 1370760
Friedman RS Bangur CS Zasloff EJ Fan LQ Wang TT Watanabe Y Kalos M Molecular and immunological evaluation of the transcription factor SOX-4 as a lung tumor vaccine antigen J Immunol 2004 172 3319 3327 14978140
Bommer UA Thiele BJ The translationally controlled tumour protein (TCTP) Int J Biochem Cell Biol 2004 36 379 385 14687915 10.1016/S1357-2725(03)00213-9
Finch JS Stjohn T Krieg P Bonham K Smith HT Fried VA Bowden GT Overexpression of 3 Ubiquitin Genes in Mouse Epidermal Tumors Is Associated with Enhanced Cellular Proliferation and Stress Cell Growth Differ 1992 3 269 278 1321656
|
16115316
|
PMC1215502
|
CC BY
|
2021-01-04 16:39:08
|
no
|
Genet Vaccines Ther. 2005 Aug 22; 3:6
|
utf-8
|
Genet Vaccines Ther
| 2,005 |
10.1186/1479-0556-3-6
|
oa_comm
|
==== Front
Health Qual Life OutcomesHealth and Quality of Life Outcomes1477-7525BioMed Central London 1477-7525-3-541614455510.1186/1477-7525-3-54ResearchDevelopment and validation of a psychosocial screening instrument for cancer Linden Wolfgang [email protected] Dahyun [email protected] Maria Cristina [email protected] Regina [email protected] Richard [email protected] Psychology Department, The University of British Columbia, 2136 West Mall, Psychology/UBC, Vancouver BC, V6T 1Z4, Canada2 British Columbia Cancer Control Agency, Canada3 Health Care and Epidemiology, University of British Columbia, Canada2005 7 9 2005 3 54 54 13 4 2005 7 9 2005 Copyright © 2005 Linden et al; licensee BioMed Central Ltd.2005Linden et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
We are reporting on the development of a psychosocial screening tool for cancer patients. The tool was to be brief, at a relatively low reading level, capture psychological variables relevant to distress and health-related quality-of-life in cancer patients, possess good reliability and validity, and be free of copyright protection.
Method
Item derivation is described, data on reliability and validity as well as norms are reported for three samples of cancer patients (n = 1057; n = 570, n = 101).
Results
The resulting 21-item psychological screen for cancer (PSCAN) assesses perceived social support, desired social support, health-related quality-of-life, anxiety and depression. It has good psychometrics including high internal consistency (alpha averaging .83, and acceptable test-retest stability over 2 months (averaging r = .64). Validity has been established for content, construct and concurrent validity.
Conclusion
PSCAN is considered ready for use as a screening tool and also for following changes in patient distress throughout the cancer care trajectory. It is freely available to all interested non-profit users.
Screeningdistressdepressionanxietyhealth-related quality of lifesocial supportnormsreliabilityvalidity
==== Body
Background
Cancer is now the leading cause of early death in Canada . A diagnosis of cancer is very emotionally threatening, may provoke anxiety or depression, and is difficult to live with because all aspects of life are overshadowed by the typical prognostic uncertainty [1-3]. Nevertheless, there is great variability in how patients respond to the diagnosis and this may be partly explained by the nature and quality of support that patients have, individual coping skills and by the meaning that they can learn to assign to this threat [4]. Psychological interventions for distress reduction can enhance quality-of-life, and help patients and families better cope [2,5,6] but distress often remains unrecognized and therefore untreated [7].
A number of critical questions for clinical practice arise from these insights, namely which intervention works best, and which patients are particularly needy and potentially responsive to a professional intervention. Screening for psychological distress and identification of emotional needs have important practical implications because practitioners want to responsibly serve their clientele and, given the scarcity of professional psychological resources, they want to make these resources available in the most equitable and efficient manner possible [5]. Researchers can help by extracting critical information from basic research for the explicit purpose of informing clinicians [8] and to ascertain best patient care.
Screening research has its own theoretical basis and concerns [1,5,9,10,12]. Screening can be expensive and has built into it the moral and professional imperative that one needs to act on urgent needs once identified. Along these lines, Rodin [8] and Ryan et al. [7] stress that screening and feedback does not necessarily lead to better patient outcomes unless such measures are accepted by the institutions and are supported by corresponding allocation of resources. However, only about one in four patients who report significant distress are actually referred to psychosocial care [9,13]. This, in turn, suggests that screening research is best done in a clinical setting with the active involvement of those professionals who are also key players in its intended subsequent routine implementation.
Given the high incidence of cancer, screening for distress requires tools that are psychometrically sound, inexpensive and quick, accepted by patients and staff, and of sufficient simplicity to be accessible to as many patients as possible [5]. In a review of the most frequently used tools for psychosocial distress screening [9], it became apparent that (a) most often measured are anxiety and depression, (b) there is no agreement on the best screening tool, (c) many of these measures are too long for routine screening, and (d) almost all of the tools used are copyrighted protected and would have to be purchased for every application. For the sake of parsimony, we decided to focus on psychological concepts that are know to be particularly critical for cancer patients, namely anxiety, distress [2,9], but also wanted to measure patient characteristics that reflect more positive aspects of life namely social support and quality-of-life [4,6,11,13]. Aside from the primary objective of developing a screening tool, we hoped that the psychometrics would support use of the same tool for tracking emotional adjustment in patients. Lastly, we wanted to include an instruction that questionnaire completion would imply permission for clinicians and researchers to directly contact patients to offer services or invite research participation. If agencies and patients give such permission, both clinical service provision and patient identification for research are made much less cumbersome.
In light of these observations, we intended to develop a tool that embraced all of the desired features listed above. While our work on a brief psychosocial tool does not claim to break completely new ground (see the review of previously used screening tools [9] and the NCCN guidelines [5], we posit that it stands out because of (a) its brevity, (b) its development in the clinical context where it was to become implemented, (c) the scope of the domains included, (d) that it measures both negative and positive aspects of the patients' quality of life, and (e) its non-commercial nature. A series of three studies (subdivided into Phase Ia, Ib, and Phase II) was planned to establish psychometrics and norms in two test phases. The objectives for Phase Ia were determination of item clarity, basic reliability, and validity features including internal consistency, desired and undesired factor correlations as a test of construct validity, as well as to generate initial norms for a cancer patient population. A second sample was tested in Phase Ib to also identify cancer-type specific norms, and gender-specific norms.
For Phase II, a third (smaller) sample was evaluated initially and then retested for establishment of test-retest reliability; in addition, participants completed a larger test package that permitted concurrent validation of the PSCAN subscales with well established and substantially longer versions of tests that we considered "industry standard" in order to show adequacy of content sampling and concurrent validity for PSCAN.
Methods
Phase I
The instrument
After extensive discussion of existing instruments and their respective strengths and weaknesses, the authors and a group of practicing clinicians agreed to focus on anxiety, depression, social support, and health-related quality-of-life. The final version of the full scale as described here is found in the Additional file 1 as are instructions for how to obtain permission for its use from the authors. (Given that no acceptable measure of anxiety and depression could be found that did not have to be purchased from a commercial publisher, a number of popular scales were studied to obtain a clear sense of content domains). Five items each were written to elicit patients' level of anxiety and depression. Each item is scaled as 1–5 ('not at all' to 'very much so')
The social support items are derived from a social support scale used in the Epidemiological Study of the Elderly [14] that is considered to be in the public domain. It provides 5 items that, when clustered, are referred to as Social Network and Support Assessment (SNSA) which taps into available informational, instrumental, and emotional support. One additional item (not part of the SNSA) asks how much social support people desire (Item 6). The SNSA has been reported to have internal consistencies of .47 to .61 [15]. The SNSA was found to predict mortality in epidemiological studies [14] and has also been used in cancer populations [15]; the existence of distinct subscales was shown to possess discriminant validity because therapy-induced changes were apparent on instrumental and informational support but not on other items [15].
Given that desired support and received support generally do not intercorrelate and that mismatched support attempts are not constructive [6,16], the 'desired support' item was written as a single item scale with a 1–9 rating ('not at all' to 'very much so') to permit variability in ratings. This single item scale has already shown remarkable clinical usefulness because Krumholz et al. [17] found that, in a sample of 292 elderly patients with heart failure, those patients not seeking and not receiving support were three times more likely to be alive one year later that those patients who did seek support but did not receive it. Apparently, social support is only useful when its availability is actually also desired. All other social support items (items 1–5) are rated as yes/no and coded 0 or 1. While the direction of scoring is ultimately arbitrary, we treated positive support ratings as high scores.
Quality-of-life measures are generally distinguished as being either broad and generic or disease-specific with researchers favoring the inherent sharper focus of disease-specific tools. That notwithstanding, we did not decide on a highly disease-specific measure because many distressing physical aspects of cancer (like pain and functional limitations) are only salient in late stage cancer and are, fortunately, of limited importance to the lives of early-stage cancer patients who form the majority of study participants. We therefore decided to keep the assessment of health-related quality-of-life (HRQoL) sufficiently broad and generic to embrace cancer patients in all stages. The chosen items are from the Health Related Quality of Life questionnaire developed by the Centers for Disease Control [18] and are part of the Behavioural Risk Factor Surveillance System since 1993. The questions seek to learn about HRQoL by distinguishing between global, self-rated health and numbers of days negatively affected by poor mental or physical health (PSCAN items 7–11). Test-retest stability of this tool was established in a sample of 868 adults with kappas ranging from = .58 to .75; reliability was not affected by gender or race [19]. The scale has also been used in a Canadian sample of 926 men and women age 65 or over and revealed that those with poor self-rated health showed a 17-fold increase in the number of unhealthy days [20].
Study population
Sample 1
Participants were 1057 consecutive cancer patients coming into first contact with the Vancouver Cancer Centre (i.e., after a positive diagnosis had been established). No demographic information was collected at this time because the data were only to be used in the item development process and not any kind of hypothesis testing. Participants were asked by the receptionist to complete the 21-item PSCAN after they had read the instructions and agreed to participate. The plan was to collect initial data for a 3-month time period rather than setting a particular sample size as a target; the rationale for this decision was to allow us later determination what percentage of the total number of eligible patients had actually participated; of all patients who had made initial contact with the cancer agency during this same time period, about 90% had indeed completed PSCAN which suggests that this sample is quite representative of the typical patient population seen by this agency.
Feedback from patients suggested that the wording of two items was somewhat ambiguous and these were subsequently changed prior to the recruitment of sample 2.
Sample 2
Participants were 547 consecutive cancer patients (average age 66.5 yrs (SD = 14.5), 304 women and 243 men) coming into first contact with the Fraser Valley Cancer Centre (i.e., after a positive diagnosis had been established). The procedure was the same as the one applied to sample 1.
Phase II
The sample consisted of 101 cancer patients making first contact with the BC Cancer Agency at the Vancouver Center (41 male, 60 female). Eligible patients were recruited consecutively by two trained research assistants over a period of one month. The research assistants were physically located in the reception area, were alerted about potentially eligible patients by the receptionist, and then approached patients individually to explain the study, seek consent, and request completion of a test package. Patients were explicitly recruited to participate in a test-retest study and indicated whether they preferred recontact by mail or telephone. Two months later, patients were recontacted according to their preferred method. If no contact could be made by telephone after three attempts, no further attempts were made. Patients who received the questionnaire package by mail, were not further reminded to return them. This relatively 'low pressure approach' still resulted in a 65% return rate of completed retest materials. The questionnaire package consisted of the PSCAN as described above, the Hospital Anxiety and Distress Scale (HADS) and the ENRCHD Social Support Instrument [ESSI; [21,22]].
The HADS is a very frequently used 14-item scale tapping anxiety and depression. Bjelland et al. [23] provided a review of the psychometrics of HADS based on 747 published studies and reported Cronbach's alphas of .68 to .93 for anxiety and .67 to .90 for depression. Factor analyses routinely confirm the underlying 2-factor structure.
The ESSI is a 7-item instrument with strong test-retest reliability (means one month apart were 27.8 and 27.8), and internal consistency in a sample of 271 cardiac patients was .88. Concurrent validity was shown by relating ESSI scores to established psychiatric diagnoses of depression and an index of social functioning [21,22].
Results
The findings obtained from the three samples during Phase I and II testing are presented in aggregated form that presents findings organized around the test's properties regarding means and reliability (Tables 1, 2, 3), and then reports on evidence of validity. Finally, raw means and standard deviations are provided for each cancer type and each gender group (Table 4). This presentational approach appeared more parsimonious and provided a more logical organization than a mere sequential listing of each temporal step of the result finding process.
Table 1 Subscale Descriptive Data (Means, and variability), Phase I
Range of scores Mean sample 1 SD sample 1 Mean sample 2 SD sample 2
SS_Total 0–5 4.6 .82 4.5 .99
SS_Desired 0–10 4.0 3.5 3.7 3.4
QOL, Perceived 0–20 6.5 5.4 6.7 5.5
QOL, Days 0–120 31.6 29.3 24.0 24.3
Anxiety 5–24 8.1 3.8 8.2 4.2
Depression 5–25 8.1 3.9 8.2 5.1
Table 2 Internal Consistency
Internal Consistency Alpha Sample 1
QOL_Days 0.79
QOL_Perceived 0.89
Anxiety 0.83
Depression 0.79
Table 3 Test-retest stability
Subscale Stability coefficient r M and SD Time 1 M and SD Time 2
Social Support Total .87 5.3 (.73) 5.3 (.74
Social support desired .59 4.2 (3.3) 4.4 (3.4)
Anxiety .67 7.4 (3.3) 6.6 (2.4)
Depression .61 7.7 (3.8) 7.0 (2.4)
QoL Perceived .59 4.2 (4.3) 5.6 (5.6)
QoL Days .49 22.3 (26.5) 20.5 (27.4)
Table 4 Subscale means for each gender and cancer type
Cancer Type Subscale – Mean (SD)
Depression Anxiety QoL-P QoL-D SS-Tot SS-Des
Gastro-Intestinal
Female (n = 39) 18.3 (12.9) 8.4 (3.7) 6.2 (5.7) 26.7 (23.9) 4.9 (.5) 4.0 (3.9)
Male (n = 39) 13.3 (4.4) 6.7 (2.7) 7.1 (4.6) 31.3 (21.7) 4.9 (.4) 2.7 (3.5)
Lung
Female (n = 35) 17.6 (8.6) 9.2 (4.5) 8.7 (4.9) 30.6 (22.9) 4.0 (1.2) 4.4 (3.6)
Male (n = 42) 17.0 (6.9) 8.9 (5.6) 11.7 (6.1) 41.3 (22.8) 4.6 (.7) 3.9 (3.2)
Lymphoma
Female (n = 12) 22.6 (8.6) 12.1 (4.4) 9.9 (7.9) 35.6 (28.4) 4.9 (.3) 5.9 (4.0)
Male (n = 13) 19.6 (11.9) 9.6 (6.4) 11.0 (6.6) 35.0 (24.6) 4.7 (.5) 4.2 (3.9)
Breast
Female (n = 133) 15.6 (6.5) 7.9 (4.3) 5.0 (4.4) 21.3 (21.3) 4.5(.9) 3.8(3.4)
Male (n = 1) 10.0 (na) 5.0 (na) 2.0 (na) 0 (na) 5.0 (na) 5.0 (na)
Melanoma
Female (n = 9) 14.5 (7.1) 8.3 (5.1) 2.1 (3.4) 12.3 (21.5) 4.7 (.5) 5.0 (3.5)
Male (n = 12) 13.2 (4.2) 6.6 (1.8) 3.9 (2.6) 4.2 (5.2) 4.6 (.7) 1.7 (2.2)
Head & Neck
Female (n = 12) 26.1 (16.0) 11.0 (3.7) 8.7 (5.9) 25.0 (21.9) 4.3 (.8) 5.6 (3.2)
Male (n = 18) 15.3 (5.8) 8.3 (4.0) 6.6 (5.8) 20.8 (24.7) 4.5 (1.2) 3.6 (3.1)
Genito-Urinary (incl Prostate)
Female (n = 5) 12.0 (2.8) 6.2 (2.2) 6.4 (3.8) 17.4 (24.8) 4.8 (.5) 1.3 (1.5)
Male (n = 83) 14.6 (6.5) 7.4 (3.5) 5.9 (5.1) 14.6 (27.6) 4.4 (1.2) 2.8 (3.4)
QoL-P = Quality-of-Life Perceived; QoL-D = Quality-of-Life Days; SS-Tot = Social Support Total; SS-Des = Social Support Desired
As the mean scores in Table 1 show, indices of variability reveal that participants used a wide range of scores that in turn suggests that PSCAN is sensitive in discriminating among patients. Means for the two samples were quite similar.
Reliability
As Table 2 reveals the internal consistencies for these four subscales are high and satisfy traditional cutoffs for full-length tests despite their brevity in PSCAN. Internal consistency was not computed for the Social Support items, because each item was designed to tap somewhat different dimensions of support and the yes/no scoring method did not create much item response variability that could then be meaningfully analyzed.
Scores for social support variables were very stable over 2 months whereas QOL and distress-related variables showed less stability although they were still moderately high (Table 3).
Validity
A number of steps were undertaken to establish construct validity. Basic requirements for test creation [24] are stated below and corresponding results are listed for each:
(a) the items that make up a distinct subscale (and that presumably reflect an underlying 'factor') intercorrelate with each other (but not so highly that they suggest duplication) and that they load (i.e., correlate) with the total subscale score. Given the complexity of results, they are not reported in detail but the pattern of results clearly indicates that this requirement was met. For example, the Quality-of-Life items correlated between r = .49 and r = .95 with the average of inter-item correlations being r = .73.
(b) the subscale scores themselves should correlate with each other if they are conceptually related. Based on extant literature, it is expected that anxiety and depression will partly overlap, and high social support and QoL should correlate to some degree with low depression and anxiety. This was confirmed with r's ranging from .55 to .92.
(c) generally, subscale scores across domains should not highly intercorrelate with each other because that would suggest redundancy that is especially undesirable in a brief screening tool. These test properties were determined with a series of correlational analyses and supported the notion of minimal overlap in general [20]. The data suggest that the three 'QoL days' items and the two 'QoL Perceived' items overlap but still tap different aspects of QoL; each respective item correlates highly with its own subscale total score. This finding supports the continued inclusion of these two sets of QoL items.
Anxiety and depression were predictably intercorrelated (r = .71 and r = .55 in sample 1 and 2 respectively) and explain about 30–50% of each other's variance.
Correlation coefficients of the rather short PSCAN subscales with equivalent longer version from established tests strongly support concurrent validity. The r-scores for samples 1 and 2 respectively were .72 and .82 for anxiety; .59 and .75 for depression, and .61 and .61.
Lastly, we computed means for men and women and each subscale as a function of cancer type. Only those types of cancer were listed where at least 5 men and 5 women were found to fill each cell (with the exception of frequent cancers that are only found in one gender). Results are displayed in Table 4. No inferential testing was conducted because the power for tests varies greatly as a function of the varying sample sizes per cell; we do, however, report effect sizes because this display of psychological profiles for all cancer types can serve as a hypothesis generator for future studies and may enable power calculations for sample size estimation.
It is tempting to interpret the results shown in Table 4 but given that no inferential tests were computed, any interpretation is speculative should be kept at a minimum. It appears that men typically report less negative affect than women and that there is considerable variability in distress and QoL as a function of cancer type. This suggests an urgent need for the accrual of a larger sample including all cancer types such that sufficiently powered inferential tests can be conducted.
Discussion
The objective of this tool development process was to gather enough information so that readers could potentially make a decision to adopt the scale for their own use with cancer populations, knowing that adequate reliability and validity testing had been undertaken. We consider the psychometric characteristics of PSCAN to be satisfactory especially when considering that it is a very brief tool [6]; the scale characteristics typically met even the desired standards for longer scales. Subscale means for two large samples recruited at different sites were very similar suggesting stability of the scale scores. Practitioners can now choose to use PSCAN instead of the recommended single-item distress thermometer [5] or use it in a complementary fashion, as a second step, if the single item distress thermometer suggests elevated distress. A second advantage of PSCAN is that the standard instructions can include a statement about the patient having consented to be contacted for offers of professional help or participation in research. This feature has been found very useful in clinical settings where research is also being conducted because Ethics Reviews Boards do not usually permit direct contacting (i.e., "cold calls") of patients for study recruiting.
Reliability
Internal consistency was good across the two independent samples and test-retest stability was also acceptable. Note that at the time of recruitment for the test-retest study (Phase II), patients had come for their first visit to the cancer center. During the 8-week interval for test and retest, these patients typically learned more about their diagnosis and many began active treatment. It was therefore expected that these experiences would affect distress and quality-of-life, and that the test-retest scores would only be moderately high. In fact, had the test-retest scores come close to r = 1, this would have suggested that PSCAN was insensitive to measuring change and that would have been indicative of a significant weakness in the test.
Validity
Individual item correlations with their respective subscale scores were high, suggesting that they load appropriately on the constructs to be measured. All relationships between constructs measured by the PSCAN were of a strength and direction so as to be consistent with the literature and that, in turn, suggests that PSCAN possesses construct validity. The newly written anxiety and depression items not only formed cohesive subscales with some predicted overlap of anxiety and depression but also correlated highly with longer versions of anxiety and depression scales of well established tools thus indicating high concurrent validity. There is no firm consensus on how to establish content validity other than finding that experts agree. The research group who developed PSCAN, the cancer agency staff who worked with it, and the patients who responded, all considered the items as meaningful and sensitive. It was also interesting to see that once Phase I had been completed and PSCAN-derived patient distress information became known to the psychosocial support staff, they reported quickly occurring changes in their clientele's composition because the patients now seen by family and patient counseling included more men and more minority patients who apparently had gone undetected by previous practice patterns. Interestingly, the use of PSCAN and especially the inclusion of an item on suicidal thinking, also led the agency' s staff to review and clarify their policy on how to respond to highly distressed, potentially suicidal patients. Hence, the development of this screening tool by a 'mixed' team of researchers and service providers also led to ready acceptance of the tool by service providers and triggered prompt changes in their practice patterns. This could be considered a form of ecological validity.
Conclusion
We conclude that there is sufficient psychometric evidence to support PSCAN's use as a screening tool and possibly as a tool for tracking patient changes in emotional well-being. Consistent with the larger literature on distress screening, it is likely that PSCAN's strength is in sensitivity of detection, not in specificity; in order to advance to a full-fledged psychiatric diagnosis, further standardized testing as well as individual histories and assessments will need to be considered. At this time, there are enough normative data available so that one can decide on a clear, replicable cutoff based on percentile scores such that an agency might decide to offer psychological services to all patients scoring above 80th percentile of depression for example. However, we believe strongly that more information is needed before establishing a cutoff that also has a clinical meaning, like a demonstration of the health consequences of not offering services to patients above a given cutoff. We also recommend that norms be compiled for a large, representative cancer sample that fairly represent both sexes, all cancer subtype populations, and cultural minorities. Compilation of norms for different stages in the trajectory of cancer care should equally be considered. This, in turn, will aid empirically-based, future decisions about use of cutoffs [25]. Furthermore, it will be important to collect normative data from a physically and psychologically healthy sample and another medical sample so as to place the data obtained from cancer patients in a larger population context, and to serve as reference levels. Although PSCAN was explicitly developed and validated for cancer patients, the measured concepts are not uniquely relevant for cancer but appear pertinent for other chronic disease populations as well [1].
Lastly, PSCAN, is freely available to non-profit users who need, however, to contact the copyright holder for permission and conditions of use via our website: .
Instructions for scoring and status quo of knowledge about norms and cutoffs will be made available at that time.
List of abbreviations
ESSI ENRICHD Social Support Instrument
HADS Hospital anxiety and Distress Scale
PSCAN Psychosocial Screen for Cancer
HRQoL Health-Related Quality of Life
SNSA Social Network and Support Assessment
Supplementary Material
Additional File 1
Linden additional file.doc PSCAN – Psychological Screening Tool.
Click here for file
==== Refs
Coyne JC Thompson R Klinkman MS Nease DE Emotional disorders in primary care J Cons Clin Psychol 2002 70 798 809 10.1037//0022-006X.70.3.798
DeBoer MF Ryckman RM Pruyn JFA Van den Borne HW Psychosocial correlates of cancer relapse and survival: A literature review Patient Educ Couns 1999 37 215 230 14528548 10.1016/S0738-3991(99)00029-4
Zabora J BrintzenhofeSzoc K Curbow B Hooker C Piantadosi S The prevalence of psychological distress by cancer site Psychooncology 2001 10 19 28 11180574 10.1002/1099-1611(200101/02)10:1<19::AID-PON501>3.3.CO;2-Y
Brennan J Adjustment to cancer – Coping or personal transition? Psychooncology 2001 10 1 18 11180573 10.1002/1099-1611(200101/02)10:1<1::AID-PON484>3.3.CO;2-K
NCCN practice guidelines for the management of psychosocial distress National Comprehensive Cancer Network Oncology 1999 13 113 147 10370925
Helgeson VS Cohen S Schulz R Yasko J Group support interventions for women with breast cancer: Who benefits from what? Health Psychol 2000 19 107 114 10762094 10.1037//0278-6133.19.2.107
Ryan H Schofield P Cockburn J Butow P Tattersall M Turner J Girgis A Bandaranayake D Bowman D How to recognize and manage psychological distress in cancer patients Eur J Cancer Care 2005 14 7 15 10.1111/j.1365-2354.2005.00482.x
Butow P Devine R Boyer M Pendlebury S Jackson M Tattersall MHN Cancer consultation preparation packages: Changing patients but not physicians is not enough J Clin Oncol 2004 22 4401 4409 15514382 10.1200/JCO.2004.66.155
Carlson LE Bultz BD Cancer distress screening; needs, models, and methods J Psychosom Res 2003 55 402 409 10.1016/S0022-3999(03)00514-2
Palmer SC Coyne JC Screening for depression in medical care: Pitfalls, alternatives, and revised priorities J Psychosom Res 2003 54 279 287 12670603 10.1016/S0022-3999(02)00640-2
Rodin GM Suffering and adaptation to cancer: What to measure and when to intervene J Psychosom Res 2003 55 399 401 14581093 10.1016/S0022-3999(03)00527-0
Maunsell E Brisson J Deschenes L Frasure-Smith N Randomized trial of a psychological distress screening after breast cancer: Effects on quality of life J Clin Oncol 1996 14 2747 2755 8874336
Carlson LE Bultz BD Benefits of psychosocial oncology care: Improved quality of life and medical care offset Health Qual Life Outcomes 2003 1 8 12756059 10.1186/1477-7525-1-8
Seeman TE Berkman LF Structural characteristics of social networks and their relationship with social support in the elderly: Who provides support? Soc Sci Med 1988 23 737 749 3358145 10.1016/0277-9536(88)90065-2
Fobair P Koopman C Dimiceli S O'Hanlan K Butler LD Classen C Drooker N Davids HR Loulan J Wallsten D Spiegel D Psychosocial intervention for lesbians with primary breast cancer Psychooncology 2002 11 427 438 12228876
Reynolds JS Perrin NA Mismatches in social support and psychosocial adjustment to breast cancer Health Psychol 2004 23 425 430 15264980 10.1037/0278-6133.23.4.425
Krumholz HM Butler J Miller J Vaccarino V Williams CS Mendes de Leon CF Seeman TE Kasl SV Berkman LF Prognostic Importance of Emotional Support for Elderly Patients Hospitalized With Heart Failure Circulation 1998 97 958 964 9529263
Centers for Disease Control and Prevention (CDC) Behavioral Risk Factor Surveillance System Survey Questionnaire 1993 Atlanta, Georgia: U.S. Department of Health and Human Services, Centers for Disease Control and Prevention
Andresen E Catlin T Wyrwich K Retest reliability and validity of a surveillance measure of health-related quality of life Qual Life Res 1993 10 199
Ôunpuu S Chambers LW Patterson C Chan D Yusuf S Validity of the US Behavioral Risk Factor Surveillance System's Health Related Quality of Life Survey Tool in a group of older Canadians Chron Dis Canada 2001 22 93 101
Barefoot JC Burg MM Carney RM Cornell CE Czajkowski SM Freedland KE Hosking JD Khatri P Rogers Pitula C Sheps D Aspects of social support associated with depression at hospitalization and follow-up assessment among cardiac patients J Cardiopul Rehab 2003 23 404 412 10.1097/00008483-200311000-00002
Vaglio J JrConard M Poston WS O'Keefe J Haddock K House J Spertus JA Testing the performance of the ENRCHD Social Support Instrument in cardiac patients Health Qual Life Outcomes 2004 2 24 15142277 10.1186/1477-7525-2-24
Bjelland I Dahl AA Tangen Haug T Neckelmann D The validity of the Hospital Anxiety and Distress Scale J Psychosom Res 2002 52 69 77 11832252 10.1016/S0022-3999(01)00296-3
Aiken LR Psychological Testing and Assessment 1988 Allyn and Bacon, Boston, MA
Santor DA Coyne JC Evaluating the continuity of symptomatology between depressed and nondepressed individuals J Abn Psychol 2001 110 216 225 10.1037//0021-843X.110.2.216
|
16144555
|
PMC1215503
|
CC BY
|
2021-01-04 16:38:15
|
no
|
Health Qual Life Outcomes. 2005 Sep 7; 3:54
|
utf-8
|
Health Qual Life Outcomes
| 2,005 |
10.1186/1477-7525-3-54
|
oa_comm
|
==== Front
Health Qual Life OutcomesHealth and Quality of Life Outcomes1477-7525BioMed Central London 1477-7525-3-561615329410.1186/1477-7525-3-56ResearchHealth-related quality of life is related to COPD disease severity Ståhl Elisabeth [email protected] Anne [email protected] Sven-Arne [email protected]önmark Eva [email protected] Klas [email protected] Fredrik [email protected]öfdahl Claes-Göran [email protected]äck Bo [email protected] Department of Respiratory Medicine and Allergology, University Hospital, SE-221 85 Lund, Sweden2 AstraZeneca R&D Lund, SE-221 87 Lund, Sweden3 The OLIN Studies, Department of Medicine, Sunderby Central Hospital of Norrbotten, SE-971 80 Luleå, Sweden4 Department of Respiratory Medicine and Allergy, University Hospital, SE-901 85 Umeå, Sweden5 Lung and Allergy Research, National Institute of Environmental Medicine, the Karolinska Institute, SE-171 77 Stockholm, Sweden2005 9 9 2005 3 56 56 13 7 2005 9 9 2005 Copyright © 2005 Ståhl et al; licensee BioMed Central Ltd.2005Ståhl et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
The aim of this study was to evaluate the association between health-related quality of life (HRQL) and disease severity using lung function measures.
Methods
A survey was performed in subjects with COPD in Sweden. 168 subjects (70 women, mean age 64.3 years) completed the generic HRQL questionnaire, the Short Form 36 (SF-36), the disease-specific HRQL questionnaire; the St George's Respiratory Questionnaire (SGRQ), and the utility measure, the EQ-5D. The subjects were divided into four severity groups according to FEV1 per cent of predicted normal using two clinical guidelines: GOLD and BTS. Age, gender, smoking status and socio-economic group were regarded as confounders.
Results
The COPD severity grades affected the SGRQ Total scores, varying from 25 to 53 (GOLD p = 0.0005) and from 25 to 45 (BTS p = 0.0023). The scores for SF-36 Physical were significantly associated with COPD severity (GOLD p = 0.0059, BTS p = 0.032). No significant association were noticed for the SF-36, Mental Component Summary scores and COPD severity. Scores for EQ-5D VAS varied from 73 to 37 (GOLD I-IV p = 0.0001) and from 73 to 50 (BTS 0-III p = 0.0007). The SGRQ Total score was significant between age groups (p = 0.0047). No significant differences in HRQL with regard to gender, smoking status or socio-economic group were noticed.
Conclusion
The results show that HRQL in COPD deteriorates with disease severity and with age. These data show a relationship between HRQL and disease severity obtained by lung function.
Health-related quality of lifeCOPDdisease severityepidemiological, Global Initiative for Chronic Obstructive Lung Disease (GOLD)St George's Respiratory Questionnaire (SGRQ)
==== Body
Background
Chronic obstructive pulmonary disease (COPD) is a major cause of morbidity and mortality worldwide and is currently the fourth leading cause of death in the US [1]. It is a slowly progressive disease, characterized by lung function impairment with airway obstruction [2,3]. Common symptoms are cough, sputum production and shortness of breath. Smoking and different air pollutants, such as are well-known risk factors for COPD [3,2].
The prevalence of COPD varies considerably between countries and areas, from 3% in India [4] to 23% in the inner-city population of Manchester, UK [5]. The US National Health and Nutrition Examination Survey (NHANES) III survey puts the prevalence of COPD in the US at 7% [6]. The figure in Spain is similar, 9% [7]. In Sweden, the prevalence of COPD in those aged above 45 years was estimated to be 8% according to the British Thoracic Society (BTS) criteria and 14% according to the Global Initiative for Chronic Obstructive Lung Disease (GOLD) guidelines [8]. However, there are a considerable number of subjects with COPD who have not been diagnosed as such. In Europe and also in Sweden only one-quarter to one-third of those with COPD have been diagnosed as having COPD or with different labelling of the disease [8-11].
Over the past decade, more and more research on the development and validation of questionnaires has been undertaken to quantify the impact of disease on daily life and well-being from the COPD subject's point of view [12]. Health-related quality of life (HRQL), and preference-based HRQL instruments (utility instruments) are increasingly used in clinical studies. Although their use is established in many fields, such as oncology and gastrointestinal disease, questionnaires are rarely used as primary endpoints in randomised clinical studies of respiratory disease. One possible reason may be the lack of information about the patients' deterioration in HRQL when the disease progresses. The Medical Outcomes Study Short Form 36 (SF-36) and St George's Respiratory Questionnaire (SGRQ) are generic and disease-specific HRQL questionnaires, respectively [13,14]. The SF-36 has been used in a number of therapeutic areas, including COPD, while the SGRQ has been widely used in both COPD and asthma research. The EQ-5D is a generic, preference-based utility measure and has been used in a number of therapeutic areas [15].
The aim of the present study was to evaluate the association between HRQL and COPD stages using forced expiratory volume in one second as a percentage of predicted normal values (FEV1 % predicted) by means of two clinical guidelines for COPD, taking into account the influence on HRQL of age, gender, smoking status and socio-economic background. The association between HRQL and forced vital capacity as a percentage of predicted normal values (FVC % predicted) was also evaluated.
Methods
Study sample
A total of 202 subjects with COPD, recruited from a representative sample of the general population in northern Sweden, were invited; 176 subjects took part in this survey and data from 168 subjects were available [16]. The study cohort was derived from the Obstructive Lung Disease in Northern Sweden (OLIN) Studies [8,9], which has previously been described in detail [16].
Procedure
After initial instruction from the administrator, a qualified nurse, the questionnaires were completed unaided by subjects in the order SF-36, SGRQ and EQ-5D. A few subjects did not complete all questionnaires.
Definition and severity of COPD
The subjects were divided into four severity groups according to FEV1% predicted (pre-bronchodilator) using two different guidelines: the updated version (not yet published) of the GOLD guidelines [3] and the BTS guidelines [2]. The definition and severity criteria are described in Table 1. Calculation of FEV1predicted normal values for FEV1 was based on the reference values from ERS guidelines. In addition, levels of FVC % predicted were also used in the analysis instead of COPD severity stages.
Table 1 Severity criteria of COPD
Global Initiative for Chronic Obstructive Lung Disease, GOLD [3]: FEV1/FVC < 70%
I: Mild COPD FEV1 ≥ 80% predicted
II: Moderate COPD FEV1 50- < 80% predicted
III: Severe COPD FEV1 30- < 50% predicted
IV: Very severe COPD FEV1 < 30% predicted
British Thoracic Society, BTS [2]: FEV1/VC < 70% and FEV1 < 80% predicted
I: Mild COPD FEV1 60- < 80% predicted
II: Moderate COPD FEV1 40-59% predicted
III: Severe COPD FEV1 < 40% predicted
A group labelled BTS stage 0 was created for subjects with FEV1 ≥ 80% predicted: i.e. identical with mild COPD according to the GOLD criteria.
HRQL questionnaires
Short Form 36
The most widely used generic questionnaire, the Medical Outcomes Study Short Form 36 (SF-36), has been widely accepted in recent years as the best generic HRQL measurement. It contains 36 items divided into eight domains: Physical Functioning (PF), Role-Physical (RP), Bodily Pain (BP), General Health (GH), Vitality (VT), Social Functioning (SF), Role-Emotional (RE) and Mental Health (MH). These domains create a profile of the subject. Two summary scores can also be aggregated, the Physical Component Summary (PCS) and the Mental Component Summary (MCS). Scores range from 0 to 100, with higher scores representing better HRQL.
St George's Respiratory Questionnaire
The best-known and most frequently used disease-specific HRQL questionnaire for respiratory diseases, is the St George's Respiratory Questionnaire (SGRQ) [14,17]. The SGRQ is a standardized, self-administered questionnaire for measuring impaired health and perceived HRQL in airways disease. It contains 50 items, divided into three domains: Symptoms, Activity and Impacts. A score is calculated for each domain and a total score, including all items, is also calculated. Each item has an empirically derived weight. Low scores indicate a better HRQL. Recent publications by the developer (PW Jones) have confirmed that the minimal important difference relevant to the patients (MID) is 4 on a scale of 0 to 100 [18,19].
EQ-5D
The EQ-5D is a generic, preference-based utility questionnaire and consists of two parts, the EQ-5D VAS and the EQ-5D index. The EQ-5D has been used in a number of therapeutic areas and contains a vertical rating scale from 0 to 100 (EQ-5D VAS), with 0 = death/worst possible health and 100 = best possible health. The EQ-5D index is a five-item questionnaire ranging from 0 to 1. The items consist of mobility, self-care, usual activity, pain/discomfort and anxiety/depression. Each item has three levels: no problem, some problem and severe problem [15]. For the EQ-5D index, 0.03 has been regarded as the MID [20].
Statistical analysis
Statistical analysis was performed using an analysis of covariance model with HRQL scores as dependent variable. Three different approaches to analysis were performed using different classification of severity of COPD from GOLD and BTS guidelines. This classification was used as factor in the analysis. In all cases age, gender, smoking status and socio-economic background was used as covariates. These variables showed sign of influence on the HRQL measures and for the sake of comparability a unified model was selected for the analysis. An additional classification of severity based on FVC % predicted normal was also investigated with the same model with classification into four groups: stage I: > 95%, stage II: 95-81%, stage 3: 80-66% and stage IV; < 66%. These levels were chosen to have approximately equal number of patients in each group. Data presented in tables are adjusted least-square means from the adopted model.
Results
Subject characteristics
The mean age of the 168 subjects (70 women) was 64.3 years (range: 28–80 years). In the six age groups (the lowest < 45 and the highest > 79 years), 57 of the subjects were smokers and 85 were ex-smokers. Three socio-economic groups were identified (manual employees, non-manual employees and unemployed including housewives). Of the 138 'employees', 65 were still working and 73 had retired, and of these, 40 had retired before the normal age of retirement. Table 2 shows the subjects' characteristics.
Table 2 Subject characteristics
Characteristic Mean data (range)
n, total number agreed 176
Men/women n = 168 98/70
Mean age, years (range) n = 168 64.3 (28–80)
FEV1, L (range) n = 159 1.76 (0.46–4.12)
FEV1 % predicted (range) n = 159 62 (18–118)
Smoking status n = 171 Smoker, n = 57 Non-smoker, n = 29 Ex-smoker, n = 85
Socioeconomic group n = 174 Manual employees, n = 78 Non-manual employees, n = 60 Unemployed incl. housewives, n = 36
HRQL in relation to COPD severity according to GOLD
The differences in SF-36 PCS between the four severity groups were statistically significant (p = 0.0059). The scores for SF-36 (PCS) were 42 in the stage I group and 29 in the stage IV group. The corresponding scores for SF-36 MCS were 55 and 48 in stages I and IV respectively (p = 0.19) (Table 3).
Table 3 Health-related quality of life scores, adjusted mean values (± SD) – GOLD criteria
Scale FEV1 % predicted
≥ 80% Stage I n = 26 79-50% Stage II n = 91 49-30% Stage III n = 33 < 30% Stage IV n = 9 p-value (all stages)
SF-36 PCS 42(12) 42(12) 40(10) 29(12) 0.0059
SF-36 MCS 55(8) 51(11) 52(12) 48(18) 0.19
SGRQ Total 25(20) 32(20) 36(20) 53(23) 0.0005
EQ-5D VAS 73(21) 65(24) 62(21) 37(28) 0.0001
EQ-5D index 0.84(0.15) 0.73(0.23) 0.74(0.25) 0.52(0.26) 0.0008
There was also a statistically significant difference in the SGRQ scores between the severity groups (p = 0.0005). The severity grades affected the level of SGRQ Total as follows: stage I: 25, stage II: 32, stage III: 36 and stage IV: 53 (Table 3, Figure 1).
Figure 1 SGRQ, Total score (adjusted mean values) in GOLD and BTS stages. p-values by test for trend.
The scores for EQ-5D VAS were 73 in stage I and 37 in stage IV (p = 0.0001). EQ-5D index showed the following scores: stage I: 0.84 and stage IV: 0.52 (p = 0.0008) (Table 3).
HRQL in relation to COPD severity according to BTS
The scores for SF-36 (PCS) were 42 in the group labelled stage 0 and 35 in stage III (p = 0.032). The corresponding scores for SF-36 MCS were 55 and 50 in stages 0 and III, respectively (p = 0.29) (Table 4).
Table 4 Health-related quality of life scores, adjusted mean values (± SD) – BTS criteria
Scale FEV1 % predicted
≥ 80% Stage 0 n = 26 79-60% Stage I n = 63 59-40% Stage II n = 47 < 40%Stage III n = 23 p-value (all stages)
SF-36 PCS 42(12) 43(11) 40(13) 35(11) 0.032
SF-36 MCS 55(8) 50(10) 54(11) 50(15) 0.29
SGRQ Total 25(20) 32(20) 34(19) 45(22) 0.0023
EQ-5D VAS 73(21) 68(20) 60(28) 50(25) 0.0007
EQ-5D index 0.84(0.15) 0.74(0.21) 0.72(0.28) 0.63(0.25) 0.0041
The severity grades affected the level of SGRQ Total scores as follows: stage 0: 25, stage I: 32, stage II: 34, and stage III: 45. There was a statistically significant difference in the SGRQ Total scores between the severity groups (p = 0.0023) (Table 4, Figure 1).
The scores for EQ-5D VAS were 73 in stage 0 and 50 in stage III (p = 0.0007). The EQ-5D index scores were 0.84 and 0.63 in stages 0 and III, respectively (p = 0.0041) (Table 4).
Influence of age, gender, smoking status and socio-economic group
The level of SF-36 PCS varied in the age groups from 44 ( < 45 years) to 36 ( > 79 years), with no statistical significance between the age groups. The level of SF-36 MCS was somewhat higher, 56 in the low age group and 51 in the high age group (not significant). The scores for SGRQ varied from 29 ( < 45 years) to 44 ( > 79 years) and they were statistically significant (p = 0.0047) (Figure 2). The scores for EQ-5D VAS varied as follows: 86 ( < 45 years) to 81 ( > 79 years). No statistical difference in EQ-5D VAS and EQ-5D index between the age groups could be seen.
Figure 2 SGRQ, Total score (mean values) in the six age groups. p-values by test for trend.
The gender comparison showed only a statistically significant difference in SF-36 PCS, with scores of 44 for the men and 35 for the women (p = 0.0005).
The mean scores for SGRQ Total were 26, 36 and 31 in the non-smoker, ex-smoker and smoker groups, respectively (not significant).
No significant differences were seen in the other two instruments. Socio-economic group showed no difference for any instrument.
HRQL in relation to FVC % predicted
The four stages of FVC % predicted ( > 95%, 95-81%, 80-66%, < 66%) had an impact on HRQL similar to the stages of FEV1 % predicted outlined from GOLD and BTS. SGRQ total score varied from 26 ( > 95%) to 43 ( < 66%) (p = 0.0002) (Table 5). Using the GOLD stages, the number of patients was unequally distributed and the SGRQ Total scores were 26 ( > 80%, n = 81), 40 (79-50%, n = 68) and 46 (49-30%, n = 10) (p < 0.0001). No patient had a value less than 30% predicted.
Table 5 Health-related quality of life scores, adjusted mean values (± SD) using FVC% predicted normal value
Scale FVC % predicted
> 95% n = 35 95-81% n = 33 80-66% n = 40 < 66% n = 34 p-value (all stages)
SF-36 PCS 44(11) 45(11) 38(12) 35(10) 0.0008
SF-36 MCS 53(9) 52(10) 53(11) 49(14) 0.28
SGRQ Total 26(17) 29(17) 37(22) 43(20) 0.0002
EQ-5D VAS 71(19) 69(24) 63(25) 49(25) 0.0002
EQ-5D index 0.79(0.18) 0.80(0.19) 0.71(0.27) 0.62(0.26) 0.0017
Correlations between the instruments
Table 6 shows the Pearson correlation coefficients between the different instruments and FEV1 % and FVC % predicted. All the questionnaires were correlated with each other. The correlation coefficients between SGRQ and SF-36 PCS/MCS were -0.62 and -0.42, respectively. The lowest correlation was seen between SF-36 MCS and SF-36 PCS (r = 0.22). The correlations between SGRQ and either FEV1 % predicted or FVC % predicted were similar (-0.34 and -0.37, respectively).
Table 6 Pearson's correlation coefficients (r)
SF-36 PCS SF-36 MCS SGRQ Total EQ-5D VAS EQ-5D index FEV1 % predicted FVC % predicted
SF-36 PCS 1.0 -- -- -- -- -- --
SF-36 MCS 0.22 1.0 -- -- -- -- --
SGRQ Total -0.62 -0.42 1.0 -- -- -- --
EQ-5D VAS 0.73 0.49 -0.63 1.0 -- -- --
EQ-5D index 0.64 0.58 -0.61 0.68 1.0 -- --
FEV1 % predicted 0.30 0.10 -0.34 0.38 0.26 1.0 --
FVC % predicted 0.32 0.08 -0.37 0.36 0.22 0.81 1.0
Discussion
The present study confirms that disease severity (based on FEV1) and age influenced HRQL among subjects with COPD. HRQL was strongly related to impaired FEV1 in our study, which is in contrast to some previous studies [21]. The relationship between disease severity using FEV1% predicted and HRQL was made obvious by staging the disease according to the GOLD and BTS guidelines. Once COPD has been diagnosed, neither gender, smoking status nor socio-economic group predicted the level of HRQL.
The relationship between disease severity and HRQL across different chronic conditions, such as ischemic stroke, Parkinson's disease and coronary heart disease, has been examined [22]. It was concluded that in Parkinson's disease the relationship between disease severity and HRQL is linear, whereas in other diseases, such as chronic coronary heart disease, a non-linear relationship was observed. One of the most important implications of a non-linear relationship is that similar changes in disease severity may have a different effect on measured HRQL. Comparing other studies with the present results, some results highlight the fact that physical functioning is one of the most important predictors of HRQL in older subjects. The present results add the clinical value of multidimensional and complex measures of HRQL as previously described [23]. A moderate association between HRQL and COPD severity stage using FEV1 % predicted was seen in another study; however, a large variation in deterioration was observed within each stage of severity, indicating that both clinical and HRQL measures should be considered in the assessment of these patients [24]. In a study by Mahler et al, the decline in lung function over time may predict various components of general HRQL [25].
On the other hand, only a few studies have highlighted a relationship between disease severity and HRQL in COPD. A recent publication supports our findings by showing that GOLD stages of COPD severity differ significantly in SGRQ [26]. However, it was observed that the upper limit of stage IV marks a threshold for dramatic worsening of HRQL, whereas a change from stage 0 to II does not correspond to any meaningful difference in HRQL.
A moderate relationship between the disease stage of COPD and HRQL was found [27]. Our findings confirm these results as patients with COPD have significant decreases in HRQL, and the latter deteriorates in parallel with lung function impairment. An observational study was conducted to explore the relationship between various determinants of disease severity and HRQL [28]. According to its results, lung function and HRQL express several different aspects of disease severity in COPD.
As was found in a study of asthma [29], no gender difference was seen in our study. However, this is not always the case, as women tend to be more sensitive to changes in HRQL [30].
Smoking status did not affect the subjects' HRQL in the present study once COPD had been established. There are various results for the association between smoking status and HRQL. One study showed that COPD patients who continue smoking have a significantly lower HRQL than those who quit smoking [31]. On the other hand, current smoking has been associated with a better HRQL in the study by Wijnhoven et al. [28]. The explanation given was that subjects who do not quit smoking might be those with a less severe stage of disease. One limitation with the present results might be the low number of subjects in the very severe stage group, however, the ANCOVA analysis compensate for the skew distribution.
The correlations between lung function and HRQL have been shown to be weak in a number of studies [21]. In the present study the SGRQ Total score ranged from 23 in GOLD stage I to 56 in GOLD stage IV (according to BTS 23–47). The correlation between FEV1 % predicted and the HRQL measurement varied between -0.34 and 0.10, with the highest correlation (-0.34) between FEV1 % predicted and SGRQ Total score. One reason for the difference in correlation between lung function and HRQL may be the influence of psychosocial variables on the HRQL outcome. The subjects in our study seemed to score their HRQL better compared to other subject groups with similar lung function. One study supported the view that the association between lung function and HRQL can be predicted by perceived self-efficacy for functional activities [32]. That study suggested that both biomedical and psychosocial influences should be taken into account in order to provide optimum assessment and treatment. The correlations in this study were stronger than previous seen and another reason might be that disease severity was considered as a category rather than a continuous variable.
Using FVC % predicted did not add any additional information; however, it supported the view that the level of lung function measured by volume has a similar but lower association with HRQL compared with FEV1 in subjects with COPD.
Conclusion
In conclusion, the results show that the level of health-related quality of life of COPD subjects deteriorates considerably with increasing severity of disease and that the deterioration is linearly related to a decrease in FEV1 % predicted normal values. A higher age also affected the COPD subjects' HRQL, while gender, smoking status and socio-economic group did not, once COPD had been established.
Authors' contributions
Elisabeth Ståhl participated in the study design, evaluation of results and drafted the manuscript
Anne Lindberg, provided with subjects
Sven-Arne Jansson performed the interviews with the subjects
Eva Rönmark provided with subjects
Klas Svensson performed the statistical analysis
Fredrik Andersson participated in the study design
Claes-Göran Löfdahl gave support with interpretation of the results
Bo Lundbäck participated in the study design and responsible for the OLIN studies
All authors read and approved the final manuscript
Acknowledgements
This study was funded by a grant from AstraZeneca.
==== Refs
Vollmer WM Osborne ML Buist AS 20-year trends in the prevalence of asthma and chronic airflow obstruction in an HMO American Journal of Respiratory & Critical Care Medicine 1998 157 1079 1084 9563722
Pauwels RA Buist AS Calverley PMA Jenkins CR Hurd SS Global strategy for the diagnosis, management, and prevention of chronic obstructive pulmonary disease: NHLBI and WHO Global Initiative for Chronic Obstructive Lung Disease (GOLD): executive summary. Respiratory Care 2001 46 798 825 11463370
BTS BTS guidelines for the management of chronic obstructive pulmonary disease. The COPD Guidelines Group of the Standards of Care Committee of the BTS Thorax 1997 52 S1 28
Ray D Abel R Selvaraj KG A 5-yr prospective epidemiological study of chronic obstructive pulmonary disease in rural south India Indian Journal of Medical Research 1995 101 238 244 7672833
Renwick DS Connolly MJ Prevalence and treatment of chronic airways obstruction in adults over the age of 45 Thorax 1996 51 164 168 8711649
Mannino DM Gagnon RC Petty TL Lydick E Obstructive lung disease and low lung function in adults in the United States - Data from the National Health and Nutrition Examination Survey, 1988-1994 Arch Intern Med 2000 160 1683 1689 10847262 10.1001/archinte.160.11.1683
Sobradillo V Miravitlles M Jimenez CA Gabriel R Viejo JL Masa JF Fernandez-Fau L Villasante C [Epidemiological study of chronic obstructive pulmonary disease in Spain (IBERPOC): prevalence of chronic respiratory symptoms and airflow limitation] Arch Bronconeumol 1999 35 159 166 10330536
Lundback B Lindberg A Lindstrom M Ronmark E Jonsson AC Jonsson E Larsson LG Andersson S Sandstrom T Larsson K Not 15 but 50% of smokers develop COPD? Report from the Obstructive Lung Disease in Northern Sweden Studies Respiratory Medicine 2003 97 115 122 12587960 10.1053/rmed.2003.1446
Lindstrom M Jonsson E Larsson K Lundback B Underdiagnosis of chronic obstructive pulmonary disease in Northern Sweden Int J Tuberc Lung Dis 2002 6 76 84 11931405
Tirimanna PRS van Schayck CP den Otter JJ van Weel C van Herwaarden CLA van den Boom G van Grunsven PM van den Bosch WJHM Prevalence of asthma and COPD in general practice in 1992: has it changed since 1977? Br J Gen Pract 1996 46 277 281 8762742
Soriano JB Vestbo J Pride NB Kiri V Maden C Maier WC Survival in COPD patients after regular use of fluticasone propionate and salmeterol in general practice Eur Respir J 2002 20 819 825 12412670 10.1183/09031936.02.00301302
Jones PW Quality of life measurement for patients with diseases of the airways Thorax 1991 46 676 682 1835178
Ware JEJ Sherbourne CD The MOS 36-item short-form health survey (SF-36). I. Conceptual framework and item selection Med Care 1992 30 473 483 1593914
Jones PW Quirk FH Baveystock CM The St George's Respiratory Questionnaire Respiratory Medicine 1991 85 25 31 1759018
Kind P Spilker B Chapter 22. The EuroQoL instrument: an index of health-related quality of life Quality of Life and Pharmacoeconomics in Clinical Trials 1996 Second Philadelphia, Lippincott-Raven Publishers 191 201
Jansson SA Andersson F Borg S Ericsson A Jonsson E Lundback B Costs of COPD in Sweden According to Disease Severity Chest 2002 122 1994 2002 12475838 10.1378/chest.122.6.1994
Jones PW Quirk FH Baveystock CM Littlejohns P A self-complete measure of health status for chronic airflow limitation. The St. George's Respiratory Questionnaire American Review of Respiratory Disease 1992 145 1321 1327 1595997
Jones PW Health status measurement in chronic obstructive pulmonary disease Thorax 2001 56 880 887 11641515 10.1136/thorax.56.11.880
Jones PW Interpreting thresholds for a clinically significant change in health status in asthma and COPD European Respiratory Journal 2002 19 398 404 11936514 10.1183/09031936.02.00063702
Brooks R Brooks R, Rabin R and de Charro F The measurement and valuation of health status using EQ-5D: A European perspective Evidence from the EuroQol BIO MED Research Programme 2003 Dordrecht, Kluwer Academic Publishers 303
Stahl E Wadbo M Bengtsson T Strom K Lofdahl CG Health-related quality of life, symptoms, exercise capacity and lung function during treatment for moderate to severe COPD Journal of Outcomes Research 2001 5 11 24
Ferrucci L Baldasseroni S Bandinelli S De Alfieri W Cartei A Calvani D Baldini A Masotti G Marchionni N Disease severity and health-related quality of life across different chronic conditions J Am Geriatr Soc 2000 48 1490 1495 11083330
Wilson IB Cleary PD Linking clinical variables with health-related quality of life. A conceptual model of patient outcomes JAMA 1995 273 59 65 7996652 10.1001/jama.273.1.59
Ferrer M Anto JM Alonso J Quality of life in COPD patients of different stages of the disease Eur Respir J 1995 8 354s
Mahler DA Tomlinson D Olmstead EM Tosteson AN O'Connor GT Changes in dyspnea, health status, and lung function in chronic airway disease American Journal of Respiratory & Critical Care Medicine 1995 151 61 65 7812573
Antonelli-Incalzi R Imperiale C Bellia V Catalano F Scichilone N Pistelli R Rengo F Do GOLD stages of COPD severity really correspond to differences in health status? European Respiratory Journal 2003 22 444 449 14516133 10.1183/09031936.03.00101203
Ferrer M Alonso J Anto JM Relationship between chronic obstructive pulmonary disease stage and health-related quality of life Cardiology Review 1998 15 42 45 10.1159/000018585
Wijnhoven HAH Kriegsman DMW Hesselink AE Penninx BWJH de Haan M Determinants of different dimensions of disease severity in asthma and COPD - Pulmonary function and health-related quality of life Chest 2001 119 1034 1042 11296166 10.1378/chest.119.4.1034
Malo JL Boulet LP Dewitte JD Cartier A L'Archeveque J Cote J Bedard G Boucher S Champagne F Tessier G Quality of life of subjects with occupational asthma Journal of Allergy & Clinical Immunology 1993 91 1121 1127 8509574 10.1016/0091-6749(93)90313-5
Juniper EF Guyatt GH Epstein RS Ferrie PJ Jaeschke R Hiller TK Evaluation of impairment of health related quality of life in asthma: development of a questionnaire for use in clinical trials Thorax 1992 47 76 83 1549827
Prigatano GP Wright EC Levin D Quality of life and its predictors in patients with mild hypoxemia and chronic obstructive pulmonary disease Arch Intern Med 1984 144 1613 1619 6380440 10.1001/archinte.144.8.1613
Kohler CL Fish L Greene PG The relationship of perceived self-efficacy to quality of life in chronic obstructive pulmonary disease Health Psychology 2002 21 610 614 12433014 10.1037//0278-6133.21.6.610
|
16153294
|
PMC1215504
|
CC BY
|
2021-01-04 16:38:13
|
no
|
Health Qual Life Outcomes. 2005 Sep 9; 3:56
|
utf-8
|
Health Qual Life Outcomes
| 2,005 |
10.1186/1477-7525-3-56
|
oa_comm
|
==== Front
Hum Resour HealthHuman Resources for Health1478-4491BioMed Central London 1478-4491-3-71609297210.1186/1478-4491-3-7ResearchThe potential impact of the next influenza pandemic on a national primary care medical workforce Wilson Nick [email protected] Michael [email protected] Peter [email protected] Osman [email protected] Department of Public Health, Wellington School of Medicine & Health Sciences, Otago University, Wellington, New Zealand2 Public Health Consulting Ltd, Wellington, New Zealand2005 11 8 2005 3 7 7 2 12 2004 11 8 2005 Copyright © 2005 Wilson et al; licensee BioMed Central Ltd.2005Wilson et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Another influenza pandemic is all but inevitable. We estimated its potential impact on the primary care medical workforce in New Zealand, so that planning could mitigate the disruption from the pandemic and similar challenges.
Methods
The model in the "FluAid" software (Centers for Disease Control and Prevention, CDC, Atlanta) was applied to the New Zealand primary care medical workforce (i.e., general practitioners).
Results
At its peak (week 4) the pandemic would lead to 1.2% to 2.7% loss of medical work time, using conservative baseline assumptions. Most workdays (88%) would be lost due to illness, followed by hospitalisation (8%), and then premature death (4%).
Inputs for a "more severe" scenario included greater health effects and time spent caring for sick relatives. For this scenario, 9% of medical workdays would be lost in the peak week, and 3% over a more compressed six-week period of the first pandemic wave. As with the base case, most (64%) of lost workdays would be due to illness, followed by caring for others (31%), hospitalisation (4%), and then premature death (1%).
Conclusion
Preparedness planning for future influenza pandemics must consider the impact on this medical workforce and incorporate strategies to minimise this impact, including infection control measures, well-designed protocols, and improved health sector surge capacity.
==== Body
Background
It is probably only a matter of time before the next influenza pandemic. The only uncertainties are its timing and impact. Effective planning for public health interventions before and during a pandemic is likely to reduce its impact [1]. A pandemic is likely to be extremely disruptive, particularly for the health sector. Not only will there be a surge in demand for health services (preventive as well as curative), but the health workforce is likely to have higher exposure and incidence rates.
We estimated for the impact of pandemic influenza on the primary care medical workforce (i.e., general practitioners) for a single country – New Zealand. The estimates inform planning for the pandemic as well as for other new emerging infectious disease threats, including those from bioterrorism, by providing some estimates for the level of "surge capacity" that must be built into the health sector.
New Zealand has approximately 23 000 health practitioners plus around 30 000 support workers delivering services in the community [2]. Around 40% of its medical practitioners and 23% of its nurses work in primary care settings. The population to general practitioner (GP) ratio varies considerably across different territorial authorities from about 450 to about 2300 [3]. New Zealand's Primary Health Care Strategy [4], released in 2001, places primary care at the centre of the country's health system. It defines a future for primary care where, increasingly, primary care and public health strategies are expected to be coordinated and intermeshed, with the overall objective of improving population health and reducing health inequalities.
The Strategy led to the formation of new non-profit umbrella organisations, called Primary Health Organisations (PHOs) [5]. PHOs are responsible for ensuring that their constituent general practices and community organisations provide comprehensive, continuing and coordinated care to their enrolled populations, including health promotion and prevention programmes. Increasingly, PHOs are held accountable to their funders for a range of population health outcomes. The development of PHOs mirrors, to an extent, the development over the past five years of primary care groups and trusts in the United Kingdom [6].
Methods
The FluAid model
The United States Centers for Disease Control and Prevention (CDC) have developed a relatively simple deterministic model, "FluAid" (on freely available software), for analysing the impact of future influenza pandemics [7]. The output of the model is the number of deaths, hospitalisations, and illnesses requiring medical consultations for a wave of pandemic influenza. These outputs were used to estimate lost workdays in this analysis. The model assumes that no public health interventions (e.g., limitations of movement, vaccine, or antiviral drugs) are used to control disease spread. Specific details on the FluAid software and the various assumptions in the model are detailed on the CDC website [8] and other documents [9,10]. This model has been used in other studies [11-13].
Baseline model assumptions
The default values used in FluAid were used for the mortality rates, the hospitalisation rates and the rates of illness. The default values for the incidence rates of clinical illness were also used (i.e., 15% and 35% for "most likely" results). Working doctors are likely to be in better health than the general population (the "healthy worker effect"), and have fewer risk factors associated with severe sequelae from influenza infection. For example, they are probably far less likely to have chronic respiratory disease, since they have markedly lower smoking rates in New Zealand than the rest of the population [14]. Therefore, the proportion of doctors assumed to be in the "high-risk" category was arbitrarily halved (i.e., from 14.4% for 19–64-year-olds down to 7.2%). This may be overly conservative, but we wished to systematically err towards underestimating the impact of a pandemic on this workforce in the baseline model.
The length of time associated with hospitalisation (average of eight days) and from clinical illness (two days) was based on the United States data in a previously published model [10]. In addition to this, it was assumed that there would be one day of convalescence for clinical illness and three days convalescence after hospitalisation (i.e., before returning to work). To determine the working days lost, these figures were adjusted by the proportion of a typical week that is spent at work (i.e., five out of seven days).
Population data
The latest available national figure for the total number of registered medical practitioners working in primary care was 3074 (i.e., those working four or more hours per week in 2002 and who are classified as working in "general practice" or "primary care") [15]. The average hours worked per week by these doctors is 42 hours, and it was assumed that they would work full time during the pandemic period (unless affected by illness).
Time distribution
The FluAid model does not consider the time frame of the pandemic within an affected region. The length of influenza epidemics is highly variable [16,17]. For the baseline analysis the distribution of cases and a duration of eight weeks was used, based on the results of a stochastically simulated influenza pandemic [18].
"More severe" scenario assumptions
For this scenario the pandemic wave was assumed to last only six weeks and the upper range "maximum" values from the FluAid model were used (for the 35% incidence rate scenario). The proportion of cases in the peak week was raised to 40% (from 32.3%), the upper limit of the days of hospitalisation was used (13 days [10]), and days lost from illness was doubled relative to the baseline model (i.e., to four days). In addition, it was assumed that every doctor would spend an average of 0.5 days during the pandemic wave period caring for sick relatives or household members.
Results
Baseline assumptions result in 584 to 1320 lost workdays for 15% and 35% incidence rates respectively (Table 1). The lost work time was 1.2% to 2.7% of maximal capacity at the time of estimated peak pandemic impact (week 4). An estimated 88% of lost workdays arose from illness not requiring hospitalisation, 8% from hospitalised cases, and 4% from deaths caused by influenza (when using the 35% incidence rate).
Table 1 Predicted impact of pandemic influenza on the population of active and registered primary care medical practitioners based on modelling with FluAid (n = 3074 doctors, 15% and 35% incidence rates)
Week of pandemic in NZ Deaths (No.) Hospital – isations (No.) Illnesses (No.)a Lost workdays from deaths (No.)b Lost workdays from hospitalisations (No.) Lost workdays from illness (No.)a Total lost workdays (No.) Reduction in days worked (%)c
1 0.0 0 5 – 11 0 1 – 2 10 – 23 11 – 25 01 – 0.2%
2 0.1 1 26 – 61 1 5 – 11 56 – 130 62 – 142 0.4 – 0.9%
3 0.2 1 – 3 59 – 137 2 12 – 25 126 – 294 139 – 321 0.9 – 2.1%
4 0.3 2 – 4 77 – 180 3 15 – 33 165 – 386 184 – 422 1.2 – 2.7%
5 0.2 1 – 2 43 – 101 4 9 – 19 93 – 217 106 – 240 0.7 – 1.6%
6 0.1 1 – 1 20 – 47 5 4 – 9 44 – 101 52 – 115 0.3 – 0.7%
7 0.0 0 5 – 12 5 1 – 2 11 – 26 17 – 33 0.1 – 0.2%
8 0.0 0 3 – 7 5 1 7 – 16 12 – 22 0.1%
Totald 1 6 – 13 239 – 556 25 47 – 102 512 – 1192 584 – 1320 0.5 – 1.1%
aFor those with clinical illness that is severe enough to require a medical consultation – but which does not result in hospitalisation.
bThe impact is cumulative for deaths in terms of lost workdays.
cRelative to the full workforce working for five days per week.
dThe figures in the columns do not add up exactly to the totals due to rounding.
The "more severe" model inputs resulted in 3591 lost workdays (Table 2), with a loss of 9.3% of maximal capacity in the peak week and 2.9% over the six-week period. The lost workdays mainly arose from the impact of illness not requiring hospitalisation (64%), then the time spent caring for others (31%), the impact of hospitalisation (4%), and then the impact of premature death (1%).
Table 2 Predicted impact of pandemic influenza on the population of active and registered primary care medical practitioners using "more severe" scenario assumptions and based on modelling with FluAid (n = 3074 doctors)
Week of pandemic in NZ Deaths (No.) Hospital-isations (No.) Illness (No.)a Lost workdays from deaths (No.)a Lost workdays from hospitalisations (No.) Lost workdays from illness (No.)a Lost workdays from caring for others Total lost workdays (No.) Reduction in days worked (%)a
1 0.2 1 64 1 13 184 88 285 1.9%
2 0.4 3 161 3 32 459 220 713 4.6%
3 0.8 6 321 7 64 918 439 1428 9.3%
4 0.4 3 161 9 32 459 220 719 4.7%
5 0.2 1 64 10 13 184 88 294 1.9%
6 0.1 1 32 10 6 92 44 152 1.0%
Total* 2 14 803 39 160 2294 1098 3591 2.9%
aSee footnotes for Table 1.
Discussion
Impact on health and workdays
The model results suggest a substantial impact on general practitioners, even with very conservative assumptions. For the "more severe" scenario a mortality rate of 65 per 100 000 is predicted (albeit for just one pandemic wave). This is much less than the total population rate for the 1918 pandemic in New Zealand of 745 per 100 000 [19], but it is more than United States total population rates for the 1957 Asian flu pandemic (22 per 100 000) and the 1968 Hong Kong flu pandemic (14 per 100 000) [20].
The results suggest that the major contributor to lost workdays will be episodes of uncomplicated illness that do not require hospitalisation. If time spent caring for sick relatives is considered (i.e., as in the "more severe" scenario) then this also made a substantial contribution to the total workdays lost. The impact of lost workdays will be magnified by the increased demand on the medical workforce, as has recently been modelled for primary care consultations and hospitalisations in New Zealand [12], and for critical care services in both New Zealand and Australia [21].
Implications for the health sector
There are several broad strategies to reduce the impact of an influenza pandemic on health care workers. First, infection control strategies aimed at doctors need to be in place. These measures include basic hygiene practices and also mask use may be appropriate (depending on risk [1]). Health authorities and doctors themselves could also stockpile and then use antivirals at the appropriate time. Such stockpiling has already commenced at a national level in New Zealand and various other countries [22]. Recent modelling work indicates that access to enough antivirals could substantially reduce the number of clinical cases and hospitalisations in the population [23].
Second, pandemic planning needs to include specific measures to maintain the functional capacity of health care workers, bearing in mind that the impact of an influenza pandemic is likely to vary between urban and rural areas. While exposure to infection may be less in relatively isolated rural areas, such areas generally have far less "spare" health care capacity, should GPs be incapacitated. General practices and health authorities can consider plans to provide care for the ill dependents of their medical staff so as to reduce absenteeism rates. Through other pandemic planning activities they can also potentially reduce the overall impact of a pandemic and hence demands on their staff. For example, rapid action at the start of the pandemic to cancel elective procedures could enhance workforce capacity. Establishing dedicated primary care assessment centres for patients with suspected influenza could also reduce overall GP workload.
Third, strategies are needed to manage the psychological impact of pandemic influenza on health care workers. Surveys of such workers show that they report a lower willingness to report for duty for infectious diseases epidemics (SARS, smallpox) than for most other forms of catastrophic disasters (environmental disasters, mass casualty incidents) [24]. Experience with SARS also demonstrated the psychological importance of having well-designed policies and protocols in place. Even in situations where health care workers perceive themselves to be at increased risk, they report feeling reassured by simple protective measures based on sound epidemiological principles, when implemented in a timely manner [25]. A review of the foundations for a SARS preparedness and response plan has specifically highlighted the importance of both appropriate staffing and support [26].
Fourth, improving health sector surge capacity now would be desirable as the New Zealand health sector is often running at stretched capacity (e.g., especially emergency departments [27]). Expanding existing services such as the "Healthline" (a free telephone information service to the public staffed by nurses) may also be worthwhile. Similarly, active promotion of key websites with information on managing influenza (e.g., as per the CDC website [28]) could be publicised each winter season. All such measures would benefit the public prior to a pandemic as well as potentially reducing the demands on the medical workforce in the primary care and secondary care settings during a pandemic.
Finally, a greater focus on the primary care nursing workforce would be of benefit. Following the implementation of the Primary Health Care Strategy there has been a rapid shift to capitation funding of general practices, and an attendant increased focus on team-based primary care (principally GPs and practice nurses). This trend raises the possibility of increasing substitution of GP work roles by nurses. This type of substitution has occurred for a decade or more in a range of community-governed non-profit practices and other capitation-funded practices [29,30]. A recent review of the medical workforce in New Zealand also highlights the potential efficiencies from some role shifting from doctors to other health workers [31]. Expanding such a non-medical health workforce, while also vulnerable to the infection during a pandemic, would provide a buffer for the GP workforce in the event of attrition of GP capacity.
Limitations with the modelling
The uncertainties associated with pandemic influenza mean that estimating its future impact is problematic. This model could substantially underestimate the true impact because the new strain may be particularly infectious and/or virulent, and the incidence rate for clinical illness might be higher for doctors given their likely occupational risk [32]. For example, one review of nosocomial outbreaks reported a health care worker incidence rate to be as high as 60% [33]. Furthermore, doctors may be relatively slow to seek care for themselves – especially at the time of a national crisis when their professional obligations are greatest. Other parameters used in the modelling may also have been overly conservative, such as the extent of the healthy worker effect among doctors and the amount of time off work taken to care for sick relatives (which was zero in the baseline model and fairly small at 0.5 days in the "more severe" scenario). There was also no consideration in the model of absenteeism effects from fear of infection (e.g., in the case of particularly virulent strains). Indeed, this absenteeism effect could be more important than actual disease in reducing health sector capacity.
Although we consider that the baseline results are more likely to underestimate than to overestimate the impact of a future influenza pandemic, there are still plausible reasons why they could be overestimates. These include the following:
• various international and national public health interventions (as recommended by WHO [1]) may reduce the impact of pandemic influenza;
• at least for subsequent pandemic waves, an appropriate vaccine may be available;
• antivirals could prevent infection and reduce morbidity amongst the medical workforce and the rest of the population [23];
• improved treatment could lower hospitalisation and mortality rates (relative to the figures used in this model).
Further research
This modelling could be further refined to address some of the limitations detailed above. Clarifying the prevalence of "high-risk" conditions among the medical workforce would be a particularly important refinement along with improving the estimates of time off work to care for relatives (or even absenteeism from fear of infection). Expanding such modelling to other parts of the health sector workforce is also desirable, along with exploring the extent that such research is generalisable to other threats (e.g., from other new emerging infectious threats, including those from bioterrorism).
Summary
This modelling work has a number of limitations and so these results could still substantially overestimate or underestimate the impact of the next influenza pandemic on the primary care medical workforce. Nevertheless, this modelling work highlights the importance of infection control strategies for health care workers, pandemic planning, and improving current health sector surge capacity.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
All authors contributed to the writing and NW, MB and OM also contributed to the design. NW conducted the analyses. All authors read and approved the final manuscript.
Acknowledgements
Preliminary aspects of this work were funded in part by the New Zealand Ministry of Health, but the views presented in this final article do not necessarily represent Ministry of Health policy. We thank the reviewers Alan Hampson and Lance Jennings for their helpful comments on the draft.
==== Refs
World Health Organization WHO consultation on priority public health interventions before and during an influenza pandemic 2004 Geneva, World Health Organization
Health Workforce Advisory Committee The New Zealand Health Workforce, Framing Future Directions, Discussion Document. 2002 Wellington, Health Workforce Advisory Committee
Brabyn L Barnett R Population need and geographical access to general practitioners in rural New Zealand N Z Med J 2004 117 U996 15475979
King A The Primary Health Care Strategy. 2001 Wellington, Ministry of Health
Hefford M Crampton P Foley J Reducing health disparities through primary care reform: the New Zealand experiment Health Policy 2005 72 9 23 15760695 10.1016/j.healthpol.2004.06.005
Weiner J Gillam S Lewis R Organization and financing of British primary care groups and trusts: observations through the prism of US managed care J Health Serv Res Policy 2002 7 43 50 11822261 10.1258/1355819021927665
Meltzer MI Shoemake HA Kownaski M Crosby R FluAid 2.0: A manual to aid state and local-level public health officials plan, prepare and practice for the next influenza pandemic (Beta test version). 2000 Centers for Disease Control and Prevention, U.S. Department of Health and Human Services
CDC (NationalVaccine Program office) FluAid
Meltzer MI Cox NJ Fukuda K The economic impact of pandemic influenza in the United States: priorities for intervention Emerg Infect Dis 1999 5 659 671 10511522
Meltzer MI Cox NJ Fukuda K Modeling the economic impact of pandemic influenza in the United States: Implications for setting priorities for intervention. Background paper. 1999 Atlanta, CDC
Schopflocher DP Russell ML Svenson LW Nguyen TH Mazurenko I Pandemic influenza planning: using the U.S. Centers for Disease Control FluAid Software for small area estimation in the Canadian context Ann Epidemiol 2004 14 73 76 14664783 10.1016/S1047-2797(03)00243-6
Wilson N Mansoor O Baker M Estimating the impact of the next influenza pandemic on population health and health sector capacity in New Zealand N Z Med J 2005 118 U1346 15778744
Wilson N Mansoor O Lush D Kiedrzynski T Modeling the impact of pandemic influenza on Pacific Islands Emerg Infect Dis 2005 11 347 349 15759341
Hay DR Cigarette smoking by New Zealand doctors and nurses: results from the 1996 population census N Z Med J 1998 111 102 104 9577462
Medical Council of New Zealand The New Zealand Medical Workforce in 2002. 2004 Wellington, Medical Council of New Zealand
Flahault A Dias-Ferrao V Chaberty P Esteves K Valleron AJ Lavanchy D FluNet as a tool for global monitoring of influenza on the Web JAMA 1998 280 1330 1332 9794312 10.1001/jama.280.15.1330
Spicer CC The mathematical modelling of influenza epidemics Br Med Bull 1979 35 23 28 760930
Longini IMJ Halloran ME Nizam A Yang Y Containing pandemic influenza with antiviral agents Am J Epidemiol 2004 159 623 633 15033640 10.1093/aje/kwh092
Rice G Black November: The 1918 influenza epidemic in New Zealand. 1988 Wellington, Allen & Unwin
Glezen WP Emerging infections: pandemic influenza Epidemiol Rev 1996 18 64 76 8877331
Anderson TA Hart GK Kainer MA Pandemic influenza-implications for critical care resources in Australia and New Zealand J Crit Care 2003 18 173 180 14595570 10.1016/j.jcrc.2003.08.008
Enserink M Buckheit K Pandemic influenza: Global update Science 2005 309 370 371 16020707 10.1126/science.309.5733.370
Gani R Hughes H Fleming D Griffin T Medlock J Leach S Potential impact of antiviral use during influenza pandemic Emerg Infect Dis 2005 11 (Expedited, in press). [http://www.cdc.gov/ncidod/EID/vol11no09/04 1344.htm]
Qureshi K Gershon RR Sherman MF Straub T Gebbie E McCollum M Erwin MJ Morse SS Health Care Workers' Ability and Willingness to Report to Duty During Catastrophic Disasters J Urban Health 2005
Koh D Lim MK Chia SE Ko SM Qian F Ng V Tan BH Wong KS Chew WM Tang HK Ng W Muttakin Z Emmanuel S Fong NP Koh G Kwa CT Tan KB Fones C Risk perception and impact of Severe Acute Respiratory Syndrome (SARS) on work and personal lives of healthcare workers in Singapore: what can we learn? Med Care 2005 43 676 682 15970782 10.1097/01.mlr.0000167181.36730.cc
Srinivasan A McDonald LC Jernigan D Helfand R Ginsheimer K Jernigan J Chiarello L Chinn R Parashar U Anderson L Cardo D Foundations of the severe acute respiratory syndrome preparedness and response plan for healthcare facilities Infect Control Hosp Epidemiol 2004 25 1020 1025 15636287
Ardagh M Richardson S Emergency department overcrowding - can we fix it? N Z Med J 2004 117 U774 15014563
CDC Influenza (Flu): What to Do If You Get Sick: Overview (Interim Guidance).
Crampton P Davis P Lay-Yee R Primary care teams: New Zealand's experience with community-governed non-profit primary care Health Policy 2005 72 233 243 15802157 10.1016/j.healthpol.2004.08.003
Crampton P Dowell A Woodward A Third sector primary care for vulnerable populations Soc Sci Med 2001 53 1491 1502 11710424 10.1016/S0277-9536(00)00438-X
Health Workforce Advisory Committee Fit for Purpose and for Practice: A Review of the Medical Workforce in New Zealand 2005 Wellington, Health Workforce Advisory Committee
Sepkowitz KA Occupationally acquired infections in health care workers. Part I Ann Intern Med 1996 125 826 834 8928990
Bridges CB Kuehnert MJ Hall CB Transmission of influenza: implications for control in health care settings Clin Infect Dis 2003 37 1094 1101 14523774 10.1086/378292
|
16092972
|
PMC1215505
|
CC BY
|
2021-01-04 16:37:43
|
no
|
Hum Resour Health. 2005 Aug 11; 3:7
|
utf-8
|
Hum Resour Health
| 2,005 |
10.1186/1478-4491-3-7
|
oa_comm
|
==== Front
Int J Health GeogrInternational Journal of Health Geographics1476-072XBioMed Central London 1476-072X-4-181607639110.1186/1476-072X-4-18ResearchA comparison of six analytical disease mapping techniques as applied to West Nile Virus in the coterminous United States Griffith Daniel A [email protected] Ashbel Smith Professor, School of Social Sciences, University of Texas at Dallas, Richardson, TX, USA2005 2 8 2005 4 18 18 9 6 2005 2 8 2005 Copyright © 2005 Griffith; licensee BioMed Central Ltd.2005Griffith; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
West Nile Virus has quickly become a serious problem in the United States (US). Its extremely rapid diffusion throughout the country argues for a better understanding of its geographic dimensions. Both 2003 and 2004 percentages of deaths by numbers of reported human cases, for the 48 coterminous US states, are analyzed with a range of spatial statistical models, seeking to furnish a fuller appreciation of the variety of models available to researchers interested in analytical disease mapping. Comparative results indicate that no single spatial statistical model specification furnishes a preferred description of these data, although normal approximations appear to furnish some questionable implications. Findings also suggest several possible future research topics.
==== Body
Background
West Nile Virus (WNV [1,2]), first isolated in the West Nile District of Uganda in 1937, is a flavivirus transmitted by a mosquito vector, with a general incubation period of 2–14 days following a bite by an infected mosquito, and is closely related to the St. Louis encephalitis virus that also is found in the United States (US). WNV can infect humans, birds, mosquitoes, horses and some other mammals, with mosquitoes becoming infected after feeding on the blood of birds that carry the virus (this virus enters and circulates in a mosquito's bloodstream for a few days before it settles in the insect's salivary glands); of particular concern is that the adult WNV-carrying Culex species of mosquito is able to survive through winters. WNV primarily results in bird mortality, and human and equine encephalitis. In temperate latitudes, West Nile encephalitis cases occur primarily in the late summer or early fall; WNV tends to be carried by less than 1 out of every 100 mosquitoes residing in geographic regions in which it actively circulates. WNV, with its first detected US case on Long Island in 1999, has swiftly diffused across the continental US (Figure 1a and Table 1) as well as elsewhere in the Western Hemisphere. This virus enjoyed a surprisingly rapid rate of diffusion, spreading from the New York City area to nearby localities contagiously, as well as leaping across space in a hierarchical fashion through, for example, bird migration routes (see ). Although presently a person has a low risk of contracting WNV, many people infected with this virus – more than 16,000 have tested positive to date – tend to experience mild (e.g., flu-like symptoms such as fever, headache, body ache and skin rash) or no symptoms (i.e., never realizing that they have been exposed to WNV), with less than 1% of those infected developing serious illness (e.g., high fever, severe headache, stiff neck, disorientation, tremors, muscle weakness, paralysis and coma), and even fewer dying from the virus (roughly 650 to date). People at higher risk of developing potentially serious conditions include the elderly (age 50 and older) and those with lowered immune systems, and some categories of outdoor workers. Meanwhile, because 1999 signifies the beginning of a logistic growth curve for WNV cases in the US, during the very early years of the new millennium, public health officials mostly were concerned about whether or not this virus had diffused to their localities. Presence or absence was the geographic quantification of choice. But once WNV appears in an area, infected people begin to die, with human deaths and cases becoming a geographic quantification of choice. In response to this public health problem, by 2002 CDC was releasing numbers of cases and of deaths, for US states; today these data also are being released for US counties. (See .) Data aggregated by state that were extracted from this web site and used in the paper for analysis purposes appear in additional file 1a (Cases of and deaths attributed to WNV); the associated geographic connectivity matrix (i.e., matrix C) appears in additional file 1b (US state geographic connectivity matrix).
Figure 1 WNV in the US. Left: (a) diffusion of the virus: years of presence is directly proportional to darkness of the gray scale. Right: (b) 2002 presence of WNV by state, through August.
Table 1 Expansion of WNV across the coterminous US.
Year # states reporting cases # human cases # human deaths
1999 4 62 7
2000 4 21 2
2001 10 50 5
2002 44 4,156 284
2003 45 9,862 264
2004 40 2,539 100
Because of the inherently geographic nature of WNV containment and eradication efforts, the objective of this paper is to summarize some initial spatial statistical analyses of the georeferenced virus in its US context using disease map modeling. In equilibrium, such features of WNV as percentage deaths should contain positive spatial autocorrelation (i.e., similar percentages of deaths should cluster together on a map) because: common local weather patterns tend to spatially cluster and partially govern mosquito population dynamics; bird populations tend to have distinct migratory routes and locational preferences; socio-economic/demographic attributes of the human population cluster geographically; and, by their very administrative unit jurisdictional nature, mosquito control programs tend to be geographically concentrated. One outcome of this work is a set of research hypotheses that should be analyzed as WNV becomes more strongly entrenched in the US geographic landscape.
Results and discussion: a comparison of disease map modeling results
Griffith [3, pp. 78–79, 114–116] reports an analysis of the presence/absence of WNV by state for January through August, 2002 (see Figure 1b). A normal probability model approximation cannot be implemented here because the response variable is binary. The join count statistics indicate the presence of strong, positive spatial autocorrelation. A visual inspection of the map furnished by the Centers for Disease Control (CDC) indicates that only the 11 most western states in the coterminous US were absent of WNV. A Markov chain Monte Carlo (MCMC) estimated auto-logistic model reveals the presence of strong, positive spatial autocorrelation. The spatial filter logistic model (a generalized linear model specification) furnishes a third description of these data, one involving eigenvector E1. Estimation attempts of both spatial filter and proper conditional autoregressive (PCAR) specifications with Bayesian inference using Gibbs sampling (BUGS) repeatedly failed because of encountered phase transitions [4]. One interesting feature of the generalized linear spatial filter model is its suggestion that California (a leap across geographic space), New Mexico, Montana and Washington should have had WNV present. The end-of-the-year CDC map indeed reveals WNV presence in these states. A second interesting feature of this spatial filter model is that it out-performs both the pseudo-likelihood and the MCMC estimated auto-logistic models.
An analysis of the numbers of deaths attributable to WNV standardized by the corresponding number of reported cases (i.e., percentage data) for 2003 (overall 2.7%) and 2004 (overall 3.6%), by state (see Figure 2), reveals the presence of weak-to-negligible spatial autocorrelation (see Tables 2 and 3). An indicator variable has been included to differentiate the states without reported cases from the remaining states. Noteworthy findings here include: (1) the intercept term is approximately -3.1, regardless of model specification or year; (2) the normal probability model approximations highlight the 0-case states (3 in 2003, and 8 in 2004) as being statistically significant, whereas the remaining four specifications find them to be consistent with the other data; (3) the auto-binomial model fails to furnish a reasonable description of these 2004 data; (4) the spatial filter models indicate that both positive and negative spatial autocorrelation effects are present in these data – these countervailing geographic trends could be why the global spatial autocorrelation indices appear to be nonsignificant; and, (5) the PCAR hierarchical generalized linear model (HGLM) uncovers statistically nonsignificant spatial autocorrelation, whereas the spatial filter HGLM uncovers statistically significant spatial autocorrelation – perhaps again reflecting the mixture of positive and negative spatial autocorrelation components.
Figure 2 The geographic distribution of percentages of WNV cases resulting in death. Left (a): 2003. Right (b): 2004.
Table 2 Parameter estimation results for observed 2003 WNV death percentages.
Gaussian SAR Gaussian spatial filter binomial spatial filter
Statistic estimate se estimate se estimate se
Spatial autocorrelation 0.08 0.101 1 (spatial filter MC = 0.243) 0.144 1 (spatial filter MC = 0.359) 0.090
intercept -3.16 0.164 -3.15 0.105 -3.11 0.078
I0 2.32 0.629 2.32 0.419 -18.21 25149
approximate dfs 45 39 40
Residual P(S-W) 0.029 0.212 0.522
Residual MC 0.0433 -0.0431 (z = -0.2) -0.0927
Residual GR 1.0295 0.9720 0.9782
deviance 6.1357 4.2542 1.0198
pseudo-R2 0.425 0.643 0.292
predicted-observed regression slope 1.05 0.76 0.69
HGLM
proper CAR spatial filter
Statistic estimate se estimate se
spatial autocorrelation 0.09 0.104 0.999 (spatial filter MC = 0.359) 0.094
intercept - 3.25 0.187 -3.12 0.072
I0 -80.11 57 -78.86 60
approximate dfs 26 40
residual P(S-W) 0.018 0.001
residual MC 0.0119 0.0350
residual GR 0.7141 0.7494
deviance 0.4089 1.1660
pseudo-R2 0.697 0.442
predicted-observed regression slope 1.82 0.97
Table 3 Parameter estimation results for observed 2004 WNV death percentages.
Gaussian SAR Gaussian spatial filter binomial spatial filter
Statistic estimate se estimate se estimate se
Spatial autocorrelation 0.01 0.101 1 (spatial filter MC = 0.030) 0.135 6.07 4.701
intercept -3.30 0.172 -3.31 0.124 -3.48 0.182
I0 2.63 0.491 2.62 0.323 -20.01 41223
approximate dfs 45 39 45
Residual P(S-W) 0.249 0.256 < 0.0001
Residual MC -0.0224 -0.0753 (zMC = -0.3) -0.073
Residual GR 1.1361 1.0788 1.088
deviance 0.9001 0.4686 1.0346
pseudo-R2 0.494 0.754 0.07
predicted-observed regression slope 1.31 0.72 0.87
HGLM
proper CAR spatial filter
Statistic estimate se estimate se
spatial autocorrelation -0.08 0.148 0.999 (spatial filter MC = 0.762) 0.226
intercept -3.28 0.122 -2.99 0.122
I0 -81.63 62 -80.94 60
approximate dfs 45 44
residual P(S-W) <0.0001 0.192
residual MC 0.0596 0.0130
residual GR 0.9417 0.9270
deviance 0.9715 0.4992
pseudo-R2 0.328 0.411
predicted-observed regression slope 30.60c 0.99
Even after employing a variable transformation to stabilize variance, the Gaussian (i.e., normal probability) approximations continue to display marked overdispersion (i.e., excessive variability) in 2003. Residuals for these models contain only trace amounts of spatial autocorrelation, and the models themselves account for nearly half of the variability in percentage deaths from WNV. But these pseudo-R2 values are inflated by the inclusion of the indicator variable denoting states with no cases. And, the simultaneous autoregressive (SAR) model predicted and observed values match well for 2003 but are underestimated by roughly 30% for 2004, whereas the spatial filter predicted values tend to be overestimated by roughly 40%, on average.
According to the 2004 data (for which pseudo- and maximum likelihood estimates are equivalent here), not only does the auto-binomial model furnish a poor description of these data, but it also is plagued by overestimation of the predicted values. Meanwhile, the spatial filter generalized linear model has residuals that contain only trace negative spatial autocorrelation, furnishes a respectable description of these data, lacks overdispersion, and provides predicted and observed values that are well matched for 2003 and overestimated for 2004.
The HGLMs have effective degrees of freedom associated with them, in order to account for parameter estimates as well as random effects estimates. Although the PCAR specification furnishes the best overall description of these data for 2003, it does so by consuming considerably more degrees of freedom. This HGLM tends to exhibit considerable underdispersion (i.e., insufficient variability), a signature of negative spatial autocorrelation. Spatial autocorrelation may well remain in the 2003 HGLM residuals, with the Moran coefficient (MC) and the Geary ratio (GR) values being inconsistent in their implications. The PCAR model also dramatically underestimates predicted probabilities. In contrast, the spatial filter HGLM produces predicted values that match well with their corresponding observed values, furnishing respectable descriptions of both data sets.
Overall, no single model specification is superior to all others. The discrepancies between specification results emphasize differences in modeling assumptions, error structure, and detailed treatment of latent spatial autocorrelation, resulting in different nuances and idiosyncrasies of the data being highlighted.
Conclusion
Several implications can be drawn from the research summarized in this paper. Foremost is that while the initial spread of WNV across the US had a prominent geographic dimension, such a dimension has not fully materialized yet for the percentage of deaths from detected cases. In addition, the spatial filter model specification is appealing because of its simplicity, and because it is able to sense the presence of a mixture of positive and negative spatial autocorrelation components latent in the geographic distribution of WNV deaths. This model feature is not shared by the more conventional specifications. These findings suggest the following research hypothesis: spatial autocorrelation contained in a disease map has its negative components fade, and its positive components strengthen, through time. This hypothesis can be tested by evaluating the regression coefficients of spatial filter eigenvectors as annual WNV data become available. The following is a second research hypothesis suggested by Tables 2 and 3: half of the variability in percentages of WNV deaths may be accounted for with yet-to-be-determined socio-economic/demographic attribute variables. Again, this hypothesis can be tested by computing the pseudo-R2 model values as annual WNV data become available. It also can be assessed by uncovering the unknown covariates. Although this is a descriptive feature of WNV here, this characterization is expected to typify expansion diffusion following an initial invasion of some territory [5]. And, the following is a third hypothesis: states without reported WNV cases are from the same statistical population as states with reported cases. This hypothesis can be tested by evaluating the zero-case indicator variable generalized linear model regression coefficient as annual WNV data become available.
Because WNV risk factors include being elderly, the size of local bird populations, and exposure to certain species of mosquitoes, geographic distributions of these groups ultimately should impact upon the geographic distribution of WNV deaths. The elderly tend to be clustered nationally, with this effect perhaps being controllable by age-standardization of case counts. The use of door and window screens, mosquito repellant, and adult, larvae and breeding site mosquito control programs tend to have socio-economic/demographic dimensions with spatial expressions. All of these factors tend to impact upon contagion diffusion, inducing positive spatial autocorrelation. Meanwhile, factors such as migratory bird routes result in leaps across geographic space (i.e., hierarchical diffusion), which initially introduce a negative spatial autocorrelation dimension into the geographic distribution of WNV deaths (i.e., a location with cases being surrounded by neighboring locations with no cases). As these diffusion paths become reinforced through cyclical repetition over time, accompanied by repeated annual waves of local contagion diffusion, a more uniform geographic distribution of WNV reservoirs should materialize, causing the negative spatial autocorrelation dimension to fade away. In the end, WNV should be characterized by positive spatial autocorrelation reflecting annual weather map patterns that promote, and the effectiveness of public health programs that attempt to minimize, the size of mosquito populations. Given the figures reported in Table 1, these patterns should contain a fair degree of variability. A similar argument pertains to horse deaths from WNV, too.
The principal covariate in the spatial statistical model specifications estimated in this paper is spatial autocorrelation. Based on the pseudo-R2 values reported in Tables 2 and 3, this covariate tends to account for about 50% of the variability in the percentages of WNV deaths. This finding is a clue that socio-economic/demographic attributes – presumably associated with the use of door and window screens, mosquito repellant, and adult, larvae and breeding site mosquito control programs – should be explored in order to identify those that statistically describe this variability. Of course, part of this covariation may disappear by using age-standardized figures. Part may disappear over time as health professionals increasingly gain experience in detecting and treating WNV cases. And, part may disappear as the hierarchical diffusion component of WNV expansion disappears.
Now that WNV has appeared at one time or another in each of the coterminous US states, years in which no cases are detected in a given state have become a feasible outcome within the same statistical population. These zero-cases can naturally occur when weather patterns suppress mosquito populations, or can result from temporary effectiveness of human uses of screens and mosquito repellants, as well as governmental adult, larvae and breeding site mosquito control programs. They also can result from cyclical biological processes in the various animal populations involved in WNV transmission. When inspecting Table 1, one should not be surprised that the more severe outbreak year of 2003 is accompanied by only 3 states having zero cases, while the less severe outbreak year of 2004 is accompanied by 8 states having zero cases. Of note is that this scenario is somewhat incomplete, since the spatial diffusion of WNV was reaching geographic saturation during these two years. Nevertheless, more severe outbreak years should tend to be accompanied by a more widespread geographic distribution of the virus.
Another general finding is that the normal probability approximation model specification tends to overemphasize the statistical significance of those states with no cases. In part this result may link to the use of translation parameters that convert the odds ratio for these states to roughly 0.5. Meanwhile, one implementation data adjustment made for the other model specifications involved recoding 0s to 1s for the cases variable, which does not alter the corresponding percentage (i.e., the number of deaths remained 0). This data adjustment will be dispensed with once WNV becomes more prevalent across the US, and hence states will not be without cases, but most likely will have to be retained for initial county-level geographic resolution studies. And, because of the size of its autoregressive parameter, the auto-binomial model proves to be of less interest for model comparison purposes.
Overall general findings suggest several rules of thumb that should help guide an analyst in his/her disease map modeling efforts. Foremost, switching between model specifications should yield similar intercept values; if markedly different values are obtained, an analyst should be suspicious and ascertain why. Second, non-normal data are best described with non-normal probability models; an analyst always should be aware of nontrivial specification error. Third, a Gaussian approximation spatial filter model can be used to more quickly explore whether both positive and negative spatial autocorrelation components underpin a disease map; a spatial filter model specification enables a detailed understanding of latent spatial autocorrelation. And, fourth, a spatial filter can be used to more quickly explore spatial structuring of random effects in a Bayesian analysis involving a large n; a spatial filter model specification dramatically reduces the numerical intensity of MCMC computations.
Finally, popular individual observation diagnostic statistics may be evaluated in terms of their covariations with spatial autocorrelation by regressing them on the candidate spatial filter eigenvectors. DFBETA diagnostic statistics, one for each attribute variable, specify the standardized differences in regression estimates for assessing the effects of individual observations on the estimated regression parameters in a fitted model. And, the HI diagnostic statistic specifies the diagonal element of the hat matrix for detecting extreme points in a regressor attribute variable matrix. For the 2003 data and the reduced-form logistic regression model, eigenvectors E1, E7, E12 and E26 account for roughly 40% of the variance in the intercept DFBETA statistic, as do eigenvectors E3, E7, E18, E27 and E29 for the indicator variable DFBETA statistic. Eigenvectors E1 and E27 account for roughly 20% of the variance in the HI statistic. These three sets of eigenvectors overlap with those selected for the affiliated generalized linear spatial filter model only in E1 and E18. Meanwhile, for the 2004 data and the reduced-form logistic regression model, eigenvectors E1, E3, E10, E15 and E22 account for roughly 45% of the variance in the intercept DFBETA statistic, as do eigenvectors E1, E5, E14, E19, E21, E27 and E29 for the indicator variable DFBETA statistic. Eigenvectors E4, E4, E7 and E24 account for roughly 30% of the variance in the HI statistic. These three sets of eigenvectors overlap with those selected for the affiliated generalized linear spatial filter model only in E1 and E15. Future research should include scrutinizing the full battery of such diagnostics in this manner, in order to better articulate relationships between local and global information. Subsequent research also needs to address the problem of states with very few cases, whose predicted values will tend to have excessive variability, and states with very large numbers of cases, whose predicted values will tend to be significant by default.
Together these findings suggest several implications about the diffusion of WNV, whose initial spread across the coterminous US at the state level of geographic resolution now is complete. In years to come, diffusion across states will be in terms of waves of re-infection. But infill contagious diffusion still is occurring at the county and finer resolutions. Findings summarized in this paper imply that for this level of diffusion, the geographic distribution of county populations should introduce both positive and negative spatial autocorrelation components into the resulting map patterns. Furthermore, one prominent socio-economic covariate should be the difference between rural and urban locations. Others should be covariates of the willingness of local populations to accept and fund aggressive mosquito control programs, as well as individually adopting measures to prevent mosquito bites.
Methods: spatial statistical modeling approaches
An analyst can choose from a variety of analytical spatial statistical tools to study a disease map. The first of these to be developed historically is the spatial autoregressive model, which is based upon a normal probability model, and hence requires disease map data to conform to a bell-shaped curve; often this requirement necessitates the use of a Box-Cox type of power transformation. Recent quantitative geography methodological developments have supplemented this approach with the spatial filter model specification [6]. One advantage of a spatial filter approach is that it also enables use of a generalized linear model specification [7,8], which for disease mapping purposes is based upon the binomial, Poisson, or negative binomial probability models (depending upon whether a disease map is expressed in terms of a binary, a percentage or a count variable). Recent MCMC methodology also enables the use of the binomial, Poisson, or negative binomial probability models with a spatial autoregressive specification [9-11]. In addition, MCMC has made Bayesian analysis implementable and hence more accessible, enabling researchers to estimate both conditional autoregressive [12,13] and spatial filter HGLM specifications. Because these modeling approaches involve different data assumptions, especially in terms of error, researchers interested in analytical disease mapping need a fuller appreciation of the variety of models they can employ in their analyses; six specifications are treated here, namely three conventional and their three spatial filter counterparts. Furnishing the basis for this appreciation is one of the purposes of this paper.
A sizeable part of spatial statistics is concerned with accounting for observation correlational effects arising from the geographic configuration of data. Quantitatively characterizing this configuration commonly is achieved in one of two ways: (1) establishing a geocoded coordinate for each areal unit, and then computing inter-point distances; and, (2) establishing a surface partitioning, and then constructing an n-by-n binary matrix C, whose cell entries are cij = 1 if areal units i and j share a boundary (employing analogies with chess: if it is non-zero in length, then the linkages are referred to as the rook's case; if it is both zero and non-zero in length, then the linkages are referred to as the queen's case), and cij = 0 otherwise. This queen's adjacency formulation is employed here.
The traditional spatial autoregressive model
Spatial statistics addresses the issue of observational correlation amongst georeferenced observations, which is known as spatial autocorrelation; this type of correlation can be indexed with a MC or a GR. This autocorrelation often is positive in nature, with most phenomena exhibiting a moderate tendency for their similar values to cluster in geographic space. Occasionally, the tendency is for dissimilar values to cluster in geographic space, representing negative spatial autocorrelation.
An autoregressive model specification accounts for spatial autocorrelation by including a variable on the right-hand side of an equation that is a function of the neighboring Y values; in other words, the disease map variable, Y, appears on both sides of an equation. When coupled with regression and the normal probability model, this specification results in a covariation term characterizing spatial autocorrelation in one of two popular ways. Denoting the autoregressive parameter that captures spatial autocorrelation with ρ, a conditional autoregressive (CAR) covariance specification involves the matrix (I - ρ C), where I is an n-by-n identity matrix. Because matrix C is raised to the power 1 (i.e., only adjacent neighbors are involved in the autoregressive function), this expression is considered a first-order specification, with the autoregressive term being CY. Because I is the identity matrix, individual areal unit variance is conditionally constant.
An important matrix can be constructed from C1, which is the vector of number of neighbors. If the inverse of the elements of C1 are inserted into the diagonal of a diagonal matrix, say D-1, then W = D-1C becomes a stochastic matrix (i.e., each of its row sums equals 1). One appealing feature of this matrix is that the autoregressive term becomes WY, which renders averages, rather than sums, of neighboring values. Because a covariance matrix must be symmetric, a matrix W specification can be used with a CAR model only by making the individual areal unit variance nonconstant: (I - ρ D-1C)D-1 = (D-1 - ρ D-1CD-1). One appealing feature of this version is that it restricts positive values of the autoregressive parameter to the more intuitively interpretable range of 0 ≤ ≤ 1.
The SAR model (see additional files 2b: Data input and preparation for the SAR model estimation with SAS; 2b: SAR model estimation with SAS) furnishes an alternative specification that frequently is written in terms of matrix W. As such, its spatial covariance is a function of the matrix (I - ρ CD-1)(I - ρD-1C) = (I - ρ WT)(I - ρ W), where T denotes matrix transpose. The resulting matrix is symmetric, is considered a second-order specification because it includes the product of two spatial structure matrices (i.e., WTW) – adjacent areal units as well as those having a single intervening unit are involved in the autoregressive function – and also restricts positive values of the autoregressive parameter to the more intuitively interpretable range of 0 ≤ ≤ 1.
For the percentage of deaths associated with diagnosed WNV cases, the log-odds ratio requires the following Box-Cox type of transformations in order to conform to a bell-shaped curve:
Non-zero translation parameters primarily are due to the presence of 0 cases, and partially due to the presence of 0 deaths for some non-zero cases. WNV deaths are a rare event, with the probability of death being highly skewed. These constants shift the locations of the empirical probabilities within the interval [0, 1] in order to have the data better conform to a normal frequency distribution. Symmetricizing effects of these transformations are portrayed in Figure 3. Zero-zero states, whose ratio becomes roughly 0.5, can be differentiated from the remaining states with an indictor variable designating them as potentially coming from a different statistical population. Because these transformations to normality should stabilize variance, arguably estimation can be done without a weighting scheme [e.g., following the variability of a binomial probability, the appropriate weighting scheme would involve division by ]; comparisons with and without the use of weights revealed no real differences in results.
Figure 3 Boxplots of raw and transformed percentage deaths from diagnosed WNV cases.
The spatial filter model
Spatial filtering involves regressing a disease map variable on a set of synthetic variates representing distinct map patterns that accounts for spatial autocorrelation; each of the three preceding spatial statistical model specifications can be replaced with a spatial filter model specification. Griffith [3] develops one form of spatial filtering whose synthetic variates are the set of n eigenvectors extracted from matrix (I - 11T/n)C(I - 11T/n), the matrix appearing in the numerator of the MC index of spatial autocorrelation, where 1 is an n-by-1 vector of ones (see additional file 3: Minitab 14.13 code for computing spatial filter eigenvectors). This procedure is similar to executing a principal components analysis in which the covariance matrix is given by (I - 11T/n)C(I - 11T/n). But rather than using the resulting eigenvectors to construct linear combinations of attribute variables, the eigenvectors themselves (instead of principal components scores) are the desired synthetic variates, each containing n elements, one for each areal unit. The extracted eigenvector relates to the mean response, and the remaining (n-1) extracted eigenvectors relate to distinct map patterns characterizing latent spatial autocorrelation – whose MCs are given by standardizing their corresponding eigenvalues [14] – that can materialize with matrix C. Furthermore, for a given geographic landscape surface partitioning, the eigenvectors represent a fixed effect in that matrix (I - 11T/n)C(I - 11T/n) does not, and hence they do not, change from one attribute variable to another.
Because this eigenfunction decomposition yields n eigenvectors, a disease map analyst needs to restrict attention to only those eigenvectors describing substantive positive/negative spatial autocorrelation (e.g., MC > 0.25 – a value that tends to relate to about 5% of the variance in Y being attributable to redundant information arising from latent spatial autocorrelation, given a particular areal unit neighborhood configuration), reducing the candidate set to a more manageable number for describing a given disease map. Supervised stepwise selection from this set of eigenvectors is a useful and effective approach to identifying the subset of eigenvectors that best describes latent spatial autocorrelation in a particular disease map. This procedure begins with only the intercept included in a regression specification. Next, at each step an eigenvector is considered for addition to the model specification. For the stepwise linear Gaussian model, commonly the eigenvector having the largest partial correlation with variable Y is selected, but only if its corresponding F-ratio achieves or surpasses a prespecified level of significance; this is the criterion used to establish statistical importance of an eigenvector. Meanwhile, in stepwise generalized linear modeling regression, the eigenvector that produces the greatest reduction in the log-likelihood function chi-square test statistic is selected, but only if it produces at least a prespecified minimum reduction; as before, this is the criterion used to establish statistical importance of an eigenvector. In each statistical procedure, at each step all eigenvectors previously entered into a spatial filter equation are reassessed, with the possibility of removal of vectors added at an earlier step. The forward/backward stepwise procedure terminates automatically when some prespecified threshold values (respectively for F-ratios and chi-square statistics) are encountered for entry and removal of all candidate eigenvectors. The ultimate inclusion criterion is determined by the MC value of the residuals, which should indicate an absence of spatial autocorrelation. Satisfying this MC condition sometimes requires supervised backward elimination of marginally selected eigenvectors because their inclusion has forced the residual MC value to decrease too far below 0. This final stopping criterion for the linear Gaussian model is relatively easy to implement because MC distributional theory is known for linear regression residuals; a corresponding stopping rule for generalized linear modeling regression is far more difficult to implement because of a lack of such distributional theory.
Spatial filters for both 2003 and 2004 maps of WNV deaths are a mixture of eigenvectors representing positive as well as negative spatial autocorrelation. The 2003 Gaussian analysis (see additional file 4: Data input, preparation, and estimation of the Gaussian spatial filter model with SAS) identifies the following as prominent eigenvectors, with the nature of their respective spatial autocorrelation denoted in parentheses:
E1 (+), E2 (+), E6 (-), E10 (+), E15 (-), E21 (-) and E25 (-).
Meanwhile, the 2003 generalized linear model logistic regression analysis (see additional file 5: Data input, preparation, and estimation of the logistic spatial filter model with SAS) identifies the following as prominent eigenvectors:
E1 (+), E6 (-), E11 (-), E15 (-), E18 (+) and E25 (-).
Eigenvectors E1, E6, E15 and E25 are common to these two sets. The 2004 Gaussian analysis identifies the following as prominent eigenvectors:
E1 (+), E3 (+), E15 (-), E24 (-), E26 (-), E27 (+) and E28 (-).
Meanwhile, the 2004 generalized linear model logistic regression analysis identifies the following as prominent eigenvectors:
E1 (+), E3 (+) and E15 (-).
Eigenvectors E1, E3 and E15 are common to these two sets. Maps of eigenvectors E1 and E15, common to all four of these sets, appear in Figure 4; these represent prominent map patterns underlying the geographic distribution of WNV death percentages.
Figure 4 Maps of selected eigenvectors. Vector element values are directly proportional to darkness of the gray scale. Left (a): marked positive spatial autocorrelation (MC = 1.06). Right (b): strong negative spatial autocorrelation (MC = -0.50).
The MCMC technique
MCMC provides a mechanism for taking spatially dependent samples from probability distributions in situations where the usual sampling is difficult, if not impossible. Many auto-models fall into this category, particularly because the normalizing constants for their joint or posterior probability distributions are either too difficult to calculate or analytically intractable. MCMC is used to simulate from some n-by-1 joint probability distribution p known only up to a constant factor, c. That is, p = cq, where q is known but c is unknown and an intractable mathematical expression [see [15], pp. 428 and 431, for mathematical statements of c for auto-Poisson and auto-binomial models]. MCMC sampling begins with conditional (marginal) probability distributions, and parameter estimates that are obtained using pseudo-likelihood estimation (i.e., an autoregressive term is estimated with a conventional regression procedure; see additional file 6: Data input, preparation, and pseudo-likelihood estimation of the autologistic model with SAS). This involves estimating covariate coefficients (β) and ρ as though observations are independent. MCMC outputs a sample of values for each parameter drawn from the joint posterior probability distribution.
Gibbs sampling is a MCMC scheme for simulation from p where the Markov chain transition matrix (M) is defined by the n conditional probability distributions of p. It is a stochastic process that returns a different result with each execution, a method for generating a joint empirical distribution of several variables from a set of modeled conditional distributions for each variable when the structure of data is too complex to implement mathematical formulae or directly simulate. It is a recipe for producing a Markov chain that yields simulated data that have the correct unconditional model properties, given the conditional distributions of those variables under study [16]. The principal idea behind it is to convert a multivariate problem into a sequence of univariate problems, which then are iteratively solved to produce a Markov chain. The following Gibbs sampling algorithm description [17] is for a selected auto-model and uses the pseudo-likelihood parameter estimates of the parameters β and ρ:
Step 1: Initialize a map (k = 0) by taking i = 1, ..., n independent random samples {yi,k=0} from a chosen probability model (e.g., a binomial model).
Step 2: Obtain new values (initially k = 1) yi,k by sequentially moving from one location (i) to another (j) on the initial map and randomly sampling from the appropriate auto-model (e.g., the auto-binomial model) using the pseudo-likelihood parameter estimates. Site selection for this process of obtaining {yi,k=1} from {yi,k=0} can follow random permutations of location sequences. The value at each location is updated immediately after it is computed.
Step 3: Obtain new values (initially k = 2) yi,k+1 by sequentially moving from one location to another on the kth map, again randomly sampling from the appropriate auto-model, and immediately updating the value at each location.
Step 4: Repeat step 3 for iterations k = 3, 4, ..., until convergence of the sufficient statistics of the parameters of interest occurs.
The final output then can be used to compute maximum likelihood estimates of parameters.
Once a Markov chain transition matrix is constructed, a sample of (correlated) drawings from a target distribution can be obtained. This is done by simulating the Markov chain a large number of times (say, 525,000), removing a "burn-in" set of iterations (say, 25,000), weeding it (select only every, say, 100th result), and recording its sufficient statistics. Convergence needs to be monitored, and hence the sufficient statistics need to be recorded. This recording should be done after the completion of each iteration. A suitable burn-in period is needed in order to generate M, the limiting Markov chain transition probability matrix, and hence before collecting statistics, and because samples are serially correlated, the chain needs to be weeded.
One difficulty with estimation of an auto-binomial model, which is supported by MCMC techniques, is that the relationship between its intercept and autoregressive parameter is established by the global percentage of deaths for a disease map. The ideal situation occurs when ρ = -α/2 [[18], p. 3]; this point is illustrated here in Table 3. Because this parameter estimate is inconsistent with the other model specifications, this model is not treated in great detail here. Of note is that the WinBUGS software package cannot be used to estimate this model because the spatial lag variable must be recomputed at each MCMC step.
The impact of spatial autocorrelation on the frequency distribution of a binomial random variable is illustrated in Figure 5. For a 100-by-100 lattice, pseudo-random binomial counts were generated for N = 1,000 and p = 0.05; each count then was divided by 10,000. As conventional statistical theory states, the distribution of these values is approximately normally distributed with mean 0.05 and variance (0.05)(1-0.05)/100. The principal impact of spatial autocorrelation is to reduce the more central frequencies and increase the tail frequencies. As spatial statistical theory states, the mean remains 0.05, but the variance increases (here by a factor of 1.55), regardless of whether only positive or only negative spatial autocorrelation is contained in the probabilities. The effect of mixing an equal amount of positive and negative spatial autocorrelation is to have some of the autocorrelation effects cancel out; now the variance is inflated by a factor of 1.35, with the equal mixture better preserving the original kurtosis.
Figure 5 Frequency distributions for simulated binomial probabilities (N = 10,000, p = 0.05, 100-by-100 lattice). Top left (a): random simulated data. Top right (b): simulated data with marked positive spatial autocorrelation embedded (MC = 1.01). Bottom left (c): simulated data with marked negative spatial autocorrelation embedded (MC = -1.01). Bottom right (d): simulated data with an equal mixture of marked positive and marked negative spatial autocorrelation embedded.
The spatial Bayesian HGLM
Meanwhile, Bayesian random-effects HGLM specifications also can be used to deal with non-normal data. One appealing feature of this approach is that spatial autocorrelation in a non-normal georeferenced random variable can be captured without having to derive an explicit multivariate generalization of its distributional form. WinBUGS [12] and GeoBUGS [13] support implementation of this model for disease mapping purposes. When analyzing maps of disease, the random effects can be spatially structured and/or unstructured. The CAR model is one way spatial structuring is included; the spatial filter is another way. If the autoregressive parameter ρ is estimated, the specification is called a PCAR model. If the autoregressive parameter value is set equal to 1, then a second unstructured random effect term is included, and the specification is called an improper CAR (ICAR) model. Parameters are estimated with MCMC techniques.
Because Bayesian statistical analysis is involved, prior distributions need to be posited for each varying quantity: the response variable, each variable coefficient, the spatial autoregressive parameter, the error variance, and the random error term. The response variable prior distribution includes the model statement. The random error term may be posited as a PCAR or an ICAR specification. If either a ICAR or a spatial filter term is included, then a prior distribution for an unstructured random effect must be included. For a disease map, the response variable prior distribution frequently will be Poisson (for counts) or binomial (for presence/absence or percentages). The accompanying CAR/PCAR frequently involves an auto-normal specification. The spatial autoregressive parameter prior distribution usually is uniform. The normal distribution furnishes a feasible prior distribution for covariate coefficients, including the spatial filter. And, the error variance prior distribution often is the gamma distribution.
The following binomial HGLM involving the PCAR model was estimated using the US WNV data and GeoBUGS (see additional file 7: 2003 data input and estimation of the PCAR logistic spatial filter model with GeoBUGS):
where α is the intercept, I0 is the binary 0–1 indicator variable for 0-case states, β1 denotes its regression coefficient, and νi denotes unobserved US state-specific random effects. The prior distributions attached to this log-mean response equation are:
Di ~ binomial(pi, Ci),
α ~ normal(0, 0.0001),
β1 ~ normal(0, 0.0001),
νi ~ auto-normal , with a conditional autoregressive model specification corresponding to a proper multivariate Gaussian distribution with a full-rank covariance matrix (I - ρ C),
~ gamma(0.5, 0.0005), and
ρ ~ uniform(1/λ48, 1/λ1), where λ48 and λ1 respectively are the smallest and largest eigenvalues of matrix C,
where ~ denotes "distributed as," Di and Ci respectively denote the number of deaths and the number of cases in state i, and σε is the standard deviation of the random effects term.
The spatial filter version (see additional file 8: 2003 data input and estimation of the logistic spatial filter model with WinBUGS) removes the CAR specification, and adds a spatial filter term together with its regression coefficient β2 ~ normal(0, 0.0001); the random effects now are independent normal(0, ).
Supplementary Material
Additional File 1
Cases of and deaths attributed to WNV; US state geographic connectivity matrix. Tabulated state-by-attribute data (a), and tabulated state-by-state binary geographic connectivity matrix data (b).
Click here for file
Additional File 2
Data input and preparation for the SAR model estimation with SAS; SAR model estimation with SAS. SAS computer code, in which the input data file paths and file names may need to be changed (a), for estimating a simultaneous spatial autoregressive model (b).
Click here for file
Additional File 3
Minitab 14.13 code for computing spatial filter eigenvectors. Minitab computer code, in which the input data file paths and file names may need to be changed; the number of areal units is stored in K1, the geographic connectivity matrix is stored in M1, and the spatial filter eigenvectors are stored in M2.
Click here for file
Additional File 4
Data input, preparation, and estimation of the Gaussian spatial filter model with SAS. SAS computer code, in which the input data file paths and file names may need to be changed, for estimating a linear regression spatial filter model.
Click here for file
Additional File 5
Data input, preparation, and estimation of the logistic spatial filter model with SAS. SAS computer code, in which the input data file paths and file names may need to be changed, for estimating a generalized linear (logistic) regression spatial filter model.
Click here for file
Additional File 6
Data input, preparation, and pseudo-likelihood estimation of the auto-logistic model with SAS. SAS computer code, in which the input data file paths and file names may need to be changed, for estimating a generalized linear auto-logistic regression model.
Click here for file
Additional File 7
2003 data input and estimation of the PCAR logistic spatial filter model with GeoBUGS.
Click here for file
Additional File 8
2003 data input and estimation of the logistic spatial filter model with WinBUGS.
Click here for file
Acknowledgements
This research was supported by NIH grant #P20RR20770, and presented to the P20 Exploratory Centers Program Workshop, University of Miami, May 22–25, 2005.
==== Refs
Emerging Infectious Diseases special issue 2001 7
Zielinski-Gutiérrez E O'Connell J West Nile Virus & other mosquito-borne infections The Health Care of Homeless Persons: A Manual of Communicable Diseases and Common Problems in Shelters and on the Streets Part I: Mosquito-borne Infections 2004 The Boston Health Care for the Homeless Program 181 186
Griffith D Spatial Autocorrelation and Spatial Filtering: Gaining Understanding through Theory and Scientific Visualization 2003 Berlin, Springer-Verlag
Guyon X Random Fields on a Network : Modeling, Statistics, and Applications 1995 Berlin, Springer-Verlag
Gould P Spatial Diffusion 1969 Washington, DC, Association of American Geographers
Griffith D A linear regression solution to the spatial autocorrelation problem J of Geographical Systems 2000 2 141 156
Griffith D A spatial filtering specification for the auto-Poisson model Statistics & Probability Letters 2002 58 245 251 10.1016/S0167-7152(02)00099-8
Griffith D A spatial filtering specification for the auto-logistic model Environment & Planning A 2004 36 1791 1811 10.1068/a36247
Gilks R Richardson S Spiegelhalter J (eds) Markov Chain Monte Carlo in Practice 1996 London, Chapman & Hall
Huffer F Wu H Markov chain Monte Carlo for autologistic regression models with application to the distribution of plant species Biometrics 1998 54 509 524
Kaiser M Cressie N Modeling Poisson variables with positive spatial dependence Statistics & Probability Letters 1997 35 423 432 10.1016/S0167-7152(97)00041-2
Cowles M Review of WinBUGS 1.4 The American Statistician 2004 58 330 336 10.1198/000313004X8515
Thomas A Best N Lunn D Arnold R Spiegelhalter D GeoBUGS User Manual, version 12 2004 on 3/4/2005
Tiefelsdorf M Boots B The exact distribution of Moran's I Environment and Planning A 1995 27 985 999
Cressie N Statistics for spatial data 1993 New York, Wiley
Robert C Casella G Monte Carlo Statistical Methods 1999 Berlin, Springer-Verlag
Casella G George E Explaining the Gibbs sampler The American Statistician 1992 46 67 174
Graham J Monte Carlo Markov chain likelihood ratio test and Wald test for binary spatial lattice data Mimeographed paper Raleigh 1994 NC: North Carolina State University, Department of Statistics
|
16076391
|
PMC1215506
|
CC BY
|
2021-01-04 16:39:06
|
no
|
Int J Health Geogr. 2005 Aug 2; 4:18
|
utf-8
|
Int J Health Geogr
| 2,005 |
10.1186/1476-072X-4-18
|
oa_comm
|
==== Front
Int Semin Surg OncolInternational seminars in surgical oncology : ISSO1477-7800BioMed Central London 1477-7800-2-161613139910.1186/1477-7800-2-16ResearchOur local experience with the surgical treatment of ampullary cancer Botsios Dimitrios [email protected] Emmanouil [email protected] Ioannis [email protected] Kostas [email protected] Emmanouil [email protected] Stavros [email protected] Evangelos [email protected] Dimitrios [email protected] Ioannis [email protected] 4th Surgical Department, Aristotle University of Thessaloniki, 'G. Papanikolaou' General Regional Hospital, Exohi, Thessaloniki 57010, Greece2005 30 8 2005 2 16 16 25 5 2005 30 8 2005 Copyright © 2005 Botsios et al; licensee BioMed Central Ltd.2005Botsios et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
The aim of this study is to report the outcome after surgical treatment of 32 patients with ampullary cancers from 1990 to 1999.
Methods
Twenty-one of them underwent pancreaticoduodenectomy and 9 local excision of the ampullary lesion. The remaining 2 patients underwent palliative surgery.
Results
When the final histological diagnosis was compared with the preoperative histological finding on biopsy, accurate diagnosis was preoperatively established in 24 patients. The hospital morbidity was 18.8% as 9 complications occurred in 6 patients. Following local excision of the ampullary cancer, the survival rate at 3 and 5 years was 77.7% and 33.3% respectively. Among the patients that underwent Whipple's procedure, the 3-year survival rate was 76.2% and the 5-year survival rate 62%.
Conclusion
In this series, local resection was a safe option in patients with significant co-morbidity or small ampullary tumors less than 2 cm in size, and was associated with satisfactory long-term survival rates.
==== Body
Background
Carcinoma of the ampulla of Vater is an entity distinct from neoplasms arising from the periampullary area. Ampullary cancer accounts for some 7% of peripancreatic tumors. It is less aggressive and has a better survival than carcinomas arising from the pancreas or common bile duct [1,2]. The 5-year survival rate reported after Whipple's resection for ampullary cancer varies from 22% to 55% [3,4], whereas the relevant rate for pancreatic carcinoma is reported not to be higher than 22% to 26% [5]. Even though the outcome of patients after resection of ampullary cancer is more favorable, almost half of them will die from tumor recurrence [6]. On the other hand, pancreaticoduodenectomy has been reported to result in morbidity of 43% and mortality of 11% [7]. This fact has led to interest in local resection of ampullary tumors as described by Halstead in 1899 [8]. After local resection, it has been reported that the mortality rate reaches 7.1% and the 5-year survival rate 35% [9].
We report the outcome after surgical treatment of patients with cancer of the ampulla of Vater, with Whipple's procedure, local resection or palliative by-pass surgery.
Methods
From 1990 to 1999, 205 patients diagnosed with periampullary neoplasms were treated in our Department. Our study population consisted of 32 of these patients that proved to have carcinoma of the ampulla of Vater, and underwent surgical treatment. They comprised 18 (56%) men and 14 (44%) women of mean age 68.5 (range: 42–77) years.
Only lesions confined to the ampulla or clearly invading the surrounding tissues from the ampulla were designated as ampullary carcinomas. Eight patients who were found histologically to have ampullary adenomas were excluded from the study as our aim was to focus on the outcome of patients with malignant lesions of the ampulla of Vater.
All the patients underwent standard diagnostic imaging investigations, which included conventional ultrasonography (US), computed tomography (CT) scan and endoscopic retrograde cholangiopancreaticography (ERCP) with multiple biopsies taken from the ampulla of Vater. Thirty (30) of the 32 patients underwent potentially radical surgery: 21 underwent pancreaticoduodenectomy (Whipple's procedure) and 9 local excision of the ampullary lesion. The remaining 2 patients underwent palliative surgery and particularly choledochojejunostomy Roux en Y. No patient received adjuvant chemotherapy or radiotherapy in the post-operative period. During the same time period, 9 patients with non-resectable ampullary cancer were not fit for operation and were treated with ERCP and stent placement. These patients that did not undergo surgical treatment were also excluded from the study. None of them received adjuvant chemotherapy or radiotherapy in the post-operative period, as well.
Pancreaticoduodenectomy was the first choice as the type of surgical treatment. Local resection was the preferable treatment when the ampullary lesion was less than 2 cm in diameter, the pre-operative biopsy showed a pT1 cancer or adenoma of the ampulla of Vater and/or the patient's concomitant medical illness or age contraindicated a major operation such as Whipple's procedure. Finally, palliative by-pass surgery was reserved for patients whose tumor was larger than 5 cm according to the findings of the preoperative imaging investigations, obstruction of the portal vein, invasion of the superior mesenteric artery and/or vein, metastatic liver disease or distant metastases.
All patients underwent regular follow-up examinations post-operatively on a 3-month basis for the first year, on a 6-month basis for the following 4 years, and annually thereafter. Follow-up included clinical examination, blood tests (CA 19-9, serum bilirubin, alkaline phosphatase), abdominal ultrasound and chest radiography.
The statistical methods employed were Fisher's exact test for comparison of proportions. Differences among groups with respect to continuous variables were tested using the Kruskal-Wallis test, whereas pairwise differences were compared by the Mann-Whitney test, at a Bonferroni-adjusted significance level. The survival curves among groups were compared with the Log-rank test and they were presented graphically with the Kaplan-Meier plots. Analyses were conducted in SPSS 11.0 (SPSS, Inc., Chicago, IL). All reported p-values are two-tailed.
Results
All tumors were adenocarcinomas originating in the ampulla of Vater. The median size of the tumors as measured by the pathologist was 2.9 (0.8–5.1) cm. The tumor grade was recorded as well differentiated in 10, moderately differentiated in 13 and poorly differentiated in 9 patients including the two patients with unresectable tumors, according to pre-operative biopsy. Nodal involvement was microscopically found in 14 (46.6%) patients. The tumor was classified as pT1 in 8 patients and pT2-T4 in the remaining 22 patients with resectable tumors. All patients with pT1 tumors underwent local excision, while all patients with pT2-T4 tumors underwent pancreaticoduodenectomy, except one who underwent local excision.
When the final histological diagnosis after surgical treatment was compared with the preoperative histological finding on biopsy, accurate diagnosis was preoperatively established in 24 patients. In the remaining patients, the diagnosis was established intra-operatively by examination of frozen section in 3, and post-operatively by examination of the specimen in 3 patients.
The in-hospital, as well as the 30-day overall mortality rate, was 0% as no death occurred among the patients of the study. The overall hospital morbidity was 18.8%, as 9 complications occurred in 6 patients post-operatively. All the complications occurred in patients who underwent pancreaticoduodenectomy resulting in a morbidity of 28.6% in this group of patients. Although this morbidity was substantially higher compared to the group undergoing local excision and palliative surgery, this difference was not statistically significant (p = 0.129, Table 1).
Table 1 Comparison of different parameters among the different surgical approach groups.
Whipple's procedure (n = 21) Local excision (n = 9) Palliative surgery (n = 2) p-value
Hospital morbidity
n (%) 6 (28.6) 0 (0) 0 (0) 0.129a
Hospital stay, days
Median (range) 16 (9–32) 8 (7–10) 7.5 (7–8) <0.001b
a Fisher's exact test
b Kruskal-Wallis test for the overall comparison (The significant pairwise comparisons with Bonferroni adjusted p-value were Whipple's procedure vs. the other two groups for both the hospital stay and operation time [p < 0.05]).
In particular, one patient presented with moderate leakage of the choledochojejunostomy, which was treated conservatively. Pancreatic fistulas were seen in two patients, which were also treated conservatively. One patient developed a sub-hepatic abscess on the 20th post-operative day, but ultrasound-guided drainage was adequate treatment in this case. Surgical intervention was required in a patient who presented with intra-abdominal bleeding due to septic erosion of gastro-duodenal artery and rupture of the pancreatojejunostomy and choledochojejunostomy on the 13th post-operative day. Re-operation in this case included ligation of gastro-duodenal artery, reconstruction of choledochojejunostomy and drainage of the ruptured pancreatojejunostomy. The post-operative period in this case was uneventful while the pancreatic fistula subsided within two months. Finally, two patients experienced moderate wound infection and drainage of the wound abscess was adequate treatment in these cases. On the other hand, the post-operative course was uneventful in patients who underwent either local excision or palliative operation.
Seventeen (17) patients have died during follow-up, and all but three died because of recurrence. Following palliative operation (choledochojejunostomy Roux en Y), both patients died 7 and 13 months after operation. Following local excision of the ampullary cancer in 9 patients, the survival rate at 3 years was 77.7% (7 patients) and 33.3% (3 patients) at 5 years. In the patients that achieved 5-year survival after local excision the resection was R0, the tumor was graded as pT1, N0, M0 and, moreover, it was well-differentiated.
Among 21 patients that underwent Whipple's procedure, the 3-year survival rate was 76.2% (16 patients) and the 5-year survival rate was 62% (13 patients). One patient is alive 10 years after the operation and considered cancer free. As shown in Figure 1, the median survival after Whipple's procedure was significantly higher than the median survival after local resection (65.2 vs. 38.6 months, p < 0.001).
Figure 1 Time to death for patients who underwent pancreaticoduodenectomy, local excision and palliative surgery. Log-rank test p < 0.001.
The overall median hospital stay was 13.5 (range: 7–32) days. The median in-patient stay after Whipple's procedure was significantly greater than the median in-patient stay after local excision and palliative operation (16 vs. 8 and 7.5 days respectively, p < 0.001, Table 1).
Discussion
Periampullary pancreatic neoplasms rank as the fifth leading cause of cancer death behind lung, breast, colorectal and prostate cancer, causing more than 30,000 deaths per year in the United States [10]. Ampullary carcinomas account for 6%–20% of all periampullary tumors and for 10.2%–36% of all operable pancreatoduodenal neoplasms [2,3,11]. They account for about 0.2% of all gastrointestinal malignancies, with a median incidence of about 57 cases per million of population per year [12,13]. In our study, cancer of the ampulla of Vater represented 15.6% of all periampullary carcinomas treated during the same time period in our Department.
A major problem in dealing with ampullary tumors is to differentiate between an adenoma and a carcinoma. Modern imaging studies such as US, CT scan, magnetic resonance imaging (MRI) and magnetic resonance cholangiopancreaticography (MRCP), have significantly improved the diagnostic accuracy in the pre-operative period. Moreover, endoscopic diagnostic techniques, such as duodenoscopy with multiple biopsies and endoscopic ultrasonography are extremely helpful tools in order to determine the nature and the extend of the tumor pre-operatively. However, despite the fact that endoscopic ultrasonography may be helpful in detecting the invasion of the tumor into the surrounding tissues and the occurrence of lymph node enlargement, it does not allow differentiation between an adenoma and a pT1 carcinoma [14]. Furthermore, duodenoscopy with multiple biopsies has been reported to miss the diagnosis of malignancy in 12% to 40% of cases [15]. Because a negative biopsy does not exclude the presence of an invasive carcinoma within the adenoma, complete excision has been recommended in all cases [16]. The above-mentioned recommendation was also the policy in our study, as we performed either Whipple's resection or local excision in all cases with a resectable ampullary tumor. In our series, duodenoscopy with biopsies failed to reveal malignancy in 20% of the patients. Endoscopic ultrasonography was not included in our diagnostic procedure due to lack of experience and relevant equipment.
Options for excision of ampullary carcinomas include local excision and pancreaticoduodenectomy. Halsted performed the first local excision for ampullary carcinoma in 1899 [8]. Although there are many case reports and a few series on the treatment of ampullary neoplasms by local ampullary resection, the criteria used to decide when local excision is suitable for certain patients are controversial, and not well addressed. For small lesions, thought to be benign pre-operatively according to endoscopic appearance and biopsy, ampullary resection is generally well accepted [17]. Bottger et al [18] stated that the indications for local excision should be that the tumor is completely removed (R0), limited to the ampulla of Vater (pT1), not poorly differentiated and with no venous/lymphatic infiltration in patients with ASA grade IV, regardless of their age. Similarly, Beger et al [19] reported that local resection is indicated in cases with pT1, N0, M0 cancer of the ampulla of Vater, excluding patients with tumor poorly differentiated. Moreover, he stated that if intra- or even post-operative histological findings show a cancer more advanced than pT1, node-positive, or poorly differentiated tumor, the procedure should be extended to a pancreaticoduodenectomy and that in patients with pT1 cancer, local excision should be always combined with a local lymph node dissection. Regarding age as an indication for local excision, we can only analyze the reported evidence about age as a contraindication for Whipple's procedure. In many authors opinion, there is rarely any justification for performing major resection in a patient over 75 years of age due to high morbidity and mortality, as well as short survival [20]. However, Kairaluoma et al [21] suggested that age is not a limiting factor for pancreatic resection and it can be performed with acceptable survival rates even in patients over 70 years of age. In general, local ampullary resection is accompanied by significantly less morbidity and mortality than pancreaticoduodenectomy [19,22]. In our study, mortality as well as morbidity was 0% among 9 patients that underwent local excision for ampullary cancer. Our criteria for performing local excision were tumor size less than 2 cm at duodenoscopy, the patient's poor medical fitness, age > 75 years and pre-operative biopsy showing a pT1 cancer or adenoma of the ampulla of Vater. In a patient less than 75 years of age, the final histological examination showed a cancer more advanced than pT1, but we did not proceed to pancreaticoduodenectomy as his comorbidity contraindicated a major operation. Finally, during local excision we performed local lymph node dissection only if the nodes were enlarged.
Pancreaticoduodenectomy is undoubtedly the procedure of choice in the management of ampullary cancer in patients who are medically fit, and this fact has been reinforced by the declining mortality after the procedure during the past decades. This was also the policy in our study; pancreaticoduodenectomy with or without pylorus preservation was the first choice of surgical treatment in patients with tumors more advanced than pT1, unless it was contraindicated by the patient's comorbidity or age (>75 years), at which point a local excision was performed. Hospital mortality after pancreaticoduodenectomy is less than 5% in recently published series, whereas the morbidity remains high and varies from 25 to 65% [6,18-20,23]. The related mortality in our series was 0% but morbidity was as high as 28.6%. The pattern of failure after surgical resection for ampullary cancer is a crucial point to assess the efficacy of adjuvant therapy for patients receiving resection for cure. No patient in our study received adjuvant chemotherapy and this is in accordance with recent published studies showing no benefit of such therapy after curative or non-curative resection of ampullary cancer [24].
Survival after local resection it is difficult to estimate given the small number of patients, but it has been reported to be 40% to 50% at 5 years [25-29]. This figure is comparable to 37.5% to 62.7% 5-year survival rate reported in the much larger pancreaticoduodenectomy series [6,18-20,23]. However, it is important to mention that the pancreaticoduodenectomy series included high-risk lesions (T3 or T4, involved nodes, poor differentiation) which are excluded from the local excision series, and this might be an explanation for the comparable survival rates. In our study which included a small number of patients, following local excision of the ampullary cancer, the survival rate at 3 and 5 years was 77.7% and 33.3% respectively. Among the patients that underwent Whipple's procedure, the 3-year survival rate was 76.2% and the 5-year survival rate 62%.
Conclusion
In this series, local resection was a safe option in patients with significant co-morbidity or small ampullary tumors (<2 cm), and was associated with satisfactory long-term survival rates.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
DB: conceived the study, performed part of the operations and coordinated the preparation of the manuscript for submission.
EMZ: did the literature search and drafted the manuscript.
IL: participated in the design of the study, collection of data and preparation of manuscript for publication.
KT: Performed part of the operations, participated in the design of the study and preparation of manuscript for publication.
EC: Performed part of the operations, participated in the design of the study and preparation of manuscript for publication.
SK: participated in the design of the study and preparation of manuscript for publication.
EVZ: participated in the literature search and preparation of manuscript for publication and helped to draft the manuscript.
DB: has given final approval of the version to be published.
ID: has given final approval of the version to be published.
==== Refs
Conlon KC Dougherty E Klimstra DS Coit DG Turnbull AD Brennan MF The value of minimal access surgery in the staging of patients with potentially resectable peripancreatic malignancy Ann Surg 1996 223 134 140 8597506 10.1097/00000658-199602000-00004
Willett CG Warshaw AL Convery K Compton CC Patterns of failure after pancreatoduodenectomy for ampullary carcinoma Surg Gynecol Obstet 1993 173 33 38 8093983
Jones BA Langer B Taylor BR Girotti M Periampullary tumors: which one should be resected? Am J Surg 1985 149 46 52 3966641
Trede M Schwall G Saeger HD Survival after pancreatoduodenectomy 118 curative resections without an operative mortality Ann Surg 1990 211 447 458 2322039
Ishikawa O Ohigashi H Imaoka S Sasaki Y Iwanaga T Matayoshi Y Inoue T Is the long-term survival rate improved by preoperative irradiation prior to Whipple's procedure for adenocarcinoma of the pancreatic head? Arch Surg 1994 129 1075 1080 7944938
Allema JH Reinders ME van Gulik TM van Leeuwen DJ Verbeek PC de Wit LT Gouma DJ Results of pancreaticoduodenectomy for ampullary carcinoma and analysis of prognostic factors for survival Surgery 1995 117 247 253 7878528
Bakkevold KE Kambestad B Morbidity and mortality after radical and palliative pancreatic cancer surgery Ann Surg 1993 217 356 368 7682052
Haelsted WS Contributions to the surgery of the bile passages, especially of the common bile duct Boston Med Surg 1899 141 645 652
Bottger T Potratz D Wellek S Ochmann M Stockle M Klupp J Jungiger T Stellenwert der bildanalytischen DNS-Cytometrie beim carcinom der ampulla vateri Chirurg 1993 64 476 485 8359060
Yeo C Sohn T Cameron J Hruban R Lillemoe K Pitt H Periampullary adenocarcinoma: analysis of 5-year survivors Ann Surg 1998 227 821 831 9637545 10.1097/00000658-199806000-00005
Yamaguchi K Enjoji M Carcinoma of the ampulla of Vater: a clinicopathologic study and pathologic staging of 109 cases of carcinoma and 5 cases of adenoma Cancer 1987 59 506 515 3791159
Roder JD Schneider PM Stein HJ Siewert JR Number of lymph node metastases is significantly correlated with survival in patients with radically resected carcinoma of the ampulla of Vater Br J Surg 1995 82 1693 1696 8548244
Neoptolemos JP Talbot IC CarrLocke DL Treatment and outcome in 52 concecutive cases of ampullary carcinoma Br J Surg 1987 74 957 961 3664230
Nattermann C Dancygier H Endosonographie bei tumoren des pancreas und der gallenwege Leber Magen Darm 1993 5 13 20 8445972
Classen M Endoscopic approach to papillary stenosis Endoscopy 1981 13 154 156 7250083
Declore R Connor CS Thomas JH Friesen SR Hermreck AS Significance of tumor spread in adenocarcinoma of the ampulla of Vater Am J Surg 1989 158 593 597 2589595 10.1016/0002-9610(89)90201-8
Clary B Tyler D Dematos P Gottfried M Pappas T Local ampullary resection with careful intra-operative frozen section evaluation for presumed benign ampullary neoplasms Surgery 2000 127 628 633 10840357 10.1067/msy.2000.106532
Bottger TC Boddin J Heintz A Junginger T Clinicopathological study for the assessment of resection for ampullary carcinoma World J Surg 1997 21 379 383 9143568
Beger HG Treitschke F Gansauge F Harada N Hiki N Mattfeldt T Tumor of the ampulla of Vater Arch Surg 1999 134 526 532 10323425 10.1001/archsurg.134.5.526
Su CH Shyr YM Lui WY P'eng FK Factors affecting morbidity, mortality and survival after pancreaticoduodenectomy for carcinoma of the ampulla of Vater Hepato-Gastroenterology 1999 46 1973 1979 10430380
Kairaluoma MI Kiviniemi H Stahlberg M Pancreatic resection for carcinoma of the pancreas and the periampullary region in patients over 70 years of age Br J Surg 1987 74 116 118 3815026
Rattner D Fernandez-del Castillo C Brugge W Warshaw A Defining the criteria for local resection of ampullary neoplasms Arch Surg 1996 131 366 371 8615720
Sperti C Pasquali C Piccoli A Sernagiotto C Pedrazzoli S Radical resection for ampullary carcinoma: long-term results Br J Surg 1994 81 668 671 7913860
Takada T Amano H Yasuda H Is postoperative adjuvant chemotherapy useful for gallbladder carcinoma? A phase III multicentre prospective randomized controlled trial in patients with resected pancreaticobiliary carcinoma Cancer 2002 95 1685 1695 12365016 10.1002/cncr.10831
Klein P Reingruber B Kastl S Dworak O Hohenberger W Is local excision of pT1 ampullary carcinomas justified? Eur J Surg Oncol 1996 22 366 371 8783654 10.1016/S0748-7983(96)90286-3
Knox R Kingston R Carcinoma of the ampulla of Vater Br J Surg 1986 73 72 73 3947884
Newman R Pittam M Local excision in the treatment of carcinoma of the ampulla of Vater J R Coll Surg Edinb 1982 27 154 157 7108832
Robertson J Imrie C Local excision of ampullary carcinoma Acta Chir Scand 1986 152 537 539 3788399
Tarazi R Hermann R Vogt D Results of surgical treatment of periampullary tumors: a thirty-five-year experience Surgery 1986 100 716 722 3764694
|
16131399
|
PMC1215507
|
CC BY
|
2021-01-04 16:38:37
|
no
|
Int Semin Surg Oncol. 2005 Aug 30; 2:16
|
utf-8
|
Int Semin Surg Oncol
| 2,005 |
10.1186/1477-7800-2-16
|
oa_comm
|
==== Front
Int Semin Surg OncolInternational seminars in surgical oncology : ISSO1477-7800BioMed Central London 1477-7800-2-161613139910.1186/1477-7800-2-16ResearchOur local experience with the surgical treatment of ampullary cancer Botsios Dimitrios [email protected] Emmanouil [email protected] Ioannis [email protected] Kostas [email protected] Emmanouil [email protected] Stavros [email protected] Evangelos [email protected] Dimitrios [email protected] Ioannis [email protected] 4th Surgical Department, Aristotle University of Thessaloniki, 'G. Papanikolaou' General Regional Hospital, Exohi, Thessaloniki 57010, Greece2005 30 8 2005 2 16 16 25 5 2005 30 8 2005 Copyright © 2005 Botsios et al; licensee BioMed Central Ltd.2005Botsios et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
The aim of this study is to report the outcome after surgical treatment of 32 patients with ampullary cancers from 1990 to 1999.
Methods
Twenty-one of them underwent pancreaticoduodenectomy and 9 local excision of the ampullary lesion. The remaining 2 patients underwent palliative surgery.
Results
When the final histological diagnosis was compared with the preoperative histological finding on biopsy, accurate diagnosis was preoperatively established in 24 patients. The hospital morbidity was 18.8% as 9 complications occurred in 6 patients. Following local excision of the ampullary cancer, the survival rate at 3 and 5 years was 77.7% and 33.3% respectively. Among the patients that underwent Whipple's procedure, the 3-year survival rate was 76.2% and the 5-year survival rate 62%.
Conclusion
In this series, local resection was a safe option in patients with significant co-morbidity or small ampullary tumors less than 2 cm in size, and was associated with satisfactory long-term survival rates.
==== Body
Background
Carcinoma of the ampulla of Vater is an entity distinct from neoplasms arising from the periampullary area. Ampullary cancer accounts for some 7% of peripancreatic tumors. It is less aggressive and has a better survival than carcinomas arising from the pancreas or common bile duct [1,2]. The 5-year survival rate reported after Whipple's resection for ampullary cancer varies from 22% to 55% [3,4], whereas the relevant rate for pancreatic carcinoma is reported not to be higher than 22% to 26% [5]. Even though the outcome of patients after resection of ampullary cancer is more favorable, almost half of them will die from tumor recurrence [6]. On the other hand, pancreaticoduodenectomy has been reported to result in morbidity of 43% and mortality of 11% [7]. This fact has led to interest in local resection of ampullary tumors as described by Halstead in 1899 [8]. After local resection, it has been reported that the mortality rate reaches 7.1% and the 5-year survival rate 35% [9].
We report the outcome after surgical treatment of patients with cancer of the ampulla of Vater, with Whipple's procedure, local resection or palliative by-pass surgery.
Methods
From 1990 to 1999, 205 patients diagnosed with periampullary neoplasms were treated in our Department. Our study population consisted of 32 of these patients that proved to have carcinoma of the ampulla of Vater, and underwent surgical treatment. They comprised 18 (56%) men and 14 (44%) women of mean age 68.5 (range: 42–77) years.
Only lesions confined to the ampulla or clearly invading the surrounding tissues from the ampulla were designated as ampullary carcinomas. Eight patients who were found histologically to have ampullary adenomas were excluded from the study as our aim was to focus on the outcome of patients with malignant lesions of the ampulla of Vater.
All the patients underwent standard diagnostic imaging investigations, which included conventional ultrasonography (US), computed tomography (CT) scan and endoscopic retrograde cholangiopancreaticography (ERCP) with multiple biopsies taken from the ampulla of Vater. Thirty (30) of the 32 patients underwent potentially radical surgery: 21 underwent pancreaticoduodenectomy (Whipple's procedure) and 9 local excision of the ampullary lesion. The remaining 2 patients underwent palliative surgery and particularly choledochojejunostomy Roux en Y. No patient received adjuvant chemotherapy or radiotherapy in the post-operative period. During the same time period, 9 patients with non-resectable ampullary cancer were not fit for operation and were treated with ERCP and stent placement. These patients that did not undergo surgical treatment were also excluded from the study. None of them received adjuvant chemotherapy or radiotherapy in the post-operative period, as well.
Pancreaticoduodenectomy was the first choice as the type of surgical treatment. Local resection was the preferable treatment when the ampullary lesion was less than 2 cm in diameter, the pre-operative biopsy showed a pT1 cancer or adenoma of the ampulla of Vater and/or the patient's concomitant medical illness or age contraindicated a major operation such as Whipple's procedure. Finally, palliative by-pass surgery was reserved for patients whose tumor was larger than 5 cm according to the findings of the preoperative imaging investigations, obstruction of the portal vein, invasion of the superior mesenteric artery and/or vein, metastatic liver disease or distant metastases.
All patients underwent regular follow-up examinations post-operatively on a 3-month basis for the first year, on a 6-month basis for the following 4 years, and annually thereafter. Follow-up included clinical examination, blood tests (CA 19-9, serum bilirubin, alkaline phosphatase), abdominal ultrasound and chest radiography.
The statistical methods employed were Fisher's exact test for comparison of proportions. Differences among groups with respect to continuous variables were tested using the Kruskal-Wallis test, whereas pairwise differences were compared by the Mann-Whitney test, at a Bonferroni-adjusted significance level. The survival curves among groups were compared with the Log-rank test and they were presented graphically with the Kaplan-Meier plots. Analyses were conducted in SPSS 11.0 (SPSS, Inc., Chicago, IL). All reported p-values are two-tailed.
Results
All tumors were adenocarcinomas originating in the ampulla of Vater. The median size of the tumors as measured by the pathologist was 2.9 (0.8–5.1) cm. The tumor grade was recorded as well differentiated in 10, moderately differentiated in 13 and poorly differentiated in 9 patients including the two patients with unresectable tumors, according to pre-operative biopsy. Nodal involvement was microscopically found in 14 (46.6%) patients. The tumor was classified as pT1 in 8 patients and pT2-T4 in the remaining 22 patients with resectable tumors. All patients with pT1 tumors underwent local excision, while all patients with pT2-T4 tumors underwent pancreaticoduodenectomy, except one who underwent local excision.
When the final histological diagnosis after surgical treatment was compared with the preoperative histological finding on biopsy, accurate diagnosis was preoperatively established in 24 patients. In the remaining patients, the diagnosis was established intra-operatively by examination of frozen section in 3, and post-operatively by examination of the specimen in 3 patients.
The in-hospital, as well as the 30-day overall mortality rate, was 0% as no death occurred among the patients of the study. The overall hospital morbidity was 18.8%, as 9 complications occurred in 6 patients post-operatively. All the complications occurred in patients who underwent pancreaticoduodenectomy resulting in a morbidity of 28.6% in this group of patients. Although this morbidity was substantially higher compared to the group undergoing local excision and palliative surgery, this difference was not statistically significant (p = 0.129, Table 1).
Table 1 Comparison of different parameters among the different surgical approach groups.
Whipple's procedure (n = 21) Local excision (n = 9) Palliative surgery (n = 2) p-value
Hospital morbidity
n (%) 6 (28.6) 0 (0) 0 (0) 0.129a
Hospital stay, days
Median (range) 16 (9–32) 8 (7–10) 7.5 (7–8) <0.001b
a Fisher's exact test
b Kruskal-Wallis test for the overall comparison (The significant pairwise comparisons with Bonferroni adjusted p-value were Whipple's procedure vs. the other two groups for both the hospital stay and operation time [p < 0.05]).
In particular, one patient presented with moderate leakage of the choledochojejunostomy, which was treated conservatively. Pancreatic fistulas were seen in two patients, which were also treated conservatively. One patient developed a sub-hepatic abscess on the 20th post-operative day, but ultrasound-guided drainage was adequate treatment in this case. Surgical intervention was required in a patient who presented with intra-abdominal bleeding due to septic erosion of gastro-duodenal artery and rupture of the pancreatojejunostomy and choledochojejunostomy on the 13th post-operative day. Re-operation in this case included ligation of gastro-duodenal artery, reconstruction of choledochojejunostomy and drainage of the ruptured pancreatojejunostomy. The post-operative period in this case was uneventful while the pancreatic fistula subsided within two months. Finally, two patients experienced moderate wound infection and drainage of the wound abscess was adequate treatment in these cases. On the other hand, the post-operative course was uneventful in patients who underwent either local excision or palliative operation.
Seventeen (17) patients have died during follow-up, and all but three died because of recurrence. Following palliative operation (choledochojejunostomy Roux en Y), both patients died 7 and 13 months after operation. Following local excision of the ampullary cancer in 9 patients, the survival rate at 3 years was 77.7% (7 patients) and 33.3% (3 patients) at 5 years. In the patients that achieved 5-year survival after local excision the resection was R0, the tumor was graded as pT1, N0, M0 and, moreover, it was well-differentiated.
Among 21 patients that underwent Whipple's procedure, the 3-year survival rate was 76.2% (16 patients) and the 5-year survival rate was 62% (13 patients). One patient is alive 10 years after the operation and considered cancer free. As shown in Figure 1, the median survival after Whipple's procedure was significantly higher than the median survival after local resection (65.2 vs. 38.6 months, p < 0.001).
Figure 1 Time to death for patients who underwent pancreaticoduodenectomy, local excision and palliative surgery. Log-rank test p < 0.001.
The overall median hospital stay was 13.5 (range: 7–32) days. The median in-patient stay after Whipple's procedure was significantly greater than the median in-patient stay after local excision and palliative operation (16 vs. 8 and 7.5 days respectively, p < 0.001, Table 1).
Discussion
Periampullary pancreatic neoplasms rank as the fifth leading cause of cancer death behind lung, breast, colorectal and prostate cancer, causing more than 30,000 deaths per year in the United States [10]. Ampullary carcinomas account for 6%–20% of all periampullary tumors and for 10.2%–36% of all operable pancreatoduodenal neoplasms [2,3,11]. They account for about 0.2% of all gastrointestinal malignancies, with a median incidence of about 57 cases per million of population per year [12,13]. In our study, cancer of the ampulla of Vater represented 15.6% of all periampullary carcinomas treated during the same time period in our Department.
A major problem in dealing with ampullary tumors is to differentiate between an adenoma and a carcinoma. Modern imaging studies such as US, CT scan, magnetic resonance imaging (MRI) and magnetic resonance cholangiopancreaticography (MRCP), have significantly improved the diagnostic accuracy in the pre-operative period. Moreover, endoscopic diagnostic techniques, such as duodenoscopy with multiple biopsies and endoscopic ultrasonography are extremely helpful tools in order to determine the nature and the extend of the tumor pre-operatively. However, despite the fact that endoscopic ultrasonography may be helpful in detecting the invasion of the tumor into the surrounding tissues and the occurrence of lymph node enlargement, it does not allow differentiation between an adenoma and a pT1 carcinoma [14]. Furthermore, duodenoscopy with multiple biopsies has been reported to miss the diagnosis of malignancy in 12% to 40% of cases [15]. Because a negative biopsy does not exclude the presence of an invasive carcinoma within the adenoma, complete excision has been recommended in all cases [16]. The above-mentioned recommendation was also the policy in our study, as we performed either Whipple's resection or local excision in all cases with a resectable ampullary tumor. In our series, duodenoscopy with biopsies failed to reveal malignancy in 20% of the patients. Endoscopic ultrasonography was not included in our diagnostic procedure due to lack of experience and relevant equipment.
Options for excision of ampullary carcinomas include local excision and pancreaticoduodenectomy. Halsted performed the first local excision for ampullary carcinoma in 1899 [8]. Although there are many case reports and a few series on the treatment of ampullary neoplasms by local ampullary resection, the criteria used to decide when local excision is suitable for certain patients are controversial, and not well addressed. For small lesions, thought to be benign pre-operatively according to endoscopic appearance and biopsy, ampullary resection is generally well accepted [17]. Bottger et al [18] stated that the indications for local excision should be that the tumor is completely removed (R0), limited to the ampulla of Vater (pT1), not poorly differentiated and with no venous/lymphatic infiltration in patients with ASA grade IV, regardless of their age. Similarly, Beger et al [19] reported that local resection is indicated in cases with pT1, N0, M0 cancer of the ampulla of Vater, excluding patients with tumor poorly differentiated. Moreover, he stated that if intra- or even post-operative histological findings show a cancer more advanced than pT1, node-positive, or poorly differentiated tumor, the procedure should be extended to a pancreaticoduodenectomy and that in patients with pT1 cancer, local excision should be always combined with a local lymph node dissection. Regarding age as an indication for local excision, we can only analyze the reported evidence about age as a contraindication for Whipple's procedure. In many authors opinion, there is rarely any justification for performing major resection in a patient over 75 years of age due to high morbidity and mortality, as well as short survival [20]. However, Kairaluoma et al [21] suggested that age is not a limiting factor for pancreatic resection and it can be performed with acceptable survival rates even in patients over 70 years of age. In general, local ampullary resection is accompanied by significantly less morbidity and mortality than pancreaticoduodenectomy [19,22]. In our study, mortality as well as morbidity was 0% among 9 patients that underwent local excision for ampullary cancer. Our criteria for performing local excision were tumor size less than 2 cm at duodenoscopy, the patient's poor medical fitness, age > 75 years and pre-operative biopsy showing a pT1 cancer or adenoma of the ampulla of Vater. In a patient less than 75 years of age, the final histological examination showed a cancer more advanced than pT1, but we did not proceed to pancreaticoduodenectomy as his comorbidity contraindicated a major operation. Finally, during local excision we performed local lymph node dissection only if the nodes were enlarged.
Pancreaticoduodenectomy is undoubtedly the procedure of choice in the management of ampullary cancer in patients who are medically fit, and this fact has been reinforced by the declining mortality after the procedure during the past decades. This was also the policy in our study; pancreaticoduodenectomy with or without pylorus preservation was the first choice of surgical treatment in patients with tumors more advanced than pT1, unless it was contraindicated by the patient's comorbidity or age (>75 years), at which point a local excision was performed. Hospital mortality after pancreaticoduodenectomy is less than 5% in recently published series, whereas the morbidity remains high and varies from 25 to 65% [6,18-20,23]. The related mortality in our series was 0% but morbidity was as high as 28.6%. The pattern of failure after surgical resection for ampullary cancer is a crucial point to assess the efficacy of adjuvant therapy for patients receiving resection for cure. No patient in our study received adjuvant chemotherapy and this is in accordance with recent published studies showing no benefit of such therapy after curative or non-curative resection of ampullary cancer [24].
Survival after local resection it is difficult to estimate given the small number of patients, but it has been reported to be 40% to 50% at 5 years [25-29]. This figure is comparable to 37.5% to 62.7% 5-year survival rate reported in the much larger pancreaticoduodenectomy series [6,18-20,23]. However, it is important to mention that the pancreaticoduodenectomy series included high-risk lesions (T3 or T4, involved nodes, poor differentiation) which are excluded from the local excision series, and this might be an explanation for the comparable survival rates. In our study which included a small number of patients, following local excision of the ampullary cancer, the survival rate at 3 and 5 years was 77.7% and 33.3% respectively. Among the patients that underwent Whipple's procedure, the 3-year survival rate was 76.2% and the 5-year survival rate 62%.
Conclusion
In this series, local resection was a safe option in patients with significant co-morbidity or small ampullary tumors (<2 cm), and was associated with satisfactory long-term survival rates.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
DB: conceived the study, performed part of the operations and coordinated the preparation of the manuscript for submission.
EMZ: did the literature search and drafted the manuscript.
IL: participated in the design of the study, collection of data and preparation of manuscript for publication.
KT: Performed part of the operations, participated in the design of the study and preparation of manuscript for publication.
EC: Performed part of the operations, participated in the design of the study and preparation of manuscript for publication.
SK: participated in the design of the study and preparation of manuscript for publication.
EVZ: participated in the literature search and preparation of manuscript for publication and helped to draft the manuscript.
DB: has given final approval of the version to be published.
ID: has given final approval of the version to be published.
==== Refs
Conlon KC Dougherty E Klimstra DS Coit DG Turnbull AD Brennan MF The value of minimal access surgery in the staging of patients with potentially resectable peripancreatic malignancy Ann Surg 1996 223 134 140 8597506 10.1097/00000658-199602000-00004
Willett CG Warshaw AL Convery K Compton CC Patterns of failure after pancreatoduodenectomy for ampullary carcinoma Surg Gynecol Obstet 1993 173 33 38 8093983
Jones BA Langer B Taylor BR Girotti M Periampullary tumors: which one should be resected? Am J Surg 1985 149 46 52 3966641
Trede M Schwall G Saeger HD Survival after pancreatoduodenectomy 118 curative resections without an operative mortality Ann Surg 1990 211 447 458 2322039
Ishikawa O Ohigashi H Imaoka S Sasaki Y Iwanaga T Matayoshi Y Inoue T Is the long-term survival rate improved by preoperative irradiation prior to Whipple's procedure for adenocarcinoma of the pancreatic head? Arch Surg 1994 129 1075 1080 7944938
Allema JH Reinders ME van Gulik TM van Leeuwen DJ Verbeek PC de Wit LT Gouma DJ Results of pancreaticoduodenectomy for ampullary carcinoma and analysis of prognostic factors for survival Surgery 1995 117 247 253 7878528
Bakkevold KE Kambestad B Morbidity and mortality after radical and palliative pancreatic cancer surgery Ann Surg 1993 217 356 368 7682052
Haelsted WS Contributions to the surgery of the bile passages, especially of the common bile duct Boston Med Surg 1899 141 645 652
Bottger T Potratz D Wellek S Ochmann M Stockle M Klupp J Jungiger T Stellenwert der bildanalytischen DNS-Cytometrie beim carcinom der ampulla vateri Chirurg 1993 64 476 485 8359060
Yeo C Sohn T Cameron J Hruban R Lillemoe K Pitt H Periampullary adenocarcinoma: analysis of 5-year survivors Ann Surg 1998 227 821 831 9637545 10.1097/00000658-199806000-00005
Yamaguchi K Enjoji M Carcinoma of the ampulla of Vater: a clinicopathologic study and pathologic staging of 109 cases of carcinoma and 5 cases of adenoma Cancer 1987 59 506 515 3791159
Roder JD Schneider PM Stein HJ Siewert JR Number of lymph node metastases is significantly correlated with survival in patients with radically resected carcinoma of the ampulla of Vater Br J Surg 1995 82 1693 1696 8548244
Neoptolemos JP Talbot IC CarrLocke DL Treatment and outcome in 52 concecutive cases of ampullary carcinoma Br J Surg 1987 74 957 961 3664230
Nattermann C Dancygier H Endosonographie bei tumoren des pancreas und der gallenwege Leber Magen Darm 1993 5 13 20 8445972
Classen M Endoscopic approach to papillary stenosis Endoscopy 1981 13 154 156 7250083
Declore R Connor CS Thomas JH Friesen SR Hermreck AS Significance of tumor spread in adenocarcinoma of the ampulla of Vater Am J Surg 1989 158 593 597 2589595 10.1016/0002-9610(89)90201-8
Clary B Tyler D Dematos P Gottfried M Pappas T Local ampullary resection with careful intra-operative frozen section evaluation for presumed benign ampullary neoplasms Surgery 2000 127 628 633 10840357 10.1067/msy.2000.106532
Bottger TC Boddin J Heintz A Junginger T Clinicopathological study for the assessment of resection for ampullary carcinoma World J Surg 1997 21 379 383 9143568
Beger HG Treitschke F Gansauge F Harada N Hiki N Mattfeldt T Tumor of the ampulla of Vater Arch Surg 1999 134 526 532 10323425 10.1001/archsurg.134.5.526
Su CH Shyr YM Lui WY P'eng FK Factors affecting morbidity, mortality and survival after pancreaticoduodenectomy for carcinoma of the ampulla of Vater Hepato-Gastroenterology 1999 46 1973 1979 10430380
Kairaluoma MI Kiviniemi H Stahlberg M Pancreatic resection for carcinoma of the pancreas and the periampullary region in patients over 70 years of age Br J Surg 1987 74 116 118 3815026
Rattner D Fernandez-del Castillo C Brugge W Warshaw A Defining the criteria for local resection of ampullary neoplasms Arch Surg 1996 131 366 371 8615720
Sperti C Pasquali C Piccoli A Sernagiotto C Pedrazzoli S Radical resection for ampullary carcinoma: long-term results Br J Surg 1994 81 668 671 7913860
Takada T Amano H Yasuda H Is postoperative adjuvant chemotherapy useful for gallbladder carcinoma? A phase III multicentre prospective randomized controlled trial in patients with resected pancreaticobiliary carcinoma Cancer 2002 95 1685 1695 12365016 10.1002/cncr.10831
Klein P Reingruber B Kastl S Dworak O Hohenberger W Is local excision of pT1 ampullary carcinomas justified? Eur J Surg Oncol 1996 22 366 371 8783654 10.1016/S0748-7983(96)90286-3
Knox R Kingston R Carcinoma of the ampulla of Vater Br J Surg 1986 73 72 73 3947884
Newman R Pittam M Local excision in the treatment of carcinoma of the ampulla of Vater J R Coll Surg Edinb 1982 27 154 157 7108832
Robertson J Imrie C Local excision of ampullary carcinoma Acta Chir Scand 1986 152 537 539 3788399
Tarazi R Hermann R Vogt D Results of surgical treatment of periampullary tumors: a thirty-five-year experience Surgery 1986 100 716 722 3764694
|
16150156
|
PMC1215508
|
CC BY
|
2021-01-04 16:38:37
|
no
|
Int Semin Surg Oncol. 2005 Sep 9; 2:18
|
latin-1
|
Int Semin Surg Oncol
| 2,005 |
10.1186/1477-7800-2-18
|
oa_comm
|
==== Front
J Autoimmune DisJournal of Autoimmune Diseases1740-2557BioMed Central London 1740-2557-2-71613524910.1186/1740-2557-2-7HypothesisTransfer of efficient anti-melanocyte T cells from vitiligo donors to melanoma patients as a novel immunotherapeutical strategy Palermo Belinda [email protected] Silvia [email protected] Stefania [email protected] Claudia [email protected] Experimental Immunology Laboratory, IRCCS Maugeri Foundation, Pavia, Italy2 Department of Clinical and Biological Sciences, University of Turin, Turin, Italy2005 31 8 2005 2 7 7 13 6 2005 31 8 2005 Copyright © 2005 Palermo et al; licensee BioMed Central Ltd.2005Palermo et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Vitiligo is a relatively common progressive depigmentary condition that is believed to be due to the autoimmune-mediated loss of epidermal melanocytes. High frequencies of self-reactive T lymphocytes directed toward melanocyte differentiation antigens are found in vitiligo patients and might be directly responsible for the pathogenesis of the disease. An interesting aspect of vitiligo is its relation to melanoma: cytotoxic T lymphocytes directed to self antigens shared by normal melanocytes and melanoma cells are found in both conditions, but the resulting immune reactions are completely different. From this standpoint, the selective destruction of pigment cells that occurs in cases of vitiligo is the therapeutic goal sought in melanoma research.
Presentation of the hypothesis
Our working hypothesis is that vitiligo patients might represent a unique source of therapeutic cells to be used in allo-transfer for HLA-matched melanoma patients. The adoptive transfer of ex-vivo generated autologous tumor-specific T cells is a therapy that has met with only limited success, essentially because of inability to isolate therapeutically valuable T cells from the majority of tumor patients. Ideally, model systems where strong and efficient responses against the same (tumor) antigens are achieved would represent a better source of therapeutic cells. We believe it is possible to identify one such model in the melanoma-vitiligo dichotomy: T lymphocytes specific for different melanocyte differentiation antigens are found in vitiligo and represent the effective anti-melanocyte reactivity that is often ineffective in melanoma.
Testing the hypothesis
Melanocyte-specific T cell clones can be isolated from the peripheral blood of vitiligo patients and tested for their capacity to efficiently expand in vitro without loosing their cytotoxic activity and to migrate to the skin. Cytotoxicity against melanoma patients' non-tumor cells can also be tested. In addition, it would be interesting to attempt an in vivo animal model. If the results obtained from these validation steps will be satisfactory, it might be possible to plan the clinical grade preparation of relevant clones for transfer.
Implications of the hypothesis
When translated into a clinical trial, the possibility of in vitro selecting few effective tumor-specific T cell clones for infusion, inherent with this approach, could enhance the therapeutic graft-versus-tumor effect while possibly decreasing the risk of graft-versus-host disease.
==== Body
Background
Vitiligo is a common skin disease characterized by the development of white macules and patches associated with local melanocyte loss [1]. Its etiology is not completely known, but the observation of circulating antimelanocytic antibodies and of lymphocytic infiltrations at the margins of lesions in the majority of patients has lent support to the hypothesis that it is an autoimmune disease [2-5].
Many self proteins expressed by melanocytes in the skin of healthy donors, in non-depigmented skin of vitiligo patients and in melanoma patients are demonstrably immunogenic [6-8]. Antigens present in both tumor cells (melanoma) and their normal cellular counterpart (melanocytes) are known as melanocyte differentiation antigens. Among these, Melan-A/MART-1 is a melanosomal protein whose immunodominant epitope for HLA-A*0201 was identified by a screening with cytotoxic T cells [9].
High frequencies of melanocyte-specific CD8+ T lymphocytes are found in vitiligo patients. Using HLA/epitope tetramers, an instrument for measuring the frequency of antigen-specific T cells independently of their functional state [10], Ogg and co-authors [11] directly demonstrated for the first time the presence of high frequencies of CD8+ T cells specific for melanocytic antigens in the peripheral blood of HLA-A*0201+ patients with autoimmune vitiligo. Following this pioneering study, other groups [reviewed in [12-14]] contributed, through the use of tetramers directly ex vivo, without any antigen-specific stimulation, to the demonstration that melanocyte-specific CD8+ cells are present in the peripheral blood of both melanoma and vitiligo patients. In particular, high numbers of Melan-A/MART1-specific cells were detected in the majority of patients. Besides circulating cells, melanocyte-specific CD8+ T lymphocytes were also observed in situ in both depigmenting lesions of patients with vitiligo [15,16].
These melanocyte-specific CD8+ T Lymphocytes might be relevant for the pathogenesis of vitiligo. The first suggestions came from rare case reports on inflammatory vitiligo [17,18] and immunohistochemical studies later on confirmed the presence of infiltrating T cells in apposition to perilesional melanocytes [19]. Importantly, similar in situ T cell infiltrates were also detected in the more common form of the disease, generalized vitiligo [20-22]. Further indications favoring a pathogenetic role for melanocyte-specific, CD8+ T cells in vitiligo came from the direct correlation between their frequency within the total T cell pool and disease activity [11,23], as well as from their capacity to kill HLA-matched tumor cells [11,24,25] and, most notably, normal matched melanocytes [16,26].
Melanoma is an aggressive form of tumor whose incidence increases by 5% per year. Although the presence of melanoma-specific CTLs in cancer patients demonstrate that tumor cells may not completely evade immune recognition, the patient's immune system can only rarely counteract tumor growth [27-29]. An unusual facet of vitiligo is its relation to melanoma: cytotoxic T lymphocytes directed at self antigens shared by normal melanocytes and melanoma cells are found in both conditions and suggest a breakdown of tolerance [11,30-34], yet the resulting immune reaction is the opposite [35-37]. In vitiligo, natural immune tolerance is over-ridden such that the host immune system can orchestrate melanocyte destruction, whereas in melanoma, an immune effector function of potential benefit to the host, i.e., efficient destruction of transformed melanocytes, does occur very rarely. In this respect, it would seem that reactivity to vitiligo melanocytes may be the effective variant of an immune response often ineffective in melanoma. The mechanisms causing these opposite effects are not known but these data, together with the resistance of melanoma to conventional chemotherapeutic and radiotherapeutic approaches, have made the melanoma/vitiligo dichotomy an important model for immunologic investigation.
Avidity of antigen recognition is an important feature of tumor-specific T lymphocytes, determining their capability to kill tumor cells. Notably, a few recent studies indicated that Melan-A/MART1-specific CD8+ T cells isolated from vitiligo patients possess an increased avidity and exert a superior anti-tumor activity than those from melanoma [24,25]. Further indications come from animal models. In one recent study [38] MT-ret transgenic mice, a model for human cutaneous melanoma, were used to investigate the natural anti-tumor T cell response. A large proportion of these animals developed melanoma-associated vitiligo and a good correlation was found between vitiligo development and melanoma control. Interestingly, T cells that secreted IFN-γ in response to melanoma cells were statistically more frequent in melanoma mice that developed vitiligo than in mice that did not, suggesting that vitiligo-associated T cells possessed an increased functional avidity. These data suggest that a qualitative difference exists between the anti-melanocyte cytotoxic T cell responses found in vitiligo versus melanoma that might explain the opposite immunologic outcomes.
Presentation of the Hypothesis
Autoimmune conditions stem from a break of tolerance to defined autoantigens and this allows for the production of high avidity antigen-specific responses. If these antigens are also relevant to tumor immunity, autoimmune cells can be exploited for tumor intervention. Our working hypothesis is that vitiligo patients might represent a unique source of therapeutic cells for HLA-matched melanoma patients, essentially due to their superior TCR affinity and increased tumoricidal potential.
The adoptive transfer of ex-vivo generated autologous tumor-specific T cells is a potentially potent therapy that has met with only limited success. An important limitation to such treatment is inability to isolate and generate therapeutically valuable T cells from the majority of tumor patients. The causes of these limitations are not well known, but might include the inability of the tumor cells to trigger a T-cell response (ignorance) or their ability to actively suppress or delete T cells of the highest avidity (tolerance). In addition, T cells derived from melanoma patients may often display functional impairments as a consequence of the tumor environment, which inhibits their acquisition of final effector functions [32]. Indeed, in vitro culture of these tumor-sensitized T cells with appropriate activation stimuli has been shown to partially restore normal functional properties [39] and to confer effector activity that could potentially sustain tumor rejection upon re-infusion [40]. However, in the majority of cases, autologous T cells are found to exert only poor cytotoxic activity toward HLA matched melanoma cells. On the other hand, CTLs against different melanoma associated antigen are found in vitiligo and mediate autologous melanocyte destruction [24-26,30,31]. Notably, the Melan-A/MART1-specific CD8+ T cells isolated from vitiligo patients appear to possess an increased avidity and to exert a superior anti-tumor activity than those from melanoma [24-26].
How can this knowledge be applied to the definition of new immunotherapeutic strategies for melanoma patients? A first possibility is an allo-transfer approach, where very efficient anti-tumor CD8+ T lymphocytes from vitiligo donors can be transferred into HLA-matched melanoma patients. High-avidity and tumor-specific T cell clones can be isolated from the peripheral blood of HLA-A2 vitiligo patients; if these could be proven to efficiently expand in vitro without loosing their cytotoxic activity and to maintain their skin-homing capability, they might represent a valuable source of therapeutic cells.
Testing the Hypothesis
In order to test the hypothesis, melanocyte-specific T cell clones could be tetramer-sorted from the peripheral blood of vitiligo patients and tested for their capacity to efficiently expand in vitro without loosing their cytotoxic activity. They will also be tested for their skin-homing activity by FACS-evaluating surface expression of the cutaneous lymphocyte antigen, CLA. As a first effort in determining the effects of host cellular environment on allo-transfer, these cells can be mixed with PBMC from HLA-matched melanoma patients and the anti-tumor potential of the vitiligo/melanoma pools assayed. Possible immunosuppressor effects exerted by either the melanoma patients' sera [41] or the tumor itself might be addressed through co-cultures. Cytotoxicity against melanoma patients' non-tumor cells (peripheral blood, skin biopsies) can also be tested. In addition, it would be interesting to attempt an in vivo animal model using immunodeficient SCID or RAG knockout mice, rendered transgenic for human HLA and transplanted with melanoma. These mice might be treated with the clone from a vitiligo patient, with or without melanoma patient's lymphocytes. If the results obtained from both the in vitro validation step and the in vivo animal model will be satisfactory, it will be possible to plan the clinical grade preparation of relevant clones for transfer. A major drawback of this approach remains achievement of HLA match between a T cell donor and a melanoma patient. Only weak associations between vitiligo and several HLA-class I and -class II alleles have been reported to date, and essentially no significant association exists with melanoma. So, as patients with vitiligo represent about 1% of the general population, the probability to find a perfect match can be estimated to be a hundred-time lower than the one theoretically achieved for allogeneic transplantation involving the entire population as potential donors.
Implications of the Hypothesis
If this hypothesis were true, at least some autoimmune diseases might be seen as novel sources of therapeutic cells for tumor patients. When translated into a clinical trial, the possibility of in vitro selecting few tumor-specific T cell clones for infusion, inherent with this approach, could enhance the therapeutic graft-versus-tumor effect while possibly decreasing the risk of graft-versus-host disease.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
B.P. participated in the design of the study; S.G. and S.M. participated in analysis and interpretation of data; C.G. conceived the study and drafted the manuscript.
Acknowledgements
This work was partially supported by Associazione Italiana per la Ricerca sul Cancro (AIRC). S.G. was supported by a fellowship from Fondazione Italiana per la Ricerca sul Cancro (FIRC). We are grateful to melanoma and vitiligo patients for their generous participation in this research project.
==== Refs
Njoo MD Westerhof W Vitiligo. Pathogenesis and treatment Am J Clin Dermatol 2001 2 167 181 Review 11705094
Bystryn JC Immune mechanisms in vitiligo Clin Dermatol 1997 15 853 861 Review 9404688 10.1016/S0738-081X(97)00126-0
Le Poole IC Wankowicz-Kalinska A van den Wijngaard RM Nickoloff BJ Das PK Autoimmune aspects of depigmentation in vitiligo J Invest Dermatol Symp Proc 2004 9 68 72 10.1111/j.1087-0024.2004.00825.x
Ongenae K van Geel N Naeyaert J-M Evidence for an autoimmune pathogenesis of vitiligo Pigment Cell Res 2003 16 90 100 12622785 10.1034/j.1600-0749.2003.00023.x
van den Wijngaard R Wankowicz-Kalinska A Pals S Weening J Das P Autoimmune melanocyte destruction in vitiligo Lab Invest 2001 81 1061 1067 11502857
Anichini A Maccalli C Mortarini R Salvi S Mazzocchi A Squarcina P Herlyn M Parmiani G Melanoma cells and normal melanocytes share antigens recognized by HLA-A2-restricted cytotoxic T cell clones from melanoma patients J Exp Med 1993 177 989 998 8459226 10.1084/jem.177.4.989
Le Poole IC Mutis T van den Wijngaard RM Westerhof W Ottenhoff T de Vries RR Das PK A novel, antigen-presenting function of melanocytes and its possible relationship to hypopigmentary disorders J Immunol 1993 151 7284 7292 8258725
Meije CB Mooi WJ Le Poole IC Van Muijen GN Das PK Micro-anatomy related antigen expression in melanocytic lesions J Pathol 2000 190 572 578 10727983 10.1002/(SICI)1096-9896(200004)190:5<572::AID-PATH570>3.0.CO;2-9
Kawakami Y Eliyahu S Sakaguchi K Robbins PF Rivoltini L Yannelli JR Appella E Rosenberg SA Identification of the immunodominant peptides of the MART-1 human melanoma antigen recognized by the majority of HLA-A2-restricted tumor infiltrating lymphocytes J Exp Med 1994 180 347 352 7516411 10.1084/jem.180.1.347
Altman JD Moss PA Goulder PJ Barouch DH McHeyzer-Williams MG Bell JI McMichael AJ Davis MM Phenotypic analysis of antigen-specific T lymphocytes Science 1996 274 94 96 8810254 10.1126/science.274.5284.94
Ogg GS Dunbar R Romero P Chen JL Cerundolo V High frequency of skin-homing melanocyte-specific cytotoxic T lymphocytes in autoimmune vitiligo J Exp Med 1998 188 1203 1208 9743539 10.1084/jem.188.6.1203
Ogg GS Detection of antigen-specific cytotoxic T lymphocytes: significance for investigative dermatology Clin Exp Dermatol 2000 25 312 316 Review 10971494 10.1046/j.1365-2230.2000.00651.x
Parmiani G Melanoma antigens and their recognition by T cells Keio J Med 2001 50 86 90 Review 11450597
Palmowski M Salio M Dunbar RP Cerundolo V The use of HLA class I tetramers to design a vaccination strategy for melanoma patients Immunol Rev 2002 188 155 163 Review 12445289 10.1034/j.1600-065X.2002.18814.x
Becker JC Guldberg P Zeuthen J Brocker EB Straten PT Accumulation of identical T cells in melanoma and vitiligo-like leukoderma J Invest Dermatol 1999 113 1033 1038 10594748 10.1046/j.1523-1747.1999.00805.x
Wankowicz-Kalinska A van den Wijngaard RM Tigges BJ Westerhof W Ogg GS Cerundolo V Storkus WJ Das PK Immunopolarization of CD4+ and CD8+ T cells to Type-1-like is associated with melanocyte loss in human vitiligo Lab Invest 2003 83 683 695 12746478
Buckley WR Lobitz WC Jr [Vitiligo with a raised inflammatory border.] AMA Arch Derm Syphilol 1953 67 316 320 13029928
Michaelsson G Vitiligo with raised borders Report of two cases Acta Derm Venereol 1968 48 158 161
Le Poole IC van den Wijngaard RM Westerhof W Das PK Presence of T cells and macrophages in inflammatory vitiligo skin parallels melanocyte disappearance Am J Pathol 1996 148 1219 1228 8644862
al Badri AM Todd PM Garioch JJ Gudgeon JE Stewart DG Goudie RB An immunohistological study of cutaneous lymphocytes in vitiligo J Pathol 1993 170 149 155 8345407 10.1002/path.1711700209
Abdel-Naser MB Kruger-Krasagakes S Krasagakis K Gollnick H Abdel-Fattah A Orfanos CE Further evidence for involvement of both cell mediated and humoral immunity in generalized vitiligo Pigment Cell Res 1994 7 1 8 8072943
van den Wijngaard R Wankowicz-Kalinska A Le Poole C Tigges B Westerhof W Das P Local immune response in skin of generalized vitiligo patients. Destruction of melanocytes is associated with the prominent presence of CLA+ T cells at the perilesional site Lab Invest 2000 80 1299 1309 10950121
Lang KS Caroli CC Muhm A Wernet D Moris A Schittek B Knauss-Scherwitz E Stevanovic S Rammensee HG Garbe C HLA-A2 restricted, melanocyte-specific CD8(+) T lymphocytes detected in vitiligo patients are related to disease activity and are predominantly directed against MelanA/MART1 J Invest Dermatol 2001 116 891 897 11407977 10.1046/j.1523-1747.2001.01363.x
Palermo B Campanelli R Garbelli S Mantovani S Lantelme E Brazzelli V Ardigo M Borroni G Martinetti M Badulli C Specific cytotoxic T lymphocyte responses against Melan-A/MART1, tyrosinase and gp100 in vitiligo by the use of major histocompatibility complex/peptide tetramers: the role of cellular immunity in the etiopathogenesis of vitiligo J Invest Dermatol 2001 117 326 332 11511311 10.1046/j.1523-1747.2001.01408.x
Mantovani S Palermo B Garbelli S Campanelli R Robustelli Della Cuna G Gennari R Benvenuto F Lantelme E Giachino C Dominant TCR-alpha requirements for a self antigen recognition in humans J Immunol 2002 169 6253 6260 12444131
Mantovani S Garbelli S Palermo B Campanelli R Brazzelli V Borroni G Martinetti M Benvenuto F Merlini G Robustelli della Cuna G Molecular and functional bases of self-antigen recognition in long-term persistent melanocyte-specific CD8+ T cells in one vitiligo patient J Invest Dermatol 2003 121 308 134 12880423 10.1046/j.1523-1747.2003.12368.x
Ferrone S Marincola FM Loss of HLA class I antigens by melanoma cells: molecular mechanisms, functional significance and clinical relevance Immunol Today 1995 16 487 494 Review 7576053 10.1016/0167-5699(95)80033-6
Hahne M Rimoldi D Schroter M Romero P Schreier M French LE Schneider P Bornand T Fontana A Lienard D Melanoma cell expression of Fas(Apo-1/CD95) ligand: implications for tumor immune escape Science 1996 274 1363 1366 8910274 10.1126/science.274.5291.1363
Maeurer MJ Gollin SM Martin D Swaney W Bryant J Castelli C Robbins P Parmiani G Storkus WJ Lotze MT Tumor escape from immune recognition: lethal recurrent melanoma in a patient associated with downregulation of the peptide transporter protein TAP-1 and loss of expression of the immunodominant MART-1/Melan-A antigen J Clin Invest 1996 98 1633 1641 8833913
Yee C Thompson JA Roche P Byrd DR Lee PP Piepkorn M Kenyon K Davis MM Riddell SR Greenberg PD Melanocyte destruction after antigen-specific immunotherapy of melanoma: direct evidence of t cell-mediated vitiligo J Exp Med 2000 192 1637 1644 11104805 10.1084/jem.192.11.1637
Le Gal FA Avril MF Bosq J Lefebvre P Deschemin JC Andrieu M Dore MX Guillet JG Direct evidence to support the role of antigen-specific CD8(+) T cells in melanoma-associated vitiligo J Invest Dermatol 2001 117 1464 1470 11886510 10.1046/j.0022-202x.2001.01605.x
Rivoltini L Carrabba M Huber V Castelli C Novellino L Dalerba P Mortarini R Arancia G Anichini A Fais S Immunity to cancer: attack and escape in T lymphocyte-tumor cell interaction Immunol Rev 2002 188 97 113 Review 12445284 10.1034/j.1600-065X.2002.18809.x
Romero P Valmori D Pittet MJ Zippelius A Rimoldi D Levy F Dutoit V Ayyoub M Rubio-Godoy V Michielin O Antigenicity and immunogenicity of Melan-A/MART-1 derived peptides as targets for tumor reactive CTL in human melanoma Immunol Rev 2002 188 81 96 Review 12445283 10.1034/j.1600-065X.2002.18808.x
Mandelcorn-Monson RL Shear NH Yau E Sambhara S Barber BH Spaner D DeBenedette MA Cytotoxic T lymphocyte reactivity to gp100, MelanA/MART-1, and tyrosinase, in HLA-A2-positive vitiligo patients J Invest Dermatol 2003 121 550 556 12925214 10.1046/j.1523-1747.2003.12413.x
Das PK van den Wijngaard RM Wankowicz-Kalinska A Le Poole IC A symbiotic concept of autoimmunity and tumour immunity: lessons from vitiligo Trends in Immunology 2001 22 130 136 11286727 10.1016/S1471-4906(00)01844-5
Ramirez-Montagut T Turk MJ Wolchok JD Guevara-Patino JA Houghton AN Immunity to melanoma: unraveling the relation of tumor immunity and autoimmunity Oncogene 2003 22 3180 3187 12789294 10.1038/sj.onc.1206462
Wankowicz-Kalinska A Le Poole C van den Wijngaard R Storkus WJ Das PK Melanocyte-specific immune response in melanoma and vitiligo: two faces of the same coin? Pigment Cell Res 2003 16 254 260 Review 12753398 10.1034/j.1600-0749.2003.00038.x
Lengagne R Le Gal FA Garcette M Fiette L Ave P Kato M Briand JP Massot C Nakashima I Renia L Spontaneous vitiligo in an animal model for human melanoma: role of tumor-specific CD8+ T cells Cancer Res 2004 64 1496 1501 14973052
Rivoltini L Kawakami Y Sakaguchi K Southwood S Sette A Robbins PF Marincola FM Salgaller ML Yannelli JR Appella E Induction of tumor-reactive CTL from peripheral blood and tumor-infiltrating lymphocytes of melanoma patients by in vitro stimulation with an immunodominant peptide of the human melanoma antigen MART-1 J Immunol 1995 154 2257 2265 7868898
Rosenberg SA Yannelli JR Yang JC Topalian SL Schwartzentruber DJ Weber JS Parkinson DR Seipp CA Einhorn JH White DE Treatment of patients with metastatic melanoma with autologous tumor-infiltrating lymphocytes and interleukin 2 J Natl Cancer Inst 1994 86 1159 1166 8028037
Juranic ZD Stanojevic-Bakic N Zizak Z Babovic N Radovic-Kovacevic V Stanojkovic T Dzodic R Antimelanoma immunity in vitiligo and melanoma patients Neoplasma 2003 50 305 309 12937846
|
16135249
|
PMC1215509
|
CC BY
|
2021-01-04 16:38:03
|
no
|
J Autoimmune Dis. 2005 Aug 31; 2:7
|
utf-8
|
J Autoimmune Dis
| 2,005 |
10.1186/1740-2557-2-7
|
oa_comm
|
==== Front
J Inflamm (Lond)Journal of Inflammation (London, England)1476-9255BioMed Central London 1476-9255-2-81604580010.1186/1476-9255-2-8ReviewMesenchymal stem cells avoid allogeneic rejection Ryan Jennifer M [email protected] Frank P [email protected] J Mary [email protected] Bernard P [email protected] Institute of Immunology, National University of Ireland, Maynooth, Co. Kildare Ireland2 Regenerative Medicine Institute (REMEDI), National Centre for Biomedical Engineering Science, National University of Ireland, Galway, Ireland2005 26 7 2005 2 8 8 1 4 2005 26 7 2005 Copyright © 2005 Ryan et al; licensee BioMed Central Ltd.2005Ryan et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Adult bone marrow derived mesenchymal stem cells offer the potential to open a new frontier in medicine. Regenerative medicine aims to replace effete cells in a broad range of conditions associated with damaged cartilage, bone, muscle, tendon and ligament. However the normal process of immune rejection of mismatched allogeneic tissue would appear to prevent the realisation of such ambitions. In fact mesenchymal stem cells avoid allogeneic rejection in humans and in animal models. These finding are supported by in vitro co-culture studies. Three broad mechanisms contribute to this effect. Firstly, mesenchymal stem cells are hypoimmunogenic, often lacking MHC-II and costimulatory molecule expression. Secondly, these stem cells prevent T cell responses indirectly through modulation of dendritic cells and directly by disrupting NK as well as CD8+ and CD4+ T cell function. Thirdly, mesenchymal stem cells induce a suppressive local microenvironment through the production of prostaglandins and interleukin-10 as well as by the expression of indoleamine 2,3,-dioxygenase, which depletes the local milieu of tryptophan. Comparison is made to maternal tolerance of the fetal allograft, and contrasted with the immune evasion mechanisms of tumor cells. Mesenchymal stem cells are a highly regulated self-renewing population of cells with potent mechanisms to avoid allogeneic rejection.
==== Body
Review
Introduction: What are Stem Cells?
The term "stem cell" can be applied to a remarkably diverse group of cells. These cells, regardless of their source, share two characteristic properties. Firstly, they have the capacity for prolonged or unlimited self-renewal under controlled conditions, and secondly they retain the potential to differentiate into a variety of more specialized cell types [1,2]. The stem cells that arise during the first days of mammalian embryonic development are pluripotent and are referred to as embryonic stem (ES) cells. These are usually derived from the inner cell mass of the pre-implantation embryo, at the blastocyst stage[3]. However stem cells are not confined to tissues of early development, but can also be found at various sites in the adult mammal. Adult stem cells are more differentiated then ES cells but can still give rise to specialized lineages[1,2]. The best-described populations to date are the hematopoietic stem cells (HSC) of the bone marrow that can generate various blood cells[4]. However the bone marrow also contains a population of mesenchymal stem cells (MSC) [1,2]. These cells, first characterized by Friedenstein and colleagues more than thirty years ago, are multipotent cells capable of differentiating into several lineages including; cartilage, bone, muscle, tendon, ligament and adipose tissue[2,5,6]. In their undifferentiated state, MSC are spindle-shaped and resemble fibroblasts[5,6] (Fig 1). There are no cell surface markers that specifically and uniquely identify MSC, and their characterization in the literature lacks consistency. The diversity of characteristics associated with MSC can be explained by differences in tissue origin, isolation methods and culture conditions between laboratories, in addition there appear to be strain-to-strain differences in murine derived MSC[2,7-9]. Whilst there is an obvious need for standardization between research groups, some consensus can be found among the conflicting data. In broad terms, MSC expanded in vitro do not express the hematopoietic or endothelial surface markers CD11b, CD14, CD31, CD34 or CD45 but stain positive for CD29, CD44, CD73, CD105, CD106 and CD166 [2,5,10]. The non-embryonic source of this population, the reduced likelihood of neoplasia, and the more limited differentiation potential, have made these cells attractive candidates for application in cell based therapies usually termed "regenerative medicine"[2]. There is one confounding influence on this approach; whilst self derived MSC pose few immunological problems, in practice regenerative medicine is likely to rely on mismatched (allogeneic) cells to repair or replace damaged tissue. Normally, allogeneic cells are deleted by host immune responses. The major surprise to Immunologists working in this field have been findings that suggest that MSC do not obey the normal "rules" of allogeneic rejection. This review will survey recent data, which convincingly indicate the mechanisms by which MSC escape the normal process of alloantigen recognition.
Figure 1 Human mesenchymal stem cells (MSC) are spindle shaped, fibroblast-like cells. Original magnification × 100, phase-contrast light microscopy, scale bar represents 50 μm.
MSC evade allorejection
The major limit to solid organ graft survival is T cell recognition by the recipient of alloantigen (dominated by, but not confined to MHC/HLA antigens)[11]. There are two mechanisms mediating this powerful rejection response; "direct" recognition, involving recognition by recipient CD8+ or CD4+ T cells of donor MHC class I and class II molecules; and "indirect" mechanisms involving recognition of peptides from the allogeneic tissue[11]. Recipient antigen presenting cells (APC) such as dendritic cells (DC) process alloantigen into peptides and present these to naive T cells on self-MHC molecules [12]. However there are notable exceptions to these allorejection processes; the fetal allograft evades rejection by the mother through a complex series of actions (reviewed in[13]), similarly tissue which has limited lymphatic drainage is less prone to allorejection[14]. Interestingly tumor cells, whilst not allogeneic, are in many cases both "altered-self" and immunogenic but often actively modulate immune responsiveness to evade immune surveillance[15]. Thus mechanisms of tumor evasion of the immune system may provide insight into how allogeneic MSC are tolerated by the mismatched host.
There is supporting evidence for the use of allogeneic MSC from both in vitro and in vivo studies that show MSC avoid normal alloresponses. A small number of in- vivo studies suggest that MSC play a role in enabling alloantigen tolerance. Koc et al, showed no evidence of alloreactive T cells and no incidence of graft v host disease when allogeneic MSC were infused into patients with Hurler's syndrome or metachromatic leukodystrophy[16]. In a previous study by the same group, autologous culture-expanded MSC were infused to breast cancer patients to investigate whether MSC would enhance the engraftment of peripheral blood stem cells after myeloablative therapy [17]. Results showed rapid hematopoietic recovery and no signs of toxicity from MSC infusion[17]. Horwitz and colleagues, reported that donor MSC contributed to bone remodelling after allogeneic stem cell transplantation in three children with osteogenesis imperfecta (OI)[18], a rare genetic disorder of type I collagen. This is supported by data from Bartholomew et al who showed that in-vivo administration of allogeneic MSC prolonged 3rd party skin graft survival in animal models[19]. Furthermore, Saito et al, demonstrated that MSC undergoing differentiation to a cardiac phenotype were tolerated in a xenogeneic environment, retaining their ability to be recruited to the injured myocardium[20]. More recent work by Aggarwal and Pittenger supported the feasibility of MSC-transplantation showing that MSC altered the phenotypes of specific immune cell subtypes thereby creating a tolerogenic environment[21]. These reports suggest that transplantation of MSC could be beneficial in patients with various disorders requiring tissue regeneration, and provide evidence supporting the tolerance of allogeneic MSC by recipients.
Data supporting the contention that MSC avoid allogeneic responses has also come from a large body of in vitro experiments, usually involving co-culture or mixed lymphocyte reactions (MLR). Evidence from these studies indicate that the use of mismatched MSC does not provoke a proliferative T cell response in allogeneic MLR, thus suggesting an immunosuppressive role for MSC[19,22-26]. Le Blanc et al, showed that MSC failed to elicit proliferation of allogeneic lymphocytes[27]. Additionally, they demonstrated that MSC remained immunosuppressive even after IFN-γ stimulation[27]. Evidence from Krampera et al confirms these findings, they showed that murine MSC lack MHC class II and inhibited T cell proliferation[25]. Tse et al, also showed that human MSC fail to elicit allogeneic T cell response in a MLR even when MHC class II was upregulated[28]. Consistent with these studies, Bartholomew et al showed that allogeneic baboon MSC suppressed the proliferative activity of lymphocytes in vitro and prolonged graft survival[19]. These findings support the view that MSC can be transplanted between MHC-incompatible individuals. Although these data show that successful use of allogeneic MSC in regenerative therapy is possible, such approaches are unlikely to be broadly acceptable until it is understood why MSC are not rejected. This question has been the subject of intense recent study and three candidate mechanisms are emerging. MSC appear to evade allogeneic rejection by a) being hypoimmunogenic; b) modulating T cell phenotype and c) creating an immunosuppressive local milieu. These mechanisms are inter-related and will involve cell contact dependent and independent interactions. The challenge facing the field is to unravel the contribution of these diverse interactions.
MSC are hypoimmunogenic
There is controversy surrounding the cell surface expression of MHC alloantigens by MSC. Although conflicting evidence exists, most studies describe human MSC as MHC class I positive and MHC class II negative (Fig 2). The data conflicting with these findings may represent different stem cell lineages or be the result of the recently described process of cell-cell transfer [29-31]. The expression of MHC class I by MSC is important because expression protects MSC from certain NK cell mechanisms of deletion. For instance, a major function of NK and NK-like cells is to kill tumor cells that have downregulated class I [32]. HLA-G is an MHC-like protein that is known to protect the fetal allograft against NK mediated rejection[33,34]. This protein has been shown to bind to the two major inhibitory NK receptors, KIR1 and KIR2, and to inhibit NK killing [35-37]. However no studies of HLA-G expression by MSC have been reported to date.
Figure 2 Human MSC cultured according to [106, 107] are A) MHC-I positive (HLA-A,B,C, antibody W6/32-FITC), B) MHC class II negative (HLA-DR, antibody LN-3-PE); C) CD14 negative (antibody MEM-18-FITC), D) CD86 negative (antibody IT2.2-PE); and E) CD40L/ CD154 (antibody 24-31-FITC), F) CD95L (FasL) negative (antibody NOK-1-PE). Isotype matched control antibody labelling are shown as unshaded plots, FITC conjugates are shown in blue, PE conjugates shown in pink. Flow cytometry performed according to methods previously described [108-110].
As MHC class II proteins are potent alloantigens, the expression by MSC is another important factor. Again there is some controversy over expression, which may be explained by the diversity of models described above. However there are widespread observations that under non-inflammatory conditions, human MSC are MHC-II negative, supporting a role for MSC as having reduced immunogenicity through the control of alloantigen expression [38-40]. The absence of MHC class II gives MSC the potential to escape recognition by alloreactive CD4+ T cells. In addition to being MHC II negative, MSC do not appear to express the co-stimulatory molecules CD40, CD40L, CD80 or CD86 required for effector T cell induction[28,39]. The absence of co-stimulatory molecules is a significant observation. It implies that any residual engagement of the T cell receptor on Th cells would result in anergy and contribute to tolerance rather than allogeneic responses. Although this is a comforting scenario, based largely on in vitro studies, it cannot fully explain the evasion of alloreactivity demonstrated by MSC. Experiments involving allogeneic co-cultures or MLR have demonstrated that both cell-cell contact and action by soluble factors contribute to the immunomodulatory function of MSC[25,41-43]. Thus it is likely that evasion of alloreactivity is a result of both MSC hypoimmunogenicity, modulation of T cell immune induction and the creation of a suppressive milieu around MSC. Although the mechanisms governing the suppressive effect are not fully understood, several studies have given indicators to the processes involved.
MSC interfere with DC maturation and function
Dendritic cells (DC) are the most influential APC, playing a key role in directing cellular and humoral immune responses against self and non-self antigens [44]. DC contribute to the establishment of tolerance, especially in the periphery[45]. Immature DC are not fully differentiated to carry out their known roles as inducers of immunity[45]. Despite this, immature DC circulate through tissues and the lymph system, capturing self and non-self antigens[45]. Immature DC that are loaded with antigen can silence T cells by deletion or by expanding regulatory T cell populations[45,46]. It has long been believed that this process contributes to graft survival during transplantation [14]. The capacity of DC to induce peripheral tolerance is a potential mechanism by which MSC could manipulate immunity in order to escape T cell recognition. Thus MSC could prevent normal allogeneic responses either through modulation of DC function or by direct effects on T cells. Indications from different studies encourage this hypothesis. Zhang et al [24] provides evidence that MSC interfere with DC maturation. Co-culture experiments showed that MSC down-regulate CD1a, CD40, CD80, CD86, and HLA-DR expression during DC maturation[24]. This is also shown by Beyth et al. [42], who suggest that human MSC converted APC into an inhibitory or suppressor phenotype via cell-to-cell contact, thus locking DC into a semi-mature state and thereby inducing peripheral tolerance. Their findings also show reduced IFN-γ, IL-12 and TNF-α in human MSC/monocyte co-culture [42]. Similarly Jiang et al reported that MSC maintain DC in an immature state[26] and show that MSC inhibit up regulation of IL-12p70 [26]. These results suggest that MSC mediate allogeneic tolerance by directing APC towards a suppressor or inhibitory phenotype that results in an attenuated or regulatory T cell response.
MSC modulate CD4+ T cell responses
Evidence has emerged that MSC interact directly with T cells to suppress alloreativity[25]. Krampera et al showed that MSC impair T cell contact with APC in a non-cognate but transient fashion[25]. This supported work from Bartholomew et al showing that the addition of IL-2 to MLR/MSC co-cultures reduced MSC suppression and restored T cell proliferation[19]. Taken together, these results strongly support a role for either a direct (T cell phenotype) or indirect (DC phenotype) mechanism of immune modulation directed by MSC.
MSC modulation of CD4+ T cell responses is more extensive than the straightforward effect described above. The regular process of antigen specific CD4+ T cell induction requires antigen capture and processing by DC (or other amenable cells), followed by a process of maturation and trafficking to local lymph nodes[14,47-49]. There is evidence that MSC prevent normal allogeneic responses by directing CD4+ T cells to a suppressive or counter-regulatory phenotype[46,50]. Di Nicola et al, showed that MSC strongly suppressed CD4+ (and CD8+) T cells in MLR[43], findings supported by Tse et al, who showed that MSC suppress the proliferation of T-cell subsets[28]. Studies of T cell differentiation have shown that in the presence of human MSC, Th1 cell secretion of IFN-γ dropped by 50% compared to cultures without MSC. Conversely, effector T cells undergoing Th2 differentiation when co-cultured with human MSC showed a significant increase in IL-4 production compared to controls[21]. These findings suggest that MSC exert a counter regulatory, anti-inflammatory role by directing cytokine-mediated immunity[21].
A strategy of regulation and deletion of specific T cells is an effective control against unwanted immune responsiveness especially after transplantion[51]. Consequently, enormous interest has focused on the possibility of Treg cells as a marker for T cell tolerance during transplantation. Treg can act directly on other T cells or indirectly through APC[46]. Aggrawal et al, demonstrated that CD4+ CD25+ T reg populations increased significantly in MLR when MSC were present compared to controls[21]. However, data exists showing that human MSC-mediated inhibition is not suppressed by removing T reg cells from co-cultures [25,42]. Nevertheless a role for these cells can not be excluded, it is possible that an incomplete replication of the suppressive microenvironment in vitro or indeed the diversity of Treg cell populations mean that these studies do not fully explore the potential role of suppressive or regulatory T cells in promoting MSC tolerance.
MSC influence control over cell division cycle pathways in cells of immunological relevance. Glennie et al have shown that T cells stimulated in co-cultures with MSC exhibit an extensive inhibition of cyclin D2 and upregulation of the cyclin dependent kinase inhibitor p27kip1 [52]. As T cell inhibition could not be reversed, these cells were not interpreted as anergic in the classical sense. The authors suggest that MSC are most likely inducing the alternative condition of divisional arrest anergy in T cells, an occurrence usually associated with CTLA-4 signalling[53]. In addition, removal of MSC from the system only restored IFN-γ production but not T cell proliferation[52]. This suggests that MSC induce a condition similar to split anergy[54] or split tolerance[55,56]. The key point is that this work demonstrates that MSC exert veto effects on T cells and it is significant in demonstrating that the mechanisms inducing MSC tolerance are not confined to patterns of cytokine secretion but extend to direct modulation of T cell division.
MSC modulate CD8+ T cell and NK cell activity
The impact of MSC on CD8+ CTL and NK cells has also been addressed. CTL can lyse allogeneic cells after recognition of cognate alloantigen, by the release of cytotoxic effectors such as, perforins, serine esterases, IFN-γ and TNF-α [57] whereas NK cells do not require antigen processing[58]. Consequently both effector cells can operate in tandem, with NK cells providing a first line defence killing target cells that escape CTL recognition or show inadequate expression of self-MHC[58]. There is evidence that MSC inhibit the formation of CTL and appear to evade NK cell targeting mechanisms. Djouad et al showed that CD8+ cells are suppressed by MSC in MLR[41]. Rasmusson supported these findings and further showed that NK cells in co-culture did not recognize MSC although lytic capability was still present[59]. This effect appeared to be mediated by soluble factors[50,59]. Thus MSC interact and suppress cell-mediated immune responses directly and through soluble factors. The targets for this suppression are DC, CD4+ Th, CD8+ CTL and NK cells; in effect MSC silence each aspect of the cellular rejection process.
MSC secrete soluble factors to create an immunosuppressive milieu
The characterisation of cytokines produced by MSC is still provisional and is hindered by the lack of standardisation in isolation and culture conditions, which have given rise to multiple findings and interpretations. It is evident that MSC do not constitutively express IL-2, IL-3, IL-4 and IL-5[60,61]. However some reports show that MSC do constitutively express mRNA for cytokines such as interleukin (IL)-6, -7, -8, -11, -12, -14, -15, -27, leukaemia inhibitory factor, macrophage colony-stimulating factor, and stem cell factor[62,63]. Some of these cytokines provide critical cell-cell interactions and promote HSC differentiation, however caution should be exercised before over interpreting these findings. Protein secretion does not always mirror mRNA levels and most workers in the field would adopt a more conservative profile of cytokine and growth factor production by MSC.
Despite these caveats, certain MSC secreted products such as Hepatocyte growth factor, (HGF) are likely to contribute to creating a local immunosuppressive environment. HGF induces mitogenic and antiapoptotic activity in different systems [64-66] and has a well-characterized role in wound repair [66-68], effects that are consistent with a role for MSC in regenerative medicine. Although some groups do not detect HGF in MSC co-cultures [41] more reports suggest that HGF is constitutively expressed by MSC [13,43,69,70]. Indications that MSC produce HGF [13,43,69,70] encourage a role for these cells in tissue repair [70]. Studies by Chunmeng et al, demonstrated that rat dermal derived "multipotent" cells secrete HGF and promote wound healing[68]. Interestingly, Azuma et al, showed that HGF treatment prevents chronic allograft nephropathy in rats[71]. Taken together these results suggest that HGF may contribute to the ability of MSC to avoid allorejection.
IL-10 has a well-documented role in T cell regulation and in the promotion of a "regulatory" or suppressor phenotype. In our hands human MSC constitutively produce IL-10 whereas Rasmusson et al and Beyth et al only detected IL-10 in co-culture experiments [42,72]. In either case, IL-10 is likely to be suppressing potential allo-responsiveness because it is a recognized growth factor for regulatory T cells [73]. IL-10 can antagonize IL-12 during induction of inflammatory immune responses [74-79]. This is supported by studies showing that MSC partially mediate suppression through IL-10 secretion in MLR cultures[42,72]. Similarly transforming growth factor (TGF)-β1 also plays a role in T cell suppression. This cytokine as well as IL-10 influences cell lineages broader than lymphocytes [74,80,81]. However constitutive expression of TGF-β1 has not been detected from our own studies on human MSC[13]. This is in line with Le Blanc who found no difference in TGF-β1 concentration in co-cultures with or without MSC [69]. In contrast Beyth et al showed that TGF-β1 was secreted in media from co-cultures of human MSC and immune cells but again co-culture did not augment TGF-β1 concentration[42]. Although a number of studies suggest no role for TGF-β1 in evasion of allogeneic responsiveness[42,69,72], it has been suggested that HGF in combination with TGF-β promotes the allo-escaping phenotype[43]. Di Nicola et al showed that neutralizing antibodies to HGF and TGF-β restored the proliferative response in MLR, suggesting that these factors are at least partially responsible[43].
MSC constitutively express the eicosanoid Prostaglandin E (PGE)-2 [82]. This may be upregulated in co-culture[21,28] or downregulated on differentiation[82]. PGE-2 influences numerous immune functions including suppression of B cell activation[83] and induction of regulatory T cells[84]. Although there is evidence for PGE-2 secretion by MSC, there is controversy surrounding a role for PGE-2 as a mediator for suppression of alloresponses in MLR. Studies from Tse, suggested that PGE-2 is not a significant component of suppression[28]. Supporting these findings Rasmusson et al showed that blocking PGE-2 production did not restore allogeneic MLR responses but did influence mitogen driven proliferation[72]. Although the present opinions are conflicting, it should be highlighted that other possible prostaglandins and eicosanoids could be influencing alloresponses[85]. Analysis of these other immunomodulatory molecules could provide further clues as to how MSC escape the immune system.
In contrast to immunosuppression through the secretion of soluble factors, suppression may be mediated by withdrawal of factors in the micro-environment necessary for active immune responses. Indoleamine 2,3-dioxygenase (IDO) is an enzyme that catabolizes L-Tryptophan, thereby depleting an essential amino acid from the local environment [86-89]. Recent evidence has shown that this mechanism is exploited by the mammalian fetal allograft to suppresses T cell activity and prevent rejection [86-89]. Although not a soluble factor, the expression of IDO may contribute to a tolergenic environment. This is of great relevance and has obvious parallels with MSC. Meisel et al showed that IDO is not constitutively expressed by MSC but can be induced by IFN-γ[90], thereby inhibiting allogeneic T cell responses by Tryptophan depletion[90]. Other findings have suggested that IDO-mediated tryptophan depletion inhibits allogeneic T-cell responses by multiple pathways[91]. The discovery of this mechanism, which shows parallels to the creation of a "Tryptophan desert" at the materno-fetal interface[13], provides a further feasible mechanism by which MSC avoid alloreactivity. However, IDO expression is not essential to the maintenance of tolerance against MSC. Tse et al showed that an IDO inhibitor or supplementary Tryptophan addition to MLR did not restore PBMC proliferation [28].
MSC control surface marker expression to exhibit a hypoimmunogenic or tolerogenic phenotype. MSC can also modulate T cell induction directly or via DC and secrete a battery of immunosuppressive factors. It is apparent that the question facing the application of regenerative medicine is no longer "how do MSC escape alloreactivity?" but rather "what is the hierarchy of signals that control immunosuppression?" In this regard, research from other fields has been informative. We have previously proposed that maternal acceptance of the fetal allograft provides indicators of how this process is controlled[13]. However, insight could also come from another avenue of inquiry. The mechanisms of tumor evasion may reflect the survival mechanisms of MSC.
MSC avoidance of alloreactivity shows parallels to tumor evasion
Escape from immune surveillance is believed to be a primary feature of malignant disease in humans. The immune effector response is sub-optimal because tumors develop multifactorial strategies to escape immune deletion[92,93]. These strategies may provide clues to how MSC promote tolerogenic mechanisms during allogeneic engraftment (Fig. 3). Modulation of tumor antigen expression, particularly MHC class I and II is a particularly common component of tumor immune evasion[93]. This is often accompanied by poor or non-expression of co-stimulatory molecules, which not only limits clonal expansion of tumor-specific CD4+ T cells, but also hinders the production of cytokines, and the development of CTL[44,94,95]. Similarly MSC show no expression of co-stimulatory molecules (Fig. 2) [28,39]. In addition to reduced immunogenicity, tumor cells can directly modulate DC and T cell function. Studies from patients with hepatocellular carcinoma showed that neoplasia induced a defect of DC maturation[96]. This parallels findings by Beyth et al [42] suggesting that human MSCs interfere with normal APC maturation, thereby indirectly influencing T-cell activation. Freshly isolated tumor-infiltrating T cells are usually inactive against autologous cancer cells but can be reactivated in-vitro by the addition of IL-2[97]. Studies of MSC by Le Blanc et al showed striking parallels to this form of suppression[69]. They suggest that MSC act by preventing expression of CD25 (IL-2 receptor) thereby limiting T cell activation[69]. Other work has shown that exogenous IL-2 addition to co-cultures containing MSC reversed the suppressive effect[19,69]. Similarly, antigen-specific CD4+ CD25+ regulatory T cells also suppress tumor-specific CD8 T cell cytotoxicity although this mechanism relies on TGF-β secretion by regulatory cells[98,99].
Figure 3 MSC and tumor cells create a suppressive microenvironment. There are fundamental differences between tumor cells (A) and MSC (B) with respect to control of cell division, however many mechanisms exploited by the former to evade immune deletion are also used by MSC to avoid allogeneic rejection. Details of mechanisms and associated references are supplied in the body of the text and Table 1.
Tumors can suppress CD4+ T cell activity and CTL tumor lysis directly through secretion of immunosuppressive factors including TGF-β1 but also PGE-2, and IL-10. Van der Pouw Kraan et al, showed that tumor-derived prostaglandins increased the production of inhibitory cytokines such as IL-10, while suppressing IL-12[100], which is necessary for effective host-cell-mediated anti-tumor immune response[75,93]. Likewise, TGF-β production has been reported from a number of tumors, contributing to immune evasion. Intriguingly in this context it also inhibits CTL differentiation [101]. Although there is little evidence that MSC secrete TGF-β1, the bone marrow is rich in this cytokine, suggesting that MSC reside in a compartment with immunosuppressive qualities.
Although there are striking parallels between MSC and some tumor cells, it is not our contention that these cells are directly related. Indeed there are distinct differences between the populations (Table 1). The fundamental difference between the cell types resides in the control of cell division and apoptosis, which are tightly regulated in MSC but dysregulated in transformed cells. Furthermore, it is well documented that some tumors exploit FasL (CD95L) expression to facilitate immune escape [102-104]. However, our own studies show that human MSC do not express FasL (Fig 2) and although there is some evidence from immortalized mini-pig derived MSC to indicate a role for FasL in suppression[105], it seems that direct induction of apoptotic deletion is not a factor involved in MSC interaction with T cells in the broader literature. The parallels between neoplastic cells and MSC lie in the expressed phenotypes rather than in any direct lineage relation. It appears that MSC retain certain aspects of the fetal allograft that promote tolerance, some of these mechanisms may be reactivated in neoplasia, the key difference being that MSC perform these functions in an ordered and controlled way whereas tumor cells do so in a manner that by definition has escaped normal controls on apoptosis or cell division.
Table 1 Comparison of MSC and Tumor cellsa
Characteristic MSC Tumor cells References
Cell Division Controlled Uncontrolled [5, 7, 111]
MHC I expression + Variable Fig 2 & [25, 27, 28, 39, 93, 111, 112]
MHC II expression - Variable Fig 2 & [2, 25, 27, 39, 93, 111, 112]
CD80 expression - - [25, 28, 39, 44, 94, 95]
CD86 expression - - Fig 2 & [25, 28, 39, 44, 94, 95]
FasL expression - + Fig. 2 & [102-104]
Prostaglandin secretion + + [21, 28, 82, 100]
IDO expression + Variable [28, 43, 59, 87, 90]
TGF-β secretion Variable + [42, 43, 59, 101, 105]
IL-10 secretion + + [13, 42, 72, 100]
DC modulation + + [24, 26, 42, 96]
Veto effects on T cells + + [23, 112]
a Descriptions of MSC in the literature are diverse and many populations have been described which show different patterns of expression. In particular work in mice appears to be strain dependent, but further variation arises from differences in isolation, culture, timing and methodology. Likewise the characteristics of neoplastic cells will vary greatly between different tumors. This table lists those characteristics where at least some cells from each diverse population show either comparative or contrasting features.
Conclusion
Current research on the interaction between MSC and T cells support the potential use of allogeneic MSC in regenerative medicine. Studies showing enhanced MSC engraftment of bone, muscle, heart etc encourage the translation of recent research into therapy. The future holds much promise for the use of allogeneic MSC and whilst obstacles exist, the potential for alloreactivity does not seem to be a major problem. From the research standpoint, MSC appear to use a surprising array of mechanisms to avoid deletion by the host including hypoimmunogenicity, modulation of DC and T cell function, as well as the creation of a suppressive microenvironment. The challenge is now to unravel the timing and control of these mechanisms in an inflammatory situation typical of the recipient patient.
List of Abbreviations
APC, antigen presenting cells; DC, dendritic cell; ES, embryonic stem; HGF, hepatocyte growth factor; HSC, hematopoietic stem cells; IDO, indoleamine 2,3,dioxygenase; KIR, killer inhibitory receptor; MLR, mixed lymphocyte-like reaction; MSC, mesenchymal stem cells; OI, osteogenesis imperfecta; PBMC, peripheral blood mononuclear cells; PGE-2, prostaglandin E2.
Competing interests
JMR and BPM have no competing interests. FPB and JMM have received salary from an organization and hold stocks or shares in an organization that may gain or lose financially from the publication of this manuscript.
Authors' contributions
FPB and BPM conceived the review; JMR performed the microscopy and flow cytometry. All authors provided interpretation of published stem cell data, and have made intellectual contributions to the content of the paper. All authors read and approved the final manuscript.
Supplementary Material
Additional File 1
Ryan et al library.
Click here for file
Acknowledgements
This work was supported by the Science Foundation Ireland Centres for Science Engineering and Technology (CSET) funding of the Regenerative Medicine Institute (REMEDI). Bernard Mahon is a Wellcome Trust/HRB "New Blood" Fellow. Ms Karen English is thanked for assistance in preparation of this manuscript.
==== Refs
Mosca JD Hendricks JK Buyaner D Davis-Sproul J Chuang LC Majumdar MK Chopra R Barry F Murphy M Thiede MA Junker U Rigg RJ Forestell SP Bohnlein E Storb R Sandmaier BM Mesenchymal stem cells as vehicles for gene delivery Clin Orthop Relat Res 2000 S71 90 11039755
Barry FP Murphy JM Mesenchymal stem cells: clinical applications and biological characterization Int J Biochem Cell Biol 2004 36 568 584 15010324
Drukker M Benvenisty N The immunogenicity of human embryonic stem-derived cells Trends Biotechnol 2004 22 136 141 15036864
Shizuru JA Negrin RS Weissman IL Hematopoietic stem and progenitor cells: Clinical and Preclinical Regeneration of the Hematolymphoid System Annu Rev Med 2005 56 509 538 15660525
Pittenger MF Mackay AM Beck SC Jaiswal RK Douglas R Mosca JD Moorman MA Simonetti DW Craig S Marshak DR Multilineage potential of adult human mesenchymal stem cells Science 1999 284 143 147 10102814
Friedenstein AJP Petrokova KV Osteogenesis in transplants of bone marrow cells Journal of Embyological Experimental Morphology 1966 16 381 390
Jiang Y Jahagirdar BN Reinhardt RL Schwartz RE Keene CD Ortiz-Gonzalez XR Reyes M Lenvik T Lund T Blackstad M Du J Aldrich S Lisberg A Low WC Largaespada DA Verfaillie CM Pluripotency of mesenchymal stem cells derived from adult marrow Nature 2002 418 41 49 12077603
Reyes M Verfaillie CM Characterization of multipotent adult progenitor cells, a subpopulation of mesenchymal stem cells Ann N Y Acad Sci 2001 938 231 233 11458512
Jorgensen C Gordeladze J Noel D Tissue engineering through autologous mesenchymal stem cells Curr Opin Biotechnol 2004 15 406 410 15464369
Jorgensen C Djouad F Apparailly F Noel D Engineering mesenchymal stem cells for immunotherapy Gene Ther 2003 10 928 931 12732877
Pawelec G Rehbein A Schlotz E Friccius H Pohla H Cytokine modulation of TH1/TH2 phenotype differentiation in directly alloresponsive CD4+ human T cells Transplantation 1996 62 1095 1101 8900309
Wood KJ Sakaguchi S Regulatory T cells in transplantation tolerance Nat Rev Immunol 2003 3 199 210 12658268
Barry FP Murphy JM English K Mahon BP Immunogenicity of adult mesenchymal stem cells: lessons from the fetal allograft. Stem cells and development 2005 in press
Niederkorn JY Peeler JS Ross J Callanan D The immunogenic privilege of corneal allografts Reg Immunol 1989 2 117 124 2701814
van den Eynde B Gaugler B van der Bruggen P Coulie P Brichard V Boon T Human tumor antigens recognised by T cells: perspectives for new cancer vaccines Biochemical Society Transactions 1995 23 681 686 8566443
Koc ON Day J Nieder M Gerson SL Lazarus HM Krivit W Allogeneic mesenchymal stem cell infusion for treatment of metachromatic leukodystrophy (MLD) and Hurler syndrome (MPS-IH) Bone Marrow Transplant 2002 30 215 222 12203137
Koc ON Gerson SL Cooper BW Dyhouse SM Haynesworth SE Caplan AI Lazarus HM Rapid hematopoietic recovery after coinfusion of autologous-blood stem cells and culture-expanded marrow mesenchymal stem cells in advanced breast cancer patients receiving high-dose chemotherapy J Clin Oncol 2000 18 307 316 10637244
Horwitz EM Prockop DJ Gordon PL Koo WW Fitzpatrick LA Neel MD McCarville ME Orchard PJ Pyeritz RE Brenner MK Clinical responses to bone marrow transplantation in children with severe osteogenesis imperfecta Blood 2001 97 1227 1231 11222364
Bartholomew A Sturgeon C Siatskas M Ferrer K McIntosh K Patil S Hardy W Devine S Ucker D Deans R Moseley A Hoffman R Mesenchymal stem cells suppress lymphocyte proliferation in vitro and prolong skin graft survival in vivo Exp Hematol 2002 30 42 48 11823036
Saito T Kuang JQ Bittira B Al-Khaldi A Chiu RC Xenotransplant cardiac chimera: immune tolerance of adult stem cells Ann Thorac Surg 2002 74 19 24 12118756
Aggarwal S Pittenger MF Human mesenchymal stem cells modulate allogeneic immune cell responses Blood 2005 105 1815 1822 15494428
Deng W Han Q Liao L Li C Ge W Zhao Z You S Deng H Zhao RC Allogeneic bone marrow-derived flk-1+Sca-1- mesenchymal stem cells leads to stable mixed chimerism and donor-specific tolerance Exp Hematol 2004 32 861 867 15345288
Potian JA Aviv H Ponzio NM Harrison JS Rameshwar P Veto-like activity of mesenchymal stem cells: functional discrimination between cellular responses to alloantigens and recall antigens J Immunol 2003 171 3426 3434 14500637
Zhang W Ge W Li C You S Liao L Han Q Deng W Zhao RC Effects of Mesenchymal Stem Cells on Differentiation, Maturation, and Function of Human Monocyte-Derived Dendritic Cells Stem Cells Dev 2004 13 263 271 15186722
Krampera M Glennie S Dyson J Scott D Laylor R Simpson E Dazzi F Bone marrow mesenchymal stem cells inhibit the response of naive and memory antigen-specific T cells to their cognate peptide Blood 2003 101 3722 3729 12506037
Jiang XX Zhang Y Liu B Zhang SX Wu Y Yu XD Mao N Human mesenchymal stem cells inhibit differentiation and function of monocyte-derived dendritic cells Blood 2005 in press
Le Blanc K Tammik C Rosendahl K Zetterberg E Ringden O HLA expression and immunologic properties of differentiated and undifferentiated mesenchymal stem cells Exp Hematol 2003 31 890 896 14550804
Tse WT Pendleton JD Beyer WM Egalka MC Guinan EC Suppression of allogeneic T-cell proliferation by human marrow stromal cells: implications in transplantation Transplantation 2003 75 389 397 12589164
Carlin LM Eleme K McCann FE Davis DM Intercellular transfer and supramolecular organization of human leukocyte antigen C at inhibitory natural killer cell immune synapses J Exp Med 2001 194 1507 1517 11714757
Onfelt B Nedvetzki S Yanagi K Davis DM Cutting edge: Membrane nanotubes connect immune cells J Immunol 2004 173 1511 1513 15265877
Vanherberghen B Andersson K Carlin LM Nolte-'t Hoen EN Williams GS Hoglund P Davis DM Human and murine inhibitory natural killer cell receptors transfer from natural killer cells to target cells Proc Natl Acad Sci U S A 2004 101 16873 16878 15550544
Ruggeri L Capanni M Martelli MF Velardi A Cellular therapy: exploiting NK cell alloreactivity in transplantation Curr Opin Hematol 2001 8 355 359 11604575
Hunt JS Petroff MG Morales P Sedlmayr P Geraghty DE Ober C HLA-G in reproduction: studies on the maternal-fetal interface Hum Immunol 2000 61 1113 1117 11137215
Ristich V Liang S Zhang W Wu J Horuzsko A Tolerization of dendritic cells by HLA-G Eur J Immunol 2005 35 1133 1142 15770701
Moretta L Moretta A Killer immunoglobulin-like receptors Curr Opin Immunol 2004 16 626 633 15342010
Parham P Killer cell immunoglobulin-like receptor diversity: balancing signals in the natural killer cell response Immunol Lett 2004 92 11 13 15081521
Gomez-Lozano N de Pablo R Puente S Vilches C Recognition of HLA-G by the NK cell receptor KIR2DL4 is not essential for human reproduction Eur J Immunol 2003 33 639 644 12616484
Gotherstrom C Ringden O Tammik C Zetterberg E Westgren M Le Blanc K Immunologic properties of human fetal mesenchymal stem cells Am J Obstet Gynecol 2004 190 239 245 14749666
Majumdar MK Keane-Moore M Buyaner D Hardy WB Moorman MA McIntosh KR Mosca JD Characterization and functionality of cell surface molecules on human mesenchymal stem cells J Biomed Sci 2003 10 228 241 12595759
Devine SM Hoffman R Role of mesenchymal stem cells in hematopoietic stem cell transplantation Curr Opin Hematol 2000 7 358 363 11055509
Djouad F Plence P Bony C Tropel P Apparailly F Sany J Noel D Jorgensen C Immunosuppressive effect of mesenchymal stem cells favors tumor growth in allogeneic animals Blood 2003 102 3837 3844 12881305
Beyth S Borovsky Z Mevorach D Liebergall M Gazit Z Aslan H Galun E Rachmilewitz J Human mesenchymal stem cells alter antigen-presenting cell maturation and induce T-cell unresponsiveness Blood 2005 105 2214 2219 15514012
Di Nicola M Carlo-Stella C Magni M Milanesi M Longoni PD Matteucci P Grisanti S Gianni AM Human bone marrow stromal cells suppress T-lymphocyte proliferation induced by cellular or nonspecific mitogenic stimuli Blood 2002 99 3838 3843 11986244
Guinan EC Gribben JG Boussiotis VA Freeman GJ Nadler LM Pivotal role of the B7:CD28 pathway in transplantation tolerance and tumor immunity Blood 1994 84 3261 3282 7524733
Steinman RM Nussenzweig MC Avoiding horror autotoxicus: the importance of dendritic cells in peripheral T cell tolerance Proc Natl Acad Sci U S A 2002 99 351 358 11773639
Mills KH Regulatory T cells: friend or foe in immunity to infection? Nat Rev Immunol 2004 4 841 855 15516964
Steinman RM The dendritic cell system and its role in immunogenicity Annu Rev Immunol 1991 9 271 296 1910679
Steinman RM Inaba K Turley S Pierre P Mellman I Antigen capture, processing, and presentation by dendritic cells: recent cell biological studies Hum Immunol 1999 60 562 567 10426272
Mahon BP Katrak K Nomoto A Macadam AJ Minor PD Mills KHG Poliovirus-specific CD4+ Th1 clones with both cytotoxic and helper activity mediate protective humoral immunity against a lethal poliovirus infection in transgenic mice expressing the human poliovirus receptor J Exp Med 1995 181 1285 1292 7699320
Thompson C Powrie F Regulatory T cells Curr Opin Pharmacol 2004 4 408 414 15251137
Wood KJ Jones ND Bushell AR Morris PJ Alloantigen-induced specific immunological unresponsiveness Philos Trans R Soc Lond B Biol Sci 2001 356 665 680 11375070
Glennie S Soeiro I Dyson PJ Lam EW Dazzi F Bone marrow mesenchymal stem cells induce division arrest anergy of activated T cells Blood 2005 105 2821 2827 15591115
Wells AD Walsh MC Bluestone JA Turka LA Signaling through CD28 and CTLA-4 controls two distinct forms of T cell anergy J Clin Invest 2001 108 895 903 11560959
Otten GR Germain RN Split anergy in a CD8+ T cell: receptor-dependent cytolysis in the absence of interleukin-2 production Science 1991 251 1228 1231 1900952
Nash AA Ashford NP Split T-cell tolerance in herpes simplex virus-infected mice and its implication for anti-viral immunity Immunology 1982 45 761 767 6279490
Martinez C Smith JM Split Tolerance Nature 1964 202 508 509 14167847
Masson D Tschopp J A family of serine esterases in lytic granules of cytolytic T lymphocytes Cell 1987 49 679 685 3555842
Ljunggren HG Karre K In search of the 'missing self': MHC molecules and NK cell recognition Immunol Today 1990 11 237 244 2201309
Rasmusson I Ringden O Sundberg B Le Blanc K Mesenchymal stem cells inhibit the formation of cytotoxic T lymphocytes, but not activated cytotoxic T lymphocytes or natural killer cells Transplantation 2003 76 1208 1213 14578755
Dormady SP Bashayan O Dougherty R Zhang XM Basch RS Immortalized multipotential mesenchymal cells and the hematopoietic microenvironment J Hematother Stem Cell Res 2001 10 125 140 11276366
Zhang Y Li CD Jiang XX Li HL Tang PH Mao N Comparison of mesenchymal stem cells from human placenta and bone marrow Chin Med J (Engl) 2004 117 882 887 15198892
Haynesworth SE Baber MA Caplan AI Cytokine expression by human marrow-derived mesenchymal progenitor cells in vitro: effects of dexamethasone and IL-1 alpha J Cell Physiol 1996 166 585 592 8600162
Silva WAJ Covas DT Panepucci RA Proto-Siqueira R Siufi JL Zanette DL Santos AR Zago MA The profile of gene expression of human marrow mesenchymal stem cells Stem Cells 2003 21 661 669 14595126
Kuroiwa T Kakishita E Hamano T Kataoka Y Seto Y Iwata N Kaneda Y Matsumoto K Nakamura T Ueki T Fujimoto J Iwasaki T Hepatocyte growth factor ameliorates acute graft-versus-host disease and promotes hematopoietic function J Clin Invest 2001 107 1365 1373 11390418
Taniguchi F Harada T Deura I Iwabe T Tsukihara S Terakawa N Hepatocyte growth factor promotes cell proliferation and inhibits progesterone secretion via PKA and MAPK pathways in a human granulosa cell line Mol Reprod Dev 2004 68 335 344 15112327
Xin X Yang S Ingle G Zlot C Rangell L Kowalski J Schwall R Ferrara N Gerritsen ME Hepatocyte growth factor enhances vascular endothelial growth factor-induced angiogenesis in vitro and in vivo Am J Pathol 2001 158 1111 1120 11238059
Ono I Yamashita T Hida T Jin HY Ito Y Hamada H Akasaka Y Ishii T Jimbow K Local administration of hepatocyte growth factor gene enhances the regeneration of dermis in acute incisional wounds J Surg Res 2004 120 47 55 15172189
Chunmeng S Tianmin C Yongping S Xinze R Yue M Jifu Q Shufen L Hui X Chengji L Effects of dermal multipotent cell transplantation on skin wound healing J Surg Res 2004 121 13 19 15313369
Le Blanc K Rasmusson I Gotherstrom C Seidel C Sundberg B Sundin M Rosendahl K Tammik C Ringden O Mesenchymal stem cells inhibit the expression of CD25 (interleukin-2 receptor) and CD38 on phytohaemagglutinin-activated lymphocytes Scand J Immunol 2004 60 307 315 15320889
Neuss S Becher E Woltje M Tietze L Jahnen-Dechent W Functional expression of HGF and HGF receptor/c-met in adult human mesenchymal stem cells suggests a role in cell mobilization, tissue repair, and wound healing Stem Cells 2004 22 405 414 15153617
Azuma H Takahara S Matsumoto K Ichimaru N Wang JD Moriyama T Waaga AM Kitamura M Otsuki Y Okuyama A Katsuoka Y Chandraker A Sayegh MH Nakamura T Hepatocyte growth factor prevents the development of chronic allograft nephropathy in rats J Am Soc Nephrol 2001 12 1280 1292 11373353
Rasmusson I Ringden O Sundberg B Le Blanc K Mesenchymal stem cells inhibit lymphocyte proliferation by mitogens and alloantigens by different mechanisms Exp Cell Res 2005 in press
Asseman C Powrie F Interleukin 10 is a growth factor for a population of regulatory T cells Gut 1998 42 157 158 9536936
Ennis DP Cassidy JP Mahon BP Prior Bordetella pertussis infection modulates allergen priming and the severity of airway pathology in a murine model of allergic asthma Clin Exp Allergy 2004 34 1488 1497 15347385
Mahon BP Ryan M Griffin F McGuirk P Mills KHG IL-12 is produced by macrophages in response to live or killed Bordetella pertussis and enhances the efficacy of an acellular pertussis vaccine by promoting the induction of Th1 cells Infect Immun 1996 64 5295 5301 8945580
Flynn MA Casey DG Todryk SM Mahon BP Efficient delivery of small interfering RNA for inhibition of IL-12p40 expression in vivo Journal of Inflammation 2004 1 4 15813981
Pretolani M Goldman M IL-10: a potential therapy for allergic inflammation. Immunol Today 1997 18 277 280 9190113
Higgins SC Lavelle EC McCann C Keogh B McNeela E Byrne P O'Gorman B Jarnicki A McGuirk P Mills KH Toll-like receptor 4-mediated innate IL-10 activates antigen-specific regulatory T cells and confers resistance to Bordetella pertussis by inhibiting inflammatory pathology J Immunol 2003 171 3119 3127 12960338
McGuirk P Mills KHG Direct anti-inflammatory effect of a bacterial virulence factor:IL-10-dependent suppression of IL-12 production by filamentous haemagglutinin from Bordetella pertussis Eur J Immunol 2000 30 415 422 10671196
Gollnick SO Cheng HL Grande CC Thompson D Tomasi TB Effects of transforming growth factor-beta on bone marrow macrophage Ia expression induced by cytokines J Interferon Cytokine Res 1995 15 485 491 7648451
Ruegemer JJ Ho SN Augustine JA Schlager JW Bell MP McKean DJ Abraham RT Regulatory effects of transforming growth factor-beta on IL-2- and IL-4-dependent T cell-cycle progression J Immunol 1990 144 1767 1776 2407783
Arikawa T Omura K Morita I Regulation of bone morphogenetic protein-2 expression by endogenous prostaglandin E2 in human mesenchymal stem cells J Cell Physiol 2004 200 400 406 15254968
Roper RL Ludlow JW Phipps RP Prostaglandin E2 inhibits B lymphocyte activation by a cAMP-dependent mechanism: PGE-inducible regulatory proteins Cell Immunol 1994 154 296 308 8131204
Akasaki Y Liu G Chung NH Ehtesham M Black KL Yu JS Induction of a CD4+ T regulatory type 1 response by cyclooxygenase-2-overexpressing glioma J Immunol 2004 173 4352 4359 15383564
Colgan SP Lipid mediators in epithelial cell-cell interactions Cell Mol Life Sci 2002 59 754 760 12088276
Munn DH Mellor AL Rossi M Young JW Dendritic cells have the option to express IDO-mediated suppression or not Blood 2005 105 2618 15746089
Munn DH Mellor AL IDO and tolerance to tumors Trends Mol Med 2004 10 15 18 14720581
Mellor AL Munn DH Tryptophan catabolism prevents maternal T cells from activating lethal anti-fetal immune responses J Reprod Immunol 2001 52 5 13 11600174
Mellor AL Munn DH IDO expression by dendritic cells: tolerance and tryptophan catabolism Nat Rev Immunol 2004 4 762 774 15459668
Meisel R Zibert A Laryea M Gobel U Daubener W Dilloo D Human bone marrow stromal cells inhibit allogeneic T-cell responses by indoleamine 2,3-dioxygenase-mediated tryptophan degradation Blood 2004 103 4619 4621 15001472
Frumento G Rotondo R Tonetti M Damonte G Benatti U Ferrara GB Tryptophan-derived catabolites are responsible for inhibition of T and natural killer cell proliferation induced by indoleamine 2,3-dioxygenase J Exp Med 2002 196 459 468 12186838
Strand S Galle PR Immune evasion by tumours: involvement of the CD95 (APO-1/Fas) system and its clinical implications Mol Med Today 1998 4 63 68 9547792
Mitra R Singh S Khar A Antitumour immune responses Expert Rev Mol Med 2003 2003 1 22 14987405
Fujiwara K Higashi T Nouso K Nakatsukasa H Kobayashi Y Uemura M Nakamura S Sato S Hanafusa T Yumoto Y Naito I Shiratori Y Decreased expression of B7 costimulatory molecules and major histocompatibility complex class-I in human hepatocellular carcinoma J Gastroenterol Hepatol 2004 19 1121 1127 15377288
Jung D Hilmes C Knuth A Jaeger E Huber C Seliger B Gene transfer of the Co-stimulatory molecules B7-1 and B7-2 enhances the immunogenicity of human renal cell carcinoma to a different extent Scand J Immunol 1999 50 242 249 10447932
Ninomiya T Akbar SM Masumoto T Horiike N Onji M Dendritic cells with immature phenotype and defective function in the peripheral blood from patients with hepatocellular carcinoma J Hepatol 1999 31 323 331 10453947
Rosenberg SA The development of new immunotherapies for the treatment of cancer using interleukin-2. A review Ann Surg 1988 208 121 135 3041925
Nakamura K Kitani A Fuss I Pedersen A Harada N Nawata H Strober W TGF-beta 1 plays an important role in the mechanism of CD4+CD25+ regulatory T cell activity in both humans and mice J Immunol 2004 172 834 842 14707053
Ranges GE Figari IS Espevik T Palladino MAJ Inhibition of cytotoxic T cell development by transforming growth factor beta and reversal by recombinant tumor necrosis factor alpha J Exp Med 1987 166 991 998 3498791
van der Pouw Kraan TC Boeije LC Snijders A Smeenk RJ Wijdenes J Aarden LA Regulation of IL-12 production by human monocytes and the influence of prostaglandin E2 Ann N Y Acad Sci 1996 795 147 157 8958925
Chang HL Gillett N Figari I Lopez AR Palladino MA Derynck R Increased transforming growth factor beta expression inhibits cell proliferation in vitro, yet increases tumorigenicity and tumor growth of Meth A sarcoma cells Cancer Res 1993 53 4391 4398 8364935
Walker PR Saas P Dietrich PY Role of Fas ligand (CD95L) in immune escape: the tumor cell strikes back J Immunol 1997 158 4521 4524 9144461
O'Connell J O'Sullivan GC Collins JK Shanahan F The Fas counterattack: Fas-mediated T cell killing by colon cancer cells expressing Fas ligand J Exp Med 1996 184 1075 1082 9064324
Niehans GA Brunner T Frizelle SP Liston JC Salerno CT Knapp DJ Green DR Kratzke RA Human lung carcinomas express Fas ligand Cancer Res 1997 57 1007 1012 9067260
Liu J Lu XF Wan L Li YP Li SF Zeng LY Zeng YZ Cheng LH Lu YR Cheng JQ Suppression of human peripheral blood lymphocyte proliferation by immortalized mesenchymal stem cells derived from bone marrow of Banna Minipig inbred-line Transplant Proc 2004 36 3272 3275 15686744
Mackay AM Beck SC Murphy JM Barry FP Chichester CO Pittenger MF Chondrogenic differentiation of cultured human mesenchymal stem cells from marrow Tissue Eng 1998 4 415 428 9916173
Murphy JM Fink DJ Hunziker EB Barry FP Stem cell therapy in a caprine model of osteoarthritis Arthritis Rheum 2003 48 3464 3474 14673997
Thorpe SJ Thein SL Sampietro M Craig JE Mahon BP Huehns ER Immunochemical estimation of haemoglobin types in red blood cells by FACS analysis British Journal of Haematology 1994 87 125 131 7524614
McGuirk P Mahon BP Griffin F Mills KHG Compartmentalization of T cell responses following respiratory infection with Bordetella pertussis: hyporesponsiveness of lung T cells is associated with modulated expression of the costimulatory molecule CD28 European Journal of Immunology 1998 28 153 163 9485195
Ryan M McCathy L Mahon BP Rappuoli R Mills KHG Mechanism of adjuvanticity of pertussis toxin (PT): PT potentiates Th1 and Th2 responses by stimulating regulatory and accessory cytokine secretion and enhancing expression of the co-stimulatory molecules B7-1, B7-2 and CD28 International Immunology 1998 10 651 662 9645613
Yang L Carbone DP Tumor-host immune interactions and dendritic cell dysfunction Adv Cancer Res 2004 92 13 27 15530555
Mapara MY Sykes M Tolerance and cancer: mechanisms of tumor evasion and strategies for breaking tolerance J Clin Oncol 2004 22 1136 1151 15020616
|
16045800
|
PMC1215510
|
CC BY
|
2021-01-04 16:36:23
|
no
|
J Inflamm (Lond). 2005 Jul 26; 2:8
|
utf-8
|
J Inflamm (Lond)
| 2,005 |
10.1186/1476-9255-2-8
|
oa_comm
|
==== Front
J Negat Results BiomedJournal of Negative Results in Biomedicine1477-5751BioMed Central London 1477-5751-4-71614304310.1186/1477-5751-4-7ResearchAn extended association screen in multiple sclerosis using 202 microsatellite markers targeting apoptosis-related genes does not reveal new predisposing factors Gödde René [email protected] Stefanie [email protected] Peter [email protected] Eckhart [email protected] Michael [email protected] Sebastian [email protected]üller Norbert [email protected] Jörg T [email protected] Department of Human Genetics, Ruhr-University, Bochum, Germany2 Department of Neurology, Kliniken Bergmannsheil, Ruhr-University, Bochum, Germany3 Department of Neurology, Knappschaftskrankenhaus, Ruhr-University, Bochum, Germany4 Department of Neurology, St. Josef-Hospital, Ruhr-University, Bochum, Germany5 Department of Transfusion Medicine, Universitätsklinikum Essen, Essen, Germany2005 5 9 2005 4 7 7 10 3 2005 5 9 2005 Copyright © 2005 Gödde et al; licensee BioMed Central Ltd.2005Gödde et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Apoptosis, the programmed death of cells, plays a distinct role in the etiopathogenesis of Multiple sclerosis (MS), a common disease of the central nervous system with complex genetic background. Yet, it is not clear whether the impact of apoptosis is due to altered apoptotic behaviour caused by variations of apoptosis-related genes. Instead, apoptosis in MS may also represent a secondary response to cellular stress during acute inflammation in the central nervous system. Here, we screened 202 apoptosis-related genes for association by genotyping 202 microsatellite markers in initially 160 MS patients and 160 controls, both divided in 4 sets of pooled DNA samples, respectively. When applying Bonferroni correction, no significant differences in allele frequencies were detected between MS patients and controls. Nevertheless, we chose 7 markers for retyping in individual DNA samples, thereby eliminating 6 markers from the list of candidates. The remaining candidate, the ERBB3 gene microsatellite, was genotyped in additional 245 MS patients and controls. No association of the ERBB3 marker with the disease was detected in these additional cohorts. In consequence, we did not find further evidence for apoptosis-related genes as predisposition factors in MS.
==== Body
Introduction
Multiple sclerosis (MS) is among the most common neurological diseases of primarily of young adults [1]. It has predominantly been characterized as a chronic inflammatory disease of the central nervous system (CNS) resulting in myelin and axonal damage and the formation of focal demyelinated plaques. Myelin-reactive T cells enter the CNS via the blood-brain barrier and mediate the observed inflammatory events [2]. While the contribution of dysfunctional elements from the immune system in MS disease development has been widely accepted from the early days of MS research [3-5], the influence of neuronal death and apoptosis in acute inflammatory plaques has partially been disregarded. Yet, recent insights into the pathogenesis of MS suggest miscellaneous impacts for apoptosis in this neurological disorder, although the contribution to disease susceptibility remains elusive.
Predisposition to the disease depends on both genetic and environmental factors, as demonstrated by twin studies [6] and by virtue of the latitude-dependent geographical distribution [7], thus assigning MS to the large family of common multifactorial diseases, at least in the northern hemisphere. Despite the influence of such predisposing factors, the underlying etiopathological mechanisms as well as most genetic factors responsible for the predisposition to MS remain largely undefined. Until now, the only consistent association has been demonstrated with the HLA-DRB1*1501-DQB1*0602 haplotype in MS patients of European descent [8-10].
Apoptosis, the self-controlled death of cells, is a physiological 'suicide programme' leading to selective elimination of specific cells, either because they become dispensable in their tissue environment or harmful through infection, malignant transformation or, in general, mutation. Regarding MS, impaired apoptosis might result in elevated numbers or extended persistence of myelin-reactive T cells in the CNS tissue, enhancing the observed inflammatory processes [11,12]. On the other hand, apoptosis of neuronal cells and their glial chaperones in acute and active MS lesions has recently been demonstrated and may account for most of the disability acquired over time [13-17]. Therefore, when ascertaining candidate genes for MS association studies, factors involved in the regulation and execution of programmed cell death should be considered supplementary to those acting in the dysregulation of the immune system.
We performed an association screen in 202 microsatellite markers in or near to putative MS candidate genes related to apoptosis and the immune system using specifically designed primers and pooled DNA in a case-control design as described previously [18]. Such an 'indirect' approach strictly relies on the presence of linkage disequilibrium (LD) between certain alleles of a microsatellite marker and the corresponding predisposing mutation in the nearby candidate gene. Association was tested by means of contigency tables comparing allele frequencies in MS patients and controls. Subsequently, in case associated markers were found, we performed microsatellite genotyping of individual DNAs, thereby excluding false positive associations resulting from artifact introduced by DNA pooling.
Materials and methods
Patient and control DNA samples
All individuals involved in this study gave written consent for the genetic analyses. Peripheral blood samples from > 600 healthy blood donors were provided by the department of transplantation and immunology of the University hospital Eppendorf (Hamburg, Germany) and the department of transfusion medicine of the University hospital Essen (Essen, Germany). More than 800 unrelated MS patients classified according to the Poser criteria [19] and attending the Departments of Neurology, University clinic of Bochum (Germany), were included. DNA was extracted from peripheral blood leukocytes by standard methods [20]. The quality of each individual DNA was evaluated by separation on 0.7% agarose gels.
DNA pooling
The employment of pooled DNA samples in microsatellite genotyping introduces errors [9], unless pooling is performed absolutely accurately. Concentration of DNA from each individual was quantified in triplicate using spectrophotometric measurement and then diluted to a final 50 ng/μl. After once more verifying these concentrations twice, 40 individual DNAs were combined into a DNA pool of a final concentration of 25 ng/μl. This way, 4 DNA pools were created for MS patients and controls, respectively. Using subpools prevents quantitative errors, as each allele image profile (AIP) of the respective microsatellite is statistically compared to the other subpool of the respective group.
Microsatellite markers
Intragenic microsatellites or, if not available, microsatellites localised in the immediate vicinity (< 50 kb) of the specific gene were included. For all genes represented by microsatellite markers, oligonucleotide sequences, distances to the specific gene, and additional information are presented in the Markers website . Only markers with equal "intra-subgroup" allele distributions with ≥ 2 alleles were included in the subsequent analyses. All significantly associated markers (p ≥ 0.05) were subsequently genotyped individually (see below).
Tailed primer polymerase chain reaction (PCR)
We used a universal fluorescence-labelled tailed oligonucleotide added to the 5' part of the sequence-specific primer for automatic fragment analysis. The tail (5'-CATCGCTGATTCGCACAT-3') was designed to be secondary structure prone, and its sequence was "blasted" against the NCBI human genome database [21] yielding no significant homologies. Gene-specific microsatellites were chosen applying the repeat-masker option of the Santa Cruz genome browser [22]. Primers were designed and adjusted to a melting temperature of 55°C using the Primer Express 2.0 Software (ABI). Amplification was performed using three oligonucleotides: (1) a tailed forward primer (tailed F), (2) a reverse primer and (3) a labelled primer (labelled F) corresponding to the 5'-tail sequence of tailed F. PCR conditions were as follows: 1 × PCR buffer (Qiagen), 1.5 pmol labelled F, 0.2 mM each dNTP, 3 mM MgCl2, 0.2 pmol tailed F, 1.5 pmol reverse primer, 0.25 U Qiagen Hot Start Taq (Qiagen) and 50 ng DNA. PCR reactions were performed with an initial activation step at 95°C for 15 min; 35 cycles of denaturation at 94°C for 1 min, annealing at 55°C for 1 min and extension at 72°C for 1 min; and a final extension at 72°C for 10 min.
Electrophoresis and genotyping
Electrophoresis was performed on a 96-well ABI377 slab-gel system. Aliquots of 1.0 μl PCR product and 2 μl of fluorescent ladder (MegaBACE ET400-R Size Standard; Amersham) were mixed. A 1 μl sample of this mix was loaded onto a 4.5% polyacrylamide (PAA) gel containing 5.625 ml 40% (19:1) PAA, 18 g urea, 5 ml 10× TBE buffer (90 mM Tris-borate, 2 mM EDTA, pH 8.3), 25 ml bidistilled H2O, 30 μl 10% ammoniumpersulphate and 20 μl Tetramethylethylendiamin. Prior to polymerisation, the gel mix was filtered through a 0.2-μm membrane filter. Electrophoreses were run using ABI standard protocols. Raw data were analysed using the Genotyper software (ABI), resulting in a marker-specific AIP. AIPs consist of a series of peaks with different heights that correspond to the respective allele frequency distribution within each analysed DNA pool.
Statistics for comparison of allele frequencies
Association was tested by comparison of the MS and control AIPs. Peak heights were normalized according to the number of expected alleles per pool (n = 80). Averages of each peak (each distinct allele) were calculated according to the total allele count. Alleles with frequencies < 5% were added up and considered as one allele. Case and control distributions for combined MS and control pools, respectively, were subsequently compared statistically by means of contingency tables. Hence, p values are nominal and approximate because of the use of estimated rather than observed counts for allele frequencies. In order to select markers for further investigations, non-corrected p values were ranked according to their evidence for association [23]. Markers showing the most significant differences between MS patients and controls were subsequently chosen for further analysis by individual genotyping.
Individual genotyping
PCR of pooled DNA samples can introduce artifacts that may cause an increased rate of false-positive results, i.e. differences between pools may appear exaggerated. Therefore, the most conspiciously differing markers were genotyped in individual DNA samples of patients and controls, both from the original pools (both n = 160) as well as additional patient and control cohorts (both n = 245) and under similar conditions as used for pool PCRs. Association was analysed by comparison of microsatellite allele frequencies from the MS cohort with the corresponding allele of the control group by chi-square testing.
Results and discussion
The statistical evaluation of 202 microsatellite markers in 160 MS patients and 160 controls combined in 8 DNA pools, each consisting of 40 individuals, respectively, revealed 7 markers with significant differences between allele frequencies of MS patients and controls (Tab. 1).
Table 1 Microsatellites with significant differences (p < 0.05) in allele frequencies between MS patients and controls when screened using pooled DNA samples. No correction for multiple testing was applied here.
Microsatellite p value
NOS1 0.0068
NFκB2 0.0207
FADD 0.0213
GZMB 0.0245
ERBB3 0.0249
NGFβ 0.0292
ADPRT 0.0335
However, except for NOS1, no marker exceeded borderline significance, and Bonferroni correction for multiple testing (n = 202) did eliminate all significant results. Nevertheless, the 4 most promising markers were chosen for further analysis by individual genotyping, thereby excluding possible artifacts introduced via DNA pooling and circumventing the need for massive correction: ERBB3 (V-erb-b2 erythroblastic leukemia viral oncogene homologue 3), NFκB2 (nuclear factor of κ light polypeptide gene enhancer in B-cells 2), NGFβ (nerve growth factor β) and NOS1 (nitric oxide synthase 1). The observed allele frequencies of pooled and individual DNA samples are compared in Fig. 1.
Figure 1 Allele frequencies genotyped in 4 microsatellite markers using pooled (black and white columns) and individual DNA samples (dark grey and light grey). CO: controls; MS: MS patients.
Allele frequencies were counted from individual genotypes and compared statistically according to AIP analysis resulting from pooled DNA using the same 160 MS patients and controls. In case additional alleles were detectable, only those alleles that were observed in both experiments were analysed. The results of the statistical tests are shown in table 2.
Table 2 Relation of p values between analyses based on pooled and individual DNAs using an identical set of 160 MS patient and controls, respectively.
Microsatellite p value (pools) p value (individual DNAs)
ERBB3 0.0249 0.016*
NFκB2 0.0207 0.15
NGFβ 0.0292 0.44
NOS1 0.0068 0.97
*corrected for multiple testing according Bonferroni (n = 4 simultaneous tests)
Apparently, 3 of the 4 comparisons of pooled and individual DNAs show substantial differences. Only the microsatellite near to the ERBB3 gene remained significantly associated when the same DNA samples were retyped individually. For the NFκB2, NGFβ and NOS1 genes, the comparisons of allele frequencies from pooled and individually-typed DNA samples (Fig. 1) show an important and typical artifact [9]. DNA polymerases tend to preferentially amplify short alleles in favour of longer alleles (length-dependent amplification). Therefore, in PCRs based on pooled DNA samples, the shorter alleles of a microsatellite marker will often be over-represented in the resulting PCR product. This effect is most apparent in the marker NOS1, where one of the observed alleles in the pooled experiment obviously results exclusively from the abovementioned effect. Also in NGFβ, the alleles 1 and 2 were significantly over-represented in the screen using pooled DNA, resulting in a false positive association. Only for the ERBB3 gene, the observed allele frequencies in the typing experiment based on pooled DNA adequately correspond to the individually typed frequencies.
In order to validate the association of ERBB3 with MS, we performed genotyping of another cohort of 245 MS patients and controls, respectively (frequencies shown in Fig. 2).
Figure 2 Allele frequencies genotyped in the microsatellite ERBB3 using the originally pooled (black and white columns) and the additional 490 individual DNA samples (dark grey and light grey). CO: controls; MS: MS patients.
Statistical analysis of the latter allele frequency distribution revealed a non significant p value of 0.325. Therefore, the association of the ERBB3 microsatellite could not be confirmed in the additional DNA cohorts of MS patients and controls.
In conclusion, we did not find supporting evidence for involvement of apoptosis-related genes in the predisposition to MS. Nevertheless, such a contribution cannot be excluded based exclusively on our experiments for various reasons. Only a fraction of all apoptosis-related genes has been included in our survey and, therefore, many more genes may represent auspicious candidates. Moreover, as our approach depends solely on the presence of LD between a marker and a predisposing mutation, missing LD between the microsatellite and its corresponding gene will also cause a negative result. As the HapMap-Project [24] progresses rapidly and, therefore, information about the haplotype block structure of the human genome increases substantially, it might soon be possible to reappraise our negative results with respect to the haplotype block structure of the gene under examination.
==== Refs
Poser CM The epidemiology of multiple sclerosis: a general overview Ann Neurol 1994 36 Suppl 2 S180 93 7998787 10.1002/ana.410360805
Hemmer B Cepok S Nessler S Sommer N Pathogenesis of multiple sclerosis: an update on immunology Curr Opin Neurol 2002 15 227 231 12045717 10.1097/00019052-200206000-00001
Iivanainen MV The significance of abnormal immune responses in patients with multiple sclerosis J Neuroimmunol 1981 1 141 172 6461670 10.1016/0165-5728(81)90040-0
Lisak RP Zweiman B Burns JB Rostami A Silberberg DH Immune responses to myelin antigens in multiple sclerosis Ann N Y Acad Sci 1984 436 221 230 6085227
Waksman BH Reynolds WE Multiple sclerosis as a disease of immune regulation Proc Soc Exp Biol Med 1984 175 282 294 6229799
Willer CJ Dyment DA Risch NJ Sadovnick AD Ebers GC Twin concordance and sibling recurrence rates in multiple sclerosis Proc Natl Acad Sci U S A 2003 100 12877 12882 14569025 10.1073/pnas.1932604100
Sotgiu S Pugliatti M Fois ML Arru G Sanna A Sotgiu MA Rosati G Genes, environment, and susceptibility to multiple sclerosis Neurobiol Dis 2004 17 131 143 15474351 10.1016/j.nbd.2004.07.015
Epplen C Jackel S Santos EJ D'Souza M Poehlau D Dotzauer B Sindern E Haupts M Rude KP Weber F Stover J Poser S Gehler W Malin JP Przuntek H Epplen JT Genetic predisposition to multiple sclerosis as revealed by immunoprinting Ann Neurol 1997 41 341 352 9066355 10.1002/ana.410410309
Godde R Nigmatova V Jagiello P Sindern E Haupts M Schimrigk S Epplen JT Refining the results of a whole-genome screen based on 4666 microsatellite markers for defining predisposition factors for multiple sclerosis Electrophoresis 2004 25 2212 2218 15274005 10.1002/elps.200405929
Goedde R Sawcer S Boehringer S Miterski B Sindern E Haupts M Schimrigk S Compston A Epplen JT A genome screen for linkage disequilibrium in HLA-DRB1*15-positive Germans with multiple sclerosis based on 4666 microsatellite markers Hum Genet 2002 111 270 277 12215840 10.1007/s00439-002-0801-8
Ohsako S Elkon KB Apoptosis in the effector phase of autoimmune diabetes, multiple sclerosis and thyroiditis Cell Death Differ 1999 6 13 21 10200543 10.1038/sj.cdd.4400459
Pender MP Genetically determined failure of activation-induced apoptosis of autoreactive T cells as a cause of multiple sclerosis Lancet 1998 351 978 981 9734959
Aktas O Wendling U Zschenderlein R Zipp F [Apoptosis in multiple sclerosis. Etiopathogenetic relevance and prospects for new therapeutic strategies] Nervenarzt 2000 71 767 773 11082807 10.1007/s001150050664
Meyer R Weissert R Diem R Storch MK de Graaf KL Kramer B Bahr M Acute neuronal apoptosis in a rat model of multiple sclerosis J Neurosci 2001 21 6214 6220 11487644
Oren A White LR Aasly J Apoptosis in neurones exposed to cerebrospinal fluid from patients with multiple sclerosis or acute polyradiculoneuropathy J Neurol Sci 2001 186 31 36 11412869 10.1016/S0022-510X(01)00496-8
Pender MP Oligodendrocyte apoptosis before immune attack in multiple sclerosis? Ann Neurol 2005 57 158; author reply 158 9 15622546 10.1002/ana.20352
Zipp F Apoptosis in multiple sclerosis Cell Tissue Res 2000 301 163 171 10928289 10.1007/s004410000179
Jagiello P Gencik M Arning L Wieczorek S Kunstmann E Csernok E Gross WL Epplen JT New genomic region for Wegener's granulomatosis as revealed by an extended association screen with 202 apoptosis-related genes Hum Genet 2004 114 468 477 14968360 10.1007/s00439-004-1092-z
Poser CM Paty DW Scheinberg L McDonald WI Davis FA Ebers GC Johnson KP Sibley WA Silberberg DH Tourtellotte WW New diagnostic criteria for multiple sclerosis: guidelines for research protocols Ann Neurol 1983 13 227 231 6847134 10.1002/ana.410130302
Miller SA Dykes DD Polesky HF A simple salting out procedure for extracting DNA from human nucleated cells Nucleic Acids Res 1988 16 1215 3344216
McGinnis S Madden TL BLAST: at the core of a powerful and diverse set of sequence analysis tools Nucleic Acids Res 2004 32 W20 5 15215342
Karolchik D Baertsch R Diekhans M Furey TS Hinrichs A Lu YT Roskin KM Schwartz M Sugnet CW Thomas DJ Weber RJ Haussler D Kent WJ The UCSC Genome Browser Database Nucleic Acids Res 2003 31 51 54 12519945 10.1093/nar/gkg129
Setakis E Statistical analysis of the GAMES studies J Neuroimmunol 2003 143 47 52 14575913 10.1016/j.jneuroim.2003.08.010
Consortium TIHM The International HapMap Project Nature 2003 426 789 796 14685227 10.1038/nature02168
|
16143043
|
PMC1215511
|
CC BY
|
2021-01-04 16:37:37
|
no
|
J Negat Results Biomed. 2005 Sep 5; 4:7
|
utf-8
|
J Negat Results Biomed
| 2,005 |
10.1186/1477-5751-4-7
|
oa_comm
|
==== Front
Microb Cell FactMicrobial Cell Factories1475-2859BioMed Central London 1475-2859-4-261614454510.1186/1475-2859-4-26ReviewEnsuring safety of DNA vaccines Glenting Jacob [email protected] Stephen [email protected] Bioneer A/S, DK-2970 Hørsholm, Denmark2 Danish Toxicology Centre, DK-2970 Hørsholm, Denmark2005 6 9 2005 4 26 26 25 8 2005 6 9 2005 Copyright © 2005 Glenting and Wessels; licensee BioMed Central Ltd.2005Glenting and Wessels; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
In 1990 a new approach for vaccination was invented involving injection of plasmid DNA in vivo, which elicits an immune response to the encoded protein. DNA vaccination can overcome most disadvantages of conventional vaccine strategies and has potential for vaccines of the future. However, today 15 years on, a commercial product still has not reached the market. One possible explanation could be the technique's failure to induce an efficient immune response in humans, but safety may also be a fundamental issue. This review focuses on the safety of the genetic elements of DNA vaccines and on the safety of the microbial host for the production of plasmid DNA. We also propose candidates for the vaccine's genetic elements and for its microbial production host that can heighten the vaccine's safety and facilitate its entry to the market.
==== Body
Introduction
Vaccination with purified plasmid DNA involves injection of the plasmid into the patient to elicit an immune response to a protein that is encoded on the plasmid [1]. This mini-review focuses upon several aspects of safety of the DNA molecule itself and of the microorganism used to manufacture the DNA. The review is not exhaustive but does raise very important safety issues to be kept in mind early in the development of DNA vaccines.
DNA vaccination was described in a study in 1990 that demonstrated the induction of gene expression following direct intramuscular injection of plasmid DNA in mice [2]. Since then our understanding of the immunological mechanisms behind this unexpected result has increased. This includes identification of immune stimulatory DNA sequences (ISS) that could explain how DNA vaccines can evoke an immune response without an adjuvant [3]. The advantages of DNA vaccines over the traditional attenuated or subunit vaccines are their capacity to induce a broad spectrum of cellular and humoral immune responses, their flexible genetic design and low cost of production in a microbial host. Almost two thousand papers have been published, and several clinical trials have been conducted testing DNA vaccines against infectious diseases such as HIV-1 [4], Ebola virus [5] and malaria [6], or to generate protective immunity against tumors [7]. Despite this extensive research, a commercial product has yet to come to the market. One reason for this may be their failure to induce a strong immune response in higher animals like primates [8]. Another reason for their absence from the market may be related to their safety. Indeed, international regulatory groups have recently questioned the safety of certain existing DNA vaccine constructs and their production systems [9]. While the main focus of research has previously been on their functionality and immunological mechanisms, work on safety aspects most often is put off until later in development. By then, making fundamental changes to the DNA vaccine to improve its safety can be extremely costly and time-consuming.
In the following we propose some basic choices related to safety to be made during the development of DNA vaccines. We highlight safety issues that can be addressed by the appropriate choice of the vaccine's genetic elements, of its microbial production host and of the conditions of manufacture. Special focus will be put on the use of food-grade host-vector systems that are based on our experience with the lactic acid bacterium Lactococcus lactis.
The vaccine's genetic elements
The organization of the genetic elements of a DNA vaccine reflects the plasmid's functionality, its bulk manufacture and its clinical use in the patient. Thus, the plasmid contains one unit responsible for its propagation in the microbial host and another unit that drives the expression of the vaccine gene in the cells of the patient. The genetic elements of the vaccine are shown in Figure 1, and particular safety concerns are listed in Table 1.
Figure 1 Genetic elements of a plasmid DNA vaccine. Plasmid DNA vaccines consists of a unit for propagation in the microbial host and a unit that drives vaccine synthesis in the eukaryotic cells. For plasmid DNA production a replication region and a selection marker are employed. The eukaryotic expression unit comprises an enhancer/promoter region, intron, signal sequence, vaccine gene and a transcriptional terminator (poly A). Immune stimulatory sequences (ISS) add adjuvanticity and may be localized in both units.
Table 1 The safety concerns and possible solutions for plasmid DNA vaccines and their production hosts. A priori each safety concern should be addressed as early in development as possible.
Safety concern Possible solution
Genetic elements Transfer of plasmid to host flora Narrow host-range replication region
Non-antibiotic plasmid marker
Germline integration Avoidance of mammalian replication region
Insertional mutagenesis and oncogenesis Artificial DNA for promoter, intron, and signal sequence
Avoidance of human-homologous DNA
Adverse effects of encoded peptide(s) Artificial signal sequences
Avoidance of mammalian replication region
Evaluation of vaccine peptide case-by-case
Induction of autoimmune reactions Minimized plasmids
Production host Endotoxins and biogenic amines Use of gram-positive organism
Transferable antibiotic resistance genes Determination of minimal inhibitory concentrations (MIC's)
Screening for transferability
Genetic instability Analysis of plasmid population by sequencing and mass spectrometry
Pathogenicity Use of food-grade organism
The unit responsible for plasmid propagation in the microbial host contains a replication region and a selectable marker. The replication region allows the maintenance of multiple copies of the plasmid per host cell and a stable inheritance of the plasmid during bacterial growth. Furthermore, the replication region also determines the plasmid's host-range. Because DNA vaccination involves injection of milligram quantities of plasmid, replication regions with a narrow host-range can reduce the probability for spread of the plasmid to the patient's own flora. A replication region dependent on chromosomally encoded factors restricts the replication to a single host strain. One such bio-containment system has been developed in E. coli based on trans-complementation of a repA- plasmid replication region by a repA+ host strain [10]. Here, the pWV01-derived vectors cannot replicate in the absence of the replication factor RepA and thus relies on a repA+ helper strain. Addition of another ori (origin of replication) region that is active in mammalian cells allows prolonged persistence and expression of the vaccine gene in the transfected tissue. However, uncontrolled expression of the vaccine gene may induce immunological tolerance. Furthermore, persistence and increased spread of the plasmid may lead to germline transmission as a result of transfection of sperm cells or oocytes [11]. In fact, PCR studies have detected vaccine plasmid in the gonads of vaccinated fetuses and in offspring of these fetuses [12]. A literature study has identified non-replicating plasmids as a factor that reduces risk of germline transmission [13]. Accordingly, only prokaryotic and narrow host range replication regions should be present on vaccine plasmids.
Selectable markers ensure stable inheritance of plasmids during bacterial growth (Fig. 1). Most vaccine plasmids rely for this on resistance to antibiotics. Although a powerful selection, resistance genes to antibiotics are discouraged by regulatory authorities [14]. The concern is that the plasmid may transform the patient's microflora and spread the resistance genes (Table 1). Indeed, there is much international scientific and regulatory focus on this issue [15-19]. A non-antibiotic-based marker on vaccine plasmids for use in E. coli has been developed. This system is based on the displacement of repressor molecules from the chromosome to the plasmid, allowing expression of an essential gene [20]. A selection marker developed in our laboratory uses an auxotrophic marker in L. lactis [21,22]. Here, genes encoded on the plasmid relieve the host's threonine requirement. This selection system is efficient and precludes the use of antibiotics.
The nature of the DNA between the functional genes in vaccine plasmids is also a safety concern. Specific DNA sequences or methylation patterns can induce anti-DNA antibodies and lead to the autoimmune disease systemic lupus erythematosus [23]. Gilkeson et al. showed that amongst various organisms bacterial DNA induced the highest level of DNA-specific antibodies [24]. Therefore, a reasonable strategy is to minimize the non-functional sequences in the vaccine plasmid (Table 1). Vaccine plasmids have been developed which omit the prokaryotic backbone using an integrase-mediated recombination technology [25]. In addition, these mini-circles showed higher in vivo gene expression than a standard plasmid. Alternatively, we have used a plasmid backbone derived entirely from food-grade bacterial DNA [26].
The vaccine expression unit consists of the elements necessary for high-level expression and targeting of the vaccine component (Figure 1). Most DNA vaccines harbor promoters and enhancer regions from pathogenic viruses such as cytomegalo virus (CMV), simian virus 40, or murine leukaemia virus. For instance, plasmid vaccines with the CMV promoter have been in clinical trials and are versatile due to the promoter's activity in a variety of tissues and animal models [27]. As more than 50% of the population in USA is infected with CMV and as the virus remains in the body throughout life [28], the use of its expression signals on vaccine plasmids may induce recombination events and form new chimeras of CMV. Promoters and enhancer regions have also been suggested from housekeeping genes encoding the mouse phosphoenolpyruvate carboxykinase and phosphoglycerate kinase [29]. However, due to the risk of insertional mutagenesis and oncogenesis, highly inter-species-conserved sequences like these should be avoided. This risk can be reduced by the use of novel synthetic promoters selected by bioinformatic tools to have a low homology to sequences potentially present in the recipient. To augment the promoter activity, introns are introduced, which have a beneficial effect on the in vivo expression of the vaccine gene [30]. Most often the intron A from CMV is used. Here, too, bioinformatics can aid in the design of synthetic introns thereby avoiding sequences already present in CMV-infected individuals.
For secretion of the vaccine peptide to the extra-cellular milieu, a signal sequence is positioned in front of the vaccine gene. This codes for a signal peptide of about 20–40 amino acids, often derived from bovine proteins such as the plasminogen activator [31]. However, the fusion of bovine peptides to an immunogen may induce an immunological cross-reaction. Signal peptides can themselves induce protective immunity against a microbial pathogen when administered as a gene vaccine [32]. Apparently, to avoid undesired immune responses, the nature of the signal peptide should be considered (Table 1). Statistical methods like the hidden Markov model have been used to predict and generate artificial signal peptide sequences for use in human cells [33]. Such a strategy could be applied to DNA vaccine development to create more appropriate signal peptides.
To enhance the potency of a DNA vaccine, ISS's are added to the plasmid (Figure 1). These are nucleotide hexamers that interact with Toll-like receptors and add adjuvanticity [34]. The function of the ISS is independent of its location on the plasmid and may be present in the prokaryotic backbone. In fact, Klinman eliminated ISS from the plasmid backbone and could partially restore the immunogenicity of the plasmid by exogenously added ISS DNA [35]. Therefore, changing the vector backbone or editing plasmid components may influence the immune response due to deletion of the ISS. This, too, emphasizes the importance of the proper selection of expression vector early in vaccine development.
The microbial host and production of bulk purified plasmid
The characteristics of the microbial host affect the quality of the purified DNA [36]. A number of safety concerns have been advanced concerning the microbial host. As explained in the following, these include production of toxins and biogenic amines, transferable antibiotic resistances, and genetic instability, including prophage-induced promiscuity and rearrangement of plasmid DNA (Table 1).
For reasons of efficiency, E. coli is usually chosen today as the production host, with its concomitant benefits and drawbacks. The benefits include a high DNA yield and well-established procedures for down-stream processing of the plasmid. However, as a gram-negative bacterium, E. coli contains highly immunogenic endotoxin, or lipopolysaccharides (LPS), in its outer membrane. Because of the net negative charge of both LPS and DNA, these molecules may be co-purified by the ion exchange principle used in the purification of plasmid DNA, although commercial kits do exist that can exclude LPS. On the other hand, the use of gram-positive hosts, none of which produce LPS, eliminate this dependency on the absolute efficiency of LPS-removing kits. Although not as efficient for plasmid production, L. lactis, as a gram-positive, produces neither endotoxin nor biogenic amines [37]. Assay for transferable antibiotic resistances in lactic acid bacteria is today a routine procedure; common L. lactis research strains are also genetically robust; and their prophages are of narrow host-range [38,39].
For large-scale plasmid production, often in about a thousand liters, the fermentation medium must sustain a high-level production of biomass and of plasmid DNA. At the same time the medium should be chemically defined and without components of animal origin that may contain viruses or prions [40]. Growth in a synthetic medium for many organisms results in low biomass and low plasmid yield. Indeed, switching microbial host to increase yield is complicated as it may lead to unexpected immunological results because of different DNA methylation patterns. Consequently, the production strain should be evaluated in synthetic media at an early point in development. Also here, L. lactis may be the host of choice due to its efficiency of growth in chemically defined media [41,42]. Finally, the genetic integrity of bulk purified plasmid molecules is today primarily monitored by sequence analysis. However, to reveal minor populations of molecules such as multimers or molecules with deletions and insertions, mass spectrometry should be considered [43].
Conclusion
Plasmid DNA vaccines could be the next generation of vaccines. As yet, research has focused on building functional DNA vaccines. Therefore, focus on safety has been limited. In this review we have mentioned some safety issues to be addressed early in vaccine development. Using bioinformatic tools, safe eukaryotic expression signals can be devised in synthetic DNA sequences. Safety may also be heightened by non-antibiotic plasmid selection markers, plasmid replication functions with narrow host-ranges, and minimized plasmids. Using a bio-containment strategy will also increase the safety of the microbial production host, as will avoidance of toxic substances like endotoxins. Synthetic growth media should be considered early in development and will influence choice of production host. Indeed, it can be easier to address several of these safety concerns early in vaccine development by basing the strategy on food-grade bacteria and their DNA, such as L. lactis and its DNA. Finally, the very availability of safe host-vector systems will most probably facilitate the overall acceptance of DNA vaccines.
Authors' contributions
The author(s) contributed equally to this work.
Acknowledgements
We thank Søren M. Madsen for critical reading of the manuscript. This work was partially financed by the Danish Ministry of Science, Technology and Innovation.
==== Refs
Azevedo V Levitus G Miyoshi A Candido AL Goes AM Oliveira SC Main features of DNA-based immunization vectors Braz J Med Biol Res 1999 32 147 153 10347749 10.1590/S0100-879X1999000200002
Wolff JA Malone RW Williams P Chong W Acsadi G Jani A Direct gene transfer into mouse muscle in vivo Science 1990 247 1465 1468 1690918
Sato Y Roman M Tighe H Lee D Corr M Nguyen MD Immunostimulatory DNA sequences necessary for effective intradermal gene immunization Science 1996 273 352 354 8662521
Mwau M Cebere I Sutton J Chikoti P Winstone N Wee EG A human immunodeficiency virus 1 (HIV-1) clade A vaccine in clinical trials: stimulation of HIV-specific T-cell responses by DNA and recombinant modified vaccinia virus Ankara (MVA) vaccines in humans J Gen Virol 2004 85 911 919 15039533 10.1099/vir.0.19701-0
Vastag B Ebola vaccines tested in humans, monkeys JAMA 2004 291 549 550 14762022 10.1001/jama.291.5.549
Wang R Doolan DL Le TP Hedstrom RC Coonan KM Charoenvit Y Induction of antigen-specific cytotoxic T lymphocytes in humans by a malaria DNA vaccine Science 1998 282 476 480 9774275 10.1126/science.282.5388.476
Liu M Acres B Balloul JM Bizouarne N Paul S Slos P Gene-based vaccines and immunotherapeutics Proc Natl Acad Sci USA 2004 101 14567 14571 15333750 10.1073/pnas.0404845101
Babiuk LA Pontarollo R Babiuk S Loehr B van Drunen Littel-van den Hurk Induction of immune responses by DNA vaccines in large animals Vaccine 2003 21 649 658 12531334 10.1016/S0264-410X(02)00574-1
European Medicines Agency Report from the ad hoc meeting of CPMP Gene Therapy Expert Group 2003 23–24 January 2003. EMEA-5382-03-Final
Law J Buist G Haandrikman A Kok J Venema G Leenhouts K A system to generate chromosomal mutations in Lactococcus lactis which allows fast analysis of targeted genes J Bacteriol 1995 177 7011 7018 8522504
European Medicines Agency Report from the ad hoc meeting of CPMP Gene Therapy Expert Group 2003 23–24 January 2003. EMEA-5382-03-Final
Gallot D Seifer I Lemery D Bignon YJ Systemic diffusion including germ cells after plasmidic in utero gene transfer in the rat Fetal Diagn Ther 2002 17 157 162 11914568 10.1159/000048030
European Medicines Agency Report from the ad hoc meeting of CPMP Gene Therapy Expert Group 2003 23–24 January 2003. EMEA-5382-03-Final
European Medicines Agency Note for Guidance on the Quality, Preclinical and Clinical Aspects of Gene Transfer Medicinal Products 2001 CPMP/BWP/3088/99
European Commission Communication from the Commission on a community strategy against antimicrobial resistance 2001 COM(2001) 333 final
European Food Safety Authority Working Document on the updating of the criteria used in the assessment of bacteria for resistance to antibiotics of human or veterinary importance 2004
European Medicines Agency Note for Guidance on the Quality, Preclinical and Clinical Aspects of Gene Transfer Medicinal Products 2001 CPMP/BWP/3088/99
Martinez JL Baquero F Interactions among strategies associated with bacterial infection: pathogenicity, epidemicity, and antibiotic resistance Clin Microbiol Rev 2002 15 647 679 12364374 10.1128/CMR.15.4.647-679.2002
US Food and Drug Administration Guidance for Industry #52. Assessment of the effects of antimicrobial drug residues from food of animal origin on the human intestinal flora 2004
Cranenburgh RM Hanak JA Williams SG Sherratt DJ Escherichia coli strains that allow antibiotic-free plasmid selection and maintenance by repressor titration Nucleic Acids Res 2001 29 E26 11222777 10.1093/nar/29.5.e26
Madsen SM Albrechtsen B Hansen EB Israelsen H Cloning and transcriptional analysis of two threonine biosynthetic genes from Lactococcus lactis MG1614 J Bacteriol 1996 178 3689 3694 8682767
Glenting J Madsen SM Vrang A Fomsgaard A Israelsen H A plasmid selection system in Lactococcus lactis and its use for gene expression in L. lactis and human kidney fibroblasts Appl Environ Microbiol 2002 68 5051 5056 12324356 10.1128/AEM.68.10.5051-5056.2002
Pisetsky DS Immune response to DNA in systemic lupus erythematosus Isr Med Assoc J 2001 3 850 853 11729584
Gilkeson GS Grudier JP Pisetsky DS The antibody response of normal mice to immunization with single-stranded DNA of various species origin Clin Immunol Immunopathol 1989 51 362 371 2785884 10.1016/0090-1229(89)90034-2
Darquet AM Cameron B Wils P Scherman D Crouzet J A new DNA vehicle for nonviral gene delivery: supercoiled minicircle Gene Ther 1997 4 1341 1349 9472558 10.1038/sj.gt.3300540
Glenting J Madsen SM Vrang A Fomsgaard A Israelsen H A plasmid selection system in Lactococcus lactis and its use for gene expression in L. lactis and human kidney fibroblasts Appl Environ Microbiol 2002 68 5051 5056 12324356 10.1128/AEM.68.10.5051-5056.2002
Wright A Semyonov A Dawes G Crameri A Lyons R Stemmer WP Diverse plasmid DNA vectors by directed molecular evolution of cytomegalovirus promoters Hum Gene Ther 2005 16 881 892 16000069 10.1089/hum.2005.16.881
National Center for Infectious Diseases
Cheng L Ziegelhoffer PR Yang NS In vivo promoter activity and transgene expression in mammalian somatic tissues evaluated by using particle bombardment Proc Natl Acad Sci USA 1993 90 4455 4459 8506285
Brinster RL Allen JM Behringer RR Gelinas RE Palmiter RD Introns increase transcriptional efficiency in transgenic mice Proc Natl Acad Sci USA 1988 85 836 840 3422466
Ashok MS Rangarajan PN Protective efficacy of a plasmid DNA encoding Japanese encephalitis virus envelope protein fused to tissue plasminogen activator signal sequences: studies in a murine intracerebral virus challenge model Vaccine 2002 20 1563 1570 11858863 10.1016/S0264-410X(01)00492-3
Jiang C Magee DM Ivey FD Cox RA Role of signal sequence in vaccine-induced protection against experimental coccidioidomycosis Infect Immun 2002 70 3539 3545 12065493 10.1128/IAI.70.7.3539-3545.2002
Barash S Wang W Shi Y Human secretory signal peptide description by hidden Markov model and generation of a strong artificial signal peptide for secreted protein expression Biochem Biophys Res Commun 2002 294 835 842 12061783 10.1016/S0006-291X(02)00566-1
Tudor D Dubuquoy C Gaboriau V Lefevre F Charley B Riffault S TLR9 pathway is involved in adjuvant effects of plasmid DNA-based vaccines Vaccine 2005 23 1258 1264 15652668 10.1016/j.vaccine.2004.09.001
Klinman DM CpG DNA as a vaccine adjuvant Expert Rev Vaccines 2003 2 305 315 12899580 10.1586/14760584.2.2.305
ICH Topic Q5D Quality of biotechnological products: derivation and characterisation of cell substrates use for production of biotechnological/biological products Geneva: International Conference on Harmonisation of Technical Requirements for Registration of Pharmaceuticals for Human Use (ICH) 1997
Straub BW Kicherer M Schilcher SM Hammes WP The formation of biogenic amines by fermentation organisms Z Lebensm Unters Forsch 1995 201 79 82 7571871 10.1007/BF01193205
Rakonjac J O'Toole PW Lubbers M Isolation of lactococcal prolate phage-phage recombinants by an enrichment strategy reveals two novel host range determinants J Bacteriol 2005 187 3110 3121 15838038 10.1128/JB.187.9.3110-3121.2005
European Food Safety Authority Working Document on the updating of the criteria used in the assessment of bacteria for resistance to antibiotics of human or veterinary importance 2004
ICH Topic Q5D Quality of biotechnological products: derivation and characterisation of cell substrates use for production of biotechnological /biological products Geneva: International Conference on Harmonisation of Technical Requirements for Registration of Pharmaceuticals for Human Use (ICH) 1997
Mierau I Leij P van SI Blommestein B Floris E Mond J Industrial-scale production and purification of a heterologous protein in Lactococcus lactis using the nisin-controlled gene expression system NICE: the case of lysostaphin Microb Cell Fact 2005 4 15 15921518 10.1186/1475-2859-4-15
Mierau I Olieman K Mond J Smid EJ Optimization of the Lactococcus lactis nisin-controlled gene expression system NICE for industrial applications Microb Cell Fact 2005 4 16 15921537 10.1186/1475-2859-4-16
Marvin LF Roberts MA Fay LB Matrix-assisted laser desorption/ionization time-of-flight mass spectrometry in clinical chemistry Clin Chim Acta 2003 337 11 21 14568176 10.1016/j.cccn.2003.08.008
|
16144545
|
PMC1215512
|
CC BY
|
2021-01-04 16:24:35
|
no
|
Microb Cell Fact. 2005 Sep 6; 4:26
|
utf-8
|
Microb Cell Fact
| 2,005 |
10.1186/1475-2859-4-26
|
oa_comm
|
==== Front
Mol PainMolecular Pain1744-8069BioMed Central London 1744-8069-1-241611150110.1186/1744-8069-1-24ResearchNeuropathic pain develops normally in mice lacking both Nav1.7 and Nav1.8 Nassar Mohammed A [email protected] Alessandra [email protected] L Caroline [email protected] John N [email protected] Molecular Nociception Group, and London Pain Consortium, Department of Biology, University College London, Gower Street, WC1E 6BT, London, UK2005 22 8 2005 1 24 24 4 8 2005 22 8 2005 Copyright © 2005 Nassar et al; licensee BioMed Central Ltd.2005Nassar et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Two voltage gated sodium channel α-subunits, Nav1.7 and Nav1.8, are expressed at high levels in nociceptor terminals and have been implicated in the development of inflammatory pain. Mis-expression of voltage-gated sodium channels by damaged sensory neurons has also been implicated in the development of neuropathic pain, but the role of Nav1.7 and Nav1.8 is uncertain. Here we show that deleting Nav1.7 has no effect on the development of neuropathic pain. Double knockouts of both Nav1.7 and Nav1.8 also develop normal levels of neuropathic pain, despite a lack of inflammatory pain symptoms and altered mechanical and thermal acute pain thresholds. These studies demonstrate that, in contrast to the highly significant role for Nav1.7 in determining inflammatory pain thresholds, the development of neuropathic pain does not require the presence of either Nav1.7 or Nav1.8 alone or in combination.
==== Body
Background
Voltage gated sodium channels (VGSC) underlie the electrical excitability of nerve and muscle. VGSCs consist of pore forming α-subunits and auxiliary β-subunits. There are ten cloned α-subunits and 4 β-subunits. The β-subunits modulate the localisation, expression and functional properties of α-subunits [1]. Different α-subunits have distinct electrophysiological and pharmacological properties [2]. The complex pattern of expression of α-subunits may imply special roles for particular subunits in different cell types [3]. Many loss- as well as gain-of-function mutations of α-subunits have been identified in human conditions characterised with epilepsy, seizures, ataxia and increased sensitivity to pain. This suggests that mutations of VGSC in humans are significant factors in aetiology of neuronal diseases [4].
Nociceptors are a subset of sensory neurons that respond to noxious thermal, mechanical and chemical stimuli. Nociceptors express multiple subtypes of α-subunits [3]. Tissue and nerve damage leads to changes in expression and function of α-subunits that in turn can lead to change in the excitability of sensory neurons. Changes in the excitability of sensory neurons are thought to underlie some chronic pain conditions [5,6]. Nav1.8 and Nav1.7 are two α-subunits that are abundant in nociceptive sensory neurons [3,7,8]. Nav1.8 is expressed exclusively in sensory neurons and is not found in the CNS [9]. Functional characterization of Nav1.8 positive neurons revealed that more than 85% are nociceptors [8]. Nav1.7 is expressed principally in peripheral neurons with very weak expression detected in CNS [10,11]. As there are no subunit specific blockers, gene knockouts in mice have provided insights into the role of individual α-subunits genes in pain [5]. Deletion of the Nav1.8 gene [12] and nociceptor-specific knockout of the Nav1.7 [13] gene have identified a role for these two α-subunits in setting mechanical and, to a lesser extent, thermal pain thresholds. In addition, behavioral studies have revealed deficits in inflammatory pain models, most dramatically in the nociceptor-specific Nav1.7 knockout [12,13]. These findings suggest that these α-subunits could be targets for new anti-inflammatory drugs.
Peripheral nerve injury leads to lowered pain thresholds, enhanced responsiveness and/or ectopic activity in sensory neurons that ultimately leads to hyperalgesia and allodynia [5,6]. Changes in expression of α-subunits of VGSCs have been documented in models of peripheral nerve injury [5,6]. This has lead to the hypothesis that modulation of α-subunits in sensory neurons may underlie the increased neuronal excitability of sensory neurons following peripheral nerve injury. Pharmacological blockade of sodium channel activity has been shown to attenuate ectopic activity [14,15] and reverse hyperalgesia following nerve injury [16]. While the role of Nav1.7 in neuropathic pain remains to be investigated, analysis of a Nav1.8 knockout mouse indicated that it is not involved in alteration of pain threshold following peripheral nerve injury [17]. This is in contrast to the finding of Lai et al who reported that antisense oligonucleotides directed against Nav1.8 administered intrathecally completely reverse neuropathic pain behavior [18]. It is possible that this discrepancy could be due to the up-regulation of the Nav1.7 channel seen in the Nav1.8 knockout mouse [12] which might mask an otherwise important role for Nav1.8 in neuropathic pain.
In the present study we investigated the role of the Nav1.7 channel in neuropathic pain using nociceptor-specific deletion of Nav1.7 in mouse. In addition, we readdressed the role of Nav1.8 in neuropathic pain by generating a double knockout of Nav1.8 and Nav1.7. We reasoned that the co-deletion of Nav1.7 in Nav1.8-expressing neurons should reveal any potential role for Nav1.8 in neuropathic pain.
Results
Generation of Nav1.7 and Nav1.8 double knockouts in nociceptors
Nociceptor-specific Nav1.7 knockout mice and their littermate controls were generated as described previously [13]. It is not possible to generate global knockouts of both Nav1.8 and Nav1.7 since global deletion of Nav1.7 is lethal at P0 [13]. Therefore, we generated Nav1.7 and Nav1.8 double-knockout (DKO) mice by exploiting the fact that the Nav1.8Cre allele has Cre sequence inserted in exon 1 followed by transcriptional stop signals [19]. Consequently homozygous Nav1.8Cre mice are Nav1.8 global-knockouts (Nav18 KO) and show no Nav1.8 currents in sensory neurons [19].
We compared the DKO strain to the Nav1.8 KO (homozygous Nav1.8Cre). To obtain the desired strains we mated mice homozygous for the Nav1.8Cre/heterozygous for floxed Nav1.7 allele with each other. The resulting progeny are all Nav1.8 Knockout. However, 25% will have wildtype Nav1.7 alleles, 50% will be heterozygous for the floxed Nav1.7 allele and 25% will be homozygous for the floxed Nav1.7 allele (i.e. DKO) figure 1.
Figure 1 Generation and genotyping of Nav1.7 and Nav1.8 double knockout mice. PCR genotyping of mice generated by breeding homozygous Nav1.8Cre/heterozygous floxed Nav1.7 mice with each other (lanes 2, 3, and 4). All the three groups are Nav1.8 Knockouts, i.e. positive for the Cre band (249 bp) and negative for the wildtype Nav1.8 band (460 bp). Mice homozygous for the floxed Nav1.7 allele (461 vs. 317 bp band) are nociceptor-specific Nav1.7 knockout as well (lane 4). C57BL6 wildtype control is shown in lane 1 and no DNA negative control is in lane 5.
Eight weeks old C57BL6 inbred mice were used as wildtype (WT) control (wildtype Nav1.8 and Nav1.7) because it would be impractical and undesirable to generate them as littermates to the test groups (DKO and Nav1.8 KO). The expected ratio of each of desired strains would be only 6.25% if the parents were heterozygous for both Nav1.8Cre and floxed Nav1.7 alleles compared with 25% as in our breeding strategy. The Nav1.8 KO/Nav1.7 heterozygous group was analysed in behavioural tests and was found to be similar to the Nav1.8 KO group (data not shown).
Development and motor coordination
The weight of male (WT 26.72 ± 1.60 gram, n = 6; Nav1.8 KO 23.44 ± 0.76, n = 9 and DKO 27.05 ± 1.63, n = 9) and female (WT 20.50 ± 1.77 gram, n = 3; Nav1.8 KO 21.43 ± 0.95, n= 3 and DKO 19.66 ± 1.03, n = 4) mice in each group was very similar and was not significantly different, figure 2a. Motor coordination as tested through performance on the rotarod apparatus was unchanged between the mouse groups studied. The time spent on the rotarod was not significantly different between the three mouse groups, figure 2b (WT 145.6 ± 13.12 sec, n = 4; Nav1.8 KO 146.4 ± 11.37, n = 7 and DKO 108.6 ± 11.5, n = 7).
Figure 2 Development and motor coordination. A) Males and females mice have similar weights among the three groups being compared. In males the P value for WT vs. Nav1.8 KO is 0.09, for WT vs. DKO is 0.89 and for Nav1.8 KO vs. DKO is 0.08. In females the P value for WT vs. Nav1.8 KO is 0.27, for WT vs. DKO is 0.71 and for Nav1.8 KO vs. DKO is 0.27. B) Performance on the rotarod apparatus is not significantly different between the three mice groups. The P value for WT vs. Nav1.8 KO is 0.98, for WT vs. DKO is 0.38 and for Nav1.8 KO vs. DKO is 0.39. P values were calculated using two-tailed T-test. WT in white, Nav1.8 KO in grey and DKO in black.
Acute pain thresholds
We measured thermal pain thresholds using the hotplate and Hargreave's tests. In the hotplate test, which involves supraspinal activity, the response latency was not different between the three mouse groups (WT 36.1 ± 6.5 sec, n = 7; Nav1.8 KO 29.0 ± 5.6, n = 7 and DKO 32.7 ± 6.35, n = 7), figure 3a. However, the response latency in the Hargreave's test was doubled in DKO group (15.30 ± 0.91 sec, n = 7) compared to both WT (6.88 ± 0.28, n = 11) and Nav1.8 KO (8.47 ± 0.78, n = 7) groups, figure 3b. The latency in the Nav1.8 KO was higher that that of the WT group which is in agreement with the finding of Akopian et al using the conventional Nav1.8 global knockout [12]. The nociceptor-specific Nav1.7 knockout alone shows a 40% increase in latency in the Hargreave's test [13]
Figure 3 Acute pain thresholds are increased in the DKO mice. A) Latency to respond in the hotplate test was not different between all groups. The P value for WT vs. Nav1.8 KO is 0.67, for WT vs. DKO is 0.72 and for Nav1.8 KO vs. DKO is 0.43. B) Latency to paw withdrawal in the Hargreave's test was doubled in the DKO mice. The P value for WT vs. Nav1.8 KO is 0.08, for WT vs. DKO is <0.0001 and for Nav1.8 KO vs. DKO is <0.0001. C) The 50% withdrawal threshold to stimulation with von Frey hairs was not different between all groups. The P value for WT vs. Nav1.8 KO is 0.21, for WT vs. DKO is 0.25 and for Nav1.8 KO vs. DKO is 0.84. D) The Nav1.8 KO and DKO mice showed profound analgesia to noxious mechanical pressure applied to the tail using the Randall-Selitto apparatus. The P value for WT vs. Nav1.8 KO is <0.0001, for WT vs. DKO is <0.0001 and for Nav1.8 KO vs. DKO is 0.67. P values were calculated using two-tailed T-test. WT in white, Nav1.8 KO in grey and DKO in black.
Pain thresholds to punctate mechanical stimulation was measured using calibrated von Frey hairs according to the up and down method [20]. The 50% withdrawal threshold was not different between the three groups (WT 0.56 ± 0.07 gram, n = 11; Nav1.8 KO 0.45 ± 0.05, n = 13 and DKO 0.49 ± 0.06, n = 17), figure 3c. In contrast pain threshold to noxious mechanical pressure applied to the tail using the Randall-Selitto apparatus was much higher in both the Nav1.8 KO (412.4 ± 33.97 gram, n = 7) and DKO (395.7 ± 21.03, n = 7) groups compared to that of the WT (131.4 ± 16.54, n = 6) group, figure 3d. Usually the cut off point (500 gram) was reached without an observed escape response in the Nav1.8 KO and DKO groups, indicating a high resistance to static blunt mechanical pressure. There was no difference between the Nav1.8 and the DKO mice in this test, figure 3d, and similar results have been obtained with the nociceptor-specific Nav1.7 knockout [13]
Inflammatory pain behaviour
The pain response elicited by injection of 20μl of 5% formalin intradermally in the hindpaw showed the typical biphasic response in the three groups, figure 4a. The first phase (1–10 minutes) was not different between the three groups (WT 107.8 ± 10.12 sec n = 8, Nav1.8 KO 155 ± 11.00 n = 4, DKO 120.4 ± 5.96 n = 8), figure 4b. In contrast, the second phase (10–60 minutes) was much reduced in the DKO group (105.4 ± 28.43 sec) compared to both the Nav1.8 KO (309 ± 80.85, P = 0.08) and the WT (216 ± 43.67, P = 0.0016) groups, figure 4b. This effect was even more dramatic than that observed in the nociceptor specific Nav1.7 knockout [13] We did not study other inflammatory pain models since the nociceptor-specific Nav1.7 knockout mouse is completely deficient in commonly used inflammatory pain models [13]. It would be impossible therefore to measure a further reduction in inflammatory pain behaviour in the DKO.
Figure 4 Reduced pain behavior in formalin test in DKO. A) The pain response after injection of 5% formalin in hindpaw showed the typical biphasic course. B) The second phase of the formalin response was reduced to 30% in the DKO mice compared to WT and Nav1.8 KO mice. The P value for WT vs. Nav1.8 KO is 0.94, for WT vs. DKO is 0.001 and for Nav1.8 KO vs. DKO is 0.08. The first phase was not significantly different between all groups. The P value for WT vs. Nav1.8 KO is 0.60, for WT vs. DKO is 0.30 and for Nav1.8 KO vs. DKO is 0.96. P values were calculated using two-tailed T-test. WT in open circles, Nav1.8 KO in grey triangles and DKO in black squares.
Neuropathic pain in nociceptor-specific Nav1.7 knockout mice
To study neuropathic pain behavior we induced peripheral nerve injury using to the Chung model (ligation of L5 spinal nerve) in the Nav1.7 nociceptor-specific knockout and homozygous floxed-Nav1.7 as controls. Both groups developed a robust mechanical allodynia starting form the third day post surgery, figure 5a. The extent and time course of development of increased mechanosensitivity was identical in both nociceptor-specific Nav1.7 knockout and their littermate controls.
Figure 5 Analysis of neuropathic pain in nociceptor-specific Nav1.7 knockout mice and double knockout mice. (A) Both control (white circles, mean 0.56 ± 0.9, n = 8) and nociceptor-specific Nav1.7 KO (black squares, 0.48 ± 0.8, n = 11) developed robust mechanical allodynia after ligation of spinal nerve L5. There is no difference in the extent of pain behavior at any time point (P = 0.49 ANOVA). (B) WT (open circles), Nav1.8 KO (grey triangles) and DKO (black squares) developed profound mechanical allodynia after ligation of spinal nerve L5.
Neuropathic pain in Nav1.7 Nav1.8 double knockout mice
To address the issue of compensatory up-regulation of Nav1.7 in Nav1.8 knockout we induced peripheral nerve injury in WT, Nav1.8 KO and DKO. All mice groups studied developed a robust mechanical allodynia starting form the third day post surgery, figure 5b. There were no statistically meaningful differences in the behaviour of the groups of mice.
Discussion
This study confirms that the two sodium channel α- subunits Nav1.7 and Nav1.8, expressed selectively in nociceptive sensory neurons, have important roles in nociception and in pain. In the absence of specific pharmacological blockers, the use of genetic approaches, a combination of global and nociceptor-specific knockouts, has enabled us to carry out studies exploring the contribution of these two isoforms of VGSCs in different pain conditions. Nav1.8 knockouts are already known to have deficits in inflammatory and visceral but not neuropathic pain [12]. Nav1.7 nociceptor-specific knockouts are almost totally refractory to changes in peripheral pain thresholds evoked by inflammatory mediator [13]. This mirrors the inflammatory phenotype of Nav1.7 gain-of-function mutations in erythermalgia [21]. This heritable disorder is caused by point mutations that lead to altered thresholds of activation in Nav1.7 [22]
Acute pain thresholds
analysis of acute mechanical thresholds in DKO mice confirmed previous findings that the knockout of either Nav1.8 or Nav1.7 renders the mice resistant in the Randel-Selitto test of noxious mechanical pressure while the responses to von Frey hairs remain unchanged compared to controls [12,13]. Furthermore, it has been shown that both Nav1.8 and nociceptor-specific Nav1.7 knockouts have small increases of about 20–40% in their thermal threshold [12,13]. Interestingly, we found that the thermal threshold in the Hargreave's test was doubled in the DKO compared to Nav1.8 knockout or WT controls, figure 3b. The increase is more than the sum of the phenotype in the individual knockouts and may indicate that Nav1.8 and Nav1.7 are the major VGSC isoforms present in nociceptive terminals and their deletion dramatically increases pain threshold. This is supported by the fact that Nav1.7 is transported to nerve terminals of sensory neurons [10] and that Nav1.8 currents are localized in the terminals [23].
Inflammatory pain
We have not studied pain behavior in inflammatory models where the nociceptive-specific Nav1.7 knockout has shown complete deficits in inflammatory pain behavior [13], because it would not be possible to observe further changes brought about by the co-deletion of Nav1.8. We only investigated the pain response in the formalin model. While the nociceptive-specific Nav1.7 knockout showed a reduction in the second phase to about 50% of controls [13] the DKO showed a slightly bigger reduction to 30% of controls, figure 4b. This is in contrast to Nav1.8 mice that showed no reduction in the second phase compared to controls, figure 4b. Therefore, the reduction in the second phase in the DKO is a new phenotype rather than a mere summation of the phenotype in the individual knockouts. This highlights further the predominant role for Nav1.7 in inflammatory pain. Expression of Nav1.7 has been reported to increase in sensory neurons subsequent to induction of inflammation [24].
This study and our previous [13] work have explored the role of Nav1.7 in Nav1.8-expressing neurons, since our strategy exploits the Nav1.8 promoter to drive Cre expression. We have shown previously using non-quantitative RT-PCR that not all the Nav1.7 mRNA in DRGs from the nociceptor-specific Nav1.7 knockout is deleted [13]. This indicates that there is a population of sensory neurons, the size of which has yet to be determined, that express Nav1.7 but not Nav1.8. Therefore, the role of Nav1.7 in sensory neurons where Nav1.8 is not expressed (and which do not express Cre) remains to be studied.
From all the above, our results indicate that Nav1.8 and Nav1.7 subunits contribute to the excitability of peripheral nerve terminals and their modulation is critical for peripheral sensitisation in inflammatory pain
Neuropathic pain
We investigated the role of Nav1.7 expressed in nociceptors in neuropathic pain. Analysis of neuropathic pain in the nociceptor-specific Nav1.7 knockout using the Chung model showed that they developed mechanical allodynia to the same extent as the littermate control mice (homozygous floxed Nav1.7), figure 5. This clearly shows that unlike its critical role in inflammatory pain [13], Nav1.7 does not contribute to neuropathic pain.
In addition, the Nav1.8 knockout and WT as well as the DKO mice developed a profound mechanical allodynia the level of which was indistinguishable between the three groups. This provides further evidence that Nav1.8 does not contribute to mechanical allodynia as reported by Kerr et al [17]. In addition our results rule out Nav1.7 up-regulation in the Nav1.8 knockout as the reason behind the discrepancy between the finding of Kerr et al [17] and Lai et al [18]. Surprisingly Lai et al [18] failed to detect the changes in acute pain thresholds observed in the Nav1.8 KO both in this study and that of Akopian et al [12], which would be expected to occur if Nav1.8 had been down-regulated.
The lack of changes in mechanical allodynia in any individual or even in double knockouts of Nav1.8 and Nav1.7 subunits is consistent with recent data by Flake et al. They reported that changes in sodium currents shortly after nerve injury do not correlate with an increase in neuronal excitability and that the sum of changes to ionic currents, and not a single class of voltage-gated ion channel, underlie increased neuronal excitability [25]. This does not, however, exclude a potential role for both Nav1.7 and Nav1.8 in spontaneous neuropathic pain as suggested by the finding that neuromas in the Nav1.8 knockout mouse display less ectopic discharges than wildtype littermates [26]. These studies confirm the importance of Nav1.7 and Nav1.8 in the physiological processes that underlie altered peripheral thresholds in inflammatory pain. Neuropathic pain that arises as a consequence of nerve damage and neuronal dysfunction is, however, not dependent on the presence of these two sodium channels. In terms of the contribution of other α-subunits to neuropathic pain, a strong correlation has been found between the expression of Nav1.3 and the appearance of neuropathic pain. GDNF that reverses neuropathic pain behaviour also normalises Nav1.3 expression [27], and recent antisense studies have supported the view that increased Nav1.3 expression contributes to neuropathic pain development [28]
Conclusion
In summary, our results indicate a critical role for Nav1.7 and Nav1.8 in setting pain thresholds of nociceptive nerve terminals and in their sensitisation following tissue damage and inflammation. However, the presence of these channels, or changes to their expression and function are not required for the establishment of mechanical allodynia arising from nerve injury.
Methods
Genotyping of mice strains
Nav1.8Cre and floxed Nav1.7 lines were produced as described [13]. Both strains were back-crossed at last 5 times onto a C57/BL6 background.
Genomic DNA was prepared from tail biopsies and genotyped using PCR. Primers for Nav1.7 are (CAGAGATTTCTGCATTAGAATTTGTTC) and (AGTCTTTGTGGCACACGTTACCTC) which give a WT band of 317 bp and a floxed band of 461 bp. Primers for Nav1.8 are (TGTAGATGGACTGCAGAGGATGGA) and (ttacccggtgtgtgctgtagaaag) which give a WT band of 460 bp. Nav1.8Cre was detected by primers (aaatttgcctgcattaccggtcga) and (aaatgttgctggatagtttttactgcc) located inside Cre sequence which give a band of 249 bp.
Behavioural analysis
All tests were approved by the United Kingdom Home Office Animals (Scientific Procedures) Act 1986 and performed in a Home Office designated room at 22 ± 2°C. Experiments were performed on animals of at least 8 weeks of age. Behavioral tests were done as before [19]
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
MAN generated the mice, performed behavioral tests and wrote the manuscript.
AL carried out the neuropathic pain study of the DKO and Nav1.8 KO.
LCS Analysed the Nav1.8Cre mouse.
JNW: supervised experiments and corrected the manuscript.
All authors read and approved the final manuscript
Acknowledgements
The authors would like to acknowledge Mark D Baker for critical reading of the manuscript and Elizabeth A Matthew for great help in establishing the Chung model. We also thank the Wellcome Trust, the MRC and the London Pain Consortium for funding this work.
==== Refs
Isom LL Sodium channel beta subunits: anything but auxiliary Neuroscientist 2001 7 42 54 11486343
Baker MD Wood JN Involvement of Na+ channels in pain pathways Trends Pharmacol Sci 2001 22 27 31 11165669 10.1016/S0165-6147(00)01585-6
Felts PA Yokoyama S Dib-Hajj S Black JA Waxman SG Sodium channel alpha-subunit mRNAs I, II, III, NaG, Na6 and hNE (PN1): different expression patterns in developing rat nervous system Brain Res Mol Brain Res 1997 45 71 82 9105672 10.1016/S0169-328X(96)00241-0
Meisler MH Kearney JA Sodium channel mutations in epilepsy and other neurological disorders The Journal of Clinical Investigation 2005 115 2010 2017 16075041 10.1172/JCI25466
Wood JN Abrahamsen B Baker MD Boorman JD Donier E Drew LJ Nassar MA Okuse K Seereeram A Stirling CL Zhao J Ion channel activities implicated in pathological pain Novartis Found Symp 2004 261 32 40 discussion 40–54 15469043
Lai J Porreca F Hunter JC Gold MS Voltage-gated sodium channels and hyperalgesia Annu Rev Pharmacol Toxicol 2004 44 371 397 14744251 10.1146/annurev.pharmtox.44.101802.121627
Djouhri L Newton R Levinson SR Berry CM Carruthers B Lawson SN Sensory and electrophysiological properties of guinea-pig sensory neurones expressing Nav 1.7 (PN1) Na+ channel alpha subunit protein J Physiol 2003 546 565 576 12527742 10.1113/jphysiol.2002.026559
Djouhri L Fang X Okuse K Wood JN Berry CM Lawson SN The TTX-resistant sodium channel Nav1.8 (SNS/PN3): expression and correlation with membrane properties in rat nociceptive primary afferent neurons J Physiol 2003 550 739 752 12794175 10.1113/jphysiol.2003.042127
Akopian AN Sivilotti L Wood JN A tetrodotoxin-resistant voltage-gated sodium channel expressed by sensory neurons Nature 1996 379 257 262 8538791 10.1038/379257a0
Toledo-Aral JJ Moss BL He ZJ Koszowski AG Whisenand T Levinson SR Wolf JJ Silos-Santiago I Halegoua S Mandel G Identification of PN1, a predominant voltage-dependent sodium channel expressed principally in peripheral neurons Proc Natl Acad Sci U S A 1997 94 1527 1532 9037087 10.1073/pnas.94.4.1527
Sangameswaran L Fish LM Koch BD Rabert DK Delgado SG Ilnicka M Jakeman LB Novakovic S Wong K Sze P Tzoumaka E Stewart GR Herman RC Chan H Eglen RM Hunter JC A novel tetrodotoxin-sensitive, voltage-gated sodium channel expressed in rat and human dorsal root ganglia J Biol Chem 1997 272 14805 14809 9169448 10.1074/jbc.272.23.14805
Akopian AN Souslova V England S Okuse K Ogata N Ure J Smith A Kerr BJ McMahon SB Boyce S Hill R Stanfa LC Dickenson AH Wood JN The tetrodotoxin-resistant sodium channel SNS has a specialized function in pain pathways Nat Neurosci 1999 2 541 548 10448219 10.1038/9195
Nassar MA Stirling LC Forlani G Baker MD Matthews EA Dickenson AH Wood JN Nociceptor-specific gene deletion reveals a major role for Nav1.7 (PN1) in acute and inflammatory pain Proc Natl Acad Sci U S A 2004 101 12706 12711 15314237 10.1073/pnas.0404915101
Devor M Wall PD Catalan N Systemic lidocaine silences ectopic neuroma and DRG discharge without blocking nerve conduction Pain 1992 48 261 268 1589245 10.1016/0304-3959(92)90067-L
Omana-Zapata I Khabbaz MA Hunter JC Clarke DE Bley KR Tetrodotoxin inhibits neuropathic ectopic activity in neuromas, dorsal root ganglia and dorsal horn neurons Pain 1997 72 41 49 9272786 10.1016/S0304-3959(97)00012-2
Chabal C Jacobson L Little J Intrathecal fentanyl depresses nociceptive flexion reflexes in patients with chronic pain Anesthesiology 1989 70 226 229 2913859
Kerr BJ Souslova V McMahon SB Wood JN A role for the TTX-resistant sodium channel Nav 1.8 in NGF-induced hyperalgesia, but not neuropathic pain Neuroreport 2001 12 3077 3080 11568640 10.1097/00001756-200110080-00019
Lai J Gold MS Kim CS Bian D Ossipov MH Hunter JC Porreca F Inhibition of neuropathic pain by decreased expression of the tetrodotoxin-resistant sodium channel, NaV1.8 Pain 2002 95 143 152 11790477 10.1016/S0304-3959(01)00391-8
Stirling LC Forlani G Baker MD Wood JN Matthews EA Dickenson AH Nassar MA Nociceptor-specific gene deletion using heterozygous NaV1.8-Cre recombinase mice Pain 2005 113 27 36 15621361 10.1016/j.pain.2004.08.015
Chaplan SR Bach FW Pogrel JW Chung JM Yaksh TL Quantitative assessment of tactile allodynia in the rat paw J Neurosci Methods 1994 53 55 63 7990513 10.1016/0165-0270(94)90144-9
Yang Y Wang Y Li S Xu Z Li H Ma L Fan J Bu D Liu B Fan Z Wu G Jin J Ding B Zhu X Shen Y Mutations in SCN9A, encoding a sodium channel alpha subunit, in patients with primary erythermalgia J Med Genet 2004 41 171 174 14985375 10.1136/jmg.2003.012153
Dib-Hajj SD Rush AM Cummins TR Hisama FM Novella S Tyrrell L Marshall L Waxman SG Gain-of-function mutation in Nav1.7 in familial erythromelalgia induces bursting of sensory neurons Brain 2005 128 1847 1854 15958509 10.1093/brain/awh514
Brock JA McLachlan EM Belmonte C Tetrodotoxin-resistant impulses in single nociceptor nerve terminals in guinea-pig cornea J Physiol 1998 512 211 217 9729630 10.1111/j.1469-7793.1998.211bf.x
Gould HJ 3rdEngland JD Soignier RD Nolan P Minor LD Liu ZP Levinson SR Paul D Ibuprofen blocks changes in Nav1.7 and 1.8 sodium channels associated with complete Freund's adjuvant-induced inflammation in rat J Pain 2004 5 270 280 15219259 10.1016/j.jpain.2004.04.005
Flake NM Lancaster E Weinreich D Gold MS Absence of an association between axotomy-induced changes in sodium currents and excitability in DRG neurons from the adult rat Pain 2004 109 471 480 15157708 10.1016/j.pain.2004.02.024
Roza C Laird JM Souslova V Wood JN Cervero F The tetrodotoxin-resistant Na+ channel Nav1.8 is essential for the expression of spontaneous activity in damaged sensory axons of mice J Physiol 2003 550 921 926 12824446 10.1113/jphysiol.2003.046110
Boucher TJ Okuse K Bennett DL Munson JB Wood JN McMahon SB Potent analgesic effects of GDNF in neuropathic pain states Science 2000 290 124 127 11021795 10.1126/science.290.5489.124
Hains BC Klein JP Saab CY Craner MJ Black JA Waxman SG Upregulation of sodium channel Nav1.3 and functional involvement in neuronal hyperexcitability associated with central neuropathic pain after spinal cord injury J Neurosci 2003 23 8881 8892 14523090
|
16111501
|
PMC1215513
|
CC BY
|
2021-01-04 16:40:08
|
no
|
Mol Pain. 2005 Aug 22; 1:24
|
utf-8
|
Mol Pain
| 2,005 |
10.1186/1744-8069-1-24
|
oa_comm
|
==== Front
Reprod Biol EndocrinolReproductive biology and endocrinology : RB&E1477-7827BioMed Central London 1477-7827-3-281601881410.1186/1477-7827-3-28ReviewRole of oxidative stress in female reproduction Agarwal Ashok [email protected] Sajal [email protected] Rakesh K [email protected] Center for Advanced Research in Human Reproduction, Infertility, and Sexual Function, Glickman Urological Institute and Department of Obstetrics-Gynecology; The Cleveland Clinic Foundation, Cleveland, Ohio 44195, USA2005 14 7 2005 3 28 28 9 5 2005 14 7 2005 Copyright © 2005 Agarwal et al; licensee BioMed Central Ltd.2005Agarwal et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
In a healthy body, ROS (reactive oxygen species) and antioxidants remain in balance. When the balance is disrupted towards an overabundance of ROS, oxidative stress (OS) occurs. OS influences the entire reproductive lifespan of a woman and even thereafter (i.e. menopause). OS results from an imbalance between prooxidants (free radical species) and the body's scavenging ability (antioxidants). ROS are a double-edged sword – they serve as key signal molecules in physiological processes but also have a role in pathological processes involving the female reproductive tract. ROS affect multiple physiological processes from oocyte maturation to fertilization, embryo development and pregnancy. It has been suggested that OS modulates the age-related decline in fertility. It plays a role during pregnancy and normal parturition and in initiation of preterm labor. Most ovarian cancers appear in the surface epithelium, and repetitive ovulation has been thought to be a causative factor. Ovulation-induced oxidative base damage and damage to DNA of the ovarian epithelium can be prevented by antioxidants. There is growing literature on the effects of OS in female reproduction with involvement in the pathophsiology of preeclampsia, hydatidiform mole, free radical-induced birth defects and other situations such as abortions. Numerous studies have shown that OS plays a role in the pathoysiology of infertility and assisted fertility. There is some evidence of its role in endometriosis, tubal and peritoneal factor infertility and unexplained infertility. This article reviews the role OS plays in normal cycling ovaries, follicular development and cyclical endometrial changes. It also discusses OS-related female infertility and how it influences the outcomes of assisted reproductive techniques. The review comprehensively explores the literature for evidence of the role of oxidative stress in conditions such as abortions, preeclampsia, hydatidiform mole, fetal embryopathies, preterm labour and preeclampsia and gestational diabetes. The review also addresses the growing literature on the role of nitric oxide species in female reproduction. The involvement of nitric oxide species in regulation of endometrial and ovarian function, etiopathogenesis of endometriosis, and maintenance of uterine quiescence, initiation of labour and ripening of cervix at parturition is discussed. Complex interplay between cytokines and oxidative stress in the etiology of female reproductive disorders is discussed. Oxidant status of the cell modulates angiogenesis, which is critical for follicular growth, corpus luteum formation endometrial differentiation and embryonic growth is also highlighted in the review. Strategies to overcome oxidative stress and enhance fertility, both natural and assisted are delineated. Early interventions being investigated for prevention of preeclampsia are enumerated. Trials investigating combination intervention strategy of vitamin E and vitamin C supplementation in preventing preeclampsia are highlighted. Antioxidants are powerful and there are few trials investigating antioxidant supplementation in female reproduction. However, before clinicians recommend antioxidants, randomized controlled trials with sufficient power are necessary to prove the efficacy of antioxidant supplementation in disorders of female reproduction. Serial measurement of oxidative stress biomarkers in longitudinal studies may help delineate the etiology of some of the diosorders in female reproduction such as preeclampsia.
==== Body
Review
1. Oxidative Stress
1.1 Free radicals
Free radical species are unstable and highly reactive. They become stable by acquiring electrons from nucleic acids, lipids, proteins, carbohydrates or any nearby molecule causing a cascade of chain reactions resulting in cellular damage and disease [1-4], figure 1) . There are two major types of free radical species: reactive oxygen species (ROS) and reactive nitrogen species (NOS).
Figure 1 Mechanisms of oxidative stress-induced cell damage.
1.2 Reactive oxygen species
The three major types of ROS are: superoxide (O2•-), hydrogen peroxide (H2O2), hydroxyl (OH•). The superoxide radical is formed when electrons leak from the electron transport chain [5]. The dismutation of superoxide results in the formation of hydrogen peroxide. The hydroxyl ion is highly reactive and can modify purines and pyrimidines and cause strand breaks resulting in DNA damage [6]. Some oxidase enzymes can directly generate the hydrogen peroxide radical.
ROS have been implicated in more than 100 diseases [7-10]. They have a physiological and pathological role in the female reproductive tract. Numerous animal and human studies have demonstrated the presence of ROS in the female reproductive tract: ovaries, [11-15], fallopian tubes [16] and embryos [17]. ROS is involved in the modulation of an entire spectrum of physiological reproductive functions such as oocyte maturation, ovarian steroidogenesis, corpus luteal function and luteolysis [11,12,18]. ROS-related female fertility disorders may have common etiopathogenic mechanisms. ROS may also originate from embryo metabolism and from its surroundings.
1.3 Reactive nitrogen species
Nitric oxide (NO) is synthesized during the enzymatic conversion of L-arginine to L-citrulline by nitric oxide synthase (NOS) [19-21]. With an unpaired electron, NO, which is a highly reactive free radical, damages proteins, carbohydrates, nucleotides and lipids and, together with other inflammatory mediators, results in cell and tissue damage, low-grade, sterile inflammation and adhesions [20]. NO potently relaxes arterial and venous smooth muscles and, less strongly, inhibits platelet aggregation and adhesion. NO donors, acting as vasodilating agents, are therefore a possible therapeutic approach [22]. NO acts in a variety of tissues to regulate a diverse range of physiological processes, but excess of NO can be toxic [1,20,21,23].
Reactive nitrogen species have been associated with asthma, ischemic/reperfusion injury, septic shock and atherosclerosis [24-27]. The two common examples of reactive nitrogen species are nitric oxide (NO) and nitrogen dioxide [1,3]. NO is produced by the enzyme NO synthase. There are 3 types of nitric oxide synthase (NOS) isoenzymes in mammals involving endothelial NO synthase (NO synthase 3), neuronal NO synthase (NO synthase 1) and inducible NO synthase (NO synthase 2). Neuronal NO synthase (nNOS) and endothelial NO synthase (eNOS) are constitutive NO synthases, and responsible for the continuous basal release of NO. Inducible NO synthase (iNOS) is present in mononuclear phagocytes (monocytes and macrophages) and produces a large amount of NO. This is expressed in response to proinflammatory cytokines and lipopolysaccharides [21,23,28]. Inducible NO synthase is activated by cytokines such as, interleukin-1, and TNF-α and lipopolysaccharides. Endothelial NO synthase is expressed in thecal cells, granulosa cells, and the surface of oocyte during the follicular development. In pathological conditions, inducible NO synthase might play a major role in NO production. In most organs, inducible NO synthase is expressed only in response to immunological stimuli [29].
1.4 Antioxidants
Under normal conditions, scavenging molecules known as antioxidants convert ROS to H2O to prevent overproduction of ROS. There are two types of antioxidants in the human body: enzymatic antioxidants and non-enzymatic antioxidants [1,3].
1.5 Enzymatic antioxidants
Enzymatic antioxidants are also known as natural antioxidants, they neutralize excessive ROS and prevent it from damaging the cellular structure. Enzymatic antioxidants are composed of superoxide dismutase, catalase, glutathione peroxidase and glutathione reductase, which also causes reduction of hydrogen peroxide to water and alcohol.
1.6 Non-enzymatic antioxidants
Non-enzymatic antioxidants are also known as synthetic antioxidants or dietary supplements. The body's complex antioxidant system is influenced by dietary intake of antioxidant vitamins and minerals such as vitamin C, vitamin E, selenium, zinc, taurine, hypotaurine, glutathione, beta carotene, and carotene [1-3,30]. Vitamin C is a chain breaking antioxidant that stops the propagation of the peroxidative process. Vitamin C also helps recycle oxidized vitamin E and glutathione [31]. Taurine, hypotaurine and transferrin are mainly found in the tubal and follicular fluid where they protect the embryo from OS [17]. Glutathione is present in the oocyte and tubal fluid and has a role in improving the development of the zygote beyond the 2-cell block to the morula or the blastocyst stage [32].
1.7. Oxidative stress in female reproduction
Cells have developed a wide range of antioxidants systems to limit production of ROS, inactivate them and repair cell damage [1-3,33]. OS influences the entire reproductive span of women's life and even thereafter (i.e. menopause). It has been suggested that the age-related decline in fertility is modulated by OS [34]. It plays a role during pregnancy [35] and normal parturition [36,37] and in initiation of preterm labor [38,39]. The pathological effects are exerted by various mechanisms including lipid damage, inhibition of protein synthesis, and depletion of ATP [40]. There is some understanding of how ROS affect a variety of physiologic functions (i.e. oocyte maturation, ovarian steroidogenesis, ovulation, implantation, formation of blastocyst, luteolysis and luteal maintenance in pregnancy) [14,15,18,19,41].
ROS are a double-edged sword – they serve as key signal molecules in physiological processes but also have a role in pathological processes involving the female reproductive tract. Since the balance is maintained by the presence of adequate amounts of antioxidants, measuring levels of the antioxidants, individually or as total antioxidant capacity (TAC), has also been examined [15,18,42,43]. Superoxide dismutase (SOD) enzymes, Copper-Zinc SOD (Cu-Zn SOD) and Manganese superoxide dismutase (MnSoD) have been localized in the granulose and thecal cells of the growing follicle. Selenium dependent glutathione peroxidase activity has been demonstrated in the follicular fluid and serum of patients undergoing IVF. The expression profiles of the transcripts of the antioxidant enzymes such as superoxide dismutase, glutathione peroxidase and gamma-glutamylcysteine synthetase in both human and mouse oviducts and oocytes have also been examined [16]. There is growing literature on the effects of OS in the female reproduction with involvement in the pathophsiology of pre-eclampsia [44,45], hydatidiform mole [46-48], free radical-induced birth defects [49] and other situations such as abortions [50].
1.8 Measurement of oxidative stress
The presence of ROS and antioxidants in the female reproductive tract has been demonstrated by various methodologies in animal and human studies. A number of OS biomarkers have been investigated including superoxide dismutase, glutathione peroxidase, conjugated dienes, lipid peroxides, thiobarbituric acid reactive substances, glutaredoxin, oxidative DNA adducts, follicular fluid, NO and TAC [12,15,16,19,29,42,51-58] (Table 1).
Table 1 Oxidative stress biomarkers in female reproductive tract
Study Biomarker Methodology Measurement units
Sugino et al [41] Enzymatic antioxidants: Cu SOD, Mn SOD, catalase, glutathione peroxidase Reverse transcription-polymerase chain reaction cDNA sequences
Attaran et al [4] Total antioxidant capacity Enhanced chemiluminescence assay Trolox equivalents
Jozwik et al [15] Lipid peroxides; Malondialdehyde, conjugated dienes, Thiobarbituric acid reactive substances. Thiobarbituric acid method Micromole of malondialdehyde/L
Seino et al [52]. Oxidative DNA adducts 8-hydroxy 2-deoxyguanosine Immunocytochemical staining
Metabolites of NO (nitrite and nitrate) in peritoneal fluid are determined by nitrate reductase and the Griess reaction [20,23]. Total NO (nitrite and nitrate) levels in the serum and follicular fluid assay of NO are measured via a rapid-response chemiluminescence analyzer [29]. Various biomarkers of oxidative stress have been determined in the placenta by immunohistochemistry or western blot analysis (Table 2). Oxidative DNA adducts 8-hydroxy 2-deoxyguanosine-have been studied by immunostaining in placenta [45], in patients with IUGR (intrauterine growth retardation) and patients with preeclampsia and IUGR [45]. The basal levels of ROS in the leukocytes in whole blood can be determined using the dihydroethidium and dichlorodihydrofluorescein-diacetate probes (Table 2).
Table 2 Measurement of biomarkers of oxidative stress in pregnancy
Study Biomarkers Methodology Measurement units
Jauniaux et al [59] Immunohistochemistry Heat shock protein 70, hydroxynenal, nitrotyrosine residues. Fluorescence intensity
Wang et al, Djordjevic et al [60, 61] Antioxidant enzyme activity assays Total SOD activity, catalase activity, glutathione peroxidase activity, reduced glutathione assay Change in optical density/minutes/mg protein
Wiktor et al [62] Oxidative DNA adducts 8-hydroxy-2 deoxyguanosine Micromoles/mole of 2-deoxy guanosine
Holthe et al [63] Superoxide anion, hydrogen peroxide and peroxynitrite Dihydrethidium probe Dichlorodihydrofluorescein Dihydrorhodamine 123 Spectrophotometry/flow cytometry Nanomoles/10 min/106 cells
Ishihara et al [64] Lipid peroxidation products Isoprostane, Urinary 8-epi-prostaglandin F2lpha, assayed by gas chromatography/mass spectrophotometer analysis pg/mg of creatinine
Vaisanen-Tommiska et al [65] Nitric oxide; Greiss reaction Stable end products: nitrite/nitrate Fluorometric assay, results expressed as NOx (sum of converted nitrite and very small amount of nitrite in serum).
Buhimishi et al [66] Plasma and red blood cell glutathione content Colorimetric assay Nanomoles/mgm of haemoglobin
2. Oxidative Stress & Female Infertility
Infertility is a disease defined as "the inability to conceive following 12 or more months of unprotected sex before an investigation is undertaken unless the medical history and physical findings dictate earlier evaluation and and treatment [67]. The prevalence of female infertility ranges from 7% to 28%, depending on the age of the woman. In general, an estimated 84% of couples conceive after 1 year of intercourse, and 92% of the couples conceive after 2 years [68]. Although the frequency and origin of different forms of infertility varies, 40%–50% of the etiology of infertility studied is due to female causes [69].
A primary diagnosis of male factor infertility is made in 30% of infertile couples. Combined female and male factor infertility is responsible for 20%–30% of cases. Finally, unexplained infertility affects 15% of couples [70]. If the results of a standard infertility examination are normal, a diagnosis of unexplained or idiopathic infertility is assigned [70]. Data from the National Survey for Family Growth indicate that the number of women with impaired fecundity has increased from 1982 to 1995, an increase of 35% in the number of women. Approximately 1.3 million American couples receive medical advice or treatment for infertility every year [71]. OS has a role in etiopathogenesis of endometriosis, tubal factor infertility, and unexplained infertility. Impact of OS on ART is discussed in further sections.
2.1 Pathophysiology of oxidative stress in female reproduction
Oxygen toxicity is an inherent challenge to aerobic life [72]. ROS can modulate cellular functions, and OS can impair the intracellular milieu resulting in diseased cells or endangered cell survival. The role of ROS in various diseases of the female reproductive tract has been investigated. ROS can affect a variety of physiological functions in the reproductive tract, and excessive levels can result in precipitous pathologies affecting female reproduction. The oxidant status can influence early embryo development by modifying the key transcription factors and hence modifying gene expression [73]. Concentrations of ROS may also play a major role both in the implantation and fertilization of eggs [72]. There is an increased interest to examine the role of OS in female reproduction because it may be a major link in the infertility puzzle as well as in some reproductive organ diseases such as endometriosis. Recently, OS has been reported to have an important role in the normal functioning of the female reproductive system and in the pathogenesis of female infertility [33,74].
2.2 Cytokines, oxidative stress and female reproduction
The control of ovarian stromal cells and germ cell function is a diverse paradigm and oxidative stress may be one of the modulators of ovarian germ cell and stromal cell physiology. A number of autocrine and paracrine factors affect the modulation of various ovarian functions and steroidogenesis. Cytokines are polypeptides or glycoproteins secreted into the extra cellular compartment by the leukocytes [75]. Mammalian ovulation or follicular rupture was proposed to result from the vascular changes and the proteolytic cascade [54]. The cross talk between these two cascades is mediated by cytokines, vascular endothelial growth factor (VEGF), and ROS (both reactive nitrogen and oxygen radicals). Interleukin-1β causes nitrite to accumulate in rat ovarian dispersates, demonstrating the close interaction between cytokines and NOS [76]. OS and cytokines are proposed to be interlinked and act as intercellular and intracellular messengers in the ovary. A number of investigators have investigated the synthesis of NOS and ROS in the ovaries [21,55,58].
Defective placentation leads to placental hypoxia and reperfusion injury due to ischemia and the resultant OS triggers the release of cytokines and prostaglandins, which results in endothelial cell dysfunction and plays an important role in the development of pre-eclampsia [77,78]. TNF-α a plasma cytokine, has been demonstrated to cause the endothelial cell injury [79]. A link between OS and expression of cytokine receptors in the cytotrophoblast, vascular smooth muscle cells and endometrial cells has also been proposed, further establishing that hyperactivation of ROS may result in pre-eclampsia [80].
The activation of mononuclear phagocytes can be triggered in endometriosis by a number of factors including damaged red blood cells and the apoptotic endometrial cells. A positive correlation between concentrations of tumor necrosis factor (TNF)-α in the peritoneal fluid and endometriosis has been reported [75]. Cytokines released by the macrophages influence the redox status of the ectopic endometrium in patients with endometriosis [81]. Superoxide dismutase, glutathione peroxidase activity and lipid peroxidation levels were measured in ectopic endometrial tissue obtained from ovarian endometriomas. Superoxide dismutase activity was found to be significantly higher in the ectopic endometrium than in eutopic endometrium, and a positive correlation was seen between malondialdehyde levels and plasma 17-beta estradiol levels. TNF-α has been shown to cause up regulation of expression of Manganese (Mn) superoxide dismutase in the endometrium in vitro [82]. The antioxidant MnSOD neutralizes superoxide anions generated by cytokine TNF-α. This is a self protective mechanism against TNF-α induced oxidative stress. Estrogen and progesterone withdrawal leads to stimulation of prostaglandin F2α production via ROS-induced NFkappa β activation [83]. The mechanism of menstruation is unclear, and activation of the transcription factor NFkappa β may be a piece in the puzzle.
Ovarian epithelial cancer is the most common type of ovarian cancer. Ovarian epithelial inflammation has been suggested as an etiological factor in ovarian epithelial cancer [11,84]. The mechanisms that bring about follicular rupture result in the exposure of the ovarian surface epithelial cells to deleterious agents (e.g. free radicals and TNF-α) [85,86]. Thus, incessant ovulation and its complex articulation by OS, inflammation and cytokines repeated cyclically may be involved in the etiopathogenesis of ovarian cancer [86]. Factors that inhibit ovulation such as oral contraceptives reduce the risk of epithelial ovarian cancer [87,88]. Recent studies point towards a role of genes active in the process of metabolism of oxidation products, in the etiology of ovarian cancer [89].
2.3 Reactive oxygen species and mediators of angiogenesis
Angiogenesis is a pathophysiological process involving formation of blood vessels from preexisting vessels. The induction of angiogenesis occurs when there is a deficiency of oxygen in tissues. This process of neovascularization results from hypoxia and induction of various angiogenic factors, and it has a role to play in physiological processes such as follicular development, endometrial growth, embryo development, growth of placental vessels and wound repair [90,91]. Angiogenesis is important for cyclical regeneration of endometrium in the menstrual cycle. A complex cytokine influence at the maternal-fetal interface creates the conditions that are necessary to support embryo implantation in the endometrium [92,93]. Any imbalance between the cytokines and angiogenesis factors could result in implantation failure and pregnancy loss [94]. Critical changes occur in the vascular system, and these changes accompany follicular growth. Follicular growth, selection of dominant follicle, corpus luteum formation, endometrial differentiation and embryo formation are key processes dependent on neovascularization [90,95]. As the endometrium grows in the menstrual cycle, vessel regeneration occurs (i.e. spiral arterioles and capillaries) [96]. Estrogens promote angiogenesis in the endometrium by controlling the expression of factors such as VEGF [97]. ROS generated from NADP (H) oxidase is critical for VEGF signaling in vitro and angiogenesis in vivo [98]. Small amounts of ROS are produced from endothelial NADP (H) oxidase activated by growth factors and cytokines.
ROS that are generated in and around the vascular endothelium may play a role in normal cellular signaling mechanisms. They may also be an important causative factors in endothelial dysfunction that leads to the development of atherosclerosis, diabetes complications and ischemia perfusion injury [98,99]. The molecular mechanism by which the oxidant status of cells modulates angiogenesis is not completely understood. As our understanding the role ROS-induced angiogenesis plays in atherosclerosis and myocardial angiogenesis grows, future studies should investigate the role ROS plays in the angiogenesis in the female reproductive tract.
2.4 Reactive oxygen species and the endometrium
There is a cyclical variation in the expression of superoxide dismutase (SOD) in the endometrium. SOD activity decreases in the late secretory phase while ROS levels increase [100]. These changes have been hypothesized to be important in the genesis of menstruation and endometrial shedding. The levels of prostaglandin F2 α increase towards the late secretory phase and ROS triggers the release of prostaglandin F2 α in vitro [101]. Stimulation of the cyclooxygenase enzyme is brought about by ROS via activation of the transcription factor NFKappa β, suggesting a mechanism for menstruation [83].
2.5 ROS and endometriosis
Increased generation of ROS by activated peritoneal macrophages has been reported in the peritoneal fluid [102]. Conflicting results were reported in further studies with large patient numbers, which failed to demonstrate an antioxidant or oxidant balance [74,103]. ROS levels in peritoneal fluid of patients with endometriosis were not significantly higher than controls.
An increased titer of autoantibodies related to OS has been reported in women with endometriosis resulting in an increase in serum autoantibody titers to oxidatively modified low density lipoproteins [104]. An OS-induced increase in autoantibody titers in the peritoneal fluid has been demonstrated in women with endometriosis. Elevated levels of the marker of lipid peroxidation lysophophatidyl choline, a potent chemotactic factor for monocytes/T-lymphocytes, were seen in the peritoneal fluid of women with endometriosis [105]. Non-terminal oxidation may have a role in the pathophysiology of endometriosis. Minimally oxidized low density lipoprotein (LDL) (M-LDL) is present in peritoneal fluid of women with endometriosis in place of the terminally oxidized LDL (Ox-LDL) [106]. The ratio of lysophosphatidyl choline, a breakdown product of Ox-LDL, to phosphatidyl choline suggests M-LDL rather than Ox-LDL. Modest levels of OS induced proliferation of endometrial stromal cells in vitro, was inhibited by antioxidants [107]. RU486, a potent antiprogestational agent with antioxidant activity also decreased proliferation of epithelial and stromal cells [108].
2.6 Reactive oxygen species and the ovary
Markers of oxidative stress such as superoxide dismutase, Cu-Zn superoxide dismutase, Mn superoxide dismutase, glutathione peroxidase, γ glutamyl synthetase and lipid peroxides have been investigated by immunohistochemical localization, m-RNA expression and thiobarbituric acid method [4,14,41]. The expression of various biomarkers of OS has been demonstrated in normal cycling human ovaries [13,14]. All follicular stages have been examined for SOD expression including primordial, primary, preantral and nondominant antral follicles in follicular phase, dominant follicles and atretic follicles [14]. ROS may have a regulatory role in oocyte maturation, folliculogenesis, ovarian steroidogenesis and luteolysis. There is a delicate balance between ROS and antioxidant enzymes in the ovarian tissues. The antioxidant enzymes neutralize ROS production and protect the oocyte and embryo.
The presence of superoxide dismutase in the ovary, revealed intense staining by immunohistochemistry in the theca interna cells in the antral follicles [13]. Antibody to Ad4-binding protein (Ad4BP) was utilized to localize Ad4BP in the nuclei of theca and granulosa cells. Ad4BP is a steroidogenic transcription factor that induces transcription of the steroidogenic P450 enzyme. Thus, it controls steroidogenesis in the ovaries. The correlation between Ad4BP and superoxide dismutase expression suggests an association between OS and ovarian steroidogeneis [14].
Both human granulosa and luteal cells respond to hydrogen peroxide with an extirpation of gonadotropin action and inhibition of progesterone secretion [11]. The production of both progesterone and estradiol hormones is reduced when hydrogen peroxide is added to a culture of human chorionic gonadotropin-stimulated luteal cells. Hydrogen peroxide lowers both cAMP dependent and non-cAMP dependent steroidogenesis [109]. The role of hCG (human chorionic gonadotropin) in the expression of the antioxidant enzyme superoxide dismutase (SOD) has been investigated. Corpora lutea collected from patients at hysterectomy and surgery for ectopic pregnancy were studied [14]. The Cu-Zn SOD expression in the corpora lutea paralleled levels of progesterone and these levels rose from early to mid luteal phase and decreased during the regression of the corpus luteum. However, in the corpus luteum from pregnant patients, the mRNA expression for Cu-Zn superoxide dismutase was significantly higher than that in midcycle corpora lutea. This enhanced expression of luteal Cu-Zn SOD may be due to hCG and hCG may have an important role in maintenance of corpus luteal function in pregnancy.
Levels of three oxidative stress biomarkers, conjugated dienes, lipid hydroperoxide and thiobarabituric acid were determined in preovulatory follicles. Concentration gradient was found to exist as levels of all three markers were significantly lower in the follicular fluid compared with serum levels [15]. The preovulatory follicle has a potent antioxidant defense, which is depleted by the intense peroxidation [15]. Glutathione peroxidase may also maintain low levels of hydroperoxides inside the follicle and thus play an important role in gametogenesis and fertilization [42].
2.7 Nitric Oxide synthase in female reproduction
The production of a viable oocyte is modulated by a complex interaction of endocrine, paracrine and autocrine factors leading to follicular maturation, granulosa cell maturation, ovulation and luteinization. Many hormonal and paracrine factors determine oocyte competence and embryo quality. Steroid hormones and local autocrine and paracrine factors influence ovarian stromal cells. The gonadotropins act through complex, multiple local signaling pathways. Cyclic AMP is thought to be the second messenger to bring about the effect of luteinizing hormone and follicular stimulating hormone [110]. In turn, cyclic AMP may activate other signaling pathways. Cyclic guanosine monophosphate (cGMP)-a cyclic nucleotide has also been proposed as a second messenger pathway. The effects of NO are proposed to be mediated through cGMP as a second messenger or by generation of ROS resulting from interaction of NO with superoxide radicals [111].
NO generated by macrophages in response to invading microbes acts as an antimicrobial agent [21]. Neurons, blood vessels and cells of the immune system are integral parts of the reproductive organs, and in view of the important functional role that NO plays in those systems, it is highly likely that NO is an important regulator of the biology and physiology of the reproductive system. NO has established itself as a polyvalent molecule that plays a decisive role in regulating multiple functions within the female as well as male reproductive system [21]. As a final immune effector, NO generated by inducible NO synthase, kills pathogens and abnormal cells but may play a detrimental role by damaging normal host tissue and cells, especially when inducible NO synthase is persistently expressed [20].
2.8 Nitric oxide synthase and fallopian tubes
The presence of NO synthase enzymes, both the constitutive and the inducible forms was delineated by immunhistochemistry and the presence of NADPH-diaphorase activity in human tubal cells [112,113]. The production of NO was demonstrated by positive NADPH diaphorase activity in the human fallopian tube. An endogenous NO system exists in the fallopian tubes [114]. NO has a relaxing effect on smooth muscles and it has similar effects on tubular contractility. Deficiency of NO may lead to tubal motility dysfunction, resulting in retention of the ovum, delayed sperm transport and infertility. Infertility associated with urogenital tract infections is associated with diminished sperm motility and viability. Increased NO levels in the fallopian tubes are cytotoxic to the invading microbes and also may be toxic to spermatozoa [114].
2.9 Nitric oxide synthase, endometrium, and myometrium
Expression of endothelial and inducible NO synthase have been demonstrated in the human endometrium [115], and the endometrial vessels [116]. Endothelial NO synthase, originally identified in vascular endothelial cells, is distributed in glandular surface epithelial cells in the human endometrium. NO also regulates the microvasculature of the endometrium and is important in menstruation.
Endothelial NOS like immunoreactivity has been reported in the endothelial cells lining the vessels, endometrium and endometrial glandular epithelial cells and myometrium [117]. Inducible NOS like immunoreactivity was detected in decidualised stromal cells and also expressed in tissues from first trimester of pregnancy. Thus, NO has a role to play in decidualisation of the endometrium and preparation of the endometrium for implantation.
Highest levels of expression of endothelial NOS mRNA have been reported in the late secretory phase of the endometrium [115]. In addition, NO might participate in the regulation of uterine contraction [118]. In normal fertile woman, the contractions increase throughout the proliferative and periovulatory phases, and decrease in the secretory phase. From this point, NO is synergistic with progesterone and might relax uterine contraction in the secretory phase in a paracrine fashion. Nitric oxide regulates the endometrial, myometrial and microvascular functions in the uterus by paracrine functions. The effects of the antiprogestational agent, Mifepristone on the endothelial NOS expression in the endometrium were investigated during the implantation phase [119]. Mifepristone inhibited the endometrial glandular epithelial eNOS expression and did not affect the endothelilal eNOS. The results supported the role of epithelial eNOS in endometrial receptivity in the perimplantation period. Significant increase in the levels of serum metabolites of NOS amongst users of progestin only contraceptives has been reported [120]. A positive correlation was also demonstrated between NO levels and prolonged/heavy bleeding. Thus NO may be involved in the pathophysiology of menorrhagia.
NO plays a significant role in pregnancy and labor. Expression of inducible NOS was highest in patients with preterm pregnancy and not in patients in term labor. The expression of these enzymes decreased by 75% at term and was barely detectable in preterm in labor patients or term labor patients [121] reiterating that NO has a role in maintenance of uterine quiescence. However other authors have reported conflicting results. Induction of iNOS expression was demonstrated in fetal membranes in labour and in in-vitro studies [122]. Higher NO metabolite concentrations were demonstrated in amniotic fluid collected from women in labor than in non-laboring patients, both at term (15.4 ± 1.6 vs. 6.8 ± 0.6 microM/mg creatinine) and preterm (16.7 ± 2.0 vs. 7.0 ± 0.8 microM/mg creatinine) [123].
2.10 Nitric oxide synthase and endometriosis
Endometriosis is found in about 35% of infertile women who have laparoscopy as part of their infertility workup [71]. Production of ROS by peritoneal fluid mononuclear cells was reported to be associated with endometriosis [75]. Low levels of NO are important in ovarian function and implantation and cause relaxation of oviduct musculature [112]. High levels of NO are reported as having deleterious effects on sperm motility, toxic to embryos and inhibit implantation [124,125]. In vitro fertilization, a process that avoids contact of gametes and embryos with potentially toxic peritoneal and oviductal factors associated with endometriosis (e.g., NOS, ROS) improves the chances of conception in these women. NO is a free radical with deleterious effects and is an important bioregulator of apoptosis [126]. Activation of polymorphonuclear leucocytes and macrophages leads to increased production of ROS [102]. Increase in number and activity of macrophages is accompanied by release of more cytokines and other immune mediators, such as NO. This was initially implicated in low-grade inflammation, while elevated peritoneal NO is consistent with the increased number and activity of macrophages [20]. High levels of NO, such as those produced by macrophages, can negatively influence fertility in several ways. Changes in peritoneal fluid, an environment that hosts the process of ovulation, gamete transportation, sperm oocyte interaction, fertilization, and early embryonic development, might affect all these steps of reproduction [2,20,127]. Studies investigating the association of nitric oxide levels and lipid peroxides and reactive oxygen species in peritoneal fluid did not find any significant difference in patients with or without endometriosis [103,128,129] Conflicting results were obtained in studies conducted by Szczepanska et al [2]. The total antioxidant capacity reduced and the individual antioxidant enzymes like superoxide dismutase were significantly lower in the peritoneal fluid of women with endometriosis-associated infertility. The lipid peroxide levels were the highest amongst patients with endometriosis suggesting role of ROS in the development of the disease process [2]. There is a cyclical expression of mRNA of NOS in the epithelial glands of the human endometrium. Higher amounts of NO and NOS are seen in the endometrium of women with endometriosis [28,130,131]. NOS expression in the ectopic endometrium of patients with adenomyosis is continuous throughout the menstrual cycle [132].
Peritoneal fluid NO levels, peritoneal macrophage NOS activity, and peritoneal macrophage inducible NOS protein expression has been examined in women with endometriosis-associated infertility. Peritoneal macrophages express higher levels of NOS, have higher NOS enzyme activity, and produce more NO in response to immune stimulation in vitro [23]. High levels of NO adversely affect sperm, embryos, implantation, and oviductal function, indicating that reduction in the peritoneal fluid NO production or blocking NO effects may improve fertility in women with endometriosis [23]. However, generation of peroxynitrite by ectopic endometrium has been reported in patients with adenomyosis. Expressions of endothelial and inducible NO synthase and peroxynitrite generation was markedly reduced after GnRH agonist therapy, supporting their potential role in the pathophysiology of adenomyosis [132]. Serum NO levels are suppressed by GnRH agonist and upregulated by gonadotropin stimulation during controlled ovarian stimulation in female partners from couples with male factor infertility [133]. Maximal levels were measured at the time of ovulation in the same study. Elevated NO production was not demonstrated in patients with ovarian hyperstimulation.
Increased levels of NO were demonstrated in the peritoneal fluid of patients with endometriosis [20,23]. It has also been hypothesized that ROS may have a role in formation of adhesions associated with endometriosis [134]. Patients with endometriosis show a higher 8-hydroxy 1-deoxyguanosine index compared to patients with other causes of infertility, such as tubal, male factor or idiopathic causes [52].
Expression of NOS is elevated in patients with endometriosis, and a common polymorphism of exon 7 at nucleotide 894 in the endothelial NOS gene may be associated with endometriosis [135]. Hence variations in the expression of the eNOS gene may be involved in endometrial angiogenesis and thus modulate the process of endometriosis.
Expression of endothelial NO synthase in the endometrium of patients with endometriosis or adenomyosis is persistently marked throughout the menstrual cycle [132]. Many investigators have reported increased expression of endothelial NOS in the glandular endometrium in patients with endometriosis [28,130]. Inducible NOS isoform is elevated in tissues of patients with endometriosis [131]. The endometrial development affects embryo implantation, and inconsistent development between endometrium and embryo could impede embryo implantation. Nitric oxide affects fecundity in endometriosis and adenomyosis [136]. Significant differences are seen in the uterine hyperperistalsis and dysperistalsis in patients with endometriosis compared with the control groups, and this may be responsible for disturbed sperm transport and reduced fertility [137].
Various cytokines secreted from endometrial cells, immune cells, or macrophages stimulate endothelial NO synthase to release NO [3,28,136]. These abnormal immune responses might eventually stimulate macrophages and/or endometrial cells to persistently produce a large amount of NO and inhibit implantation [138]. Increased expression of endothelial NO synthase has been reported throughout the menstrual cycle in the endometrium of women with endometriosis [139].
2.11 Nitric oxide synthase and the ovary
Ovarian follicullogenesis not only involves gonadotropins and the steroids, but it also involves local autocrine and paracrine factors. Nitric oxide radical is one of the local factors involved in ovarian follicullogeneis and steroidogenesis. Nitric oxide acts through activation of various iron containing enzymes. It binds to the heme containing enzyme guanylate cyclase, which activates the cyclic nucleotide cyclic-GMP [110]. Plasma concentrations of nitrate monitored during the follicular cycle, have revealed peak levels at ovulation [133,140]. Nitric oxide inhibits ovarian steroidogenesis [52]. The presence of endothelial NO synthase in human corpora lutea and the expression has been reported in the mid and early luteal phase and to a lesser extent in the late luteal [53]. Nitric oxide inhibits steroidogenesis in the corpus luteum and has luteolytic action mediated through increased prostaglandins and by apoptosis [53,141].
Follicular fluid NO seems to be produced by either endothelial NO synthase or induced NO synthase. However, in the normal physiological conditions follicular fluid NO seems to be synthesized from granulosa cells by endothelial NO synthase, since in isolated human follicular cells at least 90% of cells are granulosa cells even though macrophages and lymphocytes are present as well. In patients undergoing IVF, a positive correlation was determined between follicular fluid nitrate/nitrite levels and the follicular volume as well as the serum estradiol concentration [142]. In contrast to these findings, Manau et al, found no association between follicular fluid nitrite/nitrate levels and parameters of ovarian response [143]. Biomarkers like serum nitric oxide measurements cannot be used as predicting success with IVF [143,144]. Serum nitrate concentration may not be a good biomarker because of the short half-life of NO. Follicular blood flow was found to be a better prognostic factor for predicting successful outcomes with IVF than follicular NO levels [138]. Follicular fluid NO levels were altered in patients with infertility associated diseases. NO follicular fluid levels were significantly higher in patients with endometriosis or hydrosalpinx compared to patients with tubal obstruction [29]. No correlation was reported between the follicular NO levels and follicle maturity or follicle quality.
Some studies have demonstrated the relationship between NO concentrations in follicular growth and programmed follicular cell death (apoptosis). Folliculogenesis involves the participation of both growth of the follicle and apoptosis. Nitric oxide regulates both of these processes [21]. Sugino et al studied the role of nitric oxide in follicular atresia and apoptosis, in patients undergoing IVF and found that the smaller follicles had significantly elevated percentage of apoptotic granulosa cells with nuclear fragmentation [58]. Low concentrations of NO may prevent apoptosis, however pathologically high concentrations of NO, as well as increased superoxide generation by NO synthase due to lack of arginine, may promote cell death by peroxynitrite generation [21]. Nitric oxide involvement in various ovarian functions has been suggested. The presence of NO in the follicular fluid and the expression of NO synthase in follicles and corpus luteum have been reported [19,141,143,145].
Plasma concentration of NO was shown to increase in the follicular phase compared with the secretory phase and peaked at midcycle [140]. Nitric oxide elicited a positive effect on women with poor ovarian response compared to controlled ovarian stimulation [146]. Upregulated NO is harmful to implantation and pregnancy among patients with tubal factor infertility after controlled ovarian stimulation [147]. Serum NO levels were elevated amongst nonpregnant patients with tubal or peritoneal factor infertility [124].
Follicular fluid NO level is not associated with maturity or quality of oocyte and no significant differences were seen in concentrations of NO of follicular fluid among large, medium, or small follicle size. Higher TNF-α concentrations in follicular fluid correlated with poor oocyte quality [29]. Whereas, follicular fluid nitrite or nitrate levels were significantly lower in follicles containing mature oocytes that fertilized compared with those that did not [148]. Follicular NO has been reported to correlate negatively with embryo quality and the rate of embryo cleavage [124,147,148]. The beneficial effects of NO donors in patients with intrauterine growth retardation (IUGR) and inhibition of pre-term labor has been studied [149,150]. Using a nitroglycerine (NTG) patch, which is a NO donor, did not significantly affect the final outcome in patients undergoing in-vitro fertilization. In addition, neither placebo nor the nitroglycerine patch improved the flow resistance in the uterine artery [22]. NO donors and elevated serum NO was associated with implantation failure resulting in decreased fertility [138].
3. Assisted reproduction
Assisted reproductive technique (ART) involves the direct manipulation of the human oocytes, sperm or embryos outside the body, to establish a pregnancy. A variety of causative factors of infertility can be indications for ART, i.e. tubal factor, endometriosis, male factor and unexplained infertility [151,152]. Assisted reproductive techniques offer excellent opportunities to infertile couples for achieving pregnancy. There may be multiple sources of ROS in an IVF setting including the oocytes, cumulus cell mass, or spermatozoa used for insemination [153].
3.1 Oxidative stress and its impact on ART
The follicular fluid microenvironment has a crucial role in determining the quality of the oocyte. This in turn impacts the fertilization rate and the embryo quality. Oxidative stress markers have been localized in the follicular fluid in patients undergoing IVF/embryo transfer (ET) [4,51,154,155]. Low intrafollicular oxygenation has been associated with decreased oocyte developmental potential as reflected by increasing frequency of oocyte cytoplasmic defects, impaired cleavage and abnormal chromosomal segregation in oocytes from poorly vascularised follicles [156]. ROS may be responsible for causing increased embryo fragmentation, resulting from increasing apoptosis [157]. Thus increasing ROS levels are not conducive to embryo growth and result in impaired development. Current studies are focusing on the ability of growth factors to protect in vitro cultured embryos from the detrimental effects of ROS such as apoptosis. These growth factors are normally found in the fallopian tubes and endometrium. The factors being investigated are: Insulin growth factor (IGF)-1, and Epidermal growth factor (EGF) in mouse embryos, which in many respects are similar to human embryos [158].
Exogenous gonadotropin has a stimulatory effect on the follicular content of iron, which is a potent oxidant, catalyses generation of free radicals in Haber-Weiss reaction. Iron overload in thalassemia acts as a redox-active center and there is resultant increase in the production of free radicals [159]. Increase in free radicals was reported in follicular fluid of patients with thalassemia. The spectrum of initial hypogonadism and later gonadal failure in thalassemia, results from the injury mediated by free radicals.
Increase in TAC was seen in follicular fluid of oocytes that later were successfully fertilized. Therefore, lower total antioxidant capacity is predictive of decreased fertilization potential [154]. Lower levels were associated with increased viability of the embryos until the time of transfer, and the fertilization potential decreased with decreasing concentrations of total antioxidants. Similarly mean glutathione peroxidase levels were increased, in follicles yielding oocytes that were subsequently fertilized [42]. Levels of ROS were reported to be significantly lower in patients who did not become pregnant compared with those who became pregnant [4]. Thus intrafollicular ROS levels may be used as a potential marker for predicting success with IVF. Studies determining normal TAC levels of the follicular fluid in unstimulated cycles are lacking.
In addition levels of selenium in follicular fluid of women with unexplained infertility were lower than those in women with tubal factor or male factor infertility [42]. Higher levels of superoxide dismutase activity were present in fluid from follicles whose oocytes did not fertilize compared with those that did [12]. These discrepancies may be due to the fact that the studies measured different parameters. The effects of follicular OS on oocyte maturation, fertilization and pregnancy have also been studied [51]. Patients who became pregnant following IVF or ICSI had higher lipid peroxidation levels and TAC. Both markers were unable to predict embryo quality. Pregnancy rates and levels of lipid peroxidation and TAC demonstrated a positive correlation.
OS in follicular fluid from women undergoing IVF was inversely correlated with the women's age [160]. Using a thermochemiluminescence assay, the slope was found to positively correlate with maximal serum estradiol levels, number of mature oocytes and number of cleaved embryos and inversely with the number of gonadotropin ampoules used. The pregnancy rate achieved was 28% and all pregnancies occurred when the thermochemiluminescence amplitude was small. This is in agreement with another study that reported minimal levels of OS were necessary for achieving pregnancy [51]. Follicular fluid ROS and lipid peroxidation levels may be markers for success with IVF.
Oocyte quality is a very important determining factor in the outcome of IVF/ET. 8-hydroxy-2-deoxyguanosine is a reliable indicator of DNA damage caused by oxidative stress. This compound is an indicator of OS in various other disease processes i.e. renal carcinogenesis, and diabetes mellitus. Higher levels of 8 hydroxy 2-deoxyguanosine were associated with lower fertilization rates and poor embryo quality [52]. Higher levels of 8-hydroxy 2-deoxyguanosine are also seen in granulosa cells of patients with endometriosis, and this may impair the quality of oocytes.
Other OS markers such as thiobarbituric acid-reactive substances, conjugated dienes and lipid hydroperoxides have been studied in the preovulatory follicular fluid [15]. No correlation was seen between these markers and IVF outcome (fertilization rates or biochemical pregnancies) [15]. A potent antioxidant system may be present in the follicular fluid as indicated by low levels of all 3 biomarkers of oxidative stress in the follicular fluid. A recent chemiluminescence study examined the follicular fluid obtained from patients undergoing IVF. Hydrogen peroxide was utilized for the induction of chemiluminescence. The study found that a delicate balance was maintained by pro-oxidant/antioxidants in the follicular fluid [161].
Smoking has been associated with prolonged and dose-dependent adverse effects on ovarian function [162]. According to a meta-analysis, the overall value of the odds ratio for the risk of infertility associated with smoking was 1.60 [95% confidence interval (CI) 1.34–1.91]. ARTs, including IVF, are further shedding light on the effects smoking has on follicular health. Intrafollicular exposure to cotinine increases lipid peroxidation in the follicle [155]. Carotenoids have gained attention because they are similar to vitamin E in that they are very potent antioxidants and react with ROS; the presence of carotenoids has been demonstrated in the follicular fluid [163]. Concentrations of carotenoids, retinol and alpha tocopherol were found to be significantly higher in follicular fluid than in plasma.
Melatonin was investigated as a drug to improve oocyte quality in patients failing to get pregnant in earlier IVF cycles because of poor quality oocytes [164]. A significant reduction in the number of degenerate oocytes was reported, and the number of fertilized embryos increased. Increased follicular concentrations of melatonin reduced lipid peroxide concentration and may have prevented DNA damage.
3.2 Redox and early embryo development
Physiological levels of redox may be important for embryogenesis. Overproduction of ROS is detrimental for the embryo, resulting from impaired intracellular milieu and disturbed metabolism [17,165]. Superoxide anion, hydrogen peroxide and hydroxyl radical, can have detrimental effects on the fetus. Oxidative stress can be generated by spermatozoa, leucocytes, and by events such as sperm mediated oocyte activation and activation of the embryonic genome [165]. The ROS generation can result from oxidative phosphorylation occurring in the mitochondria. Electrons leak from the electron transport chain at the inner mitochondrial membranes. These electrons are transferred to the oxygen molecule, resulting in an unpaired electron in the orbit. This leads to the generation of the superoxide molecule. The other points of generation of ROS are the cytoplasmic NADPH-oxidase, cytochrome p450 enzymes and the xanthine oxidoreductase enzymes. Excessive OS can have deleterious effects on the cellular milieu and can result in impaired cellular growth in the embryo or apoptosis resulting in embryo fragmentation. Thus OS mediated damage of macromolecules plays a role in fetal embryopathies. Deficient folate levels in the mother result in elevated homocysteine levels. Homocysteine induced oxidative stress has been proposed as a potential factor for causing apoptosis and disrupting palate development and causing cleft palate [166]. Oxidative stress mediated damage of the macromolecules has been proposed as a mechanism of thalidomide induced embryopathy and other embryopathies [167,168].
Hyperglycemia/diabetes induced down regulation of cycloxygenase-2 gene expression in the embryo results in low PGE2 levels and diabetic embryopathy [169]. The protective role of the enzyme G6PD (Glucose 6-phophate dehydrogenase) against oxidative stress has been demonstrated in an animal study and embryopathies were prevented by protecting the embryos against oxidative stress [170].
3.3 Effect of oxygen concentration on in-vitro embryo development
Early embryo development in mammals occurs from fertilization through differentiation of principal organ systems in a low oxygen environment [168]. A marginal improvement in preimplantation embryonic viability has been reported under low oxygen concentrations in patients undergoing IVF and ICSI [171]. Lower concentrations of oxygen in in-vitro culture of porcine embryos decreased the H2O2 content and resulted in reduced DNA fragmentation, which thereby improved developmental ability [172]. The higher oxygen concentration of 20% have been associated with lower developmental competence. Accelerated development was seen under low (5%) oxygen concentrations.
3.4 Role of ROS in IVF media
ROS may be generated endogenously or exogenously, but either way it can affect the oocyte and embryo. IVF culture media may be the exogenous site of ROS generation affecting the oocytes and the preimplantation embryo. There are some specific events in embryo development associated with a change in the redox state. It has been suggested that redox may have a causative role in sperm mediated oocyte activation, embryonic genome activation and embryonic hatching from the zona pellucida [165]. Higher day 1 ROS levels in culture media were associated with delayed embryonic development, high fragmentation and development of morphologically abnormal blastocysts after prolonged culture. A significant correlation was reported between increased ROS levels in Day1 culture media and lower fertilization rates in patients undergoing ICSI [153]. Lower ROS levels were associated with higher fertilization rates, indicating the physiological relevance of low levels of ROS.
Incubation of poor quality embryos was associated with a decline in TAC in the preimplantation embryo culture medium after 24 hours incubation. Poor quality embryos may be associated with increased generation of ROS [173]. HEPES [(2-hydroxyethyl) piperazine-1-ethanesulfonic acid] was found to be the most potent protector compared to human tubal fluid media and polyvinyl alcohol against DNA damage occurring in spermatozoa, as determined by plasmid relaxation assay which measures the plasmid DNA damage [174]. IVF media supplemented with human serum albumin (10 mgm/mL), glucose (2.78 Mmol), 1% polyvinyl alcohol, 5% polyvinylpyrrolidone, sucrose (100 mM), 60% Percoll, human tubal fluid, human tubal fluid media, catalase (1 and 10 IU), and HEPES (21 mMol) scavenge ROS and confer protection from DNA damage [174].
3.5 Strategies to overcome oxidative stress in infertility/ART
Considerable interest has been generated in the use of antioxidants to overcome the adverse and pathological results of OS. Oxidative stress leads to luteal regression, [43] resulting in a lack of luteal support to a pregnancy [33]. OS can damage oocytes in developing follicles, oocytes and spermatozoa in the peritoneal cavity, or embryo in fallopian tube [17,153] or through redox (pro-oxidant and antioxidant) imbalance. OS can be overcome by reducing generation of ROS or increasing the amounts of antioxidants available. The literature contains studies that used nutritional supplements and antioxidants like vitamin C supplementation to protect against ROS and OS. However, there is lack of consensus on the type and dosage of antioxidants to be used. Clinical evidence on the benefits of antioxidant supplementation is equivocal.
Current evidence supports the use of systemic antioxidants for management of selected cases of male infertility [175]. Randomised controlled trials investigating antioxidants in female infertility are few and lack power because of the small patient numbers. In a recent randomized controlled, multi-center study, the effect of vitamin C supplementation (750 mg/day) in patients with a luteal phase defect was reported; pregnancy rates were higher in the treatment group than in the controls [176]. Similarly, concentrations of antioxidants were found to be significantly lower in women with a history of recurrent miscarriages and luteal phase defects than in healthy women [177]. Vitamin C concentrations were higher in the follicular fluid of patients supplemented with vitamin C than that of the controls. Pregnancy rate was higher in the supplemented group than in the control group although the difference was not statistically significant [178].
In a double blinded, placebo-controlled pilot study, the impact of a nutritional supplement containing vitamin E, iron, zinc, selenium, L-arginine was examined [179]. The mean mid-luteal progesterone levels increased from 8.2 ng/mL to 12.8 ng/mL, (p = 0.08), and the patients had a significant increase in ovulation and pregnancy rates (33% pregnant, p < 0.01) [179]. In a study looking at short-term supplementation with high doses of ascorbic acid during the luteal phase in IVF, the clinical pregnancy rate and implantation rate did not improve [180]. There is lack of consensus on antioxidant supplementation in idiopathic infertility and randomized controlled trials need to be designed with sufficient power and patient numbers to investigate this issue.
Fertilization and embryo development in vivo occurs in an environment of low oxygen tension [168]. During ART, it is important to avoid conditions that promote ROS generation and expose gametes and embryos to ROS. During culture, low oxygen tension is more effective at improving the implantation and pregnancy rate than higher oxygen tension [181]. Similarly, higher implantation and clinical pregnancy rates are reported when antioxidant supplemented media is used rather than standard media without antioxidants. Metal ions can sometimes result in the production of oxidants. Metal ions can also increase the production of ROS directly through the Haber-Weiss reaction. It may be useful to add metal ion chelating agents to the culture media to decrease the production of oxidants [181].
Amino acids added to the IVF media also have antioxidant properties. Adding ascorbate during cryopreservation reduces hydrogen peroxide levels and thus the oxidative distress in mammalian embryos [182]. As a consequence, the embryo development improved with enhanced blastocyst development rates. A significant negative association has been reported between duration of smoking and fertilization rates in IVF procedures. Eliminating the smoking factor would help improve fertility and ART outcomes [178]. Because a history of smoking is associated with high concentrations of OS, in-vivo antioxidants can be recommended in infertile women who smoke [155].
Follicular vascularity determines the intrafollicular oxygen content and the developmental potential of the oocyte [156,183]. Intrafollicular hypoxia results in chromosomal segregation disorders and deleterious mosaicisms in the embryo. Sildenafil, an inhibitor of phosphodiesterase enzyme, prevents the breakdown of cGMP and potentiates the effects of NO on vascular smooth muscle. Vaginal Sildenafil and L-arginine have been investigated for their potential to improve intrafollicular blood flow by potentiating the actions of NO on vascular smooth muscle. It augments the effect of NO in inducing vasodilatation and thus improving uterine blood flow. A recent study reported that Sildenafil, administered on day 3 of the menstrual cycle, appeared effective in improving uterine artery blood flow and endometrial development [184]. The same group in a subsequent cohort of 105 patients with infertility and previous failures at IVF were able to achieve higher implantation and pregnancy rates with vaginal Sildenafil [185]
Oral L-arginine supplementation in poor responder patients, during controlled ovarian stimulation may improve ovarian response, endometrial receptivity and pregnancy rate by increasing the flow around the follicles, and uterine flow [146]. Follicular fluid concentrations of nitrite/nitrate inversely correlated with embryo quality. Although the embryo quality was poor, L-arginine supplementation in normally responding patients resulted in higher follicular fluid arginine levels compared to the poor responders and increased follicular recruitment [147]. NO derivatives in higher doses in follicular fluid may cause cytostatic and cytotoxic effects and may have detrimental consequences on embryo quality, implantation and pregnancy rate.
Mechanical removal of ROS in IVF/ET has been examined [186]. Cumulus oophorus rinsing is performed to overcome the deleterious effects of ROS in patients with ovarian endometriosis [186]. ROS has deleterious effects on both the oocyte and the embryo quality. The deleterious effects of TNF-α cytokines and reactive oxygen species, which were increased in the peritoneal fluid of patients with endometriosis and unexplained infertility, were prevented, by the rinsing procedure.
3.6 Critical review of OS, ovary and ART
A comprehensive review of the published literature reveals that the role of oxidative stress is controversial due to the differences in the nature of materials examined, (i.e. follicular fluid, embryos, and culture mediums). We can conclude the number of articles on oxidative stress in the last 5 years have significantly increased compared to the previous 5 years indicating that more studies are being conducted to understand the role of oxidative stress in female reproduction. The effects of ROS studied and its ability to influence female reproduction have been studied on various endpoints in terms of the oocyte, fertilization, embryo and pregnancy. Different markers of oxidative stress are reported in various studies and the sensitivity and specifity of the various biomarkers are not known. While some research is focused on studying the antioxidant capacity others focus on studying and determining the levels of oxidative stress markers. Also, there has been an assumption in the studies measuring the amount and type of antioxidants that there is an inverse correlation between oxidative stress markers and antioxidants. These studies have also variations because some have measured the total antioxidant capacity and some have measured individual enzymes like superoxide dismutase. Further studies need to be designed to validate the results of the earlier studies, with elimination of various factors leading to bias. Eliminating the bias will make the comparison of different studies acceptable and provide support to the evidence based approach. The biomarkers of oxidative stress that are studied should be similar across different studies to make the results comparable. Prospective, randomized controlled trials with stringent inclusion criteria are needed to determine the effects of antioxidants in overcoming redox in infertility patients.
4. Age related fertility decline, Menopause and ROS
There is an age related decline in the number and quality of follicles in females. ROS may damage the oocytes [187]. The age related decline in oocyte quality also results in increased incidence of congenital anomalies in children. The ageing of the oocytes affects many biochemical pathways which have a deleterious effect on pre- and post implantation development of the embryo [188]. The pre- and postovulatory ageing of the oocytes have also been associated with congenital anomalies, behavorial alterations, and learning disabilities in later life and constitutional diseases such as diabetes mellitus, and schizophrenia. Oxidative stress occurs at menopause because of loss of estrogens, which have antioxidant effect on low-density lipoproteins. Estrogens confer cardioprotection by lowering protein oxidation and antioxidant properties [189]. Diminished antioxidant defense is associated with osteoporosis in post-menopause. Modulation of the estrogen receptors α and β has been reported to be effected in vitro by oxidative stress [190].
5. Oxidative stress and pregnancy
5.1 Oxidative stress and pre-eclampsia
Pre-eclampsia is associated with severe maternal and fetal morbidity and mortality [191]. Overall pre-eclampsia complicates 5% of all pregnancies and 11% of all first pregnancies. Recent evidence suggests the role of oxidative stress in pre-eclampsia. There is a reduced antioxidant response inpatients with pre-eclampsia [192,193] and reduced levels of antioxidant nutrients [194] and increased lipid peroxidation [45,194].
5.2 Placental oxidative stress and Pre-eclampsia
Incomplete trophoblast invasion leads to failure of conversion of thick walled tortous spiral arteries to low resistance flaccid sinusoidal vessels [195,59]. The incomplete invasion results in impaired placental perfusion. The hypoxia/reperfusion injury leads to increased expression of xanthine oxidase and NADP (H) oxidase and resultant increased generation of superoxide anion. The increased generation of pro-oxidants tilts the balance in favor of oxidative stress, which results in increased lipid peroxidation. Biomarkers of lipid peroxidation are elevated in the placenta [45,60].
5.3 Interventions to overcome oxidative stress in pre-eclampsia
There is currently no accepted method of prevention of pre-eclampsia. Antioxidants vitamin C and vitamin E have been studied in some trials for preventing pre-eclampsia. Early intervention at 16–22 weeks of pregnancy with supplementation of vitamin E and C resulted in significant reduction of pre-eclampsia in the supplemented group [196]. Supplementation in women with established pre-eclampsia did not result in any benefit [197]. Recent report of a randomized trial failed to find beneficial effects of vitamin C and E supplementation in preventing preeclampsia [198].
5.4 Redox and miscarriage
Human placenta is classified as hemomonochorial. Maternal blood directly bathes the fetal trophoblast. Establishment of the maternal placental circulation is influenced by the trophoblastic invasion. Extravillous trophoblastic invasion transforms the small caliber high resistance spiral arteries into large caliber, low resistance, and high capacitance uteroplacental arteries. Abnormal placentation has been implicated in the pathogenesis of pre-eclampsia and miscarriage [199]. Pre-eclampsia is unique to human species and miscarriage is very rare in other species [200]. Abnormal placentation leads to placental oxidative stress with resultant detrimental effects on the syncitiotrophoblast and it has been proposed as a mechanism involved in the etiopathogenesis of abortion. A sharp peak in the expression of the markers of oxidative stress in the trophoblast was detected in normal pregnancies and this oxidative burst if excessive was speculated to be a cause of early pregnancy loss [168].
The etiology of recurrent pregnancy loss remains unclear and is a scientific challenge. Oxidative stress may have a role in the etiology of recurrent pregnancy loss with no known etiology. Glutathione and glutathione transferase family of enzymes have been investigated in patients who experience recurrent abortions [201,202]. Glutathione and glutathione peroxidase are both antioxidants that neutralize the free radicals and lipid peroxides to maintain the intracellular homeostasis and redox balance.
The etiology of recurrent pregnancy loss is multifactorial and involves genetic and environmental factors [203]. In a large case controlled study, gene polymorphisms of enzymes of the glutathione family, glutathione S-transferase class mu (GSTM1) were studied. Elevated risk of recurrent pregnancy loss was found to be associated with the GSTM1 genotype null polymorphism, in patients with recurrent pregnancy loss. Elevated glutathione levels in pregnant patients with history of recurrent pregnancy loss were associated with poor outcomes (i.e. abortion) [201].
5.5 Term labor and the role of oxidative stress
There is increased generation of free radicals superoxide and nitric oxide in pregnancy, which results in oxidative stress [35]. Term labor induces increased lipid peroxidation, as evidenced by increased levels of the biomarker, malondialdehyde [37]. In a case controlled study, the serum levels of hydroperoxides were higher in patients in labor, compared to the controls, who were not in labor [36]. Term labor was demonstrated to cause an up regulation of the antioxidant reserve in the fetal compartment [66]. The role of oxidative stress in initiation of labor is not known.
F2-isoprostanes, reliable biomarkers of oxidative stress were shown to be significantly elevated in plasma of neonates compared to adults [204]. The study also demonstrated an inverse correlation between gestational age and plasma isoprostane levels.
5.6 Interventions to overcome oxidative stress during pregnancy
Based on the understanding of the pathophysiological role of NO in the female reproductive tract, NO donors have been studied for cervical ripening at term. In a randomized controlled study, vaginal administration of isosorbide dinitrate induced cervical ripening at term [205]. Oxidative stress leads to focal collagen damage in the fetal membranes and result in preterm labor [39,206]. Antioxidant supplementation has been investigated in preterm labor and pre-eclampsia for beneficial effects [196,207]. Another randomized double-blinded placebo controlled trial initiated in 2003 will examine women with type-1 diabetes. These women were randomized to receive antioxidant supplementation with vitamin C and vitamin E. Benefits of antioxidant supplementation will be investigated in patients with type-1 diabetes and the incidence of pre-eclampsia in this group of patients will be studied [208]. In addition the secondary outcomes of birth weight centile and the endothelial activation indicated by PAI-1/PAI-2 (plasminogen activator-1/plasminogen activator-2) ratio will also be studied [208].
6. Future perspectives in antioxidant therapy
Antioxidants prevent the actions of the free radicals of oxidizing the substrate. Studies conducted in humans aimed at delineating the association of TAC content of food with incidence of chronic diseases [209]. The nutrients that are being studied for their effects on chronic diseases are vitamin C, Vitamin E, carotenoids and selenium. Pregnant women with HIV infection, selenium deficiency or micronutrient deficiencies like vitamin C and vitamin A, were found to have adverse clinical outcomes in large prospective studies [210,211]. There is increasing argument for increasing the selenium intake in these patients. There is emerging enthusiasm in the use of antioxidants, natural or synthetic. Small molecules that mimic antioxidant enzymes are the new tools being developed in the antioxidant armamentarium [212]. These are cell membrane permeable unlike the natural superoxide dismutase. Antioxidants targeting cellular organelles like mitochondria are also being investigated. Gene polymorphisms of the glutathione S-transferase family and myeloperoxides and their association with endometriosis, is an area of recent interest, which is promising [26].
7. Conclusion
The literature provides some evidence of oxidative stress influencing the entire reproductive span of a woman, even the menopausal years. OS plays a role in multiple physiological processes from oocyte maturation to fertilization and embryo development. There is burgeoning literature on the involvement of OS in the pathoysiology of infertility, assisted fertility and female reproduction. Infertility is a problem with a large magnitude. In this review we attempted to examine the various causes of female infertility and the role of OS in various etiologies of infertility. OS can arise as result of excessive production of free radicals and/or impaired antioxidant defense mechanism. An increasing number of published studies have pointed towards increased importance of the role of OS in female reproduction. Clearly, we have much to learn, but what we do know is that the role of OS in female reproduction cannot be underestimated. There is evidence that OS plays a role in conditions such as abortions, pre-eclampsia, hydatidiform mole, fetal embryopathies, preterm labor and pre-eclampsia and gestational diabetes, which lead to an immense burden of maternal and fetal morbidity and mortality. The review addresses the issue that both NOS and ROS species can lead to infertility problems and a spectrum of female reproductive disorders. We emphasize that free radicals have important physiological functions in the female reproductive tract as well as excessive free radicals precipitate female reproductive tract pathologies.
Reference values for ROS and NOS, minimum safe concentrations or physiologically beneficial concentrations have yet not been defined. Patients should be assessed according to the etiological factors and analyzed separately. Most of the published studies on oxidative stress are either observational or case control studies. Newer studies should be designed with more patient numbers; similar outcome parameters and uniform study populations so that results can be more easily compared. Measurement of OS in vivo is controversial. The sensitivity and specificity of various oxidative stress markers is not known. Measurement of biomarkers of OS is subject to interlaboratory variations, and interobserver differences. A uniform method with comprehensive assessment of the OS biomarkers should be used so that the results can be compared across the studies. Treatment strategies of antioxidant supplementation, directed toward reducing OS need to be investigated in randomized controlled trials. Antioxidants maybe advised when specific etiology cannot be identified as in idiopathic infertility as there is no other evidence based treatment for idiopathic infertility and reports indicate the presence of OS. Strategies to overcome OS in-vitro conditions and balancing between in vivo and in vitro environments can be utilized in ART, to successfully treat infertility. Interventions for overcoming oxidative stress in conditions such as abortions, preeclampsia, preterm labor and gestational diabetes and intrauterine growth retardation are still investigational with various randomized controlled trials in progress.
Legend
Reprinted from an article in Reproductive BioMedicine Online by Agarwal and Allamaneni, 2004, with permission from Reproductive Healthcare Ltd [33].
Acknowledgements
The authors wish to thank Robin Verdi for her secretarial support.
==== Refs
Pierce JD Cackler AB Arnett MG Why should you care about free radicals? RN 2004 67 38 42 quiz 43 14979192
Szczepanska M Kozlik J Skrzypczak J Mikolajczyk M Oxidative stress may be a piece in the endometriosis puzzle Fertil Steril 2003 79 1288 1293 12798872 10.1016/S0015-0282(03)00266-8
Van Langendonckt A Casanas-Roux F Donnez J Oxidative stress and peritoneal endometriosis Fertil Steril 2002 77 861 870 12009336 10.1016/S0015-0282(02)02959-X
Attaran M Pasqualotto E Falcone T Goldberg JM Miller KF Agarwal A Sharma RK The effect of follicular fluid reactive oxygen species on the outcome of in vitro fertilization Int J Fertil Womens Med 2000 45 314 320 11092702
Halliwell B Gutteridge JM Cross CE Free radicals, antioxidants, and human disease: where are we now? J Lab Clin Med 1992 119 598 620 1593209
Mello Filho AC Hoffmann ME Meneghini R Cell killing and DNA damage by hydrogen peroxide are mediated by intracellular iron Biochem J 1984 218 273 275 6712613
Gibson GE Huang HM Mitochondrial enzymes and endoplasmic reticulum calcium stores as targets of oxidative stress in neurodegenerative diseases J Bioenerg Biomembr 2004 36 335 340 15377868 10.1023/B:JOBB.0000041764.45552.f3
Madamanchi NR Vendrov A Runge MS Oxidative stress and vascular disease Arterioscler Thromb Vasc Biol 2005 25 29 38 15539615
Aitken RJ Baker MA Oxidative stress and male reproductive biology Reprod Fertil Dev 2004 16 581 588 15367373 10.1071/RD03089
Evans MD Dizdaroglu M Cooke MS Oxidative DNA damage and disease: induction, repair and significance Mutat Res 2004 567 1 61 15341901 10.1016/j.mrrev.2003.11.001
Behrman HR Kodaman PH Preston SL Gao S Oxidative stress and the ovary J Soc Gynecol Investig 2001 8 S40 42 11223371 10.1016/S1071-5576(00)00106-4
Sabatini L Wilson C Lower A Al-Shawaf T Grudzinskas JG Superoxide dismutase activity in human follicular fluid after controlled ovarian hyperstimulation in women undergoing in vitro fertilization Fertil Steril 1999 72 1027 1034 10593376 10.1016/S0015-0282(99)00411-2
Shiotani M Noda Y Narimoto K Imai K Mori T Fujimoto K Ogawa K Immunohistochemical localization of superoxide dismutase in the human ovary Hum Reprod 1991 6 1349 1353 1770125
Suzuki T Sugino N Fukaya T Sugiyama S Uda T Takaya R Yajima A Sasano H Superoxide dismutase in normal cycling human ovaries: immunohistochemical localization and characterization Fertil Steril 1999 72 720 726 10521117 10.1016/S0015-0282(99)00332-5
Jozwik M Wolczynski S Szamatowicz M Oxidative stress markers in preovulatory follicular fluid in humans Mol Hum Reprod 1999 5 409 413 10338363 10.1093/molehr/5.5.409
El Mouatassim S Guerin P Menezo Y Expression of genes encoding antioxidant enzymes in human and mouse oocytes during the final stages of maturation Mol Hum Reprod 1999 5 720 725 10421798 10.1093/molehr/5.8.720
Guerin P El Mouatassim S Menezo Y Oxidative stress and protection against reactive oxygen species in the pre-implantation embryo and its surroundings Hum Reprod Update 2001 7 175 189 11284661 10.1093/humupd/7.2.175
Ishikawa M Oxygen radicals-superoxide dismutase system and reproduction medicine Nippon Sanka Fujinka Gakkai Zasshi 1993 45 842 848 8371013
Vega M Johnson MC Diaz HA Urrutia LR Troncoso JL Devoto L Regulation of human luteal steroidogenesis in vitro by nitric oxide Endocrine 1998 8 185 191 9704576 10.1385/ENDO:8:2:185
Dong M Shi Y Cheng Q Hao M Increased nitric oxide in peritoneal fluid from women with idiopathic infertility and endometriosis J Reprod Med 2001 46 887 891 11725732
Rosselli M Keller PJ Dubey RK Role of nitric oxide in the biology, physiology and pathophysiology of reproduction Hum Reprod Update 1998 4 3 24 9622410 10.1093/humupd/4.1.3
Ohl J Lefebvre-Maunoury C Wittemer C Nisand G Laurent MC Hoffmann P Nitric oxide donors for patients undergoing IVF. A prospective, double-blind, randomized, placebo-controlled trial Hum Reprod 2002 17 2615 2620 12351537 10.1093/humrep/17.10.2615
Osborn BH Haney AF Misukonis MA Weinberg JB Inducible nitric oxide synthase expression by peritoneal macrophages in endometriosis-associated infertility Fertil Steril 2002 77 46 51 11779590 10.1016/S0015-0282(01)02940-5
Reynaert NL Ckless K Wouters EF van der Vliet A Janssen-Heininger YM Nitric oxide and redox signaling in allergic airway inflammation Antioxid Redox Signal 2005 7 129 143 15650402 10.1089/ars.2005.7.129
Schulman IH Zhou MS Raij L Nitric oxide, angiotensin II, and Reactive Oxygen Species in Hypertension and Atherogenesis Curr Hypertens Rep 2005 7 61 67 15683588
Hsieh YY Chang CC Tsai FJ Lin CC Chen JM Tsai CH Glutathione S-transferase M1*null genotype but not myeloperoxidase promoter G-463A polymorphism is associated with higher susceptibility to endometriosis Mol Hum Reprod 2004 10 713 717 15299090 10.1093/molehr/gah095
Schrier RW Wang W Acute renal failure and sepsis N Engl J Med 2004 351 159 169 15247356 10.1056/NEJMra032401
Ota H Igarashi S Hatazawa J Tanaka T Endothelial nitric oxide synthase in the endometrium during the menstrual cycle in patients with endometriosis and adenomyosis Fertil Steril 1998 69 303 308 9496346 10.1016/S0015-0282(97)00478-0
Lee KS Joo BS Na YJ Yoon MS Choi OH Kim WW Relationships between concentrations of tumor necrosis factor-alpha and nitric oxide in follicular fluid and oocyte quality J Assist Reprod Genet 2000 17 222 228 10955247 10.1023/A:1009495913119
Agarwal A Saleh RA Bedaiwy MA Role of reactive oxygen species in the pathophysiology of human reproduction Fertil Steril 2003 79 829 843 12749418 10.1016/S0015-0282(02)04948-8
Chan AC Partners in defense, vitamin E and vitamin C Can J Physiol Pharmacol 1993 71 725 731 8313238
de Matos DG Furnus CC The importance of having high glutathione (GSH) level after bovine in vitro maturation on embryo development effect of beta-mercaptoethanol, cysteine and cystine Theriogenology 2000 53 761 771 10735042 10.1016/S0093-691X(99)00278-2
Agarwal A Allamaneni SS Role of free radicals in female reproductive diseases and assisted reproduction Reprod Biomed Online 2004 9 338 347 15353087
de Bruin JP Dorland M Spek ER Posthuma G van Haaften M Looman CW te Velde ER Ultrastructure of the resting ovarian follicle pool in healthy young women Biol Reprod 2002 66 1151 1160 11906936
Myatt L Cui X Oxidative stress in the placenta Histochem Cell Biol 2004 122 369 382 15248072 10.1007/s00418-004-0677-x
Fainaru O Almog B Pinchuk I Kupferminc MJ Lichtenberg D Many A Active labour is associated with increased oxidisibility of serum lipids ex vivo BJOG 2002 109 938 941 12197375
Mocatta TJ Winterbourn CC Inder TE Darlow BA The effect of gestational age and labour on markers of lipid and protein oxidation in cord plasma Free Radic Res 2004 38 185 191 15104212 10.1080/10715760310001646048
Wall PD Pressman EK Woods JR Jr Preterm premature rupture of the membranes and antioxidants: the free radical connection J Perinat Med 2002 30 447 457 12530100 10.1515/JPM.2002.071
Pressman EK Cavanaugh JL Mingione M Norkus EP Woods JR Effects of maternal antioxidant supplementation on maternal and fetal antioxidant levels: a randomized, double-blind study Am J Obstet Gynecol 2003 189 1720 1725 14710104 10.1016/S0002-9378(03)00858-5
Ray SD Lam TS Rotollo JA Phadke S Patel C Dontabhaktuni A Mohammad S Lee H Strika S Dobrogowska A Bruculeri C Chou A Patel S Patel R Manolas T Stohs S Oxidative stress is the master operator of drug and chemically-induced programmed and unprogrammed cell death: Implications of natural antioxidants in vivo Biofactors 2004 21 223 232 15630201
Sugino N Takiguchi S Kashida S Karube A Nakamura Y Kato H Superoxide dismutase expression in the human corpus luteum during the menstrual cycle and in early pregnancy Mol Hum Reprod 2000 6 19 25 10611256 10.1093/molehr/6.1.19
Paszkowski T Traub AI Robinson SY McMaster D Selenium dependent glutathione peroxidase activity in human follicular fluid Clin Chim Acta 1995 236 173 180 7554284 10.1016/0009-8981(95)98130-9
Garcia-Pardo L Granados MD Gaytan F Padilla CA Martinez-Galisteo E Morales C Sanchez-Criado JE Barcena JA Immunolocalization of glutaredoxin in the human corpus luteum Mol Hum Reprod 1999 5 914 919 10508218 10.1093/molehr/5.10.914
Tranquilli AL Bezzeccheri V Giannubilo SR Scagnoli C Mazzanti L Garzetti GG Amniotic vascular endothelial growth factor (VEGF) and nitric oxide (NO) in women with subsequent preeclampsia Eur J Obstet Gynecol Reprod Biol 2004 113 17 20 15036704 10.1016/S0301-2115(03)00369-5
Takagi Y Nikaido T Toki T Kita N Kanai M Ashida T Ohira S Konishi I Levels of oxidative stress and redox-related molecules in the placenta in preeclampsia and fetal growth restriction Virchows Arch 2004 444 49 55 14574573 10.1007/s00428-003-0903-2
Harma M Kocyigit A Demir N Role of plasma nitric oxide in complete hydatidiform mole Eur J Gynaecol Oncol 2004 25 333 335 15171312
Harma M Kocyigit A Comparison of protein carbonyl and total plasma thiol concentrations in patients with complete hydatidiform mole with those in healthy pregnant women Acta Obstet Gynecol Scand 2004 83 857 860 15315598 10.1111/j.0001-6349.2004.00608.x
Harma M Erel O Increased oxidative stress in patients with hydatidiform mole Swiss Med Wkly 2003 133 563 566 14691728
Loeken MR Free radicals and birth defects J Matern Fetal Neonatal Med 2004 15 6 14 15101606 10.1080/14767050310001650662
Lagod L Paszkowski T Sikorski R Rola R The antioxidant-prooxidant balance in pregnancy complicated by spontaneous abortion Ginekol Pol 2001 72 1073 1078 11883212
Pasqualotto EB Agarwal A Sharma RK Izzo VM Pinotti JA Joshi NJ Rose BI Effect of oxidative stress in follicular fluid on the outcome of assisted reproductive procedures Fertil Steril 2004 81 973 976 15066450 10.1016/j.fertnstert.2003.11.021
Seino T Saito H Kaneko T Takahashi T Kawachiya S Kurachi H Eight-hydroxy-2'-deoxyguanosine in granulosa cells is correlated with the quality of oocytes and embryos in an in vitro fertilization-embryo transfer program Fertil Steril 2002 77 1184 1190 12057726 10.1016/S0015-0282(02)03103-5
Vega M Urrutia L Iniguez G Gabler F Devoto L Johnson MC Nitric oxide induces apoptosis in the human corpus luteum in vitro Mol Hum Reprod 2000 6 681 687 10908276 10.1093/molehr/6.8.681
Tsafriri A Reich R Molecular aspects of mammalian ovulation Exp Clin Endocrinol Diabetes 1999 107 1 11 10077349
Van Voorhis BJ Dunn MS Snyder GD Weiner CP Nitric oxide: an autocrine regulator of human granulosa-luteal cell steroidogenesis Endocrinology 1994 135 1799 1806 7525252 10.1210/en.135.5.1799
Tamate K Sengoku K Ishikawa M The role of superoxide dismutase in the human ovary and fallopian tube J Obstet Gynaecol 1995 21 401 409 8775911
Takayama K Sasano H Fukaya T Morohashi K Suzuki T Tamura M Costa MJ Yajima A Immunohistochemical localization of Ad4-binding protein with correlation to steroidogenic enzyme expression in cycling human ovaries and sex cord stromal tumors J Clin Endocrinol Metab 1995 80 2815 2821 7673429 10.1210/jc.80.9.2815
Sugino N Takiguchi S Ono M Tamura H Shimamura K Nakamura Y Tsuruta R Sadamitsu D Ueda T Maekawa T Kato H Nitric oxide concentrations in the follicular fluid and apoptosis of granulosa cells in human follicles Hum Reprod 1996 11 2484 2487 8981140
Jauniaux E Watson AL Hempstock J Bao YP Skepper JN Burton GJ Onset of maternal arterial blood flow and placental oxidative stress. A possible factor in human early pregnancy failure Am J Pathol 2000 157 2111 2122 11106583
Wang Y Walsh SW Antioxidant activities and mRNA expression of superoxide dismutase, catalase, and glutathione peroxidase in normal and preeclamptic placentas J Soc Gynecol Investig 1996 3 179 184 8796828 10.1016/1071-5576(96)00015-9
Djordjevic A Spasic S Jovanovic-Galovic A Djordjevic R Grubor-Lajsic G Oxidative stress in diabetic pregnancy: SOD, CAT and GSH-Px activity and lipid peroxidation products J Matern Fetal Neonatal Med 2004 16 367 372 15621558 10.1080/14767050400018270
Wiktor H Kankofer M Schmerold I Dadak A Lopucki M Niedermuller H Oxidative DNA damage in placentas from normal and pre-eclamptic pregnancies Virchows Arch 2004 445 74 78 15133663
Holthe MR Staff AC Berge LN Lyberg T Leukocyte adhesion molecules and reactive oxygen species in preeclampsia Obstet Gynecol 2004 103 913 922 15121565
Ishihara O Hayashi M Osawa H Kobayashi K Takeda S Vessby B Basu S Isoprostanes, prostaglandins and tocopherols in pre-eclampsia, normal pregnancy and non-pregnancy Free Radic Res 2004 913 918 15621708 10.1080/10715760412331273421
Vaisanen-Tommiska M Nuutila M Ylikorkala O Cervical nitric oxide release in women postterm Obstet Gynecol 2004 103 657 662 15051555
Buhimschi IA Buhimschi CS Pupkin M Weiner CP Beneficial impact of term labor: nonenzymatic antioxidant reserve in the human fetus Am J Obstet Gynecol 2003 189 181 188 12861160 10.1067/mob.2003.357
ASRM Definition of "infertility" Fertil Steril 2004 82 S206 15363731
Yu SL Yap C Investigating the infertile couple Ann Acad Med Singapore 2003 32 611 613 quiz 614 14626788
Duckitt K Infertility and subfertility Clin Evid 2003 2044 2073 12967463
DeCherney A Nathan L Eskandari M Cadieux M Current Obstetric & Gynecologic Diagnosis & Treatment McGraw Hill, New York 2003 6 979 990
ASRM Guidelines for the provision of infertility services Fertil Steril 2004 82 S24 S25 15363689
Sharma RK Agarwal A Role of reactive oxygen species in gynecologic diseases Reprod Med Bio 2004 3 177 199 10.1111/j.1447-0578.2004.00068.x
Dennery PA Role of redox in fetal development and neonatal diseases Antioxid Redox Signal 2004 6 147 53 14713346 10.1089/152308604771978453
Bedaiwy MA Falcone T Sharma RK Goldberg JM Attaran M Nelson DR Agarwal A Prediction of endometriosis with serum and peritoneal fluid markers: a prospective controlled trial Hum Reprod 2002 17 426 431 11821289 10.1093/humrep/17.2.426
Bedaiwy MA Falcone T Peritoneal fluid environment in endometriosis. Clinicopathological implications Minerva Ginecol 2003 55 333 345 14581858
Ben-Shlomo I Kokia E Jackson MJ Adashi EY Payne DW Interleukin-1 beta stimulates nitrite production in the rat ovary: evidence for heterologous cell-cell interaction and for insulin-mediated regulation of the inducible isoform of nitric oxide synthase Biol Reprod 1994 51 310 318 7524699
Malek A Sager R Schneider H Effect of hypoxia, oxidative stress and lipopolysaccharides on the release of prostaglandins and cytokines from human term placental explants Placenta 2001 22 S45 50 11312628 10.1053/plac.2001.0635
Bilodeau JF Hubel CA Current concepts in the use of antioxidants for the treatment of preeclampsia J Obstet Gynaecol Can 2003 25 742 750 12970809
Hung TH Charnock-Jones DS Skepper JN Burton GJ Secretion of tumor necrosis factor-alpha from human placental tissues induced by hypoxia-reoxygenation causes endothelial cell activation in vitro: a potential mediator of the inflammatory response in preeclampsia Am J Pathol 2004 164 1049 1061 14982858
Banerjee S Smallwood A Moorhead J Chambers AE Papageorghiou A Campbell S Nicolaides K Placental expression of Interferon-gamma (IFN-gamma) and its receptor (IFN-gamma) R2 fail to switch from early hypoxic to late normotensive development in pre-eclampsia J Clin Endocrinol Metab 2005 90 944 952 15585559 10.1210/jc.2004-1113
Oner-Iyidogan Y Kocak H Gurdol F Korkmaz D Buyru F Indices of oxidative stress in eutopic and ectopic endometria of women with endometriosis Gynecol Obstet Invest 2004 57 214 217 14970668 10.1159/000076691
Sugino N Karube-Harada A Sakata A Takiguchi S Kato H Nuclear factor-kappa B is required for tumor necrosis factor-alpha-induced manganese superoxide dismutase expression in human endometrial stromal cells J Clin Endocrinol Metab 2002 87 3845 3850 12161520 10.1210/jc.87.8.3845
Sugino N Karube-Harada A Taketani T Sakata A Nakamura Y Withdrawal of ovarian steroids stimulates prostaglandin F2alpha production through nuclear factor-kappaB activation via oxygen radicals in human endometrial stromal cells: potential relevance to menstruation J Reprod Dev 2004 50 215 225 15118249 10.1262/jrd.50.215
Ness RB Endometriosis and ovarian cancer: thoughts on shared pathophysiology Am J Obstet Gynecol 2003 189 280 294 12861175 10.1067/mob.2003.408
Murdoch WJ Martinchick JF Oxidative damage to DNA of ovarian surface epithelial cells affected by ovulation: carcinogenic implication and chemoprevention Exp Biol Med (Maywood) 2004 229 546 552 15169974
Ness RB Cottreau C Possible role of ovarian epithelial inflammation in ovarian cancer J Natl Cancer Inst 1999 91 1459 1467 10469746 10.1093/jnci/91.17.1459
Modugno F Ness RB Allen GO Schildkraut JM Davis FG Goodman MT Oral contraceptive use, reproductive history, and risk of epithelial ovarian cancer in women with and without endometriosis Am J Obstet Gynecol 2004 191 733 740 15467532 10.1016/j.ajog.2004.03.035
WHO: Epithelial ovarian cancer and combined oral contraceptives. The WHO Collaborative Study of Neoplasia and Steroid Contraceptives Int J Epidemiol 1989 18 538 545 2807655
Young TW Mei FC Yang G Thompson-Lanza JA Liu J Cheng X Activation of antioxidant pathways in ras-mediated oncogenic transformation of human surface ovarian epithelial cells revealed by functional proteomics and mass spectrometry Cancer Res 2004 64 4577 4584 15231669 10.1158/0008-5472.CAN-04-0222
Geva E Jaffe RB Role of vascular endothelial growth factor in ovarian physiology and pathology Fertil Steril 2000 74 429 438 10973633 10.1016/S0015-0282(00)00670-1
Zbucka M Koda M Tomaszewski J Przystupa W Sulkowski S Wolczynski S Angiogenesis in the female reproductive processes Ginekol Pol 2004 75 649 657 15517790
Krussel JS Bielfeld P Polan ML Simon C Regulation of embryonic implantation Eur J Obstet Gynecol Reprod Biol 2003 110 S2 S9 12965085 10.1016/S0301-2115(03)00167-2
Sherer DM Abulafia O Angiogenesis during implantation, and placental and early embryonic development Placenta 2001 22 1 13 11162347 10.1053/plac.2000.0588
Choi HK Choi BC Lee SH Kim JW Cha KY Baek KH Expression of angiogenesis- and apoptosis-related genes in chorionic villi derived from recurrent pregnancy loss patients Mol Reprod Dev 2003 66 24 31 12874795 10.1002/mrd.10331
Gordon JD Mesiano S Zaloudek CJ Jaffe RB Vascular endothelial growth factor localization in human ovary and fallopian tubes: possible role in reproductive function and ovarian cyst formation J Clin Endocrinol Metab 1996 81 353 359 8550777 10.1210/jc.81.1.353
Bausero P Cavaille F Meduri G Freitas S Perrot-Applanat M Paracrine action of vascular endothelial growth factor in the human endometrium: production and target sites, and hormonal regulation Angiogenesis 1998 2 167 182 14517472
Albrecht ED Babischkin JS Lidor Y Anderson LD Udoff LC Pepe GJ Effect of estrogen on angiogenesis in co-cultures of human endometrial cells and microvascular endothelial cells Hum Reprod 2003 18 2039 2047 14507818 10.1093/humrep/deg415
Ushio-Fukai M Alexander RW Reactive oxygen species as mediators of angiogenesis signaling: role of NAD(P)H oxidase Mol Cell Biochem 2004 264 85 97 15544038 10.1023/B:MCBI.0000044378.09409.b5
Eyries M Collins T Khachigian LM Modulation of growth factor gene expression in vascular cells by oxidative stress Endothelium 2004 11 133 139 15370072 10.1080/10623320490482691
Sugino N Shimamura K Takiguchi S Tamura H Ono M Nakata M Nakamura Y Ogino K Uda T Kato H Changes in activity of superoxide dismutase in the human endometrium throughout the menstrual cycle and in early pregnancy Hum Reprod 1996 11 1073 1078 8671393
Sugino N Karube-Harada A Kashida S Takiguchi S Kato H Reactive oxygen species stimulate prostaglandin F2 alpha production in human endometrial stromal cells in vitro Hum Reprod 2001 16 1797 1801 11527878 10.1093/humrep/16.9.1797
Zeller JM Henig I Radwanska E Dmowski WP Enhancement of human monocyte and peritoneal macrophage chemiluminescence activities in women with endometriosis Am J Reprod Immunol Microbiol 1987 13 78 82 3605484
Wang Y Sharma RK Falcone T Goldberg J Agarwal A Importance of reactive oxygen species in the peritoneal fluid of women with endometriosis or idiopathic infertility Fertil Steril 1997 68 826 830 9389810 10.1016/S0015-0282(97)00343-9
Murphy AA Palinski W Rankin S Morales AJ Parthasarathy S Evidence for oxidatively modified lipid-protein complexes in endometrium and endometriosis Fertil Steril 1998 69 1092 1094 9627298 10.1016/S0015-0282(98)00087-9
Shanti A Santanam N Morales AJ Parthasarathy S Murphy AA Autoantibodies to markers of oxidative stress are elevated in women with endometriosis Fertil Steril 1999 71 1115 1118 10360920 10.1016/S0015-0282(99)00145-4
Murphy AA Santanam N Morales AJ Parthasarathy S Lysophosphatidyl choline, a chemotactic factor for monocytes/T-lymphocytes is elevated in endometriosis J Clin Endocrinol Metab 1998 83 2110 2113 9626147 10.1210/jc.83.6.2110
Foyouzi N Berkkanoglu M Arici A Kwintkiewicz J Izquierdo D Duleba AJ Effects of oxidants and antioxidants on proliferation of endometrial stromal cells Fertil Steril 2004 82 1019 1022 15474067 10.1016/j.fertnstert.2004.02.133
Murphy AA Zhou MH Malkapuram S Santanam N Parthasarathy S Sidell N RU486-induced growth inhibition of human endometrial cells Fertil Steril 2000 71 1014 1019 11056251 10.1016/S0015-0282(00)01606-X
Vega M Carrasco I Castillo T Troncoso JL Videla LA Devoto L Functional luteolysis in response to hydrogen peroxide in human luteal cells J Endocrinol 1995 147 177 182 7490532
LaPolt PS Leung K Ishimaru R Tafoya MA You-hsin Chen J Roles of cyclic GMP in modulating ovarian functions Reprod Biomed Online 2003 6 15 23 12626139
Hanafy KA Krumenacker JS Murad F NO, nitrotyrosine, and cyclic GMP in signal transduction Med Sci Monit 2001 7 801 819 11433215
Ekerhovd E Brannstrom M Alexandersson M Norstrom A Evidence for nitric oxide mediation of contractile activity in isolated strips of the human Fallopian tube Hum Reprod 1997 12 301 305 9070716 10.1093/humrep/12.2.301
Rosselli M Dubey RK Rosselli MA Macas E Fink D Lauper U Keller PJ Imthurn B Identification of nitric oxide synthase in human and bovine oviduct Mol Hum Reprod 1996 2 607 612 9239673
Rosselli M Dubey RK Imthurn B Macas E Keller PJ Effects of nitric oxide on human spermatozoa: evidence that nitric oxide decreases sperm motility and induces sperm toxicity Hum Reprod 1995 10 1786 1790 8582980
Tseng L Zhang J Peresleni T Goligorsky MS Cyclic expression of endothelial nitric oxide synthase mRNA in the epithelial glands of human endometrium J Soc Gynecol Investig 1996 3 33 38 8796805 10.1016/1071-5576(95)00039-9
Taguchi M Alfer J Chwalisz K Beier HM Classen-Linke I Endothelial nitric oxide synthase is differently expressed in human endometrial vessels during the menstrual cycle Mol Hum Reprod 2000 6 185 190 10655461 10.1093/molehr/6.2.185
Telfer JF Lyall F Norman JE Cameron IT Identification of nitric oxide synthase in human uterus Hum Reprod 1995 10 19 23 7538147
Norman JE Thompson AJ Telfer JF Young A Greer IA Cameron IT Myometrial constitutive nitric oxide synthase expression is increased during human pregnancy Mol Hum Reprod 1999 5 175 181 10065874 10.1093/molehr/5.2.175
Sun X Qiu X Gemzell-Danielsson K Effects of mifepristone on expression of endothelial nitric oxide synthase in human endometrium during the implantation phase Fertil Steril 2003 80 1454 60 14667883 10.1016/j.fertnstert.2003.05.001
Shaamash AH Zakhari MM Increased serum levels of nitric oxide metabolites among users: a possible role in progestin-induced bleeding Hum Reprod 2005 20 302 306 15471931 10.1093/humrep/deh541
Bansal RK Goldsmith PC He Y Zaloudek CJ Ecker JL Riemer RK A decline in myometrial nitric oxide synthase expression is associated with labor and delivery J Clin Invest 1997 99 2502 2508 9153294
Seyffarth G Nelson PN Dunmore SJ Rodrigo N Murphy DJ Carson RJ Lipopolysaccharide induces nitric oxide synthase expression and platelet-activating factor increases nitric oxide production in human fetal membranes in culture Reprod Biol Endocrinol 2004 2 29 15191613 10.1186/1477-7827-2-29
Marinoni E Di Iorio R Villaccio B Alberini A Rota F Cosmi EV Amniotic fluid nitric oxide metabolite levels and nitric oxide synthase localization in feto-placental tissues are modified in association with human labor Eur J Obstet Gynecol Reprod Biol 2000 89 47 54 10733023 10.1016/S0301-2115(99)00186-4
Lee TH Wu MY Chen MJ Chao KH Ho HN Yang YS Nitric oxide is associated with poor embryo quality and pregnancy outcome in in vitro fertilization cycles Fertil Steril 2004 82 126 131 15237000 10.1016/j.fertnstert.2004.02.097
Oztezcan S Turkoglu UM Kervancioglu E Kocak T Kocak-Toker N Aykac-Toker G In vitro effects of peroxynitrite on human spermatozoa Andrologia 1999 31 195 198 10470408 10.1046/j.1439-0272.1999.00279.x
Chung HT Pae HO Choi BM Billiar TR Kim YM Nitric oxide as a bioregulator of apoptosis Biochem Biophys Res Commun 2001 282 1075 1079 11302723 10.1006/bbrc.2001.4670
Polak G Koziol-Montewka M Gogacz M Blaszkowska I Kotarski J Total antioxidant status of peritoneal fluid in infertile women Eur J Obstet Gynecol Reprod Biol 2001 94 261 263 11165736 10.1016/S0301-2115(00)00352-3
Arumugam K Dip YC Endometriosis and infertility: the role of exogenous lipid peroxides in the peritoneal fluid Fertil Steril 1995 63 198 199 7805914
Ho HN Wu MY Chen SU Chao KH Chen CD Yang YS Total antioxidant status and nitric oxide do not increase in peritoneal fluids from women with endometriosis Hum Reprod 1997 12 2810 2815 9455858 10.1093/humrep/12.12.2810
Khorram O Lessey BA Alterations in expression of endometrial endothelial nitric oxide synthase and alpha(v)beta(3) integrin in women with endometriosis Fertil Steril 2002 78 860 864 12372469 10.1016/S0015-0282(02)03347-2
Wu MY Chao KH Yang JH Lee TH Yang YS Ho HN Nitric oxide synthesis is increased in the endometrial tissue of women with endometriosis Hum Reprod 2003 18 2668 2671 14645189 10.1093/humrep/deg484
Kamada Y Nakatsuka M Asagiri K Noguchi S Habara T Takata M Kudo T GnRH agonist-suppressed expression of nitric oxide synthases and generation of peroxynitrite in adenomyosis Hum Reprod 2000 15 2512 2519 11098019 10.1093/humrep/15.12.2512
Ekerhovd E Enskog A Caidahl K Klintland N Nilsson L Brannstrom M Norstrom A Plasma concentrations of nitrate during the menstrual cycle, ovarian stimulation and ovarian hyperstimulation syndrome Hum Reprod 2001 16 1334 1339 11425809 10.1093/humrep/16.7.1334
Portz DM Elkins TE White R Warren J Adadevoh S Randolph J Oxygen free radicals and pelvic adhesion formation: I. Blocking oxygen free radical toxicity to prevent adhesion formation in an endometriosis model Int J Fertil 1991 36 39 42 1672675
Zervou S Karteris E Goumenou AG Vatish M Koumantakis EE Hillhouse EW The Glu298-->Asp polymorphism of the endothelial nitric oxide synthase gene is associated with endometriosis Fertil Steril 2003 80 1524 1525 14667898 10.1016/S0015-0282(03)02204-0
Ota H Igarashi S Hatazawa J Tanaka T Endometriosis and free radicals Gynecol Obstet Invest 1999 48 29 35 10559662 10.1159/000052866
Leyendecker G Kunz G Wildt L Beil D Deininger H Uterine hyperperistalsis and dysperistalsis as dysfunctions of the mechanism of rapid sperm transport in patients with endometriosis and infertility Hum Reprod 1996 11 1542 1551 8671502
Kim KH Oh DS Jeong JH Shin BS Joo BS Lee KS Follicular blood flow is a better predictor of the outcome of in vitro fertilization-embryo transfer than follicular fluid vascular endothelial growth factor and nitric oxide concentrations Fertil Steril 2004 82 586 592 15374700 10.1016/j.fertnstert.2004.02.120
Song IO Huh Y Yoo KJ Choi BC Paik EC Son IP Increased expression of endothelial nitric oxide synthase in endometrium of infertile women with endometriosis or hydrosalpinx during the window of implantation Abstract of the 15th Annual Meeting of the ESHRE Tours, France June 14–17, 2000, Abstract No. O-035
Cicinelli E Ignarro LJ Lograno M Galantino P Balzano G Schonauer LM Circulating levels of nitric oxide in fertile women in relation to the menstrual cycle Fertil Steril 1996 1036 1038 8941078
Friden BE Runesson E Hahlin M Brannstrom M Evidence for nitric oxide acting as a luteolytic factor in the human corpus luteum Mol Hum Reprod 2000 6 397 403 10775642 10.1093/molehr/6.5.397
Anteby EY Hurwitz A Korach O Revel A Simon A Finci-Yeheskel Z Mayer M Laufer N Human follicular nitric oxide pathway: relationship to follicular size, oestradiol concentrations and ovarian blood flow Hum Reprod 1996 11 1947 1951 8921069
Manau D Balasch J Jimenez W Fabregues F Casamitjana R Creus M Arroyo V Vanrell JA Adrenomedullin and nitric oxide in menstrual and in vitro fertilization cycles. Relationship to estradiol Acta Obstet Gynecol Scand 1999 78 626 631 10422910 10.1034/j.1600-0412.1999.780711.x
Fabregues F Balasch J Manau D Creus M Jimenez W Carmona F Casamit J Vanrell JA Circulating levels of nitric oxide in successful and unsuccessful implantation after in vitro fertilization and embryo transfer. Relationship to estradiol and progesterone Acta Obstet Gynecol Scand 2000 79 564 569 10929956 10.1034/j.1600-0412.2000.079007564.x
Manau D Balasch J Jimenez W Fabregues F Civico S Casamitjana R Creus M Vanrell JA Follicular fluid concentrations of adrenomedullin, vascular endothelial growth factor and nitric oxide in IVF cycles: relationship to ovarian response Hum Reprod 2000 15 1295 1299 10831558 10.1093/humrep/15.6.1295
Battaglia C Salvatori M Maxia N Petraglia F Facchinetti F Volpe A Adjuvant L-arginine treatment for in-vitro fertilization in poor responder patients Hum Reprod 1999 14 1690 1697 10402369 10.1093/humrep/14.7.1690
Battaglia C Regnani G Marsella T Facchinetti F Volpe A Venturoli S Flamigni C Adjuvant L-arginine treatment in controlled ovarian hyperstimulation: a double-blind, randomized study Hum Reprod 2002 17 659 665 11870119 10.1093/humrep/17.3.659
Barrionuevo MJ Schwandt RA Rao PS Graham LB Maisel LP Yeko TR Nitric oxide (NO) and interleukin-1beta (IL-1beta) in follicular fluid and their correlation with fertilization and embryo cleavage Am J Reprod Immunol 2000 44 359 364 11200815 10.1111/j.8755-8920.2000.440607.x
Hata T Hashimoto M Manabe A Aoki S Iida K Masumura S Miyazaki K Maternal and fetal nitric oxide synthesis is decreased in pregnancies with small for gestational age infants Hum Reprod 1998 13 1070 1073 9619572 10.1093/humrep/13.4.1070
Lees C Campbell S Jauniaux E Brown R Ramsay B Gibb D Moncada S Martin JF Arrest of preterm labour and prolongation of gestation with glyceryl trinitrate, a nitric oxide donor Lancet 1994 343 1325 1326 7910325 10.1016/S0140-6736(94)92468-6
Braude P Rowell P Assisted conception. II – in vitro fertilisation and intracytoplasmic sperm injection BMJ 2003 327 852 855 14551103 10.1136/bmj.327.7419.852
Isaksson R Tiitinen A Present concept of unexplained infertility Gynecol Endocrinol 2004 18 278 290 15346664 10.1080/0951359042000199878
Bedaiwy MA Falcone T Mohamed MS Aleem AA Sharma RK Worley SE Thornton J Agarwal A Differential growth of human embryos in vitro: role of reactive oxygen species Fertil Steril 2004 82 593 600 15374701 10.1016/j.fertnstert.2004.02.121
Oyawoye O Abdel Gadir A Garner A Constantinovici N Perrett C Hardiman P Antioxidants and reactive oxygen species in follicular fluid of women undergoing IVF: relationship to outcome Hum Reprod 2003 18 2270 2274 14585872 10.1093/humrep/deg450
Paszkowski T Clarke RN Hornstein MD Smoking induces oxidative stress inside the Graafian follicle Hum Reprod 2002 17 921 925 11925382 10.1093/humrep/17.4.921
Van Blerkom J Antczak M Schrader R The developmental potential of the human oocyte is related to the dissolved oxygen content of follicular fluid: association with vascular endothelial growth factor levels and perifollicular blood flow characteristics Hum Reprod 1997 12 1047 1055 9194664 10.1093/humrep/12.5.1047
Yang HW Hwang KJ Kwon HC Kim HS Choi KW Oh KS Detection of reactive oxygen species (ROS) and apoptosis in human fragmented embryos Hum Reprod 1998 13 998 1002 9619561 10.1093/humrep/13.4.998
Kurzawa R Glabowski W Baczkowski T Wiszniewska B Marchlewicz M Growth factors protect in vitro cultured embryos from the consequences of oxidative stress Zygote 2004 12 231 240 15521713 10.1017/S0967199404002783
Reubinoff BE Har-El R Kitrossky N Friedler S Levi R Lewin A Chevion M Increased levels of redox-active iron in follicular fluid: a possible cause of free radical-mediated infertility in beta-thalassemia major Am J Obstet Gynecol 1996 174 914 918 8633668
Wiener-Megnazi Z Vardi L Lissak A Shnizer S Zeev Reznick A Ishai D Lahav-Baratz S Shiloh H Koifman M Dirnfeld M Oxidative stress indices in follicular fluid as measured by the thermochemiluminescence assay correlate with outcome parameters in in vitro fertilization Fertil Steril 2004 82 1171 1176 15474091 10.1016/j.fertnstert.2004.06.013
Torok A Nemeth P Torok B Berki T Tinneberg HR Bodis J Organic hydroperoxide-induced chemiluminescence of follicular fluid and blood serum samples obtained from women pretreated for in vitro fertilization Gynecol Obstet Invest 2004 57 72 79 14671414 10.1159/000075381
Shiverick KT Salafia C Cigarette smoking and pregnancy I: ovarian, uterine and placental effects Placenta 1999 20 265 272 10329346 10.1053/plac.1998.0377
Schweigert FJ Steinhagen B Raila J Siemann A Peet D Buscher U Concentrations of carotenoids, retinol and alpha-tocopherol in plasma and follicular fluid of women undergoing IVF Hum Reprod 2003 18 1259 1264 12773456 10.1093/humrep/deg249
Takasaki A Nakamura Y Tamura H Shimamura K Morioka H Melatonin as a new drug for improving oocyte quality Reprod Med Biol 2003 2 139 144 10.1111/j.1447-0578.2003.00035.x
Harvey AJ Kind KL Thompson JG REDOX regulation of early embryo development Reproduction 2002 123 479 486 11914110 10.1530/rep.0.1230479
Knott L Hartridge T Brown NL Mansell JP Sandy JR Homocysteine oxidation and apoptosis: a potential cause of cleft palate In Vitro Cell Dev Biol Anim 2003 39 98 105 12892533 10.1290/1543-706X(2003)039<0098:HOAAAP>2.0.CO;2
Parman T Wiley MJ Wells PG Free radical-mediated oxidative DNA damage in the mechanism of thalidomide teratogenicity Nat Med 1999 5 582 585 10229238 10.1038/8466
Burton GJ Hempstock J Jauniaux E Oxygen, early embryonic metabolism and free radical-mediated embryopathies Reprod Biomed Online 2003 6 84 96 12626148
Wentzel P Welsh N Eriksson UJ Developmental damage, increased lipid peroxidation, diminished cyclooxygenase-2 gene expression, and lowered prostaglandin E2 levels in rat embryos exposed to a diabetic environment Diabetes 1999 48 813 820 10102698
Nicol CJ Zielenski J Tsui LC Wells PG An embryoprotective role for glucose-6-phosphate dehydrogenase in developmental oxidative stress and chemical teratogenesis Faseb J 2000 14 111 127 10627286
Dumoulin JC Meijers CJ Bras M Coonen E Geraedts JP Evers JL Effect of oxygen concentration on human in-vitro fertilization and embryo culture Hum Reprod 1999 14 465 469 10099995 10.1093/humrep/14.2.465
Kitagawa Y Suzuki K Yoneda A Watanabe T Effects of oxygen concentration and antioxidants on the in vitro developmental ability, production of reactive oxygen species (ROS), and DNA fragmentation in porcine embryos Theriogenology 2004 62 1186 1197 15325546 10.1016/j.theriogenology.2004.01.011
Paszkowski T Clarke RN Antioxidative capacity of preimplantation embryo culture medium declines following the incubation of poor quality embryos Hum Reprod 1996 11 2493 2495 8981142
Ermilov A Diamond MP Sacco AG Dozortsev DD Culture media and their components differ in their ability to scavenge reactive oxygen species in the plasmid relaxation assay Fertil Steril 1999 72 154 157 10428165 10.1016/S0015-0282(99)00204-6
Agarwal A Nallella KP Allamaneni SS Said TM Role of antioxidants in treatment of male infertility: an overview of the literature Reprod Biomed Online 2004 8 616 627 15169573
Henmi H Endo T Kitajima Y Manase K Hata H Kudo R Effects of ascorbic acid supplementation on serum progesterone levels in patients with a luteal phase defect Fertil Steril 2003 80 459 461 12909517 10.1016/S0015-0282(03)00657-5
Vural P Akgul C Yildirim A Canbaz M Antioxidant defence in recurrent abortion Clin Chim Acta 2000 295 169 177 10767402 10.1016/S0009-8981(99)00255-7
Crha I Hruba D Ventruba P Fiala J Totusek J Visnova H Ascorbic acid and infertility treatment Cent Eur J Public Health 2003 11 63 67 12884545
Westphal LM Polan ML Trant AS Mooney SB A nutritional supplement for improving fertility in women: a pilot study J Reprod Med 2004 49 289 293 15134155
Griesinger G Franke K Kinast C Kutzelnigg A Riedinger S Kulin S Kaali SG Feichtinger W Ascorbic acid supplement during luteal phase in IVF J Assist Reprod Genet 2002 19 164 168 12036083 10.1023/A:1014837811353
Catt JW Henman M Toxic effects of oxygen on human embryo development Hum Reprod 2000 15 199 206 11041525
Lane M Maybach JM Gardner DK Addition of ascorbate during cryopreservation stimulates subsequent embryo development Hum Reprod 2002 17 2686 2693 12351549 10.1093/humrep/17.10.2686
Chui DK Pugh ND Walker SM Gregory L Shaw RW Follicular vascularity – the predictive value of transvaginal power Doppler ultrasonography in an in-vitro fertilization programme: a preliminary study Hum Reprod 1997 12 191 196 9043927 10.1093/humrep/12.1.191
Sher G Fisch JD Vaginal sildenafil (Viagra): a preliminary report of a novel method to improve uterine artery blood flow and endometrial development in patients undergoing IVF Hum Reprod 2000 15 806 809 10739824 10.1093/humrep/15.4.806
Sher G Fisch JD Effect of vaginal sildenafil on the outcome of in vitro fertilization (IVF) after multiple IVF failures attributed to poor endometrial development Fertil Steril 2002 78 1073 1076 12413996 10.1016/S0015-0282(02)03375-7
Lornage J Biological aspects of endometriosis in vitro fertilization J Gynecol Obstet Biol Reprod (Paris) 2003 32 S48 50 14968070
Tarin JJ Potential effects of age-associated oxidative stress on mammalian oocytes/embryos Mol Hum Reprod 1996 2 717 724 9239688
Tarin JJ Perez-Albala S Cano A Consequences on offspring of abnormal function in ageing gametes Hum Reprod Update 2000 6 532 549 11129687 10.1093/humupd/6.6.532
Arteaga E Villaseca P Rojas A Arteaga A Bianchi M Comparison of the antioxidant effect of estriol and estradiol on low density lipoproteins in post-menopausal women Rev Med Chil 1998 126 481 487 9731428
Tamir S Izrael S Vaya J The effect of oxidative stress on ERalpha and ERbeta expression J Steroid Biochem Mol Biol 2002 81 327 232 12361722 10.1016/S0960-0760(02)00115-2
Villar J Say L Shennan A Lindheimer M Duley L Conde-Agudelo A Merialdi M Methodological and technical issues related to the diagnosis, screening, prevention, and treatment of pre-eclampsia and eclampsia Int J Gynaecol Obstet 2004 85 S28 S41 15147852 10.1016/j.ijgo.2004.03.009
Harma M Erel O Measurement of the total antioxidant response in preeclampsia with a novel automated method Eur J Obstet Gynecol Reprod Biol 2005 118 47 51 15596272 10.1016/j.ejogrb.2004.04.012
Aydin S Benian A Madazli R Uludag S Uzun H Kaya S Plasma malondialdehyde, superoxide dismutase, sE-selectin, fibronectin, endothelin-1 and nitric oxide levels in women with preeclampsia Eur J Obstet Gynecol Reprod Biol 2004 113 21 25 15036705 10.1016/S0301-2115(03)00368-3
Mikhail MS Anyaegbunam A Garfinkel D Palan PR Basu J Romney SL Preeclampsia and antioxidant nutrients: decreased plasma levels of reduced ascorbic acid, alpha-tocopherol, and beta-carotene in women with preeclampsia Am J Obstet Gynecol 1994 171 150 157 8030691
Pijnenborg R Dixon G Robertson WB Brosens I Trophoblastic invasion of human decidua from 8 to 18 weeks of pregnancy Placenta 1980 1 3 19 7443635
Chappell LC Seed PT Briley AL Kelly FJ Lee R Hunt BJ Parmar K Bewley SJ Shennan AH Steer PJ Poston L Effect of antioxidants on the occurrence of pre-eclampsia in women at increased risk: a randomised trial Lancet 1999 354 810 816 10485722
Gulmezoglu AM Hofmeyr GJ Oosthuisen MM Antioxidants in the treatment of severe pre-eclampsia: an explanatory randomised controlled trial Br J Obstet Gynaecol 1997 104 689 696 9197872
Beazley D Ahokas R Livingston J Griggs M Sibai BM Vitamin C and E supplementation in women at high risk for preeclampsia: a double-blind, placebo-controlled trial Am J Obstet Gynecol 2005 192 520 521 15695996 10.1016/j.ajog.2004.09.005
Jauniaux E Hempstock J Greenwold N Burton GJ Trophoblastic oxidative stress in relation to temporal and regional differences in maternal placental blood flow in normal and abnormal early pregnancies Am J Pathol 2003 162 115 125 12507895
Jauniaux E Cindrova-Davies T Johns J Dunster C Hempstock J Kelly FJ Burton GJ Distribution and transfer pathways of antioxidant molecules inside the first trimester human gestational sac J Clin Endocrinol Metab 2004 89 1452 1458 15001647 10.1210/jc.2003-031332
Miller H Wilson R Jenkins C MacLean MA Roberts J Walker JJ Glutathione levels and miscarriage Fertil Steril 2000 74 1257 1258 11119764 10.1016/S0015-0282(00)01585-5
Sata F Yamada H Kondo T Gong Y Tozaki S Kobashi G Kato EH Fujimoto S Kishi R Glutathione S-transferase M1 and T1 polymorphisms and the risk of recurrent pregnancy loss Mol Hum Reprod 2003 9 165 169 12606593 10.1093/molehr/gag021
Cramer DW Wise LA The epidemiology of recurrent pregnancy loss Semin Reprod Med 2000 18 331 339 11355791 10.1055/s-2000-13722
Comporti M Signorini C Leoncini S Buonocore G Rossi V Ciccoli L Plasma F2-isoprostanes are elevated in newborns and inversely correlated to gestational age Free Radic Biol Med 2004 37 724 732 15288129 10.1016/j.freeradbiomed.2004.06.007
Ekerhovd E Weijdegard B Brannstrom M Mattsby-Baltzer I Norstrom A Nitric oxide induced cervical ripening in the human: Involvement of cyclic guanosine monophosphate, prostaglandin F(2 alpha), and prostaglandin E(2) Am J Obstet Gynecol 2002 186 745 750 11967501 10.1067/mob.2002.121327
Plessinger MA Woods JR JrMiller RK Pretreatment of human amnion-chorion with vitamins C and E prevents hypochlorous acid-induced damage Am J Obstet Gynecol 2000 183 979 985 11035350 10.1067/mob.2000.106676
Siega-Riz AM Promislow JH Savitz DA Thorp JM JrMcDonald T Vitamin C intake and the risk of preterm delivery Am J Obstet Gynecol 2003 189 519 525 14520228 10.1067/S0002-9378(03)00363-6
Holmes VA Young IS Maresh MJ Pearson DW Walker JD McCance DR The Diabetes and Pre-eclampsia Intervention Trial Int J Gynaecol Obstet 2004 87 66 71 15464786 10.1016/j.ijgo.2004.06.015
Mayne ST Wright ME Cartmel B Assessment of antioxidant nutrient intake and status for epidemiologic research J Nutr 2004 134 3199S 3200S 15514306
Kupka R Msamanga GI Spiegelman D Morris S Mugusi F Hunter DJ Fawzi WW Selenium status is associated with accelerated HIV disease progression among HIV-1-infected pregnant women in Tanzania J Nutr 2004 134 2556 2560 15465747
Fawzi W Msamanga G Micronutrients and adverse pregnancy outcomes in the context of HIV infection Nutr Rev 2004 62 269 275 15384917 10.1301/nr.2004.jul.269-275
Mates M Effects of antioxidant enzymes in the molecular control of reactive oxygen species toxicology Toxicology 2000 153 83 104 11090949 10.1016/S0300-483X(00)00306-1
|
16018814
|
PMC1215514
|
CC BY
|
2021-01-04 16:37:13
|
no
|
Reprod Biol Endocrinol. 2005 Jul 14; 3:28
|
utf-8
|
Reprod Biol Endocrinol
| 2,005 |
10.1186/1477-7827-3-28
|
oa_comm
|
==== Front
Reprod Biol EndocrinolReproductive biology and endocrinology : RB&E1477-7827BioMed Central London 1477-7827-3-341609297110.1186/1477-7827-3-34ResearchInterferon-gamma alters the phagocytic activity of the mouse trophoblast Albieri Andréa [email protected] Mara S [email protected] Sonia M [email protected] Eduardo C [email protected] Ises [email protected] Anne [email protected] Ali A [email protected] Estela [email protected] Department of Cell and Developmental Biology, Institute of Biomedical Sciences, University of São Paulo 05508-900 São Paulo, Brazil2 Department of Immunology, Institute of Biomedical Sciences, University of São Paulo 05508-900 São Paulo, Brazil3 Department of Morphology, University of Ibirapuera, São Paulo, Brazil4 Department of Obstetrics, Gynecology and Women's Health, University of Medicine and Dentistry of New Jersey, Newark, USA5 Depatment of Biomedical Sciences, University of Guelph, Ontario, Canada6 Department of Pathology and Molecular Medicine, McMaster University, Hamilton, Ontario, L8N 3Z5 Canada7 Present address: Department of Anatomy and Cell Biology, Queen's University, Kingston, Ontario, K7L 3N6 Canada2005 10 8 2005 3 34 34 2 6 2005 10 8 2005 Copyright © 2005 Albieri et al; licensee BioMed Central Ltd.2005Albieri et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Interferon-gamma (IFN-gamma) mediates diverse functions in bone marrow-derived phagocytes, including phagocytosis and microbe destruction. This cytokine has also been detected at implantation sites under both physiological and pathological conditions in many different species. At these particular sites, the outermost embryonic cell layer in close contact with the maternal tissues, the trophoblast exhibits intense phagocytic activity. To determine whether IFN-gamma affects phagocytosis of mouse-trophoblast cells, ectoplacental cone-derived trophoblast was cultured and evaluated for erythrophagocytosis. Phagocytic activity was monitored ultrastructurally and expressed as percentage of phagocytic trophoblast in total trophoblast cells. Conditioned medium from concanavalin-A-stimulated spleen cells significantly enhanced trophoblast phagocytosis. This effect was blocked by pre-incubation with an anti-IFN-gamma neutralizing antibody. Introduction of mouse recombinant IFN-gamma (mrIFN-gamma) to cultures did not increase cell death, but augmented the percentage of phagocytic cells in a dose-dependent manner. Ectoplacental cones from mice deficient for IFN-gamma receptor alpha-chain showed a significant decrease of the phagocytosis, even under mrIFN-gamma stimulation, suggesting that IFN-gamma-induced phagocytosis are receptor-mediated. Reverse transcriptase-PCR analyses confirmed the presence of mRNA for IFN-gamma receptor alpha and beta-chains in trophoblast cells and detected a significant increase in the mRNA levels of IFN-gamma receptor beta-chain, mainly, when cultured cells were exposed to IFN-gamma. Immunohistochemistry and Western blot analyses also revealed protein expression of the IFN-gamma receptor alpha-chain. These results suggest that IFN-gamma may participate in the phagocytic activation of the mouse trophoblast, albeit the exact mechanism was not hereby elucidated. Protective and/or nutritional fetal benefit may result from this physiological response. In addition, our data also shed some light on the understanding of trophoblast tolerance to inflammatory/immune cytokines during normal gestation.
==== Body
Background
During implantation, mouse trophoblast exhibits an intrinsic potential for phagocytosis, which peaks between days 7 and 9 of pregnancy and is most pronounced in the outermost primary and secondary trophoblast giant cells of the ectoplacental cone [1-4]. This activity declines during a normal gestation, but can resurge under certain conditions [5,6].
Phagocytic activity in post-implantation trophoblast internalizes maternal components, such as uterine epithelial and decidual cells, that are present along the invasion pathway of the trophoblast [3,4]. A role in providing nutrition and space for the early embryonic development is generally attributed to this activity. Intense hemophagocytosis also occurs, which is involved in iron uptake for fetal hemopoiesis [7,8]. Protection against pathogens at the maternofetal interface and immunoregulation of pregnancy has also been implicated as functions for this phagocytosis [6,9-14].
Compared with the phagocytic cells derived from bone marrow, the phagocytic and regulatory processes in the trophoblast have similarities. As with macrophages, phorbol myristate acetate (PMA), all-trans-retinal and complement component 3 enhance trophoblast phagocytosis and trigger the production and release of reactive oxygen species [15-18]. In addition, IFN-γ increases production of nitric oxide by trophoblast cells [19]. The molecular mechanisms involved in the phagocytic process exhibited by trophoblast, however, are not well known. In macrophages, phagocytosis is initiated via a plasma membrane signal that, after activating a JAK-STAT pathway, triggers a sequence of events leading to the internalization of particles [20-22]. Production of IFN-γ by activated, type 1 T lymphocytes and NK cells in response to inflammatory or immune challenges is one of the most effective regulatory signals in this process.
In many different species, IFN-γ is found at the maternofetal interface at specific intervals during normal pregnancy, produced by uterine activated T lymphocytes and natural killer cells [23-26], or even by trophoblast cells [27,28]. In mice, uterine natural killer (uNK) cells seem to be the principal maternal cells producing IFN-γ [24,25,28,29]. The effects of IFN-γ on pregnancy outcomes however, can be pathological or physiological depending upon several factors such as the susceptibility of the mice strain, the concentration of IFN-γ, the stage of pregnancy, the degree of differentiation of the cells at the maternofetal interface and the co-expression with other inflammatory cytokines [30-35]. In vitro, IFN-γ also exhibits a potent ability to induce differentiation in cytokeratin-positive ectoplacental cell populations [36].
Those findings, coupled with the high trophoblast-cell potential for phagocytosis, prompted us to examine a possibility of regulatory roles for IFN-γ in the maternofetal interface. In the present study, we use cultured ectoplacental cone-derived trophoblast to explore the effect of IFN-γ as a phagocytic stimulator at the maternofetal interface.
Methods
Mice and collection of ectoplacental cone tissue
To obtain pregnant females, three separate groups of two-month-old mice were mated as follows: a) F1 [NZW × AKR] females with BALB/c males; b) IFN-γRα-/- females with IFN-γRα-/- males; c) 129/SvJ females with 129/SvJ males. The latter two, which are congenic strains, were purchased from the Jackson Laboratory (Bar Harbor, ME, USA) and bred at the University of Guelph, Ontario, Canada. The F1 mice were obtained from Animal Care Facility of the University of São Paulo Institute Of Biomedical Sciences and mated in Sao Paulo. Estrous females were overnight caged with males (1:1) and examined on the following morning for presence of a vaginal plug indicative of mating. The morning of plug appearance was designated gestation day (gd) 0.5. On gd 7.5, mated females were sacrificed by cervical dislocation and their uteri dissected in phosphate-buffered saline (PBS; Gibco BRL, Grand Island, NY, USA) containing bovine serum albumin (BSA). Embryos at the primitive streak stage were then gently extracted from their decidual capsules. Embryonic ectoderm and the extraembryonic membranes (amnion and chorion) were dissected free from the ectoplacental cones and the cones were collected for the experimental protocols. As previously shown through morphological analysis and cytokeratin immunolocalization [18,19], the dissection of the ectoplacental cones at this time of gestation provides pure preparations of trophoblast cells. All procedures were conducted under aseptic conditions using sterile PBS and 10 % fetal bovine serum (wt/vol, FBS).
The Animal Ethical Committees of both the Institute of Biomedical Sciences of the University of São Paulo and the University of Guelph granted approval for the procedures and experiments.
Ectoplacental cone cultures
Glass coverslips (18 mm in diameter) were placed into the wells of 12-well culture plates, and the wells were filled with 1 mL of standard medium: Dulbecco's Modified Eagle's Medium (D-MEM) (Sigma, Chemical Co., St. Louis, MO, USA), supplemented with 10 % FBS (vol/vol), 0.5 mg/mL lactic acid, 0.2 ng/mL Mito™+ and antibiotics (0.05 mg/mL streptomycin and 50 IU/mL-penicillin). Three to five ectoplacental cones were introduced into each prepared well and cultured under standard conditions (36.5°C, 5 % CO2 and humidity) for 48 h to allow the trophoblast cells to attach to the coverslips and begin their outgrowth. Ectoplacental cones from IFN-γR-/- mice and their congenic 129/SvJ controls were only available for experiments involving treatment with mrIFN-γ. Ectoplacental cones of fetuses from the allogeneically mated F1 mice were used in all experiments.
Baseline and stimulated phagocytic assays
Erythrocytes were used as target cells for trophoblast phagocytosis. The erythrocytes were collected from adult male mice genetically matched to the trophoblast and prepared as described previously [15]. Briefly, the animals were anesthetized and the blood was collected by cardiac puncture. The blood sample was then transferred to heparinized 1.5-mL tubes (Liquemine, Roche Quim., Rio de Janeiro, Brazil), and centrifuged to isolate the erythrocytic fraction. Plasma and white blood cell fractions were discarded and the erythrocytes were harvested and re-suspended in D-MEM. After several cycles of centrifugation and re-suspension in D-MEM, erythrocytes were re-suspended. This was done in order to remove any remaining plasma or white blood cell contamination. The erythrocytes were then counted using a hemocytometer and added to cultures at a concentration of 2 × 107 cells/mL.
Ectoplacental cones were cultured 48 h prior to addition of either erythrocytes or stimulatory agents. The medium was then replaced with 1 mL of standard medium containing the agents to be evaluated (as phagocytic stimulators) and the erythrocytes. This combination was then cultured for an additional 24-h period. At the end of the culture period, specimens were processed for light and electron microscopy.
Erythrocytes were added to all cultures. For stimulation of erythrocyte uptake by trophoblast cells, three experimental groups were performed: i) Concanavalin A (Con A)-stimulated leukocyte-conditioned medium (LCM); ii) LCM+ antibody against IFN-γ and iii) mouse rIFN-γ (proven lipopolysaccharide and endotoxin free, Sigma Chemical Co., St. Louis, MO, USA). The (i) Con A-stimulated LCM was prepared from 24-h cultures of Balb/c splenocytes treated with 5 μg/mL Con-A and added (at 200 μL/mL) to the ectoplacental cone cultures. The (ii) LCM+ antibody against IFN-γ was produced by incubating 200 μL of LCM with monoclonal anti-IFN-γ (XMG 1.2, 200 μg/mL; AbIFN-γ) for 15 min at room temperature prior to its addition to the culture. The (iii) mrIFN-γ was used in the cultures at a concentration of 100 U/mL of medium. Non-stimulated cultures, containing only standard medium, were analyzed as controls in each experiment. Control cultures were also carried out using standard medium (without IFN-γ) and either an irrelevant antibody (rat IgG anti-CD44 glycoprotein) or monoclonal anti-IFN-γ (XMG 1.2) at concentrations of 200 μg/mL.
The concentration of mrIFN-γ was established from dose-response curves based on phagocytic indices evaluated at concentrations of 10, 100 and 500 U/mL. The final values for m rIFN-γ were selected on the basis of maximal dependent stimulation for erythrophagocytosis. The neutralizing dose of AbIFN-γ (200 μg/mL) was estimated based upon dose-response curves (0, 20, 50, 100 and 200 μg/mL) as the dose capable of neutralizing the maximal phagocytic activity in the ectoplacental cone cells.
A minimum of three independent experiments made in triplicate was used for each experimental treatment.
Light and transmission electron microscopic analysis
After the phagocytic assays, at least two samples from each experimental group were fixed in 2 % paraformaldehyde (wt/vol, EMS, Fort Washington, PA, USA) in 0.1 M PBS, pH 7.2. Fixed cultures were either stained with toluidine blue and observed under light microscopy or processed for cytokeratin A and IFN-γα chain receptor immunohistochemistry localization. In addition, one unfixed sample per group was also stained with YOPRO-1 for apoptosis detection.
For the remaining samples, the ectoplacental cones were first scraped from the culture wells and then fixed in 2.5 % glutaraldehyde (vol/vol, Sigma Chemical Co., St. Louis, MO, USA) in 0.1 M PBS, pH 7.4, for 30 min. The samples were post-fixed in osmium tetroxide (Polysciences, Warrington, PA, USA) in PBS, dehydrated in a graded ethanol series and embedded in Spurr's resin. Sections of 1-μm thickness were cut, stained with toluidine blue and examined under light microscopy. Thin sections were also prepared and stained with 2 % uranyl acetate (wt/vol) and 0.5 % lead acetate (wt/vol) before examination with a Jeol CX-100 transmission electron microscope.
All experiments were prepared and performed in triplicate on three different occasions so that, from each experimental group, 9 samples, containing an average of 27 ectoplacental cones, were available for quantitative analysis.
Measurement of phagocytic activity and statistical analysis
Ectoplacental cones were completely sectioned in semi-serial 1 μm-thickness sections. From each sample, the slide showing the greatest portion of the ectoplacental cone was selected for quantitative analysis. Phagocytic activity was then measured in toluidine blue-stained sections in a Nikon Optiphot light microscope, employing a 40x objective at a final magnification of × 400. Trophoblast erythrophagocytosis was estimated as the percentage of trophoblast giant cells (TGC) that contained one or more erythrophagosomes.
The phagocytic index (PI) was calculated as follows:
The PIs of ectoplacental cones sharing the same well were analyzed and presented as a group mean (n). For each experimental variant, a minimum of 3 groups was analyzed and the results presented as mean ± SD. Differences among the groups were analyzed by ANOVA followed by Tukey-Kramer multiple comparisons test or Student t-test. Differences were considered significant at p ≤ 0.05.
Immunohistochemistry
Ectoplacental cone cultures grown on glass coverslips were washed in PBS and fixed for 10 min at -20°C in 1 % paraformaldehyde in methanol (vol/vol). After fixation, the cells were washed in Tris-buffered saline (TBS), pH 8.2, at 4°C and then immunostained using the following incubation sequence: a) TBS-0.2 % glycine (vol/vol) for 10 min at room temperature, followed by acetone at -20°C for 15 min and TBS-0.05 % saponin (wt/vol, for recovery of antigenic sites) for 5 min; b) 5 % aqueous acetic acid (wt/vol, to block endogenous alkaline phosphatase activity) for 8 min; c) 1 % goat non-immune serum (wt/vol) at 37°C for 1 h; d) biotin-conjugated monoclonal rat anti-mouse CD 119 (1:10, IFN-γ receptor α chain-GR20; BD Biosciences, Palo Alto, CA, USA) or rat anti-mouse A cytokeratin (1:5, cytokeratin Endo-A; Developmental Studies Hybridoma Bank, Iowa City, IA, USA) at 4°C overnight; e) biotinylated goat anti-rat IgG (1:300, Vector Lab., Burlingame, CA, USA) at 37°C for 1 h. The reaction signal was amplified using the ABC system (Biomeda, Foster City, CA, USA), revealed with fast red/naphthol (Sigma) and counterstained with Mayer's hematoxylin. Substituting normal rat serum for the primary antibodies carried out negative control.
Western Blot analysis
In two independent experiments, total protein extracted from 121 cultured ectoplacental cones was mixed with an equal volume of SDS-PAGE sample buffer, boiled for 10 min, and then separated on 12 % SDS-PAGE gels. After electrophoresis, proteins were transferred to membranes (Hybond-ECL; Amersham, UK), which were then blocked in 5 % dry milk in TBS-Tween 20 (1 h) at room temperature, rinsed, and incubated (12 h, 4°C) with the primary antibody rat monoclonal IgG anti-IFN-γ receptor α chain (GR20; 2.0μg/mL in TBS; BD Biosciences, Palo Alto, CA, USA). The primary antibodies were removed and the membranes were washed in TBS and sequentially incubated with TBS-Tween 20-anti-rat IgG-biotinylated secondary antibody (1:1000, Vector Lab., Ca, USA), and with avidin conjugated to horseradish peroxidase (Vectastain ABC kit, Vector Lab., Ca, USA) according to manufacturer's instructions. The membranes were washed again and the bound enzyme detected by chemiluminescence following the manufacturer's protocols (Amersham Biosciences, UK).
Analysis of cell death
Detection of cell death was carried out using the fluorescent probe YOPRO-1 (Molecular Probes, Eugene, OR, USA), which selectively binds to apoptotic nuclei (37). After the stimulatory assay using different concentrations of mrIFN-γ (0, 100, 1,000 and 10,000 U/mL), treated and untreated cultures were washed in standard medium and incubated with YOPRO-1 (10 μM/mL in the same medium, for 10 min). Apoptotic cells were identified by green fluorescent nuclei on a Zeiss confocal laser scanning microscope (Zeiss LSM 510, EXmax/Exmin = 494/518 nm), using the 488-nm line of the Argon laser and connected to an inverted fluorescence microscope (Zeiss Axiovert 100 M). The number of reactive cells per ectoplacental cone in three independent experiments was analyzed and the results presented as mean ± SD. per experimental treatment. Differences among the groups were analyzed by ANOVA followed by Tukey test and considered significant at p ≤ 0.05.
Semiquantitative RT-PCR analysis
RNA was isolated from ectoplacental cones cultured with or without 100 U/mL IFN-γ. The cultures were washed using 0.1 M PBS and total RNA was isolated using Trizol (Gibco BRL) according to manufacturer's instructions. The RNA was diluted in diethyl pyrocarbonate-treated (DEPC)-water, quantified by spectrometry and used for reverse transcriptase PCR (RT-PCR).
Reverse transcriptase and PCR assays were performed according to manufacturer instructions, employing the Superscript Preamplification System for First Strand cDNA Synthesis (Gibco BRL). Briefly, cDNA was synthesized from 1 μg of total RNA after priming with oligo (dT). The final reaction volume was diluted to 20 μL, and 2 μL of each cDNA sample were used as a template for gene-specific PCR amplifications. PCR amplification was performed in a 50-μL final reaction mixture using a thermocycler (BioRad, Hercules, CA, USA). Cyclophilin was used as a gene amplification control and as the reference for quantitative analysis. The PCR products were separated by electrophoresis using a 1 % agarose gel and revealed through ethidium bromide staining. The PCR cycle consisted of an initial single denaturation at 94°C for 5 min, followed by denaturation at 94°C for 30 sec, annealing at 60°C for 30 sec and elongation at 72°C for 60 sec. A total of 30 cycles were run. Cycle number was selected to allow quantitative comparison of the samples in a linear way. The primers used for the amplification of IFN-γRα and β chain primers (Life Technologies, Gibco BRL, NY, USA) were those described by Lucas et al. [22] and contained in the Gene Bank database of the National Center for Biotechnology Information . The primers sequences and the respective fragment lengths are: IFN-γRα (608 bp): sense – 5'-CGGTCGAAAAAGAAGAGTGTA-3'; antisense – 5'-tcgggagtgataggcggtgag-3', IFN-γRβ (528 bp): sense- 5'-TACACTTCTCCCCTCCCTTTG-3'; antisense- 5'-ACATCATCTCGGTCCTTTTCT-3'. cyclophilin [[38], 668 bp] sense- 5'-CTTGCTGCAGACATGGTC-3'; antisense- 5'-GCAATCCTGCTAGACTTG-3'.
The density of each band was measured using a Kodak Digital Science™, ID Image Analysis Software (Eastman Kodak Company, Rochester, NY) and expressed as optical density. Data are presented as a ratio compared to cyclophilin expression. This particular approach to semiquantitative RT-PCR was based upon analytical data published by Kinoshita et al. [39]. Statistical analysis of the data for each gene was performed using the Student's t-test to compare two groups. Data are presented as mean ± SD. A value of p < 0.05 was considered to indicate a statistically significant difference.
Results
Trophoblast cell characteristics in IFN-γ-treated and untreated cultures
The general morphology of cultured ectoplacental cone-trophoblast cells was similar to that described previously [15,18]. After 48 h of culture, the peripheral trophoblast cells had already formed an extensive monolayer of large cells and surrounded a core of small cells. Cytokeratin A was detected in all cells of the ectoplacental cone, indicating that only trophoblast cells had been isolated and cultured (Figs. 1A–C). Fluorescent staining with YOPRO-1, nuclear dye that does not label living cells, showed few trophoblast-giant cells in the process of programmed cell death in the control (1.66 ± 0.57/ectoplacental cone) and IFN-γ-treated groups (Figs. 1D–I) at concentrations of 100 and 1,000 U/mL (respectively, 2.33 ± 0.57 and 3.33 ± 1.52/ectoplacental cone). At concentrations of 10,000 U/mL, the dying cells increased (12.33 ± 2.51/ectoplacental cone).
Figure 1 Forty eight hour-cultured ectoplacental cone cells. Samples were stained for cytokeratin-A intermediate filaments typical of trophoblast cells (A-C) or YO-PRO to localize cell death (D-I). IFN-γ treatments were: D-E, no treatment;F-G, 100 U/mL and,H-I, 10,000 U/mL. The arrows in figure C highlight cytoplasmic filaments and in figures D-I, dead cells. A-C, Alkaline phosphatase immunostaining. B, Negative control in which the primary antibody was omitted. D, F and H, Fluorescence microscopy. E,G and I, Nomarski imaging of D,F and H, respectively. The unique bar in figure A represents: 300 μm in A, 160 μm in B, and 80 μm in C.
Trophoblast cells from either IFN-γRα-/- mice (Fig. 2D) or from normal animals (Figs. 2A–C), from LCM mrIFN-γ treated and untreated groups showed very similar cellular and subcellular characteristics, except for the prominence of different stages of erythrocyte internalisation and erythrophagosomes in the trophoblast giant cells of the treated groups (Figs. 2B–D).
Figure 2 Cultured trophoblast cells from NZW × AKR F1 (A-C) and IFN-γRα-/- mice (D). (A) Cells of ectoplacental cone cultured in the presence of erythrocytes in standard culture medium show no phagosomes (final magnification, × 8,500), (B-D) whereas the cytoplasm of trophoblast cells stimulated with mrIFN-γ exhibit large numbers of erythrophagosomes (*). In the electron micrograph (C), different stages of erythrocyte internalization are seen (↑). (▲, free erythrocytes) (B, toluidine blue, × 640; C, × 5,500; D, × 16,500).
Trophoblast cell phagocytosis in stimulated cultures
Table 1 provides an analysis of the phagocytic behavior of ectoplacental cones under each treatment employed. The addition of LCM to ectoplacental cone/erythrocyte cultures significantly enhanced erythrophagocytosis in comparison to that observed in non-stimulated control cultures. Use of a neutralizing antibody for IFN-γ significantly reduced LCM-stimulated activity, suggesting that much of the activity in the LCM was due to IFN-γ. In addition, mrIFN-γ also caused a synchronous increment in trophoblast giant cell PI.
Table 1 Erythrophagocytic activity of cultured F1-trophoblast giant cells treated with leukocyte conditioned medium and mrIFN-γ
Culture treatment n PI (%)
Standard culture medium (SCM) 5 5.1 ± 0.95
SCM + AbIFN-γ 3 7.14 ± 1.93
SCM + AbC 4 6.44 ± 3,34
Leukocyte conditioned medium (LCM) 7 38.95 ± 3.50*
LCM+AbIFN-γ 4 18.55 ± 0.22 */**
mrIFN-γ 100 U/mL 3 34.99 ± 1.40*/***
Values significantly different at p < 0.001: *compared to the control containing only standard culture medium and erythrocytes (SCM); ** compared to treatment with leukocyte conditioned medium (LCM) and, *** compared to treatment with LCM (200 μL/mL) plus AbIFN-γ. PI, phagocytic index calculated as the percentage of trophoblast giant cells containing erythrophagosomes. AbIFN-γ, monoclonal antibody XMG 1.2 against IFN-γ (200 μg/mL). AbC, isotype matched control antibody, (200 μg/mL). mrIFN-γ, mouse recombinant IFN-γ (100 U/mL).
The doses used for mrIFN-γ were established from dose-response curves (Fig. 3A), which indicated doses of 100 U/mL for IFN-γ as the most effective for enhancing phagocytosis in trophoblast cells. On the other hand, LCM used in combination with different concentrations of the neutralizing antibody against IFN-γ reduced the trophoblast phagocytic response in comparison to LCM alone (Fig. 3B). The predicted effect of immunoglobulin in the phagocytic activity of the trophoblast was also tested using a non-relevant antibody or the neutralizing AbIFN-γ in a standard culture system in the presence or absence of LCM. The phagocytosis indices, however, did not vary significantly in relation to each respective control. In the system in which LCM was introduced, the phagocytic indices remained as high as those found when no antibody was introduced (Table 1).
Figure 3 Effect of (A) mrIFN-γ and (B) monoclonal antibody XMG 1.2 anti-IFN-γ on the erythrophagocytosis exhibited by NZW × AKR F1 gd-7.5 trophoblast cells. (A) Ectoplacental cones were cultured for 48 h in the mrIFN-γ concentrations indicated in the graphics. Note that the highest phagocytic activity occurs at the concentration of 100 U/mL. The cultures that received the monoclonal antibody (B, 0–200 μg/mL), also received 200 μL/mL of leukocyte-conditioned medium (LCM). Data represent the mean ± SD of three separate experiments and asterisks show statistical differences in relation to the control (p < 0.05).
Trophoblast giant cells from both IFN-γRα-/- and 129/SvJ mice exhibited very low PIs under non-stimulated conditions. The PIs were slightly higher in the IFN-γRα-/- ectoplacental cone cultures supplemented with mouse rIFN-γ but were still lower than PI values seen in trophoblast cells from congenic controls receiving similar treatment (Table 2).
Table 2 Effects of mouse recombinant IFN-γ on erythrophagocytosis exhibited by trophoblast cells from IFN-γRα-/- mice
Culture treatment n PI (%)
Rα-/- mice SCM (standard culture medium) 4 14.72 ± 4.05
SCM + mrIFN-γ (100 U/mL) 5 14.99 ± 10.88
Congenic animals SCM (standard culture medium) 4 11.52 ± 4.65
SCM + mrIFN-γ (100 U/mL) 3 33.46 ± 5.72 *, **
* Values significantly different compared to cultures using IFN-γ-treated and untreated Rα-/- mice (p <0.05), **compared to cultures of congenic animals that did not receive IFN-γ (p < 0.01) (ectoplacental cone from wild type mice). PI, phagocytic index calculated as the percentage of trophoblast giant cells containing erythrophagosomes. SCM, standard culture medium.
Analysis of the expression of IFN-γRα and β chains in trophoblast cell cultures
To evaluate whether the phagocytic response of trophoblast was mediated via IFN-γ receptors, we evaluated receptor expression at the mRNA and protein levels.
IFN-γ α-chain receptor was immunolocalized mainly to spread trophoblast cells. Many had giant cell features (Fig. 4). In the proliferative central region of the cultured ectoplacental cone, very few cells were reactive. This pattern of reactivity was maintained in cultures exposed or not exposed to IFN-γ.
Figure 4 Protein expression of IFN-γ receptor α chain in cultured ectoplacental cone cells. Photomicrographs A, B and C show trophoblast reactive cells (arrows) for IFN-γ receptor α chain in mrIFN-γ-treated (A-B) and non-treated cultures (C). (D) Control reaction in which, rat non-immune serum was substituted for the primary antibody. No positive cells were observed. The insert in panel C (upper left) displays a Western blot analysis for IFN-γ receptor α chain protein expression. Lane 1 contains the molecular weight marker and Lanes 2 and 3 total proteins from 121-cultured ectoplacental cones probed with anti-IFN-γ receptor α chain antibody (A, C, × 100; B,D, × 880).
Total RNA was isolated from ectoplacental cones cultured with or without IFN-γ. This RNA was used for RT-PCR evaluation IFN-γRα and β chains. mRNA for both receptor chains (α and β) was detected in trophoblast. In addition, semi-quantitative analyses for both receptor chains normalized against cyclophilin expression demonstrated elevated transcription after mrIFN-γ treatment. Greater expression of IFN-γ receptor β chain than α chain was observed (Fig. 5).
Figure 5 Gene expression of IFN-γα and β chain receptor mRNA in cultured ectoplacental cones. (A) A representative RT-PCR analysis of mRNA transcript in ectoplacental cone outgrowth cultures lacking (lanes 2 to 4) or with (lanes 5 to 7) mrIFN-γ. Lane 1 represents the DNA ladder, lanes 2 and 5, cyclophilin, used as an internal cDNA control (668 bp), lanes 3 and 6, IFN-γ receptor α chain (608 bp) and lanes 4 and 7, IFN-γ receptor β chain (529 bp). (B) Semi-quantitative RT-PCR analysis of IFN-γ receptor chain expression. A summary of 3 experiments is shown in the bar graph (mean and S.D.). Transcripts were expressed as a ratio between band intensities of the IFN-gamma receptor chain and cyclophilin. Asterisk shows statistical differences in relation to the control (c) in which no IFN-γ has been added (p < 0.05). Numbers represent: 1 and 2, IFN-γ α chain receptor expression in the absence (c) and presence of IFN-γ in the culture medium, and 3 and 4, IFN-γ β chain receptor expression in the absence (c) and presence of IFN-γ.
Discussion
This study confirms data from the literature showing that trophoblast is able to react to immune regulatory molecules by altering its phagocytic behavior [5,11]. Here, we showed that mouse ectoplacental-cone cultures treated with either Con A-activated leukocyte-culture supernatants or low doses of mrIFN-γ contain more trophoblast cells in active phagocytosis than untreated control cultures. Among cytokine mixture present in culture supernatants from stimulated leukocytes, special emphasis was given to IFN-γ and its roles. The phagocytic activity of treated trophoblast cells was specifically neutralized in the presence of antibodies against IFN-γ. This supported the hypothesis that in the conditioned medium, IFN-γ was the molecule most effective in mediating trophoblast erythrophagocytosis. Direct use of low doses of mrIFN-γ in the trophoblast cultures extended this hypothesis.
On the other hand, in the present study we cannot role out the possibility of other unidentified regulatory molecules might be also involved, since we find that the phagocytic activity of ectoplacental cone cells relevantly decreases, but not to control baseline levels in the presence of leukocyte conditioned medium plus neutralizing antibody.
Another physiological question addressed herein was the pathway by which IFN-γ acts on trophoblast cells. Expression of IFN-γ receptor chains was demonstrated at both the mRNA and protein levels suggesting that trophoblast phagocytic activity in the presence of IFN-γ is a receptor-mediated response. A second set of experiments, employing mice that do not express the IFN-γ receptor also predicted indicated the phagocytic response was receptor-dependent. IFN-γR mRNA has previously been detected in pre-implantation mouse embryos and in placentas from gd 14 until birth [40,41]. Taken together with our results, these findings indicate that IFN-γR expression is a fundamental characteristic of trophoblast cells throughout pregnancy. Higher expression of IFN-γ receptor alpha and beta chains in IFN-γ-treated cultures shows that trophoblast cells also respond to IFN-γ in a manner analogous to the responses of macrophages to this cytokine [22].
Although further analyses are required to define the molecular mechanisms involved in IFN-γ-stimulated phagocytosis, our results, suggest that IFN-γ may trigger an autocrine cascade in trophoblast giant cells. IFN regulatory factor-1 (IRF-1) is highly expressed in the mouse trophoblast [42,43]. This IFN-γ-induced transcription factor activates the promoters of many interferon-regulated genes [43], and may a key to understanding the responsiveness of the trophoblast to IFN-γ. Alternatively, IFN-γ stimulated trophoblast cells produce and release nitric oxide [19]. In macrophages, this is one of the most important events triggered by IFN-γ receptor stimulation [44,45].
Despite the numerous studies that show the presence of IFN-γ-producing cells in the pregnant uterus [24,26-29,36,46-48] and, that IFN-γ can be beneficial to pregnancy if secreted at appropriate times, concentrations and locations [29], this cytokine is considered an abortion-inducing factor. [30,32,49-54]. In mice, high, in vivo doses of IFN-γ (3 × 105 IU/mL) are deleterious to early embryos [55,56] while in vitro treatments reduce trophoblast outgrowth, limiting invasive potential. At low doses (100 IU/mL) IFN-γ increases total cell numbers in cultured ectoplacental cones and promotes their differentiation. In humans, IFN-γ has been associated with direct effects on the viability of cultured term trophoblast cells, whether or not TNF-α was present in the medium [53,57]. Mechanisms participating in IFN-γ-induced fetal death have not been clearly defined and it is quite probable that other types of cells, in addition to trophoblast cells, participate in the in vivo outcomes. Clark et al. found [52,58] that an abnormal increase in endogenous IFN-γ (1,000 IU) combined with high TNF-α (2,000 IU) on gd 7.5 led to abortion within 48 h. These cytokine levels correlated with strong expression of fgl2 prothrombinasein decidua as well as in trophoblast suggesting a maternal vascular etiology with thrombosis and ischemia. It is therefore likely that the concentration of IFN-γ, its balance with other pro- and anti-inflammatory cytokines, and the stage of gestation at which is, produced are fundamental in defining whether IFN-γ plays a physiological or pathological role during pregnancy.
Acknowledgements
We wish to express our appreciation to Dr. Patricia Gama and Eunice R. de Andrade Sá for helping with Western blot analysis, to Dr. John McNamara for the English revision of the manuscript and to FAPESP, CNPq, NSERC and the Government of Iran for financial grants and fellowships.
==== Refs
Tachi C Tachi S Lindner HR Ultrastructural features of blastocyst attachment and trophoblastic invasion in the rat J Reprod Fertil 1970 21 37 56 5461005
Welsh AO Enders AC Light and electron microscopic examination of the mature decidual cells of the rat with emphasis on the antimesometrial decidua and its degeneration Am J Anat 1985 172 1 29 3969928 10.1002/aja.1001720102
Bevilacqua E Abrahamsohn PA Ultrastructure of trophoblast giant cell transformation during invasive stage of implantation of the mouse embryo J Morphol 1988 198 341 351 3221406 10.1002/jmor.1051980308
Bevilacqua E Abrahamsohn PA Trophoblast invasion during implantation of mouse embryo Arch Biol Med Exp 1989 22 107 118 2619314
Athanassakis I Bleackely RC Paetkau V Guilbert L Barr PJ Wegmann TG The immunostimulatory effect of T cells and T cell lymphokines on murine fetally derived placental cells J Immunol 1987 138 37 44 3491160
Amarante-Paffaro A Queiroz GS Corrêa ST Spira B Bevilacqua E Phagocytosis as a potential mechanism for microbial defense of mouse placental trophoblast cells Reproduction 2004 128 207 218 15280560 10.1530/rep.1.00214
Billington WD Biology of the trophoblast Adv Reprod Physiol 1971 5 27 66 4354660
Van Dijk JP van Noort WL Kross MJ Starreveld JS van Eijk HG Isotransferrins and pregnancy: a study in the guinea pig Clin Chem Acta 1991 203 1 16 10.1016/0009-8981(91)90151-2
Pavia CS Expression of cell-mediated antimicrobial immunity by mouse trophoblast monolayers J Infect Dis 1983 146 1006 1010 6343521
Drake BL Rodger JC Phagocytic properties of cultured murine trophoblast Placenta 1987 8 129 134 3615373
Guilbert L Robertson AS Wegmann TG The trophoblast as an integral component of macrophage-cytokine network Immunol Cell Biol 1993 71 49 57 8436411
Mitchell MD Trautman MS Dudley DJ Cytokine networking in the placenta Placenta 1993 14 249 275 8367410
Delgado MA Santos-Buch CA Transplacental transmission and fetal parasitosis of Trypanosoma cruzi in outbred white Swiss mice Am J Trop Med Hyg 1978 27 1108 1115 103444
Loke YM Transmission of parasites across the placenta Adv Parasitol 1982 21 155 228 6762065
Albieri A Bevilacqua E Induction of erythrophagocytic activity in cultured mouse trophoblast cells by phorbol myristate acetate and all-trans-retinal Placenta 1996 17 507 512 8899880 10.1016/S0143-4004(96)90033-8
Gagioti S Colepicolo P Bevilacqua E Post-implantation mouse embryo has the capability to generate and release reactive oxygen species Reprod Fertil Dev 1995 7 1111 1116 8848578
Gagioti S Colepicolo P Bevilacqua E Reactive oxygen species and phagocytosis process of hemochorial trophoblast J Braz Assoc Adv 1996 48 37 42
Albieri A Kipnis T Bevilacqua E A possible role for activated complement component 3 in phagocytic activity exhibited by the mouse trophoblast Am J Reprod Immunol 1999 41 343 352 10378030
Gagioti S Scavone C Bevilacqua E Participation of the mouse implanting trophoblast in nitric oxide production during pregnancy Biol Reprod 2000 62 260 268 10642561
Lewis CE McGee JOD The Macrophage 1992 Oxford: IRL Press
Darnell JE Jr Kerr IM Stark GR Jak-STAT pathways and transcriptional activation in response to IFNs and other extracellular signaling proteins Science 1994 264 1415 1421 8197455
Lucas DM Lokuta MA McDowell MA Doan JE Paulnock DM Analysis of the IFN-gamma-signaling pathway in macrophages at different stages of maturation J Immunol 1998 160 4337 4342 9574537
Guimond MJ Luross JA Wang B Terhorst C Danial S Croy BA Absence of natural killer cells during murine pregnancy is associated with reproductive compromise in TgE26 mice Biol Reprod 1997 56 169 179 9002646
Croy BA Chantakru S Esadeg S Ashkar AA Wei Q Decidual natural killer cells: key regulators of placental development (a review) J Reprod Immunol 2002 57 151 168 12385840 10.1016/S0165-0378(02)00005-0
Ashkar AA Croy BA Interferon-γ contributes to the normalcy of murine pregnancy Biol Reprod 1999 61 493 502 10411532
Arck PC Ferrick DA Steele-Norwood D Croitoru K Clark DA Regulation of abortion by γδT cells Am J Reprod Immunol 1997 37 87 93 9138458
Fowler AK Reed CD Giron DJ Identification of an interferon in murine placentas Nature 1980 286 266 267 6157101 10.1038/286266a0
Platt JS Hunt JS Interferon-gamma gene expression in cycling and pregnancy mouse uterus: temporal aspects and cellular localization J Leukoc Biol 1998 64 393 400 9738667
Ashkar AA Di Santo JP Croy BA Interferon γ contributes to initiation of uterine vascular modification, decidual integrity, and uterine Natural Killer cell maturation during normal murine pregnancy J Exp Med 2000 192 259 269 10899912 10.1084/jem.192.2.259
Chaouat G Menu E Clark DA Dy M Minkowski M Wegmann TG Control of fetal survival in CBA × DBA/2 mice by lymphokine therapy J Reprod Fertil 1990 89 447 458 2119428
Chaouat G Meliane AA Martal J Raghupathy R Elliot J Mosmann T Wegmann TG IL-10 prevents naturally occurring fetal loss in the CBA × DBA/2 mating combination, and local defect in IL-10 production in this abortion-prone combination is corrected by in vivo injection of IFN-τ J Immunol 1995 154 4261 4268 7722286
Vassiliadis S Tsoukatos D Athanassakis I Interferon-induced class II expression at the spongiotrophoblast zone of the murine placenta is linked to fetal rejection and developmental abnormalities Acta Physiol Scand 1994 151 485 495 7976422
Krishnan L Guilbert LJ Wegmann TG Belosevic M Mosmann TR T helper 1 response against Leishmania major in pregnant C57BL/6 mice increases implantation failure and fetal resorptions. Correlation with increased IFN-gamma and TNF and reduced IL-10 production by placental cells J Immunol 1996 156 653 662 8543817
Clark DA Chaouat G Arck PC Mittruecker HW Levy GA Cutting edge: Cytokine-dependent abortion in CBA × DBA/2 mice is mediated by the procoagulant fgl2 prothombinase J Immunol 1998 160 545 549 9551885
Raghupathy R Pregnancy: success and failure within the Th1/Th2/Th3 paradigm Sem Immunol 2001 13 219 227 10.1006/smim.2001.0316
Athanassakis I Papadimitriou L Bouris G Vassiliadis S Interferon-γ induces differentiation of ectoplacental cone cells to phenotypically distinct trophoblast Dev Comp Immunol 2000 24 663 672 10831800 10.1016/S0145-305X(00)00023-9
Idziorek T Estaquier J De Bels F Ameisen JC YOPRO-1 permits cytofluorometric analysis of programmed cell death (apoptosis) without interfering with cell viability J Immunol Methods 1995 185 249 258 7561136 10.1016/0022-1759(95)00172-7
Weisinger G Gavish M Mazurika C Zinger O Transcription of actin, cyclophilin and glyceraldehyde phosphate desidrogenase genes: tissue and treatment specificity Biochem Bioph Acta 1999 1446 225 232
Kinoshita T Imamura J Nagai H Shimotohno K Quantification of gene expression over a wide range by the polymerase chain reaction Anal Biochem 1992 206 231 235 1280006 10.1016/0003-2697(92)90358-E
Chen HL Kanath R Pace JL Russel SW Hunt JS Expression of the interferon-gamma receptor gene in mouse placentas is related to stage of gestation and is restricted to specific subpopulations of trophoblast cells Placenta 1994 15 109 121 8008727
Truchet S Wietzerbin J Debey P Mouse oocytes and preimplantation embryos bear the two sub-units of interferon-gamma receptor Mol Reprod Dev 2001 60 319 330 11599043 10.1002/mrd.1094
Ashkar AA Croy BA Interferon regulatory factor-1 has important functions in placental growth and modifications of decidual arterioles Placenta 1999 20 A11
Kamijo R Harada H Matsuyama T Bosland M Gerecitano J Shapiro D Le J Koh SI Kimura T Green SJ Requirement for transcription factor IRF-1 in NO synthase induction in macrophages Science 1994 263 1612 1615 7510419
Corradin SB Buchmuller-Rouillier Y Mauel J Phagocytosis enhances murine macrophage activation by interferon-gamma and tumor necrosis factor Eur J Immunol 1991 21 2553 2558 1915557
Ryu SY Jeong KS Kang BN Park SJ Yoon WK Kim SH Kim TH Modulation of transferrin synthesis, transferrin receptor expression, iNOS expression and NO production in mouse macrophages by cytokines, either alone or in combination Anticancer Res 2000 20 3331 3338 11062761
Delassus S Coutinho GC Saucier S Darche S Kourilsky P Differential cytokine expression in maternal blood and placenta during murine gestation J Immunol 1994 152 2411 2420 8133052
Makhseed M Raghupathy R Azizieh F Farhat R Hassan N Bandar A Circulating cytokines and CD30 in normal human pregnancy and recurrent spontaneous abortions Hum Reprod 2000 15 2011 2017 10967006 10.1093/humrep/15.9.2011
Athanassakis I Iconomidou B Cytokine production in the serum and spleen of mice from day 6 to 14 of gestation: cytokines/placenta/ spleen/serum Dev Immunol 1996 4 247 255 8924760
Haimovich F Hill JA Anderson DJ The effects of soluble products of activated lymphocytes and macrophages on blastocyst implantation events in vitro Biol Reprod 1991 44 69 75 2015353
Mattson R Holmdahl R Scheynius A Bernadotte F Mattson A Avan der Meide PH Placental MHC class I antigen expression is induced in mice following in vivo treatment with recombinant interferon-gamma J Reprod Immunol 1991 19 115 129 1901087 10.1016/0165-0378(91)90012-F
Raghupathy R Makhseed M Azizieh F Hassan N Al-Azemi M Al-Shamali E Maternal Th1 and Th2 type reactivity to placental antigens in normal human pregnancy and unexplained recurrent spontaneous abortions Cell Immunol 1999 196 122 130 10527564 10.1006/cimm.1999.1532
Clark DA Lea RG Podor T Daya S Banwatt D Harley C Cytokines determining the success or failure of pregnancy Ann NY Acad Sci 1991 626 524 536 1829345
Yui J Garcia-Lloret M Wegmann TG Guilbert LJ Cytotoxicity of tumour necrosis factor-alpha and gamma-interferon against primary human placental trophoblasts Placenta 1994 15 819 835 7886023
Daher S Denardi KAG Blotta MHSL Mamoni RL Reck APM Camano L Mattar R Cytokines in recurrent pregnancy loss J Reprod Immunol 2004 62 151 157 15288190 10.1016/j.jri.2003.10.004
Fontana V Choren V Vauthay L Calvo JC Calvo L Cameo M Exogenous interferon-γ alters murine inner cell mass and trophoblast development. Effect on the expression of ErbB1, ErbB4 and heparin sulfate proteoglycan (perlecan) Reproduction 2004 128 717 725 15579589 10.1530/rep.1.00335
Cameo M Fontana V Cameo P Vauthay LG Kaplan J Tesone M Similar embryotoxic effects of sera from infertile patients and exogenous interferon-γ on long-term in-vitro development of mouse embryo Hum Reprod 1999 14 959 963 10221227 10.1093/humrep/14.4.959
Pijnenborg R Luyten C Vercruysse L Keith JC Van Assche FA Cytotoxic effects of tumor necrosis factor (TNF)-α and interferon-γ on cultured human trophoblast are modulated by fibronectin Mol Hum Reprod 2000 6 635 641 10871651 10.1093/molehr/6.7.635
Clark DA Ding J-W Yu G Levy GA Gorcynski RM Fgl2 prothrombinase expression in mouse trophoblast and decidua triggers abortion but may be countered by OX-2 Mol Hum Reprod 2001 7 185 194 11160845 10.1093/molehr/7.2.185
|
16092971
|
PMC1215515
|
CC BY
|
2021-01-04 16:37:11
|
no
|
Reprod Biol Endocrinol. 2005 Aug 10; 3:34
|
utf-8
|
Reprod Biol Endocrinol
| 2,005 |
10.1186/1477-7827-3-34
|
oa_comm
|
==== Front
Reprod Biol EndocrinolReproductive biology and endocrinology : RB&E1477-7827BioMed Central London 1477-7827-3-351609553710.1186/1477-7827-3-35ResearchPrevalence of ultrasonography proved polycystic ovaries in North Indian women with type 2 diabetes mellitus Zargar Abdul H [email protected] Vipin K [email protected] Arshad I [email protected] Shariq R [email protected] Mir I [email protected] Bashir A [email protected] Mohammad A [email protected] Mohammad [email protected] Departments of Endocrinology and Immunology, Sheri-Kashmir Institute of Medical Sciences Srinagar, J&K, India2005 11 8 2005 3 35 35 6 4 2005 11 8 2005 Copyright © 2005 Zargar et al; licensee BioMed Central Ltd.2005Zargar et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Polycystic ovaries (PCO) and their clinical expression (the polycystic ovary syndrome [PCOS]) as well as type 2 diabetes mellitus (T2DM) are common medical conditions linked through insulin resistance. We studied the prevalence of PCO and PCOS in women with diet and/or oral hypoglycemic treated T2DM and non-diabetic control women.
Design
Prospective study.
Methods
One hundred and five reproductive age group women with diet and /or oral hypoglycemic treated T2DM were the subjects of the study. Sixty age-matched non-diabetic women served as controls. Transabdominal ultrasonographic assessment of the ovaries was used to diagnose PCO. Clinical, biochemical and hormonal parameters were also noted.
Results
Ultrasonographic prevalence of PCO was higher in women with diabetes than in non-diabetic subjects (61.0% vs. 36.7%, P < 0.003) whereas that of PCOS was 37.1% in diabetic subjects and 25% in non-diabetic controls (P > 0.1). Diabetic women with PCO had diabetes of significantly longer duration than those without PCO (4.19±2.0 versus 2.9±1.6 yrs; p < 0.05). Among both diabetic and non-diabetic women, those with PCO had significantly higher plasma LH, LH/FSH ratio, total testosterone and androstenedione levels.
Conclusion
This study demonstrates a higher prevalence of PCO in women with T2DM as compared to non-diabetic subjects.
==== Body
Introduction
Polycystic ovary syndrome (PCOS) has, since its first description by Stein and Leventhal in 1935, become one of the commonest disorders in women, affecting 5 to 10% in the reproductive age-group [1-3]. This common medical condition has brought gynecologists, endocrinologists, cardiologists, pediatricians, and dermatologists together [4]. The disorder manifests as obesity, hirsutism, menstrual disturbances, acne vulgaris, male-pattern baldness, recurrent abortions, infertility, an-ovulation, and psychological and psychosexual morbidity [5]. It is an important cause of hirsutism and of infertility in our population [6,7]. With the advances and refinement in ultrasound technology, the presence of polycystic appearance of ovaries has increasingly been accepted as an essential part of the diagnosis of this syndrome and ultrasound suggestion of PCO has recently been introduced as one of the three diagnostic criteria of PCOS in Rotterdam consensus workshop 2003 [8]. Using the most conservative criteria for transabdominal studies, such as presence of ≥ 10 peripheral cysts per ovary, each 2–8 mm, and thickened stroma, some authors have shown that as many as 21–23% of normal female population have PCO on ultrasound [9-11]. Some women with otherwise confirmed PCOS may have normal sized ovaries on ultrasound and normal sized ovaries without increased folicllarity don't preclude the diagnosis in proper clinical settings [12-14]. The pathophysiology of PCOS, one of the commonest endocrinopathies, is far from clear. A plethora of data point to insulin resistance (IR) and hyperinsulinemia as the central factor, a suggestion supported by clinical benefits of insulin sensitizers [3,5,15,16].
T2DM, which accounts for 90–95% of all diabetes mellitus, is a common metabolic disorder, characterized by a combination of IR and insulin secretory defects, occurring in varying proportions [17]. Both T2DM and PCOS are now considered to be two metabolically similar but phenotypically different expressions of the same syndromic continuum with IR being the common and pivotal link, which may be genetically determined [15,18]. Women with PCOS are also at increased risk for cardiovascular disease (CVD), given the high prevalence of the metabolic syndrome among them [19]. Many studies have revealed an increased prevalence of various abnormalities of glucose tolerance in women with PCOS [3,20,21]. The scarcity of converse data on the prevalence of PCO among women with T2DM prompted this study on the prevalence of PCO in a cohort of premenopausal women with insulin naive T2DM.
Subjects and Methods
The study was conducted in the department of Endocrinology of Sheri-Kashmir Institute of Medical Sciences Srinagar Kashmir (India), a tertiary care center. One hundred and sixty five premenopausal (≤ 45 years) women, from Kashmir region of India, were studied for sonographic suggestion of PCO in addition to recording of clinical, biochemical and hormonal manifestations of PCOS. Informed consent was obtained from all subjects and the institutional review board, that looks in to ethical aspects of human experimentation approved the study.
Subjects
a. Patient group: This group comprised of 105 non-pregnant women with T2DM of at least one year duration treated with medical nutrition therapy and/or oral hypoglycemic drugs. Women with current or past history of insulin treatment, oral contraceptive use, and suspicion of nephropathy were excluded from the study.
b. Control group: Sixty non-diabetic, healthy pre-menopausal volunteers served as controls.
Methods
Each patient and control received a detailed clinical examination and underwent a relevant laboratory evaluation.
a. Clinical assessment: The history focused on age, menstrual pattern, fertility status, duration and treatment of diabetes (in diabetic women), and other relevant drug history. Menstrual pattern was characterized as regular (cycles recurring every 21–35 days), oligommenorrhea (cycle length over 35 days and under six months), polymenorrhea (cycles occurring more frequently than every 21 days), menorrhagia (heavy menstruation requiring intervention) and amenorrhoea (absence of menstruation for six months or longer). Fertility status was classified as fertile (had a previous pregnancy with no subsequent infertility), infertile (primary or secondary infertility of at least 1 year duration) and unproven (pregnancy not attempted).
The physical examination, apart from a general review of the systems, focused on the assessment of androgen status (hirsutism, temporal recession of hair, acne, muscle bulk, clitoromegaly), evidence of IR (acanthosis nigricans), and anthropometry, (body mass index (BMI), waist circumference). Ferriman and Gallwey scoring system was used to assess hirsutism, [22]. and a score of 7 or more was taken as significant.
b. Laboratory evaluation. The baseline investigations included blood counts, chest x-ray, electrocardiogram, liver and kidney function tests, glucose and lipid (total cholesterol, HDL, and triglycerides) levels. A pool of three samples taken 15 minutes apart during early follicular phase was analyzed for LH, FSH, total testosterone and androstenedione levels. The blood sample for all investigations were taken in the fasting state in a single visit. The samples were taken in early follicular phase (day 3 – day 5) in women with regular cycles in all subjects and controls.
c. Ultrasonography. A specifically trained radiologist, using Siemens Sonoline Adara with 3.5 MHz convex electronic probe performed the transabdominal ultrasonography. Local socio-cultural constraints precluded a vaginal approach. The diagnosis of PCO was made, if 10 or more follicles, each 2–8 mm diameter were present in the ovarian periphery, and stroma was echo dense [23]. The ovarian volume was calculated by measuring the diameters in three dimensions and assuming an ellipsoid shape using the formula, volume = height × width × depth × π/6.
d. Hormonal assays. All hormone estimations were done with radioimmunoassay. Total testosterone and androstenedione were measured using commercial kits from DPC, Los Angeles CA. Serum LH and FSH was assayed by immunoradiometric assays using kits from Medicorp Inc., Montreal, Canada. Sensitivity, specificity, and inter- and intra-assay coefficients of variation were within the prescribed limit as provided by the manufacturer.
Statistical analysis
SPSS 10.0 package was used for analysis of data. Correlations were done by Pearson's test. Categorical variables were compared by using Chi-square test, and for continuous variables, 't' test was used for comparing two groups and one way ANOVA when more than two groups were involved. Two tailed significance was calculated and a P value of <0.05 was taken as significant.
Results
The baseline characteristics of diabetic and control women (Table 1) were comparable except for significantly higher systolic and diastolic blood pressures in the former. Table 2 compares the various metabolic and hormonal parameters between diabetic and non-diabetic women. While the mean fasting blood glucose and LDL levels were significantly higher in diabetic women, the mean levels of cholesterol, HDL, triglycerides and VLDL were similar in diabetic and control groups. The diabetic women had a significantly higher mean ovarian volume than controls (6.15 ± 1.4 ml vs. 5.66 ± 1.4 ml, P < 0.04). On ultrasonography, PCO was found in 64 (60.9%) and 22 (36.7%) of diabetic and control women respectively, giving a significantly higher prevalence in the diabetic women (P < 0.05). In women with PCO, those with and without T2DM had comparable mean ovarian volumes (5.80 ± 1.18 vs. 5.98 ± 1.28 ml; P > 0.5). Among women with T2DM, those with PCO had a significantly higher mean ovarian volume than those without PCO (5.80 ± 1.18 vs. 4.13 ± 1.07 ml; P < 0.001).
Table 1 Clinical profile of study subjects
Clinical parameter Patients (n = 105) Controls (n = 60) p
Age (yrs) 36.19 ± 4.37* 34.98 ± 4.60 0.096
SBP (mmHg) 138.95 ± 12.39 132.75 ± 10.17 0.01†
DBP (mmHg) 84.84 ± 7.72 81.13 ± 8.57 0.05†
BMI (kg/m2) 25.42 ± 3.2 25.72 ± 3.2 0.57
W/H ratio 0.94 ± 0.7 0.93 ± 0.6 0.92
* Mean ± SD; †Significant
Table 2 Metabolic and hormonal profile of study subjects
Parameter Patients (n = 105) Controls (n = 60) P
Fasting glucose (mg/dl) 120.61 ± 34.79* 88.07 ± 14.18 0.001†
Total cholesterol (mg/dl) 193.51 ± 47.88 183.63 ± 39.72 0.177
Triglycerides (mg/dl) 193.69 ± 68.80 185.68 ± 70.89 0.478
LDL cholesterol (mg/dl) 128.96 ± 61.9 102.78 ± 40.1 0.004†
HDL cholesterol (mg/dl) 42.52 ± 7.56 43.99 ± 8.92 0.262
VLDL cholesterol (mg/dl) 38.7 ± 13.7 37.1 ± 14.1 0.478
LH (IU/L) 5.91 ± 5.72 4.67 ± 6.44 0.20
LH/FSH 1.01 ± 0.73 0.84 ± 0.66 0.132
Testosterone (ng/ml) 0.29 ± 0.22 0.27 ± 0.14 0.335
Androstendione (ng/ml) 1.87 ± 1.70 1.54 ± 0.64 0.157
* Mean ± SD; †Significant
Table 3 shows clinical features in subjects with and without PCO on ultrasonography in both groups. Overall, PCO-syndrome (PCOS) was seen in 39 (37.1%) subjects with T2DM and 15 (25.0%) controls. However, this difference was not statistically significant (P > 0.1). The mean duration of diabetes and mean serum LH, testosterone, androstenedione, and LH/FSH ratio were significantly higher in diabetic women with than in those without PCO (Tables 3 &4). While non diabetic controls with and without PCO had comparable age, age at menarche, diastolic blood pressure, BMI and waist hip ratios, the former had a significantly higher mean systolic blood pressure (Table 3). As in the diabetic women, non-diabetic controls with PCO had higher LH, LH/FSH ratio, testosterone and androstenedione levels than those without PCO; among both diabetic and non diabetic women, the blood glucose, and lipid levels were comparable in those with and without PCO (Table 4). The clinical and biochemical parameters of diabetic and non-diabetic women with PCO were comparable (Tables 3 &4). Of 105 diabetic subjects, 55 were overweight-obese (BMI ≥ 25) whereas of 60 controls, 30 were overweight-obese. Table 5 shows clinical features and frequency of PCO and PCOS in these obese and non-obese subjects.
Table 3 Clinical profile of diabetic patients and controls with and without PCO
Clinical Parameter Diabetic subjects Controls P value (ANOVA)
PCO (n = 64) Non-PCO (n = 41) PCO (n = 22) Non-PCO (n = 38)
Age (years), Mean ± SD 36.8 ± 4.28* 35.2 ± 4.4 34.2 ± 4.5 35.5 ± 4.6 0.000
Duration of DM (y), Mean ± SD 4.19 ± 2.0 2.9 ± 1.6 a -- -- 0.000
Age of menarche (y), Mean ± SD 13.5 ± 0.8 13.50 ± 0.6 13.1 ± 0.8 13.1 ± 0.7 0.015
Menstrual irregularity, 24 (37.5%) 4 (9.8%) 0 0 0.000
Oligomenorrhea 12 (18.8%) 1 (2.4%)
Polymenorrhea 1 (1.6%) 1 (2.4%)
Mennorrhagia 11 (17.2%) 2 (4.9%)
Infertility 4 (6.2%) 0 0 0 0.368
Hirsuitism 9 (14.1%) 0 4 (18.2%) 2 (5.3%) 0.030
Acne 3 (4.7%) 0 1 (4.5%) 1 (2.6%) 0.564
Acanthosis 2 (3.1%) 0 0 0 0.368
History of hypertension 40 (78%) 30 (73%) 0 0 0.000
SBP (mmHg), Mean ± SD 140 ± 12.7 139 ± 12.3 138 ± 9.5 130 ± 9.3 b 0.000
DBP (mmHg), Mean ± SD 85.8 ± 7.8 83.3 ± 7.3 82.1 ± 10.2 80.58 ± 7.5 0.012
BMI (kg/m2), Mean ± SD 25.8 ± 3.1 24.7 ± 3.3 26.36 ± 2.8 25.34 ± 3.3 0.180
Waist to hip ratio, Mean ± SD 0.94 ± 0.6 0.92 ± 0.7 0.94 ± 0.6 0.92 ± 0.6 0.426
Ovarian volume (ml), Mean ± SD 5.80 ± 1.18 4.13 ± 1.07 5.98 ± 1.28 3.89 ± 0.88 0.000
PCOS† 39 (61%) 0 15 (68%) 0 0.000
* on ultrasonography
a: P < 0.05, in the subject group, b: P < 0.05, in control group
† PCOS by Rotterdam ASRM/ESHRE consensus criteria
Table 4 Showing comparison of metabolic and hormonal profile in diabetic subjects and non-diabetic controls on the basis of presence or absence of ultrasonic suggestion of polycystic ovaries (PCO).
Clinical parameter Diabetic subjects Controls P (ANOVA)
PCO (n = 64) Non-PCO (n = 41) PCO (n = 22) Non-PCO (n = 38)
Fasting glucose (mg/dl) 123.8 ± 36.28 115.55 ± 32.0 92.14 ± 12.7 85.71 ± 14.6 0.000
Total cholesterol (mg/dl) 189.4 ± 40.9 199.8 ± 57.0 183.27 ± 37.6 183.87 ± 41.3 0.375
Triglycerides (mg/dl) 194.4 ± 69.9 192.4 ± 67.8 182.27 ± 52.1 187.65 ± 80.2 0.895
HDL cholesterol (mg/dl) 42.0 ± 6.6 43.3 ± 8.8 44.00 ± 11.0 43.99 ± 7.6 0.593
LDL cholesterol (mg/dl) 127.5 ± 55.2 131.1 ± 71.9 97.74 ± 40.7 105.76 ± 40.0 0.033
VLDL cholesterol (mg/dl) 38.8 ± 13.9 38.4 ± 13.5 36.45+10.4 37.53+16.0 0.895
LH (IU/L) 7.51+6.65 3.42 +2.21a 7.93 ± 9.52 2.77 ± 2.17b 0.000
FSH (IU/L) 5.82+2.80 7.23+4.86 5.50 ± 2.3 5.11 ± 2.9 0.039
LH/FSH ratio 1.31 ± 0.78 0.55 ± 0.26a 1.3 ± 0.85 0.5 ± 0.26b 0.000
Testosterone (ng/ml) 37.4 ± 23.0 16.9 ± 7.3a 39.6 ± 13.3 18.8 ± 7.5b 0.000
Androstendione (ng/ml) 2.40 ± 1.9 1.04 ± 0.36a 2.10 ± 0.6 1.21 ± 0.40b 0.000
* Mean ± SD
a: P < 0.05, in the subject group, b: P < 0.05, in control group
Table 5 Clinical features and frequency of PCO/PCOS in obese and non-obese subjects
Characteristic Diabetic cases Controls P value
Obese (n = 55) Non-obese (n = 50) Obese (n = 30) Non-obese (n = 30)
Menstrual irregularity
Oligomenorrhea 9 (16.4%) 4 (8.0%) 0 0 0.014
Polymenorrhea 2 (3.6%) 0 0 0 0.256
Mennorrhagia 9 (16.4%) 4 (8.0%) 0 0 0.014
Hirsutism 7 (12.7%) 2 (4.0%) 2 (6.7%) 4 (13.3%) 0.346
Acne 2 (3.6%) 1 (2.0%) 1 (3.3%) 1 (3.3%) 0.966
Acanthosis nigricans 2 (3.6%) 0 0 0 0.256
PCO 34 (61.8%) 30 (60.0%) 14 (46.7%) 8 (26.7%) 0.009
PCOS 22 (40.0%) 17 (34.0%) 10 (33.3%) 8 (16.7%) 0.181
Discussion
PCOS and T2DM are common medical conditions and both are associated with IR and compensatory hyperinsulinemia [24]. The clinical and biochemical features of PCOS are heterogeneous and there is much debate whether it represents a single or several disorders [2]. In recent years it has become apparent that PCOS not only is the most frequent cause of anovulation and hirsutism, but also is associated with characteristic metabolic disturbances that may have implications for long-term health [2,4]. Recently IR has been clearly documented by many investigators in PCOS, although the cause of IR is unknown and its relationship with PCOS as an etiological agent or an epiphenomenon remains unsettled [3,15,25,26]. T2DM is a common disorder characterized by IR and compensatory hyperinsulinemia [17]. An increased incidence of glucose intolerance and DM has been reported in women with a history of PCOS compared to controls in long-term studies [3,20,21]. If IR and hyperinsulinemia have an important role in determining ovarian morphology, PCO and their clinical expression can be expected to be more common among women with T2DM. We compared the prevalence of PCO on ultrasound in a defined clinic population of women with T2DM and in a group of age matched non-diabetic control women.
In our study 60.95% of the 105 premenopausal women with T2DM were documented to have PCO compared to 36.66% of the 60 non-diabetic controls. This prevalence in our study is much higher than the reported prevalence of 21–23% among normal female population [9-11]. In previous studies we observed PCOS to be second common cause of both hirsutism and infertility in our population [6,7]. Rodin et al reported an overall prevalence of 52% of PCO in women of Indian subcontinent origin living in England [27]. Conn et al studied 38 premenopausal women using the same ultrasonographic criteria and reported prevalence 82% among women with T2DM [24]. Not much data are available on the prevalence of PCO among diabetic women.
In our study diabetic females with PCO had significantly higher mean LH, total testosterone and androstenedione than diabetic females without PCO. Conn and colleagues found comparable mean levels of LH and testosterone in diabetic women with and without PCO and significantly higher levels of androstenedione in the former [24]. Another study, including women with and without diabetes, found significantly higher mean testosterone levels in women with PCO than in those without [27]. No statistically significant differences were found between mean systolic and diastolic blood pressures, BMI, waist hip ratios, HDL-cholesterol and triglyceride levels in diabetic women with and without PCO like observations made earlier by Conn et al [24]. We did not find any difference in the prevalence of PCO among obese (BMI ≥ 25 kg/m2) and non-obese (BMI <25 kg/m2) women with T2DM. No study has compared the prevalence of PCO among obese and non-obese type 2 diabetic women. As in the patient group, non-diabetic controls with PCO had significantly higher mean LH, testosterone and androstenedione levels and mean ovarian volume than those without PCO; all clinical and metabolic parameters were comparable between these two subgroups of non-diabetic controls except mean systolic blood pressure which was significantly more among controls with than without PCO (138.1 vs. 129.6 mmHg).
We found a significantly higher mean LDL levels in: a) diabetic women than non-diabetic women (128 mg/dl vs. 102 mg/dl), b) women with PCO and T2DM than women with PCO and without T2DM (127 mg/dl vs. 97 mg/dl) and c) obese women with PCO and T2DM than obese women with PCO but without T2DM (139 mg/dl vs. 87 mg/dl). This likely reflects the effect of diabetes on plasma lipids. An earlier study by Rodin found no difference in respect to LDL, HDL, triglycerides or total cholesterol in women of Indian sub continental ancestry with and without PCO [27]. Comparing women with PCO from the diabetic and control groups, a significant difference was found in the mean age of women from these two groups (36.83 vs. 34.18 years; P < 0.02). The mean LDL level in women with T2DM and PCO was significantly higher than in controls with PCO (127 vs. 97 mg/dl; P < 0.03); other parameters including diastolic blood pressure, BMI, waist hip ratio, cholesterol, triglycerides, HDL, VLDL, LH, testosterone and androstenedione and ovarian volume were comparable between women with PCO and with or without T2DM.
The significantly longer duration of diabetes among diabetic patients with PCO compared to those without PCO in our study probably reflects longer duration of insulin resistance/hyperinsulinemia in the former. IR is considered central to the pathogenesis of PCOS and south Asian women with PCOS have been reported to be more insulin resistant, seek treatment at a younger age and have more severe symptoms than Caucasians [28]. Hyperinsulinemia augments androgen production in PCOS directly by augmenting LH activity through stimulation of ovarian receptors of insulin and insulin-like growth factors or indirectly, by enhancing the amplitude of serum LH pulses [5]. Furthermore, hyperinsulinemia decreases SHBG levels increasing the amount of free testosterone available to act on target organs [24]. Legro and colleagues reported that 50–70% of women with PCOS have IR which probably contributes to hyperandrogenism underlying the signs and symptoms of PCOS [29]. While IR is a common feature of both PCOS and T2DM, persistent reproductive dysfunction appears to be limited to the former raising the possibility that IR in the ovary itself may confer this susceptibility. The exact role and mechanism(s) thereof IR and hyperinsulinemia are far from fully understood. While Conn et al,24 found fasting insulin levels in diabetic women with and without PCO to be comparable, Rodin et al [27]. observed in a more genetically homogenous group of Asian women, that those with both PCO and T2DM were more insulin resistant than those with diabetes alone.
In conclusion our study demonstrates a higher prevalence of PCO and its clinical expressions in women with T2DM compared to non-diabetics. The precise role of hyperinsulinemia in development and expression of PCO remains largely unresolved and needs elucidation.
==== Refs
Stein LF Leventhal MC Amenorrhea associated with bilateral polycystic ovaries Am J Obstet Gynecol 1935 29 181 191
Franks S Polycystic ovary syndrome N Engl J Med 1995 333 853 861 7651477 10.1056/NEJM199509283331307
Dunaif A Insulin resistance and the polycystic ovary syndrome: Mechanism and implications for pathogenesis Endocr 1997 18 774 800
De Leo V Musacchio MC Morgante G La Marca A Petraglia F Polycystic ovary syndrome and type 2 diabetes mellitus Minerva Ginecol 2004 56 53 62 14973410
Tsilchorozidou T Overton C Conway GS The pathophysiology of polycystic ovary syndrome Clin Endocrinol (Oxf) 2004 60 1 17 14678281 10.1046/j.1365-2265.2003.01842.x
Zargar AH Wani AI Masoodi SR Laway BA Bashir MI Salahuddin M Epidemiologic and etiologic aspects of hirsutism in Kashmiri women in the Indian Subcontinent Fertil Steril 2002 77 674 678 11937114 10.1016/S0015-0282(01)03241-1
Zargar AH Wani AI Masoodi SR Laway BA Salahuddin M Epidemiologic and etiologic aspects of infertility in the Kashmir region of India Fertil Steril 1997 68 637 643 9341602 10.1016/S0015-0282(97)00269-0
The Rotterdam ESHRE/ASRM – sponsored PCOS Consensus Workshop Group The Netherlands. Revised 2003 consensus on diagnostic criteria and long term health risks related to polycystic ovary syndrome Fertil Steril 2004 81 19 25
Polson DW Wordsworth J Adams J Franks S Polycysic ovaries – a common finding in normal women Lancet 1988 1 870 872 2895373 10.1016/S0140-6736(88)91612-1
Farquhar CM Birdsall M Manning P Mitchell JM France JT The prevalence of polycystic ovaries on ultrasound scanning in a population of randomly selected women Australia & New Zealand J Obstet Gynaecol 1994 34 67 72
Clayton RN Ogden V Hodgkinson J Worswick L Rodin DA Dyer S Meade TW How common are polycystic ovaries in normal women and what is their significance for the fertility of the population? Clin Endocrinol (Oxf) 1992 37 127 134 1395063
Atiomo WU Pearson S Shaw S Prentice A Dubbins P Ultrasound criteria in the diagnosis of polycystic ovary syndrome (PCOS) Ultrasound Med Biol 2000 26 977 980 10996697 10.1016/S0301-5629(00)00219-2
Fox R Transvaginal ultrasound appearances of the ovary in normal women and hirsute women with oligomenorrhea Australian and New Zealand J Obstet Gynecol 1999 39 63 68
Hann LE Hall DA McArdle CR Seibel M Polycystic ovarian disease: sonographic spectrum Radiology 1984 150 531 534 6691114
Sam S Dunaif A Polycystic ovary syndrome: syndrome XX? Trends Endocrinol Metab 2003 14 365 370 14516934 10.1016/j.tem.2003.08.002
Seli E Duleba AJ Treatment of PCOS with metformin and other insulin-sensitizing agents Curr Diabet Reports 2004 4 69 75
The Expert Committee on the Diagnosis and Classification of Diabetes Mellitus Follow-up report on the diagnosis of diabetes mellitus Diabet Care 2003 26 3160 3167
Waterworth DM Benett ST Gharani N McCarthy MI Hague S Batty S Conway GS White D Todd JA Franks S Williamson R Linkage and association of insulin gene VNTR regulatory polymorphism with PCOS Lancet 1997 349 986 990 9100625 10.1016/S0140-6736(96)08368-7
Legro RS Polycystic ovary syndrome and cardiovascular disease: a premature association? Endocr Rev 2003 24 302 312 12788801 10.1210/er.2003-0004
Legro RS Kunselman A Dodson WC Dunaif A Prevalence and predictors of risk for type 2 diabetes mellitus and impaired glucose tolerance in polycystic ovary syndrome: a prospective, controlled study in 254 affected women J Clin Endocrinol Metab 1999 84 165 169 9920077 10.1210/jc.84.1.165
Ganie MA Khurana ML Eunice M Ammini AC Prevalence of glucose intolerance among adolescent and young women with polycystic ovary syndrome in India Program and abstracts of 11th International Congress on Hormonal Steroids and Hormones and Cancer 2002 Fukouka, Japan 133 [abstract no. P3–30]
Ferriman D Gallwey JD Clinical assessment of body hair growth in women J Clin Endocrinol Metab 1961 21 1440 1447 13892577
Conn JJ Jacobs HS Conway GS The prevalence of polycystic ovaries in women with type 2 DM Clinical Endocrinol 2000 52 81 86 10.1046/j.1365-2265.2000.00884.x
Adams J Polson DW Franks S Prevalence of polycystic ovaries in women with anovulation and idiopathic hirsutism Br Med J 1986 293 355 359 3089520
Cordain L Eades MR Eades MD Hyperinsulinemic diseases of civilization: more than just Syndrome X Comp Biochemistry Physiol A Mol Integr Physiol 2003 136 95 112 10.1016/S1095-6433(03)00011-4
Prelevic GM Insulin resistance in polycystic ovary syndrome Curr Opinions Obstet Gynecol 1997 91 193 201
Rodin DA Bano G Bland JM Taylor K Nussery SS Polycystic ovaries and associated metabolic abnormalities in Indian subcontinent Asian women Clinical Endocrinol 1998 49 91 99 10.1046/j.1365-2265.1998.00492.x
Wijeyaratne CN Balen AH Barth JH Belchetz PE Clinical manifestations and insulin resistance (IR) in polycystic ovary syndrome (PCOS) among South Asians and Caucasians: is there a difference? Clin Endocrinol (Oxf) 2002 57 343 350 12201826 10.1046/j.1365-2265.2002.01603.x
Legro RS Castracane VD Kaufmann RP Detecting insulin resistance in polycystic ovary syndrome: purposes and pitfalls Obstet Gynecol Surv 2004 59 141 154 14752302 10.1097/01.OGX.0000109523.25076.E2
|
16095537
|
PMC1215516
|
CC BY
|
2021-01-04 16:37:13
|
no
|
Reprod Biol Endocrinol. 2005 Aug 11; 3:35
|
utf-8
|
Reprod Biol Endocrinol
| 2,005 |
10.1186/1477-7827-3-35
|
oa_comm
|
==== Front
Reprod Biol EndocrinolReproductive biology and endocrinology : RB&E1477-7827BioMed Central London 1477-7827-3-421613733310.1186/1477-7827-3-42ResearchPeptidylarginine deiminase (PAD) is a mouse cortical granule protein that plays a role in preimplantation embryonic development Liu Min [email protected] Andrea [email protected] Patricia [email protected] Michiyuki [email protected] Scott A [email protected] Prue [email protected] Department of Cell Biology and Neuroscience, University of California, Riverside, California 92521, USA2 Department of Anatomy and Medicine, School of Medicine, University of California, San Francisco, California 94143, USA3 Graduate School of Integrated Science, Yokohama City University, Yokohama, 236-0027 Japan4 Weill Medical College of Cornell University, New York, NY 10021, USA2005 1 9 2005 3 42 42 18 7 2005 1 9 2005 Copyright © 2005 Liu et al; licensee BioMed Central Ltd.2005Liu et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
While mammalian cortical granules are important in fertilization, their biochemical composition and functions are not fully understood. We previously showed that the ABL2 antibody, made against zona free mouse blastocysts, binds to a 75-kDa cortical granule protein (p75) present in a subpopulation of mouse cortical granules. The purpose of this study was to identify and characterize p75, examine its distribution in unfertilized oocytes and preimplantation embryos, and investigate its biological role in fertilization.
Results
To identify p75, the protein was immunoprecipitated from ovarian lysates with the ABL2 antibody and analyzed by tandem mass spectrometry (MS/MS). A partial amino acid sequence (VLIGGSFY) was obtained, searched against the NCBI nonredundant database using two independent programs, and matched to mouse peptidylarginine deiminase (PAD). When PAD antibody was used to probe western blots of p75, the antibody detected a single protein band with a molecular weight of 75 kDa, confirming our mass spectrometric identification of p75. Immunohistochemistry demonstrated that PAD was present in the cortical granules of unfertilized oocytes and was released from activated and in vivo fertilized oocytes. After its release, PAD was observed in the perivitelline space, and some PAD remained associated with the oolemma and blastomeres' plasma membranes as a peripheral membrane protein until the blastocyst stage of development. In vitro treatment of 2-cell embryos with the ABL2 antibody or a PAD specific antibody retarded preimplantation development, suggesting that cortical granule PAD plays a role after its release in preimplantation cleavage and early embryonic development.
Conclusion
Our data showed that PAD is present in the cortical granules of mouse oocytes, is released extracellularly during the cortical reaction, and remains associated with the blastomeres' surfaces as a peripheral membrane protein until the blastocyst stage of development. Our in vitro study supports the idea that extracellular PAD functions in preimplantation development.
==== Body
Background
Mammalian cortical granules are membrane-bound organelles located in the cortex of unfertilized oocytes [1,2]. Following gamete membrane fusion, cortical granules undergo exocytosis, and some of the released components block polyspermy by modifying the zona pellucida [3-14]. In addition, some cortical granule proteins remain associated with the embryo and appear to regulate embryogenesis, since in vitro culture of 2-cell embryos in the presence of antibodies specific to these proteins inhibited embryo cleavage [15-17]. While most cortical granules are released after fertilization, a subpopulation of Lens culinaris agglutinin (LCA)-binding cortical granules are released around the cleavage furrow during first polar body extrusion [18]. While the biological significance of this pre-fertilization release is not yet known, it likely plays a role in fertilization since it occurs at a specific time and place and involves a specific population of cortical granules. These prior studies show that mammalian cortical granules are released both before and after fertilization and that their functions are probably more complex than previously realized.
The total number of mammalian cortical granule proteins has been estimated to be between four and fourteen or more [10,19,20]. Several specific proteins have been identified as cortical granule proteins [21]. N-acetylglucosaminidase was detected in exudates of ionophore-activated mouse oocytes using an enzymatic assay and was localized in the cortical granules at the electron microscopic level [13]. Approximately 90% of oocyte N-acetylglucosaminidase was released following in vivo fertilization and was shown using competitive inhibitors or anti-N-acetylglucosaminidase antibodies to be responsible for the zona block to polyspermy [13]. Ovoperoxidase was detected in the cortical granules of unfertilized mouse oocytes at the ultrastructural level using the 3.3'-diaminobenzidine (DAB) [7,8]. Following artificial activation, ovoperoxidase was present on the oocyte's surface, in the perivitelline space, and in the zona pellucida. Following fertilization, the enzyme was inferred to harden the zona pellucida, since both peroxidase inhibitors and tyrosine analogs prevented hardening [8]. Calreticulin, an endoplasmic reticulum protein involved in calcium storage, was demonstrated in granules in the cortex of hamster oocytes by indirect immunofluorescence [22]. However, a subsequent study showed that most of the granules containing calreticulin did not label with the lectin LCA, a classical marker for mouse oocyte cortical granules [23]. This lead to the conclusion that calreticulin is localized in a population of granules that is distinct from classical cortical granules.
In addition, several proteins (p32, p56, p62, and p75) have been localized immunocytochemically in cortical granules, but their identities have not yet been established [17,19,20]. p32 was recognized on western blots by a monoclonal antibody (3E10) made against mouse cortical granule exudates and was localized immunohistochemically to cortical granules in germinal vesicle intact and metaphase II stage mouse oocytes [19]. Interestingly, p32 was not detected in 3E10 labeled fertilized oocytes and preimplantation embryos following the cortical reaction. While the function of p32 is not known, treatment of unfertilized oocytes with the 3E10 antibody did not increase polyspermy, indicating that for the experimental conditions used, p32 did not function in blocking polyspermy. The polyclonal antibody ABL2, which was made against zona free mouse blastocysts and which immunoprecipitates a 75-kDa protein from mouse oocytes, reacts immunocytochemically with cortical granules [20]. The protein is released following in vitro fertilization and artificial activation [20]. In hamster oocytes, a pair of cortical granule proteins designated p56 and p62, was recognized on western blots by the ABL2 antibody [16]. These two ABL2 specific hamster cortical granule proteins are related to sea urchin hyalin since they are also recognized by the S. purpuratus hyalin specific antibody IL2 [17]. p56 and p62 are retained in the perivitelline space and on the oolemma after fertilization. These proteins appear to be involved in early embryogenesis since in vivo treatment of 2-cell embryos with IL2 or ABL2 antibodies inhibited blastomere cleavage [16,17]. In vitro treatment of 2-cell mouse embryos with the ABL2 antibody showed similar inhibition of development [15]. Although experimental and immunohistochemical work has been done on these cortical granule proteins, they have not yet been identified biochemically or characterized functionally.
The purpose of this study was to identify the mouse cortical granule protein p75, to characterize its distribution in unfertilized oocytes and preimplantation embryos, and to examine its function in fertilization. To accomplish this, p75 was immunoprecipitated from an ovarian lysate, isolated using SDS-PAGE, then analyzed using tandem mass spectrometry. A partial peptide sequence of the protein was obtained and used to identify p75 as a member of the peptidylarginine deiminase (PAD) family of enzymes that catalyze the conversion of arginine to citrulline [24].
Materials and methods
Chemicals and Supplies
Chemicals used to make all media, polyvinylpyrrolidone (PVP), bovine serum albumin (BSA), pregnant mare's serum gonadotropin (PMSG), human chorionic gonadotropin (hCG), bovine hyaluronidase, protein A-sepharose beads, M16 medium, paraformaldehyde, Triton-X 100, α-D-mannose, N-acetylglucosamine, β-D-galactose, and N-acetylgalactosamine were purchased from Sigma Chemical Company (St. Louis, MO). HEPES buffer, light mineral oil, slides, and coverslips (#1.5) were purchased from Fisher (Tustin, CA). Lens culinaris agglutinin (LCA), streptavidin conjugated to Texas Red, and Vectashield mounting medium were purchased from Vector Laboratories (Burlingame, CA). SYTOX orange nucleic acid stain and Alexa-488 conjugated to goat anti-rabbit IgG were obtained from Molecular Probes (Eugene, OR). PAD V (N) antibody was made against recombinant human PAD V and affinity purified on an N-terminal PAD V fragment (1–262) bound column as previously described [25]. ePAD antibody was made against the N-terminal fragment (1–200) of mouse recombinant ePAD [26].
Animals
NIH Swiss white mice were purchased from Harlan (San Diego, CA). Mice were housed in a University of California at Riverside vivarium with a 14-hour light and 10-hour dark cycle and fed water and Purina rodent chow (Ralston-Purina, St. Louis, MO) ad libitum. Protocols used in this study were approved by the campus Committee on Animal Care.
Media and Fixatives
For dissection and oocyte collection, Earle's balanced salt solution with 28.18 mM of sodium bicarbonate and 24.98 mM of HEPES free acid (EBSS-H), pH 7.4 supplemented with 0.3% of polyvinylpyrrolidone (EBSS-H/0.3% PVP) was made as previously described [27]. For immunoprecipitation, lysis buffer was made with 150 mM NaCl, 10% NP-40, 0.5% sodium deoxycholate, 0.1% SDS, 50 mM Tris-HCl, pH 7.5, and a protease inhibitor cocktail as previously described [13]. For egg activation, calcium and magnesium free EBSS-H (EBSS-Ca/Mg-H) was used as previously described [19]. High salt-containing solution was made by increasing the sodium chloride concentration in EBSS-H/0.3% PVP to 300 mM. For embryo culture, M16 medium was pregassed in 37°C humidified incubator (5% CO2, 95% air) overnight before use. For confocal scanning laser microscopy, Dulbecco's phosphate buffered saline (DPBS), pH 7.4 or phosphate buffered saline (PBS), pH 7.4 was used. DPBS was made with 90.9 mM CaCl2, 2.68 mM KCl, 1.47 mM KH2PO4, 0.49 mM MgCl2·6H2O, 136.89 mM NaCl, and 8.06 mM Na2HPO4·7H2O. PBS was made as described previously [25]. For fixation, 4% paraformaldehyde was made in DPBS, pH 7.4, or in PBS, pH 7.4. Blocking solution was made in DPBS, pH 7.4 supplemented with 7.5 mg/ml glycine and 3 mg/ml BSA immediately prior to use. In some cases, blocking solution was made in PBS. 10 mM citrate buffer pH 7.0 was made with 3.78 g of citric acid and 2.411 g of sodium citrate in 1 L of H2O. To remove peripheral ABL2 specific antigen following egg activation, high salt-containing EBSS-H/0.3% PVP containing 300 mM NaCl was used. For confocal scanning laser microscopy, labeling solution was made by supplementing DPBS, pH 7.4 with 30 mg/ml BSA (DPBS/3% BSA). For LCA blotting, Tris-buffered saline (TBS), pH 7.6 was used (147 mM NaCl; 20 mM Tris-base)
Oocyte and Embryo Collection
For epifluorescence microscopy, confocal scanning laser microscopy, and gel electrophoresis, oocytes and preimplantation embryos were collected in EBSS-H/0.3% of PVP at room temperature. To collect germinal vesicle intact oocytes, female mice were injected intraperitoneally with 10 IU of PMSG (Sigma, St Louis, MO). Oocytes were collected 60 hours later from the ovaries and mechanically denuded of their cumulus cells with a thin-bore glass pipette. Unfertilized mature metaphase II oocytes were collected from female mice that were primed with 10 IU of PMSG at 2200 hours on day 1 followed by 10 IU of hCG (Sigma, St. Louis, MO) 46 hours later. For egg activation, oocytes were flushed out from oviducts with collection medium 16 to 18 hours post the hCG injection. To collect in vivo fertilized oocytes and preimplantation embryos, female mice were superovulated by intraperitoneal injection of 10 IU of PMSG at 1430 hours on day 1 followed by 10 IU of hCG 46 hours later and then placed in cages containing 2–3 male mice. The following day, fertilized oocytes were collected by flushing the oviduct with collection medium. Only oocytes with two pronuclei were used. Two-cell preimplantation embryos were collected by flushing the oviduct with collection medium 2 days after mating. Four- and eight-cell preimplantation embryos were collected by flushing the oviduct or the uterine horns with collection medium 3 days after the mating. Blastocysts were collected by flushing the uterine horns 4 days after mating.
For mature metaphase II oocytes and in vivo fertilized oocytes, cumulus cells were removed by incubating oocytes in collection medium containing 100 IU of hyaluronidase for 5 minutes at room temperature. In some experiments, zonae pellucidae were removed with 0.25% pronase in collection medium.
Human Peripheral Blood Cell Collection
Human peripheral blood cells were obtained from an informed and consenting healthy donor. Red blood cells were removed by sedimentation with dextran 200,000, and the remaining cells were then subjected to Percoll density-gradient centrifugation. Layers containing granulocytes were collected, and cells were then spread onto glass slides by cytospinning.
Immunoprecipitation
For immunoprecipitation, all steps were carried out in lysis buffer unless otherwise specified. Ovaries from adult female mice were dissected out in EBSS-H and homogenized on ice. The homogenate was kept on ice for one hour then centrifuged at 30,000 g at 4°C for 30 minutes to remove any insoluble material. The supernatants of ovarian homogenate were saved for immunoprecipitation. Homogenates of other tissues were also prepared as described above. Non-specific binding was reduced by incubation of the extracts with normal rabbit serum at 4°C with constant agitation for 90 minutes. To remove any protein-A and Sepharose bead binding proteins before using ABL2, protein A-Sepharose beads were then added and incubated with the extracts at 4°C with constant agitation for 30 minutes. The beads were pelleted by a low-speed centrifugation and supernatant was collected. The clean ovarian extracts were incubated overnight with ABL2 at a final concentration of 0.37 mg/ml at 4°C with constant agitation. Fresh protein A-Sepharose beads were added and incubated with the ovarian extracts at 4°C for 90 minutes on the next day. Beads were pelleted by a low-speed centrifugation, and the ovarian extracts were discarded. Beads were rinsed three times for a total of 45 minutes at room temperature, and sample buffer [28] was added.
Mass Spectrometry
The ABL2 immunoprecipitate was excised from the silver stained gel and the sample was sent to W.M. Keck Foundation Biotechnology Resource Laboratory (Yale University, New Haven, CT) for MS/MS identification. The procedures used at the Keck Laboratory are available on the website of the facility . Briefly, in gel trypsin digestion was performed, and protein was eluted with 50% acetonitrile and 0.1% formic acid. The eluted sample was desalted and was then subjected to nanospray MS/MS to obtain amino acid sequences of the tryptic digest.
Egg Activation
To examine release of PAD from live oocytes using immunofluorescence microscopy, oocytes were activated by incubating them in hyaluronidase for 10–15 minutes. The concentration of hyaluronidase used (approximately 200–250 units) was higher and the length of exposure was longer than is normally used to remove cumulus cells. These conditions of hyaluronidase treatment resulted in activation of most of the oocytes.
To determine if PAD remains associated with the plasma membrane as a peripheral protein after its release from cortical granules, zona free unfertilized metaphase II oocytes were incubated in EBSS-Ca/Mg-H supplemented with 0.3% PVP for 15 min at 37°C, and oocytes were artificially activated with 2 μM ionomycin for two minutes at 37°C. Control oocytes were incubated with 0.1% of DMSO for two minutes at 37°C. Activated oocytes were transferred to fresh EBSS-H supplemented with 0.05% PVP droplets under light mineral oil and incubated for 15 minutes at 37°C. Oocytes were then incubated in high salt-containing solution for 2 minutes at room temperature with constant pipetting to remove exocytosed materials from the oocyte surface. Some control oocytes were treated as mentioned above.
In Vitro Embryo Culture
Zona intact 2-cell preimplantation embryos were collected as described above in the oocyte and embryo collection section. Embryos were cultured in 50 μl of M16 supplemented with 0.02% of gentamycin under mineral oil at 37°C in the incubator (5% CO2, 95% air) for three days. The amount of antibody added to the droplet on day one as indicated below: 5 μg for polyclonal rabbit IgG, 1:100 dilution for polyclonal guinea pig IgG, 5 μg for anti-alpha integrin antibody, 5 μg for the antibody ABL2, and 1:100 dilution for anti-ePAD antibody. In some experiments, no antibody was added to the droplets. The embryos were checked everyday and total percentage of embryos that reached the blastocyst stage was recorded for each experimental group on day three.
Confocal Scanning Laser and Epifluorescent Microscopy
All procedures for CSLM were carried out at room temperature under light mineral oil unless otherwise specified. All samples for LCA and ABL2 labeling were fixed with 4% paraformaldehyde in DPBS, pH 7.4 for 30 minutes and most samples for PAD labeling were fixed with 4% paraformaldehyde in PBS, pH 7.4 for 30 minutes. Following fixation, samples were washed in blocking solution for a total of 30 minutes and then permeabilized with 0.1% Triton X-100 in blocking solution for 5 minutes. All samples were labeled in labeling solution and each labeling incubation was followed by several washes in fresh labeling solution for a total of 30 minutes. For ABL2 labeling, samples were incubated with a 1:300 dilution (40 μg/ml) of ABL2 for 30 minutes followed by 30 minutes of incubation in goat anti-rabbit IgG conjugated to Alexa 488 with a 1:300 dilution (6.6 μg/ml). Control samples were incubated with a 1:1000 dilution (28.3 μg/ml) of preimmune rabbit IgG for 30 minutes followed by goat-anti-rabbit Alexa 488. For LCA labeling, samples were incubated with 10 μg/ml of biotinylated LCA for 30 minutes followed by 30 minutes of incubation in 5 μg/ml of Texas Red-streptavidin. Control samples were incubated with 10 μg/ml of LCA that had been preincubated with 100 mM α-methyl-mannopyranoside for 30 minutes followed by 30 minutes of incubation with 5 μg/ml of Texas Red-streptavidin. To double label oocytes or preimplantation embryos, samples were first incubated with ABL2 followed by the goat anti-rabbit IgG conjugated to Alexa 488 then incubated with LCA followed by Texas Red-streptavidin as described previously. For PAD labeling, fixed samples were treated with 10 mM citrate buffer for 15 minutes at 95°C, incubated with 2 M Tris-HCl, pH 7.4, for 15 minutes, and then permeabilized with 0.1% Triton X-100 in PBS for 10 minutes. Samples were blocked with 2% normal goat serum and 2% BSA in PBS for 60 minutes and incubated with 1.5 μg/ml of rabbit anti-PAD V overnight. On the following day, samples were incubated in goat anti-rabbit IgG conjugated to Alexa 488 with a 1:300 dilution (6.6 μg/ml) for three hours at room temperature. For LCA and PAD double labeling, samples already labeled with PAD antibody were incubated with 10 μg/ml of LCA for 30 minutes and followed by 30 minutes of incubation with 5 μg/ml of Texas Red-streptavidin on following day. Control samples, non-permeabilized or permeabilized, were incubated with goat anti-rabbit IgG conjugated to Alexa 488 or Texas Red-streptavidin only. All labeled samples were examined using a Zeiss LS 510 confocal scanning laser microscope the next day. Samples were entirely sectioned optically with a space interval determined according to the pinhole setting. For some samples, two-dimensional projections of z-stacks were generated.
To label live unfertilized, activated, or fertilized oocytes with anti-ePAD, samples were incubated at room temperature in M16 culture medium containing anti-ePAD (1:100) for 45 minutes, washed in M16, and incubated 45 minutes at room temperature in M16 containing anti-guinea pig IgG conjugated to Alexa 488 (1:100). Oocytes were then washed and immediately viewed with a Nikon inverted epifluorescence microscope.
For in vitro cultured embryos, live embryos that had been incubated in a primary antibody (ABl2, anti-ePAD, or anti-integrin) were washed in M16 then incubated in either goat anti-rabbit IgG conjugated to Alexa 488 with a 1:100 dilution (19.8 μg/ml) or goat anti-guinea pig IgG conjugated to FITC with a 1:100 dilution for 1 hour at room temperature. After washing, live samples were examined, and images were taken with a Zeiss epifluorescence microscope.
Gel Electrophoresis and Lectin Blotting
Protein samples were solubilized with reducing and denaturing Laemmli sample buffer [28] prior to electrophoresis. Samples and biotinylated standards were run in one-dimensional SDS-PAGE Doucet gels (4% stacking/7.5% separating) [29] at 70 V and 140 V respectively and separated proteins were blotted onto nictrocellulose at 100 V for 1 hour [30]. For protein identification by mass spectrometry, the gel was silver stained after electrophoresis as previously described [31]. For lectin blotting, blots were washed in Tris-buffer saline (TBS) for 15 minutes at room temperature and then blocked with 0.5% Tween-20 in Tris-buffer saline (TBT) for 1 hour at room temperature. 1–10 μg/ml of the appropriate biotinylated lectin in TBT was added to the blot for overnight incubation at 4°C with constant agitation. For each control blot, biotinylated lectin was preabsorbed with 100 mM of control sugar for 2 hour at room temperature prior to the overnight incubation. Blots were washed with TBT four times for 60 minutes on the following day and then incubated in a 1:20,000 dilution of HRP-streptavidin in TBT for 40 minutes at room temperature. For PAD immunoblotting, blots were first blocked with 5% nonfat dry milk in PBS with 0.05% Tween 20 (PBT) for 30 minutes at room temperature and then washed with fresh PBT for 15 minutes. Blots were incubated with a 1:4000 dilution of anti-ePAD guinea pig IgG in PBT overnight at 4°C with constant agitation. For controls, all blots were either incubated with a 1:4000 dilution of preimmune guinea pig IgG in PBT or in PBT without antibody added. On the following day, blots were washed for 15 minutes with PBT and incubated with 1:2000 dilution of goat anti-guinea pig IgG conjugated with peroxidase for 2 hours at room temperature. For both lectin and PAD blots, enhanced chemiluminescence (Amersham, Piscataway, NJ) was used to detect bands of interest and band images were captured using Kodak X-Omat autoradiographic films. The molecular weight of protein was calculated using biotinylated standards.
Statistical Analyses
The percentage of 2-cell preimplantation embryos reaching the blastocyst stage in the presence of different antibodies and the percentage of 2-cell preimplantation embryos reaching the blastocyst stage in the absence of any antibody (control) were analyzed statistically using a one-way analysis of variance (ANOVA) followed by Dunnet's post-hoc test when results of the ANOVA were significant. In both the ANOVA and Dunnet's test, results were considered significant when p ≤ 0.05.
Results
The ABL2 antibody recognizes a 75-kDa ovarian protein that is present in cortical granules of mouse oocytes
The ABL2 antibody precipitates a 75 kDa protein (p75) from mouse oocytes [20]. To determine if other tissues express p75, various mouse tissue extracts were used to perform ABL2 immunoprecipitation. p75 was immunoprecipitated from the ovary by ABL2 (Fig. 1A, lane 4), but not from brain, liver, muscle, oviduct, or testis (Fig. 1A, lanes 1–3 and lanes 5–6). Both the ABL2 antibody (Figs. 1B, ABL2) and the lectin Lens culinaris agglutinin (LCA) (Fig. 1B, LCA) labeled granules in the cortex of oocytes. Many granules showed co-localization of the two probes in merged images (Fig. 1B, ABL2 / LCA), demonstrating p75 to be a mouse cortical granule protein. Co-localization of two probes was also observed in pre-translocated cortical granules located in the cytoplasm of germinal vesicle intact oocytes (Figs. 1 and 2 in [18]). Cryosections of mouse ovary did not show ABL2 labeling anywhere in the ovary except in the cortical granules (data not shown).
Figure 1 Tissue distribution of the ABL2 antigen. (A) Silver-stained SDS-PAGE gel loaded with the ABL2 immunoprecipitate from mouse brain (lane 1), liver (lane 2), skeletal muscle (lane 3), ovary (lane 4), oviduct (lane 5), and testis (lane 6). The ABL2 antibody immunoprecipitated a 75-kDa protein from the ovarian lysate but not from other tissues. Other bands in the gel are from the antibody used for immunoprecipitation. (B) Confocal scanning laser micrographs of germinal vesicle intact mouse oocytes double labeled with the lectin LCA (LCA) and the ABL2 antibody (ABL2). The merged image (LCA + ABL2) showed co-localization of LCA and ABL2 in some cortical granules. These images were digitally enlarged for better visualization. (C) Western blots in which ABL2 immunoprecipitate was probed with the lectins ConA, LCA, WGA, PNA, and DBA. Control blots were probed with lectins preabsorbed with the appropriate control sugar. Positive controls (blots with rabbit IgG) were included for each lectin to show that the blotting condition was optimized.
Figure 2 Identification of the ABL2 antigen using tandem mass spectrometry. (A) Silver-stained SDS-PAGE gel loaded with ABL2 immunoprecipitate from mouse ovarian lysate (lane A) and molecular weight standards (lane B). In this experiment, a protein with molecular weight of 65-kDa co-precipitated with p75. (B) MS/MS spectrum of a peptide obtained from trypsin-digested p75; the sequence of this peptide was determined to be VLIGGSFY. (C) Western blots of an ABL2 immunoprecipitate from a mouse ovarian lysate probed with guinea pig anti-ePAD IgG, preimmune guinea pig IgG, or goat anti-guinea pig IgG conjugated to peroxidase only.
Since cortical granule proteins are secreted and most secreted proteins are glycosylated, we performed lectin blotting on immunoprecipitates from ovarian lysates to determine if p75 is glycosylated [32,33]. Blots with p75 were probed with α-D-mannose-specific ConA and LCA, N-acetylglucosamine-specific WGA, β-D-galactose-specific PNA, and N-acetylgalactosamine-specific DBA. None of these lectins bound to p75 on the blots (Fig. 1C, p75 + lectins), indicating that p75 is probably not glycosylated. Blots with rabbit IgG were used as a positive control to optimize the blotting condition for each lectin and to demonstrate that the assay was working (Fig. 1C, positive control). Control blots probed with lectins preabsorbed with the appropriate sugar under the same blotting conditions did not show binding to rabbit IgG (Fig. 1C, sugar controls), demonstrating the specificity of each lectin.
Identification of p75 using mass spectrometry
To identify p75, the protein was immunoprecipitated from ovarian lysates with the ABL2 antibody and analyzed using mass spectrometry. Generally immunoprecipitation yields a single band of 75 kDa; however, occasionally a second band of 65 kDa is also obtained as shown in Figure 2A. High-energy collision-induced dissociation (CID) spectra of the trypsin-digested of peptides from each protein band was obtained, and partial amino acid sequences of the peptides were deduced. For the 65-kDa band, three peptide sequences were obtained (LVQEVTDFAK/APQVSTPTLVEARAR/LSQTFPNADFAEITK) from the spectra. When sequences were searched separately using BLAST against the NCBI nonredundant database, they all matched serum albumin precursor [GenBank:P07724]. For p75, a CID mass spectrum of the parent peptide ion (at m/z 1468.8+2) was obtained and used to deduce the amino acid sequence (Fig. 2B). The spectrum showed a series of peptide ions of decreasing mass generated from the parent peptide. The mass difference between each consecutive peptide ion was used to determine the parent peptide sequence, and a partial amino acid sequence, VLIGGSFY, was then obtained as shown in Figure 2B. The VLIGGSFY sequence matched several mouse peptidylarginine deiminases (PAD) when searched using BLAST against the NCBI nonredundant database. These included a putative mouse PAD type V-like protein [GenBank:XP_144067] predicted by NCBI automated gene predicting algorithm, an egg and embryo abundant PAD [GenBank:AH53724], and a recently characterized mouse oocyte protein, ePAD [GenBank:NP_694746]. Although the egg and embryo abundant PAD (AAH53724) and ePAD (NP_694746) are listed under different entries in the database, they may be the same since their protein sequences are identical except for three amino acids; however, we can not exclude the possibility that they are duplicated genes. In addition, Sonar MS/MS (Genomic Solutions), another software tool designed for mass spectrometric protein identification, was used to search the NCBI nonredundant database. Unlike most database search algorithms that perform protein identification based exclusively on amino acid sequence, Sonar MS/MS includes additional information such as the mass-to-charge (m/z) ratio of the original parent peptide ion to perform identification. This information becomes essential for validating positive protein identification when only a partial amino acid sequence can be obtained from the original parent peptide, as had been the case in this study. The result obtained using Sonar MS/MS showed that the sequence VLIGGSFY was matched to PADs, as had been demonstrated with the BLAST search. To confirm the MS/MS identification of p75, we used an antibody that was made against mouse ePAD [26] to probe blots of the ABL2 immunoprecipitate. The ePAD antibody detected a single protein band with a molecular weight of 75 kDa (Fig. 2C, ePAD). No bands were detected when preimmune IgG or goat anti-guinea pig IgG conjugated to peroxidase alone were used (Fig. 2C, PI and anti-guinea pig IgG). These results demonstrate that the p75 immunoprecipitated by ABL2 is indeed a PAD and confirm our MS/MS identification of p75.
Amino acid sequence comparison of different PADs
Using the MultiAlin program [34], we constructed protein sequence alignments of nine mammalian PAD proteins including all mouse PADs (five characterized mouse PADs: PAD I – IV and ePAD; two uncharacterized mouse PADs, rat PAD VI, and human PAD V [GenBank: NP_035189, NP_032838, NP_035190, AAH53724, XP_144067, NP_694746. XP_233601, NP_036519] (Fig. 3). Sequence residues that are in high consensus are shown in red and sequence residues that are in low consensus are shown in blue. Gaps (-) are introduced for optimal alignment. The multiple alignments of the nine mammalian PADs show that approximately 40% – 50% of the amino acid sequences in these PADs are identical, indicating strong homologies among members of this family. Two predictive algorithms (SignalP V2.0 and TargetP V1.0) [35-37] were used to determine that a putative signal peptide and a cleavage site exist in ePAD and AAH53724 (an egg and embryo abundant peptidylarginine deiminase), indicating they are likely secreted proteins (Fig. 3 arrow). Human PAD V has a monopartite nuclear localization sequence motif [25], and it is the only type of PAD that has been localized to the nuclei of cells (Fig. 3 underline). Only ePAD, AAH53724 (an egg and embryo abundant peptidylarginine deiminase), and XP_144067 (peptidylarginine deiminase type V-like protein) have residues that exactly match the VLIGGSFY sequence (Fig. 3 asterisks). Interestingly, rat PAD VI also has a sequence match to the peptide VLIGGSFY obtained from p75 MS/MS analysis except for the first residue V (Fig. 3 asterisks). PAD I is derived from a gene predictive program, and its sequence is 80% identical to that of ePAD or AAH53724 (an egg and embryo abundant peptidylarginine deiminase).
Figure 3 Multiple alignments of mammalian PAD protein sequences. The sequences were aligned using the program MultiAlin available at . The peptide sequence (VLIGGSFY) of p75 obtained from MS/MS analysis was searched against listed PADs and residues that were matched to it are marked (*). The signal peptide cleavage site is marked with an arrow. The monopartite nuclear localization sequence in human PAD V is underlined. High consensus sequences are in red (90% of amino acids are identical or have biochemically similar R-groups) and low consensus sequences are in blue (50% of amino acids are identical or have biochemically similar R-groups). The abbreviations of species are listed as followed: Mm = M. musculus; Rn = R. norvegicus; Hs = H. sapiens. Two putative mouse PAD sequences are referred with their accession numbers (GenBank/NCBI). Accession numbers (GenBank/NCBI) of other PADs are as followed: NP_694746; XP_233601; NP_035189; NP_035190; NP_036519; NP_032838.
Mouse cortical granules contain PAD
To ascertain if mouse cortical granules contain PAD, antibodies made against mouse ePAD and human recombinant PAD V (anti-PAD V (N)) were used to label in vivo matured germinal vesicle intact and metaphase II mouse oocytes. The ePAD antibody had been used previously [26] and showed strong labeling in the cortex. When we adjusted labeling conditions to optimize cortical labeling, both granular and cytoplasmic labeling were observed in the cortex with anti-ePAD; however, the high level of cytoplasmic labeling made it difficult to resolve individual granules and to demonstrate co-localization with LCA, a cortical granule binding lectin (not shown). Therefore the antibody to human PAD-V, which gave a cleaner signal in the cortex, was also used to localize PAD in cortical granules (Fig. 4).
Figure 4 Confocal scanning laser micrographs of (A) human blood cells and (B – D) in vivo matured mouse oocytes labeled with anti-PAD V (N), (E – H) in vivo matured mouse oocytes double labeled with anti-PAD V (N) and LCA, and (I – L) double labeled with ABL2 and LCA. All anti-PAD V labeling is shown in green, except in A where it is red. DNA stain in A is green. ABL2 labeling is green and LCA labeling is red in all figures. (A) Cytospin preparations of the granulocyte fraction were stained with anti-PAD V (N), and their nuclei were stained with SYTOX green nucleic acid stain. The merged image shows nuclear localization of PAD (yellow) in a human granulocyte. (B, C) Germinal vesicle intact mouse oocytes and metaphase II oocytes were labeled with anti-PAD V (N), (D) Metaphase II mouse oocyte did not show labeling with goat anti-rabbit IgG conjugated to Alexa 488 alone. (E, F) Polar sections of germinal vesicle intact mouse oocytes double labeled with LCA (red) and anti-PAD V (N) (green). These images were digitally enlarged 2× for better visualization. (G) Merged image of both LCA and anti-PAD V (N) showed co-localization (yellow) of labels. (H) Merged image of equatorial section of metaphase II mouse oocytes double labeled with anti-PAD V (N) and LCA showing co-localization. (I, J) Metaphase II oocytes double labeled with LCA (red) and ABL2 (green). (K) Merged image of both LCA and ABL2 showed co-localization (yellow). The inserts of I, J, and K showed the polar view of the oocyte. (L) Control oocytes were not labeled with LCA pre-absorbed with α-D-methyl-mannopyranoside. All samples were imaged at same magnification and the scale bar applies to all figures.
Human peripheral blood cells were first used as a positive control and to optimize labeling conditions with anti-human PAD-V. The antibody labeled only the granulocytes (neutrophils and eosinophils), and labeling was localized to the nuclei of the cells (Fig. 4A), as reported previously [25]. When germinal vesicle intact oocytes were then labeled, immunoreactivity was localized in the nucleus and also in granules in the cortex (Fig. 4B). In metaphase II oocytes, the antibody labeled granules in the cortex; except in the area of the cortical granule free domain which was devoid of PAD labeling (Fig. 4C). In the metaphase II oocytes, the nuclear envelope had broken down, and thus there was no nuclear staining; however, the cytoplasm of metaphase II oocytes was more intensely labeled than that of germinal vesicle intact oocytes, suggesting that nuclear PAD was now dispersed in the cytoplasm (Figs. 4B, C). These results demonstrate that PAD is present in the cortical granules, nucleus, and cytoplasm of unfertilized mouse oocytes. Control oocytes were not labeled with goat anti-rabbit IgG conjugated to Alexa 488 alone (Fig. 4D).
To confirm that anti-PAD V (N) is labeling cortical granules in the oocyte's cortex and that PAD is present in these granules, anti-PAD V (N) and LCA were used to double label germinal vesicle intact and metaphase II oocytes, and their labeling pattern was compared to that of ABL2 and LCA double labeled oocytes. Both anti-PAD V (N) and LCA labeled granules (arrow) in the cortex of germinal vesicle intact oocytes (Figs. 4E, F). When images of both probes were merged, many granules appeared orange or yellow indicating co-localization of these probes (Fig. 4G), and similar co-localization of granules was also observed when metaphase II oocytes were used (Fig. 4H). In the metaphase II oocytes, an area devoid of signal corresponding to the cortical granule free domain was observed (Fig. 4H), and this domain was not labeled by either anti-PAD V (N) or LCA. When ABL2 and LCA were used to double label metaphase II oocytes, both probes labeled the granules in the cortex and showed co-localization of granules (Figs. 4I–K), as had been observed with anti-PAD V (N) and LCA. Besides the granules in the cortex, anti-PAD V (N) also labeled cytoplasm near the cortical granules; however, this labeling is diffuse and less granular than the cortical granule labeling. This diffuse cytoplasmic labeling did not co-localize with LCA labeling (Figs. 4F, G, arrowhead). Control oocytes labeled with LCA pre-absorbed with α-D-methyl-mannopyranoside showed no labeling (Fig. 4L). Taken together, these results demonstrate that antibodies to PAD label cortical granules of mouse oocytes as had been observed with the ABL2 antibody and that PAD (ABL2 antigen, p75) is present in the cortical granules of mouse oocytes.
Localization of PAD (p75) after artificial activation and fertilization
To demonstrate that PAD is released from cortical granules when they undergo exocytosis, unfertilized, hyaluronidase activated, and in vivo fertilized oocytes were compared using immunofluorescence microscopy (Fig. 5). All oocytes were labeled live (non-permeabilized) with the primary and secondary antibody and were imaged using an inverted epifluorescent microscope to minimize damage to the living oocytes. Since only extracellular PAD was imaged in this experiment, anti-ePAD was used, and cortical cytoplasmic labeling did not interfere with interpretation of the images, as had occurred when oocytes were permeabilized and imaged with confocal microscopy (see previous section). Secondary antibody alone did not label unfertilized or fertilized oocytes (Figs. 5A–B, C–D). Unfertilized live oocytes did not show extracellular fluorescence when labeled with both anti-ePAD and the secondary antibody (Figs 5E–F), Oocytes caught in various stages of activation showed distinct patterns of extracellular labeling with anti-ePAD (Figs 5G–J). In early stages of activation, numerous extracellular granules were labeled in the perivitelline space (Figs. 5H–I). Many of these granules were the size of cortical granules suggesting they were recently exocytosed (Fig 5H). Other granules had begun to disperse and were larger in diameter (Fig 5I). At later times after activation, granular content had dispersed completely within the perivitelline space, and some labeling appeared associated with the oolemma (Fig 5J). Similar to activated oocytes, fertilized oocytes that were recovered from oviducts of mated females had labeled granules in the perivitelline space (Fig 5K). At later stages, the contents of the granules had dispersed to fill the perivitelline space (Fig. 5L).
Figure 5 Live, non-permeabilized activated and in vivo fertilized oocytes showing release of PAD from the cortical granules. A and B are the same unfertilized oocyte viewed with Hoffman optics (A) or epifluorescence microscopy (B) after labeling with the secondary antibody only. No non-specific labeling is observed in B. C and D are the same fertilized oocyte viewed with Hoffman optics (C) or fluorescence microscopy (D) after labeling with secondary antibody only. No non-specific labeling is observed after fertilization. E and F are the same unfertilized oocyte viewed with Hoffman microscopy (E) or epifluorescence microscopy (F) after labeling with both the anti-ePAD and the secondary antibody. No labeling is observed around the unfertilized oocyte (F). G, H, and I are the same early activated oocyte viewed with Hoffman optics (G) or after labeling with both anti-ePAD and secondary antibody (H, I). H is focused close to the surface of the oocyte and shows numerous small labeled granules in the perivitelline space. I is focused near the equator of the oocyte and shows larger dispersing granules and diffuse label in the perivitelline space. J shows a different oocyte at a later state of activation. Label has adhered to the surface of the oocyte and the polar body (arrow). Diffuse label is present in the perivitelline space. K and L are double labeled fertilized oocytes recovered from oviducts of naturally mated females. K shows an earlier stage after fertilization in which some label in the perivitelline space is still granular. L shows a later stage after fertilization in which label is diffuse in the perivitelline space. In K, a sperm tail in the perivitelline space has apparently absorbed PAD (arrow).
Localization of PAD versus other cortical granule components during preimplantation development
To follow the fate of secreted PAD during preimplantation development and to compare the fate of secreted PAD to glycosylated cortical granule components, fixed in vivo fertilized oocytes and in vivo matured preimplantation embryos were double labeled with the ABL2 antibody and LCA (Fig. 6). LCA would be expected to localize glycosylated cortical granule components, while ABL2 would localize secreted PAD. ABL2 was used to localize PAD in this experiment since anti-ePAD images were difficult to interpret using fixed permeabilized samples and since the method used for anti-PAD V labeling removed the zona which precluded tracing secreted material into the perivitelline space or zona.
Figure 6 Confocal scanning laser micrographs comparing distribution of PAD and LCA-binding cortical granule components in in vivo fertilized oocytes and in vivo matured 2-cell embryos. Optical sections (A – C and G – I) and two-dimensional projections of z-series (D – F and J – L) of zona intact fertilized oocytes (A – F) and zona intact 2-cell embryos (G – L) labeled with LCA (red) and ABL2 antibody (green). A, D, G, and J show the LCA labeling on the surface of the oocyte (white arrowhead), in the perivitelline space (arrow), and in the zona pellucida (yellow arrowhead). B, E, H, and K show the ABL2 labeling. C, F, I, and L are merged confocal images and two-dimensional projections showing both LCA and ABL2 labeling.
After fertilization, the released LCA-binding cortical granule components were present mainly on the surface of oocytes (Fig. 6A white arrowhead; Fig. 6D) and in the zona pellucida (Fig. 6A yellow arrowhead; Fig. 6D). Less intense LCA labeling was observed in the perivitelline space (Fig. 6A arrow; Fig. 6D). In contrast, ABL2 labeling was detected on the surface of fertilized oocytes (Fig. 6B, E) and in the perivitelline space (Fig. 6E), but not in the zona pellucida of the fertilized oocytes (Fig. 6B, E), in agreement with Figure 4. Fixed non-permeablized fertilized oocytes showed the same LCA and ABL2 labeling patterns (data not shown).
In merged images of double-labeled fertilized oocytes, much of the LCA and ABL2 labeling on the oocyte's surface was co-localized (Fig. 5C). In two-dimensional projections of z-series of the double labeled fertilized oocytes that contained pronuclei and two polar bodies, the LCA and ABL2 labeling present on the surface of the oocytes was granular, and hence similar in appearance to the cortical granules of unfertilized oocytes (Figs. 6D–F). Many of these extracellular granules were larger than the cortical granules in unfertilized oocytes, indicating they had dispersed slightly by this stage, as seen previously in unfixed activated and fertilized oocytes (Figs. 5H–J).
At the 2-cell stage, some of the LCA-binding cortical granule components (Fig. 6G white arrowhead) and all of the ABL2 antigen (Fig. 6H) remained associated with the blastomeres' plasma membranes with some labeling observed between the blastomeres (Figs. 6G–I). Two-dimensional projections of series of z-stacks revealed that LCA and ABL2 labeling on the blastomeres' surfaces was diffuse, not granular, at this stage (Figs. 6J, K, L). In merged images, LCA and ABL2 labeling showed less co-localization on the 2-cell preimplantation embryos' surface than on the fertilized oocytes' surface (compare Fig. 6C and Fig. 6I). Unlike ABL2, LCA-binding cortical granule components were evenly dispersed in the perivitelline space (Figs. 6G arrow, I, J, L) and in the zona pellucida (Figs. 6G yellow arrowhead; I, J, L).
At the 8-cell stage, LCA-binding components were found in the zona pellucida (Figs. 7A, C, D, F), but not in the perivitelline space or on the blastomeres' plasma membranes (Figs. 7A, C). ABL2 labeling was still associated only with the blastomeres' plasma membranes, where it appeared diffuse, not granular (Figs. 7B, C, E). Lack of co-localization of LCA and ABL2 on the blastomeres surface at this time supports our data (Fig. 1C) that the ABL2 antigen (PAD) is not glycosylated. At the 8 cell stage, ABL2 also labeled the subcortical region of blastomeres (Fig. 6B arrow), consistent with previous reports at this stage [15,38].
Figure 7 Confocal scanning laser micrographs comparing distribution of PAD and LCA-binding cortical granule components in in vivo matured 8-cell embryos and blastocysts. Optical sections (A – C and G – I) and two-dimensional projections of z-series (D – F and J – L) of zona intact 8-cell embryos (A – F) and blastocysts (G – L) labeled with LCA (red) and ABL2 (green). A and D show the LCA labeling in the zona pellucida (arrowhead). B and E show ABL2 labeling on the plasma membranes and in the subcortical region (arrow) of blastomeres. G and J show LCA labeling of the trophoblast cells and the inner cell mass cells. H and K show ABL2 labeling of the trophoblast cells. C, F, I, and L are merged confocal images and two-dimensional projections showing both LCA and ABL2 labeling.
At the early blastocyst stage, LCA labeling was present on the surface of the trophoblast and the inner cell mass cells (Figs 7G, I, J, L), while ABL2 labeling was found only on the surface of the trophoblast cells (Figs. 7H, I, K, L). The LCA labeling on the trophoblast was patchy (Figs. 7G, J), whereas the ABL2 labeling was evenly dispersed (Figs. 7H, K). It is probable that the LCA staining observed at this stage did not exclusively represent LCA-binding cortical granule proteins but rather newly synthesized surface proteins. Neither LCA nor ABL2 labeling was detected in the perivitelline space or in the zona pellucida (Figs. 7G–L). Control fertilized oocytes and preimplantation embryos were not labeled by preimmune IgG, LCA pretreated with α-methyl-mannopyranoside, or Texas Red-streptavidin alone (data not shown).
The previous experiment demonstrated that secreted PAD was in the perivitelline space and on the oolemma immediately after fertilization, but by the 2 cell stage was found only on the oolemma. In contrast, other cortical granule components that label with LCA passed into the perivitelline space and zona pellucida immediately after fertilization and were not found on the blastomeres' plasma membranes by the 8 cell stage. To confirm the localization of PAD on plasma membranes, in vivo matured preimplantation embryos were also labeled with anti-PAD V. In addition to labeling secreted PAD, anti-PAD V would also be expected to label nuclear and cytoplasmic forms of PAD as shown in Figure 4. Anti-PAD V was used in this experiment since it gave cleaner, more interpretable images, especially near the surface of oocytes, than anti-ePAD. Anti-PAD V (N) labeled the surface of fertilized oocytes (Figs. 8A–C arrowheads). The zona and hence the perivitelline space were lost from these oocytes during immunolabeling, therefore localization in the perivitelline space could not be confirmed with this antibody. After fertilization, PAD labeling was evenly distributed and continuous on the surface of oocytes, in contrast to unfertilized oocytes in which an area devoid of PAD labeling (cortical granule free domain) was observed above the spindle (Fig. 4C). After fertilization, both the pronuclei (arrow) and the cytoplasm were also labeled which would be expected for the anti-PAD V (N) antibody (Figs. 8A–C).
Figure 8 Confocal scanning laser micrographs showing distribution of PAD in in vivo fertilized oocytes and in vivo matured pre-implantation embryos labeled with anti-PAD V (N). (A – C) Three different focal sections of a zona free fertilized oocyte showing PAD labeling on the oocyte's surface (arrowhead) and in both pronuclei (arrows). (D, E) A 2-cell embryo and an 8-cell embryo showed PAD labeling on the blastomeres' surfaces (arrowhead) and in their nuclei (arrows). (F) A blastocyst showing that anti-PAD V (N) labeled the trophoblast cells, and inner cell mass cells, and nuclei of inner cell mass cells (arrow).
At the 2-cell stage, PAD labeling still remained on the blastomeres' plasma membranes, and some immunoactivity was found between the two blastomeres (Fig. 8D arrowhead) as had been observed with the ABL2 antibody (Fig. 6H). In addition, both the nuclei (Fig. 8D arrow) and to a lesser extent the cytoplasm of two blastomeres were stained by anti-PAD V (N). At the 8-cell stage, anti-PAD V (N) labeling was still associated with the blastomeres' plasma membranes, and the label was diffuse around the blastomeres surface (Fig. 7E, arrowhead). In addition, PAD labeling was also found deeper in the cortical cytoplasm (Fig. 8E), as was observed with the ABl2 antibody at the 8 cell stage (Fig. 7H) [39]. Anti-PAD V (N) also strongly labeled the nuclei (arrow) and weakly labeled the cytoplasm of each blastomere at the 8 cell stage (Fig. 8E). Finally, in blastocysts, anti-PAD V (N) labeled the cytoplasm of both trophoblast and inner cell mass cells with equal intensity (Fig. 8F). This antibody also labeled nuclei of inner cell mass cells (Fig. 8F arrow) and trophoblast cells (not shown in this focal section) at this stage. It was not possible to determine if the surface of the blastocyst was labeled by anti-PAD-V at this stage due to cytoplasmic labeling that extended to the periphery of the cells.
Figures 5, 6, 7, 8, above demonstrate that PAD is released from cortical granules following fertilization and that at least some PAD remains associated with the oocyte and blastomeres' surfaces during preimplantation development. Since PAD was observed in the perivitelline space of living activated and fertilized oocytes, it is possible that some of this secreted PAD binds back to the oolemma as a peripheral membrane protein. Alternatively, the PAD associated with the oolemma may represent a different isoform of PAD that is in fact an integral membrane protein. To distinguish between these two possibilities, we treated artificially activated oocytes with high salt-containing solution. If PAD is peripherally associated with the oolemma following exocytosis, high salt-containing solution should remove it from the oocyte's surface. If PAD is an integral membrane protein, it should remain on the surface following this treatment. Both anti-PAD V (N) and ABL2 labeled the surface of artificially activated oocytes (Figs 9A, E), and the labeling was removed from the surface when activated oocytes were treated with high salt-containing solution (Figs. 9B, F). Control non-activated oocytes showed PAD and ABL2 labeling in the cortical granules (Figs. 9C, G), and treatment of non-activated oocytes with high salt-containing solution did not modify this labeling (Figs. 9D, H). The control was conducted to ensure that the PAD and ABL2 labeling were removed from the activated oocyte's surface by high salt treatment and not by DMSO that was used to dissolve ionomycin for artificial activation. These results show that PAD on the oolemma behaves as a peripheral membrane protein after it is released from cortical granules by artificial activation.
Figure 9 Confocal scanning laser micrographs of in vivo matured metaphase II oocytes labeled with either anti-PAD V (N) (A – D) or ABL2 (E – H). showing that PAD is a peripheral membrane protein. A and E are artificially activated oocytes showing released PAD and ABL2 antigen (p75) on the oocyte's surface. B and F are artificially activated oocytes treated with a high salt solution that removed PAD and p75 from oocytes' surfaces. C and G are DMSO control oocytes showing PAD and ABL2 antigen in cortical granules. D and H are DMSO control oocytes treated with a high salt solution confirming that PAD and ABL2 antigen were still detected in cortical granules.
Role of PAD in preimplantation embryonic development
Previously, the ABL2 antibody was shown to inhibit hamster and mouse preimplantation embryonic development [16]. To determine if antibodies to PAD show a similar inhibitory effect on mouse preimplantation development, in vivo matured zona intact 2-cell embryos were incubated in vitro in the presence of PAD or control antibodies, and the number of blastocysts was counted on day 3 (Fig. 10A). When control embryos were cultured in the absence of any antibodies, most embryos (90%) showed normal development to the blastocyst stage. Both ABL2 and anti-ePAD inhibited development of 2-cell embryos. In the presence of ABL2 and ePAD antibodies, only 60% or 55% respectively of 2-cell embryos reached the blastocyst stage. Higher concentrations of anti-ePAD produced significantly greater inhibition of development with only 22% of the 2 cells stage becoming blastocysts (not shown). In the ABL2 and anti-ePAD treatment groups, most embryos that did not develop to the blastocyst stage were either in the 8-cell or the morula stage on day 3 (data not shown), indicating that cleavage divisions were inhibited in the presence of ABL2 and PAD antibodies. In vitro treatment of 2-cell embryos with preimmune rabbit IgG, preimmune guinea pig IgG, and function-blocking rabbit anti-β1 integrin IgG did not significantly affect embryo development (Fig. 10A) demonstrating that the results seen with the ABL2 and PAD antibodies were specific and not simply due to IgG binding to the cell surface. To show that the antibodies to PAD and β1 integrin bound to the blastomeres' surfaces, live 8-cell embryos treated with each antibody were subsequently incubated with anti-rabbit or anti-guinea pig IgG conjugated to FITC. Epifluorescent micrographs showed that all of these antibodies bound to the blastomeres' surfaces (Fig. 10B). These results demonstrate that cleavage divisions and blastocyst formation were significantly inhibited by a PAD specific antibody.
Figure 10 (A) Effects of ePAD antibody and ABL2 antibody on preimplantation development. A shows the percentage of blastocyst formation in the presence of different antibodies. The number of experiments and the total number of oocytes are shown in the figure for each experimental group. (B) Epifluorescent micrographs of in vitro matured live 8-cell embryos treated with ABL2 antibody, anti-ePAD, and β1 integrin antibody followed by the appropriate secondary antibody. In each case label is present on the blastomeres' surfaces. (***) P < 0.01.
Discussion
In the present study, a mouse cortical granule protein, p75, was immunoprecipitated from ovarian lysate, microsequenced by tandem mass spectrometry, and identified as peptidylarginine deiminase (PAD) using two independent software tools (BLAST and Sonar MS/MS). PAD was secreted from cortical granules following artificial activation or fertilization. Secreted PAD was present in the perivitelline space and on the oolemma of freshly activated or fertilized oocytes. Unlike LCA-binding cortical components which diffused into the zona pellucida, PAD remained attached to the plasma membrane of blastomeres at later times in preimplantation development. PAD on the plasma membrane of activated oocytes could be removed by high salt treatment indicating that it was a peripheral membrane protein. PAD appears to be a non-glycosylated secretory protein as it did not bind any tested lectin in blots and it did not bind LCA in confocal sections of 8 cell embryos. In vitro experiments with antibodies to PAD suggest that cortical granule PAD plays a role, after its release at fertilization, in cleavage and early development.
Various lines of evidence support the conclusion that PAD is a cortical granule protein equivalent to p75, the antigen immunoprecipitated by the ABL2 antibody. First, a PAD specific antibody (anti-ePAD) recognized p75 on Western blots. Secondly, oocyte PAD and p75 immunoprecipitated with the ABL2 antibody from mouse oocytes have the same molecular weights (75 kDa) and similar isoelectric points (pI) on 2-dimensional gels (5 to 5.5 for PAD and 4.9 to 5.3 for p75) [20,26]. Following the cortical reaction, both p75 and PAD remained associated with the plasma membrane during early preimplantation embryonic development. In the embryo culture experiments, antibodies to both p75 and PAD inhibited cleavage and preimplantation development. Finally, PAD and the lectin LCA were co-localized in some, but not all, cortical granules in agreement with our earlier observation using the ABL2 antibody [18]. When taken together, the above evidence supports the conclusion that p75 is PAD, which is localized in mouse cortical granules.
Cortical granule PAD is the first member of the PAD family that has been reported to be secreted. While most secreted proteins are glycosylated, our lectin blots suggest that the cortical granule PAD is not, a conclusion also supported by the observation that the molecular weight of cortical granule PAD immunoprecipitated from mouse ovaries (75 kDa) is similar to the molecular weight of PAD computed from its amino acid sequence (76.7 kDa). Moreover, LCA and ABl2 did not co-localize at the 8 cell stage of preimplantation development, further indicating that LCA does not bind to PAD. Several other non-glycosylated proteins, such as chemokines, albumin, and transcobalamin II, are also secreted [40-42].
PADs are a family of calcium dependent enzymes that catalyze the conversion of arginine into citrulline in proteins. In the mammalian PAD family, approximately 50% of the amino acids are identical among different isoforms within one species, and 70% to 95% of the amino acids are identical among the same isoforms in different mammals [24]. Five isoforms of PAD (PAD I, PAD II, PAD III, PAD IV, and ePAD) have been cloned and sequenced in mice. PAD II is found in various tissues including skeletal muscle, uterus, spinal cord, salivary glands, and pancreas [43]. PAD I and PAD III are expressed in epidermis and hair follicles (PAD III) [43]. Mouse PAD IV has a potential nuclear localization sequence and is likely to be present in nuclei. ePAD has been localized in mammalian oocytes and embryos [26]. The isoform of PAD immunoprecipitated by ABL2, which is the isoform secreted from cortical granules, was only found in ovarian tissue.
PAD makes up about 1% of the total protein in mouse oocytes [26], and current data indicate that oocytes contain multiple isoforms of PAD. Immunohistological data obtained with three different PAD antibodies (anti-ePAD, anti-PAD-V, and ABl2) suggest mouse oocytes contain at least four isoforms of this protein. These isoforms are located in the nucleus (anti-PAD V), non-cortical cytoplasm (anti-PAD V), cortical cytoplasm (anti-ePAD) (current study and 27), and the cortical granules (anti-ePAD, anti-PAD-V, and ABl2). In an earlier study on p75 (now identified as PAD), whole oocyte extracts subjected to 2-dimensional gel electrophoresis revealed four species of p75 with pIs of 4.9 to 5.3 [20]. Likewise, a train of proteins designated PAD with pIs ranging from 5 to 5.5 was also observed in mouse oocytes [26]. These observations support the conclusion that mouse oocytes have at least four isoforms of PAD, which are localized in the cortical granules, cytoplasm (cortical and non-cortical), and nucleus of germinal vesicle intact oocytes.
The isoform found in the cortical granules is likely ePAD (or the egg and embryo abundant PAD, AAH53724) which is the only isoform predicted to be a secreted protein with a signal sequence by both the hidden Markov model (SignalP) and neural networks (TargetP) algorithms. ePAD was also predicted to be a non-transmembrane protein (data not shown) using TMHMM software, [44,45], which would be consistent with our observation that cortical granule PAD behaves as a peripheral membrane protein following exocytosis. While, several egg proteins without signal sequences have been identified on the extracellular surface of mouse oocytes [46], it is improbable that any of the PAD isoforms without signal sequences is the secreted isoform based on pI data. Of the four species of p75 (PAD) with pIs of 4.9 to 5.3 in mouse oocytes [20], only the one with a pI of 5.3 was released and detected in cortical granule exudates (Fig. 8D in [20]). Of the various PADs that could be present in oocytes, ePAD has a pI (5.36) which is most similar to the pI (5.3) of the secreted form of p75. From the above evidence, ePAD appears to be the best candidate for the cortical granule PAD.
It is probable that the nuclear PAD observed in germinal vesicle intact oocytes and preimplantation embryos is mouse PAD IV, which has a classic monopartite nuclear localization sequence motif (PPVKK_ST, Fig. 3 underline) in the same region as human PAD V [25]. Since human PAD V and mouse PAD IV are 70% identical in their amino acid sequence (Fig. 3), it is not surprising that polyclonal antibodies made against human PAD V react with mouse PAD IV immunocytochemically. Interestingly, the cytoplasmic immunoreactivity of the PAD antibody observed in the metaphase II oocytes was brighter than of that in the germinal vesicle intact oocytes, suggesting that the nuclear PAD IV became redistributed to the cytoplasm following germinal vesicle breakdown.
The cytoplasmic PAD that we observed in mouse oocytes is most likely PAD I, PAD II, PAD III, and/or the mouse PAD type V-like protein (XP_144067) since these isoforms are predicted to have neither a signal sequence nor nuclear localization sequence. Of these four isoforms, only the mouse PAD type V-like protein (XP_144067) has the VLIGGSFY sequence. ePAD was previously interpreted to be in sheets of intermediate filaments based on immunoelectron microscopic data using the ePAD antibody [26]. However, the ePAD antibody reacted in 2D gel electrophoresis with all isoforms of PAD and stained the nuclei, cortex, and interior cytoplasm of germinal vesicle intact oocytes [26], indicating that the antibody reacts with more than one isoform of PAD, as occurred with the anti-PAD V antibody in our study. Future studies will be necessary to fully identify and characterize the function of each of the PAD isoforms in oocytes.
Our data show that, following exocytosis, some cortical granule PAD remains associated with the plasma membranes as a peripheral protein. Interestingly, the ABL2 antibody recognizes a pair of mouse embryonic glycoproteins with approximate molecular weights of 65 and 70 kDa, that were localized to the cortical cytoplasm of preimplantation embryos at both the light and electron microscopic levels [15,38,39]. It is probable that the p65/p70 found in preimplantation embryos are different from the p75 (PAD) found in oocytes for several reasons. First p75 (PAD) is not glycosylated [20], while p65/70 are glycosylated [15]. Secondly, the molecular weights and the pIs of the two embryonic glycoproteins (65 to 70 kDa/pI = 6 to 7) and the oocyte protein p75 (75 kDa/pI = 4.9 to 5.3) are different [20,38]. Lastly, p65/70 are synthesized in a stage specific manner between the 2-cell and morula stages, while the synthesis of p75 (PAD) is detected in the early stages of oogenesis and increases during oocyte growth [38,39,47]. The above data support the conclusion that p75 and p65/70 are not identical. P75 (PAD) is present in the cortical granules of unfertilized oocytes, while p65/70 appear to be cytoplasmic proteins found in preimplantation embryos. It is possible p65/70 are cytoplasmic isoforms of PAD or that both the ABL2 and PAD antibodies cross react with another type of protein that shares an epitope (s) with the PAD family. In either case, the exact identity of p65/70 remains to be determined.
Most known substrates of PADs are either intermediate filaments or filament-associated proteins that have structural function. Recently, H3 and H4 histones were shown to be substrates of human PAD IV (equivalent to human PAD V) which is involved in regulating histone arginine methylation by converting methyl-arginine to citrulline [48]. In our study, cortical granule PAD appeared to play a role in early embryogenesis since PAD antibodies, but not control antibodies or a function blocking antibody to another oolemma protein (anti-β1 integrin), inhibited embryonic development. This finding agrees with two prior studies that documented the inhibitory effect of the ABL2 antibody on preimplantation development in mice and hamsters [16,20] and identifies a new potential function for PAD. Since PAD is a calcium dependent protein, it is possible that cortical granule PAD is activated by extracellular calcium when it is exocytosed at fertilization. The activated PAD could then citrullinate and consequently activate an extracellular mitogen(s) or membrane protein(s) that is involved in regulating early embryogenesis [49-52]. Alternatively, it is possible that the cortical granule PAD affects early embryogenesis by providing a microenvironment to protect the developing preimplantation embryos. For example, as a consequence of citrullination, the PAD target protein(s) may unfold, as occurs with PAD III, and thereby be rendered ready for cross-linking, which could form a more protective extracellular matrix in perivitelline space.
Conclusion
Our study demonstrates that mouse oocytes contain multiple isoforms of PAD that are present in the nucleus, cortical cytoplasm, non-cortical cytoplasm, and cortical granules prior to fertilization. One isoform of PAD is released from the cortical granules at fertilization, and after its release, it remains associated with the zygote's and blastomeres' plasma membranes where it appears to play a role in preimplantation development.
Authors' contributions
ML performed most the experiments and prepared the manuscript. AO and PT performed the experiments in plate 5. PC, MY, and SC provided the antibodies (ABL2 antibody, PAD V (N) antibody, and ePAD antibody, respectively) and critically reviewed the manuscript. PT supervised all the work and assisted in writing the manuscript.
Acknowledgements
We would like to thank Paul Jaegu Kim for his invaluable technical assistance. We would also like to extend our gratitude to Yuhuan Wang and Zhen Wu for their assistance in obtaining Figures 3 and 9, and Norton Kitagawa for his invaluable advice in tandem mass spectrometric protein identification. Finally, we would like to thank Dr. Bradley Hyman and Dr. Leah Haimo for critical reading of the manuscript. This study was supported in part by a grant from NIH and by grants from the Academic Senate of UC, Riverside.
==== Refs
Austin CR The cortical granules of hamster eggs Expt Cell Res 1956 10 533 540 10.1016/0014-4827(56)90025-8
Cran DG Esper CR Cortical granules and the cortical reaction in mammals J Reprod Fertil Suppl 1990 42 177 188 2077122
Cran DG Cheng WT-K Changes in cortical granules during porcine oocyte maturation Gamete Res 1985 11 311 319 10.1002/mrd.1120110311
Gwatkin RBL Williams DT Hartmann JF Kniazuk M The zona reaction of hamster and mouse eggs:production in vitro by a trypsin-like protease from cortical granules J Reprod Fertil 1973 32 259 265 4734596
Wolf DP Hamada M Induction of zonal and egg plasma membrane blocks to sperm penetration in mouse eggs with cortical granule exudate Biol Reprod 1977 17 350 354 901888
Schuel H Secretory functions of egg cortical granules in fertilization and development Gamete Res 1978 1 299 382 10.1002/mrd.1120010311
Gulyas BJ Schmell ED Ovoperoxidase activity in ionophore treated mouse eggs. I. Electron microscopic localization Gam Res 1980 3 267 277 10.1002/mrd.1120030309
Schmell ED Gulyas BJ Ovoperoxidase activity in ionophore treated mouse eggs. II. Evidence for the enzyme's role in hardening the zona pellucida Gam Res 1980 3 279 290 10.1002/mrd.1120030310
Cherr GN Lambert H Meizel S Katz DF In vitro studies of the golden hamster sperm acrosome reaction: completion on the zona pellucida and induction by homologous soluble zonae pellucidae Dev Biol 1986 114 119 131 3514315 10.1016/0012-1606(86)90388-X
Moller CC Wassarman PM Characterization of a proteinase that cleaves zona pellucida glycoprotein ZP2 following activation of mouse eggs Dev Biol 1989 132 103 112 2492959 10.1016/0012-1606(89)90209-1
Tawia SA Lopata A The fertilization and development of mouse oocytes following cortical granule discharge in the presence of a protease inhibitor Hum Reprod 1992 7 1004 1009 1430117
Zhang X Rutledge J Khamsi F Armstrong DT Release of tissue-type plasminogen activator by activated rat eggs and its possible role in the zona reaction Mol Reprod Dev 1992 32 28 32 1515147 10.1002/mrd.1080320106
Miller DJ Gong X Decker G Shur BD Egg cortical granule N-acetylglucosaminidase is required for the mouse zona block to polyspermy J Cell Biol 1993 123 1431 1440 8253842 10.1083/jcb.123.6.1431
Barros C Yanagimachi R Induction of zona reaction in golden hamster eggs by cortical granule material Nature 1971 233 268 269 4938367 10.1038/233268a0
Johnson LV Calarco PG Immunological characterization of embryonic cell surface antigens recognized by antiblastocyst serum Dev Biol 1980 79 208 223 6997116 10.1016/0012-1606(80)90084-6
Hoodbhoy T Dandekar P Calarco P Talbot P p62/p56 are cortical granule proteins that contribute to formation of the cortical granule envelope and play a role in mammalian preimplantation development Mol Reprod Dev 2001 59 78 89 11335949 10.1002/mrd.1009
Hoodbhoy T Carroll EJ Talbot P Relationship Between p62 and p56, Two Proteins of the Mammalian Cortical Granule Envelope, and Hyalin, the Major Component of the Echinoderm Hyaline Layer Biol Reprod 2000 62 979 987 10727267
Liu M Sims D Calarco P Talbot P Biochemical heterogeneity, migration, and pre-fertilization release of mouse oocyte cortical granules Reprod Biol Endocrinol 2003 1
Gross VS Wessel G Florman HM Ducibella T A Monoclonal Antibody That Recognizes Mammalian Cortical Granules and a 32-Kilodalton Protein in Mouse Eggs Biol Reprod 2000 63 575 581 10906067
Pierce KE Siebert MC Kopf GS Schultz RM Calarco PG Characterization and localization of a mouse egg cortical granule antigen prior to and following fertilization or egg activation Dev Biol 1990 141 381 392 1698670 10.1016/0012-1606(90)90392-V
Hoodbhoy T Talbot P Mammalian cortical granules: contents, fate, and function Mol Reprod Dev 1994 39 439 448 7893493 10.1002/mrd.1080390413
Munoz-Gotera RJ Hernandez-Gonzalez EO Mendoza-Hernandez G Contreras RG Mujica A Exocytosis of a 60 kDa protein (calreticulin) from activated hamster oocytes Mol Reprod Dev 2001 60 405 413 11599052 10.1002/mrd.1103
Tutuncu L Stein P Ord TS Jorgez CJ Williams CJ Calreticulin on the mouse egg surface mediates transmembrane signaling linked to cell cycle resumption Dev Biol 2004 270 246 260 15136153 10.1016/j.ydbio.2004.02.008
Vossenaar ER Zendman AJ van Venrooij WJ Pruijn GJ PAD, a growing family of citrullinating enzymes: genes, features and involvement in disease Bioessays 2003 25 1106 1118 14579251 10.1002/bies.10357
Nakashima K Hagiwara T Yamada M Nuclear localization of peptidylarginine deiminase V and histone deimination in granulocytes J Biol Chem 2002 277 49562 49568 12393868 10.1074/jbc.M208795200
Wright PW Bolling LC Calvert ME Sarmento OF Berkeley EV Shea MC Hao Z Jayes FC Bush LA Shetty J Shore AN Reddi PP Tung KS Samy E Allietta MM Sherman NE Herr JC Coonrod SA ePAD, an oocyte and early embryo-abundant peptidylarginine deiminase-like protein that localizes to egg cytoplasmic sheets Dev Biol 2003 256 73 88 12654293 10.1016/S0012-1606(02)00126-4
Knoll M Talbot P Cigarette smoke inhibits oocyte cumulus complex pick-up by the oviduct independent of ciliary beat frequency Reproductive Toxicology 1998 12 57 68 9431573 10.1016/S0890-6238(97)00100-7
Laemmli UK Cleavage of structural proteins during the assembly of the head of bacteriophage T4 Nature 1970 227 680 685 5432063 10.1038/227680a0
Doucet JP Trifaro JM A discontinuous and highly porous sodium dodecyl sulfate-polyacrylamide slab gel system of high resolution Anal Biochem 1988 168 265 271 3364726 10.1016/0003-2697(88)90317-X
Towbin H Staehelin T Gordon J Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheets: procedure and some applications Proc Natl Acad Sci U S A 1979 76 4350 4354 388439
O'Connell KL Stults JT Identification of mouse liver proteins on two-dimensional electrophoresis gels by matrix-assisted laser desorption/ionization mass spectrometry of in situ enzymatic digests Electrophoresis 1997 18 349 359 9150913 10.1002/elps.1150180309
Roth J Protein N-glycosylation along the secretory pathway: relationship to organelle topography and function, protein quality control, and cell interactions Chem Rev 2002 102 285 303 11841244 10.1021/cr000423j
Dempski RE JrImperiali B Oligosaccharyl transferase: gatekeeper to the secretory pathway Curr Opin Chem Biol 2002 6 844 850 12470740 10.1016/S1367-5931(02)00390-3
Corpet F Multiple sequence alignment with hierarchical clustering Nucleic Acids Res 1988 16 10881 10890 2849754
Nielsen H Engelbrecht J Brunak S von Heijne G Identification of prokaryotic and eukaryotic signal peptides and prediction of their cleavage sites Protein Eng 1997 10 1 6 9051728 10.1093/protein/10.1.1
Nielsen H Krogh A Prediction of signal peptides and signal anchors by a hidden Markov model Proc Int Conf Intell Syst Mol Biol 1998 6 122 130 9783217
Emanuelsson O Nielsen H Brunak S von Heijne G Predicting subcellular localization of proteins based on their N-terminal amino acid sequence J Mol Biol 2000 300 1005 1016 10891285 10.1006/jmbi.2000.3903
Johnson LV Calarco PG Stage-specific embryonic antigens detected by an antiserum against mouse blastocysts Dev Biol 1980 79 224 231 7409322 10.1016/0012-1606(80)90085-8
Polak-Charcon S Calarco-Gillam P Johnson L Intracellular localization and surface expresssion of a stage-specific embryonic glycoprotein Gamete Res 1985 12 329 343 10.1002/mrd.1120120402
Seetharam B Li N Transcobalamin II and its cell surface receptor Vitam Horm 2000 59 337 366 10714245
Minghetti PP Ruffner DE Kuang WJ Dennison OE Hawkins JW Beattie WG Dugaiczyk A Molecular structure of the human albumin gene is revealed by nucleotide sequence within q11-22 of chromosome 4 J Biol Chem 1986 261 6747 6757 3009475
Figarella-Branger D Civatte M Bartoli C Pellissier JF Cytokines, chemokines, and cell adhesion molecules in inflammatory myopathies Muscle Nerve 2003 28 659 682 14639580 10.1002/mus.10462
Terakawa H Takahara H Sugawara K Three types of mouse peptidylarginine deiminase: characterization and tissue distribution J Biochem (Tokyo) 1991 110 661 666 1778991
Krogh A Larsson B von Heijne G Sonnhammer EL Predicting transmembrane protein topology with a hidden Markov model: application to complete genomes J Mol Biol 2001 305 567 580 11152613 10.1006/jmbi.2000.4315
Sonnhammer EL von Heijne G Krogh A A hidden Markov model for predicting transmembrane helices in protein sequences Proc Int Conf Intell Syst Mol Biol 1998 6 175 182 9783223
Calvert ME Digilio LC Herr JC Coonrod SA Oolemmal proteomics – identification of highly abundant heat shock proteins and molecular chaperones in the mature mouse egg and their localization on the plasma membrane Reprod Biol Endocrinol 2003 1 27 12646049 10.1186/1477-7827-1-27
Pierce KE Grunvald EL Schultz RM Kopf GS Temporal pattern of synthesis of the mouse cortical granule protein, p75, during oocyte growth and maturation Dev Biol 1992 152 145 151 1628752 10.1016/0012-1606(92)90164-C
Wang Y Wysocka J Sayegh J Lee YH Perlin JR Leonelli L Sonbuchner LS McDonald CH Cook RG Dou Y Roeder RG Clarke S Stallcup MR Allis CD Coonrod SA Human PAD4 regulates histone arginine methylation levels via demethylimination Science 2004 306 279 283 15345777 10.1126/science.1101400
Dedhar S Integrins and signal transduction Curr Opin Hematol 1999 6 37 43 9915552 10.1097/00062752-199901000-00007
Juliano R Cooperation between soluble factors and integrin-mediated cell anchorage in the control of cell growth and differentiation Bioessays 1996 18 911 917 8939069 10.1002/bies.950181110
Martins-Green M Bissell M Cell-ECM interactions in development Sem Dev Biol 1995 6 149 159
Schmidt A Hall MN Signalling to the actin cytoskeleton Annual Review of Cell and Development Biology 1998 14 305 338 10.1146/annurev.cellbio.14.1.305
|
16137333
|
PMC1215517
|
CC BY
|
2021-01-04 16:37:11
|
no
|
Reprod Biol Endocrinol. 2005 Sep 1; 3:42
|
utf-8
|
Reprod Biol Endocrinol
| 2,005 |
10.1186/1477-7827-3-42
|
oa_comm
|
==== Front
Reprod Biol EndocrinolReproductive biology and endocrinology : RB&E1477-7827BioMed Central London 1477-7827-3-461615689010.1186/1477-7827-3-46ReviewArterial oxygen saturation in healthy newborns delivered at term in Cerro de Pasco (4340 m) and Lima (150 m) Gonzales Gustavo F [email protected] Amelia [email protected] Department of Biological and Physiological Sciences. Faculty of Sciences and Philosophy, Universidad Peruana Cayetano Heredia, Av. Honorio Delgado 430. Urb. Ingenieria. Lima, Peru. PO Box 1843. Lima, Peru2 Instituto de Investigaciones de la Altura. Universidad Peruana Cayetano Heredia, Lima, Peru2005 12 9 2005 3 46 46 1 8 2005 12 9 2005 Copyright © 2005 Gonzales and Salirrosas; licensee BioMed Central Ltd.2005Gonzales and Salirrosas; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
High altitude is associated with both low pulse oxygen saturation at birth and more pre-term deliveries. The present study was performed to determine pulse oxygen saturation in newborns at term in Cerro de Pasco (4340 m) and Lima (150 m) to test the hypothesis that low pulse oxygen saturation at birth at high altitudes was not observed at term deliveries.
Methods
The present study was designed to determine pulse oxygen saturation values through 1 minute to 24 hours and values of Apgar score at 1 and 5 minutes in newborns delivered at term in Cerro de Pasco (4340 m) and Lima (150 m). Pulse oxygen saturation was recorded in 39 newborns from Cerro de Pasco (4340 m) and 131 from Lima (150 m) at 1, 2, 3, 4, 5, 10, 15, 30 minutes and 1, 2, 8 and 24 hours after delivery. Apgar score was assessed at 1 and 5 minutes after birth. Neurological score was assessed at 24 h of birth by Dubowitz exam.
Results
Pulse oxygen saturation increased significantly from 1 to 15 min after birth at sea level and from 1 to 30 minutes at Cerro de Pasco. Thereafter, it increased slightly such that at 30 min at sea level and at 60 minutes in Cerro de Pasco it reached a plateau up to 24 hours after birth. At all times, pulse oxygen saturation was significantly higher at sea level than at high altitude (P < 0.01). At 1 minute of life, pulse oxygen saturation was 15% lower at high altitude than at sea level. Apgar score at 1 minute was significantly lower at high altitude (P < 0.05). Neurological score at 24 hours was also lower at high altitude than at sea level. Head circumference, and Apgar score at 5 minutes were similar at sea level and at high altitude (P:NS). Incidence of low birth-weight (<2500 g) at high altitude (5.4%) was similar to that observed at sea level (2.29%) (P:NS). Incidences of low pulse oxygen saturation (<30%), low Apgar score at first minute (<7) and low neurological score at 24 h (<19) were significantly higher at high altitude than at sea level (P < 0.0001; P < 0.0001; and P < 0.001, respectively).
Conclusion
From these analyses may be concluded that pulse oxygen saturation at 4340 m was significantly low despite the fact that births occurred at term. Apgar scores at first minute and neurological scores were also lower at high altitudes.
==== Body
Background
Post-ductal pulse oxygen saturation (SpO2) in healthy newborns infants two minutes after birth was 67% in average and it takes 14 minutes to reach an oxygen saturation of 95% [1]. Harris et al [2] found at sea level a gradual rise in mean SpO2 value from 61% to 5 min after an uncomplicated vaginal birth.
Studies on pulse oxygen measurements during the first 30 minutes of life at high altitudes are few. Gonzales and Salirrosas [3] showed that mean pulse oxygen saturation at first minute of life was 41.3% in Cerro de Pasco, Peru (4340 m) compared with 60.4% in Lima (150 m). Ramirez-Cardich et al [4] (2004) measured pulse oxygen saturation at 10 minutes of life in La Oroya, Peru (3,800 m), but data were not compared with those obtained at sea level. All of the other studies are related to values after 30 minutes of life. In these reports, pulse oxygen saturation was significantly lower in infant populations at high altitudes around the world [5-10] as compared with values at sea level. Pulse oxygen saturation values in infants decreased as altitude increased. At 2640 m, In Bogota, Colombia, SpO2 in children aged 5 days to 24 months was 93.3%, whereas in El Alto, Bolivia (4018 m) the mean value of SpO2 in children aged 0–5 months was 87.3% [7]. Low arterial oxygen saturation at 30 min of life has also been observed in Morococha, Peru (4540 m) [11].
The importance to study pulse oxygen saturation in newborns at high altitude is that increased neonatal and infant mortality have been associated with reduced oxygen availability at high altitude [9]. SpO2 cutoff value of 30% has been considered reasonable in clinical trials as values below 30% were related to low umbilical arterial blood pH [12,13]. In Cerro de Pasco (4340 m), 20.3% of newborns had low pulse oxygen saturation (<30% SpO2) whereas in Lima the incidence was of 2.1% [3].
Why at the same oxygen partial pressure some individuals are more hypoxemic than others seems to be related to antiquity in the place of birth [5,9,14], high incidence of prematurity at high altitude [15] or intrauterine growth restriction [16].
Antiquity of the generations living at high altitudes seems to be related to the pulse oxygen saturation values [5,9,14]. To more antiquity the life at high altitude, the pulse oxygen saturation value will be higher. For instance, In Lhasa, Tibetan newborns had higher arterial oxygen saturation at birth and during the first four months of life than Han newborns. Native Tibetans have inhabited the Himalayan Plateau for approximately 25,000 years, whereas people of Han ancestry had moved there only 50 years ago [9].
In both Huancayo (3280 m) and Cerro de Pasco (4340 m) located in the Central Andes of Peru pre-term deliveries (<37 weeks of gestational age) were higher than at Lima (150 m) [15,17]. These findings contrast with large population-based studies which refer that altitude produce fetal growth restriction [16].
In Peru, populations from Central Andes as Cerro de Pasco are more recently inhabited (about 400 years) whereas those from the Southern Andes live there for more than 10,000 years [18]. The low antiquity of life at Cerro de Pasco or the high incidence of pre-term deliveries may explain the low pulse oxygen saturation values in newborn in that place.
In an attempt to control for the effect of pre-term deliveries, we have examined if pulse oxygen saturation in newborns at term in Cerro de Pasco (4340 m) at the Central Andes of Peru was different to that in another population at term born at sea level.
Methods
The study was performed in 39 newborn infants immediately after birth in Cerro de Pasco and 131 in Lima. Sample size was calculated with a limit of confidence of 95% (z = 1.96) and a power of 15%. The main outcome variable to determine sample size was the incidence of low pulse oxygen saturation at Cerro de Pasco (20%) and low pulse oxygen saturation at Lima (2%) [3]. Minimum sample size was 35. With 10% of missing data, the minimum final sample size was 39 at each place. In the present study we have included 3 cases at sea level by each case at high altitude. For the analysis two infants from Cerro de Pasco were excluded because data from gestational age measured by Usher method was 37 and 38 weeks.
Deliveries in Lima (150 m) occurred at the Hospital Nacional María Auxiliadora and in Cerro de Pasco (4340 m) at the Hospital Daniel Alcides Carrion. Both are public hospitals attending by a population of low socioeconomic status. Mothers in the present study live in districts near the public hospitals, and these districts are classified as belonging to the low socioeconomic strata.
During the last 60 years, significant migration has occurred from the Peruvian highland to the coast [19]. Most of the migrants are settled in places named cones. One of these cones corresponds to the district in which the Hospital Maria Auxiliadora is located. From this, it results that population at sea level studied has the same ethnicity as those from Cerro de Pasco. In the present study 34.35% of mothers at sea level had at least one Quechua surname, whereas 27.94% of mothers at Cerro de Pasco had at least one Quechua surname.
Weeks of gestation for the mothers were 37–42 weeks (39.33 ± 0.12 weeks for Lima and 40.49 ± 0.16 weeks for Cerro de Pasco, mean ± SEM) based on the last normal menstrual period (LNMP). This value was confirmed with Usher method [20]. Mean vales for gestational age are observed in Table 1. It has been demonstrated that the large majority of deliveries occurring at or near term showed LNMP-based gestational age estimates that were valid within plus or minus seven days of the ultrasound estimates (Gold standard). There were excluded from the study neonates with malformations, and those whose parents did not accept voluntarily to participate in the study.
Table 1 Characteristics of the mothers at sea level and at high altitude (4340 m).
Lima (n = 131) Cerro de Pasco (n = 37)
mean ± SEM mean ± SEM
Age (years) 25.15 ± 0.58 25.46 ± 0.89
Body Weight (Kg) 54.98 ± 0.69 53.31 ± 1.13
Height (m) 1.55 ± 0.01 1.53 ± 0.01
Age at menarche (years) 13.16 ± 0.16 14.22 ± 0.25*
Hemoglobin at second trimester gr% 10.53 ± 0.27 (32) 13.42 ± 0.64 (6)*
Hemoglobin at third trimester gr% 10.87 ± 0.14 (80) 13.36 ± 0.42 (14)*
Gestational age (Usher) weeks 40.17 ± 0.03 39.51 ± 0.16*
Number of subjects is between parentheses. *P < 0.01 with respect to values in Lima (150 m)
Pediatricians, medical students or nurses provided post-birth care, and assigned an Apgar score at 1 and 5 minutes. No fetal monitoring and respiratory rate data were available.
Data on maternal age, body weight, height and age at menarche were obtained from the clinical records. In the cases where mother received pre-natal care, hemoglobin measurements were also recorded.
This study was approved by the Institutional Review Board of the Scientific Research Direction at the Universidad Peruana Cayetano Heredia.
Techniques
Oximetry was performed using a pulse oximeter NELLCOR N-20 (Nellcor, Inc., Hayward, Calif) with a sensor cable model OC-3 for OXICLIQ-N sensor (oxygen transducer) applied to the first finger of the left foot immediately after delivery and after a segment of the umbilical cord was clamped. Post-ductal pulse oxygen saturation measurements were performed at 1, 2, 3, 4, 5, 10, 15, 30, 60 minutes, and 2, 8 and 24 hours after delivery. The sensor was changed almost every 8 neonates.
Oxygen saturation was monitored continuously by one of the researchers (AS) in both, Lima and Cerro de Pasco. The value observed at each time of monitoring was recorded. To assurance that readings were valid a regression analysis between pulse oxygen saturation at first minute of life and Apgar score at first minute was performed after controlling maternal age, gestational age, birth-weight, and height at birth.
The neurological evaluation was assessed at 24 hours after birth by using the neurological part of the Dubowitz's exam [21]. The neurological part of the exam was scored from 0 to 35. The neurological exam was applied when the newborn infant was in peaceful conditions. The Dubowitz exam included the following neurological criteria: posture, square window, ankle dorsiflexion, arm recoil, leg recoil, popliteal angle, scarf sign, heel to ear, head lag and ventral suspension.
The modified Dubowitz score was validated with the Scanlon neurobehavioral testing, which, it is specific for neurological assessment shortly after birth. The Scanlon test could be assessed as early as eight hours after birth [22]. The modified Dubowitz score (MDS) showed a significant correlation with Scanlon Neurobehavioral Testing: r = 0.64; P < 0.001 [23]. The test was performed by one of the researchers: AS.
Statistical Analysis
Values for pulse oxygen saturation were averaged for each neonate according to place of study and time after birth. Data are reported as group means ± SEM. Data were analyzed by two-way ANOVA. Differences between pair of means were performed with the Scheffee test.
Low birth-weight was defined as values below 2500 g; low Apgar at first or fifth minute of life was defined as values below 7; low pulse oxygen saturation at first minute was defined as values <30%; low neurological scores at 24 h was defined when values were <19. Cases of low pulse oxygen saturation at first minute, low Apgar at first minute, and low neurological score were assessed as proportions. Differences between proportions were analyzed by Fisher's exact test. A multiple regression analysis was also performed to determine the independent effect of changes in SpO2 during first three minutes on neurological score.
A P-value was considered significant when was below 0.05.
Results
Characteristics of the mothers
Mothers that were included in this study had ages that ranged from 15 to 41 years (25.15 ± 0.58, mean ± SEM) in Lima and 16–42 years (25.46 ± 0.89) in Cerro de Pasco. Mother's ages, heights and body weights were similar at sea level and at high altitude (Table 1). Age of menarche was delayed at high altitude related to sea level (P < 0.01). Gestational age estimate by Usher method was 0.66 weeks lower at high altitude than at sea level (P < 0.01) (Table 1). Premature rupture of membranes was recorded in 20.6% of mothers at sea level and 16.2% at high altitude. At sea level, 87.03% of women had prenatal care, whereas at high altitude 70.27% of women received prenatal care.
Hemoglobin values were obtained at the second trimester in 32 cases in Lima, and 6 cases in Cerro de Pasco, and for the third trimester in 80 cases in Lima and 14 cases in Cerro de Pasco. Hemoglobin values at second and third trimester of pregnancy were also higher at high altitude than at sea level (P < 0.01).
Characteristics of the newborns
Despite that both groups were conformed by newborns delivered at term (40–41 weeks of gestation), the birth-weight (P < 0.01), height (P < 0.01), and thoracic circumference (P < 0.05) were significantly lower at Cerro de Pasco (4340 m) than at Lima (150 m). Birth weight at high altitude represented in average 400 grams less than at sea level, whereas height represented 1.5 cm less at high altitude than at sea level. Apgar at first minute of life (P < 0.05), pulse oxygen saturation at first minute (P < 0.01), and neurological score at 24 hours (P < 0.01) were also lower at high altitude than at sea level. Head circumference, and Apgar score at 5 minutes were similar at sea level and at high altitude (P: NS) (Table 2). No infant included in the present study required oxygen therapy as part of delivery room care.
Table 2 Characteristics of the newborns delivered at 40–41 weeks of gestational age in Lima (150 m) and Cerro de Pasco (4340 m)
Lima (n = 131) Cerro de Pasco (n = 37)
mean ± SEM mean ± SEM
Birthweight (g) 3329.31 ± 34.84 2925.41 ± 43.63*
Height (cm) 50.54 ± 0.15 48.99 ± 0.22*
Head circumference (cm) 34.30 ± 0.11 34.31 ± 0.20
Thoracic circumference (cm) 33.55 ± 0.16 32.77 ± 0.22**
Apgar At first minute 7.95 ± 0.06 7.03 ± 0.35**
Apgar at five minutes 8.96 ± 0.03 8.68 ± 0.14
SpO2 at first minute (%) 60.60 ± 1.20 42.32 ± 2.54*
Neurological score at 24 h 27.44 ± 0.30 22.00 ± 0.67*
*P < 0.01, **P < 0.05 with respect to values in Lima (150 m)
Pulse oxygen saturation at first minute was significantly related to Apgar score at first minute of life after controlling for maternal age, gestational age, birth-weight and height at birth. The coefficient of determination was R2 = 0.30, P < 0.003. At high altitude, SpO2 at first minute in neonates to mothers with prenatal care was similar to that obtained in neonates to mothers without prenatal care (41.13 ± 3.19%, mean ± SEM and 44.82 ± 3.75%, respectively; P:NS). Similarly, in cases of premature rupture of membranes, the SpO2 in neonates at first minute of life was 43.25 ± 2.88% similar to that obtained when premature rupture of membranes was present, 40.10 ± 4.86% (P:NS).
Incidence of neonates with low birth-weight (<2500 g) at high altitude (5.4%) was similar to that observed at sea level (2.29%) (P:NS). Incidence of low pulse oxygen saturation (<30%), low Apgar score at first minute (<7) and low neurological score at 24 h (<19) was significantly higher at high altitude than at sea level (Odds Ratio, OR: 8.76, 9.96, and 35.86, respectively) (Table 3).
Table 3 Incidence of low birth-weight, low pulse oxygen saturation at first minute, low Apgar at first minute, and low neurological score at 24 hours.
Incidence of Lima Cerro de Pasco OR Confidence Interval (at 95%) P
Low birthweight 3/131 (2.29%) 2/37 (5.40%) 2.44 0.39–15.18 0.3
Low pulse oxygen saturation 4/131 (3.05%) 8/37 (21.62) 8.76 2.47–31.07 0.0001
Low Apgar at first minute 3/131 (2.29%) 7/37 (18.92%) 9.96 2.43–40.77 0.0001
Low neurological score at 24 h 1/131 (0.76%) 8/37 (21.62%) 30.33 3.60–255.92 0.0001
Cases are in the numerator. OR: Odds Ratio. P: Probability.
Pulse oxygen saturation in newborn infants
Figure 1 shows SpO2 measured by pulse oximetry from 1 minute to 24 hours of life. Oxygen saturation increased with high slope from 60.60 ± 1.20% (mean ± SEM) at first minute of life to 91.10 ± 0.5% at 15 minutes of life in Lima and from 45.08 ± 2.47% at first minute to 87.56 ± 1.19% at 30 minutes of life in Cerro de Pasco. Thereafter a plateau was observed. At all times SpO2 values were higher at sea level than at high altitude (P < 0.01). According to the regression analysis, SpO2 increases with time after birth (4.11 ± 0.07, coefficient of regression ± standard error; P < 0.0001) and decreases with altitude (-63.90 ± 0.98; P < 0.0001). The coefficient of determination for the model was 0.64. When the interaction place of birth X time after birth is included in the model, the increase in the coefficient of determination (R2 = 0.65) is negligible. However, this interaction was significant (1.19 ± 0.14; coefficient of regression ± standard error; P:0.0001).
Figure 1 Mean pulse oxygen saturation in newborns at 40–41 weeks of gestational age at sea level and at high altitude (4340 m). Data are mean ± SEM. F value = 2146.97, P < 0.0001 (Two-way ANOVA test).*P < 0.001 with respect to values at sea level.
Probability for low neurological score at 24 hours of life at high altitude
Almost all of the components of the Dubowitz score were significantly lower at high altitude than at sea level. At high altitude score of the leg recoil test was not related to the Dubowitz score (r = 0.28; P:NS). The highest correlation was observed with scarf sign test (r = 0.68, P < 0.001), popliteal angle (r = 0.63, P < 0.001), posture (r = 0.60, P < 0.001).
At high altitude, neurological score at 24 hours was related to SpO2 values at first minute of life (8.09 × 10-2 ± 0.04; coefficient of regression ± SE, P < 0.05). In Table 4 is presented data from the multivariate analysis to assess the independent effect of change of SpO2 from 1st to 3rd minute of life on neurological score at 24 hours. Data showed that low neurological score was related to dilatation period (P < 0.05) but not to changes in SpO2 from first to third minute of life (P:NS).
Table 4 Multiple regression analysis to assess the independent effect of pulse oxygen saturation during firsts minutes of life on neurological score at 24 hours of birth in Cerro de Pasco (4340 m).
Neurological score β ± SE P Confidence Interval at 95%
SpO2 3 min - SpO2 1 min 3.18 × 10-2 ± 0.03 NS -0.029 0.093
Period of dilatation 0.24 ± 0.11 <0.05 0.017 0.463
Constant 16.22 ± 2.62 <0.001 10.90 21.539
β ± SE: Coefficient of Regression ± standard error. P = Level of significance.
Discussion
Monitoring of pulse oxygen saturation is useful for the care of infants and children [24]. We had studied pulse oxygen saturation in a selected population of newborns delivered at term in Lima (150 m) and Cerro de Pasco (4340 m) in the Central Andes of Peru. Results demonstrated that pulse oxygen saturation was significantly lower at high altitudes than at sea level since first minute of life.
A rapid increase in SaO2 must occur within the first few minutes of extra-uterine life if viability is to be maintained. In the present study we have demonstrated that pulse oxygen saturation increases significantly from 1 minute to 15 minutes at sea level and from 1 to 30 minutes at high altitude. Others obtained similar findings in populations at sea level [1,2].
It is possible that different factors may affect pulse oxygen readings as meconium-stained amniotic fluid, child's activity during the SpO2 measurements, and an anemic condition of the mothers. In a recent study using the Nellcor pulse-oximeter, in cases with meconium as compared to clear amniotic fluid, the pulse oximetry remained unchanged [13]. The effect of sleep or awake during measurement in children between 24 hours of life was negligible. Only after one week of age do activities influence SpO2 [9]. With respect to anemic condition, our small sample may not allow an adequate analysis on anemic and non-anemic women. However, another study had demonstrated that SpO2 levels and respiratory rates of infant natives at 3750 meters were not different between infants born to non-anemic and anemic mothers [4]. We also do not found impact of pre natal care or premature rupture of membranes on SpO2 values of neonates at first minute of life.
Low oxygenation may have some implications on neurological development, particularly in cases when low oxygenation was prolonged. In fact 0.76% of newborns at sea level showed low neurological score at 24 hours, compared to 21.62% in Cerro de Pasco at 4340 m. These incidence values are similar to those observed for the incidence of low pulse oxygen saturation at minute 1 of life.
It is not clear whether low oxygen saturation at high altitude occurred intra-partum or whether it was an accumulative effect from intrauterine life. It has been observed that a single prolonged episode of fetal hypoxia may produce neurological abnormality [25]. We have previously demonstrated that at sea level maintaining low oxygen saturation for at least three minutes the neurological score at 24 hours of life was low [23]. However, the situation in Cerro de Pasco was different. In fact, changes of SpO2 from 1 to 3 minutes were not related to neurological score at 24 hours. Instead, SpO2 at first minute or dilatation period were related to the neurological score at 24 hours.
It is interesting the finding that incidence of low birth-weight (2500 g) in these selected populations at sea level and at high altitude was similar. However, the sample is small to make a statement. In fact, mean birth weight was significantly lower at high altitude than at sea level (400 g less in average). This suggests that a restriction in growth occurs at high altitude. This has been described by other authors [26-28]. Low mean birth-weight at high altitude of Cerro de Pasco in newborns at term seems to be related more to the low-pressure exposure than for socio-economic differences [29,30]. It has been suggested that altitude acts as an independent factor in determining birth weight with a reduction of 100 gr. per 1000 m elevation gain [31]. In our study, a difference of 400 grams less in average was observed in birth weight at 4340 m than at sea level.
Our data also demonstrated that head circumference was similar at sea level and at high altitude despite that low average birth-weight and height. Several studies have demonstrated that head size is relatively well preserved whereas body weight is significantly reduced at high altitudes [30,32].
The Apgar score is the traditional and most universal criterion for assessing the newborn's well being in the delivery room. The data of this study indicate that Apgar score at first but not at fifth minute was lower at high altitude than at sea level. This was also associated with a significantly high incidence of low Apgar score at first minute in high altitude (18.92%) than at sea level (2.29%) neonates. This finding is confirmatory with our previous results of higher incidence of low Apgar at Cerro de Pasco [15].
We have previously demonstrated that Apgar score at 1 minute but not at 5 minutes was related to neurological evaluation at 24 hours [23]. This means that infants with low Apgar score at first minute may recover at 5 minute but it does not improve neurological evaluation at 24 h of life. Low SpO2 (<30%) has been associated with acidemia [12]. Low arterial blood pH (<7.01) at birth was significantly related to low Apgar score [33]. Then it is not surprising that SpO2 at first minute was related to Apgar score at first minute and with low neurological score at 24 hours of life.
In summary, the study demonstrates that after controlling gestational age in the design pulse oxygen saturation and Apgar scores at 1 minute in newborns delivered at term were lower at high altitude in the Central Andes than at sea level.
Authors' contributions
GFG conceived of the study, and participated in its design and coordination and drafted the manuscript.
AS participated in the design of the study, the work of field and the statistical analysis.
All authors read and approved the final manuscript.
Acknowledgements
A Grant of Maintenance Resources from the Human Reproduction Programme of the World Health Organization, Geneva, Switzerland, and a Grant from the Consejo Nacional de Ciencia y Tecnología (CONCYTEC), Lima, Peru, supported this study.
==== Refs
Toth B Becker A Seelbach-Gobel B Oxygen saturation in healthy newborn infants immediately after birth measured by pulse oximetry Arch Gynecol Obstet 2002 266 105 107 12049291 10.1007/s00404-001-0272-5
Harris AP Sendak MJ Donham RT Changes in arterial oxygen saturation immediately after birth in the human neonate J Pediatr 1986 109 117 119 3723231
Gonzales GF Salirrosas A Ohno T, Kobayashi S, Masuyama M, Nakashima H The relationship of pulse oxygen saturation at birth and the score of neurological evaluation at 24 hours of life Progress in Mountain Medicine and High Altitude Physiology 1998 Matsumoto 327 328
Ramirez-Cardich ME Saito M Gilman RH Escate LE Strouse JJ Kabrhel C Johnson C Galchen R Bautista CT Effect of maternal anemia at high altitude on infant hematocrit and oxygenation Am J Trop Med Hyg 2004 70 420 424 15100457
Beall CM Strohl KP Blangero J Williams-Blangero S Decker MJ Brittenham GM Goldstein MC Quantitative genetic analysis of arterial oxygen saturation in Tibetan highlanders Hum Biol 1997 69 597 604 9299881
Beall CM Oxygen saturation increases during childhood and decreases during adulthood among high altitude native Tibetans residing at 3,800–4,200 m High Alt Med Biol 2000 1 25 32 11258584 10.1089/152702900320658
Gamponia MJ Babaali H Yugar F Gilman RH values for pulse oxymetry at high altitude Arch Dis Child 1998 78 461 465 9659095
Niermeyer S Shaffer EM Thilo E Corbin C Moore L Arterial oxygenation and pulmonary arterial pressure in healthy neonates and infants at high altitude J Pediatr 1993 123 767 772 8229488
Niermeyer S Yang I Shanmina Drolkar Zhuang J Moore L Arterial oxygen saturation in Tibetan and Han infants born in Lhasa, Tibet N Engl J Med 1995 333 1248 1252 7566001 10.1056/NEJM199511093331903
Lozano JM Duque OR Buitrago T Behaine S Pulse oxymetry reference values at high altitude Arch Dis Child 1992 67 299 301 1575552
Gamboa R Marticorena E Presión arterial pulmonar en el reciénnacido en las grandes Alturas Arch Inst Biol And 1971 4 55 66
Dildy GA Thorp JA Yeast JD Clark SL The relationship between oxygen saturation and pH in umbilical blood: implications for intrapartum fetal oxygen saturation monitoring Am J Obstet Gynecol 1996 175 682 687 8828434 10.1053/ob.1996.v175.a74922
Carbonne B Cudeville C Maillard F Goffinet F Predictive value of pulse oximetry and fetal scalp blood pH in the case of meconium-stained amniotic fluid Eur J Obstet Gynecol Reprod Biol 2003 109 27 32 12818439 10.1016/S0301-2115(02)00472-4
McAuliffe F Kametas N Krampl E Ernsting J Nicolaides K Blood gases in pregnancy at sea level and at high altitude Br J Obstet Gynaecol 2001 108 980 985 10.1016/S0306-5456(01)00225-X
Gonzales GF Guerra-Garcia R Gonzales GF Características hormonales y antropométricas del embarazo y del recién nacido en la altura Reproducción Humana en la Altura 1993 Lima: Ediciones IIA 125 141 15276991
Moore LG Fetal growth restriction and maternal oxygen transport during high altitude pregnancy High Alt Med Biol 2003 4 141 156 12855048 10.1089/152702903322022767
Carrillo CE Wong A Factores asociados al embarazo pre-término a nivel del mar y en la altura Acta And 1994 3 19 28
Bonavia D Peru. Man and History. From the Origin to Century XV 1991 1 Edubanco:Lima 586
Gonzales GF Ortiz I Age at menarche at sea level and high altitude in Peruvian women of different ethnic background Am J Hum Biol 1994 6 637 640 10.1002/ajhb.1310060512
Usher R McLean F Scott KE Judgment of fetal age. II. Clinical significance of gestational age and an objective method for its assessment Pediatr Clin North Am 1966 13 835 862 5946307
Dubowitz LMS Dubowitz V Goldberg C Clinical assessment of gestational age in the newborn infant J Pediatr 1970 77 1 10 5430794
Scanlon JW Brown WU JrWeiss JB Alper MH Neurobehavioral responses of newborns after maternal anesthesia Anesthesiology 1974 40 121 228 4812709
Gonzales GF Salirrosas A Pulse oxygen saturation and neurologic assessment in human neonates after vaginal and cesarean delivery Int J Gynaecol Obstet 1998 63 63 66 9849715 10.1016/S0020-7292(98)00109-X
Salyer JW Neonatal and pediatric pulse oximetry Respir Care 2003 48 386 396 12667266
Low JA Ludwin SK Fisher S Severe fetal asphyxia associated with neuropathology Am J Obstet Gynecol 1996 175 1383 1385 8942523
Keyes LE Armaza JF Niermeyer S Vargas E Young DA Moore LG Intrauterine growth restriction, preeclampsia, and intrauterine mortality at high altitude in Bolivia Pediatr Res 2003 54 20 25 12700368 10.1203/01.PDR.0000069846.64389.DC
Moore LG Young D McCullough RE Droma T Zamudio S Tibetan protection from intrauterine growth restriction (IUGR) and reproductive loss at high altitude Am J Hum Biol 2001 13 635 644 11505472 10.1002/ajhb.1102
Zamudio S Droma T Norkyel KY Acharya G Zamudio JA Niermeyer SN Moore LG Protection from intrauterine growth retardation in Tibetans at high altitude Am J Phys Anthropol 1993 91 215 224 8317562
Giussani DA Phillips PS Anstee S Barker DJP Effect of altitude versus economic status on birth weight and body shape at birth Pediatr Res 2001 49 490 494 11264431
Krampl E Lees C Bland JM Espinoza Dorado J Moscoso G Campbell S Fetal biometry at 4300 m compared to sea level in Peru Ultrasound Obstet Gynecol 2000 16 9 18 11084959 10.1046/j.1469-0705.2000.00156.x
Jensen GM Moore LG The effect of high altitude and other risk factors on birthweight: independent or interactive effects? Am J Public Health 1997 87 1003 1007 9224184
Zamudio S Moore LG Altitude and fetal growth: current knowledge and future directions Ultrasound Obstet Gynecol 2000 16 6 8 11084958 10.1046/j.1469-0705.2000.00155.x
Hegyi T Carbone T Anwar M Ostfeld B Hiatt M Koons A Pinto-Martin J Paneth N The Apgar score and its components in the pre-term infant Pediatrics 1998 101 77 81 9417155 10.1542/peds.101.1.77
|
16156890
|
PMC1215518
|
CC BY
|
2021-01-04 16:37:12
|
no
|
Reprod Biol Endocrinol. 2005 Sep 12; 3:46
|
utf-8
|
Reprod Biol Endocrinol
| 2,005 |
10.1186/1477-7827-3-46
|
oa_comm
|
==== Front
Reprod HealthReproductive Health1742-4755BioMed Central London 1742-4755-2-51609552610.1186/1742-4755-2-5ReviewIllegal births and legal abortions – the case of China Hemminki Elina [email protected] Zhuochun [email protected] Guiying [email protected] Kirsi [email protected] Research professor, National Research and Development Centre for Welfare and Health STAKES, P.O. Box 220, 00531 Helsinki, Finland2 Associate professor, School of Public Health, Fudan University, P.O. Box 250, 200032 Shanghai, China3 Senior researcher, International Institute for Applied Systems Analysis, 2361 Laxenburg, Austria4 Senior researcher, National Research and Development Centre for Welfare and Health STAKES, P.O. Box 220, 00531 Helsinki, Finland2005 11 8 2005 2 5 5 3 3 2005 11 8 2005 Copyright © 2005 Hemminki et al; licensee BioMed Central Ltd.2005Hemminki et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
China has a national policy regulating the number of children that a woman is allowed to have. The central concept at the individual level application is "illegal pregnancy". The purpose of this article is to describe and problematicize the concept of illegal pregnancy and its use in practice.
Methods
Original texts and previous published and unpublished reports and statistics were used.
Results
By 1979 the Chinese population policy was clearly a policy of controlling population growth. For a pregnancy to be legal, it has to be defined as such according to the family-level eligibility rules, and in some places it has to be within the local quota. Enforcement of the policy has been pursued via the State Family Planning (FP) Commission and the Communist Party (CP), both of which have a functioning vertical structure down to the lowest administrative units. There are various incentives and disincentives for families to follow the policy. An extensive system has been created to keep the contraceptive use and pregnancy status of all married women at reproductive age under constant surveillance. In the early 1990s FP and CP officials were made personally responsible for meeting population targets. Since 1979, abortion has been available on request, and the ratio of legal abortions to birth increased in the 1980s and declined in the 1990s. Similar to what happens in other Asian countries with low fertility rates and higher esteem for boys, both national- and local-level data show that an unnaturally greater number of boys than girls are registered as having been born.
Conclusion
Defining a pregnancy as "illegal" and carrying out the surveillance of individual women are phenomena unique in China, but this does not apply to other features of the policy. The moral judgment concerning the policy depends on the basic question of whether reproduction should be considered as an individual or social decision.
AsiaChinapopulation policyillegal pregnancyabortionsex ratio
==== Body
Introduction
A current widely accepted principle is that it is the parents' – or the woman's – societal right to have children: as many as they want and are capable of having. Restrictions to this right include age limits in marriage, forbidding marriages involving close relatives, sterilizations due to certain parental conditions, and outside evaluation when help is requested from the health or social care services, i.e. in cases of assisted reproduction or adoption. But once the woman or couple has been deemed eligible for childbearing in general, the number of children is not regulated by law. By contrast, the Western moral opinion is divided on the issue of a woman's right not to have children. The Roman Catholic Church opposes "artificial" contraception, and induced abortions are a hot political and moral issue in many countries. Abortions are clearly in the political arena, and individual women's choices are often regulated. A common legal arrangement is that abortions are allowed if performed in early pregnancy by licensed health professionals [1].
China has a national policy regulating the number of children that a woman may have. The goals of population policy and resulting fertility have been extensively described in the English-language literature [2-7], but descriptions of how the policy has been implemented are few and patchy. A crucial and innovative element has been that of "official permission to be pregnant". In most other countries these types of permission are only socially constructed. The central concept applied at the individual level is that of "illegal pregnancy" (in Chinese "ji hua wai ren shen", "unauthorized" pregnancy). We use the term "illegal pregnancy" in analogy with that of illegal abortion. "Unplanned pregnancy" is misleading to non-Chinese readers, because this term is commonly used in reference to planning by individual women, not to permission given by authorities. "Unauthorized" or "unapproved" pregnancy would be the most accurate translation, but it is not suggestive of the incentives and fines attached.
The purpose of this article is to describe and discuss the concept of illegal pregnancy and how it has been applied in practice. We will describe how it has been operationalized and how it relates to the sex distribution of newborns and to abortions. The concept will be examined from a historical perspective, after a brief description of China's population policy.
Methods
Contents of the laws and regulations were obtained from the original documents, and from previous English-language (identified in the text) and Chinese-language articles. The interpretation of the regulations is based on our general knowledge and previous research on Chinese policies. In China, state plans and decisions are frequently passed via other structures than the law, and prior to 2001 most population policy documents were decisions and official letters from the party or civil administration, called regulations in the subsequent text. Many regulations were made at provincial level (30 provinces).
National empirical data derive from published reports and statistics, identified in the text. Local data are from published research articles or from our own research, during which we collected data on pregnancy outcomes of a trial involving the introduction of comprehensive prenatal care in a rural county in Anhui Province [8].
China's population policies
Communist China was founded in 1949 after the Japanese occupation and a long and devastating civil war. Initially the Government instituted a strong pronatalistic policy encouraging child-bearing by means of child subsidies, and prohibiting contraceptives, abortions and sterilizations [9,10,2]. This policy was relaxed in 1953, and contraceptives and abortions under certain conditions became available. In 1957, as part of the government's first birth control campaign, legal access to abortion was made easier [11].
The dramatic peak in fertility observed in 1962 led the government to change its policy and family planning in densely populated urban areas was promoted. The Cultural Revolution halted this program, but in 1973 the first antinatalistic population policy was launched: overpopulation was considered a threat to modernization and development. The 1973 policy used approaches such as education to individual women and couples, provision of models of reproductive behavior, and improvement of contraceptive and abortion services [12]. The policy promoted delayed marriage, long intervals between births, and a two-child family.
The 1973 policy was strengthened in 1975 defining population growth targets and introducing locally-managed collective birth plans [13]. In 1979 a one-child policy, which previously already had been experimented in some provinces, was introduced to the whole country. In addition to the requirement of one child per couple it set national targets (population less than 1.2 billion and population growth zero in 2000). A new separate vertical service structure (family planning, FP) was formed to implement the policy and monitor it; until 1979 family planning had been a part of health care and was organized at the local level. In 1981, the State Family Planning Commission (currently called the National Population and Family Planning Commission, the equivalent of a "Population Ministry") was formed to coordinate the activities. In 1982, family planning, aiming at reducing population growth, was included in the Constitution as a basic state policy [2]. Since 1981, the marriage law has explicitly required couples to practice family planning [14]. The family planning policy did not become a national law until 2001, and each province produced its own regulations on the basis of national guidelines [15].
The one-child family policy led to forced family planning and opposition. This and the economic reforms of the early 1980s, which reduced the state's control, led to a more permissive policy in 1984 allowing modifications at the provincial level, usually around the rules of having a second child. In the late 1980s, clear provincial rules were demanded. The new regulations [16-27] focused, in addition to higher order births, on preventing early-age marriage and childbirth, and on reducing the high rate of induced abortions [28].
In the 1990s, population policy stressed stability, and only small modifications in population targets at the provincial level were made. Contraception was promoted instead of abortions, and emphasis was placed on improving regular services instead of conducting separate campaigns. High contraceptive use rates were achieved. The importance of integrating population policy into general social and economic planning was acknowledged, and the question of an aging population became an important consideration [29].
Currently, different provinces have different rules [30,2]. The commonest provincial rule for rural areas is "two children if the first is a girl". But some provinces have adopted "two children with a four-year interval" [2]. In the case of minority groups, more children are allowed. For cultural reasons very few children have been born outside marriage, and most women marry [31] and have their first child shortly after the marriage. Contraceptives are now available free of charge and are commonly used.
Illegal Pregnancy
For a pregnancy to be "legal" (i.e. approved by secular authorities), the woman has to have permission for it from the population authorities. The family planning regulations and the 2001 law define legal (authorized) births, i.e. what is allowed [32], and "illegal" (unauthorized) pregnancy (or birth following such pregnancy) must be logically concluded as the opposite. The regulations stipulate incentives, but disincentives to avoid illegal pregnancies are defined at a lower level, mostly by county FP committees. Whether the pregnancy is legal depends on whether the pregnancy history of the parents qualifies them to have a (new) child, and whether the child quota laid down in the yearly plan of the village or workplace is full or not. The rules vary from one area and workplace to another. The process of determining legality may include the issuing of a formal certificate from the local family planning authority. The description from two urban areas [33] shows that getting the permission may be complicated and is not a formality.
The Chinese family planning regulations do not include the concept of an "illegal child" (i.e. a child born out of an "illegal pregnancy"), and the law prohibits discrimination against children born outside marriage [34]. However, children from illegal pregnancies may not be registered or treated equally until their parents pay the fines imposed as punishment. Especially in urban areas registration with the local authority is required for medical care, schooling and employment. The use of incentives and disincentives in the population policy implementation has varied by time and place, and it is likely their impact has declined over time due to the large social changes, such as growth of private housing in cities, shrinking of state sector enterprises and internal migration. One incentive from 1988 is the "one-child certificate" which is a contract between a couple and the local government. It gives parents who agree to have only one child certain economic rewards, such as a monthly stipend, free obstetric care, increased maternity leave, highest priority in education and health care for the child, preferential treatment when one is applying for housing, and a supplementary pension [15,30,35]. Disincentives for violators include losing housing and school benefits or having to pay higher fees and fines. Fines, currently called extra tax, may be substantial: according to one source, they amount to 10–20 % of a family's annual income; and according to another, they exceed the average annual income [35]. Qian [28] reports fines of 2.5 times the village's per capita income. Payment period and how actively the fines are collected have varied by place [15].
Enforcement of the population policy has been pursued via the Communist Party and the State Family Planning (FP) Commission, both of which have a functioning vertical structure. The FP structure extends downwards from the State FP Commission to province, city, district (urban) or county (rural), street/township, and finally to neighborhood FP / village FP committee. FP service centers at the county and township level have taken over some services previously provided by local health services [15]. The FP organization has good connections to women's voluntary organizations [35]. In the 1990s, there were an estimated 300 000 FP officials, and hundreds of thousands of part-time FP workers in villages, called "cadres" (lay-people working part time for local authorities) [2]. Parallel to the hierarchical structure of civil servants, there is the party organization, and the leaders of the party at each level share the decision making with leaders of the civil servant departments. Cadres are usually also members of the Communist Party.
Implementation of the family planning policy has been taken seriously and extensive and ongoing ideological education activities involving authorities and cadres were introduced. Education has been provided on both, general and individual level. Individual women are persuaded to comply with the rules, as care and control are provided by the same individuals [33]. Educational programs address the general good: the importance of the population policy for the society. Yearly population plans, including numbers of births and the organization of FP services, are developed by local authorities, and state and municipal workplaces. Data collection is organized by cadres. Villages in rural areas and workplaces in urban settings take collective responsibility, and sign a contract to make the plan for targets and enforcement [33,36].
In rural areas, an extensive system has been created at the village and district level to ensure constant surveillance of contraceptive use and pregnancy status of all married women at reproductive age. It is common for married women to be requested to visit an FP station every two or three months for pregnancy testing, allowing for early pregnancy detection.
In cities, family planning officials and cadres within workplaces have a central function. The surveillance of contraceptive use may be more common than surveillance of pregnancies, as fear of loosing a job may motivate women not to have an illegal pregnancy [33]. Alongside education, accessible services, incentives, disincentives, and the surveillance of all married women at reproductive age, a "personal responsibility system" was introduced in the early 1990s to guarantee that the local birth plan is followed [15,36,35]. Under this system, the heads of both the FP and Party committees at each level are personally responsible for ensuring that the targets are met and the rules are followed. If deviations occur, the officials may be subject to penalties, including the withholding of monetary bonuses, the loss of promotion and dismissal. This has resulted in false reporting and financial negotiations with an official's constituency to compensate for financial losses [15,36].
Abortions and sex ratio
Until 1953, induced abortion was available only on the basis of protecting a woman's or her existing children's health; in addition abortions had to be performed early and a physician's certificate was required [11]. In 1957, abortion was made available at a woman's request within the first 10 weeks of pregnancy and each woman was allowed one abortion per year. The 1979 abortion law abolished most restrictions, and set 28 weeks of gestation as the upper limit for legally performing abortions [37]. Some provinces have made their own laws stipulating the place and performer of the abortion [12].
Whether a woman ends up having an abortion depends on the strength of her and her family's wish to have a further child, and her acceptance of abortion and the population control norms. It seems that the population at large is permissive towards abortion [38] (Table 1). In our study, conducted in a rural setting in Anhui Province, the outcome of pregnancies varied considerably according to the legal status of the pregnancy [8].
Table 1 Recorded induced abortions in China, 1975–1999
Induced abortions1) Live births1) Ratio
1976 475 1853 0.26
1978 539 1745 0.31
1980 953 1779 0.54
1982 1242 2238 0.56
1984 889 2055 0.43
1986 1158 2384 0.49
1988 1268 2457 0.52
1990 1349 2391 0.56
1992 1042 1759 0.59
1994 947 2104 0.45
1996 883 2067 0.43
1998 738 1991 0.37
1) in 10 000
Source: Ref. 40.
The number of abortions increased rapidly after the introduction of the 1979 population policy. Some reports suggest that women may have been forced to have an abortion if an illegal pregnancy occurred [39]. Applying pressure to undergo sterilization was apparently even more common [40]. However, there are reports that forced abortions are uncommon nowadays [40-43]. The temporary concealment of pregnancies and children lessens the pressure of having an unwanted abortion. Hidden children are usually registered later, sometimes as immigrants.
An unnatural sex ratio at birth, i.e. more baby boys than girls, has been reported from China at least since the early 20th century. The excess of male babies declined considerably between 1936 and 1960 [44], and increased from the mid-1970s onwards [45,44,46]. The sex ratio at birth is higher in rural than in urban areas [46,47], and the variation by province is large [34,44,48]. It has been shown that the higher parity is, the higher is the sex ratio, especially if there are no previous sons [34,49,50]. The unnatural sex ratio at birth can be the result of sex-selective abortions, hiding and/or not reporting baby girls, or of extra deaths among newborn girls [8,34,33,38]. Previously hiding of baby girls might have been the main cause for the high sex ratio at birth, as indicated by a higher numbers of girls at later ages as compared to births. However, as shown in our study in one rural county in Anhui province [our unpublished data], sex selective abortions may have become an important reason for the high sex ratio at birth.
Fetal sex determination is illegal in China and health professionals performing it are penalized [34,38]. The first regulation against fetal sex determination was issued by the national health ministry in 1989, and the content repeated in the 1995 law on maternal and child health care. The 2002 FP law explicitly refers to the ban of using ultrasound and other technologies to determine fetal sex. Abortions on the basis of a child's sex are illegal, and if detected the parents' right to have a second child is lost. Ultrasound scanning is now widely available [34,38], and the various regulations against its use for fetal sex determination indirectly hint that such use has occurred. Many other Asian countries and regions, including South Korea, Taiwan, Hong Kong, Singapore and Thailand have experienced a remarkable decline in fertility from the 1950s to 1990 [see e.g. [51,9]]. With the exception of Singapore, these countries had no other explicit population policies except for the promotion and wider provision of family planning services (Table 2).
Table 2 Comparison of sex ratios in China, Taiwan, and the South Korea
year Fertility (TFR) Sex ratio
China Taiwan South Korea China Taiwan South Korea
1981 2.6 .. .. 107 107 107
1982 2.9 .. 2.7 107 107 107
1983 2.4 2.2 .. 108 107 108
1984 2.4 .. 2.1 109 107 109
1985 2.2 .. .. 111 107 110
1986 2.4 1.7 .. 112 107 112
1987 2.6 1.7 1.6 111 108 109
1988 2.5 1.9 1.6 108 108 114
1989 2.4 1.7 .. 111 109 112
1990 2.3 1.8 1.6 115 110 117
1991 2.2 1.7 .. 116 110 113
1992 2.0 ---- .. 114 ----- 114
1995 1.8 ---- 1.7 117 ----- 113
2000 1.71) ---- 1.5 117 ----- 110
Source: Ref. 66, 67, 68, 69.
1) own estimate from the 2000 census.
Discussion
The concepts of illegal pregnancy and illegal birth do not exist in the Western literature. Pregnancies may be unwanted or not socially acceptable, or they may not be considered suitable for a particular woman's role, but they have the same legal status as other pregnancies. Even pregnancies starting with an illegal act, e.g. sex involving under-aged women or rape, are legal. In China, the concepts of "illegal" (unauthorized) pregnancy and birth are not mentioned in official Chinese documents, and they have to be deduced from definitions concerning what is "legal" (authorized). These concepts have not been analyzed in the English-language literature, where euphemisms such as "unplanned pregnancy" have been used to describe the Chinese situation. This causes confusion, because planning of pregnancies and births is understood in the West as being a private matter, while in Chinese literature planned pregnancies and births refer to those "approved by family planning authorities".
Previously, in Western countries, the concept of an "illegal" or "illegitimate" child was used. In older Finnish statistics, children born out of wedlock were classified as "illegitimate" [52-61]. We searched for literature from Finland to see whether some pregnancies and births had been considered illegal. This was not the case, even though sex before marriage or with a person other than one's legal partner had been a criminal offense from the end of the 17th century until 1926 (premarital sex) and 1948 (extramarital sex). Pregnancy and a child born as the result of a criminal act were protected, and abortion and infanticide were strictly condemned. "Illegitimate" children had otherwise the same legal rights than other children, but only in 1975 they were granted the same paternal inheritance rights as other children.
In the 20th century in Finland, the regulation of pregnancy in terms of who may be pregnant, dealt with the presumed health (physically or mentally) of the child to be born. The earlier methods to prevent women with certain handicaps or assumed inheritable diseases from conceiving, such as restrictions on marriage, were replaced later on by sterilizations and abortions [62]. With the new techniques for identifying fetal characteristics, interest in selective abortions has increased, but the modern fetal screening methods and selective abortions are voluntary, emphasizing the parents' freedom of choice.
Is Chinese population policy unique?
Our answer is no and yes. No, considering the fact that the distinction between illegality and strong social disapproval or lack of choice is not a major one, particularly among the poor. Both situations mean lack of choice, particularly if incentives not to have children and disincentives to have them are used. In countries in which religion is a strong social factor, from an individual's point of view the distinction between morals and laws may be hazy, and social and legal control may be hard to distinguish from one another. Population targets, widespread family planning information, easily accessible services and increased sex ratios at birth, are not unique to China. The increased sex ratio in China was not intended, and the comparison with some other Asian countries suggests that when fertility is low in societies favoring men, sex ratios are likely to be high [63].
But yes – China's population policy is distinctive in two ways. Firstly, it has created an efficient system for the regular monitoring of all married women's pregnancy status. This allows early intervention by means of abortion, and constant enforcement of contraception. Secondly, it was designed to apply to all women. Usually, when a society has used legal measures to prevent individual women and couples from having children, it has acted selectively. Grounds have included the assumed health of the future child or the childbearing capabilities of the parents, especially that of the mother. In Europe, the largest group of women to be excluded from childbearing used to be non-married women. Currently, the largest group is very young women.
Has the Chinese one-child policy been morally wrong?
The answer depends on one's view of who should decide on reproduction. Should it be freely decided upon by an individual / by a woman or whether a society or state should decide. The individualistic approach has been promoted to give women a choice. The proponents of individualism mainly live in countries in which large numbers of births have not been a societal problem. European and North American countries have had pronatalistic population policies, if any. In Western countries, terminating a pregnancy is legally controlled to a greater extent than being pregnant, although abortion laws vary considerably from one country to another [1].
When a society or its leaders feel there are too many people for the society to function well, moral judgment of actions undertaken to restrict the size of the population is influenced by the quality of implementation tools and the grounds for deciding that there are too many people. After 1979, the tools for restricting population size in China did not respect individuals, and at least in some places they were coercive. However, the worst forms, including forced abortions, are believed to have been rare [15]. The current national population policy explicitly prohibits coercion [40].
Many Westerners, with their cultural emphasis on gender equity, find sex-selective abortions disagreeable. To our knowledge there are no large-scale studies on Chinese views regarding sex-selective abortions. But a small study by Chu [38] found that even though rural women were undergoing such abortions, they did not think this was right and felt that it represented unfair treatment of girls.
Has the Chinese one-child policy been wise?
Because it is unknown what would have happened without it, it is difficult to judge. When comparing the development in China to that in South Korea in the 1970s, the comprehensive and resource-intensive one-child family population policy in China did not seem to have been fully successful – the decline in birth rates was not exceptionally large. The 1979 reform of population policy coincided with the rural economic reform, with its shift from collective to private farming ("household responsibility"). This reform may have partly contradicted the aim of restricting and monitoring population growth [51,64,2]. The rapid changes in population policy may have confused people and lessened public support. But the aim of ensuring equity in the policy may have softened the criticism: all women and couples are faced with equally harsh requirements. Nathansen's [33] study from an urban area showed that most women resigned to the one child policy, because they accepted the reasons for it, even though their personal wishes may have been different. However, variations in local application and fines have created geographical and social inequity. Yearly fluctuation and unbalanced growth in different decades and generations puts an extra burden on the provision of services such as schools and child care facilities, and social security in old age. The system of personal responsibility, under which meeting the family planning targets is one central achievement indicator for civil servants at different levels, has worsened the quality of population statistics, thus impairing monitoring and predictions.
Sex-specific abortions have been practical in helping families to have at least one son with fewer births. In the short term both, parents and family planning personnel have benefited from this. But in the future, the social implications of fewer women may be considerable in a culture that values marriage. Lack of women is likely to have an impact on prostitution and trafficking of women, and other social problems [50].
This article has mainly analyzed the situation in China as it was up to 2000. China is rapidly changing, and this also applies to its population policy. Since 1994, the population policy has been reoriented. The new official policy emphasizes on quality of services, respect for the individuals, the needs of citizens, and informed choice [65]. Reproductive health rather than population control is promoted, as well as cooperation between the family planning system and the health sector. The 2002 FP law prohibits discrimination of baby girls or their mothers. All over the world governments employ various mechanisms to regulate reproduction. Traditional methods of state control include marital laws, making infanticide illegal, protecting children from negligence, and regulating sterilizations and abortions in various ways. Reproduction is socially regulated through religion, morals, values, and laws. Modern technology, with its effective contraceptive methods, possibilities to see and study the fetus, and easier and safer abortion techniques, has made the regulation of childbearing technically easier. A comparison of China, where the state regulates who can be pregnant and, for example, Ireland, where the state regulates who can have an abortion shows how relative norms are in the area of reproduction [1,37]. The case of China leads us to believe that at the societal level there is no shared "world" view on the legal control of reproduction. Whether such a view exists in the context of the values and morals of individual women or families is not a topic of this paper.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
EH analyzed the data and had the main writing responsibility.
ZW collected the Chinese language literature, analyzed it and commented the drafts.
GC helped in describing the Chinese system, helped in location the relevant literature, and commented the drafts.
KV participated in interpretation the findings and writing the article.
All authors read and approved the final manuscript.
Acknowledgements
We thank Kaisa Rouvinen for her help in collecting Finnish data. This work was supported by grants from the Academy of Finland.
==== Refs
Rahman A Katzive L Henshaw SK A global review of laws on induced abortion, 1985–1997 International Family Planning Perspectives 1998 24 56 64
Xie ZM Peng X, Guo ZG Population policy and the family-planning programme The changing population of China 2000 Oxford: Blackwell 51 63
Cho LJ China's population: recent trends and future challenges NUPRI Research Paper Series No 50 1989 Tokyo: Nihon University, Population Research Institute
Hussain A Demographic transition in China and its implications World Development 2002 30 1823 34 10.1016/S0305-750X(02)00070-0
Lavely W First Impressions from the 2000 Census of China Population & Development Review 2001 27 755 768 10.1111/j.1728-4457.2001.00755.x
Murthi M Fertility change in Asia and Africa World Development 2002 30 1769 1778 10.1016/S0305-750X(02)00062-1
Tu P Peng X, Guo ZG Trends and regional differentials in fertility transition The changing population of China 2000 Oxford: Blackwell 22 33
Wu ZC Viisainen K Wang Y Hemminki E Perinatal mortality in rural China: retrospective cohort study British Medical Journal 2003 327 1319 1322 14656839
Jiang Z The fertility transition and prospects in China and other Eastern Asian countries Population Bulletin of the United Nations 1999, Nos 40/41 2000 New York: Department of Economic and Social Affairs, United Nations 197 211
Hardee K Xie ZM Gu BC Family planning and women's lives in rural China International Family Planning Perspectives 2003 30 68 76 15210405 10.1363/3006804
Rigdon SM Abortion law and practice in China: an overview with comparisons to the United States Social Science & Medicine 1996 42 543 560 8643980 10.1016/0277-9536(95)00173-5
Luo L Wu SZ Chen XQ Li MX Mundigo AI, Indriso C First-trimester induced abortion: A study of Sichuan Province, China Abortion in the developing world 1999 London: Zed Books 98 150
Kane P The Second Billion Population and Family Planning in China 1987 Maryborough, Vic: Penguin Books
Li VC Wong GC Qiu SH Cao FM Li PQ Sun JH Characteristics of women having abortions in China Social Science & Medicine 1990 31 445 453 2218624 10.1016/0277-9536(90)90039-U
Kaufman J Zhang ZR Qiao XJ Zhang Y Family planning policy and practice in China: A study of four rural counties Population and Development Review 1989 15 707 728
Anonymous Family planning legislation and policy in China China Population Today 1992 9 2 3
China. Ministry of public health. Circular on fetal sex testing and artificial insemination Annual Review of Population Law 1989 16 28
Anonymous Shanghai adopts family planning regulations China Population Newsletter 1990 7 10 11 12348767
China State Council Family Planning Measures 1988
China State family planning commission. Family planning demands Annual Review of Population Law 1989 16 7
China Fujian. Decision on family planning work of 7 March 1990 Daily Report China 1990 42 5 12285788
China Hunan. Circular on checking the second child trend, 1986 Annual Review of Population Law 1988 15 6 7
China. Guangdong. Family planning circular of 1990 Daily Report China 1990 46 12317499
China. Henan. Henan provincial rules and regulations on family planning Daily Report China 1990 36 42 12285789
China. Communist party central committee and China state council. Decision on strengthening the family planning programm for strict control of population growth China Population Today 1991 8 2 6
People's Congress of China. China maternal and child health care law Henan Zhengbao 2001 10 5 9 (In Chinese)
People's Congress of China. China population and family planning law Henan Zhengbao 2002 1 4 7 (In Chinese)
Qian ZC Progression to second birth in China: A study of four rural counties Population Studies 1997 51 221 227 10.1080/0032472031000149936
Qiao XC From decline of fertility to transition of age structure: ageing and its policy implications in China Genus 2001 57 57 81
Hesketh TZ Zhu WX Health in China: The one child family policy: the good, the bad, and the ugly British Medical Journal 1997 314 1685 1688 9193296
Zeng Y Peng X, Guo ZG Marriage patterns in contemporary China The changing population of China 2000 Oxford: Blackwell 91 93
Zhang CS Reading "Family planning law" Population and Family Planning 2002 2 14 (In Chinese)
Nathansen MC Accepting population control Urban Chinese women and the one-child family policy 1997 Richmond (U.K.): Curzon Press
Zeng Y Tu P Gu B Xu Y Li B Li Y Causes and implications of the recent increase in the reported sex ratio at birth in China Population and Development Review 1993 19 283 302
Doherty JP Norton EC Veney JE China's one-child policy: the economic choices and consequences faced by pregnant women Social Science & Medicine 2001 52 745 761 11218178 10.1016/S0277-9536(00)00175-1
Merli MG Raftery AE Are births underreported in rural China? Manipulation of statistical records in response to China's population policies Demography 2000 37 109 126 10748993
Department of Economic and Social Affairs: Abortion Policies: A Global review 2001 New York: United Nations 94 95
Chu J Prenatal sex determination and sex-selective abortion in rural central China Population and Development Review 2001 27 259 281 10.1111/j.1728-4457.2001.00259.x
Hardee-Cleaveland K Banister J Fertility policy and implementation in China, 1986–88 Population and Development Review 1988 14 245 286 11651940
Santana Cooney R Li J Sterilization and Financial Penalties Imposed on Registered Peasant Couples, Hebei Province, China Studies in Family Planning 2001 32 67 78 11326458 10.1111/j.1728-4465.2001.00067.x
Qiao XC Analysis of induced abortion of Chinese women Population Research No 3 Beijing 2002
Henshaw S Haas TA Berentsen K Carbone E Readings on induced abortion Volume 2: A world review 2000 2001 New York & Washington: The Alan Guttmacher Institute
Wu ZC Gao ES Ku XY Lu SY Wang MJ Hong WC Chow LP Induced abortion among unmarried women in Shanghai, China International Family Planning Perspectives 1992 18 51 53
Li Y Peng X Peng XZ, Guo ZG Age and sex structures The changing population of China 2000 Oxford: Blackwell 64 76
Coale AJ Banister J Five decades of missing females in China Demography 1994 31 459 479 7828766
Cao GY The future population of China: Prospects to 2045 by place of residence and by level of education Research Paper Series no 2 2001 Singapore: Asian MetaCentre for Population and Sustainable Development
CD of 2000 census 2003 Beijing: The All China Marketing Research Co., LTD
China statistical yearbook 2000 2000 Beijing: Statistical Press of China (In Chinese)
Hao HS Gao L Sex preference and its effects on fertility in China 1992 National Fertility and Family Planning Survey, China 1997 Atlanta: U.S. Department of Health & Human Services, Centers for Disease Control 59 98
Festini F de Martino M Twenty five years of the one child family policy in China Journal of Epidemiology and Community Health 2004 58 358 360 15082728 10.1136/jech.2003.017335
Gale Johnson D Effects of institutions and policies on rural population growth with application to China Population and Development Review 1994 20 503 530
Ala-Nikkola M Laiton lailliseksi ja kielletty sallituksi: tutkimus raskauden keskeyttämisen laillistamisesta Suomessa 1993 Tampere: Tampereen yliopisto
Miettinen T Se sokea syntinen rakkaus: aviottomuus yhteiskunnallisen ilmiönä ja elämänkohtalona Hauholla 1770–1815 1998 Tampere: Tampereen yliopisto, historiatieteen laitos
Nieminen A Taistelu sukupuolimoraalista 1951 Porvoo – Helsinki: WSOY
Pulma P Turpeinen O Suomen lastensuojelun historia 1987 Helsinki: Lastensuojelun Keskusliitto
Sirpi K Avioton äiti ja lapsi 1800-luvun maalaisyhteisössä 1994 Helsinki: Helsingin yliopisto, Historian laitos
Ruotsin valtakunnan vuoden 1734 laki: Naimisen Caari 1–10, 11–17
read 24.6.2002
Valtion säädöstietopankki Finlex read 24.6.2002
Laki 173/22 Suomen Asetuskokoelma. Helsinki 1922
Laki 24/26 Suomen Asetuskokoelma. Helsinki 1926
Hemminki E Rasimus A Forssas E Sterilizations in Finland: from eugenics to contraception Social Science & Medicine 1997 45 1875 1884 9447636 10.1016/S0277-9536(97)00126-3
Park CB Cho NH Consequences of son preference in a low-fertility society: imbalance of the sex ratio at birth in Korea Population and Development Review 1995 21 59 84
Goldstein A White M Goldstein S Migration, fertility, and state policy in Hubei Province, China Demography 1997 34 481 491 9545626
Baige Z Shaobo X Family planning and reproductive health in China International Journal of Gynecology and Obstetrics 2002 77 43 46 12053895 10.1016/S0020-7292(02)00011-5
Gu BC Comparing study on sex ratio in China, Taiwan, and South Korea Population Research, No 5 1996 Beijing: People's University (In Chinese)
The National Statistics Bureau: 1% of national population survey in 1995 1997 Beijing: Statistics Press
Shi CJ The determinants for change of sex ratio at birth in South Korea Population and Family Planning 2004 5 40 41 (In Chinese)
Li P Concerning on the sex ratio of birth 2005 China POPIN (China Population Information Center) (In Chinese)
|
16095526
|
PMC1215519
|
CC BY
|
2021-01-04 16:38:15
|
no
|
Reprod Health. 2005 Aug 11; 2:5
|
utf-8
|
Reprod Health
| 2,005 |
10.1186/1742-4755-2-5
|
oa_comm
|
==== Front
Respir ResRespiratory Research1465-99211465-993XBioMed Central London 1465-9921-6-781604276010.1186/1465-9921-6-78ReviewSerum biomarkers in interstitial lung diseases Tzouvelekis Argyris [email protected] George [email protected] Stavros [email protected] Demosthenes [email protected] Interstitial Lung Disease Unit, Royal Brompton Hospital, Imperial College, Faculty of Medicine, London, UK2 Department of Pneumonology, Medical School, Democritus University of Thrace, Greece2005 21 7 2005 6 1 78 78 1 2 2005 21 7 2005 Copyright © 2005 Tzouvelekis et al; licensee BioMed Central Ltd.2005Tzouvelekis et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
The use of biomarkers in medicine lies in their ability to detect disease and support diagnostic and therapeutic decisions. New research and novel understanding of the molecular basis of the disease reveals an abundance of exciting new biomarkers who present a promise for use in the everyday clinical practice. The past fifteen years have seen the emergence of numerous clinical applications of several new molecules as biologic markers in the research field relevant to interstitial lung diseases (translational research). The scope of this review is to summarize the current state of knowledge about serum biomarkers in interstitial lung diseases and their potential value as prognostic and diagnostic tools and present some of the future perspectives and challenges.
Serum biomarkersinterstitial lung diseasesidiopathic pulmonary fibrosissystemic sclerosissarcoidosisKL-6surfactant proteinscytokines
==== Body
Introduction
The use of biomarkers in medicine lies in their ability to detect disease and support diagnostic and therapeutic decisions. New research and novel understanding of the molecular basis of the disease reveals an abundance of exciting new biomarkers who present a promise for use in the everyday clinical practice. The initial evaluation of a serum biomarker concerns its expression in patients with the disease and in normal individuals in order to define sensitivity and specificity. The sensitivity of a test is defined as the proportion of patients with disease having a positive test whereas the specificity is the proportion of patients without the disease who have a negative or normal test (Table 1). Consequently the serum level of an ideal marker should: 1) increase pathologically in the presence of the disease (high sensitivity), 2) not increase in the absence of the disease (high specificity), 3) add information about the risk or prognosis 4) change in accordance with the clinical evolution, reflecting the current status of disease, or better 5) anticipate clinical changes, i.e. indicating the presence of relapse before it becomes obvious at a clinical level and finally 6) relate to disease burden and extent 7) be reproducible (as determined by the low coefficient of variation), 8) be of easy and cheap determination [1,2].
Table 1 Definition of sensitivity, specificity, and predictive values of a discrete test for the presence of a disease
Sensitivity, Specificity, and Predictive Value of a Discrete Test for the Presence of a Disease
Disease Present Disease Absent
Positive Test a (true positive) b (false positive)
Negative Test c (false negative) d (true negative)
Very few markers present a threshold at which the risk suddenly rises. The interplay between sensitivity and specificity and the nature of the disease under prediction assigns suitable cut-off points. Sensitivity and specificity calculated at various cut-off points give rise to a receiver-operating-characteristic (ROC) curve [2]. A clinically useful biomarker will be one with the largest area under the ROC curve. A number of novel blood biomarkers of lung disease including cytokines, enzymes, adhesion molecules, collagen relevant products and products of type II epithelial cells, have been studied for their clinical applicability.
The need for a biomarker in Interstitial Lung Diseases
In diffuse lung disease, there has been confusion for decades because of the term "disease activity" that variably means a) inflammation and responsiveness to therapy; and b) progressiveness. In diffuse lung disease, there are two separate therapeutic goals: a) a short term response as judged by improvements in pulmonary function test (PFT) and regression of imaging abnormalities as judged mainly by high resolution computed tomography (HRCT); and b) slowing or prevention of decline. It follows logically that a useful biomarker might conceivably predict either responsiveness (i.e. the inflammatory component) or the risk of progression of fibrotic disease. Therefore, the clinical utility of a biomarker is critically dependant upon whether clinicians have tools in diffuse lung disease that accurately predict: a) responsiveness and b) progression of disease. There is much less need to predict responsiveness than progressiveness. HRCT is reasonably accurate in the separation between a group of patients in which disease is clearly irreversible and a group of patients in whom responsiveness is reasonably likely [3]. Hunninghake et al. [4,5] reported that a confident HRCT pattern diagnostic of UIP can predict outcome as well as a histologic diagnosis of usual interstitial pneumonia (UIP). They also concluded that, when a HRCT pattern diagnostic of UIP is present, the extent of disease is the best predictor of mortality. However, the change in extent of disease is not yet confirmed to be predictive of survival. Moreover, even in cases in which HRCT is indeterminate in this regard the clinician has an answer very quickly with a trial of corticosteroids. By contrast, there is no reliable routine tool with which to predict the likelihood of progression of fibrotic disease with or without therapy. Even in the usually progressive disorder, (UIP), there is significant heterogeneity in the risk of progression and in other disorders. HRCT may be quite good at distinguishing between individual disorders but it does not have any track-record at all in predicting the rapidity of decline, within individual diseases. Moreover, the value of bronchoalveolar lavage (BAL) in assessment of prognosis and treatment in the strictly defined UIP subset is still an open question [6]. Thereby, clinicians need a biomarker with which to predict decline; a fibrogenetic biomarker.
Another key benefit a biomarker might provide is more accurate prognostic information from change in that biomarker, especially if the information can be obtained rapidly (i.e. short term change with treatment). The current serial tests including serial HRCT and PFT provide scattered statements regarding the prediction of long-term outcome. Flaherty et al. [7] stated that short term changes in baseline PFT are strongly predictive of long term survival in patients with well defined UIP and non-specific interstitial pneumonia (NSIP), whereas serial changes in HRCT were of limited value. In addition, Latsi et al. [8] have recently demonstrated that serial PFT is probably the best so far for this role in UIP/ NSIP patients (predicts survival and longitudinal behaviour of the disease more accurately than baseline variables) but represents only an approximate guide and you have to wait for months for the information.
Requirements for a biomarker in Interstitial Lung Diseases
On the contrary to the listed theoretical requirements reflecting the profile of an ideal marker, a useful biomarker in the real world of clinical practice, simply needs to fulfil an unmet need. If current non-invasive and semi-invasive tools predict outcome very accurately, a biomarker will add little extra information of value. Therefore, it simply needs to provide more accurate information about progressiveness than current modalities, not independent information. Is it mandatory for a biological predictor to add independent information about prognosis? Is it necessary for a clinician to score the scan or the factor in the lung function severity mathematically, every time a biomarker is to be measured? Additionally, this information should be transferable, since the definition of interstitial lung diseases (ILDs) is not yet constant and therefore test accuracy may vary considerably from one setting to another [9].
Regarding serial monitoring, whether or not a biomarker is highly reproducible is less crucial than clear quantification of the reproducibility, in order to distinguish between true change and measurement "noise". A marker might be poorly reproducible and yet highly accurate when it does clearly change. Furthermore, a poorly reproducible biomarker may be highly accurate when the range is categorized into for example a five – point semi quantitative scale, in order to address reproducibility issues.
Moreover, there is confusion between diagnostic and prognostic requirements. A diagnostic marker needs to be sensitive and specific [10]. However, a prognostic biomarker could, in principle, be normal when disease is present but stable. Alternatively, based on the conception that the best diagnosis is prognosis, the most useful diagnostic marker will be the one that separates patients against the disease-specific outcomes [11]. Additionally, relation to disease burden and extent should not represent prerequisites for a valuable biomarker. Clinicians have PFT and HRCT for this task. Many abnormalities are simply by-products of extensive fibrotic disease and are not informative about pathogenesis/progressiveness, per se. Failure to distinguish between severity and progressiveness has severely hindered clinical research in this field. Adding knowledge about progressiveness is one of the most fruitful applications. In addition, the ideal biomarker should not be confounded by severity ("current status"). The ideal marker should normalize in value when an effective treatment is found to prevent decline of extensive fibrotic disease, even though the "current status" of disease severity is unchanged.
In conclusion, the clinical need for a biomarker in diffuse lung disease regarding the presence or absence of the disease, the histospecific diagnosis and the prediction of responsiveness has probably little value. On the other hand, clinical necessity becomes invaluable regarding the baseline prediction of rate of progression and the early prediction of progression based on serial change with treatment. This will allow anti-inflammatory or other treatments to be evaluated or eventually modified before they have failed.
The scope of this review is based on the fact that although there are numerous published papers investigating the utility of biomarkers in the clinical research field, the number of review articles summarizing the current state of knowledge about the clinical applications of these molecules as diagnostic and prognostic tools in the research field relevant to common ILDs such as idiopathic pulmonary fibrosis (IPF), scleroderma, sarcoidosis, as well as other ILDs including radiation and drug- induced pneumonitis, pediatric ILDs and occupational and environmental diseases, still remains inadequately small.
Serum biomarkers in Interstitial Lung Diseases
Beyond other important functions, the lung epithelium produces complex secretions, including mucus blanket, surfactant proteins, as well as several proteins important for host defense [12].
Sampling the epithelial lining fluid by bronchoalveolar lavage (BAL) represents the common means of studying the proteins secreted by the lung epithelium and investigating their alterations in lung disorders [13]. However, the past fifteen years, scientists led by pioneering studies [14] that showed the presence of these proteins in the bloodstream as well, even though in small amounts, demonstrated using enzyme-linked immunosorbent assays (ELISA) significant variations of these proteins' levels in the serum of patients with different ILDs. The latter suggests that their assay might represent a novel approach in the assessment of lung diseases with still elusive pathogenesis, prognosis, diagnosis and therapeutic interventions. Because these proteins are mainly, if not exclusively secreted within the respiratory tract, their occurrence in the vascular compartment can be explained by several hypothetical mechanisms including [12]:
• Leakage from the lung into the bloodstream resulting from the increased permeability of lung vessels and the destruction of the barrier between alveolar epithelium and endothelium caused by injury to the basement membrane
• Increased production by the alveolar type II cells coupled with an increase in total type II cells per lung due to diffuse hyperplasia and
• Diminished clearance rates from the circulation.
In the present review article we have focused on recent publications concerning the most studied and interesting lung peripheral biologic markers in ILDs, namely: lung epithelium specific proteins (markers of epithelial damage), circulating cytokines estimating various types of inflammatory activity and finally enzymes and metabolites, products of epithelioid cells and derivatives of activated macrophages. (Table 2)
Table 2 List of studied serum biomarkers in ILDs
Lung epithelium-specific proteins Surfactant-associated proteins
• SP-A
• SP-D
Mucin-associated antigens
• KL-6/MUC1
Clara-cell protein
• CC16
Other lung epithelial markers
• CK-19
• Ca 19-9
• SLX
Cytokines and other serological parameters Chemokines and cytokines
• MCP-1
• MIP-1a
• ITAC/CXCL-11
• TNF
Anti-oxidant enzymes and collagen peptides
• Glutathione
• Type III procollagen peptide
Markers of T-cell activation
• sIL-2R
Markers of macrophage/monocyte activity
• ACE
• Neopterin
• b-glucuronidase
• LDH
Abbreviations: ACE: Angiotensin Converting Enzyme, Ca 19-9: Carbohydrate antigen Sialyl Lewis (a), CC16: Clara-cell protein 16, CK19: Cytokeratin fragment 19, CXCL-11: CXC chemokine 11, KL-6: Krebs von den Lungen-6, LDH: Lactate Dehydrogenase, MCP-1: Monocyte Chemoattractant Protein-1, MIP-1a: Monocyte Inflammatory Protein-1a, MUC: Mucin, ILDs: Interstitial Lung Diseases, ITAC: Interferon-inducible T cell-a chemoattractant, sIL-2R: soluble interleukin-2 receptor, SLX: Carbohydrate antigen Sialyl Lewis (x), TNF: Tumor Necrosis Factor
Lung epithelium-specific proteins
Surfactant-associated Proteins
Pulmonary surfactant is a complex and highly surface active material covering the alveolar space of the lung. Biochemically, surfactant is a molecular mixture composed mainly of structurally heterogeneous phospholipids. A major function of pulmonary surfactant is to reduce the surface tension at the air-liquid interface of the alveolus, thereby preventing alveolar collapse on expiration. It has also been demonstrated that the surfactant contains specific proteins [15]. Four surfactant-specific proteins with different structural and functional properties have so far been identified. They were named surfactant protein-(SP)-A, SP-B, SP-C and SP-D according to the chronologic order of their discovery [16] and have been divided in two distinctive groups, the low-molecular-weight hydrophobic SP-B and SP-C and the high-molecular-weight-hydrophilic SP-A and SP-D. The latter belong to the collectin subgroup of the C-type lectin superfamily and are produced by two types of non-ciliated epithelial cells in the peripheral airway, Clara cells and alveolar type II cells. Studies have demonstrated that these proteins play important roles in the innate immune system of the lung [17] and have been used as useful markers for confirming the diagnosis and evaluation of disease activity of various ILDs since they reflect the epithelial damage and turnover. SP-A has also been used as a marker for lung adenocarcinomas to differentiate lung adenocarcinomas from other types and metastatic cancers from other origins, and to detect metastasis of lung adenocarcinomas [18]. Even if the lung appears as the major site of their synthesis, their expression is not restricted to the respiratory tract but has been detected in several extrapulmonary tissues [19,20].
Mucin-associated Antigens
Mucins are major components of the mucus layer covering the airway epithelium. They consist of high-molecular-weight glycoproteins belonging to a broad family of mucin peptides [12]. Mucins are either associated with membranes or secreted at the surface of the respiratory tract [12]. Krebs von den Lungen-(KL)-6 is mainly associated with cellular membranes. It was initially described by Kohno et al. [14] as a high-molecular-weight glycoprotein and was classified as human MUC1 mucin. Immunohistochemistry has mainly detected KL-6 in alveolar type II and epithelial cells of the respiratory bronchioles. Although KL-6 is predominantly expressed by airway cells, however is not entirely lung specific, since it is also present on other somatic cells, such as pancreatic cells, eosophageal cells and fundic cells of the stomach [21]. Functionally KL-6 has been demonstrated to be inducible for the migration of human lung and skin fibroblasts, suggesting a potential role in the lung fibrogenic process [22]. Additionally KL-6 is a sensitive indicator of damage to alveolar type II cells, which strongly express this mucin at their surface. Type II pneumonocytes are regenerated over the alveolar basement membrane after the death of type I pneumonocytes over the first stage of lung injury. Therefore, its raise would theoretically represent the destruction of the normal lung parenchyma and architecture, the increased permeability of the air-blood barrier as long as the regenerating process as expressed by type II pneumonocytes' activity. Towards this direction KL-6 has been reported by several studies some of them cited in this review article as a sensitive marker for ILDs such as IPF, collagen vascular disease-associated interstitial pneumonia (CVD-IP), radiation pneumonitis, hypersensitivity pneumonitis and pulmonary sarcoidosis.
Clara Cell Protein (CC16)
The Clara Cell secretory protein-(CC)-16 is a low-molecular-weight protein of 16 kDa is secreted in large amounts into the lumen of the respiratory tract by nonciliated bronchiolar Clara cells in humans and rodents [23]. Immunohistochemical studies have shown that CC16 is not an entirely specific and exclusive product of Clara cells or even the lung [24,25]. The exact functions of CC16 are still elusive but there is increased knowledge that CC16 serves as an important immunosuppressive and anti-inflammatory mediator in the lung [12]. Additionally CC16 can inhibit production of interferon-γ (IFN-γ) by peripheral blood mononuclear cells [12]. Serum CC16 has been demonstrated to elevate in several conditions known to be related with an impairment of the air-blood barrier, including pulmonary fibrosis [26] or lung injury caused by firesmoke [27]. Thus, several group of scientists estimated CC16 blood levels to determine whether these are associated with the degree of lung involvement in ILDs and thereby can serve as a useful biomarker of the disease activity and severity.
Other lung-epithelial markers
Cytokeratin is a specific cytoskeletal structure expressed in epithelial cells, including bronchial epithelia [28,29]. Of the 27 known subunits of cytokeratin, cytokeratin fragment 19 (CK19) has been found soluble in serum and its levels have already been evaluated as a useful tumour marker for lung cancer [30]. Since it has been suggested that CK19 is released from injured bronchial epithelium [31], the past ten years it has been hypothesized and eventually indicated by several reports in the literature that CK19 serum levels are elevated in IPF and other ILDs and can be well correlated with the disease prognosis and diagnosis. Thus, in nowadays, there is an ongoing attempt to scrutinize the value of CK19 in evaluating the severity of lung injury as reflected by the increasing number of published papers investigating the role of this biologic marker in ILDs.
The cancer-associated antigens sialyl Lewis (a) (Ca 19-9) and sialyl Lewis (x) (SLX) are carbohydrate structures used as markers of cell differentiation and embryonic development [32]. It has been reported that a variety of these antigens is expressed on the cell surface during tumor progression [33]. Elevated circulating levels of these antigens have been associated with neoplastic transformation and metastases and therefore been used as tumor markers in the diagnosis of lung adenocarcinoma [34,35]. Serum levels of carbohydrate antigens have been found, however, also raised in some patients with non-malignant lung diseases such as IPF [36], tuberculosis [37] and diffuse parabronchiolitis [38] and moreover immunohistochemical analysis demonstrated their presence in the hyperplastic bronchiolar epithelium, on the surface epithelium cells and on exudates in air space. Repeated damage to the lungs may force these antigens into the blood circulation resulting to the elevated serum levels of these markers detected in these patients. Hence, it has been speculated that their elevation could mirror the extent of lung injury and serve as a valuable prognosticator of disease progression.
Cytokines and other serological parameters
A number of cytokines probing different aspects of the immunopathogenesis of ILDs have been tested for their clinical usefulness as serum biomarkers for monitoring disease severity and predicting response to treatment and therefore leading to early diagnosis of progressive disease and determination of therapeutic interventions.
Monocyte chemoattractant protein-(MCP)-1 and monocyte inflammatory protein- (MIP)-1a belong to the C-C subfamily of the chemokine family and appear to be important factors in the monocyte/macrophage-mediated inflammatory process in the ILDs [39,40]. It has been reported that epithelial cells, macrophages and vascular endothelial cells are the major-MCP-1-producing cells in IPF lung tissue [41]. Moreover, an elevation of MCP-1 and MIP-1a concentrations in BAL from patients with IPF and sarcoidosis has recently been demonstrated [40,42]. Tempted by the latter observation several group of scientists aimed to measure the levels of MCP-1 and MIP-1a in the serum and determine their clinical significance as a reliable and easily repeatable serological parameters for the differential diagnosis of ILDs and the monitoring of their clinical course.
IFN-inducible T cell-a chemoattractant-ITAC/CXCL-11 is a chemoattractant CXC chemokine with versatile properties including immunomodulatory, definsin-like antimicrobial, antiangiogenic and potentially antifibrotic activities. It has been reported to inhibit angiogenesis and thereby to reduce aberrant vascular remodeling resulting in diminished fibrosis [43,44]. Studies have demonstrated a direct association of this molecule with the pleiotropic cytokine IFN-γ1b [43,44] evidence that captured the interest of both clinicians and researchers to investigate the effects of this newly applied treatment on biologic markers such as ITAC/CXCL-11 associated with fibrosis, angiogenesis and immunomodulation in the plasma of patients with ILDs and ultimately reveal novel molecular pathways which support IFN-γ1b therapeutic utilities.
In addition, the activity and release of a large body of inflammatory mediators including tumor necrosis factor (TNF), antioxidant enzymes (glutathione), procollagen peptides (type III) and markers of cell damage such as lactate dehydrogenase (LDH) have been evaluated as prognostic and monitoring tools of the disease development, activity and progression in a variety of occupational and environmental studies.
Other delineated serological parameters that have been scrutinized for their clinical efficacy as serum biologic markers of the disease severity and the activation of several inflammatory cells contributing to the immunopathogenesis of the disease include markers of T-cell activation and markers for the evaluation of macrophage/monocyte activity.
Soluble IL-2 receptor (sIL-2R) represents a biomarker of the T-cell activation and can be found in BAL fluid and serum of sarcoidosis patients and it is released by activated alveolar immune cells. In addition to lymphocytes, activated sarcoid alveolar macrophages are capable of expressing increased numbers of sIL-2R upon activation [45]. There are several reports in the literature, some of them are reviewed here, that reveal an intimate relationship between this parameter and the clinical activity of the disease, providing further evidence for the close linkage between the course of sarcoidosis and the activated state of T-cells [45].
Parameters suitable for gauging the activity of the macrophage lineage in ILDs and specifically sarcoidosis and occupational diseases have been delineated and comprise the angiotensin converting enzyme (ACE) a product of epithelioid cells that reflect the granuloma burden of the entire body, neopterin a metabolite of the guanosinetriphosphate pathway that is released by activated macrophages and monocytes under the control of IFN-γ produced by T-cells [45] and b-glucuronidase a lysosomal enzyme, associated with increased phagocytic activity [46]. Serum and BALF levels of these molecules have been found elevated in patients with sarcoidosis and environmental lung disorders and have been used in the everyday clinical practice as markers for the clinical assessment and follow-up of granulomatous inflammation and dust-induced inflammatory response.
1. Serum biomarkers in Idiopathic Pulmonary Fibrosis and Collagen Vascular Disease – associated Interstitial Pneumonia (Tables 3 and 4)
Table 3 Studies measuring serum biomarkers in patients with IPF
Investigator Patients Controls Biomarker / Summary ROC curve analysis Cut-off values Specificity – Sensitivity Diagnostic accuracy Limitations
Kobayashi et al. 50 45
67 KL-6: a serum marker for interstitial pneumonia Yes
KL-6: 500–550 U/ml KL-6: 95 - 95% Non ILD-specific marker Potential influence by malignancies / Small number of patients
Takahashi et al. 56 52
108 Serum SP-A and SP-D as prognostic factors IPF and their relationship to disease extent Yes
SP-A: 45 ng/ml
SP-D: 110 ng/ml SP-A: 79 - 94%
SP-D: 85 - 95% Weak correlations with CT features / Inconclusive analysis of disease mortality / Reproducibility issues / Small number of patients
Yokoyama et al. 57 14 Circulating KL-6 predicts the outcome of rapidly progressive IPF No Not estimated Small number of patients / Outcome prediction with pretreatment levels / Non ILD-specific marker
Ishii et al. 61 49
9 High serum concentrations of SP-A in UIP compared with NSIP No Not estimated Small number of patients / Overlap of SP-A levels in UIP and NSIP / No ROC curve analysis / cut-off levels
Satoh et al. 69 41 Ca 19-9 serum levels as markers of disease activity in patients with fibrosing lung disease No Not estimated Small number of IPF patients / No serial measurement / No ROC curve analysis / cut-off levels
Takayama et al. 70 16 Elevated Ca 19-9 serum levels are well Correlated with the disease activity No Not estimated Small number of IPF patients / No serial measurement / No ROC curve analysis / cut-off levels
Yokoyama et al. 71 35
70 Superiority of KL-6 serum levels as a diagnostic marker of interstitial pneumonia Yes
KL-6: 449 U/ml
Ca 19-9: 26 U/ml
SLX: 41 U/ml KL-6: 74 - 91 - 99%
Ca 19-9: 43 - 77 - 94%
SLX: 20 - 71 - 96% Small number of IPF patients / Non ILD-specific markers / Not definitive cut-off values
Satoh et al. 72 23 Western blotting of serum SLX may serve as differentiator between lung adenocarcinoma and IPF Yes
SLX: 50 U/ml SLX: 93 - 83% Small number of patients / Low statistical power / Technical deficiencies
Suga et al. 74 86
10 Clinical significance of MCP-1 levels in BAL and serum in patients with ILDs No Not estimated No definitive relation with disease severity / Influence by steroid treatment / Potential influence by other lung disorders
Strieter et al. 75 32 Effects of IFN-γ-1b on biomarker expression in patients with IPF No Not estimated Pre- and post-treatment fluctuations / Not definitive results
Abbreviations: BAL: Bronchoalveolar lavage, Ca 19-9: Carbohydrate antigen Sialyl Lewis (a), CT: Computed Tomography, IFN-γ: Interferon-γ, ILD: Interstitial Lung Disease, IPF: Idiopathic Pulmonary Fibrosis, KL-6: Krebs von den Lungen-6, MCP-1: Monocyte Chemoattractant Protein-1, NSIP: Non-Specific Interstitial Pneumonia, ROC: Receiver Operating Characteristic, SLX: Carbohydrate antigen Sialyl Lewis (x), SP: Surfactant Protein, UIP: Usual Interstitial Pneumonia
Table 4 Studies measuring serum biomarkers in patients with IPF and CVD
Investigator Patients Controls Biomarker / Summary ROC curve analysis Cut-off values Specificity – Sensitivity Diagnostic accuracy Limitations
Ohnishi et al. 51 33
82 Comparative study of KL-6, SP-A, SP-D, MCP-1 as serum markers for ILDs Yes
KL-6: 465 U/ml
Sp-A: 48.2 ng/ml
Sp-D: 116 ng/ml
MCP-1: 1.080 ng/ml KL-6: 94 - 96- 96%
Sp-A: 86 -82 - 85%
Sp-D: 95- 70- 88%
MCP-1: 93- 52- 81% Small number of patients / Non ILD-specific markers / Potential influence by malignancies
Takahashi et al. 55 42
108 Serum levels of SP-A and SP-D are useful biomarkers for ILDs in patients with progressive SSc Yes
SP-A: 43.8 ng/ml
SP-D: 110 ng/ml SP-A: 33-100%
SP-D: 77- 100% Small number of patients
Greene et al. 58 427
95 Serum SP-A and SP-D as biomarkers in PF of different etiologies No Not estimated Evaluation of serial measurement not definitive
Yanaba et al. 59 39 Longitudinal analysis of serum KL-6 levels in patients with SSc: association with the activity of PF No Not estimated Small number of patients / Retrospective study / No ROC curve analysis / cut-off levels
Yanaba et al. 60 42 Comparative study of serum SP-D and KL-6 concentrations in patients with SSc as markers for monitoring the activity of PF No KL-6: 100 - 39%
Sp-D: 88 - 91% Small number of patients / Retrospective study
Fujita et al. 62 37
15 Elevation of CK19 serum levels in patients with IPF associated with CVD No Not estimated Small number of patients / Non ILD-specific marker
Dobashi et al. 64 27
10 Elevated serum and BAL CK19 levels in PF and AIP No Not estimated Small number of patients / Discrepancies with other serum parameters
Nakayama et al. 65 413
21 CK19 serum levels in patients with nonmalignant respiratory diseases Yes
CK19: 3.5 ng/ml CK19 : 30 - 50% Low specificity and sensitivity values / No adjustment with disease severity
Abbreviations: AIP: Acute Interstitial Pneumonia, BAL: Bronchoalveolar lavage, CK19: Cytokeratin fragment 19, CVD: Collagen Vascular Disease, ILD: Interstitial Lung Disease, IPF: Idiopathic Pulmonary Fibrosis, KL-6: Krebs von den Lungen-6, MCP-1: Monocyte Chemoattractant Protein-1, ROC: Receiver Operating Characteristic, SP: Surfactant Protein, SSc: Systemic Sclerosis
Idiopathic pulmonary fibrosis (IPF) is a refractory and lethal ILD characterized by fibroblast proliferation, extracellular matrix (ECM) deposition and progressive lung scarring. The incidence of IPF is estimated at 15–40 cases per 100.000 per year, and the mean survival from the time of diagnosis is 3–5 yr regardless of treatment [47]. On the other hand, scleroderma (progressive systemic sclerosis-SSc), is a systemic disease characterized by a progressive dermatologic abnormality. Systemic involvement may include among others, restrictive lung disease which develops in 30–60% of patients with scleroderma and progresses to severe restrictive lung disease and pulmonary fibrosis (a major cause of death in scleroderma) in 15% of these patients [48]. However, predicting the progression of IPF and SSc as well as their prognosis still remains elusive. To evaluate the activity and monitor the course of the disease HRCT, PFT, BAL and histologic features are clinically used [48]. Nonetheless, there are problems with the sensitivity, effort-dependability and mainly the ease of repetition of these examinations.
One of the pioneering studies that set the foundations for the development of a new research field with massive clinical implications was published ten years ago by Honda et al [49]. Authors were the first reported the potential usefulness of SP-D serum levels in reflecting the disease activity in a group of patients with different types of ILDs including IPF and collagen vascular disease – associated interstitial pneumonia (CVD-IP).
Moreover, KL-6 serological levels have been tested by Kobayashi et al. [50] whether they can reflect the activity of pneumonitis seen in different types of ILD and therefore used as a tool for the differential diagnosis of this large set of diseases and the assessment of their response to treatment. Data derived from this analysis utilizing ROC curves and cut-off values showed a distinct differentiation between ILDs and non-ILDs based on the peripheral KL-6 concentrations as well as a clear correlation of KL-6 levels with the clinical activity of the ILDs as defined by a series of conventional criteria. Although these results create major expectations in the diagnostic field of ILDs, however it should be noted that an elevation of KL-6 and SP-D serum levels suggests diagnosis of ILD and does not establish a specific diagnosis. For these markers to become truly specific, large evaluation and comparative studies are required to determine their usefulness in the differential diagnosis of ILDs.
The first and to best of our knowledge the only so far comparative study testing the sensitivity and specificity of three groups of molecules reported to serve as sensitive markers for ILDs was conducted by Ohnishi et al [51]. They generated ROC curves and performed a comparative analysis of the diagnostic values of KL-6, SP-A, SP-D and MCP-1 in a pool of serum samples derived from patients with IPF and CVD-IP. Intriguingly, this study demonstrated for the first time a clear superiority of KL-6 as a diagnostic marker of ILDs in terms of diagnostic accuracy, sensitivity, specificity and likelihood ratio. Interestingly, results from the statistical analysis confirmed that all four serum markers are specific markers for IPF and at least superior to previously described markers such as lactate dehydrogenase and other collagen products [52]. Although this study adds knowledge of high scientific rigidity about the status of these peripheral markers as diagnostic tools, however there are substantial weaknesses arising from the origin disadvantages of these molecules as specific biologic markers. More specifically, serum levels of all the four markers can be influenced by the presence of ILDs other than IPF as well as lung disorders other than ILDs such as malignancies, systemic inflammations and fibrosing lung infections [12,53,54]. Nonetheless, these observations are not to diminish their value as diagnostic tools but to illuminate the need for their further investigation in the context of more detailed studies searching potential correlation of these markers with the clinical and radiological findings and adding more specific information on the differential diagnosis of IPF among the other 200 members of the ILD group.
Towards this direction Takahashi et al. [55,56] were the first who attempted to prove a correlation between the SP-A and SP-D circulating concentrations and the radiologic findings as well as the disease extent in patients with CVD-IP and IPF. Authors published two really well done and heavily informative papers, both of them are reviewed here.
In the first study [55], results derived by a statistical analysis using ROC curves and cut-off levels demonstrated that elevated levels of SP-A and SP-D in patients with progressive SSc closely reflect in terms of sensitivity and specificity the presence of ILD on the basis of CT diagnosis. Moreover these markers increase even in patients with mild alveolitis not detectable by compatible chest X-ray. Additionally an interesting finding with major clinical applications pointed out by authors is that the combination assay of SP-D circulating concentrations and chest X-ray is almost equivalent to CT in the detection of ILD and can potentially serve as a contributor to reduce the risk of clinicians overlooking ILD complicated by progressive SSc.
Furthermore, Takahashi et al. [56] estimated the serological levels of SP-A and SP-D in IPF patients and correlated them with the radiological and functional parameters. Interestingly, accumulated findings from this analysis demonstrated a positive correlation between serum levels of both markers and the extent of alveolitis whereas no correlation was observed with the progression of fibrosis. Moreover authors showed a statistically significant link between increased concentrations of SP-D and rapid decline in pulmonary function tests as well as with higher rates of mortality. The above findings suggest that these proteins and especially SP-D may help scientists to shed further light in the pathologic characteristics of IPF seen in HRCT and assist as prognostic tool for the final outcome in applying optimal therapeutic approaches. However, caveats that should be taken under consideration include weak correlations of SP-A and SP-D serum levels with the CT features and inconclusive analysis of the disease mortality and the biomarker serum levels.
Several biologic markers have been studied the past few years not only to evaluate the disease activity and for differential diagnosis of ILDs but also for prediction of the disease outcome and the effectiveness of the commonly applied treatment during the accelerated phase of pulmonary fibrosis.
One of the most detailed and informative study in this field was published by Yokoyama et al [57]. Scrutinizing for early predictive markers of the therapeutic effects of high-dose corticosteroids on patients with rapidly deteriorating IPF, authors demonstrated that circulating levels of KL-6 could predict the efficacy of corticosteroids at an earlier time-point than other studied non-specific markers, when overall clinical effect may not yet be evident. However, potential criticisms of this study including the small number of a specific group of IPF patients studied, the finding that the KL-6 levels in the pretreatment period could not be used to predict outcome as well as the evidence that KL-6 concentrations are affected by certain malignancies [12,53] pose major limitations to these observations and highlight the necessity for further investigation and studies.
Currently, the hurdle faced by young physicians that limits the application of blood biomarkers in the daily clinical practice of IPF patients is whether alterations of their circulating concentrations are specific for IPF or they also reflect other lung parenchyma abnormalities seen in different ILDs. Several studies have tested the efficacy of blood biomarkers in differentiating IPF from other ILDs. Some of them are reviewed here.
One of the first and most extensive studies addressing that important issue was conducted by Greene et al. [58] in the context of a large cohort multivariate analysis including over 200 patients with IPF and progressive SSc and approximately 200 patients with other ILDs. Notably, authors showed an important correlation between elevated SP-A and SP-D serological levels in IPF and SSc patients compared to patients with sarcoidosis or beryllium disease. Furthermore, SP-D levels found to be strongly related to radiographic abnormalities in patients with IPF. The most intriguing result of this study was that both increasing SP-A and SP-D levels were highly predictive of mortality in patients with IPF and progressive SSc and this finding robust (especially SP-D levels) after adjustment for many measures of disease severity. However, a major limitation of this analysis was that serial measurement of SP-A and SP-D levels during the clinical course of patients with diffuse lung disease was not definitively evaluated. Hence, an important issue highlighted by authors that needs further investigation is the usefulness of serum alterations in reflecting the disease activity and extent during different time-points of the disease course.
To this end Yanaba et al. [59] carried out a longitudinal retrospective study in a relatively small number of subjects with SSc and found that the majority of patients with normal baseline serum KL-6 levels exhibited no deterioration or new onset of PF whereas patients with dramatically increased KL-6 levels showed a parallel progression of PF.
The latter observation concerning the utility of KL-6 as a biomarker in monitoring the clinical course of PF was further strengthened by the same group of scientists {Yanaba et al. [60]} in the context of a comparative study of SP-D and KL-6 serum levels in a limited number of patients with SSc. Moreover, by generating ROC curve analysis, authors concluded that the combined use of these two blood markers could be of higher diagnostic and monitoring value for the activity of PF than the single use of each marker. However, data derived from the last two studies was inconclusive mainly due to the fact that it was produced by retrospective analyses and the number of patients was inadequately small for any meaningful outcome.
Alternatively, Ishii et al. [61] estimated the diagnostic values of five lung peripheral biomarkers including SP-A, SP-D, KL-6 and two tumour markers in discriminating patients with ILD of various histopathologic patterns such as usual interstitial pneumonia (UIP) and non-specific interstitial pneumonia (NSIP). Ultimately authors demonstrated differential SP-A serum concentrations between patients with UIP and NSIP indicating a potential role of this protein for the discrimination of these two members of ILD group with sometimes similar pathologic and radiologic characteristics but with totally different prognosis and response to treatment. Although this study adds critical information on the field of differential diagnosis of ILDs, however it is of high risk to adapt this evidence in the everyday clinical practice and make the differential diagnosis based only on non-invasive methods. It is of high importance to note that there are substantial weaknesses including the relatively small number of patients used and the overlap in the serum SP-A levels between UIP and NSIP patients that limit the further application of these results and illuminate the need for a more extensive analysis to determine its diagnostic cut-off levels.
To gain a more comprehensive understanding on the activity of lung epithelial cell damage and repair characterizing IPF and other ILDs and to clear out whether these are related to the disease prognosis, investigators studied the significance of other lung specific epithelial proteins initially reported as tumour markers, in reflecting these pathological abnormalities. Some of the most extensively reported markers were CK19 and carbohydrate antigens sialyl Lewis (Ca 19-9, SLX).
There have been several published papers reporting elevated levels of CK19 in the serum of patients with various types of ILD [62,63]. Nevertheless, the first report in the literature demonstrating elevated CK19 serum levels in patients with different types of ILDs and a significant association with several clinical parameters known to be strong predictors of the disease clinical course was made by Dobashi et al [64]. This data combined with results from immunohistochemical analysis indicate that CK19 is a marker of lung injury. However, it is inconclusive given the small sample of patients recruited and the major discrepancies between levels of CK19 and other serological predictive parameters. Therefore, the efficacy of CK19 in the differential diagnosis of IPF from other ILDs as well as in the prediction of the disease prognosis should be further evaluated.
Fueled by this prospect, Nakayama et al. [65] were the first that revealed a positive linkage between increased CK19 serum levels and low survival rate in patients with IPF and CVD-IP suggesting a potential role for this tumour marker in reflecting the severity of the lung injury and the disease prognosis. Additionally, this study clearly documented by using ROC curves and cut-off levels the ineffectiveness of this marker to differentiate IPF from other non-malignant respiratory diseases characterized by marked epithelial cell damage. This observation coupled with substantial weaknesses including the lack of adjustment of CK19 serum levels with the disease severity, allow us to make only speculations on the exact role of this biomarker as a prognostic factor and warrant consideration.
Elevated plasma levels of Ca 19-9 were first demonstrated by Bungo et al. [66] in 1988. In addition, Mukae et al. [67] reported two cases of interstitial pneumonitis with marked increase of carbohydrate antigens in serum and BALF, the level of which changed in accordance to their clinical course. Immunohistochemical analysis showed the expression of these antigens on bronchiolar epithelial cells and regenerating epithelial cells which covered the surface of fibrosing alveolar septi or remodeling septal structures finding that was further corroborated by Shimizu et al. [68] who stated elevated serum levels of cancer-associated antigens (Ca 19-9, SLX) primarily localized in the hyperplastic bronchiolar epithelium and the surface epithelium cells of microscopic honeycombing in two different patients with IPF. However, the first study that tested the clinical usefulness of elevated blood levels of Ca 19-9 in a series of patients with fibrosing lung disease was conducted by Satoh et al [69]. In consistency with previous reports in the literature {Takayama et al. [70]}, authors demonstrated a strong correlation of serum levels with the degree of disease activity indicating a potential role of these molecules as prognostic markers of disease severity and therapeutic response. Furthermore, Yokoyama et al. [71] carried out the first comparative evaluation of the diagnostic values of three serum carbohydrate antigens (KL-6, Ca 19-9, and SLX). They plotted ROC curves and demonstrated a clear superiority in terms of specificity, sensitivity and diagnostic accuracy of KL-6 plasma levels in discriminating interstitial pneumonia from alveolar pneumonia and healthy volunteers. In another study, Satoh et al. [72] coupled SLX serum cut-off levels with Western blotting analysis and documented that this combination is of high diagnostic power in differentiating IPF from lung adenocarcinoma.
Evidence presented give credence to the view that cancer-associated antigens could serve as valuable and reliable diagnostic and prognostic markers of ILDs. On the other hand, none of the studies provided us with established diagnostic cut-off values since they exhibited low statistical power and in most of them no adjustment with the disease severity was performed. Moreover, KL-6 has proven to have higher discriminative power than carbohydrate antigens. In addition, potential arguments include the lack of knowledge regarding the role of these markers in the pathogenetic pathway of IPF. So far, the only study addressed this crucial issue was published by Obayashi et al. [73] who reported a direct correlation of elevated BALF Ca 19-9 levels with the percentage of neutrophils suggesting a potential role of carbohydrate antigens as neutrophilic chemoattractants and subsequently as contributors in the process of lung injury. Further analyses in the context of large prospective studies are warranted to elucidate this role and establish their clinical usefulness as markers of disease severity and activity.
The past few years led by the same perspective idea to find a reliable and reproducible biologic marker for the diagnosis and prediction of ILDs, several group of scientists have tested the usefulness of a slew of cytokines to play that role.
One of these reports was published by Suga et al [74]. Authors concluded that elevated serum and BAL levels of MCP-1 could be useful to discriminate IPF among other members of the ILD group. Interestingly, in circulating MCP-1 concentrations of patients with IPF underwent steroid therapy no trend towards fall with treatment was noticed whereas a major fluctuation in pre-and post-treatment levels was readily identifiable. However, these results do not prove neither the reproducibility of this biormarker nor its relationship with the disease behaviour. The latter observations coupled with the evidence that MCP-1 levels are markedly influenced by corticosteroid treatment as well as other lung disorders since it is a mediator of inflammation also produced in other parts of body, limit its specificity in the differential diagnosis of ILDs and underline the need for re-evaluation of its predictive value.
One of the most exciting roles cytokines can play is their use as indicators of the therapeutic effects of several drugs against IPF providing investigators with useful knowledge on the potential mechanisms of these drugs in terms of physiology and molecular biology highlighting novel therapeutic targets.
The first study in humans to characterize the effects of IFN-γ1b on plasma and lung biomarkers, speculated to serve as critical mediators in the pathogenesis of IPF was recently published by Strieter et al [75]. Authors documented indefinite pre- and post-treatment alterations in the peripheral concentrations of several biologic markers associated with fibrosis, aberrant vascular modelling and immunomodulatory activity that can be used only for generating hypotheses for future research. However, the strongest finding in this study was the differential elevation of blood and BAL before and after treatment levels of ITAC/CXCL-11, a chemokine with multifunctional activities including antiangiogenic and antimicrobial. Inferentially, this study comprises evidence that mortality in patients with IPF could be potentially improved through the versatile protective properties of IFN-γ1b supporting its therapeutic utility (Tables 3 and 4).
2. Serum biomarkers in occupational and environmental diseases (Table 5)
Table 5 Studies measuring serum biomarkers in patients with occupational and environmental diseases
Investigator Patients Controls Biomarker / Summary ROC curve analysis Cut-off values Specificity – Sensitivity Diagnostic accuracy Limitations
Borm et al. 80 33
58 Glutathione levels reflect early inflammatory response but are not a predictive parameter for individual susceptibility in CWP No Not estimated Retrospective analysis / No serial measurement / Small number of patients / Scarce data on the interrelationships among antioxidant enzymes
Borm et al. 81 39
27 TNF is a marker of individual susceptibility to dust-induced lung fibrosis in coal-workers No Not estimated Retrospective analysis / Small number of patients / Indefinite data on the acquired or genetically controlled differences of TNF release
Schins et al. 83 104
29 A 5-year follow-up study reveals that TNF is a reliable prognosticator of CWP No Not estimated No linkage to disease behaviour / Indefinite data on the effect of exposure / No ROC curve analysis / cut-off levels
Schins et al. 84 104
29 Serum PIIIP has a poor value as a biomarker to screen for CWP No Not estimated Studied population: retired miners / Diverse exposure to coal-dust / Heterogeneity of disease-fibrotic activity / Indefinite data on the effect of exposure and time variation in serum PIIP
Cobben et al. 86 201
48 Serum LDH levels are elevated in coalminers but are not associated with clinical variables of disease severity No Not estimated Retrospective study / No serial measurements / No ROC curve analysis / cut-off levels
Cobben et al. 87 191
48 Serum b-glucuronidase levels may be a useful biomarker in monitoring lung inflammation following coal dust exposure No Not estimated Retrospective study / No serial measurements / No ROC curve analysis / cut-off levels
Takahashi et al. 88 35
237 Elevated serum KL-6 levels in patients with FLD Yes
KL-6: 410 - 450 U/ml KL-6: 80 - 73% Small number of patients / No serial measurements
Harris et al. 89 108
20 Elevated serum neopterin levels in CBD Yes
Neopterin: 1.27 ng/ml Neopterin: 88 - 58% No serial measurements / No correlation with clinical and/or radiological parameters
Maier et al. 90 31
34 Neopterin is a useful diagnostic tool in differentiating CBD from BeS Yes
Neopterin: 2.5 ng/ml Neopterin: 100 - 80% No serial measurements Poor prognostic value Indefinite cut-off levels
Abbreviations: BeS: Beryllium Sensitization, CBD: Chronic Beryllium Disease, CWP: Coal-Workers Pneumoconiosis, FLD: Farmer's Lung Disease, KL-6: Krebs von den Lungen-6, LDH: Lactate Dehydrogenase, PIIIP: type III procollagen peptide, ROC: Receiver Operating Characteristic, TNF: Tumor Necrosis Factor
The concept of biomarkers is extensively developed in the field of occupational and environmental medicine. Historically, in epidemiologic research, the link between exposure and disease was often without knowing the mechanism or intervening events. The past few years, the identification of lower levels of exposures and the effective clinical and public health management of high risk populations has drugged much of attention. Borm [76] highlighted the need for methods to monitor early adverse effects, exposure, and/or susceptibility of individual subjects due to occupational and environmental causes and Schulte and Perera [77] nicely reviewed the need of extending the use of biomarkers to population studies. The relative inability of the current modalities including PFTs, questionnaires and physical examinations, to detect early signs of occupationally and environmentally related adverse effects has prompted much interest in using biochemical, molecular and pathologic changes as indicators for respiratory diseases [78]. The conventional approach to validate a biomarker is to relate a critical effect to exposure or calculated dose. A positive outcome judged on the marker-effect relationship leads to nomination of the event as a biomarker. An essential requirement for a successful biomarker of effect or susceptibility is that it should identify from among all exposed individuals those most likely to become diseased. Fueled by this prospect, a framework of studies has been designed to utilize biomarkers of exposure, susceptibility and pathophysiological changes as part of the array of tools available to assess environmental disease.
Borm and co-workers [79,80] evaluated the clinical usefulness of antioxidant enzymes, (TNF) and serum type III procollagen peptide in studies of coal dust-induced lung disorders, a wide spectrum of diseases including chronic inflammation and progressive massive fibrosis (PMF). In the first case-control study [79] glutathione levels were decreased in early stage coal workers pneumoconiosis (CWP) but increased in patients with PMF indicating a potential role of anti-oxidant enzymes in the early detection of inflammatory response to mineral dust exposure. However, serum alterations of glutathione failed to predict individual susceptibility since they most likely reflect a consequence of the disease. With this aim in mind, authors carried out a second case-control study [80] where TNF was studied as a potential risk factor and concluded that TNF release from peripheral blood monocytes exposed coal mine dust was a marker of individual susceptibility to dust-induced lung fibrosis. The latter observation corroborated earlier findings [81,82] and supported the notion that TNF release from monocytes or TNF in plasma are not associated with actual or cumulative exposure. However, definitive proof must come from a prospective analysis in the context of carefully designed follow-up study. The only follow-up study that has been conducted so far {Schins et al. [83]} showed that the miners that had disease progression (PMF) during 5 years already had high levels of dust-induced monocyte TNF release at the beginning and no alterations during follow-up were noticed. This highly informative study excluded TNF as an exposure marker and suggested that TNF release is a constitutional marker of the disease prognosis and is not highly affected by the disease itself. In a fourth study, Schins and Borm [84] conducted a 5-year prospective analysis and estimated whether type III procollagen peptide could serve as a valuable predictor of the fibrotic lung disease progression, outcome or activity. Since no differences between procollagen serum levels in coal-miners and non-dust-exposed controls were observed, the investigators stated that type III procollagen peptide is not a reliable marker of early effect.
To characterize the nature and extent of coal dust induced airway injury there is a need for biomarkers useful to monitor exposure effects. The potential of many cell mediators as monitoring tools i.e. CC16 [85], antioxidants [79] and several cytokines [80] has been raised many times. Towards this direction Cobben et al. [86] studied the role of LDH (a marker of cell damage) as marker of lung tissue injury. Authors conducted the first human study describing a considerable elevation of this enzyme in a group of ex-coalminers and a further association with other clinical variables. Additionally, the same group of investigators [87] evaluated the role of b-glucuronidase as marker of phagocytic activity and reported increased plasma concentrations in a group of coal-miners after 20 years of exposure whereas no correlation with clinical parameters was observed. The latter data is in line with the hypothesis that LDH and b-glucuronidase activity are conceivable markers of disease activity. Potential criticisms include the retrospective analysis, the lack of serial measurement and link to disease behaviour. Thus, these results should be interpreted with care. To determine the significance of these biological indicators with regard to the development or progression of CWP a longitudinal prospective design is necessary.
Hypersensitivity pneumonitis also called extrinsic allergic alveolitis is an ILD that may be due to a wide variety of inhalative antigenic stimuli. To date, the conventional diagnosis of hypersensitivity pneumonitis is usually made on the basis of history of periodically recurring or permanent complaints upon exposure to a specific inhalative antigen, interstitial abnormalities in both lungs by chest radiography or HRCT and detection of precipitating antibodies. However, at present, there are few diagnostic procedures available to confirm the diagnosis and to estimate the disease activity and most of them are too invasive and costly for widespread and daily use. With this aim in mind, Takahashi et al. [88] evaluated serum KL-6 measurement as a biologic marker for farmer's lung disease (FLD), a type of hypersensitivity pneumonitis caused by the inhalation of moldy antigens. Authors conducted a large cohort study and clearly documented significant higher blood KL-6 levels in patients with FLD compared to the levels of farmers with or without precipitating antibodies. In addition, major findings in this study previously unknown, include the consistency between KL-6 serum levels in FLD patients and the activity of the disease as well as the indication that elevated KL-6 concentrations coupled with conventional diagnostic criteria may detect subclinical FLD and determine early and effective therapeutic interventions. However, to validate whether serum KL-6 levels reflect the disease activity a larger evaluation of the specificity and sensitivity of this marker against the disease behaviour in combination with a definitively estimated serial measurement of its plasma levels are required.
While its direct role in pathogenesis of immune and inflammatory responses remains unclear, neopterin's ability to reflect monocyte and macrophage activation has been exploited and yet postulated as a marker of the beryllium-specific cell-mediated immune response that leads to a chronic granulomatous disease, a hypersensitivity disorder named chronic beryllium disease (CBD). In the study of Harris et al. [89] ROC curves were generated to evaluate the optimum diagnostic accuracy of neopterin's cut-off levels. Interestingly, authors have found that the combination of elevated neopterin's serum levels with the conventional screening test for CBD, beryllium lymphocyte proliferation test exhibited an optimized positive predictive value suggesting neopterin as a valuable biomarker in discriminating workers with CBD from these that are only sensitized to beryllium rendering lung biopsy unnecessary. These results were substantiated by in-vitro studies of peripheral blood mononuclear cells derived by patients with CBD or beryllium exposed workers [90] and an association between serum levels of neopterin and clinicolaboratory parameters of the disease severity was found. However, further confirmatory tests in large cohorts of patients including serial measurements and correlation of results with clinical and radiological findings are essential to determine reliable cut-off values for the diagnosis of the disease and the assessment of the progression likelihood (Table 5).
3. Serum biomarkers in other interstitial lung diseases (Table 6)
Table 6 Studies measuring serum biomarkers in patients with other ILDs
Investigator Patients Controls Biomarker / Summary ROC curve analysis Cut-off values Specificity – Sensitivity Diagnostic accuracy Limitations
Ohnishi et al. 92 30 Elevated circulating KL-6 levels in patients with drug induced pneumonitis No
KL-6: 520 U/ml Sensitivity: 53 – 89% Small number of patients / Discrepancies with CT features
Kohno et al. 93 15 Circulating antigen KL-6 and LDH for monitoring irradiated patients with lung cancer No Not estimated Small number of patients / Retrospective study / Evaluation of serial measurement not definitive
Goto et al. 94 16 Serum levels of KL-6 are useful biomarkers for severe radiation pneumonitis No Not estimated Small number of patients / Retrospective study / Chemotherapy influence
Takahashi et al.95 25 Diagnostic significance of SP-A and SP-D in sera from patients with radiation pneumonitis No Sp-A: 83 - 85%
Sp-D: 83 - 85% Small number of patients / Chemotherapy influence / No long term follow-up / Non ILD-specific markers
Al-Salmi et al. 97 10
10 Elevated serum KL-6 and SP-A and SP-D in pediatric ILDs No Not estimated Small number of patients / Diversity of the diseases studied / Poor correlation with functional and radiological parameters
Abbreviations: BAL: Bronchoalveolar lavage, CBD: Chronic Beryllium Disease CT: Computed Tomography, FLD: Farmer's Lung Disease, ILDs: Interstitial Lung Diseases IPF: Idiopathic Pulmonary Fibrosis, KL-6: Krebs von den Lungen-6, LDH: Lactate Dehydrogenase, ROC: Receiver Operating Characteristic, SP: Surfactant Protein
Since lung specific-epithelium proteins reflect the epithelial damage and turnover it has been hypothesized and ultimately demonstrated that they can be used as effective circulating markers for the diagnosis and prognosis of the clinical course of various types of interstitial pneumonitis including drug-associated, radiation-induced and hypersensitivity pneumonitis. Furthermore, pneumoproteins have already been introduced as potential valuable biomarkers in the research field of pediatric ILDs.
Numerous agents including cytotoxic and non-cytotoxic drugs exert pulmonary toxicity including interstitial pneumonitis which several times culminates to a fatal outcome [91]. Therefore early diagnosis is crucial, since withdrawal of the implicated drug is usually the most sufficient treatment for drug toxicities, whereas undiagnosed toxicity can be progressive and fatal.
Ohnishi et al. [92] in their attempt to introduce a specific and reproducible hallmark for the recognition of different types of drug-induced lung injury and the prediction of their clinical course, estimated the plasma concentrations of KL-6 in 30 patients with drug-associated pneumonitis classified into four different predominant radiographic patterns. The remarkable ascertainment of this study was the demonstration of a high sensitivity relation between elevated serum KL-6 levels and particular types of lung injury as well as with their clinical course. However, the small number of patients included in this study coupled with discrepancies between serum KL-6 levels and the disease extent as defined by CT findings comprise major caveats and illuminate the need for further prospective studies to determine whether measurement of KL-6 levels would be beneficial in the monitoring and early detection of drug-induced ILDs.
Radiation pneumonitis is the most common complication for thoracic tumours and is classified as an ILD. Since CT scanning that represents the gold standard diagnostic procedure for radiation pneumonitis is either not frequently repeatable or often exhibits non-specific findings hardly to be differentiated from the lung tumour manifestations, a lung specific laboratory test easily repeatable and reproducible is highly required for the early detection of radiation-associated lung injury. On the basis of this conception, Kohno et al. [93] were the first reported that serum KL-6 levels are a sensitive marker for detecting severe radiation pneumonitis. Nevertheless, because authors measured KL-6 serum levels at a few time-points in each patient's clinical course, data derived from this study is inconclusive and cannot be applied to firmly correlate circulating KL-6 concentrations with the clinical course of radiation pneumonitis. Therefore, Goto et al. [94] to further streamline these observations retrospectively monitored at shorter intervals blood KL-6 levels in patients with lung cancer who had received radiotherapy with or without chemotherapy and showed a correlation with the clinical course of radiation pneumonitis and the response to treatment. Interestingly, the finding that serum KL-6 levels in some patients were increased priory to the clinical and radiological diagnosis of radiation pneumonitis is particularly noteworthy and should be kept in mind.
Another study searching for potential indicators for the early detection of radiation pneumonitis and for monitoring its clinical course was conducted by Takahashi et al [95]. Remarkably almost all of the patients with radiation pneumonitis detected by HRCT exhibited significant increases in both SP-A and SP-D levels which showed high sensitivity and specificity for the early diagnosis of radiation pneumonitis compared with other conventional haematological laboratory indices. Additionally, an agreement of SP-A and SP-D plasma levels with the clinical response to steroid therapy was also noted. This data suggest a possible role of lung-specific epithelial proteins as a diagnostic tool for the recognition of radiation-induced lung injury even when its radiographic change is faint.
However, these studies exhibit major caveats that must be addressed prior to their further application in the routine clinical practice. Briefly, we report the small number of patients studied, the application of chemotherapy prior to radiotherapy, the absence of serial measurements through the clinical course of the disease and the evidence that these proteins (SP-A and KL-6) are neither organ nor disease specific since they have also been used as markers for lung adenocarcinoma [18,53]. To confirm the assumptions arising from these concerns, further studies are required.
As mentioned above the spectrum of ILDs includes a large heterogeneous group of disorders calculating over 200 members with IPF to be the most common form of idiopathic ILD. In contrast ILD in children occurs far less frequently and there are no dominant forms [96]. Although this research field is newly developed a recently published study by Al-Salmi et al. [97] revealed for the first time elevated serum levels of three candidate biomarkers of the disease activity and severity including KL-6, SP-A and SP-D in a group of children with ILDs of various histopathologic patterns. These results corroborate earlier findings by Kobayashi et al. [98] who reported elevated KL-6 serum levels in three children with ILD associated with dermatomyositis. Nonetheless it should be noted that these studies are deficient because of the small number of patients recruited and the diversity of the diseases studied. Furthermore, data described here is inconclusive and it ascertains poor correlation with the functional and the radiological parameters. More in depth analysis in a large cohort of patients is required for any meaningful outcome (Table 6).
4. Serum biomarkers in sarcoidosis (Table 7)
Table 7 Studies measuring serum biomarkers in patients with sarcoidosis
Investigator Patients Controls Biomarker / Summary ROC curve analysis Cut-off values Specificity / Sensitivity Diagnostic accuracy Limitations
Ziegenhagen et al.104 77
50 TNF-a release from alveolar macrophages and serum level of sIL-2R are prognostic markers for sarcoidosis patients No Not estimated No serial measurements / No ROC curve analysis / cut-off values
Ziegenhagen et al. 105 73
48 BAL and serological parameters reflect the severity of sarcoidosis No Not estimated
High predictive value for neopterin+ sIL-2R No serial measurements / No ROC curve analysis / cut-off values
Grutters et al. 106 47 Positive correlation between sIL-2R serum levels and the disease activity and severity in patients with sarcoidosis No Not estimated Small number of patients with serial measurement / No ROC curve analysis / cut-off values
Rothkrantz-Kos et al.109 144
282 Potential usefulness of inflammatory markers to monitor respiratory functional impairment in sarcoidosis Yes
sIL-2R: 750 U/ ml
sIL-2R: 1300 U/ml sIL-2R: 94 - 82%
sIL-2R: 82 - 94% Discrepancies between treated and untreated patients / No serial measurements / Retrospective study / No correlation with radiological findings
Hashimoto et al. 117 26 Correlation of plasma MCP-1 and MIP-1a levels with disease activity and clinical course of sarcoidosis patients No Not estimated Small number of patients / No ROC curve analysis / cut-off values / No determination of the cytokines' cellular sources
Iyonaga et al. 118 47
10 MIP-1 serum levels estimate the activity of granuloma formation in sarcoidosis No Not estimated Small number of patients / No ROC curve analysis / cut-off values / No correlation with radiological findings
Kobayashi et al. 119 47 Serum KL-6 for the evaluation of active pneumonitis in pulmonary sarcoidosis No Not estimated Small number of patients / No follow-up laboratory data
Hermans et al. 120 117
117 Serum CC16 levels, a marker of the integrity of the air-blood barrier in sarcoidosis No Not estimated Non ILD- specific marker / Potential influence by tobacco smoking / Poor discriminative value
Janssen et al. 122 79
38 Elevated serum CC16, KL-6, and SP-D levels reflect pulmonary disease severity and prognosis in sarcoidosis patients Yes
CC16: 12.7 ng/ml
KL-6: 223 U/ml
SP-D: 91.7 ng/ml CC16: 84 - 73- 73%
KL-6: 84 - 86 - 86%
SP-D: 84 - 66 - 66% Non ILD- specific markers / Retrospective study / No serial measurement
Abbreviations: CC16: Clara Cell protein 16, ILD: Interstitial Lung Disease, KL-6: Krebs von den Lungen-6, MCP-1: Monocyte Chemoattractant Protein-1 MIP-1a: Monocyte Inflammatory Protein-1a, ROC: Receiver Operating Characteristic, sIL-2R: soluble Interleukin-2 Receptor, SP: Surfactant Protein, TNF-a: Tumour Necrosis Factor-alpha
Sarcoidosis is a chronic systemic disorder characterized by the presence of non-caseating granulomas and accumulation of T-lymphocytes and macrophages in multiple organs [99]. The mechanisms leading to the persistent accumulation of inflammatory cells and maintain the alveolitis which may lead to irreversible organ damage are not fully understood. The classical parameters used in the management of pulmonary involvement are mainly radiological methods and PFTs which do not gauge alveolitis but rather pulmonary impairment. Consequently these parameters have been proved ineffective both for early diagnosis of the disease and prediction of the response to treatment. The last twenty years immunological studies performed with cells obtained by BAL have shed further light into the pathogenetic mechanisms of sarcoidosis [100,101] and formed the basis of concepts of its immunopathogenesis. To further streamline these concepts and ameliorate hardships generating from their clinical application, several serological parameters including cytokines, soluble cytokine receptors, enzymes and other serum components including lung epithelium-specific proteins were delineated to probe different aspects of inflammatory activity and therefore reflect the disease severity and help the early diagnosis of progressive disease [45]. Although levels of these markers are closely associated with the pathogenesis of the disease [102,103], however the availability of sufficient information as to which of them is valuable for assessment of the disease severity and the prediction of clinical deterioration still represents a bottleneck. From a clinical point of view it is even more important to know whether sarcoidosis is severe, rather than active. The latter observation represents the key evidence determining the initiation of treatment. Some of the most extensive and informative studies addressing this crucial issue are presented and criticized in this review article.
One of the first studies attempted to circumvent this problem was published by Ziegenhagen et al [104]. Authors scrutinized the efficacy of both BAL and serum parameters in indicating likelihood of progression in patients with sarcoidosis and revealed that almost half of the patients with no indications for steroid therapy who had elevated sIL-2R plasma levels experienced disease deterioration whereas none with normal values did. The aforementioned observation strongly suggests that this immune parameter could serve as a prognostic guide leading to the identification of patients with greater risk to relapse and maybe associated with clinical findings such as the disease course.
The effectiveness of sIL-2R in evaluating sarcoidosis severity even in the early stages of the disease was further confirmed by another study of the same group of scientists [105] who clearly reported significantly elevated sIL-2R concentrations in patients with progressive sarcoidosis. On the contrary, this evidence was followed by a surprising finding regarding the ACE serum concentrations which did not differ significantly between patients with stable or progressing disease indicating a poor predictive value of this biomarker in sarcoidosis.
In the study of Grutters et al. [106] authors evaluated the serological concentrations of sIL-2R as marker of disease activity, severity and prognosis in a well-defined group of patients with sarcoidosis. Data derived from this study partially consistent with other studies [107] asserts a positive correlation between sIL-2R serum levels and the disease activity and severity and furthermore claims a predictive value for this serum biomarker. However, this study was retrospective and major discrepancies between the latter findings and results derived from older studies [108] were also notable.
To establish a more reliable association between sIL-2R and severity of sarcoidosis Rothkrantz-Kos et al. [109] applied ROC curve analysis and estimated the potential prognostic and diagnostic value of four inflammatory markers to predict respiratory severity in a large number of sarcoidosis patients. Areas under curve indicated that sIL-2R showed the greatest ability to determine pulmonary severity as assessed by the highest sensitivity, specificity, positive and negative predictive values among the evaluated markers, including C-reactive protein, serum amyloid A and ACE. Nonetheless, discrepancies between serum levels of all markers in treated and untreated patients coupled with the lack of serial measurement through patients' clinical course pose major limitations to the reported data which is unable to determine the potential efficacy of the evaluated markers to reflect respiratory severity in sarcoidosis patients under treatment in general.
Accumulated evidence from the data presented reveal that the predictive values of serum ACE in sarcoid patients are still under consideration and findings derived from these studies are controversial concerning its usefulness in assessing the disease severity and predicting the clinical outcome. Although some investigators suggest that a rising ACE serum level can predict radiographic relapses of sarcoidosis [110] others have clearly demonstrated that ACE appears to be of poor prognostic value [111,112]. In the aforementioned studies [104,105,109] ACE does not yield prognostic information. However, several studies in the literature have addressed this issue and have reported an extraordinary high variability in serum ACE concentration in health, which is provoked by numerous factors. The most important factor, responsible for approximately 25% of this variation is a deletion/insertion polymorphism in intron 16 of the ACE gene [113-116]. The question whether genotype corrected normal values might be able to increase the reliability of ACE serological concentrations as a marker of disease activity in sarcoidosis remains to be determined by further studies.
The role of chemokines in monitoring the disease behaviour and in reflecting the activity of granuloma formation in sarcoidosis patients has been evaluated and postulated. Hashimoto et al. [117] were the first who carried out a longitudinal evaluation of plasma MCP-1 and MIP-1a levels in a rather small number of sarcoid subjects and demonstrated a closely relation between serum levels of the studied biomarkers and the clinical course of the disease as defined by radiological and laboratory parameters. However, further light should be shed onto the role of these chemokines as markers of the disease activity and severity as well as their cellular sources and their relationship with the granuloma formation.
To this end Iyonaga et al. [118] in parallel with the elevation of MCP-1 serum levels in sarcoid individuals and their linkage with radiographic abnormalities showed by both immunohistochemistry and in situ hybridization that the origin of the increased MCP-1 plasma levels were the macrophages peripheral to the granulomas suggesting a role for this molecule as an indicator in estimating the activity of granuloma formation and subsequently of the sarcoidosis activity.
Apart from their application as prognostic and diagnostic tools of IPF and other ILDs the utility of lung-epithelium specific proteins including KL-6, SP-D and CC16 as biological markers of the disease activity has been separately evaluated in sarcoidosis patients. To date, several studies have addressed this issue and much good work has been done towards this direction.
The potential usefulness of pneumoproteins and more specifically KL-6 as true marker in the assessment of sarcoidosis was initially suggested by Kobayashi et al [119]. Authors were the first demonstrating linkage between increased KL-6 plasma levels and consistent changes in conventional clinical and laboratory parameters for the evaluation of alveolitis activity in sarcoidosis patients suggesting a potential role of KL-6 as an indicator of the disease severity.
Moreover Hermans et al. [120] scrutinized the role of CC16 as a valuable tool for the non-invasive evaluation of the damage and consequently the integrity of the air-blood barrier associated with sarcoidosis, shedding further light into pathogenetic mechanisms underlying the disease severity. Among other findings authors demonstrated that CC16 serum levels are influenced by the pulmonary extent of the disease as defined by the radiological abnormalities indicating a potential role of this molecule as a non-invasive and easily reproducible parameter for the assessment of the disease severity. On the other hand, blood levels of this protein reflect not only the rate of entry into the circulation, but also the rate of clearance since it has been reported that CC16 is rapidly eliminated by glomerular filtration [121]. Authors addressed this issue and found that the increase occurred independently of the impaired renal function, attributing the elevation to the leakage of the protein into the bloodstream across the air-blood barrier. Further studies are warranted to establish this notion.
Although these markers have been studied separately [41,119,120] no study so far had performed a comparative analysis of their diagnostic and prognostic accuracy. Triggered by the latter perspective idea Janssen et al. [122] recently published a really well done and informative paper comparing the ability of these three markers to reflect pulmonary disease severity and prognosis in sarcoidosis patients. They performed a ROC curve analysis which revealed a better sensitivity for KL-6 in discriminating sarcoidosis patients, results consistent with previous findings [51]. However, as it pointed out from the authors KL-6 and CC16 are not ILD-specific markers [[26,53], and [121]] and evidence that supports their predictive value is inconclusive since the analysis was retrospective and no serial measurement was performed (Table 7).
Future challenges and limitations
ILDs represent a diverse group of lung diseases comprising over 200 different members that have been broadly classified into several categories [114]. IPF accounting for 46% of all ILD diagnoses in men and women [124] is rapidly becoming one of the most lethal non-malignant lung disorders in the Western world with incidence that overcomes 30 new cases per 100.000 persons per year and with mean survival 3–5 years regardless of treatment [47]. On the other hand several patients with sarcoidosis develop irreversible lung damage and pulmonary fibrosis which culminates to a fatal outcome. The past 15 years, their devastating incidence and clinical seriousness have stimulated innumerable research studies, with any possible approach (Figure 1). Currently, the most fruitful application is monitoring the disease activity and consequently the early detection of patients with increased likelihood of non-response to treatment and progressive disease. Nevertheless, there are problems with the sensitivity, effort-dependability and mainly the ease of repetition of the current modalities being used for this purpose, including BAL, pulmonary function tests and HRCT.
Figure 1 Diagram showing the number of published papers regarding serum biomarkers in patients with ILDs the last fifteen years.
Therefore, a major challenge would be the determination of a reliable serologic marker reflecting the disease behaviour before it becomes obvious in clinical level, easily reproducible and feasible to be measured serially, meaning to be acceptable by the patient. The early serial measurement of this biomarker may lead to an early detection of relapse and thereby allow anti-inflammatory and other treatments to be evaluated or eventually modified before they have failed. The latter components can potentially compile a clinician's "wish list". On the basis of this conception, an important number of serum markers either lung-specific proteins and cancer-associated antigens or cytokines and other serological parameters probing different aspects of the immunopathogenesis of ILDs has been delineated. Initial reports of these markers shed further light into the pathophysiology of ILDs, [75,107,108,119,120,122] improved diagnostic and prognostic capabilities [55-58,61,74,107-111] and ultimately, led to a better patient care. Although the applications of these markers in the clinical setting created major expectations, however the feeling of excitement comes in contrast with important deficiencies exhibited by the new methodologies including non-standardization techniques, lack of knowledge of reproducibility and link to disease behaviour. Moreover, most of the studies enrolled a limited number of patients, insufficient to extract any meaningful or statistically significant outcome. In addition, it should be noted that many of the caveats arising from these findings are generated by the origin disadvantages of the investigated serological parameters to serve as ILD-specific markers since the most of them are influenced by malignancies [12,34,35,53], inflammatory diseases [54], or other lung disorders [26,121]. Furthermore, although the majority of them showed a moderate to high clinical utility as detectors of lung disease and correlated relatively well with the disease severity, however they failed to serve as reliable prognosticators of decline or responsiveness to therapy whereas evidence derived from the evaluation of their histospecific diagnostic accuracy is scarce and limited [61] (Table 8). Additionally, the elevation of biomarkers designed to reflect disease severity is likely to mirror not only the increased rate of entry into the circulation across the damaged air-blood barrier, but also the decreased clearance resulting from significant renal dysfunction seen in ILD patients [120,121]. Finally, it is of high importance to note that unfortunately only the minority of the studies [42,51,55,56,65,109,122] clarified the effectiveness and the diagnostic accuracy of the biomarkers by applying ROC curve analysis which is essential to estimate the sensitivity and specificity of a marker and moreover to define its discriminative cut-off levels. The aforementioned observations coupled with the evidence that the pathogenesis or even the definitions of ILDs and mainly IPF still remain poorly understood and ambiguous and therefore the role of all these molecules associated with their pathogenetic mechanisms has not been yet well determined, may explain the several limitations in the clinical use of serum markers.
Table 8 Scoring of various clinical utilities for key biomarkers in ILDs
Clinical Utilities Key Biomarkers
KL-6 SP-A SP-D CC16 sIL-2R ACE TNFa
Detection of lung disease + + + + + + + NE + + + +
Histospecific diagnostic accuracy + / - + + / - NE NE + NE
Correlation with disease severity + + + + + + + + + + +
Prediction of response to therapy + + / - + / - NE NE + / - NE
Prediction of decline + + / - + + / - + + / - + / -
Abbreviations: ACE: Angiotensin Converting Enzyme, CC16: Clara-cell protein 16, KL-6: Krebs von den Lungen-6, LDH: Lactate Dehydrogenase, ILDs: Interstitial Lung Diseases, ILDs: Interstitial Lung Diseases, sIL-2R: soluble interleukin-2 receptor, SP: Surfactant Protein, TNF: Tumor Necrosis Factor NE: Not Evaluated, 0: No utility, + / -: Low, +: Moderate, + +: High Utility
Furthermore, the application of biological markers in the research field relevant to occupational and environmental lung disorders have raised some additional concerns. Most of the studies exhibited major limitations including the retrospective analysis, the heterogeneity of the studied population and the lack of serial measurements necessary for the adjustment to disease behaviour and the link to disease prognosis. The only so far follow-up study [83] identified cytokine release as a reliable prognosticator of the disease progression. For biomarkers to become a useful addition to the array of tools for investigating occupational and environmental lung diseases appropriate and carefully designed study populations together with interdisciplinary collaborations and financial supports are required. Nevertheless, ethical, legal and social issues of human experimentation that arise with the application of biomarkers need to be considered prior to conducting studies [78].
Collectively, these findings highlight the necessity for further validation prospective studies and the assessment of novel molecules [75] to serve as diagnostic and prognostic tools as well as markers of the disease activity and severity. The recent application of massive genome screening tools such as DNA microarrays in the respiratory research field [125] has led to an increase rate of discovery of genes involved in the disease initiation and progression. Thereby, the next challenge arising from the emerging of hundreds of candidate biological markers is the application of the genome discoveries in the clinical setting with the use of tissue microarrays [126] in order to establish their diagnostic, prognostic and therapeutic importance and lead to a better understanding of the biological characteristics of ILDs.
Conclusion
Currently, the application status for most of these biologic markers is still in its infancy and remains exploratory. Unfortunately, they do not yield indications for therapy or mark the end of the inflammatory process and their prognostic value still needs to be established. Although the majority of them have not yet lived up to the "great hype" that was generated, lung-epithelium specific proteins and mostly KL-6 and SP-D show the greatest promise in IPF and other ILDs whereas serum cytokines seem to be not ready for routine monitoring. As for sarcoidosis, despite the controversial aspects regarding the usefulness of ACE in reflecting the disease activity, it still remains the most reliable and well defined marker whereas sIL-2R represents one of the most fruitful applications. Finally, in occupational and environmental lung disorders the only so far follow-up study supports TNF release as a valuable predictor of the fibrotic disease progression, outcome and activity (Table 8).
Further prospective investigations, technical improvements and introduction of novel markers are warranted in order to elevate the association of serum biomarkers with the pathogenesis of ILDs in the same status as for tumour markers with lung cancer. Nonetheless, crossing the boundary from research to clinical application requires validation in multiple settings, experimental evidence supporting a pathophysiologic role, and ideally intervention trials showing that modification improves the outcome. The emergence of pioneering technologies including DNA and tissue microarrays which have already been applied with great success in the respiratory research field can help scientists to circumvent this problem and bridge this boundary. In the interim, these markers can be quite useful to supplement the clinical, radiological and physiological monitoring of the disease and identify high-risk patients who would benefit from aggressive management of established risk factors.
List of Abbreviations
Angiotensin converting enzyme (ACE)
Bronchoalveolar lavage (BAL)
Carbohydrate antigen sialyl Lewis a (Ca 19-9)
Chronic beryllium disease (CBD)
Clara Cell Protein (CC16)
Collagen vascular disease-associated interstitial pneumonia (CVD-IP)
Cytokeratin fragment 19 (CK19)
Coal Worker's Pneumoconiosis (CWP)
Enzyme-linked immunosorbent assays (ELISA)
Extracellular matrix (ECM)
Farmer's lung disease (FLD)
High resolution computed tomography (HRCT)
Idiopathic pulmonary fibrosis (IPF)
Interferon-γ (IFN-γ)
IFN-inducible T cell-a chemoattractant (ITAC)
Interstitial lung diseases (ILDs)
Krebs von den Lungen-(KL)-6
Lactate Dehydrogenase (LDH)
Monocyte chemoattractant protein-(MCP)-1
Monocyte inflammatory protein- (MIP)-1a
Non-specific interstitial pneumonia (NSIP)
Progressive Massive Fibrosis (PMF)
Pulmonary Function Test (PFT)
Receiver-operating-characteristic (ROC)
Sialyl Lewis x (SLX)
Soluble IL-2 receptor (sIL-2R)
Surfactant protein-(SP)
Systemic sclerosis (SSc)
Tumor Necrosis Factor (TNF)
Usual interstitial pneumonia (UIP)
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
AT and DB were involved with the study conception. AT, GK and SA performed the data acquisition and interpretation. AT prepared the manuscript. DB was involved in revising the article for important intellectual content. All authors read and approved the final manuscript.
Acknowledgements
The authors thank Katerina Antoniou (M.D) for her constructive comments. The authors are also grateful to Dr. Nikolaos Gouvas and Dr. Konstantinos Lasithiotakis for their valuable assistance in collecting the data and designing the figure.
==== Refs
Ferrigno D Buccheri G Biggi A Serum tumour markers in lung cancer: history, biology and clinical applications Eur Respir J 1994 7 186 97 8143820 10.1183/09031936.94.07010186
Manolio T Novel risk markers and clinical practice N Engl J Med 2003 349 1587 9 14573728 10.1056/NEJMp038136
Flaherty KR Thwaite EL Kazerooni EA Gross BH Toews GB Colby TV Travis WD Mumford JA Murray S Flint A Lynch JP 3rdMartinez FJ Radiological versus histological diagnosis in UIP and NSIP: survival implications Thorax 2003 58 143 8 12554898 10.1136/thorax.58.2.143
Hunninghake GW Zimmerman MB Schwartz DA King TE JrLynch J Hegele R Waldron J Colby T Muller N Lynch D Galvin J Gross B Hogg J Toews G Helmers R Cooper JA JrBaughman R Strange C Millard M Utility of a lung biopsy for the diagnosis of idiopathic pulmonary fibrosis Am J Respir Crit Care Med 2001 164 193 6 11463586
Hunninghake GW Lynch DA Galvin JR Gross BH Muller N Schwartz DA King TE JrLynch JP 3rdHegele R Waldron J Colby TV Hogg JC Radiologic findings are strongly associated with a pathologic diagnosis of usual interstitial pneumonia Chest 2003 124 1215 23 14555549 10.1378/chest.124.4.1215
Costabel U Guzman J Bronchoalveolar lavage in interstitial lung disease Curr Opin Pulm Med 2001 7 255 61 11584173 10.1097/00063198-200109000-00002
Flaherty KR Mumford JA Murray S Kazerooni EA Gross BH Colby TV Travis WD Flint A Toews GB Lynch JP 3rdMartinez FJ Prognostic implications of physiologic and radiographic changes in idiopathic interstitial pneumonia Am J Respir Crit Care Med 2003 168 543 8 12773329 10.1164/rccm.200209-1112OC
Latsi PI du Bois RM Nicholson AG Colby TV Bisirtzoglou D Nikolakopoulou A Veeraraghavan S Hansell DM Wells AU Fibrotic idiopathic interstitial pneumonia: the prognostic value of longitudinal functional trends Am J Respir Crit Care Med 2003 168 531 7 12791580 10.1164/rccm.200210-1245OC
Irwig L Bossuyt P Glasziou P Gatsonis C Lijmer J Designing studies to ensure that estimates of test accuracy are transferable BMJ 2002 324 669 71 11895830 10.1136/bmj.324.7338.669
Knottnerus JA van Weel C Muris JW Evaluation of diagnostic procedures BMJ 2002 324 477 80 11859054 10.1136/bmj.324.7335.477
Hansell DM Wells AU Towards complete and accurate reporting of studies of diagnostic accuracy: the STARD initiative Clin Radiol 2003 58 573 4 12887948 10.1016/S0009-9260(03)00259-9
Hermans C Bernard A Lung epithelium-specific proteins: characteristics and potential applications as markers: Am J Respir Crit Care Med 1999 159 646 78 9927386
Reynolds HY Newball HH Analysis of proteins and respiratory cells obtained from human lungs by bronchial lavage: J Lab Clin Med 1974 84 559 73 4412011
Kohno N Kyoizumi S Awaya Y Fukuhara H Yamakido M Akiyama M New serum indicator of interstitial pneumonitis activity. Sialylated carbohydrate antigen KL-6 Chest 1989 96 68 73 2661160
King RJ Klass DJ Gikas EG Clements JA Isolation of apoproteins from canine surface active material: Am J Physiol 1973 224 788 95 4633419
Possmayer F A proposed nomenclature for pulmonary surfactant-associated proteins: Am Rev Respir Dis 1988 138 990 8 3059887
Griese M Pulmonary surfactant in health and human lung diseases: state of the art: Eur Respir J 1999 13 1455 76 10445627
Kuroki Y Takahashi H Chiba H Akino T Surfactant proteins A and D: disease markers Biochim Biophys Acta 1998 1408 334 45 9813383
Rubio S Lacaze-Masmonteil T Chailley-Heu B Kahn A Bourbon JR Ducroc R Pulmonary surfactant protein A (SP-A) is expressed by epithelial cells of small and large intestine J Biol Chem 1995 270 12162 9 7744866 10.1074/jbc.270.20.12162
Fisher JH Mason R Expression of pulmonary surfactant protein D in rat gastric mucosa: Am J Respir Cell Mol Biol 1995 12 13 8 7811466
Kohno N Akiyama M Kyoizumi S Hakoda M Kobuke K Yamakido M Detection of soluble tumor-associated antigens in sera and effusions using novel monoclonal antibodies, KL-3 and KL-6, against lung adenocarcinoma Jpn J Clin Oncol 1988 18 203 16 3411786
Hirasawa Y Kohno N Yokoyama A Inoue Y Abe M Hiwada K KL-6, a human MUC1 mucin, is chemotactic for human fibroblasts Am J Respir Cell Mol Biol 1997 17 501 7 9376125
Bernard A Marchandise FX Depelchin S Lauwerys R Sibille Y Clara cell protein in serum and bronchoalveolar lavage Eur Respir J 1992 5 1231 8 1486970
Broers JL Jensen SM Travis WD Pass H Whitsett JA Singh G Katyal SL Gazdar AF Minna JD Linnoila RI Expression of surfactant associated protein-A and Clara cell 10 kilodalton mRNA in neoplastic and non-neoplastic human lung tissue as detected by in situ hybridization Lab Invest 1992 66 337 46 1311398
Peri A Cordella-Miele E Miele L Mukherjee AB Tissue-specific expression of the gene coding for human Clara cell 10-kD protein, a phospholipase A2-inhibitory protein J Clin Invest 1993 92 2099 109 8227325
Lesur O Bernard A Arsalane K Lauwerys R Begin R Cantin A Lane D Clara cell protein (CC-16) induces a phospholipase A2-mediated inhibition of fibroblast migration in vitro Am J Respir Crit Care Med 1995 152 290 7 7541278
Bernard A Hermans C Van Houte G Transient increase of serum Clara cell protein (CC16) after exposure to smoke Occup Environ Med 1997 54 63 5 9072037
Moll R Franke WW Schiller DL Geiger B Krepler R The catalog of human cytokeratins: patterns of expression in normal epithelia, tumors and cultured cells Cell 1982 31 11 24 6186379 10.1016/0092-8674(82)90400-7
Coulombe PA Kopan R Fuchs E Expression of keratin K14 in the epidermis and hair follicle: insights into complex programs of differentiation J Cell Biol 1989 109 2295 312 2478566 10.1083/jcb.109.5.2295
Pujol JL Grenier J Daures JP Daver A Pujol H Michel FB Serum fragment of cytokeratin subunit 19 measured by CYFRA 21-1 immunoradiometric assay as a marker of lung cancer Cancer Res 1993 53 61 6 7677981
Nakamura H Abe S Shibata Y Yuki H Suzuki H Saito H Sata M Kato S Tomoike H Elevated levels of cytokeratin 19 in the bronchoalveolar lavage fluid of patients with chronic airway inflammatory diseases – a specific marker for bronchial epithelial injury Am J Respir Crit Care Med 1997 155 1217 21 9105057
Childs RA Gregoriou M Scudder P Thorpe SJ Rees AR Feizi T Blood group-active carbohydrate chains on the receptor for epidermal growth factor of A431 cells EMBO J 1984 3 2227 33 6094178
Itzkowitz SH Yuan M Fukushi Y Lee H Shi ZR Zurawski V JrHakomori S Kim YS Immunohistochemical comparison of Lea, monosialosyl Lea (CA 19-9), and disialosyl Lea antigens in human colorectal and pancreatic tissues Cancer Res 1988 48 3834 42 3288336
Del Villano BC Brennan S Brock P Bucher C Liu V McClure M Rake B Space S Westrick B Schoemaker H Zurawski VR Jr Radioimmunometric assay for a monoclonal antibody-defined tumor marker, CA 19-9 Clin Chem 1983 29 549 52 6825270
Chia D Terasaki PI Suyama N Galton J Hirota M Katz D Use of monoclonal antibodies to sialylated Lewisx and sialylated Lewisa for serological tests of cancer Cancer Res 1985 45 435 7 3965149
Satoh H Kamma H Ogata T Yano H Ohtsuka M Hasegawa S Clinical significance of serum levels of a carbohydrate antigen, sialyl SSEA-1, in patients with fibrosing lung disease Am Rev Respir Dis 1991 144 1177 81 1683194
Ishiura Y Fujimura M Minami S Ueda A Iwata M Watanabe K Shinagawa T Yasui M Matsuda T Kitagawa M Increased CA19-9 level in serum and bronchoalveolar lavage fluid from a patient with pulmonary tuberculosis Nihon Kyobu Shikkan Gakkai Zasshi 1996 34 477 81 8691672
Mukae H Hirota M Kohno S Komori K Fukushima K Hiratani K Kadota J Hara K Elevation of tumor-associated carbohydrate antigens in patients with diffuse panbronchiolitis Am Rev Respir Dis 1993 148 744 51 8103656
Wolpe SD Davatelis G Sherry B Beutler B Hesse DG Nguyen HT Moldawer LL Nathan CF Lowry SF Cerami A Macrophages secrete a novel heparin-binding protein with inflammatory and neutrophil chemokinetic properties J Exp Med 1988 167 570 81 3279154 10.1084/jem.167.2.570
Car BD Meloni F Luisetti M Semenzato G Gialdroni-Grassi G Walz A Elevated IL-8 and MCP-1 in the bronchoalveolar lavage fluid of patients with idiopathic pulmonary fibrosis and pulmonary sarcoidosis Am J Respir Crit Care Med 1994 149 655 9 8118632
Smith RE Strieter RM Zhang K Phan SH Standiford TJ Lukacs NW Kunkel SL A role for C-C chemokines in fibrotic lung disease J Leukoc Biol 1995 57 782 7 7539030
Standiford TJ Rolfe MW Kunkel SL Lynch JP 3rdBurdick MD Gilbert AR Orringer MB Whyte RI Strieter RM Macrophage inflammatory protein-1 alpha expression in interstitial lung disease J Immunol 1993 151 2852 63 8360496
Cole AM Ganz T Liese AM Burdick MD Liu L Strieter RM Cutting edge: IFN-inducible ELR- CXC chemokines display defensin-like antimicrobial activity J Immunol 2001 167 623 7 11441062
Strieter RM Belperio JA Phillips RJ Dubinett SM Keane MP Strieter RM, Kunkel SL, Standiford TJ Chemokines in lung cancer Chemokines in the lung 2003 New York: Marcel Dekker 297 324
Muller-Quernheim J Serum markers for the staging of disease activity of sarcoidosis and other interstitial lung diseases of unknown etiology Sarcoidosis Vasc Diffuse Lung Dis 1998 15 22 37 9571999
Fishman WH The 1993 ISOBM Abbott Award Lecture. Isozymes, tumor markers and oncodevelopmental biology Tumour Biol 1995 16 394 402 7569686
Costabel U King TE International consensus statement on idiopathic pulmonary fibrosis Eur Respir J 2001 17 163 7 11334114 10.1183/09031936.01.17201630
Schwarz MI King TE Interstitial Lung Disease 2003 4 London: BC Decker 405
Honda Y Kuroki Y Matsuura E Nagae H Takahashi H Akino T Abe S Pulmonary surfactant protein D in sera and bronchoalveolar lavage fluids Am J Respir Crit Care Med 1995 152 1860 6 8520747
Kobayashi J Kitamura S KL-6: a serum marker for interstitial pneumonia Chest 1995 108 311 5 7634858
Ohnishi H Yokoyama A Kondo K Hamada H Abe M Nishimura K Hiwada K Kohno N Comparative study of KL-6, surfactant protein-A, surfactant protein-D, and monocyte chemoattractant protein-1 as serum markers for interstitial lung diseases Am J Respir Crit Care Med 2002 165 378 81 11818324
Kohno N Yokoyama A Hirasawa Y Kondo K Fujino S Abe M Hiwada K Comparative studies of circulating KL-6, type III procollagen N-terminal peptide and type IV collagen 7S in patients with interstitial pneumonitis and alveolar pneumonia Respir Med 1997 91 558 61 9415357 10.1016/S0954-6111(97)90090-1
Kohno N Akiyama M Kyoizumi S Hakoda M Kobuke K Yamakido M Detection of soluble tumor-associated antigens in sera and effusions using novel monoclonal antibodies, KL-3 and KL-6, against lung adenocarcinoma Jpn J Clin Oncol 1988 18 203 16 3411786
Inoue Y Nishimura K Shiode M Akutsu H Hamada H Fujioka S Fujino S Yokoyama A Kohno N Hiwada K Evaluation of serum KL-6 levels in patients with pulmonary tuberculosis Tuber Lung Dis 1995 76 230 3 7548906 10.1016/S0962-8479(05)80010-3
Takahashi H Kuroki Y Tanaka H Saito T Kurokawa K Chiba H Sagawa A Nagae H Abe S Serum levels of surfactant proteins A and D are useful biomarkers for interstitial lung disease in patients with progressive systemic sclerosis Am J Respir Crit Care Med 2000 162 258 63 10903251
Takahashi H Fujishima T Koba H Murakami S Kurokawa K Shibuya Y Shiratori M Kuroki Y Abe S Serum surfactant proteins A and D as prognostic factors in idiopathic pulmonary fibrosis and their relationship to disease extent Am J Respir Crit Care Med 2000 162 1109 14 10988138
Yokoyama A Kohno N Hamada H Sakatani M Ueda E Kondo K Hirasawa Y Hiwada K Circulating KL-6 predicts the outcome of rapidly progressive idiopathic pulmonary fibrosis Am J Respir Crit Care Med 1998 158 1680 4 9817725
Greene KE King TE JrKuroki Y Bucher-Bartelson B Hunninghake GW Newman LS Nagae H Mason RJ Serum surfactant proteins-A and -D as biomarkers in idiopathic pulmonary fibrosis Eur Respir J 2002 19 439 46 11936520 10.1183/09031936.02.00081102
Yanaba K Hasegawa M Hamaguchi Y Fujimoto M Takehara K Sato S Longitudinal analysis of serum KL-6 levels in patients with systemic sclerosis: association with the activity of pulmonary fibrosis Clin Exp Rheumatol 2003 21 429 36 12942693
Yanaba K Hasegawa M Takehara K Sato S Comparative study of serum surfactant protein-D and KL-6 concentrations in patients with systemic sclerosis as markers for monitoring the activity of pulmonary fibrosis J Rheumatol 2004 31 1112 20 15170923
Ishii H Mukae H Kadota J Kaida H Nagata T Abe K Matsukura S Kohno S High serum concentrations of surfactant protein A in usual interstitial pneumonia compared with non-specific interstitial pneumonia Thorax 2003 58 52 7 12511721 10.1136/thorax.58.1.52
Fujita J Dobashi N Ohtsuki Y Yamadori I Yoshinouchi T Kamei T Tokuda M Hojo S Okada H Takahara J Elevation of anti-cytokeratin 19 antibody in sera of the patients with idiopathic pulmonary fibrosis and pulmonary fibrosis associated with collagen vascular disorders Lung 1999 177 311 9 10467022
Matsubara Y Iwashita T Ishimatsu Y Shikuwa C Kadota J Kohno S Evaluation of CYFRA 21-1 and ProGRP in serum and bronchoalveolar lavage fluid of patients with benign lung disease Nihon Kokyuki Gakkai Zasshi 2000 38 659 64 11109801
Dobashi N Fujita J Ohtsuki Y Yamadori I Yoshinouchi T Kamei T Takahara J Elevated serum and BAL cytokeratin 19 fragment in pulmonary fibrosis and acute interstitial pneumonia Eur Respir J 1999 14 574 8 10543277 10.1034/j.1399-3003.1999.14c15.x
Nakayama M Satoh H Ishikawa H Fujiwara M Kamma H Ohtsuka M Sekizawa K Cytokeratin 19 fragment in patients with nonmalignant respiratory diseases Chest 2003 123 2001 6 12796181 10.1378/chest.123.6.2001
Bungo M Yamaji Y Futami H Shiotani T Irino S Okino T Ohmori M A case of idiopathic interstitial pneumonia with marked increase of serum CA-19-9 Nihon Kyobu Shikkan Gakkai Zasshi 1988 26 185 9 3165481
Mukae H Sakito O Oda H Senju R Fukushima K Hiratani K Kadota J Komori K Hirota M Hara K Two cases of interstitial pneumonitis with marked increase of tumor-associated carbohydrate antigens in serum Nihon Kyobu Shikkan Gakkai Zasshi 1991 29 611 7 1886302
Shimizu Y Hamada T Tanaka Y Sasaki A Nemoto T Colocalization of CA19-9 and KL-6 to epithelial cells in dilated bronchioles in a patient with idiopathic pulmonary fibrosis complicated by diffuse alveolar damage Respirology 2002 7 281 4 12153695 10.1046/j.1440-1843.2002.00391.x
Satoh H Kamma H Ogata T Yano H Ohtsuka M Hasegawa S Clinical significance of serum levels of a carbohydrate antigen, sialyl SSEA-1, in patients with fibrosing lung disease Am Rev Respir Dis 1991 144 1177 81 1683194
Takayama S Kataoka N Usui Y Inase N Natori Y Nakayama M Miura H Kimula Y CA 19-9 in patients with benign pulmonary diseases Nihon Kyobu Shikkan Gakkai Zasshi 1990 28 1326 31 2273661
Yokoyama A Kohno N Kondo K Ueda S Hirasawa Y Watanabe K Takada Y Hiwada K Comparative evaluation of sialylated carbohydrate antigens, KL-6, CA19-9 and SLX as serum markers for interstitial pneumonia Respirology 1998 3 199 202 9767620
Satoh H Ishikawa H Yamashita YT Ohtsuka M Sekizawa K Serum sialyl Lewis X-i antigen in lung adenocarcinoma and idiopathic pulmonary fibrosis Thorax 2002 57 263 6 11867833 10.1136/thorax.57.3.263
Obayashi Y Fujita J Nishiyama T Yoshinouchi T Kamei T Yamadori I Hojo S Ohtsuki Y Hirashima M Takahara J Role of carbohydrate antigens sialyl Lewis (a) (CA19-9) in bronchoalveolar lavage in patients with pulmonary fibrosis Respiration 2000 67 146 52 10773785 10.1159/000029502
Suga M Iyonaga K Ichiyasu H Saita N Yamasaki H Ando M Clinical significance of MCP-1 levels in BALF and serum in patients with interstitial lung diseases Eur Respir J 1999 14 376 82 10515417 10.1034/j.1399-3003.1999.14b23.x
Strieter RM Starko KM Enelow RI Noth I Valentine VG Idiopathic Pulmonary Fibrosis Biomarkers Study Group. Effects of interferon-gamma 1b on biomarker expression in patients with idiopathic pulmonary fibrosis Am J Respir Crit Care Med 2004 170 133 40 15044205 10.1164/rccm.200312-1670OC
Borm PJ Biological markers and occupational lung disease: mineral dust-induced respiratory disorders Exp Lung Res 1994 20 457 70 7843079
Schulte PA Perera FP Validation. In: Schulte PA, Perera FP, eds. Molecular epidemiology:principles and practices 1993 San Diego:Academic press 79 109
Schulte PA Use of biomarkers to investigate occupational and environmental lung disorders Chest 1996 109 9 12 8549226
Borm PJ Biological markers and occupational lung disease: mineral dust-induced respiratory disorders Exp Lung Res 1994 20 457 70 7843079
Borm PJ Meijers JM Swaen GM Molecular epidemiology of coal worker's pneumoconiosis: application to risk assessment of oxidant and monokine generation by mineral dusts Exp Lung Res 1990 16 57 71 2307148
Borm PJ Palmen N Engelen JJ Buurman WA Spontaneous and stimulated release of tumor necrosis factor-alpha (TNF) from blood monocytes of miners with coal workers' pneumoconiosis Am Rev Respir Dis 1988 138 1589 94 2849351
Lassalle P Gosset P Aerts C Fournier E Lafitte JJ Degreef JM Wallaert B Tonnel AB Voisin C Abnormal secretion of interleukin-1 and tumor necrosis factor alpha by alveolar macrophages in coal worker's pneumoconiosis: comparison between simple pneumoconiosis and progressive massive fibrosis Exp Lung Res 1990 16 73 80 2307149
Schins RP Borm PJ Epidemiological evaluation of release of monocyte TNF-alpha as an exposure and effect marker in pneumoconiosis: a five year follow up study of coal workers Occup Environ Med 1995 52 441 50 7670618
Schins RP Borm PJ Serum procollagen type III peptide in coal workers' pneumoconiosis: a five-year follow-up study Exp Lung Res 1994 20 445 55 7843078
Bernard AM Gonzalez-Lorenzo JM Siles E Trujillano G Lauwerys R Early decrease of serum Clara cell protein in silica-exposed workers Eur Respir J 1994 7 1932 7 7875262
Cobben NA Drent M Schols AM Lamers RJ Wouters EF Van Dieijen-Visser MP Serum lactate dehydrogenase and its isoenzyme pattern in ex-coalminers Respir Med 1997 91 616 23 9488895 10.1016/S0954-6111(97)90008-1
Cobben NA Drent M De Vries J Wouters EF Van Dieijen-Visser MP Henderson RF Serum beta-glucuronidase activity in a population of ex-coalminers Clin Biochem 1999 32 659 64 10638950 10.1016/S0009-9120(99)00070-3
Takahashi T Munakata M Ohtsuka Y Satoh-Kamachi A Sato R Homma Y Kawakami Y Serum KL-6 concentrations in dairy farmers Chest 2000 118 445 50 10936139 10.1378/chest.118.2.445
Harris J Bartelson BB Barker E Balkissoon R Kreiss K Newman LS Serum neopterin in chronic beryllium disease Am J Ind Med 1997 32 21 6 9131208 10.1002/(SICI)1097-0274(199707)32:1<21::AID-AJIM3>3.0.CO;2-U
Maier LA Kittle LA Mroz MM Newman LS Beryllium-stimulated neopterin as a diagnostic adjunct in chronic beryllium disease Am J Ind Med 2003 43 592 601 12768609 10.1002/ajim.10227
Foucher P Biour M Blayac JP Godard P Sgro C Kuhn M Vergnon JM Vervloet D Pfitzenmeyer P Ollagnier M Mayaud C Camus P Drugs that may injure the respiratory system Eur Respir J 1997 10 265 79 9042620 10.1183/09031936.97.10020265
Ohnishi H Yokoyama a Yasuhara Y Watanabe A Naka T Hamada H Abe M Nishimura K Higaki J Ikezoe J Kohno N Circulating KL-6 levels in patients with drug induced pneumonitis Thorax 2003 58 872 5 14514942 10.1136/thorax.58.10.872
Kohno N Hamada H Fujioka S Hiwada K Yamakido M Akiyama M Circulating antigen KL-6 and lactate dehydrogenase for monitoring irradiated patients with lung cancer Chest 1992 102 117 22 1320562
Goto K Kodama T Sekine I Kakinuma R Kubota K Hojo F Matsumoto T Ohmatsu H Ikeda H Ando M Nishiwaki Y Serum levels of KL-6 are useful biomarkers for severe radiation pneumonitis Lung Cancer 2001 34 141 8 11557124 10.1016/S0169-5002(01)00215-X
Takahashi H Imai Y Fujishima T Shiratori M Murakami S Chiba H Kon H Kuroki Y Abe S Diagnostic significance of surfactant proteins A and D in sera from patients with radiation pneumonitis Eur Respir J 2001 17 481 7 11405529 10.1183/09031936.01.17304810
Fan LL Langston C Pediatric interstitial lung disease: children are not small adults Am J Respir Crit Care Med 2002 165 1466 7 12045116 10.1164/rccm.2204012
Al-Salmi QA Walter JN Colasurdo GN Sockrider MM Smith EO Takahashi H Fan LL Serum KL-6 and Surfactant Proteins A and D in Pediatric Interstitial Lung Disease Chest 2005 127 403 407 15654008 10.1378/chest.127.1.403
Kobayashi I Ono S Kawamura N Okano M Miyazawa K Shibuya H Kobayashi K KL-6 is a potential marker for interstitial lung disease associated with juvenile dermatomyositis J Pediatr 2001 138 274 6 11174630 10.1067/mpd.2001.110324
Joint Statement of the American Thoracic Society (ATS), the European Respiratory Society and the World Association of Sarcoidosisand Other Granulomatous Disorders (WASOG) Statement on Sarcoidosis Am Respir Crit Care Med 1999 160 736 755
Hunninghake GW Crystal RG Pulmonary sarcoidosis: a disorder mediated by excess helper T-lymphocyte activity at sites of disease activity N Engl J Med 1981 305 429 34 6454846
Newman LS Rose CS Maier LA Sarcoidosis N Engl J Med 1997 336 1224 34 9110911 10.1056/NEJM199704243361706
Keicho N Kitamura K Takaku F Yotsumoto HL Serum concentration of soluble interleukin-2 receptor as a sensitive parameter of disease activity in sarcoidosis Chest 1990 98 1125 9 2171882
Lawrence EC Brousseau KP Berger MB Kurman CC Marcon L Nelson DL Elevated concentrations of soluble interleukin-2 receptors in serum samples and bronchoalveolar lavage fluids in active sarcoidosis Am Rev Respir Dis 1988 137 759 64 3128148
Ziegenhagen MW Benner UK Zissel G Zabel P Schlaak M Muller-Quernheim J Sarcoidosis: TNF-alpha release from alveolar macrophages and serum level of sIL-2R are prognostic markers Am J Respir Crit Care Med 1997 156 1586 92 9372680
Ziegenhagen MW Rothe ME Schlaak M Muller-Quernheim J Bronchoalveolar and serological parameters reflecting the severity of sarcoidosis Eur Respir J 2003 21 407 13 12661993
Grutters JC Fellrath JM Mulder L Janssen R van den Bosch JM van Velzen-Blad H Serum soluble interleukin-2 receptor measurement in patients with sarcoidosis: a clinical evaluation Chest 2003 124 186 95 12853522 10.1378/chest.124.1.186
Tsutsumi T Nagai S Imai K Setoyama Y Uchiyama T Izumi T Soluble interleukin-2 receptor in blood from patients with sarcoidosis and idiopathic pulmonary fibrosis Sarcoidosis 1994 11 102 9 7809492
Muller-Quernheim J Pfeifer S Strausz J Ferlinz R Correlation of clinical and immunologic parameters of the inflammatory activity of pulmonary sarcoidosis Am Rev Respir Dis 1991 144 1322 9 1741545
Rothkrantz-Kos S van Dieijen-Visser MP Mulder PG Drent M Potential usefulness of inflammatory markers to monitor respiratory functional impairment in sarcoidosis Clin Chem 2003 49 1510 7 12928233 10.1373/49.9.1510
Liebermann J Schleissner LA Nosal A Sasfre A Mishkin FS Clinical correlations of serum angiotensin-converting enzyme (ACE) in sarcoidosis Chest 1983 84 522 528 6313303
Ainslie GM Poulter LW du Bois RM Relation between immunocytological features of bronchoalveolar lavage fluid and clinical indices in sarcoidosis Thorax 1989 44 501 509 2669224
Prior C Barbee RA Evans PM Townsend PJ Primett ZS Fyhrquist F Gronhagen-Riska C Haslam PL Lavage versus serum measurements of lysozyme, angiotensin converting enzyme and other inflammatory markers in sarcoidosis Eur Respir J 1990 3 1146 1154 1965306
Sharma P Smith I Maguire G Stewart S Shneerson J Brown MJ Clinical value of ACE genotyping in diagnosis of sarcoidosis Lancet 1997 349 1602 1603 9174568 10.1016/S0140-6736(05)61631-5
Tomita H Ina Y Sugiura Y Sato S Kawaguchi H Morishita M Yamamoto M Ueda R Polymorphism in the angiotensin-converting enzyme (ACE) gene and sarcoidosis Am J Respir Crit Care Med 1997 156 255 259 9230757
Papadopoulos KI Melander O Orho-Melander M Groop LC Carlsson M Hallengren B Angiotensin converting enzyme (ACE) gene polymorphism in sarcoidosis in relation to associated autoimmune diseases J Intern Med 2000 247 71 7 10672133 10.1046/j.1365-2796.2000.00575.x
McGrath DS Foley PJ Petrek M Izakovicova-Holla L Kolek V Veeraraghavan S Lympany PA Pantelidis P Vasku A Wells AU Welsh KI Du Bois RM Ace gene I/D polymorphism and sarcoidosis pulmonary disease severity Am J Respir Crit Care Med 2001 164 197 201 11463587
Hashimoto S Nakayama T Gon Y Hata N Koura T Maruoka S Matsumoto K Hayashi S Abe Y Horie T Correlation of plasma monocyte chemoattractant protein-1 (MCP-1) and monocyte inflammatory protein-1alpha (MIP-1alpha) levels with disease activity and clinical course of sarcoidosis Clin Exp Immunol 1998 111 604 10 9528906 10.1046/j.1365-2249.1998.00519.x
Iyonaga K Suga M Ichiyasu H Yamamoto T Hiraga Y Ando M Measurement of serum monocyte chemoattractant protein-1 and its clinical application for estimating the activity of granuloma formation in sarcoidosis Sarcoidosis Vasc Diffuse Lung Dis 1998 15 165 72 9789895
Kobayashi J Kitamura S Serum KL-6 for the evaluation of active pneumonitis in pulmonary sarcoidosis Chest 1996 109 1276 82 8625680
Hermans C Petrek M Kolek V Weynand B Pieters T Lambert M Bernard A Serum Clara cell protein (CC16), a marker of the integrity of the air-blood barrier in sarcoidosis Eur Respir J 2001 18 507 14 11589348 10.1183/09031936.01.99102601
Doyle IR Hermans C Bernard A Nicholas TE Bersten AD Clearance of Clara cell secretory protein 16 (CC16) and surfactant proteins A and B from blood in acute respiratory failure Am J Respir Crit Care Med 1998 158 1528 35 9817704
Janssen R Sato H Grutters JC Bernard A van Velzen-Blad H du Bois RM van den Bosch JM Study of Clara cell 16, KL-6, and surfactant protein-D in serum as disease markers in pulmonary sarcoidosis Chest 2003 124 2119 25 14665489 10.1378/chest.124.6.2119
Green FH Overview of pulmonary fibrosis Chest 2002 122 334 339 10.1378/chest.122.6_suppl.334S
Coultas DB Zumwalt RE Black WC Sobonya RE The epidemiology of interstitial lung diseases Am J Respir Crit Care Med 1994 150 967 72 7921471
Tzouvelekis A Patlakas G Bouros D Application of microarray technology in pulmonary diseases Respir Res 2004 5 26 15585067 10.1186/1465-9921-5-26
Rubin MA Use of laser capture microdissection, cDNA microarrays, and tissue microarrays in advancing our understanding of prostate cancer J Pathol 2001 195 80 6 11568894 10.1002/path.892
|
16042760
|
PMC1215520
|
CC BY
|
2021-01-04 16:23:25
|
no
|
Respir Res. 2005 Jul 21; 6(1):78
|
utf-8
|
Respir Res
| 2,005 |
10.1186/1465-9921-6-78
|
oa_comm
|
==== Front
Respir ResRespiratory Research1465-99211465-993XBioMed Central London 1465-9921-6-891608351410.1186/1465-9921-6-89ResearchGlobal expression profiling of theophylline response genes in macrophages: evidence of airway anti-inflammatory regulation Yao Pei-Li [email protected] Meng-Feng [email protected] Yi-Chen [email protected] Chien-Hsun [email protected] Wei-Yu [email protected] Jeremy JW [email protected] Pan-Chyr [email protected] Department of Internal Medicine, National Taiwan University Hospital, No. 7, Chung-Shan South Rd., Taipei 100, Taiwan2 NTU Center for Genomic Medicine, National Taiwan University College of Medicine, Taipei 100, Taiwan3 Institutes of Biomedical Sciences and Molecular Biology, National Chung-Hsing University, No. 250, Kuo-Kuang Rd., Taichung 40227, Taiwan2005 8 8 2005 6 1 89 89 8 4 2005 8 8 2005 Copyright © 2005 Yao et al; licensee BioMed Central Ltd.2005Yao et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Theophylline has been used widely as a bronchodilator for the treatment of bronchial asthma and has been suggested to modulate immune response. While the importance of macrophages in asthma has been reappraised and emphasized, their significance has not been well investigated. We conducted a genome-wide profiling of the gene expressions of macrophages in response to theophylline.
Methods
Microarray technology was used to profile the gene expression patterns of macrophages modulated by theophylline. Northern blot and real-time quantitative RT-PCR were also used to validate the microarray data, while Western blot and ELISA were used to measure the levels of IL-13 and LTC4.
Results
We identified dozens of genes in macrophages that were dose-dependently down- or up-regulated by theophylline. These included genes related to inflammation, cytokines, signaling transduction, cell adhesion and motility, cell cycle regulators, and metabolism. We observed that IL-13, a central mediator of airway inflammation, was dramatically suppressed by theophylline. Real-time quantitative RT-PCR and ELISA analyses also confirmed these results, without respect to PMA-treated THP-1 cells or isolated human alveolar macrophages. Theophylline, rolipram, etazolate, db-cAMP and forskolin suppressed both IL-13 mRNA expression (~25%, 2.73%, 8.12%, 5.28%, and 18.41%, respectively) and protein secretion (<10% production) in macrophages. These agents also effectively suppressed LTC4 expression.
Conclusion
Our results suggest that the suppression of IL-13 by theophylline may be through cAMP mediation and may decrease LTC4 production. This study supports the role of theophylline as a signal regulator of inflammation, and that down regulation of IL-13 by theophylline may have beneficial effects in inflammatory airway diseases.
==== Body
Introduction
Asthma is a highly prevalent health problem worldwide that may cause significant morbidity and mortality [1,2]. The mechanisms of airflow obstruction in asthma are various, including broncho-constriction with the contraction of the airway's smooth muscle, increased secretion of mucus, mucosal edema with vascular leakage, and the infiltration of inflammatory cells [3]. The pathogenesis of asthma and its susceptibility involve a complex interplay of various genetic and environmental factors, which may modulate airway inflammation and the remodeling processes that are not only present even in mild asthma but also govern the appearance and severity of airway hyper-responsiveness [4].
The inflammatory cells involved include the infiltration of airway T cells, T helper cells, mast cells, basophils, eosinophils, and macrophages [5]. Macrophages are the predominant immune effector in the alveolar spaces and airway, and are believed to play a pivotal role in various pulmonary inflammatory disorders [6,7]. Recently, their importance in the pathogenesis of asthma has been reappraised and emphasized [8]. Although their role in asthmatic inflammation is still incompletely understood, it is clear that macrophages may participate in airway inflammation though multiple mechanisms. Furthermore, macrophages have been reported to release lukotriene B4 (LTB4), lukotriene C4 (LTC4), prostaglandin D2 (PGD2), superoxide anion, and lysosomal enzymes in response to immunoglobulin E (Ig E) [5,9,10]. They also produce inflammatory mediators, such as platelet-activating factor, interleukin 1 beta (IL1β), IL-6, IL-8, and tumor necrosis factor- alpha (TNF-α) [11-14]. These mediators may play important roles in producing broncho-constriction or causing inflammatory changes.
Theophylline is a weak and non-selective inhibitor of phosphodiesterase (PDE) in airway smooth muscle cells. In high doses, theophylline may lead to an increase in intracellular cAMP and cGMP, and mediate the relaxation of airway smooth muscles and suppress airway inflammation [15]. In chronic obstructive pulmonary disease (COPD) patients, theophylline can reduce the total number and proportion of neutrophils, the production of interleukin-8, and neutrophil chemotatic responses, further suggesting its anti-inflammatory effects [15,16]. Several studies have also demonstrated that theophylline has a steroid-sparing effect [17,18]. Theophylline inhibits the degranulation and release of mediators, including platelet-activating factor, LTC4, cationic proteins, and superoxide anion, from eosinophils, granulocytes, and alveolar macrophages in vitro [19,20]. However, the effects of theophylline on gene expressions in macrophages has not been well studied.
In this study, we analyzed the expression profiles of inflammation-related genes of macrophages in response to theophylline, using a human cDNA microarray [21,22]. We also identified differentially expressed genes in macrophages after incubating with theophylline. Our study confirmed the diverse roles of theophylline as an immune modulator, which may be helpful in improving its use in the treatment of airway inflammatory disorders.
Methods
Cell lines, alveolar macrophage isolation, and theophylline treatment
Human monocyte cell line THP-1 (ATCC TIB 202; ATCC, Manassas, VA) was grown with RPMI 1640 media (GIBCO-BRL; Gaithersburg, MD) supplemented with 1.5 g/l Na2HCO3, 4.5 g/l glucose and 10% FBS (GIBCO-BRL) and then incubated at 37°C with 20% O2 and 5% CO2· 3.2 × 10-7M PMA (SIGMA Chemical Co.; St. Louis, MO) was applied to monocyte cultures. After incubating with PMA for 24 hours, monocytes were differentiated into macrophage-like phenotypes. Macrophages were washed three times with RPMI medium containing 10% FBS and incubated for another 24 hours to eliminate the effects of PMA.
Alveolar macrophages were obtained by bronchoalveolar lavage (BAL) during routine bronchoscopic examination with written informed consent from three smoker patients with chronic bronchitis. BAL was performed from the right middle lobe or lingula using three to five successive aliquots of 20 ml of 0.9% sterile NaCl. The BAL fluid was centrifuged at 800 × g for 10 min at 4°C. After two washings, the cells were plated on plastic Petri dishes in serum-free RPMI 1640 media and allowed to adhere for 2 h at 37°C. Non-adherent cells were removed by washings with PBS. Adherent cells contained more than 95% alveolar macrophages [23,24]. The 5 × 104 cells were plated on 24 well plates with complete RPMI medium. After incubating for 24 hours, theophylline was added to the alveolar macrophages. The study protocol was approved by the National Taiwan University Hospital's Ethics Committee.
The designated concentration of theophylline (0, 2.5, 5, 10, and 20 μg/ml; SIGMA) was added to macrophages (PMA-treated THP-1 cells). The drug treatments covered a proper range of theophylline concentrations corresponding to the clinical plasma therapeutic levels for asthma patients [17,25]. After incubation for 24 hours, the cells were harvested with RNAzol B and followed by microarray experiments.
Human cDNA microarray analysis
Human EST clones with putative gene names were obtained from the IMAGE consortium libraries through its distributor (Research Genetics, Huntsville, AL). The cDNA microarray with 9,600 PCR-amplified cDNA fragments was prepared by an arraying machine. Five micrograms of mRNAs were labeled with Biotin-16-dUTP during the reverse transcription as described in our previous report [22]. All of the experiments were individually performed in triplicate. The microarray images were scanned, digitized, and analyzed using a flat scanner (PowerLook 3000, UMAX, Taipei, Taiwan) and GenePix 3.0 software (Axon, Union City, CA). The replicates were used to calculate the mean and standard deviation of gene expression and the coefficient of variation (CV) as the measurement of reproducibility. The details of target preparation, hybridization, color development, image analysis, and spot quantification have been described previously [21,22]. (see online supplemental data for additional details on the microarray system) (see additional file: 1).
Northern blotting and real-time quantitative RT-PCR
To confirm the results derived from the microarray, six differentially expressed clones were randomly selected from the cluster analysis and the entire inserts of the clones were individually PCR-amplified to serve as probes for Northern blotting. The amplified cDNA fragments were labeled with digoxigenin-11-dUTP by random primed labeling as our previous report [21]. To correct the quantity of RNA loading, the signals were normalized with the mRNA expression level of GAPDH in the same blot.
Due to the limitations of mRNA extraction from non-proliferated macrophages and low expression levels of some genes, we employed real-time quantitative RT-PCR (RTQ-RT-PCR) with SYBR Green detection to confirm the results derived from the microarray. There were eight differentially expressed clones randomly selected from the cluster analysis for RTQ-RT-PCR analyses. The TATA box binding protein (TBP) was used as an internal control. The primers were shown in Table 1 and detailed procedures have been described previously [22]. All of the experiments were performed in triplicate.
Table 1 Oligonucleotides for real-time quantitative RT-PCR
mRNA targets Oligonucleotides (5'→3') a Product size (bp)
IL-5 F180: ATAGCCAATGAGACTCTGAGGATTC 89
R268: AGTGTGCCTATTCCCTGAAAGAT
IL-13 F155: TGAGGAGCTGGTCAACATCA 76
R230: CAGGTTGATGCTCCATACCAT
IL-18 F211: GCTGAACCAGTAGAAGACAATTGC 94
R304: CCAGGTTTCATCATCTTCAGCTA
IL-13Rα1 F495: GGAATACCAGTCCCGACACTAACT 93
R587: GGCCTTCTCTAAAGATGTTTTCACA
IL-13Rα2 F44: GGCTATTTGAAGTCGCCATAACC 78
R121: AGATTTAAAACCTTGATATTGCCTCTCT
TNF-α F414: CTCGAACCCCGAGTGACAA 64
R477: AGCTGCCCCTCAGCTTGA
VEGF-a F1200: AACACACACTCGCGTTGCAA 69
R1268: CGGCTTGTCACATCTGCAAGT
VEGF-c F1193: AGATGCCTGGCTCAGGAAGA 74
R1266: ATGTCATGGAATCCATCTGTTGAGT
GM-CSF F119: GCCCTGGGAGCATGTGAA 78
R196: TTCATCTCAGCAGCAGTGTCTCTA
TBP F852: CACGAACCACGGCACTGATT 89
R940: TTTTCTTGCTGCCAGTCTGGAC
a F and R indicate forward and reverse primers, respectively. Numbers indicate the mRNA sequence position.
Western blotting analysis and ELISA
The details of nuclear extract preparation and Western blot analysis have been described previously [26]. IL-13 was detected using a 1:500 dilution of mouse monoclonal anti-IL-13 primary antibody, a 1:1000 dilution of HRP-conjugated anti-mouse IgG secondary antibody (Santa Cruz Biotech, Santa Cruz, CA), and the Western blotting luminol reagent (Santa Cruz Biotech) as detection reagent. α-tubulin, used as the control for gel loading, was detected using mouse monoclonal anti-α-tubulin primary antibody (Santa Cruz Biotech). In addition, the cultured medium was collected and centrifuged to remove cellular debris, and the supernatants were frozen at -80°C until assayed by ELISA (R&D System Inc., Minneapolis, MN, USA). IL-13 concentrations were determined by comparison to recombinant standards that run parallel with each batch of assays. Each sample was determined in duplicate. The sensitivity of this ELISA was at < 32 pg/ml.
Statistical analysis
All of the experiments were performed in triplicate and analyzed by ANOVA (Excel, Microsoft; Taipei, Taiwan). A P value < 0.05 was considered statistically significant. In an attempt to reduce variations arising from experimental results of different microarrays, the intensity values of spots from each microarray were re-scaled using a global-scale method. Detailed procedures have been described previously [21,22]. Where appropriate, the data are presented as the mean ± standard deviation. (see online supplement for additional details on the microarray data analysis) (see additional file: 1).
Results
Microarray analysis
Biotin-labeled probes deriving from mRNAs of macrophages (PMA-treated THP-1 cells) stimulated with different concentrations of theophylline were hybridized to microarrays with 9,600 putative genes to profile the gene expression patterns. The CV was 5.26% and the Pearson correlation coefficient of overall reproducibility for large-scale analyses was 0.98. The results of microarray analyses indicated that 2,724 out of 9,600 EST clones were identified, according to at least one dosage point, whose expression level is larger than the background (> 3,000 intensity units).
Among these, 341 genes displayed more than a 2-fold expression change across all five study-included dosages in theophylline treatment. 75 genes were randomly selected and sequenced retrospectively after differential expressions were found, to assure that they indeed represented the true transcript. 45 genes were up-regulated and 30 genes were down-regulated by theophylline in macrophages (PMA-treated THP-1 cells). A full list of genes and data related to treatment with theophylline were posted at our Web site. . In addition, the gene lists of suppressed and enhanced expression were shown in the online data supplement as Tables 1 and 2 (see additional file: 1).
These selected genes were grouped into eight categories by their putative functions on the basis of literature reports (Figure 1). The categories included: (1) cytoskeleton and motility related genes (n = 11), such as caveolin-1 and actin-related protein 3; (2) signal transduction related genes (n = 21), such as testis-specific kinase 1 and IL-6 signal transducer, (3) transcription regulators (n = 9), such as transforming growth β-Induced factor and Down syndrome critical region protein 1; (4) transport regulators (n = 7), such as CD36 and transcobalamin II; (5) cytokines (n = 4), such as IL-13 and vascular endothelial growth factor (VEGF)-C; (6) cell cycle regulators (n = 4), such as cyclin-dependent kinase inhibitor 1C and ecotropic viral integration site 2B; (7) metabolism related genes (n = 35), such as platelet proteoglycan 1 and eukaryotic translation initiation factor 2, subunit 3; and (8) miscellaneous genes (unknown) (n = 21), such as KIAA0703 gene and KIAA0266. We found that 51% of affected genes were related to signal transduction or metabolism. Genes with multiple roles were also included in more than one category.
Figure 1 Hierarchical clustering of the gene expression profile in macrophages with or without theophylline. 75 differentially expressed genes dose-dependently down- or up-regulated by theophylline were identified and further grouped into 8 categories. Relative expression levels of these genes are color-coded.
Northern blotting and RTQ-RT-PCR
To substantiate the results of the microarray studies, Northern blot analysis and RTQ-RT-PCR were performed. Six gene expressions that showed more than a 2-fold change, including ETIF2S3, IRF7, IL6ST, TAFII55, PRG1 and TESK1, were randomly selected and evaluated. Figure 2A shows that the results of Northern blot analyses were consistent with of the microarray studies. GAPDH was used as an internal control. The other eight genes selected from microarray analysis were also confirmed by RTQ-RT-PCR, including GMCSF, TNF-α, IL-13 Rα1, IL-13 Rα2, IL-5, IL-18, VEGF-a, and VEGF-c (Figure 2B). The IRF7, TAFII55, PRG1, GMCSF, TNF-α, IL-13 Rα1, IL-5, and IL-18 genes were suppressed by theophylline, whereas ETIF2S3, IL6ST, TESK1, IL-13 Rα2, VEGF-a, and VEGF-c were stimulated.
Figure 2 Northern blot and real-time quantitative RT-PCR analyses of differentially expressed genes. (A) Northern blot analysis of six randomly selected genes in macrophages. (B) Real-time quantitative RT-PCR analysis of eight cytokine genes. The relative amount of each cDNA level against to TBP cDNA was measured and defined by an arbitrary unit. (10 μg/ml of theophylline treatment approximately corresponds to the clinical plasma level.)
Theophylline down-regulates IL-13 expression
Microarray analysis revealed that IL-13 expression was dose-dependently suppressed by theophylline. Figure 3A revealed a collection of cropped microarray images (3 × 3 spots) showing gene expression patterns of IL-13 in macrophages (PMA-treated THP-1 cells) treated with theophylline. Northern and Western blot analyses also showed a similar suppression of IL-13 production (Figure 3B and 3C). The concentration of 10 μg/ml of theophylline approximately corresponds to the clinical plasma therapeutic level.
Figure 3 IL-13 expression in macrophages was suppressed by theophylline in a dose-dependent manner. (A) Close-up view of microarray digital image of IL-13 expression. (B) Northern blot analysis of IL-13 mRNA expression in macrophages. GAPDH was used as an internal control. (C) Western blot analysis revealed that IL-13 protein level in macrophages was decreased by theophylline. α-tubulin was used as the loading control. 10 μg/ml of theophylline treatment approximately corresponds to the clinical plasma level.
IL-13 mRNA expression in macrophages (PMA-treated THP-1 cells) with different dosages of theophylline treatment was measured by RTQ-RT-PCR, and results showed a significant suppression compared with the control (α = 0.05, p = 0.0079) (Figure 4A). ELISA showed that IL-13 protein secretion was also reduced in a dose-dependent manner (50.23%, 32.43%, 24.93%, and 5.33%, respectively, of the level seen in the absence of theophylline) (Figure 4B).
Figure 4 Effects of theophylline on IL-13 expression and protein secretion in macrophages. (A) IL-13 mRNA level was measured by RTQ-RT-PCR, and significantly decreased after treating with theophylline (down to less than 45% compared with control. *α = 0.05, p = 0.0079). (B) IL-13 protein secretion, by ELISA analysis, was also reduced in macrophages treated with theophylline. The trend was similar to that for the mRNA (down to less than 55% compared with control. *α = 0.05, p = 0.0075). (C) The IL-13 protein secretion in alveolar macrophages isolated from three patients (BAL-A, BAL-B, and BAL-C) with chronic bronchitis was also reduced when treated with 10 μg/ml theophylline (α = 0.05, p = 0.043; upper panel). The BAL-B specimens were treated with difference concentration of theophylline (0, 2.5, 5, 10, 20 μg/ml, respectively). IL-13 protein secretion was decreased in a dose-dependent manner with theophylline (lower panel). Arrow indicates the concentration of theophylline treatment corresponding to the clinical plasma levels (10 mg/L).
In this study, we also evaluated IL-13 expression in human alveolar macrophages using ELISA. Results showed that IL-13 protein secretion was reduced in alveolar macrophages when treated by 10 μg/ml theophylline. The amounts of IL-13 protein in those without theophilline treatment specimens BAL-A, BAL-B and BAL-C are 224, 283 and 191 pg/ml, respectively. In contrast, there are 86, 47, and 69 pg/ml of IL-13 in the respective theophylline treatment specimens. In alveolar macrophages from smoker patients with chronic bronchitis, IL-13 protein secretion was decreased in a dose-dependent manner with theophylline (Figure 4C).
cAMP-dependent pathways in the down-regulation of IL-13 expression
Since theophylline can effectively suppress the production of IL-13 by macrophages, we then examined whether other cAMP-related agents have the same effects. The designated dosages of two phosphodiesterase inhibitors type IV (etazolate and rolipram) and two cAMP-elevating agents (forskolin and dibutyryl-cAMP) were added to macrophages (PMA-treated THP-1 cells) separately for 24 hours. Dose-dependent suppression of IL-13 mRNA expression were observed in all four drugs that could increase intracellular cAMP levels (α = 0.05, p = 0.0009 compared to control) (Figure 5A). Similar results were obtained with ELISA (α = 0.05, p = 0.0018 compared to control) (Figure 5B).
Figure 5 Suppression of IL-13 expression in macrophages by PDE type IV inhibitors and cAMP-elevating agents. Two PDE type IV inhibitors, etazolate and rolipram, and two cAMP-elevating agents, forskolin and db-cAMP (dibutyryl-cAMP), were added to macrophage separately for 24 hours. The cells were harvested to extract RNA for RTQ-RT-PCR, and the cultured medium were used to carry out ELISA. (A) RTQ-RT-PCR analysis showed a decrease of IL-13 mRNA in a dose-dependent manner after treating with four drugs (α = 0.05, p = 0.0009). (B) The results of ELISA also revealed that IL-13 protein secretion was reduced after treatment with four drugs (α = 0.05, p = 0.0018).
Effects on LTC4 expression
The LTC4 is the downstream target of IL-13. Theophylline and other four cAMP-related drugs (etazolate, rolipram, forskolin, and db-cAMP) could dose-dependently suppress LTC4 secretion by macrophages (Figure 6). As shown in Figure 6A, LTC4 production in macrophages (PMA-treated THP-1 cells) was significantly reduced to 78.34%, 34.63%, 23.32%, and 13.51% of the levels seen in the absence of the drug, respectively, with different dosages of theophylline. Similar results were observed in macrophages (PMA-treated THP-1 cells) treated with other cAMP-related drugs (Figure 6B).
Figure 6 LTC4 secretion by macrophages was suppressed by theophylline and cAMP signaling regulators in a dose-dependent pattern. The cultured medium of macrophages treated with tested drugs was collected to perform ELISA. (A) LTC4 protein secretion was reduced by theophylline stimulation. (B) Etazolate, rolipram, forskolin, and db-cAMP (dibutyryl-cAMP) also suppressed LTC4 protein secretion. Arrow indicates the concentration of theophylline treatment corresponding to the clinical plasma levels (10 mg/L).
Discussion
Macrophages are key inflammatory cells that have been documented to play a critical role in various airway disorders [8]. In this study, we analyzed the gene expression profiles of macrophages in response to theophylline. A panel of inflammation related genes was identified, as well as genes associated with angiogenesis, cell adhesion, cell motility, signal transduction, and cell proliferation that are dose-dependently down- or up-regulated by theophylline. Our results revealed that 45 genes were up-regulated and 30 genes were down-regulated by theophylline (supplemental Tables 1 and 2). We also found that theophylline can down-regulate IL-13 expression in macrophages through cAMP mediation, which further leads to decreased LTC4 production. Our results provide positive evidence supporting the role of theophylline as a regulator of inflammation.
In this report, interferon regulatory factor 7 (IRF-7) and CD36 were both suppressed by theophylline in macrophages, especially in high dosages (Figures 1 and 2A, and Supplemental Table 2). IRF-7 has been studied extensively in viral infection [27] and can induce the gene expressions of interferon and cytokine [28]. Interestingly, an over-expression of IRF-7 can trigger monocyte differentiation towards macrophages and induce cell cycle arrest, suggesting a different function for IRF-7 in innate immunity [28]. Furthermore, CD36 is a multi-functional receptor that may play important roles in monocyte/macrophage biology, especially in atherogenic and inflammatory processes [29,30].
Airway inflammation in asthma is regulated by a complex network of cytokines. We found that the expressions of several cytokines were altered within the period of theophylline stimulation (Figure 2B and supplemental Tables 1 and 2). Theophylline can suppress IL-5 and IL-13 production by stimulating peripheral blood nuclear cells (PBMC) [31]. Decreased expression of immuno-regulatory cytokines, including IL-12, IL-18, or interferon gamma, can strengthen the inflammatory process and play regulatory roles in asthma by modifying Th2 lymphocyte responses [32]. Using a mouse model of allergic inflammation, it has been shown that GMCSF significantly contributes to the development of allergic airway inflammation, and that dexamethasone can completely inhibit GMCSF release [33]. Our findings reveal similar results in the suppression of IL-5, IL-18, and GMCSF in macrophages with theophylline (Figure 2B).
IL-13 is an immuno-regulatory cytokine secreted predominantly by activated Th2 cells [34], and induces dramatically different patterns of gene expression in primary cultures of airway epithelial cells, airway smooth muscle cells, and lung fibroblasts [35]. IL-13 expression is not only in T cells and mast cells but also in both normal alveolar macrophages and those from subjects with pulmonary fibrosis [36]. Some reports demonstrate that IL-13 is overproduced in asthma and have implicated IL13 in pathogenesis of inflammation and airway remodeling responses [37-39]. Although the contribution of macrophage derived IL-13 to disease is still not clear, it has been considered for therapy target because of its ability to stimulate inflammatory and airway hyperreactivity responses. In this study, there is strong evidence supporting that IL-13 expression is down-regulated by theophylline in a dose-dependent manner (Figures 3 and 4). We also further confirmed the mRNA expression and protein secretion of IL-13 with RTQ-RT-PCR and ELISA.
In macrophages (PMA-treated THP-1 cells), IL-13Rα1 mRNA expression was inhibited by theophylline, whereas IL-13Rα2 mRNA expression increased (Figure 2B). IL-13 modifies cell behavior by activating the signal transducer and activator of transcription 6 (STAT-6). Consequently, not only IL-13 concentration but also the density of IL-13Rα1 expression may determine the role of IL-13 in the regulation of inflammatory responses in affected tissues. However, not all responses to IL-13 on monocytes and macrophages are dependent on signaling via IL-13Rα1 and significant STAT6 activation [40]. Leukotrienes, the products of lipoxygenases, are thought to be important mediators of IL-13-induced asthma phenotype [41]. LTC4 stimulates eotaxin production by IL-13 treated fibroblasts, thereby indirectly inducing eosinophil sequestration [42]. Recently, some studies demonstrated that the regulation of cAMP level by inhibiting PDE activity appears to be involved in the regulation IL-13 release [43,44]. The type IV PDE inhibitors have the potential to exert an anti-inflammatory effect by inhibiting IL-13 production in lymphocyte and peripheral blood mononuclear cells [43,44].
In this study, we also investigated the influence of cAMP pathway on IL-13 and LTC4 expression in macrophage. We found that etazolate and rolipram, which are PDE type IV inhibitors, can significantly inhibit IL-13 and LTC4 production in mRNA and protein level. Similar suppressions are shown in treatment with PKA activator (forskolin and dibutyryl-cAMP). The results indicate that the inhibition of IL-13 and LTC4 might through cAMP and PKA mediation in macrophage. However, the role of PKA in anti-inflammatory effects through cAMP mediation is less established. Although most of the cAMP exerted its downstream effects though the PKA dependent pathway, some actions of cAMP have been reported to be independent of PKA, including the activation of small GTPase Rap1 [45].
In addition, several lines of evidence support that cAMP may act at transcription, post-transcription, or translation levels. For example, cAMP elevating agents can repress NF-kappaB dependent transcription by a variety of mechanism [46], and NF-kappaB is also known to be involved in the induction of TNF-alpha, IL-3, and IL-13 in human mast cells [47]. Although the mechanism involved in the regulation of cAMP and IL-13 is still unclear, this study suggests that a possible pathway of the suppressive effects of theophylline on IL-13 expression may be through a cAMP mediated regulation.
As shown in Figure 7, we summarized a model for the possible gene regulation in macrophages (PMA-treated THP-1 cells) stimulated by theophylline. Our results suggested that the suppression of IL-13 by theophylline may be through the cAMP pathway and further inhibits the expression of LTC4 and LTD4.
Figure 7 A model for the possible gene regulation in macrophage THP-1 stimulated by theophylline. There are many differentially expressed genes involved in the response to theophylline, such as ARP2, IL6ST, VEGF-c, and IL-13. The suppression of IL-13 by theophylline might be through cAMP pathway and further inhibits the expression of LTC4 and LTD4.
Conclusion
These data may facilitate the understanding of the diverse anti-inflammatory effects of theophylline, as well as the potential contributing role of macrophages in the pathogenesis of asthma. The importance of theophylline as a signal regulator of inflammation should be re-emphasized. Our results suggest that theophylline could down-regulate IL-13 expression in macrophages through cAMP mediation, and further lead to a decrease in LTC4 production, which may have beneficial effects on the therapeutic use of theophylline in pulmonary inflammatory diseases.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
PLY performed the RNA isolation, drug treatment and microarray analysis, and drafted the manuscript. MFT performed the alveolar macrophage isolation, culture, drug treatment, ELISA and drafted the manuscript. YCL performed the Northern blotting and real-time RT-PCR experiments. CHW performed the cell culture and real-time RT-PCR experiments. WYL performed the bronchoscopic examination and alveolar macrophage isolation. JJWC and PCY participated in the conception and design of the study as well as proof read the manuscript. All authors read and approved the final manuscript.
Supplementary Material
Additional File 1
Supplemental Methods: including microarray system, preparation of biotin-labeled cDNA targets, microarray hybridization and colorimetric detection, and image processing and data analysis. Supplemental Table 1. Differential genes up-regulated by theophylline in macrophage THP-1. Supplemental Table 2. Differential genes down-regulated by theophylline in macrophage THP-1.
Click here for file
Acknowledgements
This work was supported by the National Science Council of the Republic of China through the National Research Program for Genomic Medicine grants (NSC 91-3112-P-002-017-Y and NSC 93-3112-B-002-026-Y). The authors wish to thank the Microarray Core Facility for Genomic Medicine (supported by the National Research Program for Genomic Medicine, NSC) for microarray analysis and technical support.
==== Refs
McFadden ER JrWarren EL Observations on asthma mortality Ann Intern Med 1997 127 142 147 9230005
Myers TR Pediatric asthma epidemiology: incidence, morbidity, and mortality Respir Care Clin N Am 2000 6 1 14 10639553 10.1016/S1078-5337(05)70054-X
O'Byrne PM Leukotrienes in the pathogenesis of asthma Chest 1997 111 27S 34S 9042024
Laprise C Sladek R Ponton A Bernier MC Hudson TJ Laviolette M Functional classes of bronchial mucosa genes that are differentially expressed in asthma BMC Genomics 2004 5 21 15038835 10.1186/1471-2164-5-21
Hamid Q Tulic' MK Liu MC Moqbel R Inflammatory cells in asthma: mechanisms and implications for therapy J Allergy Clin Immunol 2003 111 S5 S17 12532083 10.1067/mai.2003.22
Shepherd MC Baillie GS Stirling DI Houslay MD Remodelling of the PDE4 cAMP phosphodiesterase isoform profile upon monocyte-macrophage differentiation of human U937 cells Br J Pharmacol 2004 142 339 351 15066910 10.1038/sj.bjp.0705770
Heller RA Schena M Chai A Shalon D Bedilion T Gilmore J Woolley DE Davis RW Discovery and analysis of inflammatory disease-related genes using cDNA microarrays Proc Natl Acad Sci USA 1997 94 2150 2155 9122163 10.1073/pnas.94.6.2150
Peters-Golden M The alveolar macrophage: the forgotten cell in asthma Am J Respir Cell Mol Biol 2004 31 3 7 15208096 10.1165/rcmb.f279
Fuller RW Morris PK Richmond R Sykes D Varndell IM Kemeny DM Cole PJ Dollery CT MacDermot J Immunoglobulin E-dependent stimulation of human alveolar macrophages: significance in type 1 hypersensitivity Clin Exp Immunol 1986 65 416 426 2947765
Joseph M Tonnel AB Capron A Voisin C Enzyme release and superoxide anion production by human alveolar macrophages stimulated with immunoglobulin E Clin Exp Immunol 1980 40 416 422 6254706
Arnoux B Duval D Benveniste J Release of platelet-activating factor (PAF-acether) from alveolar macrophages by the calcium ionophore A23187 and phagocytosis Eur J Clin Invest 1980 10 437 441 6788561
Janson RW Hance KR King TE Jr Human alveolar macrophages produce predominantly the 35-kD pro-forms of interleukin-1 alpha and interleukin-1 beta when stimulated with lipopolysaccharide Am J Respir Crit Care Med 1995 151 1613 1620 7735621
Lohmann-Matthes ML Steinmuller C Franke-Ullmann G Pulmonary macrophages Eur Respir J 1994 7 1678 1689 7995399
Wang N Tabas I Winchester R Ravalli S Rabbani LE Tall A Interleukin 8 is induced by cholesterol loading of macrophages and expressed by macrophage foam cells in human atheroma J Biol Chem 1996 271 8837 8842 8621523 10.1074/jbc.271.15.8837
Barnes PJ Theophylline: new perspectives for an old drug Am J Respir Crit Care Med 2003 167 813 818 12623857 10.1164/rccm.200210-1142PP
Wang CH Lin HC Lin CH Yu CT Liu SL Huang KH Chung KF Kuo HP Effect of theophylline and specific phosphodiesterase IV inhibition on proliferation and apoptosis of progenitor cells in bronchial asthma Br J Pharmacol 2003 138 1147 1155 12684271 10.1038/sj.bjp.0705131
Ito K Lim S Caramori G Cosio B Chung KF Adcock IM Barnes PJ A molecular mechanism of action of theophylline: Induction of histone deacetylase activity to decrease inflammatory gene expression Proc Natl Acad Sci USA 2002 99 8921 8926 12070353 10.1073/pnas.132556899
Tomita K Chikumi H Tokuyasu H Yajima H Hitsuda Y Matsumoto Y Sasaki T Functional assay of NF-kappaB translocation into nuclei by laser scanning cytometry: inhibitory effect by dexamethasone or theophylline Naunyn Schmiedebergs Arch Pharmacol 1999 359 249 255 10344522
Calhoun WJ Stevens CA Lambert SB Modulation of superoxide production of alveolar macrophages and peripheral blood mononuclear cells by beta-agonists and theophylline J Lab Clin Med 1991 117 514 522 1675244
Shute JK Tenor H Church MK Holgate ST Theophylline inhibits the release of eosinophil survival cytokines – is Raf-1 the protein kinase A target? Clin Exp Allergy 1998 28 Suppl 3 47 52 9756186
Yu SL Chen HW Yang PC Peck K Tsai MH Chen JJ Lin FY Differential gene expression in gram-negative and gram-positive sepsis Am J Respir Crit Care Med 2004 169 1135 1143 15001460 10.1164/rccm.200211-1278OC
Chen HW Yu SL Chen JJ Li HN Lin YC Yao PL Chou HY Chien CT Chen WJ Lee YT Yang PC Anti-invasive gene expression profile of curcumin in lung adenocarcinoma based on a high throughput microarray analysis Mol Pharmacol 2004 65 99 110 14722241 10.1124/mol.65.1.99
Hancock A Armstrong L Gama R Millar A Production of interleukin 13 by alveolar macrophages from normal and fibrotic lung Am J Respir Cell Mol Biol 1998 18 60 65 9448046
Carre PC Mortenson RL King TE JrNoble PW Sable CL Riches DW Increased expression of the interleukin-8 gene by alveolar macrophages in idiopathic pulmonary fibrosis. A potential mechanism for the recruitment and activation of neutrophils in lung fibrosis J Clin Invest 1991 88 1802 1810 1752942
Ohta K Fukuchi Y Grouse L Mizutani R Rabe KF Rennard SI Zhong NS A prospective clinical study of theophylline safety in 3810 elderly with asthma or COPD Respir Med 2004 98 1016 1024 15481279 10.1016/j.rmed.2004.02.020
Chen JJ Yao PL Yuan A Hong TM Shun CT Kuo ML Lee YC Yang PC Up-regulation of tumor interleukin-8 expression by infiltrating macrophages: its correlation with tumor angiogenesis and patient survival in non-small cell lung cancer Clin Cancer Res 2003 9 729 737 12576442
Hiscott J Grandvaux N Sharma S Tenoever BR Servant MJ Lin R Convergence of the NF-kappaB and interferon signaling pathways in the regulation of antiviral defense and apoptosis Ann N Y Acad Sci 2003 1010 237 248 15033728 10.1196/annals.1299.042
Lu R Pitha PM Monocyte differentiation to macrophage requires interferon regulatory factor 7 J Biol Chem 2001 276 45491 45496 11585813 10.1074/jbc.C100421200
Huh HY Pearce SF Yesner LM Schindler JL Silverstein RL Regulated expression of CD36 during monocyte-to-macrophage differentiation: potential role of CD36 in foam cell formation Blood 1996 87 2020 2028 8634453
Yesner LM Huh HY Pearce SF Silverstein RL Regulation of monocyte CD36 and thrombospondin-1 expression by soluble mediators Arterioscler Thromb Vasc Biol 1996 16 1019 1025 8696941
Kimura M Okafuji I Yoshida T Theophylline suppresses IL-5 and IL-13 production, and lymphocyte proliferation upon stimulation with house dust mite in asthmatic children Int Arch Allergy Immunol 2003 131 189 194 12876409 10.1159/000071485
McKay A Komai-Koma M MacLeod KJ Campbell CC Kitson SM Chaudhuri R Thomson L McSharry C Liew FY Thomson NC Interleukin-18 levels in induced sputum are reduced in asthmatic and normal smokers Clin Exp Allergy 2004 34 904 910 15196278 10.1111/j.1365-2222.2004.01973.x
Patel HJ Belvisi MG Bishop-Bailey D Yacoub MH Mitchell JA Activation of peroxisome proliferator-activated receptors in human airway smooth muscle cells has a superior anti-inflammatory profile to corticosteroids: relevance for chronic obstructive pulmonary disease therapy J Immunol 2003 170 2663 2669 12594295
Hershey GK IL-13 receptors and signaling pathways: an evolving web J Allergy Clin Immunol 2003 111 677 690 quiz 691 12704343 10.1067/mai.2003.1333
Lee JH Kaminski N Dolganov G Grunig G Koth L Solomon C Erle DJ Sheppard D Interleukin-13 induces dramatically different transcriptional programs in three human airway cell types Am J Respir Cell Mol Biol 2001 25 474 485 11694453
Hancock A Armstrong L Gama R Millar A Production of interleukin 13 by alveolar macrophages from normal and fibrotic lung Am J Respir Cell Mol Biol 1998 18 60 65 9448046
Wills-Karp M Luyimbazi J Xu X Schofield B Neben TY Karp CL Donaldson DD Interleukin-13: central mediator of allergic asthma Science 1998 282 2258 2261 9856949 10.1126/science.282.5397.2258
Kotsimbos TC Ernst P Hamid QA Interleukin-13 and interleukin-4 are coexpressed in atopic asthma Proc Assoc Am Physicians 1996 108 368 373 8902881
Zhu Z Ma B Zheng T Homer RJ Lee CG Charo IF Noble P Elias JA IL-13-induced chemokine responses in the lung: role of CCR2 in the pathogenesis of IL-13-induced inflammation and remodeling J Immunol 2002 168 2953 2962 11884467
Hart PH Bonder CS Balogh J Dickensheets HL Vazquez N Davies KV Finlay-Jones JJ Donnelly RP Diminished responses to IL-13 by human monocytes differentiated in vitro: role of the IL-13Ralpha1 chain and STAT6 Eur J Immunol 1999 29 2087 2097 10427971 10.1002/(SICI)1521-4141(199907)29:07<2087::AID-IMMU2087>3.0.CO;2-J
Grunig G IL-13 and adenosine: partners in a molecular dance? J Clin Invest 2003 112 329 331 12897201 10.1172/JCI200319392
Chibana K Ishii Y Asakura T Fukuda T Up-regulation of cysteinyl leukotriene 1 receptor by IL-13 enables human lung fibroblasts to respond to leukotriene C4 and produce eotaxin J Immunol 2003 170 4290 4295 12682264
Yoshida N Shimizu Y Kitaichi K Hiramatsu K Takeuchi M Ito Y Kume H Yamaki K Suzuki R Shibata E Hasegawa T Takagi K Differential effect of phosphodiesterase inhibitors on IL-13 release from peripheral blood mononuclear cells Clin Exp Immunol 2001 126 384 389 11737051 10.1046/j.1365-2249.2001.01646.x
Essayan DM Kagey-Sobotka A Lichtenstein LM Huang SK Regulation of interleukin-13 by type 4 cyclic nucleotide phosphodiesterase (PDE) inhibitors in allergen-specific human T lymphocyte clones Biochem Pharmacol 1997 53 1055 1060 9174120 10.1016/S0006-2952(97)00102-0
de Rooij J Zwartkruis FJ Verheijen MH Cool RH Nijman SM Wittinghofer A Bos JL Epac is a Rap1 guanine-nucleotide-exchange factor directly activated by cyclic AMP Nature 1998 396 474 477 9853756 10.1038/24884
Ye RD Beta-Adrenergic agonists regulate NF-kappaB activation through multiple mechanisms Am J Physiol Lung Cell Mol Physiol 2000 279 L615 L617 11000119
Lorentz A Klopp I Gebhardt T Manns MP Bischoff SC Role of activator protein 1, nuclear factor-kappaB, and nuclear factor of activated T cells in IgE receptor-mediated cytokine expression in mature human mast cells J Allergy Clin Immunol 2003 111 1062 1068 12743571 10.1067/mai.2003.1342
|
16083514
|
PMC1215521
|
CC BY
|
2021-01-04 16:23:25
|
no
|
Respir Res. 2005 Aug 8; 6(1):89
|
utf-8
|
Respir Res
| 2,005 |
10.1186/1465-9921-6-89
|
oa_comm
|
==== Front
Theor Biol Med ModelTheoretical Biology & Medical Modelling1742-4682BioMed Central London 1742-4682-2-321612021110.1186/1742-4682-2-32ResearchOn the number of founding germ cells in humans Zheng Chang-Jiang [email protected] E Georg [email protected] Breck [email protected] Suresh H [email protected] Department of Occupational and Environmental Medicine, Regions Hospital, University of Minnesota, 640 Jackson Street, Saint Paul, MN 55101, USA2 Division of Public Health Sciences Fred Hutchinson Cancer Research Center, 1100 Fairview Avenue North, Seattle, WA 98109, USA3 Department of Genome Sciences, University of Washington, Seattle, WA 98195, USA2005 24 8 2005 2 32 32 1 6 2005 24 8 2005 Copyright © 2005 Zheng et al; licensee BioMed Central Ltd.2005Zheng et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
The number of founding germ cells (FGCs) in mammals is of fundamental significance to the fidelity of gene transmission between generations, but estimates from various methods vary widely. In this paper we obtain a new estimate for the value in humans by using a mathematical model of germ cell development that depends on available oocyte counts for adult women.
Results
The germline-development model derives from the assumption that oogonial proliferation in the embryonic stage starts with a founding cells at t = 0 and that the subsequent proliferation can be defined as a simple stochastic birth process. It follows that the population size X(t) at the end of germline expansion (around the 5th month of pregnancy in humans; t = 0.42 years) is a random variable with a negative binomial distribution. A formula based on the expectation and variance of this random variable yields a moment-based estimate of a that is insensitive to the progressive reduction in oocyte numbers due to their utilization and apoptosis at later stages of life. In addition, we describe an algorithm for computing the maximum likelihood estimation of the FGC population size (a), as well as the rates of oogonial division and loss to apoptosis. Utilizing both of these approaches to evaluate available oocyte-counting data, we have obtained an estimate of a = 2 – 3 for Homo sapiens.
Conclusion
The estimated number of founding germ cells in humans corresponds well with values previously derived from chimerical or mosaic mouse data. These findings suggest that the large variation in oocyte numbers between individual women is consistent with a smaller founding germ cell population size than has been estimated by cytological analyses.
==== Body
1. Introduction
Despite great strides in our understanding of the genetic regulation of germ cell determination in recent years [1], the size of the founding germ cell population in humans remains obscure. Due to this uncertainty, it is difficult in a clinical environment to estimate the probability that a mutant allele known to be mosaic in the somatic tissues of a parent will be transmitted to offspring. Even in the mouse, where experimental approaches are feasible, the number of founding germ cells (FGCs) has proven difficult to establish. Cytochemical methods have suggested FGC numbers varying from 45 cells [2] to 193 cells [3]. On the other hand, genetic analysis of artificially generated chimerical cellular populations in the mouse indicate that there are only 2 to 9 cells that actually contribute to the germ cell population [4-10]. In this communication, we derive a new method that is applicable to both humans and laboratory animals. This approach exploits a "founder effect" phenomenon that has previously been shown to be amenable to mathematical analysis [11,12]. Specifically, such analysis has shown that, if a population is descended from a small set of ancestral founders, the population size should exhibit substantial variance. Due to exponential expansion of the germline from the small number of founding cells, modest variation of cell cycle parameters between individual founding cells would be amplified into substantially higher levels of variance at later stages of development. Using a stochastic model to reconstruct the germ-cell development in human females, we show that the large variance observed in counts of human oocytes is consistent with the initial origin of these cells from a much smaller FGC population than is often assumed.
2. Overview
Our approach enables us to derive an estimate of the initial FGC population size on the basis of reliable data describing the number of oocytes present at various later stages of human development, when ovarian tissue is more readily available for analysis. Cytological counts of oocytes have shown not only that the size of the female germ cell population varies substantially between individuals of the same age, but also that its age-dependent magnitude is biphasic [13]. Numbers of germ cells increase during the first half of fetal development and then begin a progressive decline that extends throughout the reproductive years. The initial phase initiates with the separation of the germline from the soma, probably taking place no later than the peri-implantation stage (about 9 days after fertilization) [13,14]. Following this initial establishment, mitotically active oogonia undergo an exponential increase in number while only a small fraction of them show any sign of degeneration. At about five months of fetal development (5/12 = 0.42 year), the population reaches its peak as the oogonia enter into meiotic arrest. The germ cells (now defined as primary oocytes) become invested by layers of nurturing granulosa cells to form the follicles, which are readily recognized and enumerated by microscopic examination. The second phase of female germline development, spanning the period from the late embryonic stage to the onset of menopause in adult females (t = 0.42 – 52 years), is characterized by a progressive decline in the number of follicles, largely due to apoptosis [15]. This decline is approximately exponential, but is accelerated in women older than age 38 [16]. Among the million or so oocytes present late in fetal development of the mother, the vast majority will undergo apoptosis while only 300–400 will progress fully through maturation and undergo ovulation during the woman's reproductive life. Eventually, when the number of oocytes in the resting pool falls below 1000, menopause occurs [16].
3. Germ-Cell Kinetics
Oogonium-Birth Model
The stochastic model we use to describe germline development consists of two separate dynamic components (Figure 1). During the early embryonic stage (t = 0 – 0.42 year), a pure-birth model [11] can be used to describe the rapid proliferation of oogonia. At time t = 0, the germline is founded by a ancestral cells (FGCs) that are newly separated from the soma. At time t (0 ≤ t ≤ 0.42), the number X(t) of oogonia follows a negative-binomial probability distribution [11]:
Figure 1 The number of female germ cells in humans undergoes three distinct rate changes, as diagrammed here and defined in the model. For the sake of clarity, the age coordinate is expanded artificially during the embryonic phase. The proliferative phase initiates at the time of germline-soma separation (ca. 9 days after fertilization; t = 0.0 year) and ends after 5 months of gestation (t = 0 – 0.42 year). The declining phases begin later in fetal life and continue into adulthood (t = 0.42 – 52 years) with an accelerated rate of oocyte depletion beginning at age 38 [16]. The dotted line shown during the embryonic stage emphasizes that oogonial cell counts from this period are inaccessible to reliable determination.
Here λ is the oogonial division rate. The expectation and variance of the random variable X(t) are E[X(t)] = aeλt, Var[X(t)] = aeλt(eλt-1) ≈ ae2λt, respectively. Note that the approximation holds true if eλt is large (for humans, aeλ·0.42 ≥ 106). The moment ratio a ≈ E2 [X(t)]/Var[X(t)] then yields an estimate of the number of FGCs.
If one were dependent on using the pure-birth model to estimate a, this could be accomplished by collecting gonadal tissues from a series of abortices and establishing the total number of oogonia in each specimen (xi(ti), i = 1,2,..., I) by microscopic evaluation. However, reliance on access to fetal tissue clearly has several drawbacks. First, many spontaneous abortions are associated with chromosomal aberrations [17] and may therefore display an abnormal pattern of growth kinetics. Second, access to non-diseased fetuses for research is limited by ethical concerns. And, third, microscopic examination of fetal tissues from an early stage of pregnancy is technically challenging. The boundaries of fetal gonad are not clearly demarcated from surrounding cell types and the oogonial cells are difficult to distinguish from the somatic cells. On the other hand, the ovarian follicles that arise at later stages of development are cytologically distinct and can be enumerated with precision. The following derivation of a pure-death model for germ cell dynamics enables us to use this more precise enumeration to advantage.
Oocyte-Death Model
A pure-death model, as described by Bailey [11], can be used to obtain an explicit formulation for the declining phase of germ cell numbers after proliferation has ceased and the apoptotic decline has begun (t = 0.42 – 52 years). Consistent with the findings of Faddy et al. [16], we permit the rate of oocyte loss from the resting pool to vary with age t. The cumulative rate function f(t) is defined as follows:
Conditional on the initial number X(t) = n of oocytes at t = 0.42 year, the number of oocytes in the resting pool at age t (0.42 ≤ t ≤ 52 years) now follows a binomial probability distribution [11]:
4. Estimation Methods
The unknown parameters (a, λ,μ1, μ2) can be estimated using two different methods. The moment-based method estimates only the number of FGCs (parameter a), while the maximum likelihood method estimates all four parameters (a, λ, μ1, μ2) simultaneously.
Moment-Based Method
As mentioned above regarding the oogonium-birth model, the random variable X(t) follows a negative binomial distribution, and therefore the moment ratio E2[X(t)]/Var[X(t)] yields an estimate of a. This relationship holds true with oocyte depletion (oocyte-death model) following the period of exponential growth. As a verification, notice first that the probability-generating function for the negative binomial probability distribution at t = 0.42 is PX(0.42)(s) = {1 - eλ·0.42(1 - s-1)}-a. Therefore, the probability-generating function for the binomial probability distribution at t > 0.42 (conditional on X (0.42) = n) is PX(t)|X(.42)(s) = ((1 - e- f(t)) + e- f(t)s)n. The compounded probability-generating function (t > 0.42) is then given by PX(t)(s) = . The mean and variance for oocyte population size in adult women can be derived from the first and second derivatives of , .
Maximum Likelihood Estimation (MLE)
Although the moment ratio is simple to compute, its derivation requires a large sample size and provides no estimates of the oogonium-birth rate (λ) and the oocyte-death rates (μ1, μ2). The maximum likelihood method is not subject to these difficulties. To derive the likelihood function, note first that Equation 3 (pure-death model) is a probability function conditional on Pr[X(0.42) = n] (Equation 1; pure-birth model). For each of the oocyte counts obtained from a series of autopsies (xi(ti), i = 1,2,..., I) during the post-embryonic stage (0.42 ≤ ti ≤ 52 years), we can combine them to define:
Since each observation xi(ti) makes a contribution like L(xi | a, λ, μ1, μ2) to the likelihood, the final likelihood of the entire data set is the product of all such terms (0.42 ≤ ti ≤ 52 years), such that . To maximize , note that the term is (for a fixed xi(ti) but variable n) a negative binomial up to a constant. Therefore, we use importance sampling [18] from negative binomials to evaluate the likelihood in Equation 4 numerically. In practice, we first generate samples of n from this distribution and then sum the values of the negative binomial probabilities for a given the sampled values of n. We obtain stable likelihood estimates with as few as 100 samples of n. During the process of searching the parameter space, we also restrict the parameter a to be a positive integer (i.e., a = 1,2,3,...).
5. Analysis of Oocyte-Counting Data
Using the above algorithms, we have analyzed published oocyte counts from 102 human females [16,19]. These computations yield MLEs of the four parameters, which are: = 2, = 31.2/year (95% CI: 30.3 – 31.9), = 0.079/year (0.067 – 0.090) and = 0.248/year (0.204 – 0.283). The 95% confidence intervals (95% CI) are based on Markov chain Monte Carlo methods [20]. The expected number of oocytes (solid curve) and the pointwise 80% CI (shaded region) for the predicted counts on the basis of the model are shown together with oocyte counts in Figure 2.
Figure 2 Shown in this diagram are published counts of follicles per individual [16,19] obtained by autopsy in adult stages, when each follicle contains a single oocyte. The data are analyzed with the compound birth-then-death model as described in the text. Observations with follicle counts < 100 were considered unreliable and were excluded from the analysis. The solid line represents the expected number of oocytes at each age in the postembryonic stages based on the model given these data. The shaded area is the pointwise 80% confidence interval. The MLEs of the parameters are: = 2, = 31.2/year, = 0.079/year, = 0.248/year.
The large MLE of λ justifies our use of the moment-based method to estimate the parameter a. To enable moment-based analysis we have grouped the data into 5 discrete age intervals [16] and calculated the mean and variance for each age interval. Computing the moment ratio index for each age interval yields the values shown (Table 1) and an overall average of = 2.7. This value agrees well with the MLE and with those values ( = 2 – 9) that were derived from the segregation of genetic markers in the mouse [4-10].
Table 1 Oocyte-counting data from Faddy et al. [16].
Age Sample Size Mean S.E. Ratio Index
19–30 years 5 78980 15580 5.1
31–35 years 13 25300 4860 2.1
36–40 years 14 21450 2650 4.7
41–45 years 32 7320 1450 0.8
>45 years 36 1880 310 1.0
6. Discussion
Our estimated values of the founding germ cell number a (MLE = 2, moment estimate = 2.7) differ substantially from the estimates made in previous studies [2,3] that relied on cytochemical staining of embryonic material. This discrepancy might best be explained by assuming that not all cells sharing the same cytological phenotype (alkaline phosphatase staining) in common with a FGC population proceed through that course of development. This view is supported by the observation that alkaline phosphatase is present not only in germ cells of the mouse, but also in somatic cells that surround these germ cells [21]. Recent description of cytochemical markers that are more specifically restricted to the germ cells confirms that the expression of alkaline phosphatase occurs in a wider spectrum of cell types [22].
Genetic analysis provides a more stringent approach to germ cell enumeration in the mouse, where experimental crosses permit reliable determination of marker transmission. If a mutation or other stably transmissible cellular property arises early in development, only a subset of cells will display the mutation at later stages. Such mosaic presence of the trait provides an opportunity to help define the stage at which the germ cell precursors (FGFs) had become segregated from the predominant somatic cell population [23,24]. If the extent of mosaicism were closely similar between soma and germline, this would indicate that the size of the FGC population is large, since it would have provided a representative sample of the mosaicism originally present throughout the early embryo. To the contrary, if the correlation between somatic and germline mosaicism is weak, the number of FGCs must be limited and the transmission of any somatic markers to offspring should be more stochastic. This type of analysis [4-10] generally predicts small number of FGCs, consistent with our derivation from the stochastic modeling of oocyte counts.
The present statistical analysis indicates that germ cell proliferation occurring during embryogenesis is characterized both by a small FGC population ( = 2) and by a rapid rate of cell division ( = 31.2/year). Although we have estimated all four parameters of the model (a, λ, μ1 and μ2) in concert using the oocyte-counting data collected post-embryonically, the division rate λ can be verified independently by microscopic study of embryonic tissues. In a recent publication, Bendsen et al. [25] reported their evaluation of 10 fetal gonad specimens between the ages of 6 and 9 weeks. Using morphological clues to distinguish germ cells from somatic cells, these authors were able to establish the number of germ cells per tissue sample throughout this period. We used exponential regression against fetal age t to analyze the data of Bendsen et al. and obtained a direct estimate of = 35.4/year (95% CI: 17.2/year – 53.7/year). Thus, our estimate of the division rate of the FGC is in good agreement with the experimental data.
It has recently been reported [26] that mammalian ovaries contain stem germ cells that are competent to enter into mitosis at a late stage of development. If confirmed by further work, this finding would necessitate some modification of the model proposed here. However, additional computation suggests that the potential effect on the estimate of a is likely to be small. To show this, we replace the pure death model for oocyte dynamics with a birth-and-death model. Also we assume that both the cell-death rate μ and the cell-birth rate ν (μ >ν) are constant during the post-embryonic stage (t > 0.42). The re-derived compounded probability generating function then becomes . From this, we obtain the expectation (E[X(t)] = aeλ.0.42-(μ-ν)t) and the variance of the oocyte population. Numerical evaluations of Var[X (t)] using a range of parameter values suggest that the variance and the estimate of a are largely determined by the observed difference between μ and ν (Var[X(t)] ≈ ae2(λ·0.42-(μ-ν)t)). In other words, our conclusion will change very little as long as the cell-death rate μ is substantially larger than the cell-birth rate ν (even if ν > 0). Because the observation on continued proliferation of germ cells in the adult female is new [26] and additional evidence will need to be collected, a fuller discussion of the modeling issues is deferred here.
The validity of our model clearly relies on certain assumptions that might be refuted by future analyses of tissue dynamics. Specifically, we have assumed that growth of the germ cell population is exponential and involves no significant cell death during gestation and also that the subsequent apoptotic decline is exponential. Furthermore, we have assumed that the sub-populations derived from each individual FGC grow and decline independently of one another. In addition, the model depends crucially on the concept that the cessation of proliferation and entry into meiotic arrest is controlled by the stage of development rather than by the size of the germ cell population. A more detailed analysis than the present report would be required to establish how robust the proposed mechanism may be to departures from each of these assumptions. There clearly are numerous models of greater complexity that could be proposed to account for the observed substantial variance among human oocyte populations. Our realization that the distribution of oocyte counts between individuals can be explained so simply by the computation described here encourages us to suggest that the small-founder effect may be a predominant cause of this variance.
Finally, we note that the squared root of an inversed moment ratio is mathematically equivalent to the coefficient of variation – the quotient of the standard deviation divided by the mean. This parameter is commonly used in biostatistics to characterize the extent of random variation across a broad range of biological processes. This coincidence suggests that the founder-effect interpretation that we have proposed may have broader applications.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
CJZ carried out the initial mathematical derivations and analyzed the oocytes-counting data using the moment-ratio method. EGL and SHM extended the mathematical derivations and completed the maximum-likelihood estimation. BB reviewed the biological implications of the models. All authors participated in preparations and revisions of the manuscript.
Acknowledgements
We would also like to acknowledge financial support from the National Institutes of Health (BB, EGL & SHM) and the National Institute of Occupational Safety and Health (CJZ).
==== Refs
Hubbard EJA Pera RAR A germ-cell odyssey: fate, survival, migration, stem cells and differentiation. Meeting on germ cells EMBO Rep 2003 4 352 7 12671678 10.1038/sj.embor.embor807
Lawson KA Hage WJ Clonal analysis of the origin of primordial germ cells in the mouse Ciba Foundation Symposium 1994 182 68 91 7835158
McMahon A Fosten M Monk M X-chromosome inactivation mosaicism in the three germ layers and the germ line of the mouse embryo J Embryol Exp Morphol 1983 74 207 20 6886595
Mintz B Hermaphroditism, sex chromosomal mosaicism and germ cell selection in allophenic mice J Animal Sci 1968 27 51 60
Mintz B Gene control of mammalian differentiation Annu Rev Genet 1974 8 411 70 4613263 10.1146/annurev.ge.08.120174.002211
Russell LB Locke M Genetic and functional mosaicism in the mouse Role of Chromosomes in Development 1964 Academic, New York 153 81
Russell LB Analysis of the albino-locus region of the mouse. II. Mosaic mutants Genetics 1979 91 141 7 428722
Searle AG Evidence from mutable genes concerning the origin of the germline Basic Life Sci 1978 12 209 24 752319
Soriano P Jaenisch R Retroviruses as probes for mammalian development: allocation of cells to the somatic and germ cell linkages Cell 1986 46 19 29 3013418 10.1016/0092-8674(86)90856-1
Wilkie TM Brinster RL Palmiter RD Germline and somatic mosaicism in transgenic mice Dev Biol 1986 118 9 18 3770310 10.1016/0012-1606(86)90068-0
Bailey NTJ The elements of stochastic processes with applications to the natural sciences 1964 John Wiley & Sons, New York
Yule GU A mathematical theory of evolution, based on the conclusions of Dr. J. C. Willis, F.R.S Proc R Soc London Ser B 1924 213 21 87
Gosden RG Grudzinskas JG, Yovich JL Ovulation 1: Oocyte development throughout life Gametes – The Oocyte 1995 Cambridge University Press, Cambridge 119 149
Wilcox AJ Baird DD Weinberg CR Time of implantation of the conceptus and loss of pregnancy N Engl J Med 1999 340 1796 9 10362823 10.1056/NEJM199906103402304
Tilly JL Commuting the death sentence: how oocytes strive to survive Nat Rev Mol Cell Biol 2001 2 838 48 11715050 10.1038/35099086
Faddy MJ Gosden RG Gougeon A Richardson SJ Nelson JF Accelerated disappearance of ovarian follicles in mid-life: implications for forecasting menopause Hum Reprod 1992 7 1342 6 1291557
Boue J Bou A Lazar P Retrospective and prospective epidemiological studies of 1500 karyotyped spontaneous human abortions Teratology 1975 12 11 26 1162621 10.1002/tera.1420120103
Morgan BJT Applied Stochastic Modelling 2000 Arnold Texts in Statistics, Oxford University Press
Leidy LE Godfrey LR Sutherland MR Is follicular atresia biphasic? Fertil Steril 1998 70 851 9 9806566 10.1016/S0015-0282(98)00316-1
Gilks WR Richardson S Spiegelhalter DJ eds Markov Chain Monte Carlo in Practice 1996 Chapman & Hall, New York
MacGregor GR Zambrowicz BP Soriano P Tissue non-specific alkaline phosphatase is expressed in both embryonic and extraembryonic lineages during mouse embryogenesis but is not required for migration of primordial germ cells Development 1995 121 1487 96 7789278
Saitou M Barton SC Surani MA A molecular programme for the specification of germ cell fate in mice Nature 2002 418 293 300 12124616 10.1038/nature00927
Nesbitt MN Gartler SM The Applications of Genetic Mosaicism to Developmental Problems Annu Rev Genet 1971 5 143 162 16097654 10.1146/annurev.ge.05.120171.001043
McLaren A Numerology of development Nature 1972 239 274 6 4562030 10.1038/239274a0
Bendsen E Byskov AG Laursen SB Larsen HP Andersen CY Westergaard LG Number of germ cells and somatic cells in human fetal testes duringthe first weeks after sex differentiation Hum Reprod 2003 18 13 8 12525434 10.1093/humrep/deg057
Johnson J Canning J Kaneko T Pru JK Tilly JL Germline stem cells and follicular renewal in the postnatal mammalian ovary Nature 2004 428 145 50 15014492 10.1038/nature02316
|
16120211
|
PMC1215522
|
CC BY
|
2021-01-04 16:39:25
|
no
|
Theor Biol Med Model. 2005 Aug 24; 2:32
|
utf-8
|
Theor Biol Med Model
| 2,005 |
10.1186/1742-4682-2-32
|
oa_comm
|
==== Front
Theor Biol Med ModelTheoretical Biology & Medical Modelling1742-4682BioMed Central London 1742-4682-2-331612022310.1186/1742-4682-2-33ResearchPossible Cis-acting signal that could be involved in the localization of different mRNAs in neuronal axons Aranda-Abreu Gonzalo E [email protected]ández Ma Elena [email protected] Abraham [email protected] Jorge [email protected] Instituto de Neuroetología, Universidad Veracruzana, Av. Dos Vistas S/N, km 2.5 Carr. Xalapa-Veracruz. Col. Industrial-Animas. C.P. 91190. Xalapa, Ver. México2005 24 8 2005 2 33 33 21 7 2005 24 8 2005 Copyright © 2005 Aranda-Abreu et al; licensee BioMed Central Ltd.2005Aranda-Abreu et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Messenger RNA (mRNA) comprises three major parts: a 5'-UTR (UnTranslated Region), a coding region, and a 3'-UTR. The 3'-UTR contains signal sequences involved in polyadenylation, degradation and localization/stabilization processes. Some sequences in the 3'-UTR are involved in the localization of mRNAs in (e.g.) neurons, epithelial cells, oocytes and early embryos, but such localization has been most thoroughly studied in neurons. Neuronal polarity is maintained by the microtubules (MTs) found along both dendrites and axon and is partially influenced by sub-cellular mRNA localization. A widely studied mRNA is that for Tau protein, which is located in the axon hillock and growth cone; its localization depends on the well-characterized cis-acting signal (U-rich region) in the 3'-UTR.
Methods
We compared the cis-acting signal of Tau with mRNAs in the axonal regions of neurons using the ClustalW program for alignment of sequences and the Mfold program for analysis of secondary structures.
Results
We found that at least 3 out of 12 mRNA analyzed (GRP75, cofilin and synuclein) have a sequence similar to the cis-acting signal of Tau in the 3'-UTR. This could indicate that these messengers are localized specifically in the axon. The Mfold program showed that these mRNAs have a similar "bubble" structure in the putative sequence signal.
Conclusion
Hence, we suggest that a U-rich sequence in the 3'-UTR region of the mRNA could act as a signal for its localization in the axon in neuronal cells. Sequences homologous to the DTE sequence of BC1 mRNA could direct the messenger to the dendrites. Messengers with homologues of both types of sequence, e.g. β-actin, might be located in both dendrites and axon.
mRNAU-rich regionaxon
==== Body
Background
A messenger RNA (mRNA) comprises three major parts, a 5'-UTR (UnTranslated Region), a coding region and a 3'-UTR. The 3'-UTR contains signal sequences involved in polyadenylation, degradation and localization/stabilization processes. Many studies have shown that certain sequences in the 3'-UTR are involved in localizing the mRNAs in different cells such as neurons, epithelial cells, oocytes and early embryos [1,2]. Such localization has been studied exhaustively in neurons. Neurons are polar cells, with dendrites and axon; dendrites receive information and the axon is specialized to transmit this information to the next neuron [3]. The maintenance of neuronal polarity depends on the microtubules (MTs) [3-6], which are found along both axon and dendrites, and is partially determined by subcellular mRNA localization. The mechanism responsible for creating the polarity involves synergistic controls of translation, stabilization and association with elements of the cytoskeleton.
The mRNA of tau has been studied in detail [7]. Tau is located in the axon hillock and growth cone; the well-characterized cis-acting signal (U-rich region) located in the 3'-UTR of its mRNA is responsible for its localization [8]. HuD protein interacts with this U-rich sequence to form a mRNA-protein complex that is transported toward the axon (axon hillock and growth cone) by interacting with KIF3A, a kinesin responsible for anterograde movement [9-11].
Recently, many mRNAs have been shown to be located in neuronal axons: β-actin, tropomyosin 3 (Tpm3), cofilin, vimentin, immunoglobulin heavy chain biding protein (Bip), heat shock protein 60 (HSP60), heat shock protein 70 (HSP70), heat shock protein 90 (HSP90), glucose regulated protein (grp75) and synuclein [12]. The objective of this paper is to determine, using bioinformatics tools, whether there is a cis-acting signal in all the mRNAs that are transported to the axon and whether this putative signal is similar to the U-rich region in the 3'-UTR of tau mRNA.
Results
The 3'-UTR of tau mRNA contains 3884 bases; the U-rich region (in bold) is responsible for the localization of this mRNA in the axon hillock and growth cone.
UCAGGCCCCUGGGGCCGUCACUGAUCAUGGAGAGAAGAGAGAGUGAGAGUGUGGAAAAAAAAAAAAAAAGAAUGACCUGGCCCCUCACCCUCUGCCCUCCCCGCUGCUCCUCAUAGACAG GCUGACCAGCUUGUCACCUAACCUGCUUUUGUGGCUCGGGUUUGGCUCGGGACUUCAAAAUCAGUGAUGGGAAAAAGUAAAUUUCAUCUUUCCAAAUUGAUUUGUGGGCUAGUAAUAAAA UAUUUUUAAGGAAGGAAAAAAAAAACACGUAAAACCAUGGCCAAACAAAACCCAACAUUUCCUUGGCAAUUGUUAUUGACCCCGCCCCCCCCUCUGAGUUUUAGAGGGUGAAGGAGGCUU UGGAUAGAGGCUGCUUCUGGGGAUUGGCUGAGGGACUAGGGCAACUAAUUGCCCACAGCCCCAUCUUAGGGGCAUCAGGGACAGCGGCAGACAUGAAAGACUUGGGACUUGGUGUGUUUG UGGAGCCGUAAGGCGUAUGUUAACUUUGUGUGGGUUUGAGGGAGGACUGUGAUAGUGAAGGCUGAGAGAUGGGUGGGCUGGGAGUCAGAGGAGAGAGGUGAGGAAGACAGGUUGGGAGAG GGGGCAUUGCGUCCUUGCCAAGGAGCUUGGGAAGCACAGGUAGCCCUGGCUGCAGCAGUCUUAGCUAGCACAGAUGCCUGCCUGAGAAAGCACAGUGGGGUACAGUGGGUGUGUGUGCCC CUUCUGAAGGGCAGCCCAUGGGAGAAGGGGUAUUGGGCAGAAGGAAGGUAGGCCCCAGAAGGUGGCACCUUGUAGAUUGGUUCUCUGAAGGCUGACCUUGCCAUCCCAGGGCACUGCUCC CACCCUCCAGGAGGAGGUCUGAGCUGAGGAGCUUCCUUUUCGAUCUCACAGGAAAACCUGUGUUACUGAGUUCUGAAGUUUGGAACUACAGCCAUGAUUUUGGCCACCAUACAGACCUGG GACUUUAGGGCUAACCAGUUCUUUGUAAGGACUUGUGCCUCUUGCGGGAACAUCUGCCUGUUCUCAAGCCUGGUCCUCUGGCACUUCUGCAGUGGUGAGGGAUGGGGGUGGUAUUCUGGG AUGUGGGUCCCAGGCCUCCCAUCCCUCGCACAGCCACUGUAUCCCCUCUACCUGUCCUAUCAUGCCCACGUCUGCCACGAGAGCCAGUCACUGCCGUCCGUACAUCACGUCUCACCGUCC UGAGUGCCCAGCCUCCCAAGCCCAAUCCCUGGACCCCUGGGUAGUUAUGGCCAAUCUGCUCUACACUAGGGGUUGGAGUCCAGGGAAGGCAAAGAUUUGGGCCUUGGUCUCUAGUCCUAC GUUGCCAGAAUCCAACCAGUGUGCCUCCCACAAGGAACCUUACAACCUUGUUUGGUUUGCUCCAUCAGGCGUUUGGCGCCAUCGUGGAUGGAGUCCGUGUGUGCCUGGAGAUUACCCUGG ACACCUCUGCUUUUUUUUUUUUUACUUUAGCGGUUGCCUCCUAGGCCUGACUCCUUCCCAUGUUGAACUGGAGGCAGCCAAGUUAGGUGUCAAUGUCCUGGCAUCAGUAUGAACAGUC AGUAGUCCCAGGGCAGGGCCACACUUCUCCCAUCUUCUGCUUCCACCCCAGCUUGUGAUUGCUAGCCUCCCAGAGCUCAGCCGCCAUUAAGUCCCCAUGCACGUAAUCAGCCCUUCCAUA CCCCAAUUUGGGGAACAUACCCCUUGAUUGAAAUGUUUUCCCUCCAGUCCUAUGGAAGCGGUGCUGCCUGCCUGCUGGAGCAGCCAGCCAUCUCAGAGACGCAGCCCUUUCUCUCCUGUC CGCACCCUGCUGCGCUGUAGUCGGAUUCGUCUGUUUGUCUGGGUUCACCAGAGUGACUAUGAUAGUGAAAAGAAAAAGAAAAAGAAAAAAGAAAAAAGAAAAAAAAAAAAGGACGCAUGU UAUCUUGAAAUAUUUGUCAAAAGGUUGUAGCCCACCGCAGGGAUUGGAGGGCCUGAUAUUCCUUGUCUUCUUCGUGACUUAGGUCCAGGCCGGUCGAGUGCUACCCUGCUGGACAUCCCA UGUUUUGAAGGGUUUCUUCUUCAUCUGGGACCCCUGCAGACACUGGAUUGUGACAUUGGAGGUCUAUACAUUGGCCAAGGCUGAAGCACAGGACCCGUUAGAGGCAGCAGGCUCCGACUG UCAGGGAGAGCUUGUGGCUGGCCUGUUUCUCUGAGUGAAGAUGGUCCUCUCUAAUCACAACUUCAAGUCCCACAGCAGCCCUGGCAGACAUCUAAGAACUCCUGCAUCACAAGAGAAAAG GACACUAGUACCAGCAGGGAGAGCUGUGGCCCUAGAAAUUCCAUGACUCUCCACUACUAUCCGUGGGUCCUUUCCAAGCCUUGCCUCGUCACCAAGGGCUUGGGAUGGACUGCCCCACUG AUGAAAGGGACAUCUUUGGAGACCCCCUUGGUUUCCAAGGCGUCAGCCCCCUGACCUUGCAUGACCUCCUACAGCUGAAGGAUGAGGCCUUUAAAGAUUAGGAACCUCAGGCCCAGGUCG GCCACUUUGGGCUUGGGUACAGUUAGGGACGAUGCGGUAGAAGGAGGUGGCCAACCUUUCCAUAUAAGAGUUCUGUGUGCCCAGAGCUACCCUAUUGUGAGCUCCCCACUGCUGAUGGAC UUUAGCUGUCCUUAGAAGUGAAGAGUCCAACGGAGGAAAAGGAAGUGUGGUUUGAUGGUCUGUGGUCCCUUCAUCAUGGUUACCUGUUGUGGUUUUCUCUGUAUACCCCCAUUUACCCAU CCUGCAGUUCCUGUCCUUGAAUAGGGGUGGGGGUACUCUGCCAUAUCUCUUGUAGGCAGUCAGCCCCCAAGUCAUAGUUUGGAGUGAUCUGGUCAGUGCUAAUAGGCAGUUUACAAGGAA UUCUGGCUUGUUACUUCAGUGAGGACAAUCCCCCAAGGCCCUGGCACCUGUCCUGUCUUUCCAUGGCUCUCCACUGCAGAGCCAAUGUCUUUGGGUGGGCUAGAUAGGGUGUACAAUUUG CCUGGAACCUCCAAGCUCUUAAUCCACUUUAUCAAUAGUUCCAUUUAAAUUGACUUCAAUAUAAGAGUGUAUCCAUUUGAGAUUGCUUGUGUUGUGGGGUAAAGGGGGGAGGAGGAACAU GUUAAGAUAAUUGACAUGGGCAAGGGGAAGUCUUGAAGUGUAGCAGUUAAACCAUCUUGUAGCCCCAUUCAUGAUGUUGACCACUUGCUAGAGAGAAGAGGUGCCAUAAGGCUAGAACCU AGAGGCUUGGCUGUCCACCAACAGGCAGGCUUUUGCAAGGCAGAGGCAGCCAGCUAGGUCCCUGACUUCCCAGCCAGGUGCAGCUCUAAGAACUGCUCUUGCCUGCUGCCUUCUUGUGGU GUCCAGAGCCCACAGCCAAUGCCUCCUCAAAACCCUGGCUUCCUUCCUUCUAAUCCACUGGCACAUCAGCAUCACCUCCGGAUUGACUUCAGAUCCACAGCCUACACUACUAGCAGUGGG UAAGACCACUUCCUUUGUCCUUGUCUGUUCUCCAGAAAAGUGGGCAUGGAGGCGGUGUUAAUAACUAUAGGUCUGUGGCUUUAUGAGCCUUCAAACUUCUCUCUAGCUUCUGAAAGGGUU ACUUUUGGGCAGUAUUGCAGUCUCACCCUCCGAUGGCUGUAGCCUGUGCAGUUGCUGUACUGGGCAUGAUCUCCAGUGCUUGCAAGUCCCAUGAUUUCUUUGGUGUUUUGAGGGUGGGGG GAGGGACAUGAAUCAUCUUAGCUUAGCUUCCUGUCUGUGAAUGUCCAUAUAGUGUACUGUGUUUUAACAAACGAUUUACACUGACUGUUGCUGUACAAGUGAAUUUGGAAAUAAAGUUAU UACUCUGAUUAAACAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
This cis-acting signal of tau was compared base by base with the other afore mentioned mRNAs using simple alignment. We also made comparisons with another sequence that is specific for the localization of BC1 mRNA in dendrites, the Dendritic Target Element (DTE) [13].
In the β-actin mRNA of chicken a cis-acting signal "zipcode" has been described; a zipcode binding protein binds to this sequence and this is a prerequisite for the localization of the mRNA. The sequence is a tandem repeat of an ACACCCACACCC motif. The mRNA of β-actin has been located in the axon of the neuron and in dendritic spines [12,14]. β-actin mRNA has a sequence closely similar to the tau signal in the first part of its 3'-UTR, but there is also another sequence that could participate in its localization in dendrites; this sequence is very similar to the DTE [13]. The protein tropomyosin 3 has been located in the growth cone of the neuron; its mRNA has also been detected in axons during development [15]. Both tropomyosin 3 and β-actin form parts of the cytoskeleton. No well-defined sequence signal that could be involved in the specific localization of these messengers in the axon (tpm3 and β-actin) has been identified, so it is likely that β-actin and tropomyosin are not exclusive to the axon and could be also found in dendrites.
Cofilin is a cytoskeleton modulating protein; it is also known as actin depolymerizing factor (ADF). The potential role of cofilin is to modulate the changes of actin organization that accompany neurite initiation, axonogenesis and growth cone guidance [16]. The possible signal sequence found in the 3'-UTR of cofilin mRNA is very similar to that of tau; they share a U-rich region, which indicates that this messenger might be transported to the growth cone of the developing neuron. However, a possible DTE sequence is also present, located upstream of the U-signal.
Vimentin has been located by RT-PCR in the axons of dorsal root ganglia (DRG) neurons. A possible sequence signal in vimentin mRNA shares some U with tau but also contains more purines, which might indicate that the protein is not exclusive to the axonal region [12].
Bip is a protein that binds to the immunoglobin heavy chains in pre-β cells. Its mRNA shares some U with the tau sequence; nevertheless, its sequence suggests that this mRNa, like vimentin, is probably not exclusive to the axon [17].
The heat shock proteins and grp75 messengers have similarities with the tau sequence, but once again they are probably not exclusive to the axon. They could interact with other proteins in different parts of the cell.
Synuclein is a soluble unfolded protein that can aggregate into insoluble fibrils under several pathological conditions including Parkinson's and Alzheimer's diseases [12]. The possible cis-acting signal of the mRNA for this protein is very similar to the tau signal, with only a single U to C substitution, suggesting that the synuclein messenger may be transported to the axon by a similar mechanism to the tau messenger and that aggregation and precipitation of the synuclein protein within the axon contributes to neurodegenerative disease.
These analyses carried out by alignment allowed us to show that the cis-acting signals of the mRNAs examined have some homology with that of the tau messenger.
The highest homology scores are:
The secondary structures of these four mRNAs, which showed the closest homologies to the tau sequence, were analyzed using the program Mfold (Figs. 1, 2, 3, 4). Fig. 5 shows a model of U-rich mRNAs that could be transported to the axon. The results show that the secondary structures of synuclein and cofilin mRNAs are very similar to that of the tau messenger, and the cis-acting signal sequence is inside the "bubble" according to the Mfold program. This indicates to us that both the signal sequence and the secondary structure could be determining factors in the location of these messengers in the axon region.
Figure 1 RNA secondary structure of 3'-UTR of tau mRNA. The arrow indicates the "bubble" where the HuD binds to stabilize the messenger.
Figure 2 RNA secondary structure of 3'-UTR of GRP75 mRNA. The arrow indicates the U-rich signal sequence that could be involved in the localization of the messenger.
Figure 3 RNA secondary structure of 3'-UTR of synuclein mRNA. The arrow indicates the U-rich signal sequence that could be involved in the localization of the messenger.
Figure 4 RNA secondary structure of 3'-UTR of cofilin mRNA. The arrow indicates the U-rich signal sequence that could be involved in the localization of the messenger.
Figure 5 A model of the 3'-UTR/U-rich region by virtue of which the mRNA could be transported toward the axon. The mRNAs that contain a U-rich sequence in the 3'-UTR are candidates for transport toward the axon. The model suggests that an mRNA binding protein intreacts with the signal-sequence forming a putative complex that is anchored to a kinesin protein. The mRNAs that do not contain such a U-sequence might remain in the cell body or to migrate towards the dendrites.
Discussion
The first messenger to be analyzed in the 3'-UTR with respect to its localization/stabilization was tau [9,10]. The U-rich region in its signal sequence enables the formation of a complex with HuD, a prerequisite for transport to the axon, and increases the stability of the mRNA. The basic function of tau protein is to stabilize the microtubules; it prevents depolymerization and consequent loss of neuronal polarity. Recently, several other messengers have been shown by RT-PCR to be localized in neuronal axon of the neuron, but the possibility that these mRNAs are also located in the dendrites has not been excluded.
GRP75
Glucose regulated protein 75 (GRP75) is an important molecular chaperon belonging to the heat shock protein (HSP) family. It is highly expressed in conditions of glucose deprivation of glucose. Its messenger was located in the axon and it has a U-rich region. It might not be confined exclusively to the axon because this protein responds to a metabolic stress [18].
Synuclein
Alfa-synuclein is involved in neurodegenerative diseases and its presence has been observed in the pre-synaptic and nuclear compartments, though the location in the nucleus has not been well documented. The synuclein messenger possesses a U-rich region; nevertheless a C interrupts the potential signal sequence. When it is wrongly folded, this protein may aggregate in the cell forming fibrils, typical of Alzheimer's and Parkinson's diseases. The aggregation of synuclein is similar to tau in Alzheimer patients, which could indicate similar intracellular behavior by both proteins [19].
Cofilin
The 3'-UTR of the cofilin messenger has a U-rich region very similar to the signal sequence of tau, which on the face of it suggests that it might be located exclusively in the axon. Nevertheless, recent studies demonstrate that it participates in the shrinkage of dendritic spines associated with the long-term depression of hippocampal synapses, suggesting that it is also found in dendrites. Moreover, it is involved in neuronal development, axogenesis, guidance of the growth cone and dendrite formation. Although the cofilin messenger is present in axons, the possible participation of the protein in events related to the unplugging of synapses because of its association with actin further suggests that it is not confined to the axon but also occurs in the dendrites [16].
β-actin
The 3'-UTR of the β-actin messenger is very short and shows low homology when aligned with the tau cis-acting signal. However, when it was aligned with the dendritic target element, it showed better homology. The β-actin messenger was shown to possess a zipcode that leads it towards the dendrites instead of the axon [14].
HSP70 and HSP90
Molecular chaperones and their functions in protein folding have been implicated in several neurodegenerative conditions, including Parkinson's and Huntington's diseases, which are characterized by accumulation of protein aggregates (e.g. α-synuclein and huntingtin, respectively). These aggregates have been shown in various experimental systems to respond to changes in levels of molecular chaperones, suggesting the possibility of therapeutic intervention and a role for chaperones in disease pathogenesis. It remains unclear whether chaperones also play a role in Alzheimer's disease, a neurodegenerative disorder characterized by β-amyloid and tau protein aggregates. In various cellular models, increased levels of Hsp70 and Hsp90 promote tau solubility and tau binding to microtubules, reduce insoluble tau and cause reduced tau phosphorylation. Conversely, lowered levels of Hsp70 and Hsp90 result in the opposite effects. A direct association between the chaperones and tau protein has been demonstrated. Many results suggest that the up-regulation of molecular chaperones may suppress the formation of neurofibrillary tangles by partitioning tau into a productive folding pathway and thereby preventing tau aggregation [20]. When we compared the 3'-UTRs of the messengers for these chaperones, they showed some homology with the cis-acting signal of tau because each possesses a U-rich region, which could indicate that they are found in axons.
The model
On the basis of the results we suggested a model for mRNA localization in the axon (Fig. 5).
The mRNAs containing the U-rich region could be complexed with a protein responsible for transport toward the axon, just as HuD complexes with and stabilizes the tau messenger [9]. HuD itself has the capacity to bind to different mRNAs such as GAP-43 [21], neuroserpin [22], acetylcholinesterase [23] and c-myc [24], so it might interact with other messengers with a U-rich signal, stabilizing the messenger and facilitating transport to the axon. The motor protein that translocates the complex along the axonal microtubules could belong to the kinesin family, by analogy with the translocation of the tau messenger by the kinesin KIF3A [11]. When the complex reaches the correct destination, the mRNA is translated. mRNAs that lack the U-rich sequence presumably go to another cellular compartment in the neuron; those with DTE-like signals might preferentially accumulate in the dendrites.
The mechanisms determining whether a messenger such as β-actin is transported preferentially to the axon or the dendrites are poorly understood. The existence of two potentially conflicting location signals in the 3'-UTR (one U-rich and tau-like, the other DTE-like) raises questions about how the final destination of such mRNAs is determined within the neuron.
Conclusion
In the 3'-UTRs of some mRNAs in neurons there are cis-acting signals that direct mRNAs such as tau and GAP-43 to the axon. In general, these signals are rich in uridine and do not contain guanidine. Comparison of the Dendritic Target Element (DTE) with the 3'-UTRs of several axon-located messengers showed some homology in a specific region of the 3'-UTR. Most of the 3'-UTRs studied possess homologies with the signals involved in the localization of mRNAs in axons and dendrites. This might explain why as much β-actin is present in dendrites as in axons, though the distribution mechanisms in such cases are not understood. In addition, we found a DTE homology in the 3'-UTR of HSP70 and 90. The significance of this is not clear; some messengers are transported towards the axon or towards the dendrites as required.
A sequence homologous to DTE in tau occurs near the end of the 3'-UTR, next to the polyadenylation site, which indicates that only the axon signal sequence (not the dendrite signal sequence) is functional, because mRNA degradation starts at the poly(A) site. The 3'-UTR of MAP2 [25] possesses no homology with the axon signal sequence, suggesting that as many tau as MAP2 mRNAs are transported exclusively to their respective regions inside the neuron.
Very U-rich sequences in the 3'-UTR might be signals that direct some mRNAs exclusively to the axon. If we understand which signals/sequences the neuronal cell uses for the correct location of its mRNAs, it might become possible to determine which factors lead to mislocalization of messengers and of proteins, as has recently been suggested in relation to certain neurodegenerative diseases such as Alzheimer's.
Methods
All the mRNAs analyzed in this study belong to the Rattus norvegicus genome and were located using the following GeneBank accession numbers. β-actin; NM_031144, tropomyosin 3 (Tpm3); NM_057208, cofilin; NM_017147, vimentin; NM_031140, immunoglobulin heavy chain biding protein (Bip); M14050, heat shock protein 60 (HSP60); X53585, heat shock protein 70 (HSP70); L16764, heat shock protein 90 (HSP90); S45392, glucose regulated protein (grp75); s78556, synuclein; NM_031688; NM_057114 and NM_053576 and tau; X79321. The 3'-UTRs of the mRNAs were analyzed using the program ClustalW [26], and the secondary structures were generated by the Mfold program [27].
Competing interests
The author(s) declare that they have no competing interests.
==== Refs
Wilhelm JE Vale RD RNA on the move: the mRNA localization pathway J Cell Biol 1993 123 269 274 8408211 10.1083/jcb.123.2.269
St Johnston D The intracellular localization of messenger RNAs Cell 1995 81 161 170 7736568 10.1016/0092-8674(95)90324-0
Black MM Baas PW The basis of polarity in neurons TINS 1989 12 211 214 2473556
Daniels M The role of microtubules in the growth ans stabilization of nerve fibres Ann NY Acad Sci 1975 253 535 544 167634
Sargent BP What distinguishes axons from dendrites? Neuron knows more than we do TINS 1989 12 203 205 2473553
Yamada KM Spooner BS Wessells NK Axon growth: Roles of microfilaments and microtubules Proc Natl Acad Sci 1970 66 1206 1212 5273449
Litman P Barg J Rindzoonski L Ginzburg I Subcellular localization of tau mRNA in differentiating neuronal cell culture: Implications for neuronal polarity Neuron 1993 10 627 638 8476613 10.1016/0896-6273(93)90165-N
Sadot E Marx R Barg J Behar L Ginzburg I Complete sequence of 3'-untranslated region of tau from rat central nervous system: implications for mRNA heterogeneity J Mol Biol 1994 241 325 331 8057376 10.1006/jmbi.1994.1508
Aranda-Abreu GE Behar L Chung S Furneaux H Ginzburg I Embryonic lethal abnormal vision-like RNA-binding proteins regulate neurite outgrowth and tau expression in PC12 cells J Neurosci 1999 19 6907 6917 10436048
Aronov S Aranda G Behar L Ginzburg I Axonal tau mRNA localization coincides with tau protein in living neuronal cells and depends on axonal targeting signal J Neurosci 2001 21 6577 6587 11517247
Aronov S Aranda G Behar L Ginzburg I Visualization of translated tau protein in the axons of the neuronal P19 cells and characterization of tau RNP granules J Cell Sci 2002 115 3817 3827 12235292 10.1242/jcs.00058
Willis D Ka WL Jun-Qi Z Chang JH Smith A Kelly T Merianda TT Sylvester J Minnen JV Twiss J Differential Transport and Local Translation of Cytoskeletal, Injury-Response, and Neurodegeneration Protein mRNAs in Axons J Neurosci 2005 25 778 791 15673657 10.1523/JNEUROSCI.4235-04.2005
Muslimov IA Santi E Homel P Perini S Higgins D Tiedge H RNA transport in dendrites: a cis-acting targeting element is contained within neuronal BC1 RNA J Neurosci 1997 17 4722 4733 9169532
Dhanrajan MT Oleynikov Y Kelič S Shenoy SM Hartley A Stanton PK Singer R Bassell GJ Activity-Dependent Trafficking and Dynamic Localization of Zipcode Binding Protein 1 and β-Actin mRNA in Dendrites and Spines of Hippocampal Neurons J Neurosci 2003 23 3251 12716932
Hannan AJ Gunning P Jeffrey PL Weinberger RP Structural compartments within neurons: developmentally regulated organization of microfilament isoform mRNA and protein Mol Cell Neurosci 1998 11 289 304 9698395 10.1006/mcne.1998.0693
Sarmiere PD Bamburg JR Regulation of the neuronal actin cytoskeleton by ADF/cofilin J Neurobiol 2004 58 103 117 14598374 10.1002/neu.10267
Munro S Pelham HR An Hsp70-like protein in the ER: identity with the 78 kd glucose-regulated protein and immunoglobulin heavy chain binding protein Cell 1986 46 291 300 3087629 10.1016/0092-8674(86)90746-4
Liu Y Liu W Song XD Zuo J Effect of GRP75/mthsp70/PBP74/mortalin overexpression on intracellular ATP level, mitochondrial membrane potential and ROS accumulation following glucose deprivation in PC12 cells Mol Cell Biochem 2005 268 45 51 15724436 10.1007/s11010-005-2996-1
Specht CG Tigaret CM Rast GF Thalhammer A Rudhard Y Schoepfer R Subcellular localisation of recombinant alpha- and gamma-synuclein Mol Cell Neurosci 2005 28 326 334 15691713 10.1016/j.mcn.2004.09.017
Dou F Netzer WJ Tanemura K Li F Hartl FU Takashima A Gouras GK Greengard P Xu H Chaperones increase association of tau protein with microtubules Proc Natl Acad Sci 2003 100 721 726 12522269 10.1073/pnas.242720499
Mobarak CD Anderson KD Morin M Beckel-Mitchener A Rogers SL Furneaux H King P Perrone-Bizzozero NI The RNA-binding protein HuD is required for GAP-43 mRNA stability, GAP-43 gene expression, and PKC-dependent neurite outgrowth in PC12 cells Mol Biol Cell 2000 11 3191 3203 10982410
Cuadrado A Navarro-Yubero C Furneaux H Kinter J Sonderegger P Munoz A HuD binds to three AU-rich sequences in the 3'-UTR of neuroserpin mRNA and promotes the accumulation of neuroserpin mRNA and protein Nucleic Acids Res 2002 0 32202 2211
Deschenes-Furry J Belanger G Perrone-Bizzozero N Jasmin BJ Post-transcriptional regulation of acetylcholinesterase mRNAsin nerve growth factor-treated PC12 cells by the RNA-binding protein HuD J Biol Chem 2003 278 5710 5717 12468554 10.1074/jbc.M209383200
Manohar CF Short ML Nguyen A Nguyen NN Chagnovich D Yang Q Cohn SL Cohn HuD, a neuronal-specific RNA-binding protein, increases the in vivo stability of MYCN RNA J Biol Chem 2002 277 1967 1973 11711535 10.1074/jbc.M106966200
Blichenberg A Schwanke B Rehbein M Garner CC Richter D Kindler S Identification of a cis-acting dendritic targeting element in MAP2 mRNAs J Neurosci 1999 19 8818 8829 10516301
EMBL-EBI, European Bioinformatics Institute
MFOLD: Prediction of RNA secondary structure (M. Zuker)
|
16120223
|
PMC1215523
|
CC BY
|
2021-01-04 16:39:25
|
no
|
Theor Biol Med Model. 2005 Aug 24; 2:33
|
utf-8
|
Theor Biol Med Model
| 2,005 |
10.1186/1742-4682-2-33
|
oa_comm
|
==== Front
Theor Biol Med ModelTheoretical Biology & Medical Modelling1742-4682BioMed Central London 1742-4682-2-3410.1186/1742-4682-2-34Book ReviewReview of "Systems Biology in Practice" by Edda Klipp, Ralf Hertwig, Axel Kowald, Christoph Wierling and Hans Lehrach Agutter Paul S [email protected] Theoretical and Cell Biology Consultancy, 26 Castle Hill, Glossop, Derbyshire, UK2005 26 8 2005 2 34 34 Klipp E , Herwig R , Kowald A , Wierling C , and Lehrach H .
Systems Biology in Practice .
Berlin: Wiley-VCH . 2005 . 449 pages, ISBN-10 3-527-31078-9, ISBN-13 978-3-527-31078-4. Eur 99.00 hardback . 6 8 2005 26 8 2005 Copyright © 2005 Agutter; licensee BioMed Central Ltd.2005Agutter; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
==== Body
Systems biology – the study of such functional networks as cellular metabolism, signalling and gene expression – is a rapid growth area, as illustrated by many recent publications in Theoretical Biology and Medical Modelling and other journals. Its emergence has been fostered by the development of automated, high-throughput techniques such as DNA microarrays, of the ever-increasing storage capacity and calculating speed of computers, of new mathematical tools, and of modern electronic communications. Its impact on biological research is already evident: traditional data (e.g. the results of single-gene studies) are being reinterpreted in a broader cellular context, radically new research strategies are appearing, novel insights into the integration of cell systems are being gained, and the prospects for advances in medicine, biotechnology, ecology and other areas of applied science and technology seem, to the optimistic, almost limitless. To write a book surveying the concepts, methods and potential applications of this nascent though already wide field was a challenging enterprise, but the authors of Systems Biology in Practice have largely achieved their objective. This is a timely volume that should be welcomed both by practising systems biologists and by newcomers to the field.
The book is divided into three parts. Part I (chapters 1–4), occupying about a quarter of the main text, introduces the relevant biological and mathematical concepts and experimental techniques. Part II (chapters 5–12), occupying the following two-thirds of the book, focuses on the three main facets of systems biology as it stands today (metabolism, signal transduction and gene expression), but it also includes brief surveys of the cell cycle and ageing, an interesting chapter on evolution and self-organization, a discussion of data integration methods, and a brief but lively speculation about future directions and applications. The short final part (chapters 13–14) is a descriptive list of currently-available internet databases and tools and modelling tools, assessing the advantages and disadvantages of each. Every concept and method introduced throughout the book is illustrated by at least one boxed example – a well-attested and effective pedagogical device – and most of the examples used are clear and appropriate. The overall organization of the text is lucid, the index is thorough, and the publishers are to be commended on a very well-produced volume.
Inevitably, since this is a pioneering work on the subject, there are flaws. Informed readers will notice omissions; for example, the otherwise excellent account of metabolic control theory in chapter 5 makes no mention of the biochemical systems theory of Savageau, Voit and their colleagues, which some authorities consider conceptually and methodologically superior in certain applications. The balance between sections of the text seems odd in places. In chapter 3, for instance, after clear and well-directed introductions to linear algebra and ordinary differential equations, the authors launch into an account of advanced statistical theory (section 3.4.1) that has little direct relevance to the remainder of the book – in contrast to the remainder of section 3.4, where various statistical techniques and their practical applications are well described – and would be better placed in an appendix, if not omitted. Chapter 5 begins with 20 pages of elementary thermodynamics and enzymology, surely elementary for most readers, and ends with the same amount of text – a mere 20 pages – on metabolic control theory, which is less likely to be familiar and is far more mathematically sophisticated. Much of the account of ageing in chapter 7 is devoted to the defective mitochondria theory – reasonably, in that one of the authors (Kowald) has been a major contributor to this theory, but less reasonably in that several alternative models of ageing in the literature are scarcely considered. There are several misprints: for example, equation 3.21 (p. 63) is incorrect, 'sites' is written for 'sides' (p. 67) and 'mitochondrium' for 'mitochondrion' (p. 168). In places, the writing would have benefited from the assistance of a native English speaker; odd phrasing such as 'negative definite' for 'always negative' (p. 74, example 3–10 and surrounding text) is distracting. Perhaps more importantly for some readers, the authors have concentrated throughout on ordinary differential equation models. Stochastic modelling is mentioned where it is potentially relevant, but only briefly, and this might be the most serious imbalance in the book; as the authors admit, the continuum hypothesis is unlikely to be valid in modelling many aspects of signal processing and transcriptional regulation.
Because of the rate of progress in systems biology, it seems likely that a second edition of this book will appear before many years have passed. It is to be hoped that these (mostly minor) defects will be eliminated in the next edition. However, they detract little from the value of the work as it stands. I expect to consult it regularly over the foreseeable future, and I am confident that biologists everywhere will benefit from having a copy to hand.
|
0
|
PMC1215524
|
CC BY
|
2021-01-04 16:39:25
|
no
|
Theor Biol Med Model. 2005 Aug 26; 2:34
|
utf-8
|
Theor Biol Med Model
| 2,005 |
10.1186/1742-4682-2-34
|
oa_comm
|
==== Front
Thromb JThrombosis Journal1477-9560BioMed Central London 1477-9560-3-121611531510.1186/1477-9560-3-12Original Basic ResearchAnalysis of blood coagulation in mice: pre-analytical conditions and evaluation of a home-made assay for thrombin-antithrombin complexes Sommeijer Dirkje W [email protected] Oerle René [email protected] Pieter H [email protected] Janneke J [email protected] Joost CM [email protected] Henri MH [email protected] Cate Hugo [email protected] Laboratory for Experimental Internal Medicine, Academic Medical Center of Amsterdam, Amsterdam, The Netherlands2 Department of Vascular Medicine, Academic Medical Centre of Amsterdam, Amsterdam, The Netherlands3 Laboratory for Clinical Thrombosis and Haemostasis, Department of Internal Medicine, and Cardiovascular Research Institute Maastricht, University of Maastricht, The Netherlands2005 22 8 2005 3 12 12 23 3 2005 22 8 2005 Copyright © 2005 Sommeijer et al; licensee BioMed Central Ltd.2005Sommeijer et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
The use of mouse models for the study of thrombotic disorders has gained increasing importance. Methods for measurement of coagulation activation in mice are, however, scarce. The primary aim of this study was to develop a specific mouse thrombin-antithrombin (TAT) ELISA for measurement of coagulation activation and to compare it with two commercially available assays for human TAT complexes. In addition, we aimed to improve methods for mouse plasma anticoagulation and preparation.
Methods and results
First, for the measurement of TAT-complexes in plasma a mouse specific TAT-ELISA was developed using rabbit polyclonal antibodies raised against mouse thrombin and rat antithrombin, respectively. This ELISA detected an increase in TAT levels in a mouse model of endotoxemia. Two commercial human TAT ELISAs appeared to be less specific for mouse thrombin-rat antithrombin complexes.
Second, to prevent clotting of mouse blood sodium citrate was either mixed with blood during collection in a syringe or was injected intravenously immediately prior to blood collection. Intravenous sodium citrate completely inhibited blood coagulation resulting in plasma with consistently low TAT levels. Sodium citrate mixed with blood during collection resulted in increased TAT levels in 4 out of 16 plasma samples. Third, heparinase was added to plasma samples after in vivo injection of different heparin doses to test its neutralizing effect. Heparinase neutralized up to a 20 U of heparin/mouse and resulted in accurate APTT and factor VIII determinations.
Conclusion
These procedures and reagents for plasma preparation and coagulation testing will improve studies on thrombotic disorders in mice.
==== Body
Background
Since the identification of genes encoding mouse clotting proteins and the creation of a growing number of transgenic mice, the use of mouse models for the study of thrombotic disorders has gained increasing interest [1]. While methodology for determining fibrin deposition in tissues [2] and quantification of thrombus formation [2,3] is readily available, optimal utilization of blood samples from mice remains a poorly explored area.
The availability of proper laboratory assays is essential in order to study coagulation activation in mice, comparably to humans. There is, however, a striking lack of analytical methods to test coagulation activation specifically in mice. For human studies, several commercial assays are available, including assays for thrombin-antithrombin (TAT) complexes to test in vivo coagulation activation [4,5]. When these assays, based on antibody detection of human protein epitopes, are applied on mouse plasma, it is likely that difficulties can arise depending on the degree of interspecies cross-reactivity, sensitivity, and specificity.
A practical issue of importance is that, traditionally, research on blood coagulation depends on determination of the activity of coagulation factors in a plasma sample. For this purpose specific anticoagulants, including sodium citrate or cocktails of substances need to be added to blood samples, in order to prevent ex vivo clotting. While even in humans the collection of proper plasma samples is not without practical problems, the collection of plasma from mice is a major technical challenge. As far as we know, there are no published procedures for anticoagulation that specifically address plasma sampling in mouse.
The use of mouse models enables investigations into the mechanism of coagulation activation in plasma as well as its consequences, i.e. fibrin deposition and pathological changes in organs, simultaneously. However, for proper tissue collection procedures such as heparinization are frequently required, which impairs most conventional tests of plasmatic coagulation. Human studies showed that addition of heparinase to plasma samples neutralized up to 2 U/ml heparin resulting in accurate factor VIII determinations [6]. In order to minimize interference of heparin we tested the effect of heparinase added to mouse plasma samples following in vivo heparinization.
In this study we developed a new mouse specific TAT ELISA and validated this method using a mouse model of endotoxemia, characterized by enhanced coagulation activation. Furthermore, we compared this new ELISA with two commercially available human immunoassays for measuring thrombin-antithrombin complexes in mouse plasma samples collected by different blood collection procedures.
Methods
Animals
All studies were performed in male mice of the C57Bl/6 strain (University of Maastricht or Academic Medical Center, Amsterdam), which were 8 weeks old and weighing approximately 20 grams at the start of the experiments. The animals were housed in normal cages in an environment with a 12 hour light-dark cycle, controlled temperature (20 ± 2°C) and humidity (50 ± 10%). Animals had free access to water and diets (Hope Farms, Woerden, The Netherlands). Endotoxemia studies were performed by i.p. injection of 2 mg lipopolysaccharide (LPS, E. coli, Sigma, St.Louis, MO) per kg. The Experimental Animal Ethics Committee of the University Maastricht and of the Academic Medical Center, Amsterdam approved all experimental protocols.
Chemicals
All reagents were of analytical grade or better and were from commercial suppliers.
Development of TAT-ELISA
Rabbit immunization
Mouse thrombin (Sigma, St.Louis, MO, USA) and rat antithrombin (AT) (Sigma) were dissolved in PBS (150 mmol/L NaCl, 10 mmol/L sodium phosphate, pH 7.4), to a concentration of 200 and 140 μg/ml, respectively. Fifteen-week old New Zealand White rabbits, weighing approximately 3 kilograms were obtained from Harlan (Cambridge, UK). Immediately before immunization with thrombin and AT, pre-immune sera were collected. Thereafter, each rabbit was immunized with a mixture of 500 μl Freund's Complete Adjuvant (Difco, Detroit, MI, USA) and 500 μl protein solution. For each protein, two rabbits were injected twice s.c. on the back and twice i.m. in the hind legs. The rabbits were boosted at nine and 15 weeks after the first immunization with thrombin and AT in combination with Freund's Incomplete Adjuvant (Difco). During the first three months test samples of approximately five ml blood were obtained regularly and thereafter samples of approximately 50 ml were obtained monthly from each rabbit.
Reactivity of rabbit antisera for TAT ELISA
MaxiSorp plates (Nunc A/S, Roskilde, Denmark) were coated overnight at 4°C with 100 μl thrombin or AT in a concentration of 5 μg/ml diluted in coating buffer (70 mM Na2CO3, 30 mM NaHCO3, pH 9). Subsequently, the plates were washed three times with PBS, 0.05 %(v/v) Tween (PBS/Tween) and incubated with 100 μl rabbit sera diluted in PBS, 0.05 %(v/v) Tween, 1 %(v/v) FCS (PBS/Tween/FCS). The plates were incubated for 1 hour at room temperature and washed three times with PBS/Tween. Thereafter, the plates were incubated for 1 hour with 100 μl horseradish peroxidase (HRP)-conjugated swine anti-rabbit antibodies (DAKO A/S, Glostrup, Denmark) diluted 250-times in PBS/Tween/FCS. The plates were washed three times with PBS/Tween and developed with o-phenylenediamine (OPD, Sigma) for 30 minutes. The reaction was stopped with 100 μl 1 M H2SO4 and the optical density (OD) was determined at 490 and 650 nm. Calculations were performed using the SOFTmax software from Molecular Devices Corporation (Sunnyvale, CA, USA).
Purification of rabbit antibodies and conjugation to digoxigenin (DIG) for TAT ELISA
Three ml rabbit serum, 1/10 diluted with 200 mM sodium phosphate pH 7, was applied on a HiTrap Protein A column (Pharmacia LKB, Uppsala, Sweden). Before application of the sample the column was equilibrated with five column volumes 20 mM sodium phosphate, pH 7. After application of the sample the column was washed with 5 volumes 20 mM sodium phosphate and the rabbit IgG was eluted with 100 mM citric acid, pH 5. The pH of the IgG-elution sample was adjusted with 1 M Tris-HCl to pH 7 and dialyzed overnight at 4°C against PBS. The concentration of the purified IgG samples was adjusted to 1 mg/ml. One ml sample was incubated in a glass tube at RT under continuously shaking with 47 μl DIG-NHS uitleggen? ester (5 mg/ml, Roche Diagnostics). The samples were dialyzed overnight against PBS and frozen in aliquots at -20°C.
In vitro generation of TAT-complexes
TAT-complexes were generated in vitro according to the method from Boisclair et al. (7). Briefly, 50 μg mouse thrombin and 140 μg rat antithrombin were dissolved in 330 μl TAT-buffer (50 mM Tris, 175 mM NaCl, 0.02% sodium azide, 0.1% PEG-6000, 4% BSA, pH 7.9) and incubated for 45 minutes at 37°C. The samples were aliquoted in 20 μl portions and frozen at -20°C.
Western blot analysis
Western blot samples were prepared as follows: 5 μl 5x sample buffer (250 mM Tris-HCl pH 6.8, 10% SDS, 50% Glycerol) was added to 20 μl thrombin (50 μg/ml), 20 μl AT (50 μg/ml) or 20 μl in vitro TAT-complexes. The samples were prepared without addition of β-mercaptoethanol and were not denatured by heating because this results in aggregation of the in vitro TAT complexes in the PEG-6000 containing buffer. Plasma samples were prepared by combining 20 μl plasma, 20 μl 5x sample buffer and 60 μl H2O. Ten μl of each sample was separated on a 7.5% acrylamide gel and transferred by electroblotting to a nitrocellulose filter (Schleicher and Schuell, Dassel, Germany). After washing for 5 minutes with PBS/1%Tween and blocking for 1 hour at RT in PBS/Tween/5% (v/v) Milk the filters were incubated with 1/1000 diluted rabbit serum in PBS/Tween. After three washes of 10 minutes with PBS/Tween the blots were incubated for 1 hour at RT with HRP-conjugated swine anti-rabbit IgG diluted 250-fold in PBS/Tween. Finally, the blots were washed with PBS/Tween and H2O and incubated with a chemoluminescencent substrate (LumiLight, Roche Diagnostics, Mannheim, Germany). The pre-stained low range SDS PAGE markers (Bio-Rad Laboratories, Hercules, CA, USA) were used as molecular weight marker.
Sandwich TAT-complex ELISA
TAT-levels in plasma were detected with the new mouse TAT ELISA as follows: MaxiSorp plates were coated overnight at 4°C with 100 μl purified anti-thrombin antibodies in a concentration of 1 μg/ml in coating buffer. Subsequently, the plates were washed three times with PBS/Tween and incubated with 50 μl standard or 4-fold diluted plasma samples in PBS/Tween/FCS for 1 hour on a shaker at room temperature. The plates were washed three times with PBS/Tween and incubated with 100 μl purified DIG-conjugated anti-AT antibodies in a concentration of 0.5 μg/ml for 1 hour at room temperature. The plates were washed three times with PBS/Tween and incubated for 1 hour at room temperature with 100 μl HRP-conjugated sheep F(ab)2 anti-DIG fragments (Roche Diagnostics) diluted in PBS/Tween. The plates were developed with OPD for 30 minutes and after termination of the reaction with 1 M H2SO4 the OD was determined at 490 and 650 nm. A standard for the mouse TAT ELISA was constructed with two-fold serial dilutions of mouse thrombin – rat antithrombin complexes (Sigma) in PBS/Tween/FCS buffer or mouse plasma.
Detection of TAT-complexes with three different ELISA's
The newly developed TAT-ELISA was compared with two commercially available ELISAs according to the manufacturer's instructions: Enzygnost® TAT micro (DadeBehring BV, Leusden, The Netherlands) and TAT ELISA (Enzyme Research Laboratories (ERL), South Bend, IN, USA). For this purpose, in vitro TAT complexes were measured as standard and diluted in either buffer or normal mouse pool plasma. In addition, TAT-levels measured in plasma from control and LPS-treated (2 mg/kg i.p., 6 hours) mice, were compared.
Plasma collection procedures
Anti-coagulation with sodium citrate
Two methods of anticoagulation with sodium citrate were compared. In the first procedure, 900 μl blood was drawn from the vena cava into a syringe containing 100 μl 3.2% (w/v) sodium citrate. For the second approach 3.2% (w/v) sodium citrate in a total volume of body weight (gr) / 13 × 100 μl was i.v. administrated in the vena cava 20–30 seconds prior to blood drawing from the same vein into a syringe. Blood samples were centrifuged for 15 minutes at 3000 rpm at RT, subsequently plasma was centrifuged for 5 minutes at 13 000 rpm to remove remaining cells and platelets, and immediately frozen at -80°C.
Neutralizing heparin in plasma
Mice were anticoagulated by intravenous injection of 0, 4, 20 or 400 units heparin (Leo Pharmaceuticals, Weesp, The Netherlands), diluted in approximately 200 μl water for injection. After 5 minutes blood samples of approximately 700 μl were obtained by exsanguination via the vena cava into a syringe containing 0.1 volume 3.2% (w/v) sodium citrate. The plasma samples were centrifuged for 10 minutes at 3000 rpm and immediately frozen at -80°C until use. After thawing, a quarter of a heparinase tablet (Dade Hepzyme, Dade Behring, Marburg, Germany) was added to 150 μl of plasma sample in a plastic tube, mixed gently and incubated at room temperature for 15 minutes. Then, activated partial thromboplastin time (APTT) and factor VIII were measured with standard procedures with a Behring Coagulation System (BCS, Dade Behring) using human reagents as provided by the manufacturer. For factor VIII measurements human FVIII deficient plasma (Dade Behring) was used.
Statistical analysis
Data are presented as mean with SEM. Differences between two groups were tested with students' t test. Differences between three groups were tested by one way ANOVA. P < 0.05 was set as threshold of statistical significance. All computations were performed using SPPS 11.0.
Results
Development of a mouse TAT-ELISA
Generation of rabbit antibodies against mouse thrombin and rat antithrombin
Rabbits were immunized with mouse thrombin or with rat antithrombin (AT). Before immunization neither of the rabbits exhibited immunoreactivity towards (xenogenic) thrombin and AT (data not shown). Twenty-five days after immunization the rabbits exhibited a response against the immunizing protein, which reached maximal levels 50 days after immunization. Serum immune titers of each rabbit were determined by ELISA. Serum samples of rabbits immunized with thrombin could be diluted more than 25,000-fold before the reactivity with thrombin reached background level (data not shown). Serum samples of rabbits immunized with AT could even be diluted 100,000-fold before reaching background signal (data not shown). Since antibodies from two rabbits recognized thrombin or antithrombin with equal immunoreactivity, sera from each pair of animals were pooled and used for further experiments (data not shown).
Western blot analysis showed that serum from rabbits immunized with thrombin, reacted only with thrombin (lane T) and not with antithrombin (lane AT) (Figure 1A). In the lane loaded with in vitro TAT complexes, two bands were observed of respectively 39 kD and 94 kD. The 39 kD band represents thrombin, whereas the 94 kD corresponds to the reported molecular weight of TAT complexes [7]. Besides prothrombin (72 kD), TAT-complexes were detected in plasma samples from LPS-treated mice (Figure 1A, lanes 1 through 4), indicating proper recognition of TAT-complexes in mouse plasma by the rabbit α-thrombin antibodies.
Figure 1 Antithrombin (lane AT), thrombin (lane T) and in vitro TAT complexes (lane TAT) were separated on a 7.5% acrylamide gel and transferred to a nitrocellulose filter. Plasma samples of control mice (lane 1) or mice treated with 50 μg/ml LPS for 2 (lane 2), 6 (lane 3) and 24 hours (lane 4) were separated also. The filters were incubated with 1000-fold diluted serum from a rabbit immunized with thrombin (A), or serum from a rabbit immunized with AT (B).
With serum obtained from rabbits immunized with mouse AT, opposite results were obtained. No band was detectable in the lane loaded with thrombin while a 58 kD band was detectable in the lane loaded with AT (Figure 1B). The two upper bands, also detected in the plasma samples, are multimeric complexes of AT. Again, two bands were detected in the TAT-lane: one band of 58 kD, representing uncomplexed AT, and a 94 kD TAT complex band. Western-blot analysis of plasmas obtained from LPS-treated mice showed the presence of TAT-complexes recognized by the rabbit α-AT antibodies (Figure 1B, lanes 1 through 4).
Development of TAT-complex sandwich ELISA
To discriminate and select between capture and conjugating antibody, protein-A purified rabbit α-thrombin antibodies and rabbit α-AT antibodies were conjugated to digoxigenin (DIG). In brief, using 96-wells plates coated with thrombin or AT it was confirmed that DIG was successfully conjugated to both antibodies (data not shown). Subsequently, 96-wells plates were coated with either rabbit α-thrombin antibodies or rabbit α-AT antibodies in a concentration of 1 μg/ml. Thereafter, the plates were incubated with two-fold serial dilutions of in vitro TAT-complexes. Plates coated with rabbit α-AT antibodies were subsequently incubated with serial dilutions of DIG-conjugated rabbit α-thrombin antibodies, whereas plates coated with rabbit α-thrombin antibodies were incubated with serial dilutions of DIG-conjugated rabbit α-AT antibodies.
The use of rabbit α-thrombin capture antibodies was superior to rabbit anti-rat antithrombin capture antibodies in combination with DIG-conjugated rabbit α-thrombin as detection antibodies in the detection of TAT-complexes, even at low concentration of secondary antibody (Figure 2A and 2B). Furthermore, the background signal was virtually zero in the absence of TAT-complexes and a 2000-fold dilution of DIG-conjugated rabbit α-AT antibodies. Detection of TAT-complexes using the combination of rabbit -AT antibodies as capture antibody and DIG-conjugated rabbit α-thrombin antibodies as detection antibody showed lower specificity compared to the reverse setup (Figure 2B). Therefore, for further experiments the rabbit α-thrombin antibodies and DIG-conjugated rabbit α-AT antibodies were used as capture antibody and detection antibody, respectively.
Figure 2 Selection of capture and conjugating antibody. 96-wells plates were coated with either rabbit anti-mouse thrombin antibodies (Panel A) or rabbit anti-rat antithrombin antibodies (Panel B) prior to incubation with two-fold serial dilutions of in vitro TAT-complexes. Plates coated with rabbit α-AT antibodies were subsequently incubated with serial dilutions of DIG-conjugated rabbit α-thrombin antibodies, whereas plates coated with rabbit α-thrombin antibodies were incubated with serial dilutions of DIG-conjugated rabbit α-AT antibodies.
To further prove the possibility that positive results were caused by cross reactivity of rabbit α-thrombin antibodies with AT or vice versa, plates coated with rabbit α-thrombin antibodies were incubated with serial dilutions of thrombin or AT alone and thereafter incubated with DIG-conjugated rabbit α-AT antibodies. Neither thrombin nor AT alone were detected in this analysis (data not shown).
In vitro TAT-complexes were generated in order to be used as standard. The linearity of standard diluted in buffer was similar to the linearity upon dilution in mouse plasma or pooled human pool plasma (data not shown). While lower OD's were detected for standard diluted in human plasma as compared to buffer, the slopes of the linear part of the curves of standard diluted in human plasma (slope = 0.45, r2 = 0.996) or buffer (slope = 0.43, r2 = 0.998) were similar.
Using frozen aliquots of in-vitro TAT-complexes as standard, the intra- and inter-assay coefficients of variation (CV) of the ELISA were determined. The intra-assay CV were assessed by measuring TAT values in three samples each from different aliquots (n = 5) on one plate. The mean TAT concentration in the three samples was 5.1, 11.3 and 15.4 ng/ml and the corresponding CV were 7.7, 6.6 and 5.5 %. The three samples with different TAT levels were also analyzed on five separate ELISA plates and the inter-assay CV were 7.0, 5.1 and 4.7 %, respectively.
Comparision of three TAT ELISA's in vitro
Detection of TAT-levels in standard prepared from in vitro TAT-complexes was compared between three ELISA methods: 1. the mouse specific TAT ELISA as described in Materials and Methods, 2. the human specific Enzygnost® TAT micro from Dade Behring, and 3. the human specific TAT ELISA from ERL. Using the human specific TAT ELISA from ERL, a standard curve comparable to the mouse specific TAT ELISA was obtained by serial dilution of standard in buffer (Figure 3A). In contrast to the mouse specific ELISA, both commercial kits failed in detecting in vitro prepared TAT-complexes after serial dilutions in mouse plasma (Figure 3B).
Figure 3 Comparison of three TAT ELISA's using in vitro TAT-complexes diluted in buffer (Panel A) or normal mouse pool plasma (Panel B). In vitro TAT-complexes were generated according to the method from Boisclair et al. (8). TAT-levels in serial dilutions of standards were measured by the mouse specific TAT ELISA (□), the human specific TAT ELISA from Enzyme Research Laboratories (△), and the human specific Enzygnost® TAT micro from Dade Behring (○).
Plasma collection procedures
Anticoagulation with sodium citrate
Two different techniques of anticoagulation with sodium citrate were tested in mice. Collection of venous blood from the vena cava in a syringe with 0.1 volume sodium citrate resulted in activation of blood coagulation in four out of sixteen plasma samples, as indicated by high TAT-levels > 100 ng/ml, which indicated undesired activation (Figure 4). Plasma samples obtained by i.v. injection of sodium citrate (180 μl per mouse) via the vena cava showed little or no activation of blood coagulation, since fourteen samples had undetectable TAT-levels of 0 ng/ml and two with levels of 2 and 13 ng/ml (Figure 4).
Figure 4 TAT-complex levels in mouse plasma after two different blood drawing techniques, determined with the home made TAT assay. Plasma samples were obtained by either collection of venous blood in a syringe filled with 180 μl 3.2 %(w/v) citrate (Citrate in Syringe) or by i.v. injection of 3.2% (w/v) sodium citrate (body weight (gr) / 13 × 10 μl) into the vena cave 20–30 seconds before blood drawing (Citrate in Vein).
Neutralizing heparin in plasma samples
To assess the influence of in vivo injected heparin and its neutralization with heparinase, APTT and factor VIII activity were measured. Intravenous injection of 4, 20 and 400 units of heparin prolonged APTT similarly in mouse plasma by 300 seconds. Neutralizing heparin through addition of heparinase shortened the APTT significantly (p < 0.0001) (Figure 5A). The APTT was the same (p = 0.6, tested by ANOVA) for plasma from mice treated with 4 or 20 units of heparin followed by heparinase neutralization compared to control plasma. The APTT of plasma pretreated with 400 U of heparin also shortened after heparinase addition, though it remained prolonged (169 seconds ± 47) compared to control plasma (29 seconds ± 1). Factor VIII activity was also normalized to normal levels after treatment with heparinase in plasma from mice treated with 4 or 20 units of heparin (p = 0.6, tested by ANOVA) (Figure 5B). The level of factor VIII in plasma fro mice treated with 400 units of heparin increased after heparinase addition (51 ± 5%) but not to the level of control plasma (125 ± 18%).
Figure 5 Neutralizing effect of heparinase on different amounts of heparin was tested. A) APTT returned to normal levels after heparinase addition to plasma samples from mouse treated with 4 and 20 units of heparin per mouse. B) Also, factor VIII activity returned to normal levels after heparinase treatment to plasma samples from mouse treated with 4 and 20 units of heparin.
TAT measurement in vivo: the effects of LPS administration
To study the kinetics of TAT levels in plasma samples of endotoxin-treated mice and to evaluate the TAT ELISA, plasma samples were obtained at various time points after i.p. LPS administration. In plasma of control mice the TAT concentration was 1.6 ± 0.4 ng/ml (Figure 6). A 10-fold increase in TAT values up to 20.9 ± 2.0 ng/ml was observed in plasma samples obtained 4 hours after LPS administration. After 24 hours the TAT levels were back at baseline levels. The TAT levels on t = 0.5, t = 1.5 and t = 4 hours differed significantly from control levels (p < 0.05, unpaired Student' t-test).
Figure 6 Increased TAT-complex levels in plasma of LPS-treated mice. Plasma samples were collected at 0, 0.5, 1.5, 4, 12, and 24 hours after i.p. injection of 2 mg/kg LPS. Values are means of triplicates ± SD. *; p < 0.05, and **; p < 0.0001.
It should be noted that the actual variation in peak TAT levels was very small in this experiment, which is in contrast with the observed range in values in the next experiment that was done at a later stage in different mice (Figure 7).
Figure 7 TAT-complex levels in mouse plasma without or after 6 hours of endotoxin treatment (2 mg/kg) as measured with the mouse specific TAT ELISA (panel A), the human specific TAT ELISA from Enzyme Research Laboratories (panel B), or the human specific Enzygnost® TAT micro from Dade Behring (panel C). No correlation was found between TAT-concentration measured with our own ELISA and either the human specific TAT ELISA from Enzyme Research Laboratories (slope = 33.27, r2 = 0.360), or the human specific Enzygnost® TAT micro from Dade Behring (slope = 0.73, r2 = 0.132).
For comparison of in vivo formed TAT complexes in mouse plasma by the three different ELISA methods, the standards supplied in each commercial assay were diluted according to the manufactures instructions and used to calculate TAT-concentrations. Plasma samples were obtained from control mice and from mice six hours after i.p. LPS administration (2 m/kg) by drawing venous blood after i.v. administration of sodium citrate as described before. In plasma of control mice TAT concentration was 2.9 ± 2.9 ng/ml, 9.2 ± 7.7 ng/ml, and 1.1 ± 0.20 ng/ml, measured with respectively the Dade Behring, ERL, and our own ELISA (Figure 7). After 6 hours of endotoxemia blood coagulation was activated as indicated by a 30-fold increase in TAT-levels as measured with the mouse specific TAT ELISA (35.7 ± 21.1 ng/ml; p < 0.05 as compared to control plasma) (Figure 7), whereas TAT levels measured by the ERL and Dade Behring ELISA's were 2588.5 ± 1234.0 ng/ml and 37.4 ± 32.4 ng/ml, respectively. Of note, the response detected ex vivo by the three assays was quite comparable, whereas the in vitro comparison suggested that at least in plasma our TAT assay would be superior as compared to the commercial assays. We have no proper explanation for this discrepancy. Theoretically, a difference in affinity of the assay antibodies against the mouse-mouse TAT in ex vivo plasma as compared to the mouse-rat TAT complexes in vitro may play a role. Alternatively, the commercial assays could lack specificity for mouse TAT complexes, which could also explain the considerabble differences in TAT concentrations that are between 10–100 fold higher in the ERL assay as compared to our TAT method. We have no evidence to substantiate either of these explanantions, other than that a weak correlation, as measured with Spearman correlation test, between the mouse specific TAT ELISA and the human specific ELISA from ERL r = 0.600; p = 0.051 suggests that a similar product is being detected. No correlation was found between TAT-concentration measured with the mouse specific ELISA and the Dade Behring ELISA. A correlation was found between the two commercial human specific ELISAs (r = 0.694; p < 0.05).
Discussion
Studying the relationship between coagulation changes in blood and at tissue level in mice requires the availability of tools for the analysis of coagulation activation. In this study we evaluated several procedures for plasma preparation and the measurement of coagulation activation in a mouse model of endotoxemia. First, we described the development of a specific mouse sandwich ELISA to measure the level of TAT complexes in mouse plasma. Second, we compared three different sandwich ELISAs for measurement of TAT levels in mouse plasma. Third, we tested different anticoagulation techniques to prevent ex vivo clotting.
Using the new TAT ELISA we detected changes in mouse plasma TAT content 4 hours after endotoxin infusion. Previously, it was demonstrated that endotoxin treatment of mice resulted in increased TAT complexes [8], TF mRNA synthesis [9] and enhanced tissue deposition of fibrin [10]. In human volunteers endotoxin infusion induced monocyte TF mRNA expression and plasma F1+2 and TAT levels [11]. Therefore, we expected an increase of TAT levels in plasma of mice treated with endotoxin. Indeed, basal TAT levels of 1.6 ± 0.4 ng/ml were increased in mice treated with endotoxin with maximal levels of 20.1 ± 9.9 ng/ml after 4 hours.
In addition to the mouse specific TAT ELISA, the two commercially available human TAT ELISAs, Enzygnost® TAT micro from Dade Behring and the TAT ELISA from Enzyme Research Laboratories also detected increases in TAT levels after endotoxin treatment in mice. However, significant differences between the three assays were observed. The levels of TAT complexes as measured with the ERL assay were about 70 fold higher as compared to the two other assays. Furthermore, the Dade Behring assay could not detect in vitro prepared TAT complexes diluted in either buffer or plasma and in vivo TAT levels did not correlate to the levels measured with the mouse specific ELISA. Since plasma levels of TAT complexes measured by the ERL ELISA correlated to levels measured with the mouse specific ELISA, we speculate that this assay has probably a slightly better specificity for mouse TAT complexes than the Dade Behring assay. In addition, the ERL ELISA could measure in vitro prepared TAT complexes diluted in buffer. Still, the substantial variation in measured TAT concentrations suggests that the ERL assay, developed for use in human plasma, is not sufficiently specific for mouse TAT complexes. In additioon to these practical considerations and limitations, the considerable costs of the commercial human assays makes by comparison our novel assay to a cheaper and probably reliable alternative for the analysis of mouse TAT complexes in plasma samples.
For ex vivo anticoagulation sodium citrate is the most commonly used reagent added to human plasma. Since we experienced that anticoagulation with sodium citrate or EDTA in the syringe did not always effectively prevent ex vivo clotting of mouse blood, an alternative method of anticoagulation was sought. This led to a novel method to anticoagulate the blood samples in which sodium citrate was injected intravenously into the vena cava 20–30 seconds before blood collection by exsanguination. In our experiments, TAT levels were not increased in any of the plasma samples obtained using the i.v. injection of sodium citrate method, in contrast to the standard method in which sodium citrate was added to blood in the syringe. Therefore, we conclude that anticoagulation of mouse blood by i.v. injection of sodium citrate into the circulation results in more reliable TAT data and in plasma more suitable for further analysis of coagulation and related factors.
For immunohistochemical analyses of tissues pathologists may prefer the infusion of heparin prior to sacrificing the animal. To determine whether this anticoagulant effect could be eliminated to allow other assays including plasma based clotting tests, we tested the neutralizing activity of heparinase in mouse plasma. In accordance with results in human plasma we confirmed that up to 20 units of heparin per mouse, prolonging the APTT to more than 300 seconds, could be adequately neutralized with heparinase as indicated by normalized FVIII activity and APTT levels [6].
Conclusion
In conclusion, the home made ELISA for mouse TAT complexes detects thrombin generation after endotoxin challenge in mice. Applied to blood samples obtained after systemic citrate infusion, this methodology may improve the quality of studies aimed at elucidating the pathophysiology of coagulation activation in mice.
List of abbreviations
aPTT activated partial thromboplastin time
AT antithrombin
BSA bovine serum albumin
HRP horseradish peroxidase
LPS lipopolysaccharide
PBS phosphate buffered saline
RT room temperature
TAT Thrombin-Antithrombin complex
Authors' contributions
Dirkje W. Sommeijer was involved in planning, experimental setup, development of methods, analysis of samples and writing the manuscript, René van Oerle was involved in murine blood sample collections, development of methods, and analysis of plasma samples, Pieter H. Reitsma was involved in experimental setup, analysis, and writing the manuscript, Janneke J. Timmerman was involved in design and development of antibodies against thrombin and antithrombin, Joost C.M. Meijers was involved in analysis, experimental setup, and writing of the manuscript, Henri M.H. Spronk was involved in animal work, development of methods, and writing of the manuscript, Hugo ten Cate was involved in planning, experimental setup, and writing of the manuscript.
Acknowledgements
We like to thank Willy Morriën-Salomons, Joost Daalhuisen and Ingvild Kopp from the Academic Medical Center for their assistance. Hugo ten Cate is a Clinical Established Investigator of the Netherlands Heart Foundation. A short summary of the TAT laboratory procedure was previously published in Weijer et al. 2004, Journal of Infectious Diseases, 198 (12): 2308–2317
==== Refs
Dorffler-Melly J Schwarte LA Ince C Levi M Mouse models of focal arterial and venous thrombosis Basic Res Cardiol 2000 95 503 509 11192373 10.1007/s003950070028
Kerlin B Cooley BC Isermann BH Hernandez I Sood R Zogg M Hendrickson SB Mosesson MW Lord S Weiler H Cause-effect relation between hyperfibrinogenemia and vascular disease Blood 2004 103 1728 1734 14615369 10.1182/blood-2003-08-2886
Schafer K Muller K Hecke A Mounier E Goebel J Loskutoff DJ Konstantinides S Enhanced thrombosis in atherosclerosis-prone mice is associated with increased arterial expression of plasminogen activator inhibitor-1 Arterioscler Thromb Vasc Biol 2003 23 2097 2103 14512369 10.1161/01.ATV.0000097766.36623.DF
Boisclair MD Lane DA Wilde JT Ireland H Preston FE Ofosu FA A comparative evaluation of assays for markers of activated coagulation and/or fibrinolysis: thrombin-antithrombin complex, D-dimer and fibrinogen/fibrin fragment E antigen Br J Haematol 1990 74 471 479 2189490
Pelzer H Schwarz A Heimburger N Determination of human thrombin-antithrombin III complex in plasma with an enzyme-linked immunosorbent assay Thromb Haemost 1988 59 101 106 3363526
Manco-Johnson MJ Nuss R Jacobson LJ Heparin neutralization is essential for accurate measurement of factor VIII activity and inhibitor assays in blood samples drawn from implanted venous access devices J Lab Clin Med 2000 136 74 79 10882230 10.1067/mlc.2000.107299
Lane DA Caso R Antithrombin: structure, genomic organization, function and inherited deficiency Baillieres Clin Haematol 1989 2 961 998 2688761
Sugatani J Igarashi T Munakata M Komiyama Y Takahashi H Komiyama N Maeda T Takeda T Miwa M Activation of coagulation in C57BL/6 mice given verotoxin 2 (VT2) and the effect of co-administration of LPS with VT2 Thromb Res 2000 100 61 72 11053618 10.1016/S0049-3848(00)00305-4
Mackman N Sawdey MS Keeton MR Loskutoff DJ Murine tissue factor gene expression in vivo. Tissue and cell specificity and regulation by lipopolysaccharide Am J Pathol 1993 143 76 84 8317556
Yamamoto K Loskutoff DJ Fibrin deposition in tissues from endotoxin-treated mice correlates with decreases in the expression of urokinase-type but not tissue-type plasminogen activator J Clin Invest 1996 97 2440 2451 8647936
Franco RF de Jonge E Dekkers PE Timmerman JJ Spek CA van Deventer SJ van Deursen P van Kerkhoff L van Gemen B ten Cate H van der Poll T Reitsma PH The in vivo kinetics of tissue factor messenger RNA expression during human endotoxemia: relationship with activation of coagulation Blood 2000 96 554 559 10887118
|
16115315
|
PMC1215525
|
CC BY
|
2021-01-04 16:36:24
|
no
|
Thromb J. 2005 Aug 22; 3:12
|
utf-8
|
Thromb J
| 2,005 |
10.1186/1477-9560-3-12
|
oa_comm
|
==== Front
Virol JVirology Journal1743-422XBioMed Central London 1743-422X-2-591607640310.1186/1743-422X-2-59ReviewMacrophages and cytokines in the early defence against herpes simplex virus Ellermann-Eriksen Svend [email protected] Department of Clinical Microbiology, Aarhus University Hospital, Skejby Sygehus, Brendstrupgaardsvej 100, DK-8200 Aarhus N., Denmark2005 3 8 2005 2 59 59 5 7 2005 3 8 2005 Copyright © 2005 Ellermann-Eriksen; licensee BioMed Central Ltd.2005Ellermann-Eriksen; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Herpes simplex virus (HSV) type 1 and 2 are old viruses, with a history of evolution shared with humans. Thus, it is generally well-adapted viruses, infecting many of us without doing much harm, and with the capacity to hide in our neurons for life. In rare situations, however, the primary infection becomes generalized or involves the brain.
Normally, the primary HSV infection is asymptomatic, and a crucial element in the early restriction of virus replication and thus avoidance of symptoms from the infection is the concerted action of different arms of the innate immune response. An early and light struggle inhibiting some HSV replication will spare the host from the real war against huge amounts of virus later in infection. As far as such a war will jeopardize the life of the host, it will be in both interests, including the virus, to settle the conflict amicably. Some important weapons of the unspecific defence and the early strikes and beginning battle during the first days of a HSV infection are discussed in this review.
Generally, macrophages are orchestrating a multitude of anti-herpetic actions during the first hours of the attack. In a first wave of responses, cytokines, primarily type I interferons (IFN) and tumour necrosis factor are produced and exert a direct antiviral effect and activate the macrophages themselves. In the next wave, interleukin (IL)-12 together with the above and other cytokines induce production of IFN-γ in mainly NK cells. Many positive feed-back mechanisms and synergistic interactions intensify these systems and give rise to heavy antiviral weapons such as reactive oxygen species and nitric oxide. This results in the generation of an alliance against the viral enemy.
However, these heavy weapons have to be controlled to avoid too much harm to the host. By IL-4 and others, these reactions are hampered, but they are still allowed in foci of HSV replication, thus focusing the activity to only relevant sites. So, no hero does it alone. Rather, an alliance of cytokines, macrophages and other cells seems to play a central role. Implications of this for future treatment modalities are shortly considered.
==== Body
Introduction
Virus-host interactions are crucial for the outcome of infections. Several strategies have been utilized by viruses to overcome the host defence. For the virus to be successful, these evasive strategies have to be balanced with the pathology induced and the possibilities of transmission to new susceptible individuals. The mammalian host utilizes ubiquitous and redundant antiviral defence mechanisms. In different viral infections, different parts of the host defence seem to be crucial. However, the redundancy ensures that other systems are ready to take over, if one of them fails. The final outcome of a viral infection depends on a delicate regulation and timing of these antiviral effector mechanisms in response to the invading virus.
A viral infection of an individual thus involves a conflict between the virus and the host, which could conceptually be viewed upon as a human controversy escalating to invasion and armed struggle. To understand the resulting course of events it is important to know each party of the conflict and to conduct an analysis of the powerful weapons held by each of the combatants. The present review analyzes the early non-specific events in the conflict upon herpes simplex virus (HSV) infection. Initially, each participant of the conflict, the infecting HSV and the non-specific antiviral weapons of the host, are described. Subsequently, the early events of the conflict, the armament, early strikes and the opening battle between HSV and the host are discussed. Insight into the early non-specific defence mechanisms are important for our understanding of the conflict and may indicate how to intervene during serious systemic infections.
The combatants – facts and hypotheses on function
Herpes Simplex Virus
Herpesviruses are ubiquitous viruses generally infecting humans early in life. The majority of humans has had a primary infection with one or more herpesviruses and harbour these viruses in a latent state for the rest of their lives. The initial infection is most often asymptomatic, but can be symptomatic depending on the herpesvirus in question and the age and immune status of the host. The viruses are phylogenetically old and humans and herpesviruses have evolved together [1]. This co-evolution has created viruses which are well adapted to the human host and environment. Thus, herpesviruses are capable of coping with the human immune defence in a balanced manner generally without serious threads to the life of the host. Infection with a foreign herpesvirus, normally hosted by another species, does not always hold this balance, and the pathology is unpredictable. This is seen when humans are infected with the simian B virus, which often shows serious clinical outcome [2].
The human herpes simplex viruses were initially identified by Lowenstein, who passed it onto rabbits in 1919, and found it to be sensitive to alcohol and higher temperatures [3]. The viruses were classified into two serologically different types by Schneweiss in 1962 [4], and these are now known to belong to the subfamily of Alphaherpesvirinae together with varicella-zoster virus. These alphaherpesviruses all show neurotropic latency, and mucosal or skin lesions are frequently seen as a result of viral reactivation from sensory nerves. The two types of herpes simplex virus confer the genera Simplexvirus 1 and -2, which were formally designated by the International Committee on Taxonomy of Viruses as Human herpesvirus (HHV) 1 and 2 [5].
Herpes simplex virus (HSV) type 1 and type 2 are very closely related, showing a homology at the DNA level of 83% in protein coding regions and less in noncoding regions [6]. The genetic map of the two herpes simplex viruses is colinear [6], and the genomes are of approximately the same size, HSV-1 of 152 kbp [7] and HSV-2 of 155 kbp, and code for corresponding genes [6]. The minor sequence variations give different cleavage sites for restriction endonucleases, which has been used intensively as an important epidemiological tool [8-10].
Structure of herpes simplex virus
As all other herpesviruses the herpes simplex viruses are enveloped, icosahedral DNA viruses with a capsid of approximately 100 nm (fig. 1)[1]. The envelope holds at least 10 different glycoproteins protruding from the outer side (gB, gC, gD, gE, gG, gH, gI, gK, gL, and gM). The glycoproteins are primarily responsible for attachment to cellular receptors and fusion of membranes (especially gB and gD) [11-14]. In addition, there are two unglycosylated proteins in the viral envelope. The glycoproteins of the envelope have several immunoregulatory effects besides their primary more mechanical functions in viral attachment and entry [15-19].
Figure 1 HSV composition and entry. Electron micrograph of negatively stained HSV particle with indications of major structural elements. Important mediators of adsorption to cells (1), receptor binding (2) and fusion of membranes (3) during the process of infection are drawn stylistically.
In the space between the envelope and the capsid, the complete viral particles posses an almost amorphous structure which was termed the tegument by Roizman and Furlong [20]. The tegument consists of several viral proteins involved in the initial phases of viral infection and replication such as transport of the viral DNA out of the capsid [21], early shutoff of cellular protein synthesis (vhs) [22], and initiation of transcription of viral genes (α-trans-inducing factors) [23]. Besides the tegument seen in complete viral particles, tegument-like structures are seen in enveloped particles lacking a capsid and DNA, the so called light particles [24,25].
The capsid is composed of a complex icosahedral structure of 162 capsomeres, each with a central channel running from the outside to the interior of the capsid. Inside the capsid the double stranded linear DNA is packed as a spool with the ends in close proximity [21,26,27]. The genome consists of a long (L) and a short (S) segment which are covalently linked [28], and contains a high density of genetic information with about 94 open reading frames (ORF) and encodes approximately 84 polypeptides [7,29], of which only 37 are required for replication of the virus in cultured cells [30,31]. The viral genes are expressed in a cascade in groups classified as immediate early (IE, α), early (β), early late (γ1) and late (γ2) genes, each with a certain characteristic group of promoters regulating the sequential expression [29,32]. Generally, the α-gene products are transcription inducers, the β-gene products are viral enzymes such as the thymidine kinase and the viral DNA polymerase, and the products of γ-genes are the structural proteins of the viral particle [33]. The viral transcriptional chain is closed by some of the tegument proteins (e.g. VP16/Vmw65) which are γ-gene products with structural properties in the tegument of the viral particle and besides this harbour transcription-inducing capacity upon α-gene promoters crucial in the induction of the next replication cycle of the virus [32,34].
Infection of the cell
The HSV infection is initiated by adsorption of the viral particle via gB or gC to a cellular receptor, which is a heparan sulphate chain on cellular proteoglycans [35]. Thus HSV adsorption can be inhibited by heparin and soluble heparan sulfates [36,37]. This initial binding, in which gC is important but dispensable, is of greater significance for HSV-1 than for HSV-2, a divergence which could have implications for the different pathogenic patterns of the two strains [38,39]. Furthermore, trapping of HSV to heparan sulfate motives in the tissues, e.g. basal laminas, may be of importance for containment of the infection at a specific site [40]. Binding to the heparan sulfate-containing cellular receptors, which are in size with the HSV particle itself, is reversible, and serves to concentrate the viral particle in near proximity to the cell (fig. 1) [35,39,41].
A crucial step is then conducted by gD binding to an entry receptor, of which three classes has been described [42]. These include herpes virus entry mediator (HVEM), later designated as herpes virus entry protein A (HveA), which is a member of the tumour necrosis factor receptor family, nectin-1 (HveC) and nectin-2 (HveB), both members of the immunoglobulin superfamily, and heparan sulfate sites modified by 3-O-sulfotransferases [43-46]. The differential use of these receptors is of importance for HSV entry of different cell types and infection of polarized cells [47-51], exemplified by nectin-1, which is of importance in infection of the vaginal mucosa [52]. Upon binding to one of these entry receptors, conformational changes in gD lead to interaction with gB or gH-gL dimer, which results in membrane fusion by a mechanism not known in detail (fig. 1) [41,53]. The membrane fusion can take place both with the plasma membrane on the surface of the target cell and with an endosomal membrane after intraluminal pH-reduction, as it is seen for some other enveloped viruses [50,54-56].
Following these initial steps of infection several immunomodulatory cellular events are induced, but the potential importance of signalling through receptors involved in adsorption and membrane fusion is only scarcely analysed [57]. The receptor molecule HVEM is by its normal ligand capable of inducing activation of nuclear factor κB (NF-κB) and activation of T cells. By interaction with HSV-gD these receptor responses are inhibited. Thus, the HSV interaction with at least one of its receptors has multiple potentials for modulation of the host response to the infection [58,59].
Replication and formation of progeny virus
Upon fusion, the HSV nucleocapsid is transported by microtubules to a nuclear membrane pore where the viral DNA is released into the nucleus [60,61]. Both viral tegument products and cellular kinases are responsible for the initiation of α-gene transcription [62]. In these initial events the determination of whether it will lead to a lytic infection cycle or a latent infection seems to be directed largely by the infected cell type in question [63,64]. A key event in this seems to be early induction of latency-associated transcripts (LATs) with sequences antisense to the infected cell protein null (ICP0) and ICP4 [65-67]. In the initial phase of the lytic replication cycle, the IE-gene products, besides being transcription factors for the next wave of viral proteins, intimately regulate cellular functions in favour of viral replication and immune evasion [33,68]. Of these, the ICP0, a promiscuous transactivator without much DNA-binding capacity, forces the cell to a pre-dividing state optimal for viral protein synthesis [69,70]. Furthermore, ICP0 is active in inhibiting immune mechanisms such as interferon production and antiviral effects of interferons [71-73] and induces degradation of cellular proteins, involving the proteasome [74,75].
Very early in infection, the first transcriptional activity is seen just inside the nuclear membrane at the site were the viral DNA enters the nucleus [76]. The produced ICP0 co-localizes with the promyelocytic leukaemia (PML) nuclear bodies and initiates degradation of these, an event which seems to be important for productive replication of the virus [77,78]. ICP4 binds to parental viral DNA which is juxta-localized to the PML bodies, and later, when the bodies are degraded, replication compartments are formed, in which also ICP27 can be found [76,79,80]. ICP27 affects the posttranscriptional polyadenylation and splicing of RNA, and it is thus an element of the delayed host protein shutoff [81]. Immune evasion is additionally induced by the IE protein ICP47 which binds to transporter associated with antigen processing, TAP1/TAP2 and blocks the presentation of viral peptides by the major histocompatibility complex (MHC)-I [82].
The HSV progeny is formed in the nucleus of the infected cell, where the viral DNA is packed into preformed capsids. These are assembled with the tegument proteins and bud through the inner nuclear membrane to the perinuclear space [11]. The route of virus from here to the external side of the cell is controversial. Apparently two routes of viral egress are possible [11]. One way is by continuous passage through vesicles and the Golgi apparatus, where the membrane proteins are modified. The other route is by fusion of the newly acquired envelope with the outer nuclear membrane or the membrane of a vesicle, generating naked nucleocapsids in the cytoplasm. From here a new budding event should take place, for instance into the Golgi apparatus. The progeny virus thus acquires the envelope from other membranes, than the inner nuclear lamella, as it is indicated by analysis of membrane lipids [83]. Increasing evidence is pointing at this latter possibility of de-envelopment and re-envelopment as the dominating route of HSV egress [84-86].
The progress of HSV infection in tissues is influenced by the capacity of HSV to infect adjacent cells directly through cell junctions. The virus is thus avoiding exposure to extracellular substances such as antibodies and complement. The glycoproteins gE and gI are crucial for this kind of polarised transmission which primarily takes place in epithelial infections [47,87].
Epidemiology
As it is the case at the molecular level, the two herpes simplex viruses show similarities in their clinical appearance, both giving rise to primary infections of mucosal membranes and showing latency in sensory nerve ganglia [1]. The primary infections with HSV are often asymptomatic, especially at young age, but in a minority of cases vesicular or ulcerative lesions are seen. Although HSV-1 and -2 can give rise to indistinguishable clinical infections, there are differences in the anatomical distribution of these infections, as described in 1967 by Dowdle et al. [88]. HSV-1 is predominantly giving rise to infections above the waist, and HSV-2 to infections below the waist. This pattern is, however, not as straightforward as primarily described. In the last decades changes in both prevalence and distribution of HSV infections have been seen. The overall prevalence of HSV infection is very different in different countries and ethnic and social populations [89-91]. A decline in HSV prevalence has been observed in the western countries, probably because of improved socioeconomic conditions [92-94]. In parallel to the decline in prevalence, the aetiology of herpes genitalis has changed in several countries, presumably because of altered human habits and conditions of life [92]. In some areas of the world the proportion of genital infections caused by HSV-1 is still low (4–20 %) [95,96], but in others the relative proportion of genital herpes caused by HSV-1 is increasing [97,98]. In Norway, approximately half of primary genital HSV infections are caused by the type 1 virus [99], and in young women in Edinburgh, Scotland, 60% of new cases are caused by HSV-1 [100]. This shift of aetiology is probably caused by changes in sexual behaviour, especially oral-genital contact [101,102], and by the decreased prevalence of HSV-1 seropositivity at sexual debut, leaving a larger proportion of young adults permissive for a HSV type 1 infection [99]. Seropositivity to HSV-1 does not render any protection against catching an infection with HSV-2 [92,103,104], but a higher proportion of primary genital HSV-2 infections are asymptomatic in HSV-1 seropositive individuals than in seronegative individuals [103].
The aetiology of a genital infection is not insignificant, in that the frequency of recurrence is higher in HSV type 2-infected individuals than in those infected with type 1 [89,95]. The frequency of primary and recurrent infections with both HSV-1 and -2 has been reported to be higher among women than men [97,103,105]. Overall, these epidemiological changes could have implications for the risk of neonatal infection from vaginal delivery, in that more women are seronegative at delivery and thus a higher number have the risk of caching a primary HSV infection. On the other hand, less HSV is circulating, reducing the risk of those who are susceptible.
Clinical appearance and pathogenesis
As described above, primary infection with HSV is most often asymptomatic, especially in younger children [106]. However, some individuals experience a symptomatic primary infection with vesicular herpetic gingivostomatitis or in adolescence more often a pharyngitis [107]. As it is the case with orofacial infections, a primary genital HSV infection can be both asymptomatic and symptomatic with ulcerative lesions and with or without generalized symptoms such as fever, headache etc. [108,109]. Rarely, the infection disseminates to one or several organs giving rise to infections such as necrotising hepatitis, meningitis, encephalitis or to disseminated intravascular coagulopathy [110-113]. Such a clinical course, although uncommon, is most often seen in immunosuppressed patients e.g. transplant patients, neonates or pregnant women [114-116]. In pregnancy, primary infection with HSV without previous seroconversion at the time of delivery seems to be the main risk factor for infection of the newborn [109,117]. Genital HSV reactivations at labour only seem to posses a minor risk for neonatal infection of the baby [117,118], but in spite of this, approximately 70% of neonates infected are born by asymptomatic women [63]. The amount of virus in vaginal secretions during reactivations is much lower than the amount of virus in primary infections, and in reactivated cases maternal antibodies furthermore seems to be protective for the neonate [117,119-122].
When transmitted, the course of HSV infection in the newborn varies. In the pre-acyclovir era about one third of cases were mucocutaneus infections only involving the skin, mouth and eyes, one third were infections of the central nervous system (CNS) with or without mucocutaneus involvement, and the last third were disseminated infections involving multiple organs, including the liver, lungs, adrenals, and often the CNS [119]. Of these, neonates with a generalized infection had a one-year mortality of approximately 60%, those with CNS-infections had intermediary mortality, and nearly no mortality was seen in the group of patients with only mucocutaneus involvement [119]. In infected with multi-organ involvement the deaths are often set off by infection of liver or lungs or by coagulopathy. Sequelae, such as mental or neurological disabilities are seen in some of those with CNS involvement [123].
Now a day, after initiation of high-dose acyclovir treatment, the mortality and sequela rates have dropped [124]. The clinical pattern of neonatal HSV infections has changed in that less of the mucocutaneus infections disseminate to generalized infections when treated [123]. Even with high-dose acyclovir, improvements in treatment protocols are still needed, because the mortality is still as high as 30% in disseminated infections. Reduction in the time from debut of symptoms to initiation of therapy is vital and passive immunotherapy with HSV-specific antibodies could posses a potential as adjuvant to the antiviral treatment [123,125,126]. Other adjuvant treatment modalities are still needed in both neonatal infections and in generalized infections at later ages.
The pathology of HSV infections is mainly caused by a direct cytopathic effect of the virus, resulting in cellular lysis and focal necrosis of the infected area [119,127,128]. In tissues capable of regeneration, this is not devastating, provided that the lesions do not totally destroy the organ or result in functional disability during the infection. In the brain, however, the capacity for regeneration is small, and larger necroses induced by viral infection will result in life-long sequelae [119,123]. A delicate balance exists between the direct HSV-induced pathology and the immunopathology induced by immune reactions to the virus and the toxic and functional side effects of these reactions [129]. Immunopathogenesis seems to be the main aspect of HSV stromal keratitis, which often leads to blindness [130,131]. The scarification from this infection has even been attributed to autoimmunity by molecular mimicry [132]. Weak immune response to the virus leads to severe infections because of massive viral replication and dissemination. An immense immune reaction, especially with high amounts of virus to trigger a response, can bring about increased symptoms of infection, local symptoms such as high intra-cerebral pressure or pulmonary complications, as well as generalized or septic symptoms [129,133-136].
It is thus clear that early control of HSV replication in the initial phases of infection is crucial for the host. Early containment or at least inhibition of viral replication can prevent dissemination of the infection, and the early non-specific immune reactions thus have the potential to inhibit development of a symptomatic infection. Obviously the host will benefit from an attenuated or asymptomatic course of infection, but HSV – with the potential of subsequent reactivation from a latent site – could also benefit from such a course of infection, in that the host will survive and the activity of the host in society will not be hampered by symptoms from infection. Thus, the HSV has excellent chances to reach new susceptible hosts which bring the virus and the host in a situation of mutual benefit [33].
Macrophages
Macrophages are ubiquitous cells of the mononuclear phagocyte system found throughout the body. Many attempts have been made to classify this range of cells with phagocytic activity. In 1892 Metchnikoff named them macrophages (large eaters) in contrast to microphages (the polymorphonuclear leukocytes)[137], and in 1924 Aschoff defined the reticuloendothelial system by the criteria of uptake of vital dye [138]. The macrophages are now more precisely defined as an important member of the mononuclear phagocyte system, defined in 1969 by van Furth and colleagues [139]. In the tissues they constitute a dynamic pool of cells with many functional capabilities, among which the capacity of phagocytosis, microbial killing, motility, and adherence to surfaces are classic [139].
The macrophages originate from the bone marrow, where proliferating promonocytes give rise to monocytes which enter the blood stream [140]. After a mean circulation time of approximately 11/2 day, the blood monocytes migrate to the tissues [140]. In the tissues the monocytes differentiate into macrophages with characteristics determined by the environment of the tissue in question [141]. The tissue macrophages in the major organs are represented by Kupffer cells in the liver, alveolar and interstitial lung macrophages, spleenic and sinusoidal lymph node macrophages, microglia in the brain, osteoclasts in bone, and Langerhans cells of the skin. Thus, macrophages are strategically situated all over the body taking care of debris from the organism itself and foreign material, among others invading microorganisms, including viruses [142,143]. Macrophages in different organs have different characteristics and functional capabilities and can not totally substitute one another in studies on macrophages [141,144-147]. Likewise, macrophages from different species can possess differences in their functional capability, e.g. the capacity for nitric oxide (NO) production [148,149].
Macrophages in tissues are, as described above, in part originating directly from monocytes, but they are also in part originating from local proliferation. This local proliferation in the tissues is performed by newly recruited monocytes, and in the steady state situation they only constitute a small fraction of the mononuclear phagocytes present [150]. Of the monocytes produced in the bone marrow of mice and passing through the blood, approximately half are targeting the liver, 15 % are going to the lungs, 25 % to the spleen and 7 % to the peritoneal cavity [150-152]. In the lungs, 70% of tissue macrophages in the steady-state originate from monocyte influx and 30% from local proliferation [153]. This proportion might vary between different tissues, as the lifespan of tissue macrophages in different organs also varies from around 6 days in mouse spleen to approximately one month for alveolar macrophages [151,152]. In the skin, Langerhans cells are a very stable and long-lived population of cells staying there for at least 18 month in the steady-state situation. However, in inflammation the Langerhans cells are within 2 weeks replaced and supplemented by circulating mononuclear cells [154]. When an inflammatory process is initiated, the dynamics of monocytes and macrophages are changed. Monocytes and other white blood cells are produced and recruited from the bone marrow, and the white blood cell count in the circulation is increased. The monocytes are mainly passing through the blood to become tissue macrophages, and the number of macrophages in the inflamed tissue can be increased by more than ten times [155]. In inflamed tissue the local proliferation of macrophages does not seem to increase, although the number of newly recruited cells is high, indicating that the differentiation of monocytes in the tissues is accelerated [155].
The differentiation of monocytes and activation of macrophages have been a focus of interest for many years because of the observation that macrophage activation is crucial in the defence against many intracellular pathogens [156-159]. It became clear relatively early that lymphocytes and soluble factors secreted by these (lymphokines) are important in activation of macrophages for killing of intracellular bacteria, e.g. Listeria [160]. In the killing of bacteria, interferon (IFN)-γ was shown to be an important stimulator of macrophage activation [161]. As mechanisms in performance of the killing simple toxic substances of reactive oxygen species (ROS) and nitric oxide were identified and seem to conduct their action in synergy [162-164]. The toxic substances are chemically simple, but their production and regulation in macrophages are very complex and still a matter of intense studies [149].
The state of the activated macrophage has changed conceptually from being viewed as one specific condition of the cell towards a more dynamic picture, provoked by the fact that macrophages activated by different means show different phenotypical characteristics [163,165]. The activated macrophage is now viewed as a cell with floating characteristics of many functional capacities regulated by a multitude of stimulating substances, such as the cytokine environment, hormones, and pathogenic and foreign substances [147,166]. Among variables, controlling macrophage activity in infected individuals, are the genetic constitutions of the host. The genetic background has been shown to be of importance for the regulation of both basic proliferation and function of macrophages and for the more specific antimicrobial responses [167,168].
Cytokines
Soluble mediators of lymphocyte activities were described as early as 1953, but the first lymphokines/cytokines found and characterized were the type I interferons. Soon after, many other soluble mediators of lymphocyte and monocyte/macrophage activities were found [169-171]. The term lymphokine was introduced by Dumonde et al. in 1969, to describe lymphocyte derived factors, and the term monokine was used as a description of factors coming from the mononuclear phagocyte system, both acting on many cells, primarily leukocytes [172]. Because of a broader view on origin and function of these factors, the term cytokine is now more often used. Each cytokine was originally named according to biological activity in a functional assay, which often gave several different names to one cytokine, and thus confusion at the molecular level. To straighten this out, a numerical nomenclature of interleukins (between leukocytes) was introduced in 1979 [173]. This numbering system has clarified the field, but since it has no mnemonic functional anchorage it has drawn critique since then [174-176].
The cytokines are generally smaller proteins, some composed of two subunits, utilizing specific receptors on target cells for induction of their functional effects. They are structurally related in three families, with the prototypes being IL-1, IL-2 and IL-17 [176]. Functionally, cytokines are highly potent regulatory proteins acting in a paracrine or autocrine manner at picomolar concentrations [177]. The cytokine receptors are also structurally clustered in families, and functionally utilize a battery of overlapping kinases and nuclear binding proteins in their signalling pathway and thus have overlapping functions [178]. The final functional capacity of the effector cell thus reflects the cytokine environment experienced by the cell [177]. Thus the cytokines comprise a network of factors inducing or inhibiting each others secretion and function in different cells, giving rise to a constantly floating landscape of a large array of functional capacities [177]. In the early hours of a viral infection, the cytokines produced by cells infected or coming into contact with viral products are vital in conduction of the innate immune response to the infection [168,179].
Interferons
The interferons (IFNs) were described and named in 1957 by Isaacs and Lindenmann [170], who characterized the substances involved in the previously described interference of one virus with the replication of another unrelated virus, and the interfering activity of inactivated influenza virus with the subsequent infection of chorio-allantoic membranes [180-182]. The IFNs were the first cytokines described in detail, and thus provided the fundamental basis for the understanding of the cytokine concept [183]. The IFNs are divided into three major groups. The two original groups of IFNs are designated type I and type II, type I being the so called non-immune IFN, and type II the immune IFN. Type II (IFN-γ) is produced in high amounts as part of a specific immune reaction, whereas the type I IFNs can be produced by many cell types in response to, in immunological terms, non-specific stimulation. The many functions of IFNs and the growing understanding of signalling and regulation indicate that IFN analogues may play a major role in the next generation of new antiviral compounds [171].
The type I IFNs are a diverse group of cytokines, consisting of IFN-α, IFN-β, IFN-ε, IFN-κ, IFN-ω, IFN-δ, IFN-τ, and IFN-ξ/limitin [171,184]. The first five of these are expressed in humans, and their relative production depends on the stimulus and the cell type in question. The IFN-α family consists of multiple species and some of these in different allelic forms in both humans and mice. In humans 13 IFN-α genes and one pseudogene and in mice 14 IFN-α genes and 3 pseudogenes have been identified, clustering on chromosome 4 in mouse and chromosome 9 in man [185]. The functional importance of such a diversity is largely unknown. The subtypes differ in potency and have previously been shown to vary in their profile of activities [186,187], but new studies show correlation between antiproliferative and antiviral effects of various IFN-α species [185]. Thus, it seems that the importance of the diversity could come from varying expression patterns of the different IFN-α species. Most of the α IFNs are N-glycosylated, but glycosylation does not correlate with activity of the molecule, but rather with in vivo stability, and recombinant IFNs are shown to have activity comparable with that of the naturally produced molecules [185,188]. Only one IFN-β species exists, coded by a gene situated in the IFN type I cluster on chromosome 4 in mouse and chromosome 9 in man, as described above [185].
The natural IFN-α and -β have a molecular weight of 19 – 26 kDa and most species retain stability at pH 2 [189]. All type I IFNs bind to one common receptor composed of two subunits, IFN-α-receptor(R)1 and IFN-αR2. The IFN-α/β receptor (IFNAR) signal through the JAK/STAT-pathway by phosphorylation of the Janus kinase (JAK)1, tyrosine kinase (Tyk)2, signal transducer and activator of transcription (STAT)1 and STAT2, and induces genes with an IFN-stimulated response element (ISRE) in their promoter [171,190].
Generally the type I IFNs exhibit a huge range of biological effects, such as antiviral and antiproliferative effects, stimulation of immune cells such as T cells, natural killer (NK) cells, monocytes, macrophages, and dendritic cells, increased expression of MHC-I, activation of pro-apoptotic genes and inhibition of anti-apoptotic mechanisms, modulation of cellular differentiation, and inhibition of angiogenesis [171]. The newly discovered IFN-ξ/limitin also interacts with the IFN-α/β receptor, and is regarded as a type I IFN [184,191]. Antiviral activity of IFN-ξ has been shown against many viruses including HSV, and it exhibits both immunomodulatory and anti-tumour effects, but the lymphosuppressive activity is less than that of IFN-α [184,192]. A human homolog of IFN-ξ could thus have interesting potential in the therapy of tumours and viral infections.
The type II IFN is represented by only one member, the IFN-γ [193]. Structurally, IFN-γ is distinct from the type I IFNs, and it signals through a different receptor. For many years IFN-γ was thought only to be expressed by T cells. Later the large granular lymphocytes (NK cells) were recognised as important producers by the fact that Ia-antigen (MHC-II) expression on mouse macrophages could be induced by Listeria monocytogenes infection in SCID mice lacking T cells [194-196]. In recent years it has, however, been clear that other cell types, originally thought not to be producers of IFN-γ, are in fact capable of IFN-γ expression. So now macrophages, B cells, NKT cells and professional antigen-presenting cells are also recognized as IFN-γ producers in certain situations [197-202]. Induction and production of IFN-γ in antigen-presenting cells and NK cells seem to be vital in the early non-specific response to infections and of importance in the linkage to the adaptive specific responses coming up later [202-204]. The induction of IFN-γ production in non-T cells (e.g. NK cells) is conducted by cytokines, especially IL-12 in synergy with other proinflammatory cytokines, largely produced by mononuclear phagocytes [205,206].
IFN-γ exerts its effects through a distinct class II cytokine receptor, the IFN-γ receptor (IFNGR), composed of two subunits, IFN-γR1 and IFN-γR2. Upon binding of a homodimer of IFN-γ to the receptor complex, JAK2 autophosphorylates and then transphosphorylates JAK1. Activated JAK1 in turn phosphorylates IFN-γR1, which allows binding of the STAT1 homodimer to the receptor and subsequent phosphorylation of STAT1 [204]. The IFNGR and STAT1 are preformed as hetero- and homo-dimers, and upon receptor binding, the IFN-γ-IFN-γR1-STAT1 complex seems to be internalized and translocated to the nucleus, where the activated STAT1 homodimer binds to DNA at GAS elements and induces the first wave of responses [204,207-211]. Many of these initial IFN-γ induced products are transcription factors participating in further regulation of the many IFN-induced cellular response. Among these products are the IFN regulatory factors (IRFs) which stimulate or inhibit transcription of genes possessing an ISRE in the promoter region [204,212].
For many years the key mediator of macrophage activation during antigen-induced processes was recognised as macrophage activating factor (MAF) [213]. Only later, the crucial importance of these effects was attributed to IFN-γ [214,215]. IFN-γ has antiviral activity, but the most important effects of IFN-γ seem to be activation of macrophages, antigen-presenting cells, and NK cells and inhibition of T-helper type 2 (Th2) cells, resulting in a Th1-driven cell-mediated response to infection [204]. Experiments in knock out (KO) mice with deficient IFN-γ, IFNGR, or STAT1 expression have shown that this system is of major importance, but not vital, in the host response to viral infections [216-219].
Besides the two traditional groups of IFNs, a new group of IFN-like cytokines has been described in various species and named IL-28A (IFN-λ2), IL-28B (IFN-λ3), and IL-29 (IFN-λ1) [171,220]. These cytokines are antiviral proteins interacting with a distinct heterodimeric class II cytokine receptor composed of IFN-λR1 and IL-10R2, but sharing with the type I IFNs some intracellular signalling pathways through the ISRE [221]. Thus, they have a largely similar antiviral effect as the type I IFNs [220].
Tumour necrosis factor
Tumour necrosis factor (TNF, former designated TNF-α) and lymphotoxin (LT; former TNF-β) were for many years also known as cachectin from their involvement in cachexia of cancer patients [222]. TNF is a prototype and the second member of the TNF ligand superfamily (TNFSF2), now encompassing over 40 known signalling molecules, among which the LTα, LTβ, and LIGHT (LT-like, exhibits inducible expression, and competes with HSV glycoprotein D for HVEM, a receptor expressed by T lymphocytes) are some of the more prominent ligands [58,223]. Each member is the ligand of one or two distinct receptors of the TNF receptor family sharing a high degree of homology. The current nomenclature of these ligands and receptors has now been gathered on the internet [224]. TNF is a type II transmembrane glycoprotein coded from the human chromosome 6 and from chromosome 17 in mice [223]. It is synthesized as a 26 kDa transmembrane pro-TNF, primarily located in the membranes of the Golgi apparatus [225]. The pro-TNF is cleaved by a metalloprotease releasing the 17 kDa extracellular portion of the molecule [222,226]. Production and release of TNF from the cell is regulated at both the transcriptional and translational level and by post translational modification as described above [227]. During HSV infection both pre- and post-transcriptional regulatory mechanisms are involved in TNF production [228]. TNF is produced by many cell types of immune origin, primarily mononuclear phagocytes, neutrophils, NK cells and T cells, and has diverse effects on different cells [222].
Both membrane bound and soluble TNF interact as homotrimers with two different receptors, the p55 TNFR1 (TNFRSF1A) and the p75 TNFR2 (TNFRSF1B) [222]. As most other receptors of this family, TNFR1 holds a death domain important in the pro-apoptotic pathway. TNFR1 is expressed virtually on every cell type except erythrocytes, whereas TNFR2 is mostly expressed on endothelial and bone marrow derived cells [227]. The TNFR2 activates NF-κB (p50, p65/RelA, and p52/RelB) by ubiquitin-mediated degradation of inhibitor-κB (IκB) after phosphorylation by an IκB kinase (IKK). Besides inducing apoptosis, TNFR1 also activates NF-κB (p50/p65) [229,230]. Furthermore, the activator protein 1 (AP-1) is activated by mitogen-activated protein kinases (MAPKs) and together with NF-κB primarily acts in the proinflammatory pathways. Thus, signalling from the TNF receptor family induces a delicate balance between life and death (apoptosis) of the cell. Both of the TNF receptors can by proteolytic cleavage be converted to soluble receptors with the capacity to compete with their signalling ancestors, but also act to stabilize the trimeric TNF and thus maintain its activity [227,231].
The TNF superfamily seems to have evolved with the adaptive immune system in vertebrates and is crucial for the embryonic development of lymphoid tissue [223]. Furthermore, TNF is, as a proinflammatory cytokine, involved in activation of many immune cells and is thus an important factor of both the early non-specific and the specific immune response [232]. The importance of the TNF superfamily in antiviral defence is illustrated by the fact that different viruses have developed mechanisms for interference with nearly every step of activity of this system [227,229].
Interleukin-12, IL-23 and IL-27
IL-12 is the prime member of a small group of heterodimeric cytokines, all with the capacity to induce production of IFN-γ in a variety of cells. IL-12 was first described as an NK cell stimulatory factor (NKSF) and identified as a heterodimeric molecule composed of a p40 and a p35 subunit, which are covalently linked [233]. The p35 subunit has homologies to IL-6, and p40 is homologous to the extracellular domain of the haematopoietin receptor family, particularly the IL-6Rα chain [234]. The two IL-12 subunits are coded from different chromosomes, i.e. the human chromosomes 3 and 5 and the mouse chromosomes 6 and 11, respectively [235]. These genes are regulated separately, and coordinated induction in the same cell is required for secretion of the biologically active IL-12p70 heterodimer [236]. IL-12 is produced by monocytes, macrophages, dendritic cells, neutrophils and B cells [235,237]. In the initial response of spleen cells in mice injected in vivo with extracts of toxoplasma gondii or with lipopolysaccharide (LPS), the cellular source was found to be dendritic cells, but cultured macrophages have by themselves also been shown to produce IL-12p40 upon HSV-2 infection [238,239]. Such differences could depend on variations in the signalling mechanisms involved, which is also illustrated by the observation that the production in dendritic cells and macrophages has different kinetics. This difference could be brought about by differences in the requirement for co-stimulation with IFN-γ [240]. A collaborative action of dendritic cells and macrophages could be important, as indicated for IL-12 induction by influenza virus and other inducers [241].
The receptor for IL-12 is found on NK cells, T cells and dendritic cells and consists of two subunits (β1 and β2), which signal by the β2 subunit through the JAK/STAT pathway, primarily by activated STAT4 [235]. The primary effect of IL-12 is induction of IFN-γ production in NK cells and T cells, and IL-12 activates the cytotoxic potential of these cells. The IFN-γ locus in NK cells is constitutively demethylated and is thus ready for transcription of the gene, which is in contrast to that of T cells, [242]. Macrophages and NK cells are then stimulated by IFN-γ, resulting in activation for enhanced antimicrobial capacity [243,244]. IL-12 and IFN-γ in conjunction are the main responsible factors for activation of a Th1-driven adaptive cellular immune response, important for the long-term control of intracellular pathogens [235]. IL-12 stimulates proliferation of naïve T cells, and in conjunction with IFN-γ inhibits Th2 cell differentiation and the production of Th2 cytokines (e.g. IL-4, IL-5, and IL-13) [235]. Thus IL-12 holds a key position in induction and control of the Th1 response. The IL-12-induced IFN-γ production is synergistically enhanced by other cytokines such as TNF and IL-1 [240], and IFN-γ production can even be induced in macrophages by co-stimulation with IL-18 [197,245,246], a cytokine which by itself does not possess major IFN-γ-inducing capacity [240]. A positive feed-back loop is initiated by the IL-12-induced production of IFN-γ, in that IFN-γ is an important primer of IL-12 production, thus accelerating the system [247]. Furthermore, T cells enhance IL-12 production through signals of the proinflammatory TNF family [240]. In virus-infected macrophages a similar autocrine feed-back loop involving IL-12, IL-18, IFN-α/β, and IFN-γ could be speculated [248].
This potentially harmful situation, with accelerating IFN-γ production, regulated in a positive feed-back loop by IL-12, is inhibited by cytokines possessing anti-inflammatory properties. Among these IL-10 holds a crucial position as an inhibitor of IL-12 production, an effect which is also conducted by transforming growth factor-β (TGF-β) [249-251].The Th2 cytokines of the other side of the adaptive response, IL-4 and IL-13, inhibit IL-12 induction in the early phases of stimulation, but later they can be potent inducers of IL-12 production, although they still inhibit many of the IFN-γ-induced activities [212,252,253]. Phagocytosis of apoptotic cells by macrophages inhibits production of IL-12, a regulatory mechanism which seems to be important in restriction of the damages induced by uncontrolled defence mechanisms [254]. Injection of high doses of IL-12 to virus-infected mice is toxic, and leads to death with the pathology of TNF-related toxic shock, an effect which was explained by increased sensitivity to the toxic effects of TNF, and found to be dependent on the genetic constitution of the host [255,256].
The small IL-12 cytokine family also includes two other heterodimeric cytokines, IL-23 and IL-27, and a homodimer of IL-12p40. The latter is found in vivo in mice and functions as an antagonist of IL-12, but it is debated whether it exists in humans [257,258]. IL-23 is composed of the IL-12p40 and a p19 subunit and likewise binds to a receptor with one of the IL-12 receptor subunits (IL-12Rβ1) and a distinct IL-23R subunit [240,259]. The production and function of IL-23 is quite similar to that of IL-12, but IL-23 has a unique capacity to induce proliferation of memory T cells [235], and it has been found in nervous ganglia of HSV-infected mice on day 3 of infection [260]. IL-23 drives IL-17 production of NK cells, which mobilizes neutrophils and promotes production of the proinflammatory cytokines IL-1, IL-6, and TNF [261]. IL-27 is the newest recognized member of the family, constructed of two distinct subunits (EBI3 and p28), but still with functional capacities alike those of IL-12 [262]. The functional implications of these later discovered members of the IL-12 family is not yet clear, but it seems as if they are contributors to the overall effects of the IL-12 family and fine-tune the system [235,263-266]. The induction of IFN-γ and activation of NK cells is not only mastered by members of the IL-12 cytokine family. Other cytokines, like IL-15, are also implicated in development, function, and activation of these cells [267,268]. Generally, the IL-12 cytokine family has shown itself of importance in early defence against several viral infections, and as a vital inducer and regulator of the adaptive immune response against viruses and other intracellular pathogens [219,256,261,269].
Interleukin-4 and IL-13
Upon an accelerating pro-inflammatory response induced by initial viral replication the organism has to embank the IFN-γ-activated potentially harmful actions of macrophages and NK cells. Important mediators of this embankment are IL-4 and IL-13, which as described above repress the induction of IL-12, and thus put a brake on the positive feed-back loop of IFN-γ production [249,252]. Furthermore, IL-4 suppresses the production of other pro-inflammatory cytokines such as TNF and IL-1 [270]. Most importantly, IL-4 and IL-13 are potent inhibitors of the efferent arm of the pro-inflammatory system, and thus inhibit production of reactive oxygen species and nitric oxide. The production of these two potentially harmful effector mechanisms of activated macrophages is hampered by inhibition of production of the responsible enzymes in these reactions, the NADPH oxidase and the inducible nitric oxide synthase (iNOS) [271-273].
The primary producer cells of IL-4 and IL-13 are the Th2 cells, but these cytokines are also produced by basophils and mast cells [274-276]. The receptors for IL-4 and IL-13 are expressed on most cells and are composed as dimers of four different chains. IL-4 is the ligand of two receptors: A high-affinity heterodimer of IL-4Rα and the IL-2R common γ-chain and another heterodimeric receptor composed of IL-4Rα and IL-13Rα1. IL-13 binds to three complexes: A high-affinity heterodimer of IL-13Rα and IL-4Rα and two homodimers composed of either IL-13Rα1 or IL-13Rα2, which are both coded from genes on the human X-chromosome [276]. The immunomodulatory signalling is conducted through the JAK/STAT-pathway utilizing JAK1, JAK3 and STAT6. Phosphorylated and homodimerized STAT6 binds to STAT binding elements (SBE), which includes GAS, and either trans-activates or inhibits transcription of the adjacent genes [212]. The functions of IL-4 and IL-13 are nearly overlapping with only discreet discrepancies [276,277].
IL-4 was discovered in 1982 on the basis of another important effect of the cytokine, namely the ability to induce proliferation of B cells, and it was from this effect in the early years called B cell growth factor [278]. As this, some other effects of IL-4 are stimulating, in that it furthermore activates other Th2-like effects such as B cell class-switching and expression of mannose receptor and Fc receptor for IgE on macrophages [276]. Despite the anti-inflammatory profile IL-4 has in vivo been shown to confer some resistance to HSV infection [279,280]. IL-4 is thus not only an inhibiting cytokine but essentially an immunomodulatory cytokine with regulatory effects on macrophages as well.
The armament and early strikes
The early innate defence mechanisms have for many years been regarded as important for the course of many viral infections, including infections with HSV [281]. The control of viral replication and dissemination during the first days of an HSV infection seems to be vital for the final outcome. If the viral replication is not halted by natural defence mechanisms during induction and maturation of the antigen-specific immune response, the adaptive immune system can be overwhelmed by massive viral infection at the dawn of activity of the specific reactions. The mechanisms of the anti-herpetic natural defence have been analysed extensively. It became relatively early clear that antiviral activity of macrophages [281] and NK cells [282] and early activity of the IFN-system [283] were important mediators of innate resistance to HSV. The relative contribution of each of these players in the early defence has been much debated, and as more interactions and molecular mechanisms are now elucidated, it seems clear that all of these players each hold a crucial position in an integrated antiviral natural defence system.
Early induction of IFN-α/β by HSV
An important model used in the study of resistance mechanisms in defence against generalized infection with HSV is a mouse model, where mice infected intra-peritoneally or intra-venously experience a generalized infection with HSV replication in most organs, including the liver, spleen, and eventually the brain [284]. The dissemination of infection to the brain and the severity of infection of the peripheral organs depend in part on the age of the mice, as is the case in humans, where neonates have difficulties in controlling a HSV infection [281,285-287]. The course of infection in mice also depends on the type of HSV in question. Furthermore, in 1975 Lopez described a differential susceptibility of inbred mice to generalized infection with HSV, and this genetic difference in sensitivity has since been used for analysis of resistance factors of importance for the anti-herpetic defence [288]. In generalized infections, the genetics of the relative resistance to HSV-2 was shown to segregate with the X-chromosome [289]. This pattern of resistance to the generalized infection was for both HSV-1 and -2 attributed to a genetically determined difference in the capacity for IFN-α/β production [179,290,291], and it was shown that the X-linked pattern of resistance segregated with the HSV-2-induced production of IFN-α/β in macrophages during the first hours of infection [168]. Furthermore, macrophages from female mice respond to HSV with higher IFN-α/β production than macrophages from male mice [168]. This observation is in line with female mice being more resistant to HSV infection in vivo [291].
Early production of IFN-α/β has been correlated to resistance of HSV infections in several other studies. Treatment of mice with antibodies to IFN-α/β increases and accelerates mortality of a generalized HSV-1 infection and with higher doses of virus, mice are dying already after three to four days, a period where antigen-specific mechanisms are still in the induction and proliferation phase [292]. Furthermore, mice treated with mercuric chloride showed higher titres of HSV-2 in the first days of infection, an effect which could be correlated to impaired production of IFN-α/β [293,294]. In studies on peripheral HSV infections, such as cutaneous or corneal infections, IFN-α/β has been shown to be produced locally and to restrict the local replication of HSV and infection of nervous ganglia cells of the area, an effect which has also been correlated to the genetic constitution of the host [295-298].
The genetic background for the X-linked trait of HSV resistance and IFN-α/β production of macrophages remains unravelled. Induction of IFN-α/β upon HSV infection seems to be governed by different mechanisms in different cells [299]. IFN-α/β can be induced early by both infectious and UV-inactivated HSV in various cells, with the infectious virus being the more potent inducer in mouse peritoneal macrophages, whereas the UV-inactivated virus showed most potency in human peripheral blood mononuclear cells (PBMC) [168,299-302]. Production of IFN-α/β was induced by gD of HSV-1 in PBMC, but not in murine macrophages [17,303,304]. In PBMC-derived dendritic cells, however, the cellular mannose receptor was shown to be involved [299,305]. Furthermore, different Toll-like receptors (TLRs) have been shown to react with HSV [306]. TLRs are transmembrane pattern recognition receptors (PRRs) that detect redundant microbial molecular motives and induce antiviral and proinflammatory cytokines in response to alerting signals. In dendritic cells, TLR9-signalling, induced by the GC-rich HSV genome, has been shown to govern the induction of IFN-α/β, but TLR9-KO mice are still capable of controlling HSV infections in vivo [307-309]. However, in mouse macrophages the TLRs do not seem to be crucial for IFN-α/β induction upon HSV infection [304]. This is in agreement with the observation that the majority of IFN-α/β produced by spleen cells and dendritic cells and the total production from bone marrow macrophages was independent of TLR9 or MyD88, which is necessary for signalling by most TLRs [308]. In this study, heat inactivated virus was shown still to induce IFN-α/β in cells utilizing TLR9. As resident peritoneal macrophages do not produce IFN-α/β in response to even high doses of heat inactivated HSV, this gives an additional indication of independency from TLR9 of IFN-α/β production in macrophages [300]. Moreover, efficient induction of IFN-α/β by HSV in macrophages required dsRNA-activated protein kinase (PKR) activity and infectivity of the virus [304]. This is in agreement with the observation that dsRNA, which is produced by most viruses during replication, induces IFN through PKR, and not through TLR3, which also binds dsRNA [310]. Furthermore, another mechanism of IFN induction by dsRNA through a RNA helicase has been proposed [311].
The different induction patterns in different cells types, and the fact that IFN-α/β seems largely to be induced by other mechanisms than TLRs, explain the fact that knocking out TLR-signalling by MyD88 did not influence the in vivo infection with HSV in mice [309]. Other TLRs have also been shown to mediate signals in HSV infections. In HSV encephalitis in TLR2-KO mice, viral replication seemed unchanged or slightly increased during the first 4 days of infection, and the production of IL-6 and monocyte chemoattractant protein 1 were impaired, but interestingly pathological changes and mortality were reduced [134].
In relation to the X-linked resistance pattern of HSV infection and IFN production upon HSV infection, it is interesting that some of the TLRs are coded from the X-chromosome [312]. These are the TLR7 and TLR8, which are triggered by guanosine- or uridine-rich ssRNA in the endosomal compartment of cells [313,314]. There are, however, no indications that this pathway is implicated in IFN induction in cells during HSV infection, but the question has still not been directly addressed.
Regulation of the IFN-α/β gene induction is in part governed by activation of the transcription factors IRF-3 and -7, which are induced by IFN-α/β itself, resulting in a positive feed back loop, an effect which has been known for years without knowledge of the signalling mechanisms [315-317]. Thus, one possible explanation for the genetic differences in HSV-induced IFN-α/β production could be an elevated physiological level of this IFN self-stimulating system [318-321]. An analysis of the levels of IRF-3 and -7 in normal macrophages from these mice could be of interest. Analysis of the levels of the IFN-induced enzyme 2'-5'-oligoadenylate synthetase (OAS) in uninfected cells showed low but slightly higher levels in cells from relatively resistant mice [322]. With LPS, cells from the relatively resistant (C57Bl/6) mice show an early induction pattern of IFN-α/β, peaking within 2 hours, whereas cells from the susceptible BALB/c mice demonstrate a delayed response, peaking 7 hours after induction [323].
Among other transcription factors involved in induction of the various IFN-α/β genes are the heterodimeric NF-κB family, which is activated by TLRs, IL-1R, and TNFR [324]. During a HSV infection NF-κB is activated and translocated to the nucleus [325]. Many regulatory mechanisms of NF-κB activation exist, one of them exerted through TNF, which is produced by macrophages very early during HSV infection (fig. 2) [293,300,325,326]. Thus, the responsible mechanisms might be exerted by other regulatory signals, influencing the magnitude of the HSV-triggered IFN-α/β induction pathway, and perhaps not by this pathway in itself [327].
Figure 2 First early wave of response. The very early response to HSV infection of macrophages (Mφ). During the first few hours of infection HSV induces production of IFN-α/β and TNF in macrophages. The implications of these cytokines for HSV replication in neighbouring cells and for macrophage activation and production of reactive oxygen species (ROS) are outlined. Stimulatory pathways are indicated by green arrows (→), and inhibitory pathways are drawn in red.
A number of X-linked immunodeficiencies have been described, one of them being the Wiskott-Aldrich syndrome with defects in a protein expressed in haematopoietic cells, facilitating reorganization of the actin cytoskeleton, and thus influencing the mobility of immune cells and chemotaxis of macrophages. Patients with this X-linked immunodeficiency show aggravated herpetic infections, and cells from some patients seem to produce lower amounts of IFN in response to HSV [328-330]. Cells from patients with another X-linked immunodeficiency with mutations in the CD40-ligand, a member of the TNF family, showed decreased IFN-α/β production when infected with HSV-1, but these patients apparently show a normal response to viral infections [331,332]. This supports the notion above that other regulatory signals might be involved.
Effect of early IFN-α/β on HSV replication
The overall effects of the IFN-α/β system, besides the production as described above, are determined by the sensitivity of cells to the secreted IFN-α/β. The effector mechanisms of IFN-α/β on HSV replication are not fully elucidated. Several IFN-α/β-activated systems are involved, including the dsRNA-activated PKR, which phosphorylates, and thereby inhibits, the elongation initiation factor (eIF)-2α, resulting in inhibition of translation [333]. Another important mediator of the antiviral activity is the OAS system, which activates 2'-5'oligoadenylate-dependent RNase L with the capacity to degrade single-stranded RNA [333]. Lately, the PML bodies have been described as crucial for the anti-HSV effect of IFN-α/β [334].
In mice exhibiting a relatively HSV-resistant phenotype, the direct antiviral effect of IFN-α/β in embryonic cells was found to be approximately three-fold higher than in cells from susceptible mice [322]. Data from another study showed comparable results on IFN-α/β sensitivity concerning the replication of encephalomyocarditis virus (EMCV) in cells from the same mouse strains [335]. This phenomenon was inherited as a co-dominant autosomal trait without any apparent influence of X-linked genes [322,336]. Further studies in mouse fibroblasts have revealed that TNF intensify the antiviral effect of IFN-α/β and, thus, the in vivo situation seems more complicated (fig. 2) [300,337]. In the original publication on genetics of HSV susceptibility in inbred mice, Lopez reported fibroblasts from the different mice to replicate HSV equally, and the same was found in the cells showing differential sensitivity to IFN-α/β [288,322]. In line with these results, the IFN-activated OAS, an inhibitor of HSV replication, was induced to a higher degree in cells from the resistant mice upon IFN-α/β treatment [322,333,338]. Furthermore, the level of stimulated and unstimulated OAS was generally found to vary between different inbred mouse strains [339]. Thus, the genetic difference in antiviral action of type I IFNs seems to affect the replication of several different viruses and to correlate with resistance to HSV.
The viral host protein synthesis shutoff, exerted by the HSV vhs-protein of the tegument, has major effects on the cytokine production of infected cells and reduces the effect of IFN-α/β on HSV replication [340]. Furthermore, the tegument proteins have been shown to induce cellular inhibitors of the JAK/STAT pathway, resulting in inhibition of both IFN signalling and production [341,342]. The IE protein ICP0 inhibits activation of IRF-3 and thereby also restricts IFN-induced pathways [71-73], and ICP0, ICP4 and ICP27 induce late shutoff of protein synthesis with decreased mRNA stability and thus reduced cytokine production [81,343]. As outlined, it thus seems HSV has evolved several mechanisms to evade the consequences of the IFN-α/β system, which underline the importance of these cytokines in the antiviral defence.
Early effects of HSV on macrophage activation
During HSV infection macrophages are activated and possess an increased antiviral potential [281,344]. Classically, the macrophage antiviral activity has been described as intrinsic or extrinsic [345]. Resting macrophages possess a high degree of intrinsic activity against HSV, generally being non-permissive to viral replication. The macrophages are thus a blind end for the HSV infection, and they can in that way protect other cells from infection, for example as a barrier lining the liver sinusoids [344]. The extrinsic antiviral activity refers to the ability of macrophages to inactivate virus outside the macrophage itself or to inhibit viral replication in other cells [346]. The intrinsic antiviral activity depends among other factors on macrophage differentiation and has been correlated to IFN activity, either physiological levels of "spontaneous" pre-infection-synthesized or rapidly acting autocrine IFN-α/β [344]. In that respect, macrophages from mice of the resistant phenotype showed higher intrinsic activity by being less permissive to HSV replication [281,347].
One potential antiviral mechanism of macrophages may be the production of ROS. These were originally assigned to bacterial killing, but the effect of ROS has also been correlated to antiviral functions, although they might not be of major importance [348]. The ROS are mainly produced by NADPH-oxidases (Nox), which are membrane-bound multi-component enzymes primarily situated in the phagolysosome [349]. Activation of the NAHPD-oxidase, by phosphorylation and fusion of the enzyme subunits, primarily results in production of superoxide anion (O2-), which by superoxide dismutase can be converted to hydrogen peroxide (H2O2). The H2O2 in turn is then by Fe2+ (Fenton reaction) or by Fe3+ and O2- (Haber-Weiss reaction) converted to hydroxyl radical (·OH), hydroxyl anion (OH-) and singlet oxygen (1O2), or by the myeloperoxidase to hypochlorous acid (HOCl) [349,350]. Small amounts of ROS are also produced by the mitochondria and may be of importance as signalling molecules from TNF [351,352].
During HSV infection in vivo, macrophages are activated and achieve an increased capacity to react with a respiratory burst of ROS when appropriately triggered, i.e. by phorbol esters (fig. 2) [353]. This macrophage activation is induced early in response to HSV infection, reaching a plateau within the first 12 hours of i.p. infection [353]. In vitro, macrophages were shown to be the cell type responding with an oxidative burst, and this capacity peaked after only 8 hours of infection with HSV [353]. This HSV-induced capacity for an increased respiratory burst was shown to be governed by autocrine IFN-α/β as a sine qua non phenomenon [300,353]. Nevertheless, TNF was also found to influence the macrophage activation. By itself, TNF reduced the macrophage capacity for a respiratory burst, but in combination with IFN-α/β it synergistically enhanced the IFN-induced activation [293,300]. Interestingly, a secreted portion of the HSV-gG acts as a phagocyte chemoattractant and induces production of ROS by signalling through the receptor activated by the phorbol esters [354].
The HSV-induced activation of macrophages in vivo is influenced by the genetic constitution of the host, with the most pronounced activation of macrophages originating from resistant mice, as expected on the basis of the genetics of IFN-α/β production in response to the infection. Furthermore, the genetics of the efferent part of the IFN-α/β-mediated HSV-induced activation of macrophages, displayed a co-dominant autosomal trait, as was the case with the antiviral effect of IFN-α/β in fibroblasts [336]. Thus, the genetically-determined sensitivity to IFN-α/β seems to be expressed in different cell-types. The influence of TNF on the genetics of this phenomenon has not been addressed. In Contrast to these observations, the genetics concerning the antiproliferative effect of IFN-α/β in bone marrow cells seems to be reversed [335,355]. This might, however, be linked, in that ROS are shown to activate various signalling molecules, mediate apoptosis, and exhibit antiproliferative effects depending on the dose and time of exposure [352].
Little is known on the potential antiviral effect of ROS. By examining peroxidized lipids, which is an oxidative product from ROS in tissues, it has been documented that these are produced during the acute HSV infection in vivo, and speculations on antiviral mechanisms have focused on induction of apoptosis [348,356,357]. HSV triggers apoptosis of infected cells by several pathways, and the importance of this phenomenon is indicated by the fact that the virus has evolved mechanisms to counteract each of these pathways [358]. Macrophages generally suppress apoptosis in HSV infections, as seen by increased apoptosis in macrophage-depleted mice [359].
Several studies on the mechanisms involved in the early battle against HSV, performed in in vivo animal models, have pointed to IFN-α/β as a crucial player. In adoptive transfer experiments, the effect of adult mouse spleen cells on the initial phase of a generalized HSV infection in suckling mice was conducted by IFN-α/β [360]. Furthermore, administration of a hematopoetic growth factor to neonatal mice increases the number of dendritic cells, B cells and NK cells, and confers resistance in a cutaneous model of HSV infection. The effect in this model could also largely be attributed to the actions of IFN-α/β, with some additional contributions by IFN-γ [361,362]. In KO-mice IFN-α/β was able to control the initial phase of a generalized HSV infection without contributions from NK, T- or B cells, but these latter players were necessary for survival and long term control of the infection [363].
The importance of an early, local IFN-response in models including in vivo progression and evaluation of final outcome of infection is more unclear, in that many other viral and host factors are of importance in these more complicated models with several stages of infection and involvement of different organs. Such models are, however, more close to the normal human HSV infection, starting at an epithelial surface, but to expect that one resistance factor in such a complicated system will come out clear as the responsible factor for the outcome downstream the sequence of events, is too simplistic. Nevertheless, induced expression of IFN-α/β in the eye by plasmid DNA or an adenovirus vector was shown to inhibit early local replication of HSV and the concomitant spread of virus to the brain and death from encephalitis [333,364], and in IFN-α/βR KO-mice HSV replicated to much higher titres than in normal mice [297].
This tells us that the innate and adaptive immune systems exhibit much redundancy, and that IFN-α/β is of vital importance in local inhibition of HSV replication. The multitude of antiviral mechanisms, be it innate or adaptive, have varying effects and importance in the different phases of infection, such as initial local infection, dissemination to other organs, establishment of latency and reactivation, and conclusions can not be drawn from one situation to another.
The opening battle
The reactions discussed above, involving production of IFN-α/β and TNF, take place within the first 6 to 12 hours of a HSV infection, and thus are reactions, which can execute an effect within the first replication cycle of the virus. A little later, other cytokines such as IL-12, IL-18 and IFN-γ are produced and give rise to other weapons in the battle against the virus. They will, in turn, within the next replication cycle execute their actions, with potential harmful consequences for either parts of the conflict.
IL-12 and IFN-γ production in early HSV infection
A few hours after the type I IFN and TNF response, macrophages react upon HSV infection with production of IL-12, which is seen from 8 to 12 hours after infection and on [238,365]. The same was found with other viruses 12 to 24 hours after infection [366]. In these and other studies, the producers of IL-12p40 during viral infection seem to be inflammatory cells, including macrophages, and not the infected stromal cells [365,367]. The IL-12 induction during HSV infection requires infectious virus, and it was shown to be regulated at the transcriptional level [238], as it is also the case when it is induced by LPS [247,367]. The dependence on infectivity is, however, in conflict with results from in vivo production of IL-12p40 and IFN-γ in draining lymph nodes from sites injected with UV-inactivated HSV [302]. High doses of UV-inactivated virus were used, and some minimal transcription of viral genes could have taken place, although the virus was not replication competent. Transcription of the IL-12p40 gene in macrophages requires de novo protein synthesis during the inducing HSV infection, which could explain the relatively late appearance of IL-12 production [238,367]. The κB-sequence of the IL-12p40 promoter binds NF-κB in HSV-infected cells, and the production of IL-12p40 was found to be repressed by an inhibitor of NF-κB activation [238]. Both these observations indicate that signalling through NF-κB is of significance in HSV-induced IL-12 production.
In human macrophages, TNF has been shown to inhibit IL-12p40 production, but not p35 production, by a mechanism not involving NF-κB [368]. Furthermore, IL-12 has in a mouse model been shown to stimulate TNF expression [255], indicating that TNF can participate in a negative feed-back loop in the regulation of the IL-12 system [369]. Likewise, IFN-α/β has been shown to inhibit IL-12 production in both humans and mice [370-372]. The implication of such inhibition by IFN-α/β and TNF, which are secreted very early in HSV infections, well before the production of IL-12, has so far not been elucidated.
As described earlier, the IL-12p40 induction is influenced by IFN-γ in a positive feed back loop. IFN-γ could activate IL-12 transcription through binding of IRF-1, -2, and -8 to an ISRE site in the promoter-region of IL-12 [373,374]. Upon HSV infection, IFN-γ is produced as part of the non-specific response to the virus. A marked synergism between HSV and IFN-γ in IL-12 induction has been demonstrated [238], indicating that the IL-12 / IFN-γ auto-accelerating system is of importance during HSV infections.
The IFN-γ-inducing activity of the produced IL-12 is pronounced in mouse peritoneal cells after 24 hours of infection with HSV [238]. In a study by Kirchner et al. IFN-γ was detected as early as on day 3 of in vivo HSV infection, and the IFN-γ production was correlated to the genetics of HSV resistance [375]. During HSV infection, the production of IFN-γ is mainly induced as a concerted action of several factors and not by IL-12 alone. IFN-α/β by itself was shown to be a weak inducer of IFN-γ production by NK cells, but in synergy with IL-12 the production of IFN-γ was markedly enhanced [376]. In elicited peritoneal macrophages, HSV induced efficient IFN-γ production through cooperation of IL-12, IFN-α/β and IL-18 [377]. In such a proinflammatory environment even other cells than NK and T cells, e.g. macrophages, might produce lower levels of IFN-γ [200,202,245]. IL-12 signals through STAT4, but STAT4 translocation to the nucleus of NK cells has also been seen after IFN-α/β stimulation [206,378]. Likewise, IFN-α/β induces STAT4 phosphorylation in T cells [379], indicating that IL-12 and IFN-α/β at this point act through a shared signalling pathway. Furthermore, the synergistic action of IL-12 and IL-18 in IFN-γ production by macrophages was shown to be dependent on STAT4 [197]. In addition to these factors, TNF and IL-1 have also been shown to act in synergy with IL-12 in IFN-γ induction [206,380,381] and vice versa, IFN-γ has been shown to synergize with HSV in induction of TNF production [325]. This further emphasizes the concept of positive feed-back mechanisms in the regulation of early IFN-γ production.
The important direct effect of IFN-α/β on HSV replication was found to be enhanced synergistically by IFN-γ in both cell culture and in vivo in mice [363,382-384]. This is, however, in conflict with an early study, which could not reveal any synergism between IFN-α/β and IFN-γ on the replication of HSV in human blood mononuclear cells [385]. Synergistic action of the two types of IFN is further supported by the observation of synergism between IFN-γ and TNF on HSV replication in corneal cells, and the fact that this was exerted through production of IFN-β [386-388]. The effect was, however, greatly dependent on the cell type examined, which could explain the above-mentioned inconsistency. Synergism between TNF and IFN-γ in inhibition of HSV replication has now been shown to be mediated by activation of a tryptophan-depleting enzyme [389]. Thus, relatively small amounts of early IFN-γ produced by NK cells in response to IL-12, IFN-α/β, TNF, and IL-18 could in collaboration with the already present IFN-α/β and TNF have important local effect on HSV replication in permissive cells (fig. 3). This conclusion is further supported by observations in KO mice, indicating that collaborated action of IFN-α/β and IFN-γ is of importance in control of subcutaneous HSV infections [362].
Figure 3 Second early wave of response. Regulatory pathways controlling production and action of IFN-γ during early HSV infection. When infected with HSV macrophages (Mφ) produce several cytokines, including IL-12, which stimulate production of IFN-γ, primarily in NK cells. IFN-γ then induces NO production in macrophages and stimulate the direct antiviral activity of IFN-α/β in other cells. Stimulatory pathways are indicated by green arrows (→), and inhibitory pathways are drawn in red.
In vivo studies on HSV infections in immunodeficient, KO, and antibody-treated mice have shown that the IL-12, -23 / IFN-γ system is able to control the infection, affecting both the survival rate and the HSV titres early in infection [390,391]. The effect of IL-12 in HSV infections seems to be conducted in synergy with IL-18 [390], as it has also been shown for vaccinia virus [392]. In HSV corneal infections in KO mice, IL-12 was shown to participate in the immune pathogenesis [393], but in another study utilizing IL-12 encoding plasmid DNA, corneal expression of IL-12 reduced the angiogenesis, and thus the pathology of the infection [394]. However, both studies agreed that IL-12 does not affect the local titres of HSV in the eye. After a thermal injury, wide-spread HSV infections are an important risk, and treatment of injured mice with IL-12 combined with soluble IL-4R results in augmentation of the IFN-γ production and decreased viral replication and mortality [395].
In mice infected with murine cytomegalovirus (MCMV) production of IL-12-induced IFN-γ by NK cells has been demonstrated in vivo, and the system was further shown to lower the viral titres [396,397]. The IL-12 / IFN-γ system seems, however, not to be of importance in all viral infections, in that the latter study could not detect any production of early IL-12 or IFN-γ in a model of infection with the arenavirus lymphocytic choriomeningitis virus. Analyses of the IL-12, -23 / IFN-γ system in humans with genetic defects and in KO-mice reveal more redundancy in man than in mouse and indicate that the system is of more importance in DNA- than in RNA-virus infections [219].
The producers of early IFN-γ, the NK and NKT cells, and the cytokine IL-15 and the transcription factor T-bet, which are both crucial for the differentiation and function of these cells, have all been shown to be decisive for the early control of HSV infection in vivo [268,398-400]. Although NK cells but not IFN-γ was shown to be decisive for survival from ocular infections [401], such an effect of IFN-γ has been seen by others [402]. Furthermore, a review of genetic functional NK cell defects found NK cells and their innate IFN-γ production to be of central importance in herpesvirus infections [403].
Overall, it can be concluded that the IL-12 / IFN-γ system is active in HSV infections and possesses an important antiviral potential, capable of controlling viral replication during the early phases of infection.
Production of NO in early HSV infection
In macrophages exposed to IFN-γ, the enzyme inducible nitric oxide synthase is induced, which eventually results in production of NO from molecular oxygen and a guanidino nitrogen by conversion of L-arginine to L-citrulline [404]. Upon HSV infection, the iNOS gene is induced, as shown by detection of iNOS-mRNA in infected mouse peritoneal cells and corneal neutrophils [405,406]. The production of NO in HSV-infected cultures of resting mouse peritoneal cells, which comprise a mixed population of macrophages, lymphocytes, NK cells etc., is dependent on the virus being infectious [405]. This is in line with the requirement of infectious HSV for IL-12 production and thus for production of IFN-γ as described previously [238]. NO could itself be involved in a positive feed-back, in that signalling of IL-12 utilizing Tyk2 requires the activity of NO [407]. When exogenous IFN-γ is added to virus-infected cells, a marked synergism is seen. This synergistic effect of HSV on the IFN-γ-induced NO production in macrophages was shown to be mediated by autocrine secretion of TNF [325,405]. In line with this, mice with a targeted disruption of the TNF gene showed impaired resistance to HSV and increased viral replication within the first days of infection [408], and antibodies to TNF and an inhibitor of NO production impaired early control of HSV infection in peripheral nervous tissue [409].
The induction of iNOS and the following production of NO in response to IFN-γ and HSV is a relatively slow reaction, coming up after about 18 hours of infection [405]. In in vivo vaginal HSV infections iNOS mRNA could be detected after 24 hours of infection [410]. Thus, the production of this relatively toxic substance is part of the second wave of innate defence mechanisms. The retarded production of NO and the requirement for two or more signals for induction of iNOS are logic considering the toxicity of NO and the potentially harmful consequences for the host.
In HSV-infected macrophages exposed to IFN-γ, iNOS is induced synergistically though TNF-induced NF-κB activation and translocation to the nucleus, as shown by binding of a heterodimeric complex of p55/p65 and a homodimer of p55 to the κB-site of the iNOS promoter during infection [325]. The crucial position of NF-κB in the induction of iNOS and production of NO is also indicated by experiments showing that antibodies to TNF inhibit activation of NF-κB and production of NO in HSV-infected cells and abolish the synergism between the virus and IFN-γ, an observation which was also seen with inhibitors of NF-κB activation [325]. Further analysis of the signalling mechanism has revealed that the synergism upon HSV infection is influenced by physical interaction of IRF-1 and the NF-κB subunit p65 and controlled by the ISRE-site and the distal κB-site of the iNOS promoter (fig. 5) [411]. A further support for this notion comes from the observation that the DNA-binding capacity of NF-κB and the nuclear translocation of IRF-1 have similar kinetics upon HSV infection [411] and the fact that IRF-1 is essential for iNOS induction [412]. Induction of other genes such as IFN-β and vascular cell adhesion molecule 1 also involve physical interaction of IRF-1 and NF-κB [413], and both IRF-1 and IRF-2 have in other cells types been shown to form complexes with NF-κB [414,415]. Another potential mechanism in the synergistic induction of iNOS could involve complex formation of IFN consensus sequence-binding protein (ICSBP or IRF-8) and IRF-1, which is also important for high-output NO production but has still not been studied in HSV infections [416].
Figure 4 Regulation of iNOS induction at the cellular level. Cytokines controlling the iNOS induction in macrophages (Mφ) during early HSV infection. IFN-γ, produced mainly by NK cells, stimulates iNOS production. This IFN-γ-induced production of iNOS can be inhibited by IL-4. Upon HSV infection of macrophages they produce TNF which synergizes with the IFN-γ-induced pathways and inhibits the inhibitory signals of IL-4. Thus, the virus overrules the restrictive signals and opens up for an otherwise closed pathway. Stimulatory pathways are indicated by green arrows (→), and inhibitory pathways are drawn in red.
Figure 5 Regulation of iNOS induction at the molecular level. Transcription factors controlling induction of the iNOS gene. Activated STAT1 induces transcription of the IRF-1 and iNOS genes, an effect which is competed by activated STAT6. IRF-1 interacts physically with NF-κB, binds to the distal κB-binding site of the iNOS promoter region, and stimulates transcription. Only when NF-κB is absent, IRF-2 can bind to the ISRE site and block transcription. Stimulatory pathways are indicated by green arrows (→), and inhibitory pathways are drawn in red.
Thus, high-output NO production from activated macrophages is controlled by a "double-lock" signalling mechanism restricting the production of this antiviral toxic substance to sites of active viral replication, and sparing uninfected tissue from the detrimental effects (fig. 4).
The antiviral effects of NO have been documented in several viral infections, although there clearly exist viruses and conditions where NO does not exhibit major antiviral properties. NO is thus not a magic bullet against virus infections [417]. In HSV infections, NO has been shown to confer a substantial part of the antiviral activity induced by IFN-γ in a macrophage cell line and to participate in the extrinsic anti-HSV effect of macrophages [418-421]. An exogenously added donor of NO has in several cell lines been shown to reduce the replication of HSV [422]. In vivo, analysis of mice treated with an inhibitor of NO production showed higher titres of HSV in the lungs but increased survival rates due to reduced inflammation [135]. Recently, a study using another inhibitor of NO production has confirmed the anti-herpetic effect of NO during a HSV respiratory infection, but in this study mice with inhibited NO production showed increased inflammatory responses, symptoms of infection, and mortality [423]. Replication of HSV during vaginal infection was increased in the presence of an inhibitor of NO production, and this enhanced viral replication was most prominent during the first 24 hours of infection [410]. In iNOS-KO mice, the herpes virus MCMV replicates to higher titres in various organs and in macrophages, and this results in impaired survival of the animals [424]. Weanling mice with a targeted disruption of the iNOS gene showed increased HSV replication, but apparently without differences in HSV titres during the first days of infection [425], and in adult KO mice, we could not detect any significant effect of NO during the early days of a generalised HSV infection (Ellermann-Eriksen, unpublished results). Probably, these in vivo results are due to redundancy of the antiviral system. [426].
The final effects of NO on HSV infections therefore appear to be balanced between antiviral versus toxic effects, and the final outcome seems to depend on the timing, infectious dose, and tissues involved. Thus NO production in the early phases of HSV infection is one of the effector mechanisms of the innate immune response inhibiting HSV replication, but when overproduced, NO might itself result in pathology, as discussed in the following section.
Restriction of NO production during HSV infection
As outlined above, positive feed-back mechanisms exist at the afferent side of the early cytokine response, involving especially the production of IFN-γ, IL-12, IFN-α/β and TNF, and synergisms at the efferent side, resulting in high-output NO production. As a result of coordinated induction of the iNOS gene by several transcription factors, activated by especially IFN-γ and TNF, a potent early antiviral system is activated. However, NO causes damage to DNA, proteins and lipids in cells and tissues and could thus be deleterious for the host [427-430]. A study in KO-mice indicates that NO can be responsible for inflammation and life-threatening symptoms to HSV infection of the lungs [135]. This effect of NO on pulmonary symptoms is also observed in influenza virus infections [431], although NO inhibits replication of both influenza virus and severe acute respiratory syndrome coronavirus [432,433]. Consequently, when this system is activated, it has to be controlled and eventually closed down, as it would otherwise induce unnecessary harm to the host. Such negative regulations of the iNOS gene induction in IFN-γ activated macrophages is conducted by IL-4 and IL-13 [272,273,434]. Furthermore, TGF-β can exhibit down-regulation of NO production through several post-transcriptional regulatory mechanisms, but the contribution of these pathways have not been analysed in HSV infections [435,436]. IL-4 production during HSV infection has in vaginal and CNS infections been demonstrated on day 2 of infection and to increase for the next days [437,438]. In peritoneal cells from mice infected i.p. production of IL-4 could be detected at day 5 of infection [439].
At low IFN-γ concentrations, IL-4 has been shown to inhibit iNOS induction through STAT6 competition with STAT1 binding to the GAS element of the IRF-1 promoter region. This results in reduced expression of the transcription factor IRF-1, which is crucial for induction of iNOS [440]. Generally, STAT6 was shown to be a key factor in IL-4- and IL-13-induced inhibition of iNOS gene transcription induced by IFN-γ (fig. 5) [441].
At higher IFN-γ concentrations, activated STAT6 is no longer able to compete with the high amounts of activated STAT1 dimer [440]. However, in this situation IL-4 is still able to inhibit the production of NO from IFN-γ-stimulated macrophages [272,273,434]. In the presence of high levels of IFN-γ, IL-4 is not able to alter the induction of IRF-1, but the production of IRF-2 is increased [434]. The human promoter region of IRF-2 contains a SBE, and the induction of IRF-2 could thus potentially be mediated by STAT6 binding to this element [442]. This is in agreement with the fact that IRF-2 is known to compete with the binding of IRF-1 to ISRE sites and to antagonize the trans-activating activity of IRF-1 in the regulation of other IFN-induced genes [443-445]. Inhibition of iNOS expression by high concentrations of IRF-2 relative to IRF-1 has thus been proposed as a controlling mechanism in situations with high levels of IFN-γ (fig. 5) [434]. Furthermore, another mechanism could evolve from the observation that IL-4 signalling can result in disruption of the complex formation of ICSBP and IRF-1 and thereby inhibit iNOS induction [416]. Other mediators of IL-4-induced repression of iNOS induction might exist, in that another DNA-binding transcriptional repressor competing with IRF-1 has been described [446].
In IFN-γ activated macrophages the IL-4- and IL-13-induced inhibition of iNOS induction can thus be overruled by HSV infection, leading to a sustained NO production (fig. 4) [439,447]. This effect of HSV infection is mediated through TNF production and NF-κB activation [439,447]. However, pre-treatment with IL-4 has in a Theiler's murine encephalomyelitis virus model showed inhibition of NF-κB activation [448]. In thioglycollate-induced peritoneal cells, LPS and TNF could only overcome the inhibiting effect of IL-4 in situations, where IL-4 was added simultaneously or after the stimulators [272], a sequence of events which, however, is in agreement with the sequence of cytokine production in HSV infections. When activated, the NF-κB p65 physically interacts with IRF-1 and trans-activate iNOS transcription in HSV-infected and TNF-treated cells [411,449]. It is thus tempting to speculate that the NF-κB-IRF-1 complex has higher affinity for the combined DNA-binding site and thus is able to obstruct the binding of IRF-2 to the ISRE site of the iNOS promoter and in that way turn the competition towards transcriptional activity (fig. 5) [411]. This will block the inhibiting effect of IL-4 in foci of HSV replication and open up for NO production at sites where the antiviral effect is of more importance than the potential toxicity.
Conclusions and perspectives for future clinical intervention
In treatment of HSV infections, we have for many years had a very powerful tool in the antiherpetic drug acyclovir and related compounds. But there are still therapeutic problems in the group of patients with generalized or CNS infections, and therefore it is tempting and timely to hypothesize on possible future treatment strategies. As described, it is clear that relatively discrete but early actions of the non-specific defence systems are crucial for the long term outcome of the infection. The same holds for the antiviral therapy, and early presumptive therapy and rapid diagnostics could thus potentially improve the final outcome. In the seeking for improved antiviral treatment, adjuvant therapy with anti-HSV antibodies could potentially accelerate the clearance of viral particles, and block viremic dissemination in patients, who are still seronegative at the time of treatment.
Immunomodulatory treatment modalities imitating the early non-specific antiviral defence, working as described in this review, could be considered. The key players exhibiting the least toxicity by themselves could be used, taking advantage of potential synergy with other cytokines in the foci of HSV infection. In the future, molecules with affinity for various receptors are expected to be produced, and when we know the signalling mechanisms in detail and all the potential interactions, molecular signalling could be addressed directly by pharmaceuticals.
In consequence of the crucial position of the type I IFNs in innate response to HSV, future analogues of IFN-α/β seem obvious as candidates for adjuvant treatment of severe HSV infections. This could be supplemented with IL-12, which would give the highest IFN-γ production in foci of HSV infection because of other cytokines such as IFN-α/β, TNF and IL-18 being present there. With focussed production of IFN-γ at sites of active viral replication and treatment with IFN type I analogues the focal antiviral activity could be increased markedly, without too much activity in areas without infection. To hamper systemic consequences of the enhanced proinflammatory reactions, such pro-inflammatory treatment could perhaps benefit from concomitant treatment with IL-4 or other STAT6-activating therapeutics in the future. This would further focus the activity to sites of active HSV replication. In situations with massive viral replication in nearly all organs, high-dose aciclovir should perhaps only be supplemented with anti-inflammatory medications and inhibitors of TNF, since many of these individuals risk to die from septic reactions.
Competing interests
The author(s) declare that they have no competing interests.
Acknowledgements
I wish to thank all members of the group for highly constructive discussions and especially Søren C. Mogensen for critical review of the manuscript. For excellent technical help with the electron microscopy I thank Ruth Nielsen. Furthermore, I wish to thank The Department of Clinical Microbiology, Aarhus University Hospital, Skejby and Department of Medical Microbiology and Immunology, University of Aarhus for their hospitality and support.
==== Refs
Roizman B Pellett PE Knipe DM and Howley PM The family Herpesviridae: A brief introduction Fields Virology 2001 71 Fourth Lippincott Williams & Wilkins 2381 2397
Whitley RJ Hilliard JK Knipe DM and Howley PM Cercopithecine herpesvirus (B virus) Fields Virology 2001 83 Fourth Lippincott Williams & Wilkins 2835 2848
Lowenstein A Aetiologische untersuchungen über der fieberhaften, herpes Munch Med Wochenschr 1919 66 769 770
Schneweis KE Serologische untersuchungen zur typendifferenzierung des herpesvirus hominis Z Immun Exp Ther 1962 124 24 48 13992216
Roizman B Carmichael LE Deinhardt F de The G Nahmias AJ Plowright W Rapp F Sheldrick P Takahashi M Wolf K Herpesviridae. Definition, provisional nomenclature, and taxonomy. The Herpesvirus Study Group, the International Committee on Taxonomy of Viruses Intervirol 1981 16 201 217
Dolan A Jamieson FE Cunningham C Barnett BC McGeoch DJ The genome sequence of herpes simplex virus type 2 J Virol 1998 72 2010 2021 9499055
McGeoch DJ Dalrymple MA Davison AJ Dolan A Frame MC McNab D Perry LJ Scott JE Taylor P The complete DNA sequence of the long unique region in the genome of herpes simplex virus type 1 J Gen Virol 1988 69 1531 1574 2839594
Hayward GS Frenkel N Roizman B Anatomy of herpes simplex virus DNA: strain differences and heterogeneity in the locations of restriction endonuclease cleavage sites Proc Natl Acad Sci U S A 1975 72 1768 1772 168574
Sakaoka H Kurita K Iida Y Takada S Umene K Kim YT Ren CS Nahmias AJ Quantitative analysis of genomic polymorphism of herpes simplex virus type 1 strains from six countries: studies of molecular evolution and molecular epidemiology of the virus J Gen Virol 1994 75 513 527 8126449
Buchman TG Roizman B Nahmias AJ Demonstration of exogenous genital reinfection with herpes simplex virus type 2 by restriction endonuclease fingerprinting of viral DNA J Infect Dis 1979 140 295 304 227967
Roizman B Knipe DM Knipe DM and Howley PM Herpes simplex viruses and their replication Fields Virology 2001 72 Fourth Lippincott Williams & Wilkins 2399 2459
Highlander SL Cai WH Person S Levine M Glorioso JC Monoclonal antibodies define a domain on Herpes Simplex Virus glycoprotein B involved in virus penetration J Virol 1988 62 1881 1888 2452895
Cai WH Gu B Person S Role of glycoprotein B of herpes simplex virus type 1 in viral entry and cell fusion J Virol 1988 62 2596 2604 2839688
Highlander SL Sutherland SL Gage PJ Johnson DC Levine M Glorioso JC Neutralizing monoclonal antibodies specific for herpes simplex virus glycoprotein D inhibit virus penetration J Virol 1987 61 3356 3364 2444713
Lubinski JM Wang L Soulika AM Burger R Wetsel RA Colten H Cohen GH Eisenberg RJ Lambris JD Friedman HM Herpes simplex virus type 1 glycoprotein gC mediates immune evasion in vivo J Virol 1998 72 8257 8263 9733869
Lubinski JM Jiang M Hook L Chang Y Sarver C Mastellos D Lambris JD Cohen GH Eisenberg RJ Friedman HM Herpes simplex virus type 1 evades the effects of antibody and complement in vivo J Virol 2002 76 9232 9241 12186907
Ankel H Westra DF Welling-Wester S Lebon P Induction of interferon-alpha by glycoprotein D of herpes simplex virus: a possible role of chemokine receptors Virol 1998 251 317 326
Patel A Hanson J McLean TI Olgiate J Hilton M Miller WE Bachenheimer SL Herpes simplex type 1 induction of persistent NF-kappa B nuclear translocation increases the efficiency of virus replication Virol 1998 247 212 222
Teng YC Chen YT Cheng YY Hung SL Requirements for the upregulation of interleukin-6 by herpes simplex virus-infected gingival fibroblasts Viral Immunol 2005 18 170 178 15802961
Roizman B Furlong D Fraenkel-Conrat H and Wagner RR The replication of herpesviruses. Comprehensive Virology 1974 New York, Plenum Press 229 403
Zhou ZH Chen DH Jakana J Rixon FJ Chiu W Visualization of tegument-capsid interactions and DNA in intact herpes simplex virus type 1 virions J Virol 1999 73 3210 3218 10074174
Read GS Frenkel N Herpes simplex virus mutants defective in the viron-associated shutoff of host polypeptide synthesis and exhibiting abnormal synthesis of a (immediate early) viral polypeptides J Virol 1983 46 498-512 498-512 6302315
Batterson W Roizman B Characterization of the herpes simplex virion-associated factor responsible for the induction of alpha genes J Virol 1983 46 371 377 6302308
McLauchlan J Rixon FJ Characterization of enveloped tegument structures (L particles) produced by alphaherpesviruses: integrity of the tegument does not depend on the presence of capsid or envelope J Gen Virol 1992 73 269 276 1311356
Szilagyi JF Cunningham C Identification and characterization of a novel non-infectious herpes simplex virus-related particle J Gen Virol 1991 72 661 668 1848601
Furlong D Swift H Roizman B Arrangement of herpesvirus deoxyribonucleic acid in the core J Virol 1972 10 1071 1074 4117962
Poffenberger KL Roizman B A noninverting genome of a viable herpes simplex virus 1: presence of head-to-tail linkages in packaged genomes and requirements for circularization after infection J Virol 1985 53 587 595 2982037
Wadsworth S Jacob RJ Roizman B Anatomy of herpes simplex virus DNA. II. Size, composition, and arrangement of inverted terminal repetitions J Virol 1975 15 1487 1497 167196
Rajcani J Andrea V Ingeborg R Peculiarities of herpes simplex virus (HSV) transcription: an overview Virus Genes 2004 28 293 310 15266111
Ward PL Roizman B Herpes simplex genes: the blueprint of a successful human pathogen Trends Genet 1994 10 267 274 7940755
Nishiyama Y Herpes simplex virus gene products: the accessories reflect her lifestyle well Rev Med Virol 2004 14 33 46 14716690
O'Hare P The virion transactivator of herpes simplex virus Seminars in Virology 1993 4 145 155
Whitley RJ Knipe DM and Howley PM Herpes simplex viruses Fields Virology 2001 73 Fourth Lippincott Williams & Wilkins 2461 2509
Post LE Mackem S Roizman B Regulation of alpha genes of herpes simplex virus: expression of chimeric genes produced by fusion of thymidine kinase with alpha gene promoters Cell 1981 24 555 565 6263501
Shukla D Spear PG Herpesviruses and heparan sulfate: an intimate relationship in aid of viral entry J Clin Invest 2001 108 503 510 11518721
Nahmias AJ Kibrick S Inhibitory effect of heparin on herpes simplex virus J Bacteriol 1964 87 1060 1066 4289440
WuDunn D Spear PG Initial interaction of herpes simplex virus with cells is binding to heparan sulfate J Virol 1989 63 52 58 2535752
Gerber SI Belval BJ Herold BC Differences in the role of glycoprotein C of HSV-1 and HSV-2 in viral binding may contribute to serotype differences in cell tropism Virol 1995 214 29 39
Trybala E Liljeqvist JA Svennerholm B Bergstrom T Herpes simplex virus types 1 and 2 differ in their interaction with heparan sulfate J Virol 2000 74 9106 9114 10982357
Huard J Feero WG Watkins SC Hoffman EP Rosenblatt DJ Glorioso JC The basal lamina is a physical barrier to herpes simplex virus-mediated gene delivery to mature muscle fibers J Virol 1996 70 8117 8123 8892937
Spear PG Longnecker R Herpesvirus entry: an update J Virol 2003 77 10179 10185 12970403
Spear PG Eisenberg RJ Cohen GH Three classes of cell surface receptors for alphaherpesvirus entry Virol 2000 275 1 8
Montgomery RI Warner MS Lum BJ Spear PG Herpes simplex virus-1 entry into cells mediated by a novel member of the TNF/NGF receptor family Cell 1996 87 427 436 8898196
Geraghty RJ Krummenacher C Cohen GH Eisenberg RJ Spear PG Entry of alphaherpesviruses mediated by poliovirus receptor-related protein 1 and poliovirus receptor Science 1998 280 1618 1620 9616127
Warner MS Geraghty RJ Martinez WM Montgomery RI Whitbeck JC Xu R Eisenberg RJ Cohen GH Spear PG A cell surface protein with herpesvirus entry activity (HveB) confers susceptibility to infection by mutants of herpes simplex virus type 1, herpes simplex virus type 2, and pseudorabies virus Virol 1998 246 179 189
Shukla D Liu J Blaiklock P Shworak NW Bai X Esko JD Cohen GH Eisenberg RJ Rosenberg RD Spear PG A novel role for 3-O-sulfated heparan sulfate in herpes simplex virus 1 entry Cell 1999 99 13 22 10520990
Johnson DC Webb M Wisner TW Brunetti C Herpes simplex virus gE/gI sorts nascent virions to epithelial cell junctions, promoting virus spread J Virol 2001 75 821 833 11134295
Haarr L Shukla D Rodahl E Dal Canto MC Spear PG Transcription from the gene encoding the herpesvirus entry receptor nectin-1 (HveC) in nervous tissue of adult mouse Virol 2001 287 301 309
Campadelli-Fiume G Cocchi F Menotti L Lopez M The novel receptors that mediate the entry of herpes simplex viruses and animal alphaherpesviruses into cells Rev Med Virol 2000 10 305 319 11015742
Nicola AV Hou J Major EO Straus SE Herpes simplex virus type 1 enters human epidermal keratinocytes, but not neurons, via a pH-dependent endocytic pathway J Virol 2005 79 7609 7616 15919913
Milne RS Nicola AV Whitbeck JC Eisenberg RJ Cohen GH Glycoprotein D receptor-dependent, low-pH-independent endocytic entry of herpes simplex virus type 1 J Virol 2005 79 6655 6663 15890903
Linehan MM Richman S Krummenacher C Eisenberg RJ Cohen GH Iwasaki A In vivo role of nectin-1 in entry of herpes simplex virus type 1 (HSV-1) and HSV-2 through the vaginal mucosa J Virol 2004 78 2530 2536 14963155
Perez-Romero P Perez A Capul A Montgomery R Fuller AO Herpes simplex virus entry mediator associates in infected cells in a complex with viral proteins gD and at least gH J Virol 2005 79 4540 4544 15767456
Wittels M Spear PG Penetration of cells by herpes simplex virus does not require a low pH-dependent endocytic pathway Virus Res 1991 18 271 290 1645908
Nicola AV McEvoy AM Straus SE Roles for endocytosis and low pH in herpes simplex virus entry into HeLa and Chinese hamster ovary cells J Virol 2003 77 5324 5332 12692234
Nicola AV Straus SE Cellular and viral requirements for rapid endocytic entry of herpes simplex virus J Virol 2004 78 7508 7517 15220424
Spear PG Herpes simplex virus: receptors and ligands for cell entry Cell Microbiol 2004 6 401 410 15056211
Mauri DN Ebner R Montgomery RI Kochel KD Cheung TC Yu GL Ruben S Murphy M Eisenberg RJ Cohen GH Spear PG Ware CF LIGHT, a new member of the TNF superfamily, and lymphotoxin alpha are ligands for herpesvirus entry mediator Immunity 1998 8 21 30 9462508
La S Kim J Kwon BS Kwon B Herpes simplex virus type 1 glycoprotein D inhibits T-cell proliferation Mol Cells 2002 14 398 403 12521303
Sodeik B Ebersold MW Helenius A Microtubule-mediated transport of incoming herpes simplex virus 1 capsids to the nucleus J Cell Biol 1997 136 1007 1021 9060466
Ojala PM Sodeik B Ebersold MW Kutay U Helenius A Herpes simplex virus type 1 entry into host cells: reconstitution of capsid binding and uncoating at the nuclear pore complex in vitro Mol Cell Biol 2000 20 4922 4931 10848617
Schang LM Phillips J Schaffer PA Requirement for cellular cyclin-dependent kinases in herpes simplex virus replication and transcription J Virol 1998 72 5626 5637 9621021
Kent JR Kang W Miller CG Fraser NW Herpes simplex virus latency-associated transcript gene function J Neurovirol 2003 9 285 290 12775412
Steiner I Spivack JG Deshmane SL Ace CI Preston CM Fraser NW A herpes simplex virus type 1 mutant containing a nontransinducing Vmw65 protein establishes latent infection in vivo in the absence of viral replication and reactivates efficiently from explanted trigeminal ganglia J Virol 1990 64 1630 1638 2157048
Spivack JG Fraser NW Expression of herpes simplex virus type 1 (HSV-1) latency-associated transcripts and transcripts affected by the deletion in avirulent mutant HFEM: evidence for a new class of HSV-1 genes J Virol 1988 62 3281 3287 2841480
Sawtell NM Thompson RL Herpes simplex virus type 1 latency-associated transcription unit promotes anatomical site-dependent establishment and reactivation from latency J Virol 1992 66 2157 2169 1312626
Thompson RL Sawtell NM The herpes simplex virus type 1 latency-associated transcript gene regulates the establishment of latency J Virol 1997 71 5432 5440 9188615
Jones C Herpes simplex virus type 1 and bovine herpesvirus 1 latency Clin Microbiol Rev 2003 16 79 95 12525426
Roizman B HSV gene functions: what have we learned that could be generally applicable to its near and distant cousins? Acta Virol 1999 43 75 80 10696424
Hagglund R Roizman B Role of ICP0 in the strategy of conquest of the host cell by herpes simplex virus 1 J Virol 2004 78 2169 2178 14963113
Melroe GT DeLuca NA Knipe DM Herpes simplex virus 1 has multiple mechanisms for blocking virus-induced interferon production J Virol 2004 78 8411 8420 15280450
Mossman KL Saffran HA Smiley JR Herpes simplex virus ICP0 mutants are hypersensitive to interferon J Virol 2000 74 2052 2056 10644380
Harle P Sainz BJ Carr DJ Halford WP The immediate-early protein, ICP0, is essential for the resistance of herpes simplex virus to interferon-alpha/beta Virol 2002 293 295 304
Everett RD ICP0, a regulator of herpes simplex virus during lytic and latent infection Bioessays 2000 22 761 770 10918307
Boutell C Everett RD The herpes simplex virus type 1 (HSV-1) regulatory protein ICP0 interacts with and Ubiquitinates p53 J Biol Chem 2003 278 36596 36602 12855695
Everett RD Sourvinos G Leiper C Clements JB Orr A Formation of nuclear foci of the herpes simplex virus type 1 regulatory protein ICP4 at early times of infection: localization, dynamics, recruitment of ICP27, and evidence for the de novo induction of ND10-like complexes J Virol 2004 78 1903 1917 14747555
Maul GG Everett RD The nuclear location of PML, a cellular member of the C3HC4 zinc-binding domain protein family, is rearranged during herpes simplex virus infection by the C3HC4 viral protein ICP0 J Gen Virol 1994 75 1223 1233 8207389
Everett RD Zafiropoulos A Visualization by live-cell microscopy of disruption of ND10 during herpes simplex virus type 1 infection J Virol 2004 78 11411 11415 15452264
Everett RD Sourvinos G Orr A Recruitment of herpes simplex virus type 1 transcriptional regulatory protein ICP4 into foci juxtaposed to ND10 in live, infected cells J Virol 2003 77 3680 3689 12610143
Everett RD Murray J ND10 components relocate to sites associated with herpes simplex virus type 1 nucleoprotein complexes during virus infection J Virol 2005 79 5078 5089 15795293
Sandri-Goldin RM Properties of an HSV-1 regulatory protein that appears to impair host cell splicing Infect Agents Dis 1994 3 59 67 7812656
Bauer D Tampe R Herpes viral proteins blocking the transporter associated with antigen processing TAP - from genes to function and structure Curr Top Microbiol Immunol 2002 269 87 99 12224518
van Genderen IL Brandimarti R Torrisi MR Campadelli G van Meer G The phospholipid composition of extracellular herpes simplex virions differs from that of host cell nuclei Virol 1994 200 831 836
Browne H Bell S Minson T Wilson DW An endoplasmic reticulum-retained herpes simplex virus glycoprotein H is absent from secreted virions: evidence for reenvelopment during egress J Virol 1996 70 4311 4316 8676453
Whiteley A Bruun B Minson T Browne H Effects of targeting herpes simplex virus type 1 gD to the endoplasmic reticulum and trans-Golgi network J Virol 1999 73 9515 9520 10516060
Brack AR Klupp BG Granzow H Tirabassi R Enquist LW Mettenleiter TC Role of the cytoplasmic tail of pseudorabies virus glycoprotein E in virion formation J Virol 2000 74 4004 4016 10756012
Dingwell KS Brunetti CR Hendricks RL Tang Q Tang M Rainbow AJ Johnson DC Herpes simplex virus glycoproteins E and I facilitate cell-to-cell spread in vivo and across junctions of cultured cells J Virol 1994 68 834 845 8289387
Dowdle WR Nahmias AJ Harwell RW Pauls FP Association of antigenic type of Herpesvirus hominis with site of viral recovery J Immunol 1967 99 974 980 4295116
Strutt M Bailey J Tenant-Flowers M Graham D Zuckerman M Ethnic variation in type of genital herpes simplex virus infection in a South London genitourinary medicine clinic J Med Virol 2003 69 108 110 12436485
Cowan FM French RS Mayaud P Gopal R Robinson NJ de Oliveira SA Faillace T Uuskula A Nygard-Kibur M Ramalingam S Sridharan G El Aouad R Alami K Rbai M Sunil-Chandra NP Brown DW Seroepidemiological study of herpes simplex virus types 1 and 2 in Brazil, Estonia, India, Morocco, and Sri Lanka Sex Transm Infect 2003 79 286 290 12902576
Suligoi B Torri A Grilli G Tanzi E Palu G Seroprevalence and seroincidence of herpes simplex virus type 1 and herpes simplex virus type 2 infections in a cohort of adolescents in Italy Sex Transm Dis 2004 31 608 610 15388998
Roest RW van der Meijden WI van Dijk G Groen J Mulder PG Verjans GM Osterhaus AD Prevalence and association between herpes simplex virus types 1 and 2-specific antibodies in attendees at a sexually transmitted disease clinic Int J Epidemiol 2001 30 580 588 11416087
Hashido M Lee FK Nahmias AJ Tsugami H Isomura S Nagata Y Sonoda S Kawana T An epidemiologic study of herpes simplex virus type 1 and 2 infection in Japan based on type-specific serological assays Epidemiol Infect 1998 120 179 186 9593488
Nahmias AJ Lee FK Beckman-Nahmias S Sero-epidemiological and -sociological patterns of herpes simplex virus infection in the world Scand J Infect Dis Suppl 1990 69 19 36 2175939
Theng TS Chan RK Genital herpes in a sexually-transmitted infection clinic in Singapore: a 1-year retrospective study Ann Acad Med Singapore 2004 33 200 203 15098634
Solomon L Cannon MJ Reyes M Graber JM Wetherall NT Reeves WC Epidemiology of recurrent genital herpes simplex virus types 1 and 2 Sex Transm Infect 2003 79 456 459 14663120
Tran T Druce JD Catton MC Kelly H Birch CJ Changing epidemiology of genital herpes simplex virus infection in Melbourne, Australia, between 1980 and 2003 Sex Transm Infect 2004 80 277 279 15295125
Buxbaum S Geers M Gross G Schofer H Rabenau HF Doerr HW Epidemiology of herpes simplex virus types 1 and 2 in Germany: what has changed? Med Microbiol Immunol (Berl) 2003 192 177 181 12761673
Nilsen A Myrmel H Changing trends in genital herpes simplex virus infection in Bergen, Norway Acta Obstet Gynecol Scand 2000 79 693 696 10949236
Manavi K McMillan A Ogilvie M Herpes simplex virus type 1 remains the principal cause of initial anogenital herpes in Edinburgh, Scotland Sex Transm Dis 2004 31 322 324 15107637
Lafferty WE Downey L Celum C Wald A Herpes Simplex Virus Type 1 as a Cause of Genital Herpes: Impact on Surveillance and Prevention J Infect Dis 2000 181 1454 1457 10762576
Cowan FM Copas A Johnson AM Ashley R Corey L Mindel A Herpes simplex virus type 1 infection: a sexually transmitted infection of adolescence? Sex Transm Infect 2002 78 346 348 12407237
Langenberg AGM Corey L Ashley RL Leong WP Straus SE A prospective study of new infections with herpes simplex virus type 1 and type 2. Chiron HSV Vaccine Study Group N Engl J Med 1999 341 1432 1438 10547406
Oliver L Wald A Kim M Zeh J Selke S Ashley R Corey L Seroprevalence of herpes simplex virus infections in a family medicine clinic Arch Fam Med 1995 4 228 232 7881604
Engelberg R Carrell D Krantz E Corey L Wald A Natural history of genital herpes simplex virus type 1 infection Sex Transm Dis 2003 30 174 177 12567178
Cesario TC Poland JD Wulff H Chin TD Wenner HA Six years experience with herpes simplex virus in a children's home Am J Epidemiol 1969 90 416 422 4311429
McMillan JA Weiner LB Higgins AM Lamparella VJ Pharyngitis associated with herpes simplex virus in college students Pediatr Infect Dis J 1993 12 280 284 8387178
Corey L Mandell GL, Bennett JE and Dolin R Herpes simplex virus Principles and practice of infectious diseases 2000 125 Fifth Churchill Livingstone 1564 1580
Prober CG Hensleigh PA Boucher FD Yasukawa LL Au DS Arvin AM Use of routine viral cultures at delivery to identify neonates exposed to herpes simplex virus N Engl J Med 1988 318 887 891 2832756
Nahmias AJ Disseminated herpes-simplex-virus infections N Engl J Med 1970 282 684 685 4190592
Joseph TJ Vogt PJ Disseminated herpes with hepatoadrenal necrosis in an adult Am J Med 1974 56 735 739 4363476
Keane JT Malkinson FD Bryant J Levin S Herpesvirus hominis hepatitis and disseminated intravascular coagulation. Occurrence in an adult with pemphigus vulgaris Arch Intern Med 1976 136 1312 1317 185974
Whitley RJ Roizman B Herpes simplex virus infections Lancet 2001 357 1513 1518 11377626
Flewett TH Parker RG Philip WM Acute hepatitis due to Herpes simplex virus in an adult J Clin Pathol 1969 22 60 66 4306578
Hillard P Seeds J Cefalo R Disseminated herpes simplex in pregnancy: two cases and a review Obstet Gynecol Surv 1982 37 449 453 6752785
Linnemann CCJ First MR Alvira MM Alexander JW Schiff GM Herpesvirus hominis type 2 meningoencephalitis following renal transplantation Am J Med 1976 61 703 708 185900
Brown ZA Selke S Zeh J Kopelman J Maslow A Ashley RL Watts DH Berry S Herd M Corey L The acquisition of herpes simplex virus during pregnancy N Engl J Med 1997 337 509 515 9262493
Brown ZA Benedetti J Ashley R Burchett S Selke S Berry S Vontver LA Corey L Neonatal herpes simplex virus infection in relation to asymptomatic maternal infection at the time of labor N Engl J Med 1991 324 1247 1252 1849612
Whitley RJ Herpes simplex virus infections of women and their offspring: implications for a developed society Proc Natl Acad Sci U S A 1994 91 2441 2447 8146137
Whitley RJ Neonatal herpes simplex virus infections: is there a role for immunoglobulin in disease prevention and therapy? Pediatr Infect Dis J 1994 13 432 438 8072834
Sullender WM Yasukawa LL Schwartz M Pereira L Hensleigh PA Prober CG Arvin AM Type-specific antibody to Herpes Simplex Virus type 2(HSV-2) glycoprotein G in pregnant women, infants exposed to maternal HSV- 2 infection at delivery, and infants with Neonatal Herpes J Infect Dis 1988 157 164 171 2826604
Prober CG Sullender WM Yasukawa LL Au DS Yeager AS Arvin AM Low risk of herpes simplex virus infections in neonates exposed to the virus at the time of vaginal delivery to mothers with recurrent genital herpes simplex virus infections N Engl J Med 1987 316 240 244 3025727
Kimberlin DW Neonatal herpes simplex infection Clin Microbiol Rev 2004 17 1 13 14726453
Kimberlin DW Lin CY Jacobs RF Powell DA Corey L Gruber WC Rathore M Bradley JS Diaz PS Kumar M Arvin AM Gutierrez K Shelton M Weiner LB Sleasman JW de Sierra TM Weller S Soong SJ Kiell J Lakeman FD Whitley RJ Safety and efficacy of high-dose intravenous acyclovir in the management of neonatal herpes simplex virus infections Pediatrics 2001 108 230 238 11483782
Kimberlin DW Lin CY Jacobs RF Powell DA Frenkel LM Gruber WC Rathore M Bradley JS Diaz PS Kumar M Arvin AM Gutierrez K Shelton M Weiner LB Sleasman JW de Sierra TM Soong SJ Kiell J Lakeman FD Whitley RJ Natural history of neonatal herpes simplex virus infections in the acyclovir era Pediatrics 2001 108 223 229 11483781
Kimberlin DW Advances in the treatment of neonatal herpes simplex infections Rev Med Virol 2001 11 157 163 11376479
Boddingius J Dijkman H Hendriksen E Schift R Stolz E HSV-2 replication sites, monocyte and lymphocytic cell infection and virion phagocytosis by neutrophils, in vesicular lesions on penile skin. Electronoptical studies of a biopsy J Cutan Pathol 1987 14 165 175 3038975
Straus SE Mandell GL, Bennett JE and Dolin R Herpesviridae Principles and practice of infectious diseases 2000 124 Fifth Churchill Livingstone 1557 1564
Skoldenberg B Herpes simplex encephalitis Scand J Infect Dis Suppl 1996 100 8 13 9163027
Huster KM Panoutsakopoulou V Prince K Sanchirico ME Cantor H T cell-dependent and -independent pathways to tissue destruction following herpes simplex virus-1 infection Eur J Immunol 2002 32 1414 1419 11981829
Biswas PS Banerjee K Kim B Rouse BT Mice transgenic for IL-1 receptor antagonist protein are resistant to herpetic stromal keratitis: possible role for IL-1 in herpetic stromal keratitis pathogenesis J Immunol 2004 172 3736 3744 15004178
Zhao ZS Granucci F Yeh L Schaffer PA Cantor H Molecular mimicry by herpes simplex virus-type 1: autoimmune disease after viral infection Science 1998 279 1344 1347 9478893
Adler H Beland JL Kozlow W Del Pan NC Kobzik L Rimm IJ A role for transforming growth factor-beta1 in the increased pneumonitis in murine allogeneic bone marrow transplant recipients with graft-versus-host disease after pulmonary herpes simplex virus type 1 infection Blood 1998 92 2581 2589 9746800
Kurt-Jones EA Chan M Zhou S Wang J Reed G Bronson R Arnold MM Knipe DM Finberg RW Herpes simplex virus 1 interaction with Toll-like receptor 2 contributes to lethal encephalitis Proc Natl Acad Sci U S A 2004 101 1315 1320 14739339
Adler H Beland JL Del Pan NC Kobzik L Brewer JP Martin TR Rimm IJ Suppression of herpes simplex virus type 1 (HSV-1)-induced pneumonia in mice by inhibition of inducible nitric oxide synthase (iNOS, NOS2) J Exp Med 1997 185 1533 1540 9151890
Kawada J Kimura H Ito Y Ando Y Tanaka-Kitajima N Hayakawa M Nunoi H Endo F Morishima T Evaluation of systemic inflammatory responses in neonates with herpes simplex virus infection J Infect Dis 2004 190 494 498 15243922
Metchnikoff E Lecons sur la pathologie comparée de l'inflammation 1892 Paris, Masson
Aschoff L Das reticulo-endotheliale system Ergeb Inn Med Kinderheilkd 1924 26 1 118
van Furth R Cohn ZA Hirsch JG Humphrey JH Spector WG Langevoort HL The mononuclear phagocyte system: a new classification of macrophages, monocytes, and their precursor cells Bull World Health Organ 1972 46 845 852 4538544
van Furth R Cohn ZA The origin and kinetics of mononuclear phagocytes J Exp Med 1968 128 415 435 5666958
Laskin DL Weinberger B Laskin JD Functional heterogeneity in liver and lung macrophages J Leukoc Biol 2001 70 163 170 11493607
Benacerraf B Sebestyen MM Schlossman S A quantitative study of the kinetics of blood clearance of P32-labelled Escherichia coli and Staphylococci by the reticuloendothelial system J Exp Med 1959 110 27 48 13664867
Brunner KT Hurez D McCluskey RT Benacerraf B Blood clearance of P32-labeled vesicular stomatitis and Newcastle disease viruses by the reticuloendothelial system in mice J Immunol 1960 85 99 105 13805345
Lepay DA Nathan CF Steinman RM Murray HW Cohn ZA Murine Kupffer cells. Mononuclear phagocytes deficient in the generation of reactive oxygen intermediates J Exp Med 1985 161 1079 1096 3921651
Lee SH Crocker P Gordon S Macrophage plasma membrane and secretory properties in murine malaria. Effects of Plasmodium yoelii blood-stage infection on macrophages in liver, spleen, and blood J Exp Med 1986 163 54 74 3001215
Guillemin GJ Brew BJ Microglia, macrophages, perivascular macrophages, and pericytes: a review of function and identification J Leukoc Biol 2004 75 388 397 14612429
Stout RD Suttles J Functional plasticity of macrophages: reversible adaptation to changing microenvironments J Leukoc Biol 2004 76 509 513 15218057
Denis M Human monocytes / macrophages: NO or no NO? J Leukoc Biol 1994 55 682 684 8182347
MacMicking J Xie QW Nathan C Nitric oxide and macrophage function Annu Rev Immunol 1997 15 323 350 9143691
Crofton RW Diesselhoff-den Dulk MM van Furth R The origin, kinetics, and characteristics of the Kupffer cells in the normal steady state J Exp Med 1978 148 1 17 670884
Blussé van Oud Alblas A van Furth R Origin, Kinetics, and characteristics of pulmonary macrophages in the normal steady state J Exp Med 1979 149 1504 1518 448291
van Furth R Diesselhoff-den Dulk MM Dual origin of mouse spleen macrophages J Exp Med 1984 160 1273 1283 6491600
Blussé van Oud Alblas A Mattie H van Furth R A quantitative evaluation of pulmonary macrophage kinetics Cell Tissue Kinet 1983 16 211 219 6839345
Merad M Manz MG Karsunky H Wagers A Peters W Charo I Weissman IL Cyster JG Engleman EG Langerhans cells renew in the skin throughout life under steady-state conditions Nat Immunol 2002 3 1135 1141 12415265
van Furth R Diesselhoff-den Dulk MC Mattie H Quantitative study on the production and kinetics of mononuclear phagocytes during an acute inflammatory reaction J Exp Med 1973 138 1314 1330 4762549
Mackaness GB Cellular resistance to infection J Exp Med 1962 116 381 406 14467923
Mackaness GB Blanden RV Collins FM Host-parasite relations in mouse typhoid J Exp Med 1966 124 573 583 5922285
Szuro-Sudol A Murray HW Nathan CF Suppression of macrophages antimicrobial activity by a tumor cell product J Immunol 1983 131 384-387 384-387 6408180
Andrews T Sullivan KE Infections in patients with inherited defects in phagocytic function Clin Microbiol Rev 2003 16 597 621 14557288
Mauel J Buchmuller Y Behin R Studies on the mechanisms of macrophage activation. I. Destruction of intracellular Leishmania enriettii in macrophages activated by cocultivation with stimulated lymphocytes J Exp Med 1978 148 393 407 308979
Gordon S Macrophage activation and differentiation Ann Inst Pasteur Immunol 1986 137c 197 200 3524411
Denis M Tumor necrosis factor and granulocyte macrophage-colony stimulating factor stimulate human macrophages to restrict growth of virulent Mycobacterium avium and to kill avirulent M. avium: killing effector mechanism depends on the generation of reactive nitrogen intermediates J Leukoc Biol 1991 49 380 387 1900522
Nathan CF Root RK Hydrogen peroxide from mouse peritoneal macrophages J Exp Med 1977 146 1648-1662 1648-1662 925614
Karupiah G Hunt NH King NJ Chaudhri G NADPH oxidase, Nramp1 and nitric oxide synthase 2 in the host antimicrobial response Rev Immunogenet 2000 2 387 415 11256747
Bennedsen J Riisgaard S Rhodes JM Larsen SO In vitro studies on normal, stimulated and immunologically activated mouse macrophages. III. Intracellular multiplication of Listeria monocytogenes Acta Pathol Microbiol Scand C 1977 85C 246 252 409099
Mantovani A Sica A Sozzani S Allavena P Vecchi A Locati M The chemokine system in diverse forms of macrophage activation and polarization Trends Immunol 2004 25 677 686 15530839
Sluiter W Elzenga-Claasen I van der Voort van der Kley-van Andel van Furth R Differences in the response of inbred mouse strains to the factor increasing monocytopoiesis J Exp Med 1984 159 524 536 6693834
Ellermann-Eriksen S Liberto MC Iannello D Mogensen SC X-linkage of the early in vitro alpha/beta interferon response of mouse peritoneal macrophages to herpes simplex virus type 2 J Gen Virol 1986 67 1025 1033 3011968
Marks J James DM The effect of tuberculin on sensitized and normal leucocytes J Hyg (Lond) 1953 51 340 346 13096742
Isaacs A Lindenmann J Virus interference. I. The interferon Proc R Soc Lond B Biol Sci 1957 147 258 267 13465720
Pestka S Krause CD Walter MR Interferons, interferon-like cytokines, and their receptors Immunol Rev 2004 202 8 32 15546383
Dumonde DC Wolstencroft RA Panayi GS Matthew M Morley J Howson WT "Lymphokines": non-antibody mediators of cellular immunity generated by lymphocyte activation Nature 1969 224 38 42 5822903
Mizel SB Farrar JJ Revised nomenclature for antigen-nonspecific T-cell proliferation and helper factors Cell Immunol 1979 48 433 436 92371
Waksman BH International Lymphokine Nomenclature Committee J Immunol 1980 124 1510 1511
Paul WE International Union of Immunological Societies. Nomenclature Committee Working Group on Lymphokines J Immunol Methods 1988 115 1 2 3192939
Schrader JW Interleukin is as interleukin does Trends Immunol 2002 23 573 574 12464567
Balkwill FR Burke F The cytokine network Immunol Today 1989 10 299 304 2686679
Trinchieri G Cytokines and cytokine receptors Immunol Rev 2004 202 5 7 15546382
Zawatzky R Gresser I De Mayer E Kirchner H The role of interferon in the resistance of C57BL/6 mice to various doses of herpes simplex virus type 1 J Infect Dis 1982 146 405 410 6180032
Hoskins M A protective action of neurotropic against viscerotropic yellow fever virus in Macacus rhesus Am J Trop Med 1935 15 675 680
Henle W henle G Interference of inactive virus with the propagation of virus of influenza Science 1943 98 87 89 17749157
Henle W Interference phenomena between animal viruses; a review J Immunol 1950 64 203 236 15412251
Pestka S Langer JA Zoon KC Samuel CE Interferons and their actions Annu Rev Biochem 1987 56 727 777 2441659
Oritani K Tomiyama Y Interferon-zeta/limitin: novel type I interferon that displays a narrow range of biological activity Int J Hematol 2004 80 325 331 15615256
van PV Lanaya H Renauld JC Michiels T Characterization of the murine alpha interferon gene family J Virol 2004 78 8219 8228 15254193
Evinger M Rubinstein M Pestka S Antiproliferative and antiviral activities of human leukocyte interferons Arch Biochem Biophys 1981 210 319 329 6170263
Ortaldo JR Herberman RB Harvey C Osheroff P Pan YC Kelder B Pestka S A species of human alpha interferon that lacks the ability to boost human natural killer activity Proc Natl Acad Sci U S A 1984 81 4926 4929 6589637
Runkel L Meier W Pepinsky RB Karpusas M Whitty A Kimball K Brickelmaier M Muldowney C Jones W Goelz SE Structural and functional differences between glycosylated and non-glycosylated forms of human interferon-beta (IFN-beta) Pharm Res 1998 15 641 649 9587963
Pestka S Larocca AT Borden EC Colby CB Dower SK Vries JE Taga T Kishimoto T Stanley ER Aggarwal BB Gutterman JU Aggarwal BB and Gutterman JU Human cytokines - Handbook for basic and clinical research 1992 1 Boston, Blackwell Scientific Publications 1 435
David M Signal transduction by type I interferons Biotechniques 2002 33 S58 S65
Oritani K Medina KL Tomiyama Y Ishikawa J Okajima Y Ogawa M Yokota T Aoyama K Takahashi I Kincade PW Matsuzawa Y Limitin: An interferon-like cytokine that preferentially influences B-lymphocyte precursors Nat Med 2000 6 659 666 10835682
Kawamoto S Oritani K Asada H Takahashi I Ishikawa J Yoshida H Yamada M Ishida N Ujiie H Masaie H Tomiyama Y Matsuzawa Y Antiviral activity of limitin against encephalomyocarditis virus, herpes simplex virus, and mouse hepatitis virus: diverse requirements by limitin and alpha interferon for interferon regulatory factor 1 J Virol 2003 77 9622 9631 12915574
Green JA Cooperband SR Kibrick S Immune specific induction of interferon production in cultures of human blood lymphocytes Science 1969 164 1415 1417 5783715
Bancroft GJ Bosma MJ Bosma GC Unanue ER Regulation of macrophage Ia expression in mice with severe combined immunodeficiency: induction of Ia expression by a T cell-independent mechanism J Immunol 1986 137 4 9 3086456
Bancroft GJ Schreiber RD Bosma GC Bosma MJ Unanue ER A T cell-independent mechanism of macrophage activation by interferon-gamma J Immunol 1987 139 1104 1107 3112223
Young HA Regulation of interferon-gamma gene expression J Interferon Cytokine Res 1996 16 563 568 8877725
Schindler H Lutz MB Rollinghoff M Bogdan C The production of IFN-gamma by IL-12/IL-18-activated macrophages requires STAT4 signaling and is inhibited by IL-4 J Immunol 2001 166 3075 3082 11207258
Yoshimoto T Okamura H Tagawa YI Iwakura Y Nakanishi K Interleukin 18 together with interleukin 12 inhibits IgE production by induction of interferon-gamma production from activated B cells Proc Natl Acad Sci U S A 1997 94 3948 3953 9108085
Yoshimoto T Takeda K Tanaka T Ohkusu K Kashiwamura S Okamura H Akira S Nakanishi K IL-12 up-regulates IL-18 receptor expression on T cells, Th1 cells, and B cells: synergism with IL-18 for IFN-gamma production J Immunol 1998 161 3400 3407 9759857
Gessani S Belardelli F IFN-gamma expression in macrophages and its possible biological significance Cytokine Growth Factor Rev 1998 9 117 123 9754706
Carnaud C Lee D Donnars O Park SH Beavis A Koezuka Y Bendelac A Cutting edge: Cross-talk between cells of the innate immune system: NKT cells rapidly activate NK cells J Immunol 1999 163 4647 4650 10528160
Frucht DM Fukao T Bogdan C Schindler H O'Shea JJ Koyasu S IFN-gamma production by antigen-presenting cells: mechanisms emerge Trends Immunol 2001 22 556 560 11574279
Rosenzweig SD Holland SM Congenital defects in the interferon-gamma / interleukin-12 pathway Curr Opin Pediatr 2004 16 3 8 14758107
Schroder K Hertzog PJ Ravasi T Hume DA Interferon-gamma: an overview of signals, mechanisms and functions J Leukoc Biol 2004 75 163 189 14525967
Bancroft GJ Sheehan KC Schreiber RD Unanue ER Tumor necrosis factor is involved in the T cell-independent pathway of macrophage activation in scid mice J Immunol 1989 143 127 130 2499625
Lieberman LA Hunter CA Regulatory pathways involved in the infection-induced production of IFN-gamma by NK cells Microbes Infect 2002 4 1531 1538 12505525
Krause CD Mei E Xie J Jia Y Bopp MA Hochstrasser RM Pestka S Seeing the light: preassembly and ligand-induced changes of the interferon gamma receptor complex in cells Mol Cell Proteomics 2002 1 805 815 12438563
Braunstein J Brutsaert S Olson R Schindler C STATs dimerize in the absence of phosphorylation J Biol Chem 2003 278 34133 34140 12832402
Ota N Brett TJ Murphy TL Fremont DH Murphy KM N-domain-dependent nonphosphorylated STAT4 dimers required for cytokine-driven activation Nat Immunol 2004 5 208 215 14704793
Subramaniam PS Torres BA Johnson HM So many ligands, so few transcription factors: a new paradigm for signaling through the STAT transcription factors Cytokine 2001 15 175 187 11563878
Decker T Kovarik P Meinke A GAS elements: a few nucleotides with a major impact on cytokine-induced gene expression J Interferon Cytokine Res 1997 17 121 134 9085936
Paludan SR Interleukin-4 and interferon-gamma: the quintessence of a mutual antagonistic relationship Scand J Immunol 1998 48 459 468 9822252
Nathan CF Karnovsky ML David JR Alterations of macrophage functions by mediators from lymphocytes J Exp Med 1971 133 1356 1376 5576335
Nathan CF Murray HW Wiebe ME Rubin BY Identification of interferon-gamma as the lymphokine that activates human macrophage oxidative metabolism and antimicrobial activity J Exp Med 1983 158 670 689 6411853
Schreiber RD Hicks LJ Celada A Buchmeier NA Gray PW Monoclonal antibodies to murine gamma-interferon which differentially modulate macrophage activation and antiviral activity J Immunol 1985 134 1609 1618 2578513
Bouley DM Kanangat S Wire W Rouse BT Characterization of herpes simplex virus type-1 infection and herpetic stromal keratitis development in IFN-gamma knockout mice J Immunol 1995 155 3964 3971 7561104
Durbin JE Hackenmiller R Simon MC Levy DE Targeted disruption of the mouse Stat1 gene results in compromised innate immunity to viral disease Cell 1996 84 443 450 8608598
Yu Z Manickan E Rouse BT Role of interferon-gamma in immunity to herpes simplex virus J Leukoc Biol 1996 60 528 532 8864138
Novelli F Casanova JL The role of IL-12, IL-23 and IFN-gamma in immunity to viruses Cytokine Growth Factor Rev 2004 15 367 377 15450252
Sheppard P Kindsvogel W Xu W Henderson K Schlutsmeyer S Whitmore TE Kuestner R Garrigues U Birks C Roraback J Ostrander C Dong D Shin J Presnell S Fox B Haldeman B Cooper E Taft D Gilbert T Grant FJ Tackett M Krivan W McKnight G Clegg C Foster D Klucher KM IL-28, IL-29 and their class II cytokine receptor IL-28R Nat Immunol 2003 4 63 68 12469119
Kotenko SV Gallagher G Baurin VV Lewis-Antes A Shen M Shah NK Langer JA Sheikh F Dickensheets H Donnelly RP IFN-lambdas mediate antiviral protection through a distinct class II cytokine receptor complex Nat Immunol 2003 4 69 77 12483210
Goetz FW Planas JV MacKenzie S Tumor necrosis factors Dev Comp Immunol 2004 28 487 497 15062645
Ware CF The TNF superfamily Cytokine Growth Factor Rev 2003 14 181 184 12787557
The new TNF nomenclature scheme 2005
Shurety W Merino-Trigo A Brown D Hume DA Stow JL Localization and post-Golgi trafficking of tumor necrosis factor-alpha in macrophages J Interferon Cytokine Res 2000 20 427 438 10805378
Blobel CP Metalloprotease-disintegrins: links to cell adhesion and cleavage of TNF alpha and Notch Cell 1997 90 589 592 9288739
Herbein G O'Brien WA Tumor necrosis factor (TNF)-alpha and TNF receptors in viral pathogenesis Proc Soc Exp Biol Med 2000 223 241 257 10719836
Paludan SR Ellermann-Eriksen S Kruys V Mogensen SC Expression of TNF-alpha by herpes simplex virus-infected macrophages is regulated by a dual mechanism: transcriptional regulation by NF-kappa B and activating transcription factor 2/Jun and translational regulation through the AU-rich region of the 3' untranslated region J Immunol 2001 167 2202 2208 11490006
Benedict CA Viruses and the TNF-related cytokines, an evolving battle Cytokine Growth Factor Rev 2003 14 349 357 12787571
Hatada EN Krappmann D Scheidereit C NF-kappaB and the innate immune response Curr Opin Immunol 2000 12 52 58 10679399
Moller B Ellermann-Eriksen S Storgaard M Obel N Bendtzen K Petersen CM Soluble tumor necrosis factor (TNF) receptors conserve TNF bioactivity in meningitis patient spinal fluid J Infect Dis 1996 174 557 563 8769614
Pfeffer K Biological functions of tumor necrosis factor cytokines and their receptors Cytokine Growth Factor Rev 2003 14 185 191 12787558
Kobayashi M Fitz L Ryan M Hewick RM Clark SC Chan S Loudon R Sherman F Perussia B Trinchieri G Identification and purification of natural killer cell stimulatory factor (NKSF), a cytokine with multiple biologic effects on human lymphocytes J Exp Med 1989 170 827 845 2504877
Trinchieri G Pflanz S Kastelein RA The IL-12 family of heterodimeric cytokines: new players in the regulation of T cell responses Immunity 2003 19 641 644 14614851
Watford WT Moriguchi M Morinobu A O'Shea JJ The biology of IL-12: coordinating innate and adaptive immune responses Cytokine Growth Factor Rev 2003 14 361 368 12948519
Gubler U Chua AO Schoenhaut DS Dwyer CM McComas W Motyka R Nabavi N Wolitzky AG Quinn PM Familletti PC Cately MK Coexpression of two distinct genes is required to generate secreted bioactive cytotoxic lymphocyte maturation factor Proc Natl Acad Sci U S A 1991 88 4143 4147 1674604
D'Andrea A Rengaraju M Valiante NM Chehimi J Kubin M Aste M Chan SH Kobayashi M Young D Nickbarg E Chizzonite R Wolf SF Trinchieri G Production of natural killer cell stimulatory factor (interleukin 12) by peripheral blood mononuclear cells J Exp Med 1992 176 1387 1398 1357073
Malmgaard L Paludan SR Mogensen SC Ellermann-Eriksen S Herpes simplex virus type 2 induces secretion of IL-12 by macrophages through a mechanism involving NF-kappaB J Gen Virol 2000 81 3011 3020 11086132
Reis e Sousa C Hieny S Scharton-Kersten T Jankovic D Charest H Germain RN Sher A In vivo microbial stimulation induces rapid CD40 ligand-independent production of interleukin 12 by dendritic cells and their redistribution to T cell areas J Exp Med 1997 186 1819 1829 9382881
Trinchieri G Interleukin-12 and the regulation of innate resistance and adaptive immunity Nat Rev Immunol 2003 3 133 146 12563297
Gerosa F Gobbi A Zorzi P Burg S Briere F Carra G Trinchieri G The reciprocal interaction of NK cells with plasmacytoid or myeloid dendritic cells profoundly affects innate resistance functions J Immunol 2005 174 727 734 15634892
Tato CM Martins GA High FA DiCioccio CB Reiner SL Hunter CA Cutting Edge: Innate production of IFN-gamma by NK cells is independent of epigenetic modification of the IFN-gamma promoter J Immunol 2004 173 1514 1517 15265878
Leist TP Heuchel R Zinkernagel RM Increased bactericidal macrophage activity induced by immunological stimuli is dependent on interferon (IFN)-gamma. Interference of anti-IFN-gamma but not anti-IFN-alpha/beta with modulation of macrophage activity caused by lymphocytic choriomeningitis virus infection or systemic graft-vs.-host reactions Eur J Immunol 1988 18 1295 1298 3138138
Varinou L Ramsauer K Karaghiosoff M Kolbe T Pfeffer K Muller M Decker T Phosphorylation of the Stat1 transactivation domain is required for full-fledged IFN-gamma-dependent innate immunity Immunity 2003 19 793 802 14670297
Munder M Mallo M Eichmann K Modolell M Murine macrophages secrete interferon gamma upon combined stimulation with interleukin (IL)-12 and IL-18: A novel pathway of autocrine macrophage activation J Exp Med 1998 187 2103 2108 9625771
Fukao T Frucht DM Yap G Gadina M O'Shea JJ Koyasu S Inducible expression of Stat4 in dendritic cells and macrophages and its critical role in innate and adaptive immune responses J Immunol 2001 166 4446 4455 11254700
Ma X Chow JM Gri G Carra G Gerosa F Wolf SF Dzialo R Trinchieri G The interleukin 12 p40 gene promoter is primed by interferon gamma in monocytic cells J Exp Med 1996 183 147 157 8551218
Pirhonen J Regulation of IL-18 expression in virus infection Scand J Immunol 2001 53 533 539 11422901
Matsumoto H Suzuki K Tsuyuguchi K Tanaka E Amitani R Maeda A Yamamoto K Sasada M Kuze F Interleukin-12 gene expression in human monocyte-derived macrophages stimulated with Mycobacterium bovis BCG: cytokine regulation and effect of NK cells Infect Immun 1997 65 4405 4410 9353012
D'Andrea A Aste-Amezaga M Valiante NM Ma X Kubin M Trinchieri G Interleukin 10 (IL-10) inhibits human lymphocyte interferon gamma-production by suppressing natural killer cell stimulatory factor / IL-12 synthesis in accessory cells J Exp Med 1993 178 1041 1048 8102388
Du C Sriram S Mechanism of inhibition of LPS-induced IL-12p40 production by IL-10 and TGF-beta in ANA-1 cells J Leukoc Biol 1998 64 92 97 9665281
D'Andrea A Ma X Aste-Amezaga M Paganin C Trinchieri G Stimulatory and inhibitory effects of interleukin (IL)-4 and IL-13 on the production of cytokines by human peripheral blood mononuclear cells: priming for IL-12 and tumor necrosis factor alpha production J Exp Med 1995 181 537 546 7836910
Marshall JD Robertson SE Trinchieri G Chehimi J Priming with IL-4 and IL-13 during HIV-1 infection restores in vitro IL-12 production by mononuclear cells of HIV-infected patients J Immunol 1997 159 5705 5714 9548515
Kim S Elkon KB Ma X Transcriptional suppression of interleukin-12 gene expression following phagocytosis of apoptotic cells Immunity 2004 21 643 653 15539151
Orange JS Salazar-Mather TP Opal SM Spencer RL Miller AH McEwen BS Biron CA Mechanism of interleukin 12-mediated toxicities during experimental viral infections: role of tumor necrosis factor and glucocorticoids J Exp Med 1995 181 901 914 7869050
Biron CA Orange JS IL-12 in acute viral infectious disease Res Immunol 1995 146 590 600 8839166
Heinzel FP Hujer AM Ahmed FN Rerko RM In vivo production and function of IL-12 p40 homodimers J Immunol 1997 158 4381 4388 9127002
Carra G Gerosa F Trinchieri G Biosynthesis and posttranslational regulation of human IL-12 J Immunol 2000 164 4752 4761 10779781
Oppmann B Lesley R Blom B Timans JC Xu Y Hunte B Vega F Yu N Wang J Singh K Zonin F Vaisberg E Churakova T Liu M Gorman D Wagner J Zurawski S Liu Y Abrams JS Moore KW Rennick D Waal-Malefyt R Hannum C Bazan JF Kastelein RA Novel p19 protein engages IL-12p40 to form a cytokine, IL-23, with biological activities similar as well as distinct from IL-12 Immunity 2000 13 715 725 11114383
Broberg EK Setala N Eralinna JP Salmi AA Roytta M Hukkanen V Herpes simplex virus type 1 infection induces upregulation of interleukin-23 (p19) mRNA expression in trigeminal ganglia of BALB/c mice J Interferon Cytokine Res 2002 22 641 651 12162874
Langrish CL McKenzie BS Wilson NJ de Waal MR Kastelein RA Cua DJ IL-12 and IL-23: master regulators of innate and adaptive immunity Immunol Rev 2004 202 96 105 15546388
Pflanz S Timans JC Cheung J Rosales R Kanzler H Gilbert J Hibbert L Churakova T Travis M Vaisberg E Blumenschein WM Mattson JD Wagner JL To W Zurawski S McClanahan TK Gorman DM Bazan JF de Waal MR Rennick D Kastelein RA IL-27, a heterodimeric cytokine composed of EBI3 and p28 protein, induces proliferation of naive CD4(+) T cells Immunity 2002 16 779 790 12121660
Schijns VE Haagmans BL Wierda CM Kruithof B Heijnen IA Alber G Horzinek MC Mice lacking IL-12 develop polarized Th1 cells during viral infection J Immunol 1998 160 3958 3964 9558103
Puccetti P Belladonna ML Grohmann U Effects of IL-12 and IL-23 on antigen-presenting cells at the interface between innate and adaptive immunity Crit Rev Immunol 2002 22 373 390 12803317
Abdi K IL-12: the role of p40 versus p75 Scand J Immunol 2002 56 1 11 12100467
Brombacher F Kastelein RA Alber G Novel IL-12 family members shed light on the orchestration of Th1 responses Trends Immunol 2003 24 207 212 12697453
Ohteki T Critical role for IL-15 in innate immunity Curr Mol Med 2002 2 371 380 12108948
Gill N Rosenthal KL Ashkar AA NK and NKT cell-independent contribution of interleukin-15 to innate protection against mucosal viral infection J Virol 2005 79 4470 4478 15767447
Ha SJ Kim DJ Baek KH Yun YD Sung YC IL-23 induces stronger sustained CTL and Th1 immune responses than IL-12 in hepatitis C virus envelope protein 2 DNA immunization J Immunol 2004 172 525 531 14688363
Zhou Y Lin G Baarsch MJ Scamurra RW Murtaugh MP Interleukin-4 suppresses inflammatory cytokine gene transcription in porcine macrophages J Leukoc Biol 1994 56 507 513 7930948
Zhou Y Lin G Murtaugh MP Interleukin-4 suppresses the expression of macrophage NADPH oxidase heavy chain subunit (gp91-phox) Biochim Biophys Acta 1995 1265 40 48 7857983
Bogdan C Vodovotz Y Paik J Xie QW Nathan C Mechanism of suppression of nitric oxide synthase expression by interleukin-4 in primary mouse macrophages J Leukoc Biol 1994 55 227 233 7507968
Paludan SR Lovmand J Ellermann-Eriksen S Mogensen SC Effect of IL-4 and IL-13 on IFN-gamma-induced production of nitric oxide in mouse macrophages infected with herpes simplex virus type 2 FEBS Lett 1997 414 61 64 9305732
Kurt-Jones EA Hamberg S Ohara J Paul WE Abbas AK Heterogeneity of helper/inducer T lymphocytes. I. Lymphokine production and lymphokine responsiveness J Exp Med 1987 166 1774 1787 2960773
Seder RA Paul WE Dvorak AM Sharkis SJ Kagey-Sobotka A Niv Y Finkelman FD Barbieri SA Galli SJ Plaut M Mouse splenic and bone marrow cell populations that express high-affinity Fc epsilon receptors and produce interleukin 4 are highly enriched in basophils Proc Natl Acad Sci U S A 1991 88 2835 2839 1826367
Gordon S Alternative activation of macrophages Nat Rev Immunol 2003 3 23 35 12511873
Hart PH Bonder CS Balogh J Dickensheets HL Donnelly RP Finlay-Jones JJ Differential responses of human monocytes and macrophages to IL-4 and IL-13 J Leukoc Biol 1999 66 575 578 10534111
Howard M Farrar J Hilfiker M Johnson B Takatsu K Hamaoka T Paul WE Identification of a T cell-derived B cell growth factor distinct from interleukin 2 J Exp Med 1982 155 914 923 6977612
Ghiasi H Osorio Y Perng GC Nesburn AB Wechsler SL Recombinant herpes simplex virus type 1 expressing murine interleukin-4 is less virulent than wild-type virus in mice J Virol 2001 75 9029 9036 11533166
El Ghorr AA Norval M The effect of UV-B irradiation on primary and secondary HSV-1 infections in interleukin-4 knockout mice Arch Dermatol Res 1999 291 459 465 10482018
Mogensen SC Role of macrophages in natural resistance to virus infections Microbiol Rev 1979 43 1 26 379574
Lopez C Resistance to herpes simplex virus - type 1 (HSV-1) Curr Top Microbiol Immunol 1981 92 15 24 6171384
Kirchner H Immunobiology of infection with herpes simplex virus Monographs in Virology Monographs in Virology Melnick JL 1982 1 104
Mogensen SC Role of macrophages in hepatitis induced by herpes simplex virus type 1 and 2 in mice Infect Immun 1977 15 686 691 192671
Halford WP Balliet JW Gebhardt BM Re-evaluating natural resistance to herpes simplex virus type 1 J Virol 2004 78 10086 10095 15331741
Mogensen SC Macrophages and age-dependent resistance to hepatitis induced by herpes simplex virus type 2 in mice Infect Immun 1978 19 46 50 203536
Zawatzky R Engler H Kirchner H Experimental infection of inbred mice with herpes simplex virus. III. Comparison between newborn and adult C57BL/6 mice J Gen Virol 1982 60 25 29 6178800
Lopez C Genetics of natural resistance to herpesvirus infections in mice Nature 1975 258 152 153 171586
Mogensen SC Genetics of macrophage-controled resistance to hepatitis induced by herpes simplex virus type 2 in mice Infect Immun 1977 17 268 273 197015
Zawatzky R Kirchner H DeMaeyer-Guignard J DeMaeyer E An X-linked locus influences the amount of circulating interferon induced in the mouse by herpes simplex virus type 1 J Gen Virol 1982 63 325 332 6185626
Pedersen EB Haahr S Mogensen SC X-linked resistance of mice to high doses of herpes simplex virus type 2 correlates with early interferon production Infect Immun 1983 42 740 746 6196293
Gresser I Tovey MG Maury C Bandu MT Role of interferon in the pathogenesis of virus diseases in mice as demonstrated by the use of anti-interferon serum J Exp Med 1976 144 1316 1323 186555
Ellermann-Eriksen S Christensen MM Mogensen SC Effect of mercuric chloride on macrophage-mediated resistance mechanisms against infection with herpes simplex virus type 2 Toxicology 1994 93 269 287 7526487
Christensen MM Ellermann-Eriksen S Rungby J Mogensen SC Influence of mercuric chloride on resistance to generalized infection with herpes simplex virus type 2 in mice Toxicology 1996 114 57 66 8931761
Halford WP Veress LA Gebhardt BM Carr DJ Innate and acquired immunity to herpes simplex virus type 1 Virol 1997 236 328 337
Noisakran S Campbell IL Carr DJ Ectopic expression of DNA encoding IFN-alpha 1 in the cornea protects mice from herpes simplex virus type 1-induced encephalitis J Immunol 1999 162 4184 4190 10201945
Leib DA Harrison TE Laslo KM Machalek MA Moorman NJ Virgin HW Interferons regulate the phenotype of wild-type and mutant herpes simplex viruses in vivo J Exp Med 1999 189 663 672 9989981
Simmons A La-Vista AB Neural infection in mice after cutaneous inoculation with HSV-1 is under complex host genetic control Virus Res 1989 13 263 270 2549748
Rong Q Alexander TS Koski GK Rosenthal KS Multiple mechanisms for HSV-1 induction of interferon alpha production by peripheral blood mononuclear cells Arch Virol 2003 148 329 344 12556996
Ellermann-Eriksen S Autocrine secretion of interferon-alpha/beta and tumour necrosis factor-alpha synergistically activates mouse macrophages after infection with herpes simplex virus type 2 J Gen Virol 1993 74 2191 2199 8409942
Kirchner H Engler H Schroder CH Zawatzky R Storch E Herpes Simplex virus type 1-induced interferon production and activation of natural killer cells in mice J Gen Virol 1983 64 437 441 6300294
Riffault S Carrat C Milon G Charley B Colle JH Transient IFN-gamma synthesis in the lymph node draining a dermal site loaded with UV-irradiated herpes simplex virus type 1: an NK- and CD3- dependent process regulated by IL-12 but not by IFN-alpha/beta J Gen Virol 2000 81 2365 2373 10993924
Lebon P Inhibition of herpes simplex virus type 1-induced interferon synthesis by monoclonal antibodies against viral glycoprotein D and by lysosomotropic drugs J Gen Virol 1985 66 2781 2786 2999320
Malmgaard L Melchjorsen J Bowie AG Mogensen SC Paludan SR Viral activation of macrophages through TLR-dependent and -independent pathways J Immunol 2004 173 6890 6898 15557184
Milone MC Fitzgerald-Bocarsly P The mannose receptor mediates induction of IFN-alpha in peripheral blood dendritic cells by enveloped RNA and DNA viruses J Immunol 1998 161 2391 2399 9725235
Mogensen TH Paludan SR Reading the viral signature by Toll-like receptors and other pattern recognition receptors J Mol Med 2005 83 180 192 15635478
Lund J Sato A Akira S Medzhitov R Iwasaki A Toll-like receptor 9-mediated recognition of Herpes simplex virus-2 by plasmacytoid dendritic cells J Exp Med 2003 198 513 520 12900525
Hochrein H Schlatter B O'Keeffe M Wagner C Schmitz F Schiemann M Bauer S Suter M Wagner H Herpes simplex virus type-1 induces IFN-alpha production via Toll-like receptor 9-dependent and -independent pathways Proc Natl Acad Sci U S A 2004 101 11416 11421 15272082
Krug A Luker GD Barchet W Leib DA Akira S Colonna M Herpes simplex virus type 1 activates murine natural interferon-producing cells through toll-like receptor 9 Blood 2004 103 1433 1437 14563635
Diebold SS Montoya M Unger H Alexopoulou L Roy P Haswell LE Al Shamkhani A Flavell R Borrow P Sousa R Viral infection switches non-plasmacytoid dendritic cells into high interferon producers Nature 2003 424 324 328 12819664
Yoneyama M Kikuchi M Natsukawa T Shinobu N Imaizumi T Miyagishi M Taira K Akira S Fujita T The RNA helicase RIG-I has an essential function in double-stranded RNA-induced innate antiviral responses Nat Immunol 2004 5 730 737 15208624
Du X Poltorak A Wei Y Beutler B Three novel mammalian toll-like receptors: gene structure, expression, and evolution Eur Cytokine Netw 2000 11 362 371 11022119
Heil F Hemmi H Hochrein H Ampenberger F Kirschning C Akira S Lipford G Wagner H Bauer S Species-specific recognition of single-stranded RNA via toll-like receptor 7 and 8 Science 2004 303 1526 1529 14976262
Takeda K Akira S Toll-like receptors in innate immunity Int Immunol 2005 17 1 14 15585605
Sato M Hata N Asagiri M Nakaya T Taniguchi T Tanaka N Positive feedback regulation of type I IFN genes by the IFN-inducible transcription factor IRF-7 FEBS Lett 1998 441 106 110 9877175
Dai J Megjugorac NJ Amrute SB Fitzgerald-Bocarsly P Regulation of IFN regulatory factor-7 and IFN-alpha production by enveloped virus and lipopolysaccharide in human plasmacytoid dendritic cells J Immunol 2004 173 1535 1548 15265881
Steward WE Gosser LB Lockart RZJ The effect of priming with interferon on interferon production by two lines of L cells J Gen Virol 1972 15 85 87 4337220
Belardelli F Gessani S Proietti E Locardi C Borghi P Watanabe Y Kawade Y Gresser I Studies on expression of spontaneous and induced interferons in mouse peritoneal macrophages by means of monoclonal antibodies to mouse interferons J Gen Virol 1987 68 2203-2212 2203-2212 3039046
Gresser I Vignaux F Belardelli F Tovey MG Maunoury MT Injection of mice with antibody to mouse interferon alpha/beta decreases the level of 2'-5' oligoadenylate synthetase in peritoneal macrophages J Virol 1985 53 221 227 2981340
Proietti E Vanden-Broecke C Di-Marzio P Gessani S Gresser I Tovey MG Specific interferon genes are expressed in individual cells in the peritoneum and bone marrow of normal mice J Interferon Res 1992 12 27 34 1374109
Nickolaus P Zawatzky R Inhibition by interleukin-4 of constitutive beta interferon synthesis in mouse macrophages J Virol 1994 68 6763 6766 8084010
Ellermann-Eriksen S Justesen J Mogensen SC Genetically determined difference in the antiviral action of alpha/beta interferon in cells from mice resistant or susceptible to herpes simplex virus type 2 J Gen Virol 1986 67 1859 1866 3018128
Wells CA Ravasi T Faulkner GJ Carninci P Okazaki Y Hayashizaki Y Sweet M Wainwright BJ Hume DA Genetic control of the innate immune response BMC Immunol 2003 4 5 12826024
Hayden MS Ghosh S Signaling to NF-kappaB Genes Dev 2004 18 2195 2224 15371334
Paludan SR Ellermann-Eriksen S Mogensen SC NF-kappaB activation is responsible for the synergistic effect of herpes simplex virus type 2 infection on interferon-gamma-induced nitric oxide production in macrophages J Gen Virol 1998 79 2785 2793 9820155
Paludan SR Mogensen SC Virus-cell interactions regulating induction of tumor necrosis factor alpha production in macrophages infected with herpes simplex virus J Virol 2001 75 10170 10178 11581385
Reis LFL Lee TH Vilcek J Tumor necrosis factor acts synergistically with autocrine interferon-beta and increases interferon-beta mRNA levels in human fibroblasts J Biol Chem 1989 264 16351 16354 2550437
Thrasher AJ Jones GE Kinnon C Brickell PM Katz DR Is Wiskott-Aldrich syndrome a cell trafficking disorder? Immunol Today 1998 19 537 539 9864941
Orange JS Stone KD Turvey SE Krzewski K The Wiskott-Aldrich syndrome Cell Mol Life Sci 2004 61 2361 2385 15378206
Messina C Kirkpatrick D Fitzgerald PA O'Reilly RJ Siegal FP Cunningham-Rundles C Blaese M Oleske J Pahwa S Lopez C Natural killer cell function and interferon generation in patients with primary immunodeficiencies Clin Immunol Immunopathol 1986 39 394 404 3698344
Fontana S Moratto D Mangal S De Francesco M Vermi W Ferrari S Facchetti F Kutukculer N Fiorini C Duse M Das PK Notarangelo LD Plebani A Badolato R Functional defects of dendritic cells in patients with CD40 deficiency Blood 2003 102 4099 4106 12893749
O'Sullivan B Thomas R CD40 and dendritic cell function Crit Rev Immunol 2003 23 83 107 12906261
Al-Khatib K Williams BR Silverman RH Halford W Carr DJ Distinctive roles for 2',5'-oligoadenylate synthetases and double-stranded RNA-dependent protein kinase R in the in vivo antiviral effect of an adenoviral vector expressing murine IFN-beta J Immunol 2004 172 5638 5647 15100308
Chee AV Lopez P Pandolfi PP Roizman B Promyelocytic leukemia protein mediates interferon-based anti-herpes simplex virus 1 effects J Virol 2003 77 7101 7105 12768029
Dandoy F DeMaeyer-Guignard J Bailey D DeMaeyer E Mouse genes influence antiviral action of interferon in vivo Infect Immun 1982 38 89 93 6183213
Ellermann-Eriksen S Sommerlund M Mogensen SC Differential sensitivity of macrophages from herpes simplex virus-resistant and -susceptible mice to respiratory burst priming by interferon-alpha/beta J Gen Virol 1989 70 2139 2147 2769233
Mestan J Brockhaus M Kirchner H Jacobsen H Antiviral activity of tumour necrosis factor. Synergism with interferons and induction of oligo-2',5'-adenylate synthetase J Gen Virol 1988 69 3113 3120 2462015
Harle P Cull V Agbaga MP Silverman R Williams BR James C Carr DJ Differential effect of murine alpha/beta interferon transgenes on antagonization of herpes simplex virus type 1 replication J Virol 2002 76 6558 6567 12050368
Galabru J Robert N Buffet-Janvresse C Riviere Y Hovanessian AG Continuous production of interferon in normal mice: effect of anti-interferon globulin, sex, age, strain and environment on the levels of 2-5A synthetase and p67K kinase J Gen Virol 1985 66 711 718 2580048
Suzutani T Nagamine M Shibaki T Ogasawara M Yoshida I Daikoku T Nishiyama Y Azuma M The role of the UL41 gene of herpes simplex virus type 1 in evasion of non-specific host defence mechanisms during primary infection J Gen Virol 2000 81 1763 1771 10859382
Yokota S Yokosawa N Okabayashi T Suzutani T Miura S Jimbow K Fujii N Induction of suppressor of cytokine signaling-3 by herpes simplex virus type 1 contributes to inhibition of the interferon signaling pathway J Virol 2004 78 6282 6286 15163721
Chee AV Roizman B Herpes simplex virus 1 gene products occlude the interferon signaling pathway at multiple sites J Virol 2004 78 4185 4196 15047834
Mogensen TH Melchjorsen J Malmgaard L Casola A Paludan SR Suppression of proinflammatory cytokine expression by herpes simplex virus type 1 J Virol 2004 78 5883 5890 15140986
Mogensen SC Virelizier JL The interferon-macrophage alliance Interferon 1987 8 55 84 2445689
Wu L Morahan PS Macrophages and other nonspecific defences: Role of modulating resistance against herpes simplex virus Curr Top Microbiol Immunol Herpes simplex virus - pathogenesis, immunobiology and control Rouse BT 1992 89 110 1499352
Morahan PS Morse SS McGeorge MG Macrophage extrinsic antiviral activity during herpes simplex virus infection J Gen Virol 1980 46 291 300 6247426
Domke-Opitz I Poberschin P Mittnacht S Kirchner H Role of interferon in persistent infection of macrophages with herpes simplex virus Virol 1987 159 306 311
Skulachev VP Possible role of reactive oxygen species in antiviral defense Biochemistry (Mosc ) 1998 63 1438 1440 9916163
Roos D van Bruggen R Meischl C Oxidative killing of microbes by neutrophils Microbes Infect 2003 5 1307 1315 14613774
Bogdan C Rollinghoff M Diefenbach A Reactive oxygen and reactive nitrogen intermediates in innate and specific immunity Curr Opin Immunol 2000 12 64 76 10679404
Garg AK Aggarwal BB Reactive oxygen intermediates in TNF signaling Mol Immunol 2002 39 509 517 12431383
Boonstra J Post JA Molecular events associated with reactive oxygen species and cell cycle progression in mammalian cells Gene 2004 337 1 13 15276197
Mogensen SC Ellermann-Eriksen S Sommerlund M Herpes simplex virus type 2 primes mouse macrophages for an early and genetically determined respiratory burst mediated by interferon-alpha/beta J Gen Virol 1989 70 1371 1379 2543784
Bellner L Thoren F Nygren E Liljeqvist JA Karlsson A Eriksson K A proinflammatory peptide from herpes simplex virus type 2 glycoprotein G affects neutrophil, monocyte, and NK cell functions J Immunol 2005 174 2235 2241 15699157
Gallien-Lartigue O Carrez D DeMaeyer E Strain dependence of the antiproliferative action of interferon on murine erythroid precursors Science 1980 209 292 293 6155700
Valyi-Nagy T Olson SJ Valyi-Nagy K Montine TJ Dermody TS Herpes simplex virus type 1 latency in the murine nervous system is associated with oxidative damage to neurons Virol 2000 278 309 321
Lidbury BA Ramshaw IA Rolph MS Cowden WB The antiviral activity of tumour necrosis factor on herpes simplex virus type 1: role for a butylated hydroxyanisole sensitive factor Arch Virol 1995 140 703 719 7794113
Galvan V Roizman B Herpes simplex virus 1 induces and blocks apoptosis at multiple steps during infection and protects cells from exogenous inducers in a cell-type-dependent manner Proc Natl Acad Sci U S A 1998 95 3931 3936 9520470
Irie H Kiyoshi A Koyama AH A role for apoptosis induced by acute herpes simplex virus infection in mice Int Rev Immunol 2004 23 173 185 14690859
Bukowski JF Welsh RM The role of natural killer cells and interferon in resistance to acute infection of mice with herpes simplex virus type 1 J Immunol 1986 136 3481 3485 2420892
Vollstedt S O'Keeffe M Odermatt B Beat R Glanzmann B Riesen M Shortman K Suter M Treatment of neonatal mice with Flt3 ligand leads to changes in dendritic cell subpopulations associated with enhanced IL-12 and IFN-alpha production Eur J Immunol 2004 34 1849 1860 15214033
Vollstedt S Franchini M Hefti HP Odermatt B O'Keeffe M Alber G Glanzmann B Riesen M Ackermann M Suter M Flt3 ligand-treated neonatal mice have increased innate immunity against intracellular pathogens and efficiently control virus infections J Exp Med 2003 197 575 584 12615899
Vollstedt S Arnold S Schwerdel C Franchini M Alber G Di Santo JP Ackermann M Suter M Interplay between alpha/beta and gamma interferons with B, T, and natural killer cells in the defense against herpes simplex virus type 1 J Virol 2004 78 3846 3850 15047800
Noisakran S Carr DJ Type I interferons and herpes simplex virus infection: a naked DNA approach as a therapeutic option? Immunol Res 2001 24 1 11 11485206
Kanangat S Thomas J Gangappa S Babu JS Rouse BT Herpes simplex virus type 1-mediated up-regulation of IL-12 (p40) mRNA expression. Implications in immunopathogenesis and protection J Immunol 1996 156 1110 1116 8557986
Coutelier JP Van Broeck J Wolf SF Interleukin-12 gene expression after viral infection in the mouse J Virol 1995 69 1955 1958 7853541
Kumaraguru U Rouse BT The IL-12 response to herpes simplex virus is mainly a paracrine response of reactive inflammatory cells J Leukoc Biol 2002 72 564 570 12223525
Ma X Sun J Papasavvas E Riemann H Robertson S Marshall J Bailer RT Moore A Donnelly RP Trinchieri G Montaner LJ Inhibition of IL-12 production in human monocyte-derived macrophages by TNF J Immunol 2000 164 1722 1729 10657616
Hodge-Dufour J Marino MW Horton MR Jungbluth A Burdick MD Strieter RM Noble PW Hunter CA Pure E Inhibition of interferon gamma induced interleukin 12 production: a potential mechanism for the anti-inflammatory activities of tumor necrosis factor Proc Natl Acad Sci U S A 1998 95 13806 13811 9811882
Cousens LP Orange JS Su HC Biron CA Interferon-alpha/beta inhibition of interleukin 12 and interferon-gamma production in vitro and endogenously during viral infection Proc Natl Acad Sci U S A 1997 94 634 639 9012836
Byrnes AA Ma X Cuomo P Park K Wahl L Wolf SF Zhou H Trinchieri G Karp CL Type I interferons and IL-12: convergence and cross-regulation among mediators of cellular immunity Eur J Immunol 2001 31 2026 2034 11449355
Dalod M Salazar-Mather TP Malmgaard L Lewis C Asselin-Paturel C Briere F Trinchieri G Biron CA Interferon alpha/beta and interleukin 12 responses to viral infections: pathways regulating dendritic cell cytokine expression in vivo J Exp Med 2002 195 517 528 11854364
Salkowski CA Kopydlowski K Blanco J Cody MJ McNally R Vogel SN IL-12 is dysregulated in macrophages from IRF-1 and IRF-2 knockout mice J Immunol 1999 163 1529 1536 10415056
Masumi A Tamaoki S Wang IM Ozato K Komuro K IRF-8/ICSBP and IRF-1 cooperatively stimulate mouse IL-12 promoter activity in macrophages FEBS Lett 2002 531 348 353 12417340
Kirchner H Zawatzky R Hirt HM In vitro production of immune interferon by spleen cells of mice immunized with herpes simplex virus Cell Immunol 1978 40 204 210 212209
Hunter CA Gabriel KE Radzanowski T Neyer LE Remington JS Type I interferons enhance production of IFN-gamma by NK cells Immunol Lett 1997 59 1 5 9334850
Malmgaard L Paludan SR Interferon (IFN)-alpha/beta, interleukin (IL)-12 and IL-18 coordinately induce production of IFN-gamma during infection with herpes simplex virus type 2 J Gen Virol 2003 84 2497 2500 12917471
Jewett A Gan XH Lebow LT Bonavida B Differential secretion of TNF-alpha and IFN-gamma by human peripheral blood-derived NK subsets and association with functional maturation J Clin Immunol 1996 16 46 54 8926285
Rogge L D'Ambrosio D Biffi M Penna G Minetti LJ Presky DH Adorini L Sinigaglia F The role of Stat4 in species-specific regulation of Th cell development by type I IFNs J Immunol 1998 161 6567 6574 9862683
Tripp CS Wolf SF Unanue ER Interleukin 12 and tumor necrosis factor alpha are costimulators of interferon gamma production by natural killer cells in severe combined immunodeficiency mice with listeriosis, and interleukin 10 is a physiologic antagonist Proc Natl Acad Sci U S A 1993 90 3725 3729 8097322
Shibuya K Robinson D Zonin F Hartley SB Macatonia SE Somoza C Hunter CA Murphy KM O'Garra A IL-1 alpha and TNF-alpha are required for IL-12-induced development of Th1 cells producing high levels of IFN-gamma in BALB/c but not C57BL/6 mice J Immunol 1998 160 1708 1716 9469428
Domke I Straub P Jacobsen H Kirchner H Panet A Inhibition of replication of herpes simplex virus in mouse macrophages by interferon J Gen Virol 1985 66 2231 2236 2413165
Sainz BJ Halford WP Alpha/Beta interferon and gamma interferon synergize to inhibit the replication of herpes simplex virus type 1 J Virol 2002 76 11541 11550 12388715
Halford WP Halford KJ Pierce AT Mathematical analysis demonstrates that interferons-beta and -gamma interact in a multiplicative manner to disrupt herpes simplex virus replication J Theor Biol 2005 234 439 454 15784277
Domke-Opitz I Straub P Kirchner H Effect of interferon on replication of herpes simplex virus types 1 and 2 in human macrophages J Virol 1986 60 37 42 3018299
Feduchi E Carrasco L Mechanism of inhibition of HSV-1 replication by tumor necrosis factor and interferon gamma Virol 1991 180 822 825
Chen SH Oakes JE Lausch RN Synergistic anti-HSV effect of tumor necrosis factor alpha and interferon gamma in human corneal fibroblasts is associated with interferon beta induction Antiviral Res 1993 22 15 29 8250541
Chen SH Oakes JE Lausch RN Synergistic anti-herpes effect of TNF-alpha and IFN-gamma in human corneal epithelial cells compared with that in corneal fibroblasts Antiviral Res 1994 25 201 213 7710269
Adams O Besken K Oberdorfer C MacKenzie CR Russing D Daubener W Inhibition of human herpes simplex virus type 2 by interferon gamma and tumor necrosis factor alpha is mediated by indoleamine 2,3-dioxygenase Microbes Infect 2004 6 806 812 15374002
Harandi AM Svennerholm B Holmgren J Eriksson K Interleukin-12 (IL-12) and IL-18 are important in innate defense against genital herpes simplex virus type 2 infection in mice but are not required for the development of acquired gamma interferon-mediated protective immunity J Virol 2001 75 6705 6709 11413339
Vollstedt S Franchini M Alber G Ackermann M Suter M Interleukin-12- and gamma interferon-dependent innate immunity are essential and sufficient for long-term survival of passively immunized mice infected with herpes simplex virus type 1 J Virol 2001 75 9596 9600 11559791
Gherardi MM Ramirez JC Esteban M IL-12 and IL-18 act in synergy to clear vaccinia virus infection: involvement of innate and adaptive components of the immune system J Gen Virol 2003 84 1961 1972 12867626
Osorio Y Wechsler SL Nesburn AB Ghiasi H Reduced severity of HSV-1-induced corneal scarring in IL-12-deficient mice Virus Res 2002 90 317 326 12457985
Lee S Zheng M Deshpande S Eo SK Hamilton TA Rouse BT IL-12 suppresses the expression of ocular immunoinflammatory lesions by effects on angiogenesis J Leukoc Biol 2002 71 469 476 11867684
Kobayashi H Kobayashi M Utsunomiya T Herndon DN Pollard RB Suzuki F Therapeutic protective effects of IL-12 combined with soluble IL-4 receptor against established infections of herpes simplex virus type 1 in thermally injured mice J Immunol 1999 162 7148 7154 10358160
Orange JS Wang B Terhorst C Biron CA Requirement for natural killer cell-produced interferon gamma in defense against murine cytomegalovirus infection and enhancement of this defense pathway by interleukin 12 administration J Exp Med 1995 182 1045 1056 7561678
Orange JS Biron CA An absolute and restricted requirement for IL-12 in natural killer cell IFN-gamma production and antiviral defense. Studies of natural killer and T cell responses in contrasting viral infections J Immunol 1996 156 1138 1142 8557990
Ashkar AA Rosenthal KL Interleukin-15 and natural killer and NKT cells play a critical role in innate protection against genital herpes simplex virus type 2 infection J Virol 2003 77 10168 10171 12941930
Adler H Beland JL Del Pan NC Kobzik L Sobel RA Rimm IJ In the absence of T cells, natural killer cells protect from mortality due to HSV-1 encephalitis J Neuroimmunol 1999 93 208 213 10378885
Svensson A Nordstrom I Sun JB Eriksson K Protective immunity to genital herpes simpex virus type 2 infection is mediated by T-bet J Immunol 2005 174 6266 6273 15879125
Ghiasi H Cai S Perng GC Nesburn AB Wechsler SL The role of natural killer cells in protection of mice against death and corneal scarring following ocular HSV-1 infection Antiviral Res 2000 45 33 45 10774588
He J Ichimura H Iida T Minami M Kobayashi K Kita M Sotozono C Tagawa YI Iwakura Y Imanishi J Kinetics of cytokine production in the cornea and trigeminal ganglion of C57BL/6 mice after corneal HSV-1 infection J Interferon Cytokine Res 1999 19 609 615 10433361
Orange JS Human natural killer cell deficiencies and susceptibility to infection Microbes Infect 2002 4 1545 1558 12505527
Nathan C Nitric oxide as a secretory product of mammalian cells FASEB J 1992 6 3051 3064 1381691
Baskin H Ellermann-Eriksen S Lovmand J Mogensen SC Herpes simplex virus type 2 synergizes with interferon-gamma in the induction of nitric oxide production in mouse macrophages through autocrine secretion of tumour necrosis factor-alpha J Gen Virol 1997 78 195 203 9010304
Daheshia M Kanangat S Rouse BT Production of key molecules by ocular neutrophils early after herpetic infection of the cornea Exp Eye Res 1998 67 619 624 9990326
Diefenbach A Schindler H Rollinghoff M Yokoyama WM Bogdan C Requirement for type 2 NO synthase for IL-12 signaling in innate immunity Science 1999 284 951 955 10320373
Minagawa H Hashimoto K Yanagi Y Absence of tumour necrosis factor facilitates primary and recurrent herpes simplex virus-1 infections J Gen Virol 2004 85 343 347 14769891
Kodukula P Liu T Rooijen NV Jager MJ Hendricks RL Macrophage control of herpes simplex virus type 1 replication in the peripheral nervous system J Immunol 1999 162 2895 2905 10072539
Benencia F Gamba G Cavalieri H Courreges MC Benedetti R Villamil SM Massouh EJ Nitric oxide and HSV vaginal infection in BALB/c mice Virol 2003 309 75 84
Paludan SR Malmgaard L Ellermann-Eriksen S Bosca L Mogensen SC Interferon (IFN)-gamma and Herpes simplex virus/tumor necrosis factor- alpha synergistically induce nitric oxide synthase 2 in macrophages through cooperative action of nuclear factor-kappa B and IFN regulatory factor-1 Eur Cytokine Netw 2001 12 297 308 11399519
Kamijo R Harada H Matsuyama T Bosland M Gerecitano J Shapiro D Le J Koh SI Kimura T Green SJ Mak TW Taniguchi T Vilcek J Requirement for transcription factor IRF-1 in NO synthase induction in macrophages Science 1994 263 1612 1615 7510419
Neish AS Read MA Thanos D Pine R Maniatis T Collins T Endothelial interferon regulatory factor 1 cooperates with NF-kappa B as a transcriptional activator of vascular cell adhesion molecule 1 Mol Cell Biol 1995 15 2558 2569 7537851
Drew PD Franzoso G Becker KG Bours V Carlson LM Siebenlist U Ozato K NF kappa B and interferon regulatory factor 1 physically interact and synergistically induce major histocompatibility class I gene expression J Interferon Cytokine Res 1995 15 1037 1045 8746784
Drew PD Franzoso G Carlson LM Biddison WE Siebenlist U Ozato K Interferon regulatory factor-2 physically interacts with NF-kappa B in vitro and inhibits NF-kappa B induction of major histocompatibility class I and beta 2-microglobulin gene expression in transfected human neuroblastoma cells J Neuroimmunol 1995 63 157 162 8550813
Xiong H Zhu C Li H Chen F Mayer L Ozato K Unkeless JC Plevy SE Complex formation of the interferon (IFN) consensus sequence-binding protein with IRF-1 is essential for murine macrophage IFN-gamma-induced iNOS gene expression J Biol Chem 2003 278 2271 2277 12429737
Reiss CS Komatsu T Does nitric oxide play a critical role in viral infections? J Virol 1998 72 4547 4551 9573217
Karupiah G Xie QW Buller RM Nathan C Duarte C MacMicking JD Inhibition of viral replication by interferon-gamma-induced nitric oxide synthase Science 1993 261 1445 1448 7690156
Komatsu T Bi Z Reiss CS Interferon-gamma induced type I nitric oxide synthase activity inhibits viral replication in neurons J Neuroimmunol 1996 68 101 108 8784266
Karupiah G Harris N Inhibition of viral replication by nitric oxide and its reversal by ferrous sulfate and tricarboxylic acid cycle metabolites J Exp Med 1995 181 2171 2179 7539042
Benencia F Courreges MC Nitric oxide and macrophage antiviral extrinsic activity Immunol 1999 98 363 370
Croen KD Evidence for antiviral effect of nitric oxide. Inhibition of herpes simplex virus type 1 replication J Clin Invest 1993 91 2446 2452 8390481
Gamba G Cavalieri H Courreges MC Massouh EJ Benencia F Early inhibition of nitric oxide production increases HSV-1 intranasal infection J Med Virol 2004 73 313 322 15122810
Noda S Tanaka K Sawamura S Sasaki M Matsumoto T Mikami K Aiba Y Hasegawa H Kawabe N Koga Y Role of nitric oxide synthase type 2 in acute infection with murine cytomegalovirus J Immunol 2001 166 3533 3541 11207313
MacLean A Wei XQ Huang FP Al Alem UA Chan WL Liew FY Mice lacking inducible nitric-oxide synthase are more susceptible to herpes simplex virus infection despite enhanced Th1 cell responses J Gen Virol 1998 79 825 830 9568978
Shiloh MU MacMicking JD Nicholson S Brause JE Potter S Marino M Fang F Dinauer M Nathan C Phenotype of mice and macrophages deficient in both phagocyte oxidase and inducible nitric oxide synthase Immunity 1999 10 29 38 10023768
Henry Y Lepoivre M Drapier JC Ducrocq C Boucher JL Guissani A EPR characterization of molecular targets for NO in mammalian cells and organelles FASEB J 1993 7 1124 1134 8397130
Szabo C Zingarelli B O'Connor M Salzman AL DNA strand breakage, activation of poly (ADP-ribose) synthetase, and cellular energy depletion are involved in the cytotoxicity of macrophages and smooth muscle cells exposed to peroxynitrite Proc Natl Acad Sci U S A 1996 93 1753 1758 8700830
Rubbo H Radi R Trujillo M Telleri R Kalyanaraman B Barnes S Kirk M Freeman BA Nitric oxide regulation of superoxide and peroxynitrite-dependent lipid peroxidation. Formation of novel nitrogen-containing oxidized lipid derivatives J Biol Chem 1994 269 26066 26075 7929318
Wei XQ Charles IG Smith A Ure J Feng GJ Huang FP Xu D Muller W Moncada S Liew FY Altered immune responses in mice lacking inducible nitric oxide synthase Nature 1995 375 408 411 7539113
Karupiah G Chen JH Mahalingam S Nathan CF MacMicking JD Rapid interferon gamma-dependent clearance of influenza A virus and protection from consolidating pneumonitis in nitric oxide synthase 2-deficient mice J Exp Med 1998 188 1541 1546 9782132
Rimmelzwaan GF Baars MM de Lijster P Fouchier RA Osterhaus AD Inhibition of influenza virus replication by nitric oxide J Virol 1999 73 8880 8883 10482647
Akerstrom S Mousavi-Jazi M Klingstrom J Leijon M Lundkvist A Mirazimi A Nitric oxide inhibits the replication cycle of severe acute respiratory syndrome coronavirus J Virol 2005 79 1966 1969 15650225
Paludan SR Ellermann-Eriksen S Lovmand J Mogensen SC Interleukin-4-mediated inhibition of nitric oxide production in interferon-gamma-treated and virus-infected macrophages Scand J Immunol 1999 49 169 176 10075021
Vodovotz Y Bogdan C Paik J Xie QW Nathan C Mechanisms of suppression of macrophage nitric oxide release by transforming growth factor beta J Exp Med 1993 178 605 613 7688028
Vodovotz Y Geiser AG Chesler L Letterio JJ Campbell A Lucia MS Sporn MB Roberts AB Spontaneously increased production of nitric oxide and aberrant expression of the inducible nitric oxide synthase in vivo in the transforming growth factor beta 1 null mouse J Exp Med 1996 183 2337 2342 8642342
Inagaki-Ohara K Iwasaki T Watanabe D Kurata T Nishiyama Y Effect of the deletion of US2 and US3 from herpes simplex virus type 2 on immune responses in the murine vagina following intravaginal infection Vaccine 2001 20 98 104 11567752
Ikemoto K Pollard RB Fukumoto T Morimatsu M Suzuki F Small amounts of exogenous IL-4 increase the severity of encephalitis induced in mice by the intranasal infection of herpes simplex virus type 1 J Immunol 1995 155 1326 1333 7636198
Paludan SR Ellermann-Eriksen S Malmgaard L Mogensen SC Herpes simplex virus type 2 infection of macrophages impairs IL-4- mediated inhibition of NO production through TNF-alpha-induced activation of NF-kappaB Arch Virol 2000 145 575 591 10795524
Ohmori Y Hamilton TA IL-4-induced STAT6 suppresses IFN-gamma-stimulated STAT1-dependent transcription in mouse macrophages J Immunol 1997 159 5474 5482 9548487
Takeda K Kamanaka M Tanaka T Kishimoto T Akira S Impaired IL-13-mediated functions of macrophages in STAT6-deficient mice J Immunol 1996 157 3220 3222 8871614
Harada H Takahashi E Itoh S Harada K Hori TA Taniguchi T Structure and regulation of the human interferon regulatory factor 1 (IRF-1) and IRF-2 genes: implications for a gene network in the interferon system Mol Cell Biol 1994 14 1500 1509 7507207
Harada H Fujita T Miyamoto M Kimura Y Maruyama M Furia A Miyata T Taniguchi T Structurally similar but functionally distinct factors, IRF-1 and IRF-2, bind to the same regulatory elements of IFN and IFN-inducible genes Cell 1989 58 729 739 2475256
Harada H Willison K Sakakibara J Miyamoto M Fujita T Taniguchi T Absence of the type I IFN system in EC cells: transcriptional activator (IRF-1) and repressor (IRF-2) genes are developmentally regulated Cell 1990 63 303 312 2208287
Tanaka N Kawakami T Taniguchi T Recognition DNA sequences of interferon regulatory factor 1 (IRF-1) and IRF-2, regulators of cell growth and the interferon system Mol Cell Biol 1993 13 4531 4538 7687740
Kuo TC Calame KL B lymphocyte-induced maturation protein (Blimp)-1, IFN regulatory factor (IRF)-1, and IRF-2 can bind to the same regulatory sites J Immunol 2004 173 5556 5563 15494505
Paludan SR Ellermann-Eriksen S Malmgaard L Mogensen SC Inhibition of NO production in macrophages by IL-13 is counteracted by Herpes simplex virus infection through tumor necrosis factor-alpha- induced activation of NK-kappa B Eur Cytokine Netw 2000 11 275 282 10903807
Molina-Holgado E Arevalo-Martin A Castrillo A Bosca L Vela JM Guaza C Interleukin-4 and interleukin-10 modulate nuclear factor kappaB activity and nitric oxide synthase-2 expression in Theiler's virus-infected brain astrocytes J Neurochem 2002 81 1242 1252 12068072
Saura M Zaragoza C Bao C McMillan A Lowenstein CJ Interaction of interferon regulatory factor-1 and nuclear factor kappaB during activation of inducible nitric oxide synthase transcription J Mol Biol 1999 289 459 471 10356322
|
16076403
|
PMC1215526
|
CC BY
|
2021-01-04 16:39:00
|
no
|
Virol J. 2005 Aug 3; 2:59
|
utf-8
|
Virol J
| 2,005 |
10.1186/1743-422X-2-59
|
oa_comm
|
==== Front
Virol JVirology Journal1743-422XBioMed Central London 1743-422X-2-621610517310.1186/1743-422X-2-62ResearchMimivirus relatives in the Sargasso sea Ghedin Elodie [email protected] Jean-Michel [email protected] Department of Parasite and Virus Genomics, The Institute for Genomic Research, 9712 Medical Center Drive, Rockville, MD 20850, USA2 Department of Microbiology and Tropical Medicine, George Washington University, Washington DC, USA3 Structural and Genomics Information laboratory, CNRS-UPR2589, IBSM, 13402, Marseille, France; University of Mediterranee School of Medicine, 13385, Marseille, France2005 16 8 2005 2 62 62 1 5 2005 16 8 2005 Copyright © 2005 Ghedin and Claverie; licensee BioMed Central Ltd.2005Ghedin and Claverie; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
The discovery and genome analysis of Acanthamoeba polyphaga Mimivirus, the largest known DNA virus, challenged much of the accepted dogma regarding viruses. Its particle size (>400 nm), genome length (1.2 million bp) and huge gene repertoire (911 protein coding genes) all contribute to blur the established boundaries between viruses and the smallest parasitic cellular organisms. Phylogenetic analyses also suggested that the Mimivirus lineage could have emerged prior to the individualization of cellular organisms from the three established domains, triggering a debate that can only be resolved by generating and analyzing more data. The next step is then to seek some evidence that Mimivirus is not the only representative of its kind and determine where to look for new Mimiviridae. An exhaustive similarity search of all Mimivirus predicted proteins against all publicly available sequences identified many of their closest homologues among the Sargasso Sea environmental sequences. Subsequent phylogenetic analyses suggested that unknown large viruses evolutionarily closer to Mimivirus than to any presently characterized species exist in abundance in the Sargasso Sea. Their isolation and genome sequencing could prove invaluable in understanding the origin and diversity of large DNA viruses, and shed some light on the role they eventually played in the emergence of eukaryotes.
==== Body
Introduction
The discovery and genome sequence analysis of Mimivirus [1,2], the largest of the Nucleo-cytoplasmic Large DNA Viruses (NCLDV), challenged much of the accepted dogma regarding viruses. Its particle size (>400 nm), genome length (1.2 million bp) and extensive gene repertoire (911 protein coding genes) all contribute to blur the established boundaries between viruses and the smallest parasitic cellular organisms such as Mycoplasma or Nanoarchea [2]. In the universal tree of life, the Mimivirus lineage appears to define a new branch, predating the emergence of all established eukaryotic kingdoms [2]. Although this result is compatible with various hypotheses implicating ancestral DNA viruses in the emergence of eukaryotes [3-5], it requires confirmation from additional data. An urgent task is thus to convince ourselves that Mimivirus is not the sole representative of its kind (i.e. a viral counterpart to the platypus) and to provide some rational guidance as to where to begin the search for eventual new Mimiviridae.
Mimivirus was serendipitously discovered within Acanthamoeba polyphaga, a free-living ubiquitous amoeba, prevalent in aquatic environments. Phylogenetic analysis of the most conserved genes common to all nucleo-cytoplasmic large double-stranded DNA viruses (NCLDV) [6] positions Mimivirus as an independent lineage, roughly equidistant from the Phycodnaviridae (algal viruses) and Iridoviridae (predominantly fish viruses). Given the ecological affinity of these virus families for the marine environment, we have examined the sequence data set gathered through environmental microbial DNA sampling in the Sargasso Sea [7] to look for possible Mimivirus relatives.
Results
By comparing Mimivirus ORFs to the Sargasso Sea sequence data set and to all other publicly available sequences, 138 (15%) of the 911 Mimivirus ORFs were found to exhibit their closest match (Blastp E-values ranging from 10-74 to 10-4 [8]) to environmental sequences (see Additional file 1). Even before the discovery of Mimivirus, increasingly complex large double-stranded DNA viruses have been isolated, in particular from unicellular algae. The genome analysis of these Phycodnaviruses revealed a variety of genes encoding enzymes from totally unexpected metabolic pathways [9]. Mimivirus added more unexpected genes (such as translation system components [2]) to this list. As the gene repertoire of these large viruses and the gene content of cellular organisms become increasingly comparable, we have to be cautious in the interpretation of environmental/metagenomics sequence data. To focus our study on environmental organisms most likely to be viruses, we limited further analyses to Mimivirus homologues member of the NCLDV core gene sets [2,6]. These core genes are subdivided into four classes from the most (class I) to least (class IV) evolutionarily conserved [6]. Seven of 10 Mimivirus Class I core genes (L206 to R400) have their closest homologues in the Sargasso Sea data. This is also the case for 3 of 7 class II (R450-R313)core genes, 3 of the 13 class III core genes (R429-L364) and 7 of the 16 Class IV core genes (L4-R301) (Table 1). Overall, 43% of Mimivirus core genes have their closest homologues in the Sargasso Sea data set. To further assess the viral nature of these unknown microbes, we studied the phylogenetic relationships between the corresponding Mimivirus proteins, their Sargasso Sea homologues, and the closest homologues in other NCLDVs (see Materials and Methods). Figure 1a–c exhibits three independent phylogenic trees computed using the MEGA3 software [10] for Mimivirus ORFs R449 (unknown function), R429 (unknown function) and L437 (putative virion packaging ATPase). Figure 1a shows that the closest environmental R449 homologues cluster with Mimivirus separately from the known phycodnaviruses, while other Sargasso Sea homologues cluster in a way suggesting the presence of a new clade distinct from Phycodnaviridae. The tree based on R429 and L437 (Fig. 1b,c) similarly suggests the presence of close Mimivirus relatives not belonging to the Phycodnaviridae or Iridoviridae clades.
Table 1 Matching Status of Mimivirus core genes (type 1 to 4).
ORF# Definition Best score in nr Best score in DNA viruses Best score in Sargasso Sea Status Reciprocal Best match
L206 Helicase III / VV D5 167-virus 167 214 Best ENV YES
R322 DNA pol (B family) extein 207 167 238 Best ENV YES
L437 A32 virion packaging ATPase 169-virus 169 191 Best ENV YES
L396 VV A18 helicase 200-virus 200 187 -
L425 Capsid protein 119-virus 117 142 Best ENV complex
R439 Capsid protein 164-virus 159 173 Best ENV complex
R441 Capsid protein 137-virus 147 209 Best ENV complex
R596 E10R-Thiol oxidoreductase 104-virus 105 119 Best ENV YES
R350 VV D6R – helicase 170-virus 170 102 -
R400 F10L – prot. Kinase 86-virus 86 58 -
R450 A1L-transcr factor 52-virus 47 65 Best ENV
R339 TFII-transcr. factor 62 42 66 Best ENV
L524 MuT-like NTP PP-hydrolase 40 38 39 -
L323 Myristoylated virion prot. A 43 42 40 -
R493 PCNA 92 87 154 Best ENV YES
L312 Small Ribonucl. reduct 341 338 310 -
R313 Large Ribonucl. reduct 766 741 740 -
R429 PBCV1-A494R-like 152-virus 152 216 Best ENV YES
L37 BroA, KilA-N 123-virus 124 65 -
R382 mRNA-capping enz. 86 78 166 Best ENV YES
L244 RNA pol. sub 2 (Rbp2) 727 416 508 -
R501 RNA pol. sub.1 (Rpb1) 805 415 520 -
R195 ESV128-Glutaredoxin 50 39 49 -
R622 S/Y phosphatase 75 73 65 -
R311 CIV193R BIR domain 68 44 51 -
L65 Virion memb. prot 44 44 - -
R480 Topoisomerase II 902 717 367 -
L221 Topoisomerase I bacterial 528 35 516 -
R194 Topoisomerase I pox-like 188 100 145 -
L364 SW1/SNF2 helicase 70-virus 70 72 Best ENV YES
L4 N1R/P28 DNA binding prot 123-virus 124 72 -
L540 Pre-mRNA helicase – splicing 256 136 214 -
L235 RNA pol subunit5 69 38 50 -
R354 Lambda-type exonuclease 69-virus 69 154 Best ENV YES
R343 RNAse III 129 112 131 Best ENV YES
R141 GDP mannose 4,6-dehydratase 294 68 252 -
L258 Thymidine kinase 151 140 124 -
L271 Ankyrin repeats (66 paralogs) 179 152 192 Best ENV complex
R325 Metal-dependent hydrolase 69-virus 69 105 Best ENV YES
L477 Cathepsin B 226 43 47 -
R497 Thymidylate synthase 278 242 217 -
R449 Uncharacterized prot. 69-virus 69 129 Best ENV YES
R303 NAD-dependent DNA ligase 270-virus 270 228 -
L805 MACRO domain 36 33 - -
R571 Patatin-like phospholipase 105 80 122 Best ENV YES
R301 Uncharacterized prot. 48-virus 48 65 Best ENV YES
Figure 1 Phylogenetic evidence of uncharacterized Mimivirus relatives. (a) Neighbor-joining (NJ) clustering (see Materials and Methods) of Mimivirus R449 ORF with its best matching (≈35% identical residues) environmental homologues (noted Sargasso1 to Sargasso6 according to their decreasing similarity) and closest viral orthologues (28% identical). (b) NJ clustering of Mimivirus R429 ORF with its best matching (≈50% identical) environmental homologues (noted Sargasso1 to Sargasso5) and closest viral orthologues (35% identical). (c) NJ clustering of Mimivirus putative virion packaging ATPase L437 with its best matching (≈45% identity) environmental homologues (Sargasso1 and Sargasso2) and closest viral orthologues (34% identical). Abbreviations: Phyco: Phycodnavirus; PBCV: Paramecium bursaria chlorella virus 1; EsV: Ectocarpus siliculosus virus; FsV: Feldmannia sp. virus; HaV: Heterosigma akashiwo virus; Irido: Iridovirus; LCDV: Lymphocystis disease virus 1; Frog: Frog virus 3; Amby: Ambystoma tigrinum stebbensi virus; Rana: Rana tigrina ranavirus; Chilo: Chilo iridescent virus. Bootstrap values larger than 50% are shown. Branches with lower values were condensed.
Another piece of evidence substantiating the existence of an unknown Mimivirus relative in the Sargasso Sea is the discovery of contigs built from the data that contain multiple genes with a high degree of similarity to Mimivirus genes. A spectacular case is illustrated in Figure 2. Here, a 4.5 kb scaffold (See Materials and Method) exhibits 4 putative ORFs. When compared to the whole nr database, each of them has as a best match 4 distinct Mimivirus ORFs: thiol oxidoreductase R368 (29% identical, E-value < 10-9), NTPase-like L377 (25% identical, E-value < 10-20), unknown function L375 (34% identical, E-value < 10-30), and DNA repair enzyme L687 (40% identical, E-value < 10-62). Moreover, the gene order is conserved for three of them (R368, L375, L377). Such colinearity is rarely observed between viral genomes except for members of the same family. Unfortunately, the sequences of these genes are not conserved enough to allow the construction of informative phylogenic trees that would include other NCLDV orthologues.
Figure 2 Organization of four Mimivirus ORF best matching homologues in a 4.5 kb environmental sequence scaffold (approximately to scale). The three colinear Mimivirus homologues are in green. Unmatched ORF extremities are indicated by dots. The two diagonal lines indicate where the two contigs are joined on the scaffold.
As of today, genes encoding capsid proteins are among the most unequivocal genes of viral origin. Except for cases of integrated proviral genomes, no cellular homologues of viral capsid proteins have ever been found. During our study, the closest homologues of Mimivirus capsid proteins were found to be capsid protein genes of environmental origin. For example, Mimivirus capsid protein (R441) was found to be 48.5% identical to an unknown environmental sequence, when it is only 36.2% identical to the major capsid protein Vp49 of Chlorella virus CVG-1, its best match among known viruses (Figure 3). As the environmental capsid protein sequence also shares 44.5% identical residues with the CVG-1 Vp49, the corresponding uncharacterized virus appears to lie at an equal evolutionary distance from the Mimiviridae and the Phycodnaviridae.
Figure 3 Partial 3-way alignment (N-terminus region) of Mimivirus capsid protein (R441) with it best matching homologues in the NR and Environmental sequence databases. The Mimivirus R441 protein shares 83/229 (36.2%) identical residues (colored in red or blue) with the major capsid protein Vp49 of Chlorella virus CVG-1 and 111/229 (48.5%) identical residues (indicated in red or green) with the N-terminus of a capsid protein from an unknown large virus sampled from the Sargasso Sea (Accession: EAD00518). On the other hand, the CVG-1 Vp49 and the Sargasso Sea sequence share 44.5% identical residues. By comparison, the CVG-1 Vp49 protein share 72% of identical residue with PBCV-1 Vp54, its best matching homologue among known phycodnaviruses.
Discussion
Our results predict that DNA viruses of 0.1 to 0.8 microns in size exist in the Sargasso Sea that are evolutionarily closer to Mimivirus than to any presently characterized species. These viruses are abundant enough to have been collected by environmental sampling. It must be noticed that a similar approach attempting to find relatives to two other unique NCLDVs, the African swine fever virus (the unique member of Asfarviridae) and the White spot syndrome virus, a major shrimp pathogen (the sole Nimaviridae), failed to provide convincing results (Claverie, data not shown). The identification of numerous Mimivirus-like sequences in the Sargasso Sea data is thus not simply the result of a large number of sequences been compared, but truly suggests that viruses from this clade are specifically abundant in the sampled marine environment. It is actually expected that many novel viruses will be encountered in natural waters in which they constitute the most abundant microrganisms [11,12]. There might be as many as 10 billion virus particles per litre of ocean surface waters [13]. Interestingly, the specialized literature abounds of descriptions of large virus-like particle associated with algae [e.g. [14-16]], or various marine protists [17,18]. With the exception of Phycodnaviruses [19-21], the genomic characterization of these viruses has not been attempted. Guided by the results presented here, their isolation and genome sequencing could prove invaluable in understanding the diversity of DNA viruses and the role they eventually played in the evolution of eukaryotes.
Materials and methods
The protocols used to collect Sargasso Sea environmental micro-organisms and generate DNA sequences from these samples has been described elsewhere [7]). The data analyzed here correspond to "bacteria-sized" organisms that have passed through 3 μm filters and been retained by 0.8 μm to 0.1 μm filters. Mimivirus-like particles (0.8–0.4 μm) belong in this range.
Database similarity searches were performed using the Blast suite of programs [8] (default options) as implemented on the web server and as implemented at The Institute for Genomic Research. Final similarity searches were performed on the non-redundant peptide sequence databases (nr) and environmental data (env-nr) downloaded from the National Institute for Biotechnology Information ftp server on March 14, 2005. To avoid missing potential better matches with annotated virus ORFs, all Mimivirus ORFs exhibiting a best match (blosum62 scoring scheme) in env-nr were also searched against all DNA virus genomes using TblastN (peptide query against translated nucleotide sequence). The comprehensive list of Mimivirus ORFs exhibiting a best match in the env-nr database is given in Additional file: 1. Phylogenetic analyses were conducted using MEGA version 3.0 [10] (option: Neighbor joining, 250 pseudo-replicates, and gaps handled by pairwise deletion). Tree branches were condensed for bootstrap values <50%.
Only Mimivirus ORFs with best matching homologues in DNA viruses and belonging to the nucleo-cytoplasmic large DNA virus core gene set (2, 6) were analyzed in detail. These ORFs (and matching status) are listed in Table 1. Phylogenetic analyses were limited to viral homologues and environmental sequences exhibiting a reciprocal best match relationship with the corresponding Mimivirus ORF (putative orthologues) (YES in the rightmost column). The three cases (red lines in Table 1) exhibiting the best bootstrap values are shown in Figure 1. Cases of complex relationships, for instance due to the presence of many paralogues (e.g. capsid proteins), are also indicated. These cases of non-reciprocal best matches are frequent (i.e. the closest homologue of a Mimivirus ORFs being an environmental sequence, but the latter sequence exhibiting a better match with a different ORF in the nr database).
Two environmental sampling contigs – contig IBEA_CTG_1979672 (AACY01022731, GI:44566181) and contig IBEA_CTG_1979673 (AACY01022732, GI:44566179) – are linked in a 4,465 bp scaffold (scaffold IBEA_SCF = 2208413) found to contain four ORFs with strong matches to Mimivirus peptides (R368, L377, L375, and L687). The three colinear ORFs (R368, L377, L375) are found on one contig while the orthologue to Mimivirus ORF L687 is found in the second contig. It is conceivable that the lack of colinearity for this fourth ORF is due to an assembly error.
Supplementary Material
Additional file 1
List of Mimivirus ORFs exhibiting a best match in the env-nr database
Click here for file
Acknowledgements
We are indebted to James van Etten for pointing out some ancient observations of very large virus-like particles in algae and marine protists. We thank Stéphane Audic for his help with the server and Hiroyuki Ogata and Vish Nene for reading the manuscript. This work was supported by internal funding from TIGR, CNRS, and the French National Genopole Network.
==== Refs
La Scola B Audic S Robert C Jungang L de Lamballerie X A giant virus in amoebae Science 2003 299 2033 12663918 10.1126/science.1081867
Raoult D Audic S Robert C Abergel C Renesto P The 1.2-megabase genome sequence of Mimivirus Science 2004 306 1344 1350 15486256 10.1126/science.1101485
Villarreal LP DeFilippis VR A hypothesis for DNA viruses as the origin of eukaryotic replication proteins J Virol 2000 74 7079 7084 10888648 10.1128/JVI.74.15.7079-7084.2000
Bell PJ Viral eukaryogenesis: was the ancestor of the nucleus a complex DNA virus? J Mol Evol 2001 53 251 256 11523012 10.1007/s002390010215
Takemura M Poxviruses and the origin of the eukaryotic nucleus J Mol Evol 2001 52 419 425 11443345
Iyer LM Aravind L Koonin EV Common origin of four diverse families of large eukaryotic DNA viruses J Virol 2001 75 11720 11734 11689653 10.1128/JVI.75.23.11720-11734.2001
Venter JC Remington K Heidelberg JF Halpern AL Rusch D Environmental genome shotgun sequencing of the Sargasso Sea Science 2004 304 66 74 15001713 10.1126/science.1093857
Altschul SF Madden TL Schaffer AA Zhang J Zhang Z Gapped BLAST and PSI-BLAST: a new generation of protein database search programs Nucleic Acids Res 1997 25 3389 3402 9254694 10.1093/nar/25.17.3389
Van Etten JL Unusual life style of giant chlorella viruses Annu Rev Genet 2003 37 153 195 14616059 10.1146/annurev.genet.37.110801.143915
Kumar S Tamura K Nei M MEGA3: Integrated software for molecular evolutionary genetics analysis and sequence alignment Brief Bioinform 2004 5 150 163 15260895 10.1186/1471-2105-5-150
Wommack KE Colwell RR Virioplankton: viruses in aquatic ecosytems Microbiol Mol Biol Rev 2000 64 69 114 10704475 10.1128/MMBR.64.1.69-114.2000
Ghedin E Fraser C A virus with big ambitions Trends Microbiol 2005 13 56 57 15680763 10.1016/j.tim.2004.12.008
Fuhrman JA Marine viruses and their biogeochemical and ecological effects Nature 1999 399 541 548 10376593 10.1038/21119
Mattox KR Stewart KD Probable virus infections in four genera of green algae Can J Microbiol 1972 18 1620 1621 5081127
Dodds JA Cole A Microscopy and biology of Uronema gigas, a filamentous eukaryotic green alga, and its associated tailed virus-like particle Virology 1980 100 156 165
Van Etten JL Lane LC Meints RH Viruses and virus-like particles of eukaryotic algae Microbiol Rev 1991 55 586 620 1779928
Shin W Boo SM Virus-like particles in both nucleus and cytoplasm of Euglena viridis Algological Studies 1999 95 125 131
Growing MM Large VLPs from vacuoles of phaeodarian radiolarians and from other marine samples Marine Ecology progress Series 1993 101 33 43
Van Etten JL Graves MV Muller DG Boland W Delaroque N Phycodnaviridae – large DNA algal viruses Arch Virol 2002 147 1479 516 12181671 10.1007/s00705-002-0822-6
Claverie JM Giant viruses in the oceans: the 4th algal virus workshop Virol J 2005 2 52 15967028 10.1186/1743-422X-2-52
Wilson WH Schroeder DC Allen MJ Holden MTG Parkhill J Complete Genome Sequence and Lytic Phase Transcription Profile of a coccolithovirus Science 2005 309 1090 1092 16099989 10.1126/science.1113109
|
16105173
|
PMC1215527
|
CC BY
|
2021-01-04 16:39:01
|
no
|
Virol J. 2005 Aug 16; 2:62
|
utf-8
|
Virol J
| 2,005 |
10.1186/1743-422X-2-62
|
oa_comm
|
==== Front
Virol JVirology Journal1743-422XBioMed Central London 1743-422X-2-741612239610.1186/1743-422X-2-74ResearchHuman enterovirus 71 subgenotype B3 lacks coxsackievirus A16-like neurovirulence in mice infection Chan Yoke-Fun [email protected] Sazaly [email protected] Sime Darby Technology Centre, 2, Jalan Tandang, 46050 Petaling Jaya, Selangor, Malaysia2 Department of Medical Microbiology, Faculty of Medicine, University of Malaya, 50603 Kuala Lumpur, Malaysia2005 26 8 2005 2 74 74 22 6 2005 26 8 2005 Copyright © 2005 Chan and AbuBakar; licensee BioMed Central Ltd.2005Chan and AbuBakar; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
At least three different EV-71 subgenotypes were identified from an outbreak in Malaysia in 1998. The subgenotypes C2 and B4 were associated with the severe and fatal infections, whereas the B3 virus was associated with mild to subclinical infections. The B3 virus genome sequences had ≥85% similarity at the 3' end to CV-A16. This offers opportunities to examine if there are characteristic similarities and differences in virulence between CV-A16, EV-71 B3 and EV-71 B4 and to determine if the presence of the CV-A16-liked genes in EV-71 B3 would also confer the virus with a CV-A16-liked neurovirulence in mice model infection.
Results
Analysis of human enterovirus 71 (EV-71) subgenotype B3 genome sequences revealed that the 3D RNA polymerase and domain Z of the 3'-untranslating region RNA secondary structure had high similarity to CV-A16. Intracerebral inoculation of one-day old mice with the virus resulted in 16% of the mice showing swollen hind limbs and significantly lower weight gain in comparison to EV-71 B4-infected mice. None of the mice presented with hind leg paralysis typical in all the CV-A16 infected mice. CV-A16 genome sequences were amplified from the CV-A16-infected mice brain but no amplification was obtained from all the EV-71-inoculated mice suggesting that no replication had taken place in the suckling mice brain.
Conclusion
The findings presented here suggest that EV-71 B3 viruses had CV-A16-liked non-structural gene features at the 3'-end of the genome. Their presence could have affected virulence by affecting the mice general health but was insufficient to confer the EV-71 B3 virus a CV-A16-liked neurovirulence in mice model infection.
==== Body
Background
Enterovirus 71 (EV-71) was first described in 1969 during an outbreak with central nervous system complications in California [1]. Since then, EV-71 infections have been associated with a number of outbreaks with wide clinical manifestations, ranging from mild hand, foot and mouth disease (HFMD) to severe neurological complications and deaths. These include outbreaks in Bulgaria [2], Hungary [3], Japan [4] and more recently Malaysia [5,6], Taiwan [7] and Singapore [8]. In the later three outbreaks, more than a hundred deaths in total were reported, elevating EV-71 infection as one of the most deadly virus infection to date amongst young children below the age of 3 years old in Asia. The sudden emergence of the deadly forms of EV-71 infection in Asia was puzzling, as the virus together with other human enterovirus A viruses especially coxsackievirus A5 (CV-A5), CV-A10 and CV-A16 have been noted to cause HFMD in the region for sometime [9]. During the outbreak in Malaysia, at least three different EV-71 subgenotypes were identified. The subgenotypes C2 and B4 were associated with the severe and fatal infections, whereas, mild to subclinical infections were associated with the B3 viruses [10-12]. Unlike the earlier two subgenotypes, the B3 virus circulated for only a brief period during the outbreak and they have since not been isolated from patients from the later outbreaks [11,12]. A recent study reported that the B3 virus genome sequences had ≥93% similarity to EV-71 at the 5' end whereas the P3 genome region and 3'UTR had ≥85% similarity to CV-A16 [13]. CV-A16 is known to be the most common causative agent for the self-limiting HFMD. It is usually characterized by mild fever, oral ulcers and vesicular lesions on palms and soles and is not known to cause severe and fatal CNS infections. It is not presently understood why EV-71 infections tend to cause the more severe form of HFMD in comparison to CV-A16. The findings that EV-71 B3 viruses had high sequence similarity to CV-16 at the 3' end of the genome and that the viruses were not associated with the severe form of HFMD, offered opportunities to examine the potential roles of the respective genes in determining virulence. Hence, the present study was undertaken to examine if there are characteristic similarities and differences between CV-A16, EV-71 B3 and the more virulent EV-71 B4 virus and to determine if the presence of the CV-A16-liked genes in the EV-71 B3 virus genome would also confer the virus a CV-A16-liked neurovirulence in mice.
Results and Discussion
The consensus amino acid sequences of the two available EV-71 B3 virus genomes (SHA63 and SHA66) were compared to other available subgenotype B4 and CV-A16/G10 genome sequences from the Genbank. Several amino acids (His1775, Thr1947, Ile1806, Gln1825, Thr1928, Thr1947, Asn2099, Glu2114 and Gln2159) that were characteristic of the CV-A16/G10 were found in EV-71 B3 isolates. These amino acid differences occurred only within the 3D RNA polymerase gene, suggesting that this gene is very much CV-A16 than it is EV-71. Comparisons of the EV-71 B3 amino acid sequences against all other EV-71 and CV-A16 also revealed at least 12 amino acids (Asn1124, Arg1152, Ser1335, Ser1641, Tyr1799, Asp1822, Val1860, Ser1864, Val1997, Ala2039 Asp2101 and Leu2125) that were unique to the EV-71 B3 isolates. Eight of these amino acid differences occurred within the 3D RNA polymerase gene. Two of these unique mutations found were located between amino acids 176–348 genome region essential for RNA-protein interactions [14] (Fig. 1). Alignment of the EV-71 B3 (SHA66) and EV-71 B4 (UH1) isolates RNA polymerase against the three-dimensional crystal structure of poliovirus 1 Mahoney strain 3D RNA polymerase (PDB: 1RDR) was performed to locate these mutations. Of these eight mutations in EV-71 B3 virus, three were located within the finger subdomain and two were located at the palm motif suggesting that the EV-71 B3 virus amino acid substitutions were mainly located within the 3D RNA polymerase functional domains. The highly 'flexible' finger domain is involves in modulating substrate recognition and oligomerization of the polymerase for binding to nucleotides [15]. In poliovirus, mutations within the 3D RNA polymerase located to the 3' end of the genome have been shown to affect neurovirulence [16,17]. Hence, this highlights the potential importance of the 3D RNA polymerase in determining the virus neurovirulence. It was also found that in addition to the presence of CV-A16 or CV-A16-liked 3D RNA polymerase gene sequences, the EV-71 B3 viruses also shared a similar predicted 3' UTR secondary structures with CV-A16/G10 at domain Z (Fig. 2), a domain reported as important in determining cardiovirulence of CV-B3 [18]. Mutations that affect the stem-and-loop structures have been shown earlier to abolish infectivity and virus RNA synthesis [19,20]. The predicted domain Y known to form a tertiary RNA 'kissing' structure with domain X of the EV-71 B3 virus, however, differed from the EV-71 B4 and CV-A16/G10 (Fig. 2).
Figure 1 Structural alignment of EV-71 and CV-A16 3D RNA polymerase amino acid sequences. EV-71 subgenotype B3, B4 and CV-A16/G10 amino acid sequences were aligned against the poliovirus 1 Mahoney 3D RNA polymerase template sequences (PDB: 1RDR). Conserved residues are indicated as (●) and each domain are boxed and labeled. Residues shared by EV-71 B3 virus and CV-A16 were highlighted in grey and residues unique for EV-71 B3 virus were highlighted in pink.
Figure 2 Predicted RNA secondary structures of EV-71 B3, EV-71 B4 and CV-A16/G10 3' UTR. RNA structures were predicted based on the lowest free energy, using the Zuker algorithm as implemented in RNA Structure (version 3.71). The predicted 3' UTR structures consisted of nucleotides from position 7326–7407 and additional 12 nucleotides of the poly-A tail.
Inoculation of one day-old newborn mice showed that all mice inoculated with CV-A16 had the typical signs and symptoms of CV-A16 infections by day two post-inoculation. The mice were lethargic, had floppy tails, tremoring, uncoordinated movement and reduced average body weight in comparison to EV-71 B3- or EV-71 B4-inoculated mice (Fig. 3a,3g). Approximately 17% (4/24) of the mice had hind leg paralysis by day three post-inoculation and one died (Fig. 3b,3e,3f, Additional file: 1). By day four post-inoculation, all the CVA16-inoculated mice had developed hind leg paralysis and subsequently died (Fig. 3b,3e,3f). A 150 bp enterovirus genome sequence were amplified and sequenced from the total RNA of the brain of all the CV-A16-inoculated mice confirming the presence of CV-A16 in the mice brain (Fig. 4). Mice-inoculated with EV-71 B3 and EV-71 B4 viruses also had significantly reduced average body weight in comparison to the control mock-infected mice (Student's t-test, P < 0.05, Fig. 3d,3g). Mice inoculated with EV-71 B3 virus, however, had significantly reduced average body weight in comparison to those inoculated with the EV-71 B4 virus (Fig. 3g). These mice appeared lethargic and uncoordinated beginning on day two post-inoculation. Of these, 16% (4/25) developed swollen hind legs and one subsequently died on day five post-inoculation (Fig. 3c,3e,3f). There were no hind leg paralysis noted and the remaining surviving mice recovered, fed well and regained balance after day six post-inoculation. In contrast, about 20% (6/31) of the mice inoculated with EV-71 B4 virus developed swollen fore limbs or hind legs and of these, three died after day four post-inoculation (Fig. 3e,3f). After day eight post-inoculation, the B4-inoculated mice also recovered, became more active and fed well. Pairwise comparison of the clinical illness and survival probability between the virus-inoculated groups and control were significant suggesting that the three viruses, CV-A16, EV-71 B3 and EV-71 B4 viruses caused death in mice (log rank survival analysis, P < 0.05, Fig. 3e,3f) but only infection with CV-A16 lead to 100% mortality. In contrast to CV-A16 infection, no amplification of the enterovirus sequence was detected in the selected EV-71 B4- and EV-71 B3-inoculated mice brain, suggesting that EV-71 B3 and EV-71 B4 viruses perhaps did not replicate in the mice brain when introduced intracerebrally (Fig. 4). This may help to explain the absence of hind leg paralysis in all the EV-71-infected mice and the complete recovery of all the surviving mice. Death seen amongst these mice may have been caused by infection of other tissues as manifested in mice with swollen limbs and legs. Evidence suggesting that EV-71 strains isolated during the Bulgaria poliomyelitis-like epidemic had higher tropism for mouse muscle tissues than the brain tissues [2] and EV-71 neurovirulence mimicking human infection was achieved only from using a mouse-adapted virus strain but not the parental strain [21,22] support the findings from the present study that EV-71 B3 and B4 did not infect the brain. The infection, however, manifests clinically in some mice as non-specific swollen limbs and legs. Hence it is possible that, though both EV-71 and CV-A16 viruses are closely related, different receptors are utilize for the respective virus entry into the different tissues and this could be mediated through the virus structural proteins. The mutations that occurred within the 3D RNA polymerase of the EV-71 B3 virus along with the presence of CV-A16-liked 3' UTR domain Z RNA secondary structure then could contribute to virulence but by themselves did not affect EV71 neurovirulence in mice as in contrast to CV-A16, the B3 virus lacks tropism for the mice brain. Since the major differences between the EV-71 B4 and EV-71 B3 viruses occurred at the 3' end of the genome, this support the view that the structural genes of EV-71 and CV-A16 determined tissue tropisms.
Figure 3 EV-71 and CV-A16 infections of newborn mice. One-day old newborn mice were intracerebrally inoculated with 1 × 103 PFU virus per mouse and monitored daily. CV-A16-infected mice had floppy tails on day two post-inoculation (a) and hind leg paralysis beginning on day three post-inoculation (arrow, b). Mice with swollen limbs were noted in EV-71 B3 virus infection (arrow, c) and the EV-71 B3-infected mice had significantly reduced body weight gain in comparison to the mock-infected mice (d, V = B3-infected mouse, C = mock-infected mouse). Mice with floppy tails, swollen limbs and paralysis (e) and death (f) were recorded. The weight gain of the surviving mice was also determined (g).
Figure 4 Detection of enterovirus genome sequences in infected newborn mice brain. At selected intervals post-inoculation (indicated by the number above each lane), mice were sacrificed (each mouse indicated by the alphabet above each lane) and RT-PCR was performed using an enterovirus specific primers. The presence of a 150 bp amplified DNA fragment indicates the presence of enterovirus genome, which was later confirmed by DNA sequencing.
Results from the present study, also did not support the possibility that acquisition of CV-A16-liked genome sequences alone is sufficient to confer the EV-71 B3 virus a CVA16-liked neurovirulence in mice. The significant mice weight gain differences noted between mice infected with EV-71 B3 and EV-71 B4 viruses, with the later performing much better, however, suggested that EV-71 B3 virus infection somehow did affect mice general health. As weight gain differences are the only biological parameter that differentiate between the B3 and B4 viruses, it does appears that EV-71 B3 affected mice more than the EV-71 B4 virus. It is also worth noting that in contrast to infection in mice, CV-A16 infection in human in general does not result in severe infection as oppose to EV-71, particularly the EV-71 B4 virus infection. In parallel manner, the EV-71 B3 viruses, while they affected mice, they did not cause severe or fatal infection in humans. These implied that the EV-71 B3 virus is truly different and as its genome suggested, it has to some extent features of both EV-71 and CV-A16 infection in-vivo.
Conclusion
Results from the present study suggest that EV-71 B3 virus had CV-A16-liked non-structural gene 3D RNA polymerase and 3' UTR features at the 3' end of the genome. Their presence affected virulence differently from infection with EV-71 B4 and CV-A16 by affecting the mice general health. The presence of the CV-A16-liked genes, however, was insufficient to markedly influence the neurovirulence properties of EV-71 B3 virus in mice.
Materials and methods
Viruses
Two EV-71 isolates identified from the 1997 HFMD outbreak in Malaysia were used. The subgenotype B3 isolate, SHA66 (EMBL: AJ238457) was isolated from a HFMD patient presented with mild infection [6,23]. The subgenotype B4 isolate, UH1 (EMBL: AJ238455) on the other hand, was isolated from the brain of a patient who died of EV-71-associated neurogenic pulmonary edema [5,6,24]. The CV-A16 isolate used was previously isolated from a HFMD patient seen at the University Malaya Medical Centre. This CV-A16 isolate was identified and characterized using monoclonal antibody staining (Chemicon Cat #3323, California, USA) and amplification of partial 5' UTR gene (data not shown).
Amino acid sequence analysis
Amino acid sequences were examined after stripping the 5' UTR and 3' UTR sequences and consensus sequences of EV-71 B3 and EV-71 B4 viruses were aligned and manually edited using GeneDoc software [25]. The previously published three-dimensional crystal structure of the 3D RNA polymerase was downloaded as template for the alignment. Using the WHAT IF program [26], domains that represent the conserved regions, loops, insertion or deletions were manually visualized to generate a structural alignment.
RNA secondary structure prediction
The 3' UTR RNA secondary structure was predicted using Zuker optimal and suboptimal minimal free energy folding algorithms, as implemented in RNA Structure version 3.71 software [27]. Part of the poly A tract was incorporated into the sequences.
Determination of virulence in mice
A total of 24, 25 and 31 one-day old newborn ICR mice were inoculated intracerebrally with either CV-A16 or SHA66 (B3 virus) or UH1 (B4 virus) virus inoculum. The virus inoculum with infectivity of ~1 × 103 p.f.u. was injected in a volume of 10–20 μl into the mice brain. The mice were closely monitored for any clinical symptoms, paralysis and death and the weight of each surviving mouse was recorded daily up to day 11 post-inoculation. Another litter with at least 10 one-day old newborn mice was injected with comparable growth medium and used as controls. At selected intervals post-infection, some of the mice were sacrificed and the brain tissues were harvested for total RNA using the TRI Reagent™ (Molecular Research Centre, Inc., Cincinnati, USA) following the manufacturer's recommended protocols. The RT-PCR amplification for the detection of enterovirus sequence was performed using 1 μg of RNA. Access RT-PCR kit (Promega, USA) and primer pairs, EntabF (5'-TCC TCC GGC CCC TGA ATG CGG CTA AT-3'; nucleotide positions 449–474, based on MS87 strain, Genbank: U22522) and EVRR (5'-AAT TGT CAC CAT AAG CAG GC-3'; nucleotide positions 586–606) were used. Reverse transcription was performed at 42°C for an hour followed by amplification steps; 95°C-30 seconds, 55°C-30 seconds and 72°C-30 seconds for 30 cycles and finally with 5 minutes extension at 72°C using the PTC thermal cycler (MJ Research, Massachusetts, USA). When no amplicon was obtained, the number of cycle was increased to 40. Alternatively, a second step PCR using similar parameters was performed using ten-fold diluted RT-PCR product as template. The amplified DNA fragments were electrophoresed using 2% agarose gel in 0.5 × tris-acetate EDTA buffer (0.02 M Tris base, 0.5 mM EDTA pH 8.0, 0.057% glacial acetic acid) and sequence confirmation was made by sequencing the DNA fragment.
Statistics
Student's t-test was used to evaluate if the differences in weight between the virus-inoculated mice and control mice was significant. Wilcoxon signed rank test was used to compare the survival and paralysis probability between the virus-inoculated mice and control mice. All statistical analyses were implemented using SPSS for Windows version 11.5 (SPSS Inc., Illinois, USA). All tests were two-sided and P < 0.05 was considered as statistically significant.
List of Abbreviation
CV Coxsackievirus
EV Enterovirus
HFMD Hand, foot and mouth disease
UTR Untranslated region
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
The corresponding author, Sazaly AbuBakar is the principal investigator of the study; is involved in the design, supervision, data analyses and writing of the report. Chan Y-F performed all the virological investigations, nucleotide sequencing and analyses of data. All authors were involved in the preparation of this "Research Article" and figures.
Supplementary Material
Additional File 1
Hind leg paralysis in CV-A16 infected mice. By day three post-inoculation, the mice were lethargic, tremoring and uncoordinated.
Click here for file
Acknowledgements
This study is funded in parts by grants from the Ministry of Science, Technology and Innovation, Malaysia # 06-02-09-001-BTK/TD/002.
==== Refs
Schmidt NJ Lennette EH Ho HH An apparently new enterovirus isolated from patients with disease of the central nervous system J Infect Dis 1974 129 304 309 4361245
Chumakov M Voroshilova M Shindarov L Lavrova I Gracheva L Koroleva G Vasilenko S Brodvarova I Nikolova M Gyurova S Gacheva M Mitov G Ninov N Tsylka E Robinson I Frolova M Bashkirtsev V Martiyanova L Rodin V Enterovirus 71 isolated from cases of epidemic poliomyelitis-like disease in Bulgaria Arch Virol 1979 60 329 340 228639 10.1007/BF01317504
Nagy G Takatsy S Kukan E Mihaly I Domok I Virological diagnosis of enterovirus type 71 infections: experiences gained during an epidemic of acute CNS diseases in Hungary in 1978 Arch Virol 1982 71 217 227 6285858 10.1007/BF01314873
Hagiwara A Yoneyama T Takami S Hashimoto I Genetic and phenotypic characteristics of enterovirus 71 isolates from patients with encephalitis and with hand, foot and mouth disease Arch Virol 1984 79 273 283 6320781 10.1007/BF01310816
Lum LCS Wong KT Lam SK Chua KB Goh AYT Lim WL Ong BB Paul G AbuBakar S Lambert M Fatal enterovirus 71 encephalomyelitis J Pediatr 1998 133 795 798 9842048
AbuBakar S Chee HY Al-Kobaisi MF Xiaoshan J Chua KB Lam SK Identification of enterovirus 71 isolates from an outbreak of hand, foot and mouth disease (HFMD) with fatal cases of encephalomyelitis in Malaysia Virus Res 1999 61 1 9 10426204 10.1016/S0168-1702(99)00019-2
Wu TN Tsai SF Li SF Lee TF Huang TM Wang ML Hsu KH Shen CY Sentinel surveillance for Enterovirus 71, Taiwan, 1998 Emerg Infect Dis 1999 5 458 460 10341187
Singh S Chow VTK Chan KP Ling AE Poh CL RT-PCR, nucleotide, amino acid and phylogenetic analyses of enterovirus type 71 strains from Asia J Virol Meth 2000 88 193 204 10.1016/S0166-0934(00)00185-3
Infectious agent surveillance report (IASR). Isolation and detection report of viruses by years 1982–July 2005: enterovirus 1 & enterovirus 2
McMinn P Lindsay K Perera D Chan HM Chan KP Cardosa MJ Phylogenetic analysis of enterovirus 71 strains isolated during linked epidemics in Malaysia, Singapore, and Western Australia J Virol 2001 75 7732 7738 11462047 10.1128/JVI.75.16.7732-7738.2001
Cardosa MJ Perera D Brown BA Cheon D Chan HM Chan KP Cho H McMinn P Molecular epidemiology of human enterovirus 71 strains and recent outbreaks in the Asia-Pacific region comparative analysis of the VP1 and VP4 genes Emerg Infect Dis 2003 9 461 468 12702227
Herrero LJ Lee CSM Hurrelbrink RJ Chua BH Chua KB McMinn PC Molecular epidemiology of enterovirus 71 in peninsular Malaysia, 1997–2000 Arch Virol 2003 148 1369 1385 12827466 10.1007/s00705-003-0100-2
Chan YF AbuBakar S Recombinant human enterovirus 71 in hand, foot and mouth disease patients Emerg Infect Dis 2004 10 1468 1470 15496251
Cornell CT Semler BL Subdomain specific functions of the RNA polymerase region of poliovirus 3CD polypeptide Virology 2002 298 200 213 12127783 10.1006/viro.2002.1481
Hansen JL Long AM Schultz SC Structure of the RNA-dependent RNA polymerase of poliovirus Structure 1997 5 1109 1122 9309225 10.1016/S0969-2126(97)00261-X
Christodoulou C Colbere-Garapin F Macadam A Taffs LF Marsden S Minor P Horaud F Mapping of mutations associated with neurovirulence in monkeys infected with Sabin 1 poliovirus revertants selected at high temperature J Virol 1990 64 4922 4929 2168976
Tardy-Panit M Blondel B Martin A Tekaia F Horaud F Delpeyroux F A mutation in the RNA polymerase of poliovirus type 1 contributes to attenuation in mice J Virol 1993 67 4630 4638 8392604
Merkle I van Ooij MJ van Kuppeveld FJ Glaudemans DH Galama JM Henke A Zell R Melchers WJ Biological significance of a human enterovirus B-specific RNA elements in the 3' nontranslated region J Virol 2002 76 9900 9909 12208967 10.1128/JVI.76.19.9900-9909.2002
Melchers WJ Hoenderop JG Bruins Slot HJ Pleij CW Pilipenko EV Agol VI Galama JM Kissing of the two predominant hairpin loops in the coxsackie B virus 3' untranslated region is the essential structural feature of the origin of the replication required for the negative-strand RNA synthesis J Virol 1997 71 686 696 8985400
Mirmomeni MH Hughes PJ Stanway G An RNA tertiary structure in the 3' untranslated region of enteroviruses is necessary for efficient replication J Virol 1997 71 2363 2370 9032373
Chen YC Yu CK Wang YF Liu CC Su IJ Lei HY A murine oral enterovirus 71 infection model with central nervous system involvement J Gen Virol 2004 85 69 77 14718621 10.1099/vir.0.19423-0
Wang YF Chou CT Lei HY Liu CC Wang SM Yan JJ Su IJ Wang JR Yeh TM Chen SH Yu CK A mouse-adapted enterovirus 71 strain causes neurological disease in mice after oral infection J Virol 2004 78 7916 7924 15254164 10.1128/JVI.78.15.7916-7924.2004
AbuBakar S Chee HY Shafee N Chua KB Lam SK Molecular detection of enteroviruses from an outbreak of hand, foot and mouth disease in Malaysia in 1997 Scand J Infect Dis 1999 31 331 335 10528868 10.1080/00365549950163734
AbuBakar S Shafee N Chee HY Induction of apoptosis in neurogenic pulmonary edema Med J Malaysia 1999 54 402 403 11045072
Nicholas KB Nicholas HB JrDeerfield DW GeneDoc: analysis and visualization of genetic variation EMBNEW NEWS 1997 4 14
Vriend G WHAT IF: A molecular modeling and drug design program J Mol Graph 1990 8 52 56 2268628 10.1016/0263-7855(90)80070-V
Matthews DH Sabina J Zuker M Turner DM Expanded sequence dependence of thermodynamic parameters improves prediction of RNA secondary structure J Mol Biol 1999 288 911 940 10329189 10.1006/jmbi.1999.2700
|
16122396
|
PMC1215528
|
CC BY
|
2021-01-04 16:39:00
|
no
|
Virol J. 2005 Aug 26; 2:74
|
utf-8
|
Virol J
| 2,005 |
10.1186/1743-422X-2-74
|
oa_comm
|
==== Front
Virol JVirology Journal1743-422XBioMed Central London 1743-422X-2-761613525310.1186/1743-422X-2-76Short ReportHepatitis B virus X protein interacts with β5 subunit of heterotrimeric guanine nucleotide binding protein Lwa Siew Hui [email protected] Wei Ning [email protected] School of Chemical and Biomedical Engineering and School of Biological Sciences, College of Engineering, Nanyang Technological University, Blk 1 Innovation Centre, 16 Nanyang Drive Unit 100 Level 1, Singapore 6377222005 31 8 2005 2 76 76 15 7 2005 31 8 2005 Copyright © 2005 Lwa and Chen; licensee BioMed Central Ltd.2005Lwa and Chen; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
To isolate cellular proteins interacting with hepatitis B virus X protein (HBX), from HepG2 cells infected with hepatitis B virus (HBV).
Results
HBV particles were produced in culture medium of HepG2 cells transfected with the mammalian expression vector containing the linear HBV genome, as assessed by commercially available ELISA assay. A cDNA library was made from these cells exposed to HBV. From yeast two hybrid screening with HBX as bait, human guanine nucleotide binding protein β subunit 5L (GNβ5) was isolated from the cDNA library constructed in this study as a new HBX-interacting protein. The HBX-GNβ5 interaction was further supported by mammalian two hybrid assay.
Conclusion
The use of a cDNA library constructed from HBV-transfected HepG2 cells has resulted in the isolation of new cellular proteins interacting with HBX.
==== Body
Background
Infection by hepatitis B virus (HBV), an enveloped DNA virus of the hepadnaviridae family, has been closely related to development of hepatocellular carcinoma (HCC). The role of this virus in the series of events leading to the onset of HCC has remained elusive [1]. However, it has been suggested that the smallest protein encoded by the HBV genome, HBX, is involved in the development of HCC [2].
Several proteins have been demonstrated to interact with the HBX protein through the use of the Yeast Two Hybrid system. These include the C-terminal portion of a novel human proteasome alpha subunit which possesses a protein sequence of close relationship to that of the 28 kD subunits from other organisms; PSMA7, the α-subunit of the 20S proteasome complex; PSMC1, the subunit of the 19S proteasome regulatory cap complex; XAPC7, a highly conserved proteasome subunit belonging to the α-subunit of the 20S proteasome complex. PSMA7, PSMC1 and XAPC7 were demonstrated to interact with the second Kunitz-type domain of the HBX protein [3-6]. Another two proteins, XAP2 (X-associated protein 2) and XIP (HBX-interacting protein) were found to be negative regulators of HBX. Overexpression of XAP2 abolished the transactivating function of HBX while the specific interaction of XIP to the carboxy terminus of HBX in differentiated HCC cells led to a reduction of wild-type HBV viral replication to levels similar to that observed after transfection with HBX-minus virus [7,8].
HBX was also shown to interact with XAP-1 (X-associated protein 1), a human homolog of the monkey UV-damaged DNA-binding protein (UV-DDB) which might be involved in DNA repair [9]. Besides these, HBX was able to interact and colocalise with HVDAC3 to the mitochondria, resulting in decreased mitochondrial transmembrane potential. HVDAC3 was identified as a new member of the human voltage-dependent anion channel (VDAC) family that provides pathways for ATP and metabolites across the mitochondrial membrane. It constitutes part of the permeability transition pore complex in the mitochondral membrane which regulates mitochondrial transmembrane potential and cytochrome c release [1].
However, these reported HBX interacting proteins have all been isolated from normal liver cDNA library (cells that had not been exposed to HBV), which may reflect physical but physiologically not meaningful interactions. There is therefore a need to comprehensively isolate and characterize, in a HBV-infected environment, cellular proteins interacting with the widely studied HBX.
Such an environment has previously been generated in vitro, firstly by showing that clonal cells derived from HepG2 transfected with a HBV-containing plasmid could elicit acute hepatitis in chimpanzees through secretion of hepatitis B surface antigen (HBsAg) particles and virions [10], and later by our investigation indicating the production of HBV particles in culture medium of HepG2 cells transfected with a replication competent HBV genome cloned in a mammalian expression vector [11].
The aim of this study is to isolate cellular proteins interacting with HBX in an HBV-infected environment.
Results and Discussion
Infection of HepG2 Cells by Replicative HBV Genome: an ELISA Analysis
The culture medium of the HepG2 cells transfected with infective HBV genome [11] was assayed for the presence of HBsAg prior to cell harvesting for RNA extraction. In Table 1, values D1, E1, F1, G1, H1, A2, B2, C2, D2, E2, F2, G2 and H2 corresponded respectively to the undiluted medium, its 10×, 102×, 103×, 104×, 105×, 106×, 107×, 108×, 109×, 1010×, 1011×, 1012× serial dilutions measured at an absorbance of 450 nm. The presence of HBsAg was indicated in the undiluted medium, its 10 × dilution and up to 104 × dilution. The values of these samples were respectively, 2.393, 0.464 and down to 0.186. They exceeded the calculated cut-off of 0.179, thus indicating the presence of HBsAg. The results of this assay run were valid as the criteria for the quality control had been met. The mean A450 of the negative controls was 0.129, which did not exceed 0.2. The A450 of the positive control was 1.494, which was 1.365 higher than the mean A450 of the negative controls. The value of the positive control fulfilled the quality control criteria which requires the value to be at least 0.8 higher. The cut-off of the assay reading, 0.179, was derived by adding 0.05 to the mean A450 of the negative controls. Based on this cut-off value, readings that were equivalent to or higher than 0.179 indicated reactive samples. Taken together, our results indicated the presence of viral infection process in HepG2 cells transfected with the replicative HBV genome.
Table 1 Production of HBV in Transfected HepG2 Cells by ELISA Analysis
Wells Samples Absorbance at 450 nm (A450)
A1 Negative Control 1 0.131
A2 Negative Control 2 0.126
A3 Positive Control 1.494
A4 Undiluted growth medium of HBV-transfected HepG2 cells 2.393
A5 10 × diluted growth medium 0.464
A6 102 × diluted growth medium 0.164
A7 103 × diluted growth medium 0.142
A8 104 × diluted growth medium 0.186
cDNA Library Construction
The cDNA library that would be used in the yeast-two-hybrid screening for protein-protein interactions was constructed by simultaneous transformation of double-stranded cDNA and the activation domain vector, pGADT7-Rec, into yeast strain AH109. After a 4-day incubation period, yeast plasmid DNA was extracted 10 random yeast colonies used as template for PCR. Presence of cDNA inserts of varied sizes ranging between 300 to 600 base pairs were indicated in lanes 4, 6, 7, 9, 10 and 11 as shown in Figure 1. Library transformants were then harvested and pooled.
Figure 1 Analysis of Insert Size of the cDNA Library from HBV-transfected HepG2 Cells. Lane 1: 100 bp Marker. Lanes 2–11: Yeast plasmid DNA extracted from library colonies to check for range of sizes of inserts before use of library for screening. Lanes 4, 6, 7, 9, 10 and 11 showed inserts ranging between 300 to 600 bp.
Yeast Two Hybrid Screening for HBX Interacting Proteins
The total number of colonies that resulted from the yeast two hybrid screening and the percentage number of colonies that progressively developed a blue colour over a span of 4 days was tabulated in Table 2. The absence of blue color for the 10 colonies resulting from the empty pGBKT7 suggested that 50% of the blue colonies corresponded to HBX-interacting proteins.
Table 2 Summary of Yeast Two Hybrid Screening for HBX Interacting Proteins
Transformed Y187 yeast strain Total number of colonies after interaction with cDNA library % number of blue colonies
Day
1 2 3 4
Full length HBX in pGBKT7 1.6 × 104 0 10 20 50
pGBKT7 without HBX 10 0
Blue colour development was due to the use of the X-α-Gal assay system as an indication of positive interactions in the yeast-two-hybrid system. Colonies that resulted from the screening of the cDNA library and the HBX-pGBKT7 bait as well as those that resulted from the negative control reaction which is the interaction between the empty bait and the cDNA library were observed for their development of blue colouration on SD/-Ade/-His/-Trp/-Leu/X-α-Gal agar.
Y187 yeast strain was transformed individually with each of the 3 control bait constructs empty pGBKT7 vector, full length HBX in pGBKT7 vector and negative pGBKT7-Lamin control vector. Colonies that grew on SD/-Trp were transferred to SD/-Trp/X-α-Gal. Blue colour development was observed over a period of 5 days. Transformation of these 3 negative control plasmids into Y187 yeast strain did not produce yeast colonies on SD/-Ade/-Trp and SD/-His/-Trp plates after incubation of 5 days. A comparison of the intensity of blue coloration of positive clones resulting from the library screening is shown in Fig 2. In contrast to these blue colonies, only white color was observed for the three negative control clones from the individual transformation of the empty pGBKT7 vector, full length HBX in pGBKT7 vector and negative pGBKT7-Lamin control vector.
Figure 2 Positive Yeast Colonies with Potential HBX Interacting Proteins. Comparison between control yeast colonies and positive colonies from the library screening that developed blue colour over a 5-day incubation period at 30°C. A1-5, B1-5, C1-5, D1-5, E1-5, F1, G6-10, H6-10, I6-10, J6-10, K6-10, L6: colonies that turn blue after α-galactosidase assay screening indicated positive interaction between bait and cDNA library. F2 and L7: control using bait plasmid cloned with HBX transformed into yeast strain Y187 to test if the bait alone led to transcriptional activation. α-galactosidase assay screening indicate negative interaction between bait and library. Colonies remained white. F3 and L8: control using negative control plasmid, pGBKT7-Lamin transformed into yeast strain Y187. α-galactosidase assay screening indicate negative interaction between bait and library. Colonies remained white. F4 and L9: Control using bait plasmid, pGBKT7 transformed into yeast strain Y187. α-galactosidase assay screening indicate negative interaction between bait and library. Colonies remained white.
The negative control experiment as shown in Table 2, in which an empty bait vector was used in place of the HBX bait yielded 10 colonies after the screening but none of these turned blue. Therefore, it can be concluded that the X-α-Galactosidase assay is a reliable method for determining positive interactions in the yeast-two-hybrid system. As shown in Fig. 2, the HBX bait alone did not activate transcription of the reporter gene that led to blue colour development. This confirms that the blue colonies which resulted after screening were indeed due to interactions between the AD fusion protein and the BD fusion protein in the two separate yeast strains.
Through the analysis of sequences of 9 positive clones, only one had a library insert that was in frame with the activation domain of GAL4 protein on pGADT7-Rec vector. The insert corresponded to human guanine nucleotide binding protein β subunit 5L (GNβ5). Although it has been well established that HBX interacts with many molecules of cellular signaling pathway, GNβ5 has hitherto not been identified as a HBX-interacting protein. The use of a cDNA library constructed from HBV-transfected HepG2 cells would therefore be significant in that it would enable cellular signaling events occurring after HBV infection to be traced via interactions with the HBX protein which had been implicated in HBV-related hepatocellular carcinoma. This also implies that the use of an infected library increases the likelihood of identifying new interacting partners.
Despite the stringency of nutritional selection through the use of the AH109 yeast strain which contains three reporters, namely, ADE2, HIS3 and MEL1, the yeast-two-hybrid system is only able to reduce the number of false positives. There still exists an uncertainty of whether interactions are always true positives. For this reason, the Mammalian-Two-Hybrid system was used to re-confirm the interaction. This system verifies protein-protein interactions by transcriptional activation. Similar to the yeast two hybrid system, the Mammalian-Two-Hybrid Checkmate system consists of two vectors namely pBIND and pACT. Interaction between two proteins expressed from these two vectors will be assessed by the luciferase activity following the transient transfection of both vectors into mammalian cells.
The pBIND-GNβ5 plasmid was transfected together with the pACT-HBX plasmid and the pG5luc luciferase vector to test the protein-protein interaction between partial GNβ5 sequence and full-length HBX in the Mammalian-Two-Hybrid system. The percentage activity as reflected by the luciferase reporter was plotted out in the graph shown in Figure 3. The assay value given by this interaction was compared against the positive control nteraction between pACT-MyoD and pBIND-1d and a negative control which included pBIND-GNβ5 and the empty pACT vector. The results showed that the luciferase activity triggered by interaction between the partial GNβ5 and full length HBX was twenty percent higher than that produced by the negative control reaction, thus suggesting presence of interaction. However, the luciferase activity of the interaction was only 63% of that measured for the interaction in the positive control. Thus, although the experimental value exceeds that of the negative control, the confirmation of the interaction between HBX and partial GNβ5 can be fine tuned by further tests using the same system.
Figure 3 Dual-Luciferase Reporter Assay demonstrating the interaction between GNβ5 and HBX. The pBIND-GNβ5 plasmid was transfected together with the pACT-HBX plasmid and the pG5luc luciferase vector to test the protein-protein interaction between partial GNβ5 sequence and full-length HBX in the Mammalian-Two-Hybrid system. The percentage activity as reflected by the luciferase reporter was plotted out in the graph.
Possible Role of GNβ5 in HBX-mediated Cellular Signaling Pathway
G proteins are heterotrimeric guanine nucleotide binding proteins that are involved in signal transduction. They are peripherally associated with the plasma membrane and function to couple signals to seven transmembrane-spanning surface receptors. G proteins consist of α, β and γ subunits, of which β and γ subunits are tightly associated. In a typical G protein coupled signaling pathway, the ligand-activated receptor catalyzes the exchange of guanine nucleotides in the α subunit. The GTP-bound α subunit dissociates from the receptor as well as the βγ subunit and proceeds to activate its respective effector molecule. The free βγ subunit also activates a similar or different effector molecules.
The hydrolysis of the bound-GTP to GDP by intrinsic GTPase activity of the α subunit leads to a conformational switch, resulting in the termination of its effector interaction. The resulting α-GDP re-associates with the free βγ subunit. This is followed by re-entry of the newly formed heterotrimer into the signaling cycle. To date, 17 α subunits, five β subunits and 12 γ subunits have been identified [12].
The GNβ5 isolated in our study would be of significance. Previous reports had suggested that the GNβ5 protein was essentially soluble. The GNβ5 protein displays sequence homology to a group of Gβ-like proteins known as the WD-40 repeat family [13]. It has been proposed that this repeat enables G protein β subunits to adopt multiple conformations which could greatly expand the number of signaling partners.
G proteins and Cancer Development
G proteins have also been implicated in cancer development [12]. In the classical paradigm, GPCR-mediated signal transduction involves the agonist-dependent interaction of GPCRs with G proteins at the plasma membrane, and the subsequent generation of soluble second messengers or ion currents by membrane localized effectors.
There exist two major mechanisms for transmembrane signaling in intercellular communication, mediated respectively by receptor tyrosine kinases and by GPCRs. Recent evidence shows that the two pathways can converge on the same effectors, for example, Ras and mitogen-activated protein kinase (MAPK). Both systems appear to use specific protein-protein interactions for localization of key signaling intermediates to appropriate membrane compartments. For receptor tyrosine kinases, protein-protein interactions are mediated by Src homology SH2 and SH3 while for GPCRs, interaction of the βγ complex of heterotrimeric G proteins [13].
Based on the above, HBX probably associates with GNβ5 at the early cellular stage of HBV infection. From previous reports that Ras and MAPK constitute the point of convergence of the tyrosine kinase and the G protein coupled receptor signaling pathways, it is likely that HBX plays a role in bridging and activating the Src-kinase and MAPK mediated pathways at the early stage of viral infection. Further functional studies, including the down-regulation of expression of GNβ5 using siRNA, should shed new lights on its role in HBV infection.
Methods
Cell Culture and Transfection
HepG2 cells were grown in Dulbecco's Modified Eagle's Medium (DMEM) (Gibco) supplemented with 10% fetal calf serum. HepG2 cells cultured in 10 ml of DMEM in a 9 cm × 6 cm flask were transiently transfected by Effectene (QIAGEN) with 1.5 μg of pcDNA 3.1-HBV DNA when cell confluence reached 60%. Cells were maintained in a saturated, humidified environment of 5% CO2 – 95% air at 37°C. After 36 hours of incubation, 1 ml of the cell culture medium was retained for ELISA analysis.
ELISA Analysis
The HBV DNA level in the culture medium of the transfected HepG2 cells was analysed by ELISA (Murex HBsAg Version 3, ABBOTT).
Bait Plasmid Construct
The full length HBX gene sequence of 462 base pairs was amplified by PCR using full length HBV DNA (adw2 subtype) as template. Oligonucleotide primers were designed with NcoI and PstI enzyme restriction sites as follows:
5'-TGCCATGGCAATGGCTGCTAGGCTGTACTGCC-3'
5'-AACTGCAGTTAGGCAGAGGTGAAAAAGTT-3'
The PCR product and the binding domain vector, pGBKT7 of 7.3 kb (MATCHMAKER Two-Hybrid System 3, BD Biosciences) were digested individually with enzymes, NcoI and PstI. (New England Biolabs) After each enzymatic digestion, the PCR product and vector were column-purified. (QIAGEN) 2 μl of the purified PCR product was ligated to 6 μl of the binding domain vector in a reaction mixture which included 1 μl of T4 DNA Ligase and 1 μl of T4 DNA Ligase Buffer (New England Biolabs). Incubation of the mixture was carried out overnight at 16°C. 5 μl of the ligation mixture was transformed into 50 μl of DH5α competent cells (Stratagene). Transformed cells were plated on 30 μg/ml Kanamycin LB agar plates and incubated at 37°C. Colonies were analysed for presence of correctly inserted bait DNA using the pair of primers used for amplifying the corresponding HBX bait DNA in a PCR reaction. Plasmid DNA from the colony identified to contain the correct insert was extracted using the Miniprep Plasmid DNA extraction kit (QIAGEN) and was transformed into yeast strain, Y187 (Clontech Laboratories, Inc.)
cDNA Library Construction
RNA was isolated from 2 × 107 HBV-transfected HepG2 cells in a final elution volume of 50 μl using the RNeasy Mini Kit (QIAGEN). cDNA was synthesised using the MATCHMAKER Library Construction and Screening Kit (Clontech Laboratories, Inc) First strand cDNA synthesis was carried out using 1 μl of a random CDS III/6 Primer: 5'-ATTCTAGAGGCCGAGGCGGCCGACATG-NNNNNN-3', 2 μl of RNA and 1 μl of water. The mixture was incubated at 72°C for 2 minutes followed by incubation on ice for 2 minutes. This was followed by the addition of 2 μl of 5 × First-Strand Buffer, 1 μl of DTT (20 mM), 1 μl of dNTP mix (10 mM) and 1 μl of Moloney Murine Leukemia Virus Reverse Transcriptase. This was followed by a 10 minute-incubation period at 25°C and a 10 minute-incubation period at 42°C. 1 μl of SMART III Oligonucleotide (10 μM; 5'AAGCAGTGGTATCAACGCAGAGTGGCCATTATGGCCGGG-3') was addedand the reaction mixtures were incubated at 42°C for 1 hour. The tube was incubated at 75°C for10 minutes to terminate first-strand synthesis and cooled to room temperature before adding 1 μl (2 units) of RNaseH. Next, incubation at 37°C was carried out for 20 minutes. Second strand cDNA was synthesised by Long Distance-PCR. Sufficient double-stranded cDNA was prepared for transformation by setting up two 100 μl reaction mixtures, each composed of the following: 2 μl of first-strand cDNA, 70 μl of deionised water, 10 μl of 10 × Advantage 2 PCR Buffer, 2 μl of 50 × dNTP mix, 2 μl of 5' PCR Buffer, 2 μl of 3' PCR Buffer, 10 μl of GC-Melt Solution and 2 μl of 50 × Advantage 2 Polymerase Mix. Thermal cycling was carried out using the following conditions: 95°C for 30 seconds; 24 cycles of 95°C for 10 seconds, each followed by 68°C for {6 minutes + [5(x)] seconds} where x increased per cycle from "0, 1, 2, 3,..." to 23; 68°C for 5 minutes. Double-stranded cDNA was column-purified using CHROMA SPIN + TE - 400 columns (Clontech Laboratories, Inc.).
200 μl of AH 109 yeast competent cells (BD Biosciences) was transformed with 14 μl of ds cDNA and 6 μl of pGADT7-Rec. The transformation mixture was distributed onto 130 100 mm SD/-Leu agar plates. The plates were incubated at 30°C for 5 days.
Harvesting of Yeast Transformants
1 litre of freezing medium (YPD medium with 25% v/v glycerol and sterilised at 121°C for 15 minutes) was prepared for the harvest of transformants. 5 ml of freezing medium was added to each plate and yeast colonies were scraped off the plates and pooled into a sterile 1 litre conical flask. The mixture was mixed well by swirling the flask before storing as 1 ml aliquots at -80°C.
Interaction of Yeast Strains
A single yeast colony from the transformation of the bait plasmid into Y187 yeast strain was innoculated into 50 ml of SD/-Trp liquid medium and incubated at 30°C with shaking at 270 rpm. When OD600 of the culture reached 0.8, it was combined with an 1 ml aliquot of the cDNA library in a 1-litre conical flask together with 45 ml of yeast culture medium (2 × YPDA medium with Kanamycin, 50 μg/ml). The mixture was incubated for 24 hours at 30°C with gentle swirling at 40 rpm.
A drop of the mixture was analysed under phase-contrast microscope (400×) to check for the presence of zygotes. Centrifugation of the mixture was carried out at 1000 × g for 10 minutes and the supernatant was discarded. The cell pellet was resuspended in 10 ml of 0.5 × YPDA/Kanamycin (50 μg/ml). The entire mixture was plated out onto 100 mm SD/-Ade/-His/-Leu/-Trp agar plates. 100 μl of the suspension was distributed evenly onto each plate. The plates were incubated at 30°C for 5 days.
Selection of Yeast Diploids Expressing Interacting Proteins
Only colonies that measured 2 mm or more after 5 days of incubation at 30°C were selected for further screening. 60 colonies were randomly selected for the first round of screening. These were streaked onto SD/-Leu/X-α-Gal agar plates. The agar plates were incubated at 30°C over a 4-day incubation period. 52 colonies that turned blue at the end of the 4-day incubation period on the SD/-Leu/X-α-Gal agar plates were individually innoculated into 10 ml of SD/-Leu medium and 5 ml of SD/-Ade/-His/-Leu/-Trp medium. The liquid cultures were incubated with shaking at 30°C, 280 rpm. After 48 hours, glycerol stocks of these cultures were prepared by adding 1.4 ml of each culture to 0.3 ml of glycerol and 0.3 ml of the respective medium. The stocks were stored at -80°C. The SD/-Leu cultures were centrifuged at 5500 rpm, 4°C, for 10 minutes. Yeast plasmid extraction was carried out to determine library insert sizes. 2 μl of extracted plasmid was used as template for PCR. 52 samples corresponding to bands with sizes of approximately 500 basepairs or more were purified using PCR purification columns (QIAGEN) and sequenced. These 52 colonies were also streaked onto SD/-Trp/X-α-Gal agar plates in a grid-like pattern. The following plasmids, the pGBKT7 vector, pGBKT7 cloned with full-length HBX and the pGBKT7-Lamin negative control vector (BD Biosciences) were transformed into Y187. The resulting colonies were used as controls. The agar plates were incubated at 30°C and the rate at which the colonies turned blue over a 4-day incubation period was noted.
Confirmation of Interaction by Mammalian CheckMate System
The partial human guanine nucleotide binding protein β subunit 5L (GNβ5) insert was amplified for cloning in-frame with the GAL 4 containing binding domain vector, pBIND (CheckMate System, Promega) for use in the Mammalian-two-hybrid system. The 5' and 3' primers were designed with BamHI and EcoRV digestion sites respectively as follows: 5'-GAGGATCCTCAAAGATAAGAGGAGGATCGT-3' and 5'-GAGATATCTCGGGGGCCAGGTCCAAGCAGA-3'
HBV DNA (adw2 subtype) was used as template for PCR to amplify the full length HBX sequence of 462 base pairs using the following 5' and 3' primers which were designed with SalI and EcoRV digestion sites respectively:
5'-TGGTCGACCAATGGCTGCTAGGCTGTACTGC-3'
5'-AAGATATCTTTTAGGCAGAGGTGAAAAAGTT-3'
The resulting PCR product was purified and cloned into the Herpes Simplex Virus VP16 activation domain vector, pACT, of the Mammalian-two-hybrid system (CheckMate System, Promega).
Transfection of HepG2 Cells
Transfection of HepG2 cells was carried out using Effectene reagent (QIAGEN) when cells reached 35% confluence. Duplicated reactions were carried out in two 6-well plates. For each of the positive control, negative control and experimental wells, 1.2 ng of combined DNA composing of 3 different plasmids in a 1:1:1 ratio was used. The positive control reaction consisted of 0.4 ng of each of the following vectors, pG5luc, pBIND-1d and pACT-MyoD. The negative control reaction consisted of 0.4 ng of each of the following vectors, pG5luc, empty pACT vector and the GNβ5-pBIND vector. The actual experimental reaction to confirm the interaction consisted of 0.4 ng of each of the following vectors, pG5luc GNβ5-pBIND vector and the HBX-pACT vector. Cells were incubated in a saturated, humidified environment of 5% CO2 – 95% air at 37°C. After 2 hours of incubation, the growth medium was aspirated and the cells were washed twice using 2 ml of PBS per wash. 2 ml of fresh medium was added to each well and the plates were incubated in 5% CO2 and 95% air at 37°C for 36 hours.
Harvesting and Lysis of Transfected cells
The transfected cells were trypsinised and the cell suspension of each well was collected and centrifuged at 1500 rpm at 4°C for 3 minutes. The supernatant was removed and each cell pellet was resuspended in 4 ml of PBS. Centrifugation was then carried out at 1500 rpm at 4°C for 3 minutes. The supernatant was discarded and 250 μl of 1 × Passive Lysis Buffer (Promega) was used to lyse each cell pellet.
Dual-Luciferase Reporter Assay
20 μl of each test sample was mixed with 100 μl of Luciferase Assay Reagent II (Promega) in a luminometer tube. The luminometer (Sirius Tube Luminometer, Berthold Detection Systems) was programmed to provide a 2-second pre-read delay, followed by a 10-second measurement period for each reporter assay. Upon recording the first reading, 100 μl of 1 × Stop & Glo Reagent (Promega) was promptly added to the reaction mix. After mixing, the second reading was recorded.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
SH Lwa was a recipient of a graduate research scholarship from Nanyang Technological University, and conducted experiments under the direction of Dr. Chen. Dr. Chen initiated the research, writing of the draft manuscript with subsequent editing and revisions by both authors.
Acknowledgements
This work was supported by grant 03/1/22/18/229 (WN Chen) from the Biomedical Research Council, Agency for Science, Technology and Research, Singapore.
==== Refs
Rahmani Z Huh KW Lasher R Lasher R Siddiqui A HBV X protein colocalizes to mitochondria with a human voltage-dependent anion channel, HVDAC3, and alters its transmembrane potential J Virol 2000 74 2840 2846 10684300 10.1128/JVI.74.6.2840-2846.2000
Nijhara R Jana SS Goswami SK Rana A Majumdar SS Kumar V Sarkar DP Sustained activation of mitogen-activated protein kinases and activator protein 1 by the hepatitis B virus X protein in mouse hepatocytes in vivo J Virol 2001 75 10348 10358 11581403 10.1128/JVI.75.21.10348-10358.2001
Fischer M Runkel L Schaller H HBx protein of hepatitis B virus interacts with the C-terminal portion of a novel human proteasome alpha-subunit Virus Genes 1995 10 99 102 7483296 10.1007/BF01724303
Hu Z Zhang Z Doo E Coux O Goldberg AL Liang TJ Hepatitis B virus X protein is both a substrate and potential inhibitor of the proteasome complex J Virol 1999 73 7231 7240 10438810
Huang J Kwong J Sun EC Liang TJ Proteasome complex as a potential cellular target of HBX protein J Virol 1996 70 5582 5591 8764072
Zhang Z Torii N Furusaka A Malayaman N Hu Z Liang TJ Structural and functional characterization of interaction between hepatitis B virus X protein and the proteasome complex J Biol Chem 2000 275 15157 15165 10748218 10.1074/jbc.M910378199
Kuzhandaivelu N Cong YS Inouye C Yang WM Seto E XAP2, a novel hepatitis B virus X-associated protein that inhibits X transactivation Nucleic Acids Res 1996 24 4741 4750 8972861 10.1093/nar/24.23.4741
Melegari M Scaglioni PP Wands JR Cloning and characterization of a novel hepatitis B virus x binding protein that inhibits viral replication J Virol 1998 72 1737 1743 9499022
Lee TH Elledge SJ Butel JS Hepatitis B virus X protein interacts with a probable cellular DNA repair protein J Virol 1995 69 1107 1114 7815490
Sells MA Zelent AZ Shvartsman M Acs G Replicative intermediate of hepatitis B virus in HepG2 cells that produce infectious virions J Virol 1988 62 2836 2844 2839705
Chen WN Oon CJ Changes in antigenicity of a hepatitis B virus mutant stemming from lamivudine therapy Antimicrob Agents Chemother 2000 44 1765 10896650 10.1128/AAC.44.6.1765-1765.2000
Radhikal V Dhanasekaran N Transforming G proteins Oncogene 2001 20 1607 1614 11313908 10.1038/sj.onc.1204274
Inglese J Koch WJ Touhara K Lefkowitz R G-βγ interactions with PH domains and Ras-MAPK signaling pathways TIBS 1995 20 151 155 7770915
|
16135253
|
PMC1215529
|
CC BY
|
2021-01-04 16:39:01
|
no
|
Virol J. 2005 Aug 31; 2:76
|
utf-8
|
Virol J
| 2,005 |
10.1186/1743-422X-2-76
|
oa_comm
|
==== Front
World J Surg OncolWorld Journal of Surgical Oncology1477-7819BioMed Central London 1477-7819-3-561612021810.1186/1477-7819-3-56EditorialPreoperative lymphoscintigraphy and triangulated patient body marking are important parts of the sentinel node process in breast cancer Krynyckyi Borys R [email protected] Suk Chul [email protected] Chun K [email protected] Department of Radiology, Division of Nuclear Medicine, The Mount Sinai School of Medicine, The Mount Sinai Hospital, New York, New York, USA2005 24 8 2005 3 56 56 17 6 2005 24 8 2005 Copyright © 2005 Krynyckyi et al; licensee BioMed Central Ltd.2005Krynyckyi et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Breast CancerSentinel Lymph NodeMorbidityLymphoscintigraphy
==== Body
Introduction
Failure to visualize or correctly visualize sentinel nodes (SN) during preoperative lymphoscintigraphy (LS) is a frustrating problem. Most of these instances occur due to inexperience in performing the studies, and can be realized and corrected with the use of proper technique, even in centers that do not have state of art equipment. This is now becoming a major issue, as more and more sentinel node biopsies are being performed, and will increasingly gain more importance once sentinel lymph node biopsy (SLNB) becomes the standard of care in patients with breast cancer world wide.
Pandey et al. [1] has recently reported two unfortunate experiences with LS during SLNB. From the images and the description of the cases, it appears that the studies were suboptimal in both imaging and injection technique, and understandably, completely frustrating to any surgeon.
In the second case, contamination of the patient by stray radioactivity from the perilesional injections is suggested by the authors as the cause of the superior focus that appeared in the supracalvicular region [1]. Actually, true contamination is very easy to realize with proper imaging technique [2-4]. When multiple angled views (0°, 45°, 90°) are obtained, including standing/sitting views and triangulated body marking (TBM), contamination is extremely unlikely to be missed, as it is always surface based. In over 2000 LS cases we have performed, the very rare stray activity has always been picked up for what it is, before the patient is presented to the surgeons [2-4].
In the first case, a SN was hidden by the injected perilesional activity [1]. This is a known issue when the primary lesion is located near the axilla itself. Multiple strategies exist for dealing with diffusion and scatter from the injection site hiding the adjacent SNs in the axilla, and are described below:
1) When performing perilesional injections of radiotracer, multiple angled views and standing views, which use gravity to markedly shift the injected activity inferiorly and medially away from the axillary SN compared to the supine position, are indispensable [3,4]. In addition, the perilesional injections should be on the side of the lesion away from the axilla to further lessen diffusion of activity from the injection site from obliterating the SN [3].
2) The volume of perilesional injections should be reduced when injecting close to the axilla to lessen diffusion and inherent scatter of activity (maintain below 3 ml) [3].
3) Upwardly offset energy windows can be used in the camera to reduce "relative" scatter in the images to help better see the SNs [2-4].
4) Proper image display techniques are critical; the images presented for the first case [1] appear to display suboptimal adjustment as related to upper level, contrast/gamma/threshold. Proper adjustment parameters are crucial for the delineation of a faint SN adjacent to intense activity from the injection site [4].
5) Yet, finally, an even better solution to the complexities of upper outer quadrant lesions, is to shift part or all of the injected dose to the areolar region [3-5]. This will largely or completely resolve the issues of injected activity/diffusion/scatter hiding a closely approximated SN in the axilla [3-5]. In addition, An Adaptive Injection Technique (AIT) can be employed to allow a limited control over the number of echelon nodes realized, and improve overall SN intensity. High injection volumes (0.8–1.0 ml) administered over 1–2 minutes at the "areolar-cutaneous junction", referred to by us as LymphoBoost (LB), tend to nearly always delineate SNs (96.6%–100%) [4-6], and often multiple distant echelon nodes [3-7]. This is because these very shallow LB injections are extremely efficient at delivering activity to the SN compared to perilesional intra-parenchymal injections, and even intradermal over the lesion injections [3-7]. Compared to concurrently administered perilesional and intradermal over the lesion injections, LB areolar injections, and other areolar type injections in general, delineate essentially the same primary nodes (excluding internal mammary nodes [3,5,6]) in the vast majority of patients, and do so with several times the efficiency of getting activity into the SN [3-7]. The LB areolar injections occasionally depict additional nodes closer to the injection site, the "reverse echelon node" (REN), a feature of areolar type injections in general related to injection location [3,5,8]. They also delineate additional conventional echelon nodes compared to perilesional injections [3,5-7]. There is preliminary evidence that low injection volumes (0.1–0.3 ml) at the areolar-cutaneous junction tend to delineate fewer echelon nodes (which could be viewed as advantageous by some), and also result in fainter nodes, compared to high injection volumes (0.6–1.0 ml) [[5], (data pending publication)]. In addition, low injection volumes are also more prone to failure in delineating the SNs on LS as evidenced by a low 78.5% SN visualization rate during LS in a recent study by Rousseau et al. [9] when using very low volume 0.1 ml injections of radiotracer in a manner similar to LB. Utilizing an AIT, the initial LB injection can be performed using a low or moderate volume of injection, if less echelon nodes are desired, optionally followed by an additional LB injection with a higher volume (0.8–1.0 ml), if needed, when better SN delineation is required (more or brighter SNs). The need for a second injection is based on the viewing of sequential dynamic image sets obtained during the injection process [3-7]. Alternately, only a single high volume (0.8–1.0 ml) LB injection can be performed for the study, as a means to simplify the process.
As judiciously mentioned in the case report by Pandey et al. [1], an additional cause of non-visualization of SNs on LS is extensive replacement of macrophages in the SNs by tumor. This can lead to non-visualization of SNs or SNs that are very faint, if visualized at all [9-11]. However, utilizing the AIT described above with LB injections, will usually maximize the activity in the SNs (many times more than with isolated perilesional injections) making the missed SN on LS images from tumor infiltration (because it is very faint) less likely.
We perform over half of our studies using a two day protocol, as this allows more time for the nuclear medicine department to work with the patient, allows us to take multiple angle and standing views, perform TBM, and eliminates the rush to complete the study and print the images. The two day protocol completely avoids creating any delays in the surgical schedule related to LS imaging and TBM, when surgery is performed the morning after the prior afternoon of imaging. It also allows surgery to be performed at centers remote from the location performing LS and TBM.
The often misunderstood [12-14] and least realized benefits of LS and TBM pertains to their ultimate purposes; i.e. providing a map of the nodes so the surgeons can reduce the amount of dissection performed during SLNB, thereby reducing the morbidity of the procedure of SLNB itself. Even if successful visualization of SNs occurred in only 50% of LS studies, their worth would still be priceless to these women, as a tool to further reduce morbidity during SLNB, compared to the potentially higher levels of morbidity when using only the probe without LS or TBM [15-17].
Fortunately, 94.7%–100% of studies demonstrate SNs when using optimal techniques of areolar injection and various levels of LS image acquisition [4-6,17-21], making these underutilized tools of morbidity reduction applicable to nearly all women undergoing SLNB. This view appears to be in some ways shared by the authors in the case report by Pandey et al. [1], as a triple technique consisting of images, probe and dye is still suggested, regardless of the misfortunate LS events in the two cases presented [1].
In this light, the additional utility of standing/sitting views, with the arm out 90° to the body axis, warrants further explanation. These views are the most accurate for delineating the true number of SNs in the axilla, as they separate bunched SNs that can appear as a cluster (single focus) on supine views [3,4,22-28]. They achieve this by shifting attenuating breast tissues away from the axilla (more SN counts), allowing the camera to be closer to the SNs (better SN resolution), shifting the injected activity away from the axilla (less scatter from injection site to hide SNs) and by stretching out the lymphatic channels from their overlapped/bunched state in the spine position, to their natural drawn out state when the patient is standing (better separation between SNs) [3,4,22-28]. By providing surgeons with a more accurate number and "relative" orientation of SNs in the woman's body, they can better plan on how extensive or selective their dissections are going to be, by more accurately knowing the number of nodes to expect.
Standing/sitting views are naturally not useful for TBM, which requires the patient to closely approximate the surgical position to be meaningful, i.e. triangulation performed on supine anterior and oblique views with the arm out in the 90° surgical position. TBM has been previously described, and can help provide skin references regarding SN location for surgeons before incision [2,3,26,27]. Briefly, the technique consists of the following: with the woman supine, back flat against the imaging table with the arm in the surgical position, an anterior image is obtained and the SN is noted on the monitor and a mark (tape, pen marker, electronic) is made on the monitor screen surface over the SN seen on the monitor. A radioactive point source is brought into the field of view and moved along the woman's chest until it is positioned under the reference point on the monitor surface (the tape or marker point placed earlier on the screen surface while refreshing/updating the image), at which time the woman's body is marked with a indelible color marker. This is repeated in the 45° projection, as lateral views are blocked by the arm in the surgical position. Triangulation into the body along the anterior and 45° projection of the skin surface markings estimates SN location at the crossover point in the body [2,3,16]. The TBM technique can be useful in select patients by shorting the time of the initial probe survey during surgery. It can also help when decay has reduced the counts in the SNs when using two day protocols and in obese patients, where effective directionality and sensitivity of the probe are poor at the skin surface because of attenuation and increased distances from the SNs [2,3,7,27,28]. The technique of TBM will not accurately delineate the SN position when the breast position is different, arm position has shifted, or the patients torso is rotated differently during surgery compared to the position during chest marking during LS. These technical issues must be kept in mind when using the TBM method as an additional guide during the initial probe survey, and an agreed upon routine between surgeon and imaging specialist is required.
None of these techniques require the latest gamma camera equipment, in fact nearly all our studies are performed on cameras at least 12 years old [2-7]. What is required to obtain the best LS images and the most accurate TBM, is a full understanding of the issues and details involved in the techniques of injection, LS imaging and TBM [2-7,10,16,17,24,26-28].
Abbreviations
SN Sentinel node
LS Lymphoscintigraphy
SLNB Sentinel lymph node biopsy
TBM Triangulated body marking
AIT Adaptive injection technique
LB LymphoBoost
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
BRK Conceived the manuscript.
SCK Assisted with conception, edited manuscript.
CKK Assisted with conception, edited manuscript.
==== Refs
Pandey M Deo SVS Maharajan R Fallacies of preoperative lymphoscintigraphy in detecting sentinel node in breast cancer World J Surg Oncol 2005 3 31 15927059 10.1186/1477-7819-3-31
Krynyckyi BR Miner M Ragonese JM Firestone M Kim CK Machac J Technical aspects of performing lymphoscintigraphy: optimization of methods used to obtain images Clin Nucl Med 2000 25 978 985 11129163 10.1097/00003072-200012000-00003
Krynyckyi BR Kim CK Goyenechea MR Chan PT Zhang ZY Machac J Clinical breast lymphoscintigraphy: optimal techniques for performing studies, image atlas, and analysis of images Radiographics 2004 24 121 145 discussion 139–145 14730041
Krynyckyi BR Sata S Zolty I Kim CK Knesaurek K Reducing exposure from 57Co sources during breast lymphoscintigraphy by optimizing energy windows and other suggested enhancements of acquisition and the display of images J Nucl Med Technol 2004 32 198 205 15576341
Krynyckyi BR Kim CK Mosci K Fedorciw BJ Zhang ZY Lipszyc H Machac J Areolar-cutaneous "junction" injections to augment sentinel node count activity Clin Nucl Med 2003 28 97 107 12544125 10.1097/00003072-200302000-00002
Krynyckyi BR Chun H Kim HH Eskandar Y Kim CK Machac J Factors affecting visualization rates of internal mammary sentinel nodes during lymphoscintigraphy J Nucl Med 2003 44 1387 1393 12960181
Kim S Kim CK Krynyckyi BR Areolar-cutaneous junction injection boosts activity in sentinel node by more than 50 times compared to perilesional injection: implications for morbidity reduction Am Surg 2005 71 77 80 15757063
Roumen RM Geuskens LM Valkenburg JG In search of the true sentinel node by different injection techniques in breast cancer patients Eur J Surg Oncol 1999 25 347 351 10419702 10.1053/ejso.1999.0655
Rousseau C Classe JM Campion L Curtet C Dravet F Pioud R Sagan C Bridji B Resche I The impact of nonvisualization of sentinel nodes on lymphoscintigraphy in breast cancer Annals of Surgical Oncology 2005 12 533 538 15889212 10.1245/ASO.2005.07.014
Kim SH Shim J Kim CK Machac J Krynyckyi BR Reverse echelon node and a lymphatic ectasia in the same patient during breast lymphoscintigraphy: the importance of injection and imaging technique Br J Radiol 2004 77 1053 1056 15569651 10.1259/bjr/65044256
Goyal A Douglas-Jones AG Newcombe RG Mansel RE Effect of lymphatic tumor burden on sentinel lymph node biopsy in breast cancer Breast J 2005 11 188 194 15871704 10.1111/j.1075-122X.2005.21591.x
Burak WE JrWalker MJ Yee LD Kim JA Saha S Hinkle G Olsen JO Pozderac R Farrar WB Routine preoperative lymphoscintigraphy is not necessary prior to sentinel node biopsy for breast cancer Am J Surg 1999 177 445 449 10414690 10.1016/S0002-9610(99)00088-4
McMasters KM Wong SL Tuttle TM Carlson DJ Brown CM Dirk Noyes R Glaser RL Vennekotter DJ Turk PS Tate PS Sardi A Edwards MJ Preoperative lymphoscintigraphy for breast cancer does not improve the ability to identify axillary sentinel lymph nodes Ann Surg 2000 231 724 731 10767794 10.1097/00000658-200005000-00013
Harlow SP Is preoperative lymphoscintigraphy needed for sentinel node procedures in breast cancer? Annals of Surgical Oncology 2005 12 515 516 15947867 10.1245/ASO.2005.03.900
Kim D Kim S Kim C Machac J Krynyckyi B The effects of lymphoscintigraphy (LS) on morbidity reduction during sentinel lymph node biopsy (SLNB): A comparative study of the literature [abstract] Oral presentation at the 4th Biennial International Sentinel Node Congress: Los Angeles 2004 P7 3–6 December 2004
Kaleya RN Heckman JT Most M Zager JS Lymphatic mapping and sentinel node biopsy: a surgical perspective Semin Nucl Med 2005 35 129 134 15765375 10.1053/j.semnuclmed.2004.11.004
Mariani G Erba P Villa G Gipponi M Manca G Boni G Buffoni F Castagnola F Paganelli G Strauss HW Lymphoscintigraphic and intraoperative detection of the sentinel lymph node in breast cancer patients: the nuclear medicine perspective J Surg Oncol 2004 85 112 122 14991882 10.1002/jso.20023
Pelosi E Bello M Giors M Ala A Giani R Bussone R Bisi G Sentinel lymph node detection in patients with early-stage breast cancer: comparison of periareolar and subdermal/peritumoral injection techniques J Nucl Med 2004 45 220 225 14960639
Pelosi E Ala A Bello M Douroukas A Migliaretti G Berardengo E Varetto T Bussone R Bisi G Impact of axillary nodal metastases on lymphatic mapping and sentinel lymph node identification rate in patients with early stage breast cancer Eur J Nucl Med Mol Imaging 2005 DOI: 10.1007/s00259-005-1797-9
Maza S Thomas A Winzer KJ Huttner C Blohmer JU Hauschild M Richter M Krossin T Geworski L Zander A Guski H Munz DL Subareolar injection of technetium-99 m nanocolloid yields reliable data on the axillary lymph node tumour status in breast cancer patients with previous manipulations on the primary tumour: a prospective study of 117 patients Eur J Nucl Med Mol Imaging 2004 31 671 675 14745517 10.1007/s00259-003-1447-z
Ellis RL Seifert PJ Neal CE Pavolka KR Mann JL Malafa MP Wichterman KA Ross DS Dunnington GL Periareolar injection for localization of sentinel nodes in breast cancer patients Breast J 2004 10 94 100 15009034 10.1111/j.1075-122X.2004.21264.x
Kim SH Kim SC Kim DW Kim YJ Youssef IM Kim CK Machac J Krynyckyi BR Can different arm and body positions help in detecting more sentinel lymph nodes (SN) during lymphoscintigraphy (LS) [abstract]? J Nucl Med 2005 46 405P 15750151
Kim SH Kim SC Kim DW Machac J Kim CK Krynyckyi BR The effect of different arm positions on sentinel node localization during lymphoscintigraphy [abstract] J Nucl Med 2005 46 405P 15750151
Kim S Youssef I Kim CK Machac J Krynyckyi BR Prominent lymphatic channels simulating sentinel nodes: The utility of standing and delayed views in delineating the true number and position of nodes and the implications for further morbidity reduction Clin Nucl Med 2005
Pierini A Dworkin HJ Is the upright position more sensitive than the supine position in breast cancer sentinel node lymphoscintigraphy? Clin Nucl Med 2001 26 823 825 11564917 10.1097/00003072-200110000-00003
Krynyckyi BR Kim CK Shafir MK Mosci K Machac J Leonard M Freeman Breast cancer and its management, the utility and technique of lymphoscintigraphy Nuclear Medicine Annual 2003 131 169
Krynyckyi BR Singh G Colon D Kim CK Travis A Kim SC Machac J Letter to the Editor Eur J Surg Oncol 2005 31 805 806 15975758 10.1016/j.ejso.2005.04.015
Kim S Krynyckyi BR Fedorciw BJ Kim CK Extremely indirect dermal lymphatic channels: value of areolar-cutaneous junction injection technique for breast lymphoscintigraphy Curr Surg 2005 62 335 337 15890219 10.1016/j.cursur.2004.11.021
|
16120218
|
PMC1215530
|
CC BY
|
2021-01-04 16:39:04
|
no
|
World J Surg Oncol. 2005 Aug 24; 3:56
|
utf-8
|
World J Surg Oncol
| 2,005 |
10.1186/1477-7819-3-56
|
oa_comm
|
==== Front
Nutr JNutrition Journal1475-2891BioMed Central London 1475-2891-4-231609114610.1186/1475-2891-4-23ResearchControlled, double-blind, randomized clinical trial to evaluate the impact of fruit juice consumption on the evolution of infants with acute diarrhea Valois Sandra [email protected] Hugo [email protected] Ângela [email protected] Tereza Cristina [email protected] Carlos Maurício [email protected] Fima [email protected] Department of Pediatrics, Hospital Professor Edgar Santos, Universidade Federal Da Bahia, Salvador, Bahia, Brazil2 Department of Pediatrics, Sansum Medical Research Institute, Santa Barbara CA, USA2005 9 8 2005 4 23 23 25 5 2005 9 8 2005 Copyright © 2005 Valois et al; licensee BioMed Central Ltd.2005Valois et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
In order to assess the effects of juice feedings during acute diarrhea a double-blind, randomized study was performed in 90 children, mean age of 10 ± 4.28 months. Thirty patients with acute diarrhea were fed twice-daily 15 ml/kg of Apple Juice (AJ), 30 received White Grape Juice (WGJ), and 30 were given colored and flavored water (WA) as part of their age appropriate dietary intake. The duration and severity of diarrhea were the main endpoint variables of the study performed in a metabolic unit. The patients were similar among the 3 groups, had diarrhea for 50–64 hours prior to admission, and were dehydrated when admitted to the unit for study. Half of the patients in each group were well nourished and the others had mild to moderate degrees of malnutrition. Rotavirus infection was the agent causing the illness in 63% of the patients. The infants fed juice ingested 14–17% more calories than those given WA, (those receiving AJ and WGJ ingested 95 and 98 Calories/Kg/d respectively) whereas those receiving WA consumed 81 cal/kg/d). The increased energy intake was not at the expense of other foods or milk formula. The mean body weight gain was greater among patients receiving WGJ (+ 50.7 gm) as compared with the patients in the AJ group (+ 18.3 gm) or the patients fed WA (- 0.7 gm) (p = 0.08). The duration of the illness was longer in the infants fed juice as compared with those given WA (p = 0.006), the mean +/- SD duration in hours was 49.4 ± 32.6, 47.5 ± 38.9 and 26.5 ± 27.4 in patients fed AJ, WGJ and WA respectively. All patients improved while ingesting juice and none of them developed persistent diarrhea; most recovered within 50 hours of the beginning of treatment and less than one fourth had diarrhea longer than 96 hours in the unit. The fecal losses were also increased among the juice fed patients (p = 0.001); the mean ± SD fecal excretion in g/kg/h was 3.94 ± 2.35, 3.59 ± 2.35, and 2.19 ± 1.63 in AJ, WGJ and WA respectively. The stool output was highest during the first day of treatment among all the patients, though those fed AJ had the highest volume of fecal losses and those who received WA had the lowest stool excretion. After the first day of treatment the differences in fecal excretion were not significant. The ability to tolerate carbohydrates during the illness and immediately after recovery was similar among the 3 groups of patients. Intake of juices with different fructose/glucose ratios and osmolarities resulted in more fecal losses and more prolonged diarrhea as compared with water feedings, but the patients given juice ingested more calories and gained more weight, particularly among those being fed the juice with equimolar concentrations of fructose and glucose.
==== Body
Introduction
Worldwide, acute diarrhea represents one of the leading causes of death in children less than 5 years of age. The average duration of an acute episode of this illness is 5 days, but in some cases the diarrhea may persist more than 14 days. This leads to deterioration of the nutritional status and of the prognosis of the disease. Whereas 0.7% of acute diarrhea cases may be fatal, mortality may be 10 to 35% in the persistent diarrhea patients [1]. Previous studies have demonstrated that proper management of hydration and appropriate dietary intake during the illness can decrease the stool output, reduce the risk of prolonged diarrhea, and improve the nutritional status of patients [1-3].
The World Health Organization (WHO) has recommended home fluids, including fruit juices, as appropriate feedings and as a measure to improve fluid balance during acute diarrheal episodes in children [1]. On the other hand the Provisional Committee on Quality Improvement, Sub-committee on Acute Gastroenteritis and the Committee On Nutrition of the American Academy of Pediatrics recommended that the management of diarrheal disease in young children should include early feedings of age appropriate foods, while avoiding foods high in fat and feedings with simple sugars, including teas, juices and soft drinks [4,5]. ESPGHAN's working group on acute diarrhea endorsed similar recommendations. They recommended the use of a normal diet without restrictions, including lactose [3].
However, the recommendations for or against juice feedings during acute diarrhea have not been prospectively assessed. Previous data showed that not all fruit juices are equally absorbed. Some juices, like pear (PJ) and apple juice (AJ), which contain more fructose than glucose and sorbitol, are poorly absorbed as compared with white grape juice (WGJ) which contains equimolar concentrations of fructose and glucose, without sorbitol. We recently demonstrated that children challenged with a single serving of PJ or AJ during the recovery phase of diarrhea presented recurrence of loose stools, whereas those who received WGJ did not [6]. In other reports it has been shown that apple juice ingestion may be associated with chronic diarrhea [7].
In this double blind randomized clinical trial we measured the effects of juice intake during acute diarrheal illness. AJ, WGJ, and water feedings were given twice daily as part of an age appropriate dietary intake. While feedings of juice increased the stool loses and the duration of diarrhea, juice intake also contributed to a higher energy intake and increased weight gain, particularly among those fed the juice containing equimolar quantities of fructose and glucose.
Patients and Methods
The study was conducted in a double-blind design. There were 90 infants with severe diarrhea admitted to the Fima Lifshitz Metabolic Unit at The University Hospital Professor Edgar Santos, Federal University of Bahia, Salvador, Bahia, Brazil. Thirty patients in each group were randomly assigned to receive one of three juice feedings: apple juice (AJ), white grape juice (WGJ) or water (WA). The composition of each one of them is shown in table 1. These were packaged by the manufacturer (Welch's in Concord MA) in identical bottles and were of the same appearance and color. The WA was colored and flavored to resemble juice. Thirty identically labeled bottles containing 300 ml each were provided per patient. A new bottle was opened for each of the twice daily feedings. The investigators involved with the care of the patients in the study were not aware of the code identifying the bottle content or of the type of juice that the infant was randomized to be fed. To establish the randomization list, permuted blocks of variable length, with four blocks for each group, were used in order to avoid imbalance between treatment groups.
Table 1 Carbohydrate Content of Fruit Juices
Juice Osmolality mOsm/L Fructose gm/dl Glucose gm/dl Sucrose gm/dl Sorbitol gm/dl Energy cal/dl
Apple 700 6.2 2.7 1.2 0.5 40.4
White Grape 1040 7.5 7.1 0.0 0.0 58.4
Waterx 46 -- -- -- -- --
Modified from:
Hyams JS, Etienne NL, Leichtner AM, Theuer RC, Carbohydrate malabsorption following fruit juice ingestion in young children. Pediatrics 1988;82:64-8. Hardinge MG, Swarner JS, Crooks H. Carbohydrate in foods. J AM Diet Assoc 1965;46:197-204
xColored/flavored to resemble juice
The patients fulfilled the following inclusion criteria: male, age 4 to 18 months, had an episode of acute diarrhea (defined as more than 3 watery stools in the previous 24 hours and of no more than 3 days duration prior to admission) and were moderately dehydrated. Patients presenting severe dehydration or other conditions or concurrent serious illness, and history of chronic diarrhea as well as those exclusively breast fed prior to the time of the illness were excluded from the study. An informed consent was elicited from the mothers of all patients admitted into the protocol. The study was approved by the Institutional review Board of the Hospital and of the University.
The patients were hydrated according WHO guidelines. They were treated with ORS given at a dose of 100 ml/kg over 6 hr. The maintenance hydration phase started once the acute dehydration was treated. This was continued throughout the duration of the illness. During this phase the patients received ORS solution on a volume to weight replacement of ongoing stool loses and vomit. After rehydration was achieved, the infants were started on their usual diet that consisted of age appropriate milk formula and feedings and complementary foods [8]. Additionally, all infants received a serving of 15 ml/kg of AJ, WGJ, or WA, twice daily (10 AM and 3 PM) throughout the diarrheal episode. Plain water was offered "ad libitum" between meals. Lactose free formula was utilized in two patients who had severe stool losses >10 ml/kg/hr during their milk formula feedings.
A Breath Hydrogen Test (BH2) was performed twenty-four hours after the illness improved, defined as two formed stools passed during 24 hrs., or no stools for 12 hrs. The patients were fasted for 6 hours and a juice feeding was given. Breath Hydrogen levels were measured every 30 minutes for 3 hours using a SC Microanalyzer (Quintron Instruments Co). A peak rise in H2 of at least 20 ppm was considered as a positive response [9].
Body weight was measured on admission, after rehydration, and daily thereafter until discharge. Nutritional assessment was determined using weigh-for-length with National Center for Health Statistics (NCHS) data as reference. Nutritional intake and the amount of fluids ingested were measured throughout the study. Stool weight, Urine volume, and Vomitus weight were quantitated using metabolic techniques and specially designed beds to accurately collect stool loses throughout the study. Breast milk intake was estimated by weighing the patients before and after breastfeeding. All stools were tested for pH and sugars by Clinitest tablets. Standard laboratory techniques were used for measurement of serum sodium, potassium and hemoglobin and hematocrit on admission, at 24 hours after the initiation of the study and as clinically indicated afterwards. Stool cultures for pathogens and rotavirus by ELISA were performed on admission. All patients were requested to return after one week for outpatient clinic follow-up. At this visit clinical evolution of each patient was recorded.
The following endpoint variables were quantified: duration of the illness, severity of diarrhea (assessed by the number, type and consistency of the stools) and the amounts of fecal loses measured as g/kg/day. Vomitus losses were also quantitated. The amount of fluid intake required to maintain fluid balance was also determined. Body weight changes were measured utilizing the weight of the patient attained after rehydration as compared with the one prior to discharge. The presence of carbohydrate intolerance was determined by the fecal pH and sugar excretion as well as by the breath hydrogen levels after juice intake.
The sample size was calculated by the Power program to ensure statistically significant differences on the stool output and duration of diarrhea [10]. The estimated sample size was 26 patients per group assuming a 30% clinical improvement, on the above outcomes, at a power of 80% at 0.05 significance. Data were analyzed by Analysis of Variance (ANOVA), when data distribution was not normal, a non-parametric test of Kruskal-Wallis were performed. Survival curves Kaplan-Meyers were used when appropriate [11].
Results
The clinical characteristics and the laboratory data of the patients on admission are shown in tables 2 and 3. The patients in each of the 3 groups were similar in age, duration and severity of diarrhea, presence of fever and vomiting. Also there were no differences in the proportion of patients who received breast feedings or in their nutritional status. Over 46% of patients studied in each group, were well nourished, more than 33% showed mild body weight deficits (<1 SD) and the others had mild to moderate malnutrition (>2 SD). All patients presented some dehydration (mild to moderate), and required no intravenous hydration therapy. Differences in clinical characteristics were not significant among groups.
Table 2 Clinical Characteristics of Patients on Admission
Apple Juice n = 30 White Grape Juice n = 30 Water* n= 30
χ SD χ SD χ SD
Age (months) 10.27 4.75 10.27 4.14 11.09 4.00
Diarrhea Duration (hr) 56.37 33.90 64.17 38.95 53.47 33.35
Fever Duration (hr) 37.77 33.46 47.11 49.62 41.04 38.11
Vomiting (h) 41.57 35.45 46.97 39.54 44.33 31.44
Breastfeeding 13 (n) 43.3% 16 (n) 53.3% 09 (n) 30.0 %
Well Nourished 16 (n) 53.3% 15 (n) 50.0% 14 (n) 46.6%
Nutritional Risk** 08(n) 26.6% 10(n) 33.3% 15(n) 50.0%
Mild Malnutrition 03 (n) 10.0% 05(n) 16.6% 01(n) 3.4%
Moderate/Severe Malnutrition 03(n) 10.0% 00(n) 0.0% 00(n) 0.0%
There were no statistically significant differences among groups, Analysis of Variance (ANOVA) p > 0.05. Fever was considered above 37.5 C temperature. Nutritional risk indicated body weight deficit < 1 SD, Mild malnutrition < 2 SD and Moderate/Severe Malnutrition > 2 SD.
n = number of patients and % of patients in each category.
*Colored/flavored to resemble juice
Table 3 Laboratory Data of Patients on Admission*
Apple Juice n = 30 White Grape Juice n = 30 Water* n = 30
χ SD χ SD χ SD
Serum Sodium (mEq/l) 142.03 5.45 140.96 3.51 140.51 5.39
Serum Potassium (mEq/l) 3.99 0.59 3.95 0.81 4.27 0.78
Hematocrit (%) 31.17 3.29 30.93 2.50 32.03 3.55
Hemoglobin (g/dl) 10.28 1.13 10.24 0.87 10.58 1.19
No Anemia 06(n) 20.0 % 02(n) 6.7 % 07(n) 23.3 %
Mild Anemia † 20(n) 66.7 % 26(n) 86.6% 20(n) 66.7 %
Severe Anemia† 04(n) 13.3 % 02(n) 6.7 % 03(n) 10.0 %
Rotavirus 19(n) 63.3 % 18(n) 60.0 % 18(n) 60.0 %
Parasites‡ 05(n) 16.7%. 04(n) 13.3 % 01(n) 3.3 %
*There were no significant differences among groups by ANOVA p > 0.05.
† The criteria for mild Anemia was a hemoglobin less than 11.0 g/dl and for severe anemia was a hemoglobin less than 9 g/dl (WHO, 1989)
‡ The parasites detected were: Ascaris lumbricoides (4), Giardia lamblia (1) Blastocytes hominis (1) Entamoeba coli (1) and Cryptosporidium parvum (3).
n = number of patients and % of patients in each category
xColoured/flavoured to resemble juice
The serum electrolyte levels on admission to the hospital were similar among the 3 groups of patients (Table 3). The hemoglobin (Hb) and hematocrit levels were also similar among groups, but two thirds of the patients exhibited mild degrees of anemia (Hb < 11 g), and in 9 infants there was a more severe degree (Hb < 9 g). In all instances iron supplementation was prescribed at the completion of the study. Rotavirus was identified in the stools of 55 of the patients, in 4 there was a pathogenic Escherichia Coli, in 10 there were parasites detected, and in the remaining 21 infants there were no stool pathogens identified.
The daily intake of the patients while in the study is shown in table 4. The amount of water, milk formula, and breast milk feedings did not differ among the groups. However, the infants given WGJ readily consumed more juice than those fed AJ or WA. The total energy intake was higher in the juice fed groups as compared with the WA one. The WGJ patients ingested an average of 17% more calories on a daily basis, and the AJ infants consumed a mean of 14% more than the WA group. The increased energy intake was not at the expense of the other foods; both milk formula and complementary foods were ingested in similar quantities among the 3 groups of patients. The mean body weight gain was also higher among the juice fed patients; there was a mean weight gain of 18.3 gm in the AJ fed patients and of 50.6 gm in those given WGJ, whereas there was a mean loss of body weight (- 7.0 gm) among the WA group patients (p = 0.08).
Table 4 Daily intake of patients throughout the study (Kcal/Kg/day)
Apple Juice n = 30 White Grape Juice n = 30 Water* n = 30 P value
χ SD χ SD χ SD
Total Calories 95.84 22.42 98.65 30.52 81.43 23.09 0.02*
Milk Formula 54.49 23.43 50.21 34.57 52.68 17.93 0.81
Breast Milk 08.71 11.57 15.80 18.56 05.45 10.96 0.06
Water 30.67 9.25 30.27 09.57 26.98 9.37 0.25
ORS 45.52 31.17 39.12 25.10 25.10 17.91 0.01**
"Juices" 18.61 3.93 20.98 5.35 17.32x 4.37 0.01***
Total Liquids 157.90 38.26 158.42 50.35 127.70 28.25 0.001****
Complementary Foods 23.90 11.61 25.12 13.97 26.92 11.59 0.64
* Significant differences for Juice Groups vs Water, by ANOVA (Bonferroni)
** Significant differences for Water vs both juice groups
*** Significant differences for WGJ vs each other group
**** Significant Differences for Water vs each other group
x Colored flavored water to resemble juice
Data are means +/- SD
The duration of diarrhea differed among the 3 groups of patients (Table 5). The total duration of diarrhea from the start of the illness through their recovery was decreased among the WA fed patients as compared with the AJ and WGJ: groups 111.7 ± 48.2, 105.4 ± 44.9 and 80.0 ± 39.6 hours in AJ, WGJ and WA respectively. The differences in the duration of diarrhea were more marked while being treated in the hospital, the illness being shorter among patients given water instead of juice. However the majority of the patients recovered promptly regardless of the treatment given (Figure 1). Most of the patients improved within 50 hours after treatment was instituted, less than one forth of them had diarrhea persisting more than 96 hours and no one had it for more than 7 days.
Table 5 Duration of diarrhea in hours after randomization. Duration of diarrhea
Mean SD
Apple Juice 49.4 32.6
White Grape Juice 47.5 38.9
Water * 26.5 27.4
*Significant differences detect by Kruskal-Wallis p < 0.05. Water vs Juice groups. Data are hours +/- SD
Water coloured and flavored to resemble juice.
Figure 1 Survival analysis of total of diarrhea – Kaplan-Meier, per group. * Statistic difference were found, p < 0,05, among water group and apple juice (P = 0,03) or white grape juice (P = 0,00).
The severity of diarrhea also differed among the treatment groups (Table 6). The stool output among the WA group of patients was significantly decreased as compared with those fed juice. During the first day of treatment those fed water had a mean stool output of 40% less than those fed juice. With improvement the differences among the treatments were minimized, and were no longer significant after the 2nd day of the illness. There were also significant differences in the severity of diarrhea among AJ and WGJ feedings (Table 6). During the first 24 hours of treatment of the illness the patients receiving AJ had more marked stool losses than those fed WGJ. The mean excretion of stools was 21% higher in AJ fed patients than in the WGJ fed group.
Table 6 Fecal losses throughout the study and on the first day after randomization
Total Losses (g/kg/hr) First Day Losses g/kg/hr
Mean SD Mean SD
Apple Juice 3.94 2.35 4.13** 2.90
White Grape Juice 3.59 2.35 3.28** 2.39
Waterx 2.19* 1.63 1.78*** 1.80
*Differences Among groups Water vs each of the juice groups (Kruskal-Wallis test) (p = 0.001)
** Differences between WGJ and AJ (p = 0.02)
*** Differences between Water and each of the juice groups (p = 0.001)
xColoured/flavoured to resemble juice
The stool excretion data were treated for possible confounding and co-variables which could play a role in determining the final results. The covariance analysis and the robust regression showed a possible influence of ORS intake in the differences detected among the 3 groups of patients. The stool output was not different when the ORS intake was adjusted for the 3 groups of patients and the urinary outputs were also similar. Vomitus losses were not different among the 3 groups of patients, vomiting being present primarily during the first 24 hrs. of treatment.
There were 2 patients who had severe diarrhea during the treatment period (>10 ml/kg/hr). These patients were in the AJ group, one of these patients had acid stools and none had carbohydrates in feces. These 2 patients were treated with a lactose free formula, with improvement in stool output. There were three other patients who showed carbohydrate intolerance with acid stools or sugars in feces among the 3 groups, they improved without any dietary modifications.
Vomiting was present in all groups of patients: 22 (AJ), 26 (WGJ), and 19 (WA) during the first day of treatment; nevertheless this did not represented a limitation for Oral Rehydration Therapy (ORT) or feeding regimen and no patient was withdrawn or shifted to another therapy due to vomiting.
The response to juice feedings, as determined by breath hydrogen excretion after improvement of the illness, was not different among the 3 groups of patients. Most infants did not show breath hydrogen excretion levels above 20 ppm; and there were 17, 16 and 10 patients among the AJ, WGJ and WA groups respectively, who failed to show BH2 levels above 5 ppm at any time. Only 18 patients demonstrated a delta BH2 level increase above the basal value of more than 20 ppm. Eight of them were given AJ, 6 were fed WGJ and 4 received WA. There were 11, 16, and 13 patients in each group who exhibited a BH2 value above 20 ppm independent of the delta differences from basal among the WGJ, AJ and WA fed groups respectively. The differences in BH2 levels among groups were not significant.
Discussion
This is the first double blind prospective study designed to evaluate the effects of fruit juice feedings during diarrheal disease in young children. Two commonly available juices were selected and compared with water intake, as part of an age appropriate dietary intake. One juice contained equimolar quantities of glucose and fructose (WGJ) and the other one provided a higher fructose to glucose ratio and contained sorbitol (AJ). All patients improved while being fed water or any of these 2 juices. However those receiving juice had more stool losses than those fed water, highest being among the patients fed AJ. This was a significant finding during the first day of treatment, not thereafter. On the other hand the children fed juice ingested more calories and gained more weight than those fed water, those fed WGJ having the best response.
Acute gastroenteritis continues to be a common illness among infants and children worldwide. The disease causes an estimated 2 million deaths annually among children in the developing world. In the United States diarrhea accounts for more than 1.5 million outpatient visits, 200,000 hospitalizations and 300 deaths per year [12]. Children younger than 5 years of age are at much higher risk of death from diarrhea than older children and adults [13]. Infants younger than one year of age are particularly susceptible to this disease and are at the highest risk of death, 43% to 78% of mortalities from the illness among children less than 5 years of age occur in infants less than one year [13-15].
Although the number of children currently dying from diarrhea continues to be unacceptably high, it is substantially lower than the 5 million deaths estimated 20 years ago [16]. The critical factors accounting for the reduction in mortality rates from this illness include widespread use of oral hydration solutions and the proper nutritional rehabilitation of sick infants [17].
The American Academy of Pediatrics has emphasized the importance of oral hydration and early nutritional support to aid these patients safely and effectively through the diarrheal episode [5]. A rapid realimentation with age appropriated foods and an unrestricted diet is recommended, as soon as dehydration is corrected. Nursing should be continued for those infants being breast fed and a standard full strength formula to be given to those formula fed children. The old concept of "Bowel Rest" has no scientific validity and it can serve to aggravate and increase the risks of the disease [18]. Apart from the undesirable metabolic effects of even brief fasts, withholding oral intake may further compound the intestinal absorptive processes and may lead to deterioration of the nutritional status of the patient [19]. Even though feedings increase the stool output and diarrhea, children who are fed attain higher body weights at the end of the illness than children who are not fed. This was evident in this study. Patients who were fed juice as part of the nutritional intake during the diarrheal illness had a higher body weight at recovery than those fed water, although they exhibited larger stool losses during the first 24 hours of nutritional rehabilitation.
Diarrhea, like other infections, decreases the appetite and sick infants often reject most foods, although breast milk is better accepted [20]. The lack of appetite may be mediated by interleukin 1, a hormone released by the white cells after infection [21]. The intensity of anorexia may not necessarily correlate with the severity of the illness. A child may lose his or her appetite with even mild diarrhea, with anorexia lasting from a few hours to several days [22]. As much as 20% to 70% of food available may be wasted or not eaten, during bouts of diarrhea [22]. Thus feedings of a well accepted available energy source might be desirable and necessary to enhance the nutrient intake of sick infants and young children. The patients in this study readily consumed fruit juice, and the intake of this food did not displace the consumption of other nutrients. Juice feedings resulted in a higher energy balance, particularly among the infants fed WGJ. Infants fed juice ingested 14–17% more calories than those given WA, AJ and WGJ ingested 95 and 98 calories/kg/d respectively, whereas those receiving WA consumed 81 calories/kg/d.
Previous data showed that fruit juices differ in carbohydrate composition and that juices containing equimolar concentrations of glucose and fructose were best absorbed throughout the first 5 years of life [7,23]. Similarly we have previously shown that this type of juice is better tolerated after recovery from acute diarrhea [6]. The present study confirmed that this juice was better suited during the acute stages of the illness. The fecal losses associated with consumption of WGJ juice during the treatment of acute diarrhea were lower than those observed during feedings of juice containing higher fructose to glucose ratios and sorbitol. However, the stool output was highest only during the first day of treatment, with differences in stool output rapidly disappearing with recovery from the illness. All patients improved within 3–4 days while ingesting juice and none of them developed persistent diarrhea. The ability to tolerate carbohydrates was also similar among the 3 groups of patients.
The patients were given ad libitum up to 15 ml/kg/dose of juice twice daily throughout the study. This dose of juice exceeded the recommended allowance by the AAP-CON [4] which limits the intake of juice to one serving of 4–6 oz per day to children of this age. However by allowing the patients to ingest at will the high energy drink during the illness, they did not consume the full amount of juice offered, they only ingested approximately 17 to 21 ml/kg/day, those receiving WA consuming the lesser amounts. The patients given juice feedings also ingested more fluids. One can speculate that fluid intake was higher due to fecal losses replacement and/or thirst induced by juice. A covariance analysis and a robust regression to determine a possible confounding factor did not support that possibility. No differences were found among groups by adjusting the stool output to ORS or fluid intake. However, the ORS volume represented the most important fluid intake, indicating that diarrhea duration and stool losses were the consequence of this finding.
The maintenance of a positive energy balance during the illness may be of particular importance for the vulnerable infant who is at a higher nutritional risk even before developing diarrhea [19,22]. This illness is considered to be one of the most important risks for the development of malnutrition [24]. Diarrhea and other infections affect the body's economy through a number of mechanisms including the decreased absorption of nutrients [22,24,25]. The provision of simple carbohydrates in a balanced proportion, as present in some juices, may facilitate energy balance even during the illness and may be positive for the infant's nutritional rehabilitation [6]. The ingestion of juice during the acute episode of diarrhea provided a higher average energy intake than that of those fed WA (+ 12 cal/kg/day for AJ and 18 cal/kg/day for WGJ). However ingestion of larger amounts of fruit juice has been associated with prolongation of diarrhea [7]. Additionally ingestion of juices containing high fructose and sorbitol may also be associated with other negative consequences, i.e. colic [26] and increased energy requirements [27].
The transient malabsorption during the acute phase of the illness may be overcome by absorptive advantages of the carbohydrate composition of specific feedings [6]. Similar results were found with amino-acid based ORS [28] and with the low-osmolality ORS [16,29,30], though there was a negative role of high osmolality solutions. However the most important therapy for the sick infant is the rapid rehydration, the maintenance of fluid and electrolyte balance and the provision of adequate feedings.
Conclusion
Feedings of juices with different fructose/glucose ratio, with or without sorbitol and osmolality levels resulted in more fecal losses in the first 24 hours of diarrhea as compared with water feedings. However the patients given juice ingested more calories and gained more weight, particularly those being fed the juice with equimolar concentrations of fructose and glucose without sorbitol. All patients recovered with appropriate treatment without anyone developing persistent diarrhea. Our data strongly support the present recommendation of maintaining normal dietary habits during acute diarrheal episodes. However, juice choices may vary in its effect, as they vary in their composition; thus these differences should be considered when recommendations are made for feeding sick infants.
Acknowledgements
Supported in part by Welch's Inc. and Pediatric Sunshine Academics Inc.
==== Refs
World Health Organization Programme for the control of diarrheal diseases A manual for the treatment of diarrhoea Ver 2 1990 Washignton (DC): WHO/CDD/SER/80
Brown KB Gastanaduy AS Effect of continued oral feeding on clinical and nutritional out-comes of acute diarrhea in children J Pediatr 1998 112 191 200 3339500
The Working Center Group on Acute Diarrheoa of the European Society for Pediatric Gastroenterology, Hepatology and Nutrition. Management of acutegastroenteritis in Europe and the impact of the new recommendations a multicenter study The Working Center Group on Acute Diarrheoa of the European Society for Pediatric Gastroenterology, Hepatology and Nutrition J Pediatr Gastroenterol Nutr 2000 30 522 7 10817282 10.1097/00005176-200005000-00011
American Academy of Pediatrics Provisional Commitee on Quality Improvement. Subcommittee on Acute Gastroenteritis. Practice parameter: the management of acute gastroenteritis in young children Pediatrics 1996 97 424 36 8604285
American Academy of Pediatrics Academy endorses CDC gastroenteritis guidelines Sheth AAP News 2004 25 99
Ribeiro H Ribeiro TC Valois S Mattos A Lifshitz F Incomplete carbohydrate absorption from fruit juice consumption after acute diarrhea J Pediatr 2001 139 325 7 11487766 10.1067/mpd.2001.116279
Lifshitz F Ament ME Kleinman RE Klish W Lebenthal E Perman J Udall JN Jr Role of juice carbohydrate malabsorption in chronic nonspecific diarrhea in children J Pediatr 1992 120 825 29 1578324
OPAS/OMS – Organizacion Panamericana de La salud / Organizacion Mundial de la salud Nutrición y alimentación del niño em los primeiros años de vida 1997 Washington, DC, EUA, OMS/OPAS 163 76 1308309
Perman JB Barr RG Watkins JB Sucrose malabsorption in children: Non-invasive diagnosis by interval breath hydrogen determination J Pediatr 1978 93 17 22 650340
Erdfelder E Faul F Buchner A G Power: A general power analysis program behavior Research Methods, Instruments & Computers 1996 28 1 11
Center for Disesase Control and prevention (CDC), USA WHO Geneve, Switzerland. Epi Info 6.04d: A world processing database and statistic program for public health 2001
King CK Glass R Bresee JS Duggan C Managing acute gastroenteritis among children: oral rehydration, maintenance and nutritional therapy MMWR Recomm Rep 2003 52 1 16 14627948
Fikree FF Azam SI Berendes HW Time to focus child survival programmes on the newborn: assessment of levels and causes of infant mortality in rural Pakistan Bull World Health Organ 2002 80 271 76 12075362
Baqui AH Sabir AA Begum N Arifeen SE Mitra SN Black RE Causes of childhood deaths in Bangladesh: un update Acta Paediatr 2001 90 682 90 11440104 10.1080/080352501750258775
Shamebo D Muhe L Sandstrom A Wall S The Butajira rural health project in Ethiopia : mortality pattern of the under fives Trop Pediatr 1991 37 254 61
Choice Group Efficacy and safety of oral rehydration solution with reduced osmolarity in adults with cholera: a randomised double blind clinical trial Lancet 1999 354 296 9 10440307 10.1016/S0140-6736(98)09332-5
Lifshitz F Management of acute diarrheal disease J Pediatr 1991 118 S25 26 2007953
Bhan MK Current concepts in management of acute diarrhea Indian Pediatr 2003 40 463 76 12768058
Meier R Burri E Stemerwald M The hole of nutrition in diarrhea syndromes Curr Opin Clin Nutr Meta Core 2003 6 563 7 10.1097/00075197-200309000-00010
Bhandari N Bahl R Mazumder S Martinez J Black RE Bhan MK Effect of community-based promotion of exclusive breast-feeding on diarrhoeal illness and growth: a cluster randomized controlled trial Lancet 2003 361 1418 23 12727395 10.1016/S0140-6736(03)13134-0
Rothwell NJ Luheshi GN Interleukin 1 in the brain: biology, pathology and therapeutic target Thends Neurosci 2000 23 618 25 10.1016/S0166-2236(00)01661-1
Gracey M Nutritional effects and management of diarrhea in infants Acta Paediat Suppl 1999 88 110 26 10.1080/080352599750029835
Nobrigot T Chasalow FI Lifshitz F "Carbohydrate absorption from one serving of fruit juice in young children: age and carbohydrate composition effects" J Am Coll Nutr 1997 16 152 58 9100216
Patwari AK Diarrhoea and malnutrition interaction Indian J Pediatr 1999 66 S124 34 11132459
Molla AM Molla A Sarker SA Khanun M Whole gut transit time and its relation ship to absorption of macronutrients during diarrhoea and after recovery Scand J Gastroenterol 1983 18 537 543 6669929
Duro D Rising R Cedillo M Lifshitz F Association between infantile colic and carbohydrate malabsorption from fruit juices in infancy Pediatr 2002 109 797 805 10.1542/peds.109.5.797
Valois S Rising R Duro D Cole C Cedillo M Lifshitz F Carbohydrate malabsorption may increase daily energy requirements in infants Nutrition 2003 19 832 6 14559316 10.1016/S0899-9007(03)00160-6
Ribeiro HC JrRibeiro T Mattos A Palmeira C Fernandez D Sant'ana I Rodrigues I Bendicho MT Fontaine O Treatment of acute diarrhea with oral rehydration solutions containing glutamine J Am Coll Nutr 1994 13 215 55
Choice Group Multicentre evaluation of reduced-osmolarity oral rehydration salts solution Lancet 1995 345 282 5 7837862
Choice Group Multicenter, Randomized, Double-Blind Clinical Trial to Evaluate the Efficacy and Safety of a Reduced Osmolarity Oral Rehydration Salts Solution in Children With Acute Watery Diarrhea Pediatrics 2001 107 613 8 11335732 10.1542/peds.107.4.613
|
16091146
|
PMC1216327
|
CC BY
|
2021-01-04 16:39:31
|
no
|
Nutr J. 2005 Aug 9; 4:23
|
utf-8
|
Nutr J
| 2,005 |
10.1186/1475-2891-4-23
|
oa_comm
|
==== Front
PLoS BiolPLoS BiolpbioplosbiolPLoS Biology1544-91731545-7885Public Library of Science San Francisco, USA 1616203310.1371/journal.pbio.0030331Research ArticleBioinformatics/Computational BiologyCell BiologyDevelopmentMolecular Biology/Structural BiologyDermatologyMus (Mouse)MammalsVertebratesAnimalsMolecular Dissection of Mesenchymal–Epithelial Interactions in the Hair Follicle Molecular Signatures in SkinRendl Michael
1
Lewis Lisa
1
Fuchs Elaine [email protected]
1
1Howard Hughes Medical Institute, The Rockefeller University, New York, New York, United States of AmericaHogan Brigid Academic EditorDuke University Medical CenterUnited States of America11 2005 20 9 2005 20 9 2005 3 11 e33126 4 2005 19 7 2005 Copyright: © 2005 Rendl et al.2005This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are properly credited.
Molecular Signatures of the Developing Hair Follicle
De novo hair follicle formation in embryonic skin and new hair growth in adult skin are initiated when specialized mesenchymal dermal papilla (DP) cells send cues to multipotent epithelial stem cells. Subsequently, DP cells are enveloped by epithelial stem cell progeny and other cell types to form a niche orchestrating hair growth. Understanding the general biological principles that govern the mesenchymal–epithelial interactions within the DP niche, however, has been hampered so far by the lack of systematic approaches to dissect the complete molecular make-up of this complex tissue. Here, we take a novel multicolor labeling approach, using cell type–specific transgenic expression of red and green fluorescent proteins in combination with immunolabeling of specific antigens, to isolate pure populations of DP and four of its surrounding cell types: dermal fibroblasts, melanocytes, and two different populations of epithelial progenitors (matrix and outer root sheath cells). By defining their transcriptional profiles, we develop molecular signatures characteristic for the DP and its niche. Validating the functional importance of these signatures is a group of genes linked to hair disorders that have been largely unexplored. Additionally, the DP signature reveals novel signaling and transcription regulators that distinguish them from other cell types. The mesenchymal–epithelial signatures include key factors previously implicated in ectodermal-neural fate determination, as well as a myriad of regulators of bone morphogenetic protein signaling. These findings establish a foundation for future functional analyses of the roles of these genes in hair development. Overall, our strategy illustrates how knowledge of the genes uniquely expressed by each cell type residing in a complex niche can reveal important new insights into the biology of the tissue and its associated disease states.
Determining the molecular signature of the cells that orchestrate hair follicle growth generates new insights that will aid in understanding the normal biology and disease states of this tissue.
==== Body
Introduction
During embryogenesis, hair follicle formation is dependent upon a series of reciprocal interactions between the single-layered epithelium and a dermal cell condensate. This specialized cluster of mesenchymal cells becomes enveloped by the epithelial (matrix [Mx]) cells at the base of the developing follicle, and postnatally, they persist as the dermal papilla (DP) (Figure 1A; [1,2] ).
Figure 1 Hair Follicle Morphogenesis and Differential Expression of Lef1-RFP and K14-H2BGFP in All Cells of the Mature Follicle
(A) Schematic depicts follicle morphogenesis, which begins in waves in mouse backskin at approximately embryonic day 15 (~E15) and is complete at approximately postnatal day 4 (~P4).
(B) Transgene constructs and mice. The transgenes used for injection are shown in the stick diagrams. Nucleotide residues encompassing the gene fragments cloned are noted (+1 is the transcription initiation site). For each transgene, three transgenic lines were engineered; all were phenotypically normal. Data shown are from P4 backskin follicles of mice harboring both transgenes. Left and middle images are phase contrast and epifluorescence images, respectively, of hair follicles after dispase and collagenase treatment. Right image shows 3D reconstruction of a confocal Z-stack of a follicle showing that most of the RFP resides in the center of the follicle bulb. DP, dermal papilla; Mc, melanocytes; Md, medulla; Mx, matrix; ORS, outer root sheath.
(C) Section of K14-H2BGFP follicle, which at P4 (shown) is fully mature. Images shown are DAPI and epifluorescence channels, separately, and merged. Note the approximately 3-fold higher levels of GFP in ORS nuclei (arrows) compared to Mx nuclei and its progeny. White lines demarcate the mesenchymal–epithelial boundaries, which are separated by a basement membrane of ECM.
(D) Four-color confocal image of a section of a K14-H2BGFP and Lef1-RFP double transgenic follicle at P4. In addition to H2BGFP (green) and RFP (red) epifluorescence, the follicle was labeled with DAPI (blue) and Abs against tyrosinase (secondary Abs are against Cy5 in the far-red and white was used as a pseudocolor). Note that the anti-tyrosinase antibody labels the RFP-positive Mc and demarcates them from the RFP-positive, anti- tyrosinase negative DP. Note also that Mc are located on the epithelial side of the basement membrane, denoted by the white lines.
The architecture and biology of the mature follicle is complex (Figure 1A). At the base and in close association with the DP, Mx cells are transiently proliferative and maintain a relatively undifferentiated status. As Mx cells progress upward, they differentiate into the hair shaft (cortex and medulla) and the channel or inner root sheath (IRS) that surrounds the hair. The IRS is then encased by an outer root sheath (ORS) contiguous with the epidermis. The entire structure is enclosed by a basement membrane composed of extracellular matrix (ECM) proteins that separate the skin epithelium from the dermis and DP. A small number of follicle melanocytes (Mc) reside just above this membrane in the epithelial compartment of the hair bulb.
When Mx cells exhaust their proliferative capacity, the hair stops growing, and the lower epithelial part of the follicle enters a destructive phase (catagen). As the epithelium shrinks, the basement membrane and DP move upward. Following a resting period (telogen), epithelial stem cells (SCs) at the base of the remaining hair follicle (the bulge) receive signals from the now adjacent DP, and reenter a growth phase (anagen) to regenerate the follicle and produce a new hair.
Genetic engineering has recently enabled the isolation of epithelial SCs within the bulge [3,4]. When exposed to skin dermis, the descendants of a single epithelial SC can give rise to epidermis, follicles, and sebaceous glands, when engrafted onto the backs of Nude mice lacking hair [5]. It has long been recognized that the critical mesenchymal cells in this process are the DP [1]. In contrast to dermal skin fibroblasts (3T3 cells), which only permit epidermal repair in this assay, microdissected rat whisker DP cells induce hair growth [6,7]. In vitro, the DP cells lose this ability. Co-culturing DP cells, either with epidermal keratinocytes [8], or with embryonic fibroblasts expressing a Wnt3a but not a Sonic hedgehog (Shh) transgene [9], prolongs their potential.
Knowledge of the genes expressed by the DP and its neighbors would be of obvious value in sifting through the complex mechanisms by which DP cells maintain their remarkable inductive function while in the niche, and lose them outside of it. To date, most known DP markers have been found fortuitously. Because the DP resides in the center of a diverse cellular niche, comprised of surrounding Mx and ORS cells, Mc, and dermal sheath cells, its relative inaccessibility, coupled with its loss of potential in vitro, have posed major technical hurdles in cleanly isolating populations of these cells. Thus, although microarray and cDNA library analyses have been conducted on microdissected and/or cultured whisker DP [10–12], the array data have yielded only a handful of the known DP markers, making it difficult to evaluate the potential significance of unexpectedly expressed genes from these arrays.
Recently, it was proposed that the DP might be the origin of multipotent skin-derived precursor cells (SKPs), which are cell aggregates derived from skin cultures [13]. Interestingly, SKPs bear some resemblance to neurospheres derived from cultured neural SCs, and in vivo, a few SKP markers localize to DP. This has led to the speculation that the DP might be the residence for neural progenitor cells [13]. However, these analogies are complicated by the close proximity of Mc (neural crest derived) and DP in the follicle. Additionally, in contrast to other body sites, the head dermis develops embryologically from neural crest [14], and the parallels are drawn largely from studying rodent whiskers [13]. Thus, although the existence of a population of multipotent neuroprogenitor cells in adult follicles would place the DP squarely at the center of major clinical relevance, it remains unclear as to just how similar DP cells actually are to neuronal cells.
To probe more deeply into the special features of the DP and the nature of their cross-talk with neighboring cells, we have developed a novel strategy employing double-transgenic mice, in combination with selective cell-surface labeling to facilitate the purification of backskin DP cells and the cells surrounding their niche. By employing fluorescence activated cell sorting (FACS), we purified sufficient quantities of DP and four additional cell populations to obtain their transcriptional profiles. This has allowed us to identify the defining features that distinguish the DP cell from its neighbors, including Mc, and also the epithelial progenitors that receive cues from the DP to give rise to the differentiated cells of the hair shaft and its channel. With these molecular signatures, we have gained new insights into the DP and its microenvironment. These analyses now pave the way for future dissection of the key inductive signals produced and received by the DP that are lost upon culture in vitro. In addition, the novel multicolor labeling strategy and rigorous cross-comparisons between multiple, closely interacting cell types should have broad applicability in deciphering which genes within microarrays are likely to play key functional roles within a complex cellular niche or tissue.
Results
Isolation and Purification of DP and Four Neighboring Cell Types/Lineages
The DP cells are underrepresented dermal residents that exist as the cellular nut cloaked by a microenvironment composed of other cell types. We therefore devised a novel strategy that would enable us to use FACS to purify the DP from its complex cellular surroundings. We engineered transgenic mice expressing red fluorescent protein (RFP) under the control of a human Lef1 promoter fragment [15], and mated them to mice expressing histone H2BGFP under the control of a keratin 14 (K14) promoter [4] (Figure 1B). The K14 promoter is active only in the epithelial cells of the skin [16]. Because its activity includes epithelial SCs, H2BGFP was detected in all of the follicle epithelial nuclei. However, the promoter is most strongly active in the transiently amplifying cells of the basal epidermal and ORS layer, and correspondingly, this is where the H2BGFP was most abundant (Figures 1C and S1). By contrast, H2BGFP levels were approximately 3× reduced in Mx cells, and in differentiated Mx progeny (IRS, hair shaft, companion layer) (Figure S2). These data were consistent with the marked downregulation of K14 promoter activity upon Mx cell specification [17].
In marked contrast to the H2BGFP expression pattern, cytoplasmic RFP levels were strongest in the DP (Figure 1B and 1D). The only other location of strong RFP in the skin was in the Mc, typified by their co-expression of tyrosinase (Figure 1D) and CD117 (Kit; unpublished data). Interestingly, RFP was not found in the Mx or precortex cells where the endogenous murine Lef1 gene is normally expressed [18]. Weak RFP was sporadically found in the premedulla, which was also positive for H2BGFP (double-positive FACS population in Figure 2A).
Figure 2 Isolation and Purification of Mx, ORS, DP, DF, and Mc Populations
(A) Schematic of isolation procedure. After removing subcutaneous fat by dissection, and epidermis/upper follicle segment by enzymatic digestion, single-cell suspensions were prepared from pure dermis and subjected to three FACS schemes to purify five populations of cells: Mx, GFPlowRFP−; ORS, GFPhighRFP−; DP, RFPhighGFP−CD34−CD45−CD117−; DF, RFP−GFP−CD34−CD45−CD117−; Mc, RFPhighGFP−CD117+.
(B) Immunofluorescence analyses of FACS isolated cell populations. Frozen skin sections (hair bulb) and relevant cytospin populations were stained with Abs as color-coded and indicated. At the right of each set is quantification of percentage of cells that expressed the marker. Note: ~10% of DP and DF cells lysed on cytospin. and hence did not stain with any markers. β4, β4 integrin; Tyr, tyrosinase; Vim, vimentin; white line, basement membrane.
(C) RT-PCR: cDNA fragments were resolved by agarose gel electrophoresis, and the gene detected is denoted at left. All fragments were of the expected size. Expression of Msx2, vimentin, and β4 in multiple populations was later confirmed.
(D) Cell cycle differences in cell populations. Profiles of the five purified populations were performed by FACS. Anti-BrdU immunofluorescence is from a P4 backskin follicle from a mouse injected intraperitoneally with 50 μg/g 5-bromo-2′-deoxyuridine (BrdU) (Sigma-Aldrich) and analyzed 4 h later. Note greatest incorporation in Mx and ORS.
To isolate follicles, we first treated P4 backskins with dispase to selectively remove and discard the epidermis and uppermost parts of hair follicles, and then digested the dermal ECM with collagenase (Figure 1B). After brief trypsinization, larger debris (including hair shafts) was removed by passing the cell suspension through a cell strainer, eliminating the green fluorescent protein (GFP) positive, terminally differentiated hair cells that were still largely attached to the hair shaft. The single-cell suspension was then subjected to three different FACS isolation schemes. Channels specific for GFP and RFP were employed in various combinations with antibodies (Abs) against different cell surface markers to isolate Mx, ORS, Mc, DP, and a dermal fraction (DF) enriched in fibroblasts (Figure 2A). Prior to FACS, we stained with the cell surface marker Abs to verify that these markers were not lost by the trypsinization procedure (unpublished data).
The ORS and Mx were sorted based on their 3-fold different levels of GFP expression and absence of RFP (ORS: GFPhighRFP−, Mx: GFPlowRFP−). For DP and DF isolation, whole-cell preparations were first subjected to immunolabeling and magnetic depletion of RFP-positive Mc (CD117) and of dermal endothelial cells (CD34) and immune cells (CD45). The fractions were then sorted as the RFP-positive (DP) and -negative (DF) fractions, and further distinguished by their absence of GFP. Thus, DP cells were RFPhighGFP−CD34−CD45−CD117−, while DFs were considered as those cells that were RFP−GFP−CD34−CD45−CD117−. Finally, Mc cells were selected as the RFP- and CD117-positive population in a separate immunolabeling (Mc: RFPhighGFP−CD117+).
We judged the purity of each population by immunofluorescence microscopy and RT-PCR analyses (Figure 2B and 2C). As predicted, the putative Mx fraction showed strong labeling with Abs against proliferating nuclear antigen Ki67 and weak labeling with Abs against keratin 5 (K5) and K14 (Figure 2B). Included mRNAs in this population were Wnt10b, Msx2, and Foxn1, known to be expressed in the Mx (Figure 2C). In contrast, less than 7% of the cells in this sorted population labeled with markers characteristic of the differentiating Mx progeny (Figure S2). This was true for the hair cortex (hair keratins, AE13), the IRS/medulla (trichohyalin, AE15), and the companion layer/medulla (a K6 Ab diagnostic for these cells) (Figure S2). These data corroborated our purification strategy for the Mx. The putative ORS fraction appeared to be similarly pure and distinct from the Mx pool. Thus, the ORS population was strongly positive for K5 and β4 integrin, and displayed reduced Ki67 and Msx2 and no detectable Wnt10b (Figure 2B and 2C).
We examined the purity of the CD117, RFP-positive, and GFP-negative Mc fractions by testing for tyrosinase, Kit, and melanophilin—three key Mc markers. These markers were present in the Mc fraction, but they were not detected in the other populations (Figure 2B and 2C).
We were particularly interested in the DP and in defining its unique features that distinguish these cells from dermal fibroblasts. As expected, our DP and DF fractions were both enriched for vimentin, but alkaline phosphatase (AP) activity was strong only in the DP (Figure 2B). Nearly every cell in the DP fraction exhibited some AP activity, and ~90% displayed very strong activity (Figure 2B and Figure S3A and S3B). Some AP activity was detected in the DF, which could be due to the known presence of low AP activity in some of the non-DP dermal sheath cells, at the base of the hair follicles [19]. As judged by semi-quantitative RT-PCR, the level of AP mRNAs (Akp2) was markedly higher in the DP than the DF fractions (Figure 2C). Moreover, only the DP fraction scored strongly positive for mRNAs encoding the additional known DP markers Alx4, Noggin, and Fgf7 (Figure 2C). This preliminary molecular analysis suggests a purity of this DP fraction not achieved by previous methods [9–12].
Finally, we verified functionality of the DP fraction by culturing them in vitro for 1–2 wk, and then grafting the cultured DP cells with keratinocytes onto the backs of Nude mice. In contrast to grafting either keratinocytes alone or keratinocytes in conjunction with cultured dermal fibroblasts [7], the DP fraction produced haired skin (Figure S3C). Such characteristics have only been ascribed to DP cells or so-called dermal cup cells at the base of the hair bulb [9,19]. These functional data lend further evidence of the DP character of our population.
The cell-cycle profiles of the five populations varied in accordance with the levels of anti-Ki67 labeling and BrdU incorporation in vivo (Figure 2B and 2D). Quiescent DP and Mc populations displayed less than 1% cells in S-phase. The transiently dividing populations of Mx and ORS showed ~15% of S-phase cells. When coupled with protein and mRNA expression patterns, the specificity of cell cycle profiles further validated the purification schemes and confirmed the identity of each fraction.
Molecular Signatures of the DP and its Four Neighboring Populations
By purifying all of the cell populations within the niche of the hair bulb, we were able to obtain the transcriptional information necessary to dissect the commonalities and differences of these cell types, both at a global and at a gene-by-gene basis. For each population, purifications and microarray hybridizations (Affymetrix Moe430A) were performed in duplicate. A high level of correlation (96 ±0.7%) between replicate hybridizations (Figure 3A and Table S1) and other quality-control statistics validated our performance of microarray data generation (Table S2). Raw data and normalized microarray expression data can be accessed at the Gene Expression Omnibus (http://www.ncbi.nlm.nih.gov/geo/). A cursory examination of the overall correlation of genes present in each fraction revealed the relationship of each cell population relative to the other cell types of the niche (Figure 3A). The corresponding dendrogram allowed further visual inspection of the relations between the five different cell populations of the DP niche (Figure 3A, inset). While the correlation between replicates set the standard of a near perfect match (r > 0.96), there was a remarkably high correlation between the DP and DF, and the Mx and ORS, respectively, highlighting the common mesenchymal origin of DP and DF and the close lineage relationship of Mx and ORS. Intriguingly, the lowest correlation occurred between DP and Mx, revealing striking differences between the two populations whose signaling exchange orchestrates the dynamics of the hair growth.
Figure 3 Gene Expression Patterns of the Five Hair Follicle/Skin Populations
(A) Overall correlation of gene expression profiles between different cell types. A high correlation coefficient was observed between the DP-DF and Mx-ORS populations, consistent with their mesenchymal versus epithelial characters. Intriguingly, the least similar populations were the Mx and DP, which are strongly engaged in reciprocal exchange of signaling necessary for the maintenance of both compartments. Inset: Dendrogram outlining the correlation of gene expression between the populations.
(B,C) The Venn diagram demonstrates the degree of similarity based on absolute present calls. Probe-sets were considered present only if they were called present in both replicates. Note that for each cell type, ~200–400 probe-sets showed selective hybridization under these criteria, while ~2/3 of all present probe-sets were represented in at least four or all five of the fractions and ~4,000 genes (~6,000 probe-sets) were expressed and unchanged in all five populations (“Molecular backbone”). When all five fractions were cross-compared, ~150–300 genes (“Molecular signatures”) scored as being upregulated by ≥ 2× selectively in only cell type. The lists show array predictions for differential expression of mRNAs encoding some established, defining markers for each cell type, when compared against an adjacent or related cell population. The total number of mRNAs that show at least a 2× difference between the two cell populations is provided at the top, and the fold difference in specific mRNA levels is given next to each marker. DF, dermal fraction; DP, dermal papilla; Mc, melanocytes; Mx, matrix; ORS, outer root sheath.
We next turned to high-stringency comparative analyses to uncover their common and distinguishing features. Initial inspection of the distribution of genes called “present” irrespective of their expression levels revealed that more than two-thirds of the more than 22,000 probe-sets scored as present in at least one population (Figure 3B), with the bulk of genes being present in at least four or all five fractions. Conversely, a few hundred genes were present exclusively in one fraction, and the number of genes present in the overlap between any two fractions again highlighted the close lineage relationship of Mx and ORS, and DP and DF, respectively (Figure 3B). To ensure that we did not overlook genes that are called present in more than two fractions and yet show dramatically different expression levels (e.g., 4× present: 3× < 200, 1× > 2,000) we next performed comparative analyses, providing a more robust measure of the commonalities and differences between the five cell populations. Of the more than 9,000 probe-sets called present in all fractions, ~6,000 (4,000 genes) scored also as unchanged, comparing all five populations against each other, providing a list of putative housekeeping or “molecular backbone” genes irrespective of the lineage or cell type (Figure 3C and Table S3). By contrast, only 150–300 genes scored as upregulated by at least 2-fold in one fraction relative to the other four. In many cases, these genes were also selectively called present in only one of the fractions indicative of a specialized function. These subsets provided “molecular signatures” for each population (Figure 3C).
Each signature faithfully contained many previously assigned markers for each cell type and differentiation status [2,20]. In addition, the arrays permitted comparisons of relative expression levels of these genes in different cell compartments (Figure 3C). The mRNA level for the Mx growth factor Fgf7 was more than 16× higher in DP than DF. The mRNAs encoding known transcriptional regulators of Mx cell growth and differentiation were 4–6× higher in Mx than ORS. Conversely, mRNAs encoding ORS keratins were 3–15× higher in ORS versus Mx. mRNAs required for melanin pigment granule production were 6–14× higher in Mc than DP. Comprehensive lists of all signature genes are provided in Tables S4–S7.
The presence of the expected cell type–preferred patterns of gene expression gave us the confidence to progress to novel features of the signatures. Although we used the DF fraction for comparative purposes, we concentrated on the four populations at the base of the follicle. We grouped their signature genes and the list of common, unchanged genes (molecular backbone) into putative functional categories based upon established Gene Ontology (GO) classifications (http://www.geneontology.org/) and calculated significantly enriched categories (Figure 4A and 4B and Table S8). The common, unchanged group was largely genes encoding proteins involved in basic cellular functions, such as DNA, RNA, and protein metabolism (Figure 4A). A complete list of all GO classification of the molecular backbone is provided in Table S9. Semi-quantitative RT-PCR analyses verified that these mRNAs were expressed at similar levels across the five cell populations (Figure S4).
Figure 4 GO Analyses and Functional Grouping of the Molecular Backbone and Signatures
(A) GO analyses of ~4,000 genes present and unchanged in all five fractions irrespective of lineage or cell type. Shown is the percentage of genes in a given GO category, compared to all genes of the signature of a given cell type. Note that the genes were enriched mostly in categories involved in basic cell functions representing the molecular backbone. Asterisks denote a significant increase over a whole genome prediction. NC, not changed.
(B) GO analyses of the molecular signatures. The signature was defined as the genes whose expression was upregulated by ≥ 2× in only one of the five hair/backskin populations. Each signature was categorized into groups of genes depending upon their putative cellular functions. Shown is the percentage of genes in a given GO category, compared to all genes of the signature of a given cell type. Asterisks denote a significant increase over a whole genome prediction.
(C) The molecular signatures. The gene abbreviations and/or accession numbers are according to the NCBI listings. # denotes genes implicated in skin/hair disorders. (P) denotes genes with appreciable signal but higher levels in one of the other four populations. For multiple genes in a signature, the abbreviation is listed once, followed by –x, where x is the specific gene number.
In contrast, the differential and/or overlapping enrichment of genes in the specialized categories of the signatures provided the first insights, at a genomic level, of the functional properties of the different niche cell types (Figure 4B). Genes within the most relevant categories are listed in Figure 4C. The signatures contained many novel genes associated with signal transduction pathways of hair follicle morphogenesis, cell type−specific transcriptional regulators, cytoskeletal components, and ECM and adhesion molecules (Figure 4B and 4C). A detailed list of significant GO categories for each signature is provided in Table S8, and comprehensive signature gene lists sorted by GO classifications can be found in Tables S10–S13.
To rigorously test the degree to which the microarray data faithfully recapitulated the unique expression patterns of each cell type, we performed a series of semi-quantitative RT-PCR (Figure 5A) and real-time PCR analyses (Figure S5) across all five cell populations. For these analyses, we selected a number of mRNAs encoding signaling molecules, transcription factors, and ECM/adhesion molecules that scored as preferentially upregulated in one of the populations relative to the others (Figures 5A and Figure S5). We then contrasted the fold changes of the real-time PCRs with the actual average signal values of the microarray analyses. As shown in Figure S5, the expression patterns were remarkably similar, and often indistinguishable. These data provided a graphical illustration of the degree to which the comparative analyses based upon our microarray analyses faithfully recapitulated the differential expression patterns of the signature genes within the DP niche. Overall, these expression patterns should be helpful in future studies aimed at understanding how these genes play functional roles in hair biology. Below, we highlight some key features of the signatures.
Figure 5 Implementation of Array Analyses to Examine Characteristics and Dynamics of the Follicle DP Niche
(A) Semi-quantitative RT-PCR on mRNAs isolated from each population. Shown are representative data from molecular signature genes (see Figure 4) whose expression patterns in the DP niche environment had been previously uncharacterized. In this case, categories were consolidated into three groups: Signal transduction, transcription/nuclear, and cytoskeleton/ECM/cell adhesion. For each primer set, at least three different cycles were employed, and the resulting cDNA fragments were resolved by agarose gel electrophoresis along with DNA size markers to confirm that bands were of the expected sizes. For each gene, the data presented were from the cycle that provided the most meaningful comparisons. Note: bands seen in > 1 fraction accurately reflect mRNA expression at the differences in levels shown.
(B) Immunohistochemistry and in situ hybridizations. Skin sections were taken from 2-mo-old K14-GFPactin mice [49] whose follicles were at the transition from the resting to growing (telogen to anagen) stage of the hair cycle (8 wk) or from P4 WT mice (full anagen follicles) (all others). Sections were subjected to either immunofluorescence using color-coded Abs as indicated or in situ hybridization using the indicated biotinylated cRNA probes (sense controls were negative).
The ORS and Mx Signatures
The ORS signature included genes encoding a complex array of largely unstudied putative skin transcription factors. This list contained known (Bnc, Ets2, Tcf3, Egr2/Krox-20, hairless [Hr], and vitamin D receptor [Vdr]), as well as previously unrecognized ORS transcription factors (Figures 4B, 5A, and Figure S5). The signature was further distinguished by focal adhesion and ECM genes, reflecting an ability of ORS cells to not only to adhere to, but also synthesize and remodel their adjacent basement membrane. Since ECM is composed of signaling molecules, the upregulation of these genes further suggested a possible feedback loop to reinforce cell-substratum contacts in the ORS.
In contrast to ORS, Mx cells are typified by their ability to respond to cues from their microenvironment and differentiate upward to form the six concentric rings of the hair follicle. The Mx signature revealed their status at the nexus of proliferation and differentiation (Figures 4B and S5). In addition to established Mx transcription factors (Msx2, Msx1, Ovol1, Hoxc13, Dlx3, Foxn1, Hr, Lef1, and Ap2), the signature included several forkhead cousins of Foxn1 (Nude mouse), one of which (Foxq1) has been linked to the Satin mutant mouse, defective in hair shaft differentiation [21]. Also on this list were Gcl (germ cell-less) and Tcfcp212 (grainyhead-like1), thought to function in early SC differentiation and/or lineage boundaries. The Mx signature also revealed many genes encoding members of the Fgf, Wnt, Tgfβ, Tgfα, Shh, and Bmp signal transduction pathways (Figure 4C). This was in good agreement with the established ability of Mx to orchestrate signal transduction pathways and specify the hair shaft and its channel. Additionally, the signature included genes encoding keratins and other structural proteins. In part, this could reflect early steps in lineage differentiation. However, for at least three structural genes, it is noteworthy that (a) keratin c29 is highly homologous to K17, whose absence causes premature Mx apoptosis and alopecia in mice [22]; (b) skin lacking Cldn1 (claudin 1) displays abnormally short hairs [23]; and (c) Gsdm (gasdermin) mutations have recently been linked to alopecia in mice [24].
The DP Signature: Insights into Mesenchymal–Epithelial Cross-Talk
A goal of this study was to identify novel features of the DP that might give us insights into understanding how these cells exert their power over epithelial SCs and their ORS and Mx progeny. By comparing against DF, we screened out general fibroblast features, e.g., expression of type I and type III procollagen chains, vimentin, and TGFβ1-induced genes. By contrasting the DP with ORS and Mx signatures, we could identify genes exclusively expressed in either compartment and begin to make predictions regarding the epithelial–mesenchymal cross-talk that transpires in the hair bulb.
The purity of our DP cells yielded an unprecedented sensitivity of detection. Of approximately 30 genes reported to be expressed in DP in vivo [25], 24 were either in our DP signature or expressed in DP but more abundant in one or more of the other populations (Figure 4C). By contrast, only three of these genes had appeared on the prior array list from microdissected DP [11], and only five were on the list of 309 expressed genes from cultured DP [12]. Most of the ~180 genes in our DP signature encoded novel factors involved in transcription, cell communication, and signaling (Figure 4C). Less than 5% of our DP signature genes appeared on the previously published arrays of microdissected whisker DP in vivo [11] or in vitro [12].
Given the near complete lack of overlap between our DP signature and prior published reports, it was important to verify the novel aspects of each signature, as we had already done for the well-established features. Semi-quantitative PCR confirmed that the majority of genes were expressed predominantly by only a single cell population, i.e., the hallmark of our signature lists (Figure 5A). The few exceptions were readily explained upon inspection of the gene expression profiles across the five populations. For example, Fst (follistatin) and Sostdc1 (ectodin/wise) scored as ~3× higher in DP than in ORS, but 3–30× higher in ORS than in the other three fractions. Analogously, Wnt5a and the Gata 3-like factor Trps1 (tricho-rhino-phalangeal syndrome1) scored as ~2–6× higher in DP than in Mx, respectively, but ~1.5–10× higher in Mx than in other fractions. Real-time PCR further documented the accuracy of the DP signature (Figure S5).
Finally, we showed that expression of our DP signature genes can be detected in highly enriched pelage follicle preparations (see below). For a number of novel DP genes, we also used in situ hybridization and immunofluorescence to verify mRNA expression patterns and extend our findings to the protein level (Figure 5B). That our DP signature bears strong resemblance to the list of known DP genes and bears little resemblance to previously published profiles of DP cells emphasizes the importance of conducting array analyses on purified populations of skin DP cells. The PCR, in situ hybridizations, and immunofluorescence data offer compelling evidence to attest to the faithfulness and reliability of our signatures, and provide the first clear view of the DP and its niche microenvironment.
Functional Links Between Array Data, Human/Mouse Genetic Disorders of Hair and Skin, and Epithelial–Mesenchymal Interactions in the Hair Follicle
Our array comparisons provided us with the confidence to probe more deeply into the physiological relevance of the signature lists. One of the most interesting and striking features of our array comparisons was the large number of signature genes that are associated with different genetic disorders of the hair. Denoted by a “#” in the signature lists of Figure 4C, these genes included (a) the Mx signature genes Psors1c2, Tacstd2, Notch1, Msx2, Msx1, Hoxc13, Dlx3, Foxq1, Tcfcp212, Trps1, Hr, Cldn1, and Gsdm; (b) the ORS signature genes Pthlh, Vdr, Egr2, Hr, Krt1–15, Krt1–14, Krt2–5, Col17a1, Col4a5, Lamb3, Lama5, Lama3, Itgb6, Itgb4, Itga3, and Dst; and (c) the DP signature genes Ptch, Pthr1, Fgfr1, Pdgfra, Bmp4, Fst, Nog, Tgfbr1, Trps1, Sox18, and Inhba. In addition, the hair disorder–associated genes included several genes, e.g., Kitl, Lef1, Hr, and Gli2, which were featured prominently in the arrays, but which were expressed at relatively high levels in more than one of the five cell populations, thus excluding them from the signature lists. Real-time PCR was used to confirm the expression patterns of these functionally important signature genes (Figure 6).
Figure 6 Detailed Expression Analysis of Hair/Skin Disease Genes Found in the Molecular Signatures of the Epithelial and Mesenchymal Populations of the Hair Bulb
Real-time PCR confirmation of 24 different signature genes of the Mx, ORS, or DP, which have previously been implicated in genetic disorders of the hair. Many of these genes have not been well-studied at the level of expression and function. In each case, the highest level of mRNA expressed corresponded to the cell population in which the signature gene appeared. Moreover, in cases where more than one cell population showed appreciable mRNA levels, this was also reflected in our microarray comparisons.
Even though these genes have been previously genetically linked to hair/skin disorders, only a few have been well-studied at the level of expression and function. We were particularly intrigued by DP signature genes such as Trps1, Sox18, Fst, and activinβ-A (Inhba), whose roles in hair follicle morphogenesis have remained poorly understood [26–29]. Of additional note was the DP signature gene Fgf10, recently shown to be required for embryonic whisker development [30]. Fgf10 and Fgf7 bind to the same receptor (encoded by Fgfr2 and in the Mx signature), and Fgf10′s presence in the DP signature explains why Fgf7 knockout mice display a milder hair phenotype than the conditional Fgfr2 knockout [31,32].
Further insights into the DP-Mx cross-talk came from evaluating the distribution of Shh pathway members. Whereas Shh is expressed by Mx, Shh receptor and downstream effector genes were part of DP's signature (Figure 4C). Additionally, mRNA encoding Hhip (hedgehog-interacting protein) was more than 80× higher in DP than Mx (Figures 5A and S5). By in situ hybridization and anti-Hhip immunofluorescence, we detected Hhip at the early stages of follicle downgrowth (Figure 5B). This was intriguing since in lung development, Shh signaling through Patched can accentuate Hhip expression, making the extending lung bud tip refractory to Shh signaling and permissive for Fgf10 expression [33]. Moreover, Fgf10 is known to be negatively regulated by Shh, and conversely, both mesenchymal Fgf10, and also the BMP inhibitor Noggin, can enhance epithelial Shh expression [34,35]. When taken together, our findings suggest a regulatory circuitry for sustaining expression of Fgf10/7 in Hhip-positive DP and permitting Shh in Mx. Since excess Shh would be expected to override the effects of Hhip and downregulate Fgf10 and Fgf7, this may also explain why Shh treatment per se did not maintain the inductive ability of cultured DP [9].
Given the reported effects of Wnts on the maintenance of DP potential [9] and the presence of Wnt5a in embryonic dermal condensates [36], it was interesting that Wnt5a, previously reported in postnatal hair follicles [36], was in the DP signature (Figure 4C). Equally intriguing was the presence of DP signature genes encoding both secreted Wnt inhibitors (Wif1, Sfrp2, and Frzb), as well as possible Wnt effectors. Semi-quantitative RT-PCR and real-time PCR (Figures 5A and S5), as well as anti-Wif1 immunofluorescence supported these observations (Figure 5B). Like Hhip, Wif1 expression was maintained in adult DP and present at different stages of the hair cycle.
Of all the novelties in the DP signature, we were particularly struck by the number of Bmp pathway members whose mRNA expression levels were upregulated by at least 2× in DP. Bmp4 has already been implicated in the cross-talk that specifies hair differentiation [37]. However, Bmp6 was particularly notable in that its mRNA levels were more than 10× higher in DP than the four other populations, a feature confirmed by in situ hybridization (Figure 5A and 5B, and Figure S5). All the cells within the hair bulb, including the DP, expressed the requisite BMP receptor (Bmpr1a). This said, the DP signature included a surprising number of genes encoding BMP inhibitors, including Noggin, Gdf10, Sostdc1/Ectodin/Wise [11], Prdc (protein related to Dan/Cerberus), and Bambi. Of these, only Noggin has previously been documented as a functionally important BMP inhibitor in the DP [34,38]. The preponderance of BMPs/BMP inhibitors in the DP signature suggested a greater importance of the BMP pathway in promoting DP character than had been previously appreciated.
The DP Signature: Relation between Neural SCs and DP Cells
Recently, it was reported that skin cultures contain neurosphere-like structures that can be induced to form neurons and glial cells [13,39]. Although prior array data on dissected whisker DP, and their cultures, showed no resemblance of DP to neurally derived cells [11,12], several markers expressed by the skin-derived neurospheres were traced by in situ hybridization to whisker follicles [13]. The relative lack of resemblance between these prior whisker “DP” screens and our signature containing bona fide DP markers offered a possible explanation for these discrepancies. However, since some of head mesenchyme is known to be derived from neural crest [40], a documented resemblance between whisker DP and neural progenitor cells would still not be definitive. Our array data allowed us to address more important and as yet unexplored questions: (1) Do SKPs and/or neural progenitors share similarities with DP from skin whose mesenchyme is not derived from neural crest? and (2) How does DP character compare to that of neural progenitors, nearby Mc (of known neural crest origin), and dermal fibroblasts (derived from dermamyotome)?
We first addressed the relation between DP and SKPs cultured from skin dermis [13]. Only five genes, Snai2 (slug), Twist1, Cspg2 (versican), Nexin1, and Ncam1, have been reported to be expressed in both SKPs and backskin follicles [9,13,41,42]. Four of these genes appeared on our DP signature (Figure 4C). Of the remaining known SKP-expressed genes (Shox2, Pax3, Snail1, Sox9, Nestin, Wnt-1, Sca-1/Ly6A-E, Twist2, and Fn1) [13], only Shox2 was in the DP signature, and only Fn1 scored as present in DP. Conversely, Sox2 and Ngfr (p75) were readily detected in pelage DP (Figure 7) and yet they were reported as absent in SKPs [13].
Figure 7 Neuronal and Neural Crest Genes Expressed by the DP and Hair Follicles
(A) Semi-quantitative RT-PCRs were conducted as in the legend to Figure 5, except in this case, we used oligonucleotides against neuronal and/or neural crest expressed genes. In all cases, these genes either appeared on the DP molecular signature or scored as expressed by the DP as well as one or more of the other four skin populations. (see Figure 4) Shown are representative RT-PCR data, which show an excellent correlation with the DP-preferred expression pattern of the majority of these neural genes.
(B) Immunohistochemistry and in situ hybridizations of neuronal/neural crest genes in skin. Sections of P4 backskins were subjected to either immunofluorescence using color-coded Abs as indicated or in situ hybridization using the indicated biotinylated cRNA probes (sense controls were negative). Merged images of serial sections were used to compare Ab (red) and in situ (pseudogreen) patterns. Gfra1, glial derived neurotrophic factor receptor1; Mdk, midkine; Prss12, serine protease neurotrypsin; Tyr, tyrosinase.
(C) Detection of neuronal/neural crest genes in highly enriched hair follicle preparations. Highly enriched follicle preparations were isolated by serial low-speed centrifugation following dispase and collagenase digestion of P4 backskins. After isolation and preparation of their mRNAs, semi-quantitative RT-PCR was conducted using oligonucleotides to those neuronal and neural crest markers that were found in the DP signature. As controls, oligonucleotides were used against Akp2, Alx4, Bmp6, and Fgf7, which are all markers that we mapped to DP by in situ hybridizations and/or immunofluorescence (see Figure 5). Note that the neuronal/neural crest genes appearing on the DP signature showed comparable signals to the documented DP genes.
Although differences between SKP cultures and in vivo DP expression patterns had escaped prior notice, such differences could nevertheless exist because SKPs are derived from cultures rather than a purified in vivo cell population. We therefore turned to addressing the broader relation between DP and neural SCs. In this regard, it was notable that Zic1, Zic3, and Sox2 were all part of the DP signature and absent in Mc. These mRNAs encode key transcription factors that specify neuronal fate at the expense of ectoderm [43,44]. The signature also included about ten other neural genes (Figure 4C).
We confirmed the preferred expression of these genes in DP by using semi-quantitative RT-PCR. As shown in Figure 7A, most genes were preferentially upregulated in the DP fraction relative to all of the other fractions, including Mc. An exception was Sox10, whose expression by array analyses and by RT-PCR scored as preferentially expressed in Mc. Also confirming the array data were our RT-PCR analyses of Sox9, which scored as preferentially expressed in the ORS, and Wnt5a, which scored as present in Mx and DF populations as well as in the DP (Figure 7A). We also confirmed DP localization of Prss12 (serine protease neurotrypsin), Gfra1 (glial derived neurotrophic factor receptor1), and Mdk (midkine) by in situ hybridization (Figure 7B). Co-labeling with anti-tyrosinase (Mc-specific) verified that the hybridization was in the DP and not Mc compartment. In addition, we verified the expression of these and additional neuronal/neural crest–related DP signature genes in highly purified follicle preparations (Figure 7C). Given the in vivo expression of neuronal/neural crest–related mRNAs in the DP, we could not attribute the unusual expression patterns to the presence of minor neural contaminant(s), e.g., Schwann cells, which are likely to ensheath the sensory nerve endings within the skin [45]. Rather, the in situ and immunofluorescence patterns of the neuronal component of the DP signature (Figure 7) showed a good correlation with the physical location of the DP, as did the in vivo expression pattern of the DP signaling and follicle disease genes.
Despite the alluring parallels between backskin DP and cells of neural origin, the DP signature did not strongly resemble neural crest, neural SCs, or any of the neural lineages described to date, including Mc. Additionally and equally surprising was the degree to which the DP and dermal fibroblast signatures were distinct, as the DF signature did not display these neuronal-like parallels, nor did they exhibit the bank of hair disease genes or signaling genes seen in the DP signature. Taken together, our data point to a signature unique to the DP and not shared by any of the cell populations constituting the distinctive DP niche microenvironment.
Discussion
The potent inductive ability of DP to promote follicle formation has been recognized for decades [1]. However, their minority status, coupled with their rapid loss of potential in vitro, has left their molecular nature elusive. By devising a strategy to obtain pure DP and their neighboring cells, we were able to overcome this hurdle and determine a molecular signature for DP. By comparing expressed DP genes to those of DFs, Mc, and neighboring follicle epithelial cells, we could selectively hone in on similarities and weed out genes expressed by DP but not preferentially relative to their neighbors. Interestingly, and unexpectedly, the DP signature was divergent from all cell categories to which parallels had been drawn previously.
Our DP signature contained most of the established DP markers. This was important, since no other published DP screen to date has provided a signature that accurately reflects the known DP expression program [11,12]. The failure of prior arrays to include most markers documented by in situ hybridizations and/or immunofluorescence is most likely attributable to the difficulties in purifying DP from the complex milieu of its surroundings and from the rapid loss of DP character that is known to happen when DP cells are taken out of their native niche and placed in culture. In addition, our Mx signature contained many of the established Mx markers and gave us the first glimpse at a global array profile enriched for this compartment of cells. Although the Mx itself is likely to be a mixture of early progenitor cells for all the differentiation lineages of the hair follicle, this total Mx profile will nevertheless be valuable in discerning how these cells diverge as they maintain their contact with the DP.
The ability to produce arrays that faithfully recapitulate the established programs of the DP and the Mx enabled us to capture new insights into the fascinating properties of these specialized cells and their potential for intercellular cross-talk. Amongst the most interesting features is the marked number of known hair disorder genes expressed by each of these cell types. Most of these genes, e.g., Noggin [37] and Trps1 [26], were linked to hair diseases only within the past decade, concomitant with advances in positional cloning and mouse genetics. However, there are many more spontaneous and chemically induced mutations that have yet to be mapped and that involve hair phenotypes. Our arrays will be beneficial in accelerating the rate at which these diseases are mapped in the future.
As importantly as the contributions that these gene expression patterns make to establishing links between hair/skin genetic disorders and genes are the contributions that they make to our understanding of the underlying biology. Even for those cases where links between a disease gene and a hair disorder have been established, there is often little or no reliable data available for which of the cells of the hair follicle express the gene or how defects in the gene cause the morphological defects associated with the disease state. Examples in point are Pthr1, Pdgfra, Tgfbr1, Egr2, androgen receptor,and Sox18, which have all been implicated previously in hair disorders, but not recognized as genes that are preferentially expressed in a particular follicular compartment. Knowledge of the genes that are preferentially expressed by the epithelial and mesenchymal cells of the hair follicle provides a framework for functional studies to probe more deeply into the comprehensive biology involved. In this regard, disease genes are obvious candidates for governing the maintenance and character of a particular cell type, and their prevalence in our arrays provides perhaps the best evidence that these arrays are functionally significant.
The selective presence of known DP and hair disease genes in our DP signature gave us confidence in utilizing the list to better understand DP character and how it differs from dermal fibroblasts in stimulating the Mx cells of the hair follicle to differentiate. We were especially intrigued by a resemblance between the molecular differences in Mx and DP arrays versus those that are seen when neural and non-neural ectoderm segregate during embryonic neural induction [44]. During embryogenesis, when epidermal and neural lineages diverge, the epidermal lineages are determined by Msx1, Dlx3/5, and Ap2 transcription factors, which subsequently control keratin gene expression, whereas the neural lineages are determined by Sox2, Zic1, and Zic3 transcription factors, which subsequently lead to Ncam1 and neural tubulin expression [44]. It is striking that in postnatal skin, the Mx signature of the hair follicle bulb is marked by the presence of Msx1/2, Dlx3/Dlx2, Ap2, and keratin genes, while the DP signature features Sox2, Zic1, Zic3, and Ncam1. In the future, it will be interesting to pursue the parallels between mesenchymal–epithelial cross-talk in the hair follicle and neuronal-epidermal cross-talk in embryonic development.
Our comparative analyses will also be valuable in the quest to realize the clinical potential for which the DP is known, namely for its inductive powers in hair growth. In this regard, we have uncovered a number of special features of the hair bulb microenvironment that provide tantalizing clues as to how DP cells exert their inductive powers. Among them are BMPs and BMP inhibitors, Shh inhibitory proteins, and Wnt signaling molecules. By comparing how the molecular signature changes when DP cells are removed from their niche and placed in culture, it should be possible to identify those genes whose expression is intrinsic to DP, and those whose expression is lost upon culture. Conversely, the constellation of molecular signature factors that are secreted by the native DP niche should then pave the avenue to employ a systematic approach to define the factors essential for maintenance of both the DP and the Mx signatures. A combinatorial effect will likely be necessary to fully unleash the full inductive potential of cultured DP, and knowing which surface receptors are expressed by DP and by Mx should expedite identification of these conditions.
In conclusion, our novel multicolor labeling approach demonstrates how array comparisons of highly purified cell populations can be employed to dissect the molecular make-up and cellular interactions within a complex microenvironment. Although we applied this strategy to unravel the molecular repertoire of closely interacting cell types within the hair follicle, the strategy is likely to be broadly applicable to a number of complex model systems.
Materials and Methods
Mice, cell isolation, FACS, and engraftments
For Lef1-RFP transgenic mice, a 6,713-basepair XbaI/NotI fragment of the human Lef1 promoter/5′UTR was cloned from a BAC (bacterial artificial chromosome) clone and assembled with RFP (DsRed-T1, a kind gift from B. Glick, University of Chicago, Illinois, United States). The K14-H2BGFP mice were previously generated in the lab [4]. Five backskins from P4 K14-H2BGFP/Lef1-RFP double transgenic mice were treated with dispase 4 °C, 8 h to separate epidermis/upper follicles from dermis. Dermis was digested with 0.2% collagenase at 37 °C, 40 min. Intact follicles and dermal cells were sedimented at 300 × g; follicles were obtained at 20 × g. Following trypsinization, 37 °C, 5 min, cell suspensions were strained. ORS and Mx cells were selected by FACS as GFPhighRFP− or GFPlowRFP− cells, respectively. DP were obtained after first depleting Mc (CD117+), lymphocytes (CD45+), and endothelial cells (CD34+) with biotinylated Abs (BD Pharmingen, San Diego, California, United States)/magnetic anti-biotin microbeads (Miltenyi Biotec, Bergish Gladbach, Germany), and then selecting for RFPhighGFP− cells in the FACS. The DF enriched in fibroblasts was the RFP−GFP−CD34−CD45−CD117− population. For Mc isolation, cells were incubated with CD117-biotin followed by staining with streptavidin-APC (1:200, BD Pharmingen). Mc were purified by selecting RFPhighCD117+ cells. Cells were stained, washed, and sorted in PBS/5% FCS. For dead cell exclusion, 300 ng/ml propidium iodide were added before FACS.
Cell isolations were performed on a FACSVantage SE system equipped with FACS DiVa software (BD Biosciences, Franklin Lakes, New Jersey, United States). Gates for fluorescence fractionation were set to match those approximated by semi-quantitative immunofluorescence analyses of the cell compartments. Cells were gated for single events and viability, then sorted. Cell purity was determined by postsort FACS analysis and typically was > 95%. For immunofluorescence characterization, cells were cytospun with a Cytospin4 unit (Thermo/Shandon, Pittsburgh, Pennsylvania, United States).
Engraftments were performed as described [7,9]. Experiments included a positive control of cell suspensions from freshly isolated WT dermis plus keratinocytes and a negative control of keratinocytes alone. Freshly isolated newborn keratinocytes (5–10 × 106) and DP cells (2–4 × 106) in first passage (1–2 wk of culture) were used for grafts. Hair typically appeared after 17–24 d.
RNA isolation and microarray analyses
Total RNAs from FACS cells were purified using the Absolutely RNA Microprep kit (Stratagene, La Jolla, California, United States), and fluorometrically quantified (Ribogreen, Molecular Probes, Eugene, Oregon, United States). Quality was assessed by RNA 6000 Pico Assay (Agilent Technologies, Palo Alto, California, United States), and 800 ng were primed with oligo(dT)-T7 primer and reverse transcribed (Superscript III cDNA synthesis kit; Invitrogen, Carlsbad, California, United States). One round of amplification/labeling was performed to obtain biotinylated cRNA (MessageAmp aRNA kit; Ambion, Austin, Texas, United States), and 10 μg labeled cRNA was hybridized 45 °C, 16 h to mouse genome array MOE430a (Affymetrix, Santa Clara, California, United States). Processed chips were read by an argon-ion laser confocal scanner (Genomics Core Facility, Sloan Kettering Cancer Center, New York, New York, United States). Two entirely independent datasets were obtained for the five cell populations.
Scanned microarray images were imported into Gene Chip Operating Software (GCOS, Affymetrix) to generate signal values and absent/present calls for each probe-set using the MAS 5.0 statistical expression algorithm (.chp files). Each array was scaled to a target signal of 500 using all probe-sets and default analysis parameters. For comparisons, raw data and .chp files were imported into GeneTraffic 3.8 (Iobion Informatics, La Jolla, California, United States), and replicate microarrays were grouped and compared using the Robust Multi-Chip Analysis algorithm. Gene lists were compiled containing probe-sets > 2-fold increased for one over the four other populations. Probe-sets that scored as increased, but called absent were eliminated. Genes were grouped functionally by uploading probe-set lists to the “Database for Annotation, Visualization and Integrated Discovery” (DAVID 1.0) Web tool (http://apps1.niaid.nih.gov/david/upload.asp [46]).
Semi-quantitative RT-PCR and real-time PCR
Total RNAs were purified from FACS-sorted cells as above, and after quantification with Ribogreen (Molecular Probes), normalized RNA quantities were reverse transcribed (Superscript III First-Strand Synthesis System, Invitrogen) using oligo(dT) primers. cDNAs were adjusted to equal levels by PCR amplification with primers to Gapdh. PCR amplification of genes of interest was performed using primers designed within the target sequence of the microarray probe-sets where possible, ensuring the uniqueness of the primers and the amplicon. All > 50 primer pairs were designed to work at the same settings: 3 min at 94 °C initial denaturing, 26–35 cycles of 15 s at 94 °C denaturing, 30 s at 60 °C annealing, and 25 s at 72 °C extension. For a list of primers used, see Table S14. Amplifications with minus reverse transcriptase control cDNAs yielded no products for any of the primer pairs at the cycles tested. For real-time PCR, the same primers were employed using the LightCycler System (Roche, Basel, Switzerland), LightCycler 3.5 software and the LightCycler DNA Master SYBR Green I reagents. Differences between samples and controls were calculated based on the 2−ΔΔCP method.
Cell culture
Viability of FACS-isolated DP cells was assessed by Trypan Blue (Sigma, St. Louis, Missouri, United States) staining, and equal numbers of live cells (5,000/cm2) were plated in Amniomax C-100 medium (Invitrogen), previously used for dog whisker DP cells [47]. This was the best of the five media tested.
Immunofluorescence and in situ hybridizations
Lef1-RFP positive tissues were first fixed in a 37% formaldehyde buffer for 2–3 h and then embedded in OCT and frozen. All other tissues were immediately embedded in OCT, frozen, and sectioned. Paraformaldehyde-fixed sections and cytospin preparations were subjected to immunofluorescence or in situ hybridizations essentially as described [48,49]. When applicable, the MOM basic kit (Vector Laboratories, Burlingame, California, United States) was used to prevent non-specific binding of mouse monoclonal Abs. Abs and dilutions used were: AE13 (mouse, 1:50, [50]), AE15 (mouse, 1:50, [51]), Alx4 (mouse, 1:100, Exalpha, Maynard, Massachusetts, United States), BrdU (rat, 1:200, Abcam, Cambridge, Massachusetts, United States), CD104 (rat, 1:100, BD Pharmingen), Hhip (goat, 1:200, R&D Systems, Minneapolis, Minnesota, United States), Ki67 (rabbit, 1:500, Novocastra, Newcastle upon Tyne, United Kingdom), K5 (rabbit, 1:5,000, Fuchs Lab), K6 (rabbit, 1:1000, Fuchs Lab), Tyrosinase (rabbit, 1:500, kind gift from VJ Hearing), Vimentin (rabbit, 1:500, Biomeda, Foster City, California, United States), Wif1 (goat, 1:200, R&D Systems), HoxA9 (rabbit, 1:200, R&D Systems), p75 (rabbit, 1:100, Oncogene Science, Cambridge, Massachusetts, United States), Ncam1 (rat, 1:100, Chemicon International, Temecula, California, United States),. FITC, TexasRed, or Cy5 conjugated anti-mouse, -rat, -rabbit, or anti-goat Abs (1:200, Jackson Laboratories, Bar Harbor, Maine, United States) were used as secondary Abs. For detection of AP activity, the substrate 4-Nitroblue tetrazolium chloride/5-bromo-4-chloro-3-indolyl-phosphate 4-toluidine (NBT/BCIP, Roche) was used as recommended by the manufacturer's instructions. The following probes for in situ hybridizations were generated from IMAGE cDNA clones (IMAGE consortium,
ATCC) using the DIG RNA labeling kit (SP6/T7) (Roche): Bmp6 (Image:2779955), Gfra1 (Image:6390018), Hhip (Image:6402422), Mdk (Image:4167496), and Prss12 (Image:3665834). Nuclei were stained using 4-diamidino-2-phenylindole (DAPI). Imaging was performed using an Axioskop and Axiophot microscope (Carl Zeiss, Thornwood, New York, United States) equipped with Spot RT (Diagnostic Instruments, Sterling Heights, Michigan, United States) and Axiocam (Zeiss) digital cameras, respectively, or with an LSM 510 confocal microscope (Zeiss).
Supporting Information
Figure S1 GFP High and Low Levels in Mature Anagen Phase Hair Follicles
Two full-length hair follicles are shown, labeled with DAPI to mark the nuclei, anti-keratin 5 (red) and epifluorescence to show H2BGFP (green). Note that there is an approximately 3-fold greater level of GFP fluorescence in the ORS versus the Mx. Also note that every ORS cell along the length of the hair follicle co-expresses cytoplasmic K5 and nuclear H2BGFP, but not every ORS nucleus is present in each frozen section.
(5.8 MB TIF).
Click here for additional data file.
Figure S2 Minimal Contribution of Early Differentiating Cells to the Undifferentiated Mx Fraction
To assess the contribution of terminally differentiating hair shaft cells to the undifferentiated Mx fraction, FACS-isolated cell populations were analyzed by immunofluorescence. Frozen skin sections (hair bulb) and cytospin populations were stained with Abs for AE13, AE15, and Keratin-6 (K6), which are expressed in the (pre-) cortex, IRS/medulla, and medulla/companion layer, respectively. DAPI (blue) and H2BGFP (green) are also shown in each immunofluorescence image. Cytospin quantifications (right) show that only 3–7% of the sorted Mx cells were positive for these differentiation markers. These cells most likely represent early differentiating cells that reside in the upper Mx area (arrows). Most of the other terminally differentiating cells of the hair follicle are eliminated in the trypsinization step, which is not sufficiently robust to dislodge the firmly adherent differentiating cells from the hair shaft.
(5.7 MB TIF).
Click here for additional data file.
Figure S3 Variable AP Activity of Sorted DP Cells and Functional Hair Reconstitution Assay
(A and B) AP activity in FACS isolated DP cells. Cytospun sorted DP cells showed two levels of AP activity. While the majority of cells were strongly stained, a minor fraction showed weak reactivity ([B], arrows, bottom panel). Sorted ORS cells served as control ([B], top panel). Lower AP levels were detected in vivo in the most proximal part of DP (not shown).
(C) Functional hair reconstitution with FACS-isolated DP cells. Newborn keratinocytes were grafted onto backskins of Nude mice along with FACS-purified DP cells that were cultured 1–2 wk. After 3 wk, grafts were photographed. Note: it is well-established that dermal fibroblasts do not have this ability, which is unique to the DP cells and perhaps a few mesenchymal cells associated with the DP at the base of the follicle [7,9,19].
(3.5 MB TIF).
Click here for additional data file.
Figure S4 RT-PCR Confirmation of Molecular Backbone Genes
Semi-quantitative RT-PCR of genes of the Molecular Backbone group of present and unchanged genes. Shown are representative examples of genes involved in RNA and protein metabolism.
(1.0 MB TIF).
Click here for additional data file.
Figure S5 Real-Time PCR Confirmation of Signature Genes and Correlation with Microarray Results
Real-time PCR confirmation of selected novel and control genes from Figures 2C and 5A. Note the consistent distribution of genes between cell populations using both PCR methods (top panels and Figures 2C and 5A). As a measure of the performance of the microarrays, average signal values were plotted for each gene along with the real-time PCR results (bottom panels). Note the near-perfect match at the quantitative level.
(3.5 MB TIF).
Click here for additional data file.
Table S1 Correlation Coefficients of Array Hybridizations
Raw data of correlation coefficients as shown in Figure 3A. The p-values of replicates are highlighted in bold.
(16 KB XLS).
Click here for additional data file.
Table S2 Microarray Expression Reports
Compilation of quality control statistics for each of the ten microarrays (five fractions, two replicates) arranged in separate spreadsheets.
(104 KB XLS).
Click here for additional data file.
Table S3 Molecular Backbone Genes
Complete list of Molecular Backbone Genes with average array signals and present/absent calls.
(1.5 MB XLS).
Click here for additional data file.
Table S4 Mx Signature Genes
Complete list of Signature Genes with average array signals and present/absent calls, and average fold changes compared to each other fraction. Note that for convenience of access the genes are hyperlinked to the NCBI LocusLink/EntrezGene entries.
(130 KB XLS).
Click here for additional data file.
Table S5 ORS Signature Genes
As in Table S4.
(4.4 MB XLS).
Click here for additional data file.
Table S6 DP Signature Genes
As in Table S4.
(165 KB XLS).
Click here for additional data file.
Table S7 Mc Signature Genes
As in Table S4.
(238 KB XLS).
Click here for additional data file.
Table S8 Significance Analysis of GO Categories of the Molecular Signatures and the Molecular Backbone
Molecular Signature and Backbone genes were grouped functionally into GO categories and statistically analyzed with the “Database for Annotation, Visualization and Integrated Discovery” (DAVID 1.0) Web tool. The table contains the p-values of the Fisher exact probability test and the more conservative EASE Score as a statistical measure of enrichment of genes within GO categories. The values for each Molecular Signature and the Molecular Backbone group are arranged in separate tabs.
(742 KB XLS).
Click here for additional data file.
Table S9 GO Classification of Molecular Backbone
Complete list of genes with corresponding GO classifications. Note that each gene may be in several categories. GO systems were split into separate tabs. Biological Process, BP; Molecular Function, MF; Cellular Component, CC.
(5.0 MB XLS).
Click here for additional data file.
Table S10 GO Classification of Mx Signature
Complete list of genes with corresponding GO classifications. Note that each gene may be represented in several categories. The list is sorted for classifications, and each gene is hyperlinked to the NCBI LocusLink/EntrezGene entries.
(937 KB XLS).
Click here for additional data file.
Table S11 GO Classification of ORS Signature
As in Table S10.
(1.81 MB XLS).
Click here for additional data file.
Table S12 GO Classification of DP Signature
As in Table S10.
(1.3 MB XLS).
Click here for additional data file.
Table S13 GO Classification of Mc Signature
As in Table S10.
(1.8 MB XLS).
Click here for additional data file.
Table S14 Primers for RT-PCR and Real-Time PCR
Primer sequences for RT-PCR and real-time PCR. The same primers were employed for both RT-PCR methods.
(24 KB XLS).
Click here for additional data file.
Accession Numbers
The Entrez Gene GeneID (http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=gene) accession numbers for the genes and gene products discussed in this paper are: a (50518), Akp2 (11647), Bambi (68010), Bmpr1a (12166), Bnc (12173), Cldn1 (12737), Cspg2 (13003), Dlx2 (13392), Dlx3 (13393), Egr2 (13654), Ets2 (23872), Fgf10 (14165), Fgf7 (14178), Fgfr1 (14182), Fgfr2 (14183), Fn1 (14268), Foxn1 (15218), Foxq1 (15220), Frzb (20378), Fst (14313), Gcl (23885), Gdf10 (14560), Gfra1 (14585), Gsdm (57911), Hhip (15245), Hoxc13 (15422), Hr (15460), Inhba (16323), Itgb4 (192897), Mki67 (17345), Kit (16590), Krt1–14 (16664), Krt1-c29 (16675), Krt2–5 (110308), Lef1 (16842), Mdk (17242), Mlph (171531), Msx1 (17701), Msx2 (17702), Ncam1 (17967), Nes (18008), Ngfr (18053), Nog (18121), Ovol1 (18426), Pax3 (18505), Pdgfra (18595), Prdc (23893), Prss12 (19142), Ptch (19206), Pthr1 (19228), Sca-1 (110454), Serpine2 (20720), Sfrp2 (20319), Shh (20423), Snai2 (20583), Snai1 (20613), Sostdc1 (66042), Sox10 (20665), Sox18 (20672), Sox2 (20674), Sox9 (20682), Tcf3 (21415), Tcfcp2l2 (195733), Tgfb1 (21803), Tgfbr1 (21812), Trps1 (83925), Twist1 (22160), Twist2 (13345), Tyrp1 (22178), Vdr (22337), Vim (22352), Wif1 (24117), Wnt1 (22408), Wnt10b (22410), Wnt3a (22416), Wnt5a (22418), Zic1 (22771), and Zic3 (22773).
Raw data and normalized microarray expression data have been deposited at the Gene Expression Omnibus (GEO, http://www.ncbi.nlm.nih.gov/geo/) under the accession number GSE3142.
A special thank-you goes to those who provided us with special antibodies, cells, and reagents, whose gifts are cited in the text. We are especially grateful to the following people: Rockefeller University core facility staff: Svetlana Mazel and Tamara Shengelia (FCRC Facility); Fred Quimby (LARC); Alison North (Bioimaging Facility). Memorial Sloan Kettering Cancer Research Center core facility staff: Agnes Viale, Juan Li, and Hui Zhao (Genomics Core Facility); Nicholas Socci (Computational Biology Center). Maria Nikolova, Lisa Polak, and Nicole Stokes (Fuchs lab) for their technical assistance with various phases of this work; Tudorita Tumbar, Bradley Merrill, Valerie Horsley, Valentina Greco, Hoang Nguyen, Cedric Blanpain, William Lowry, and other present and former members of the Fuchs lab for their valuable discussions and advice, as well as Michael Mildner (University of Vienna) for his technical assistance with real-time PCR. MR is a fellow of the Max Kade Foundation and a recipient of an Erwin-Schrödinger Fellowship from the Austrian Science Fund. EF is an investigator of the Howard Hughes Medical Institute. This work was supported by R01 grants from the National Institutes of Health (AR 31737 and AR050452 to EF).
Competing interests. The authors have declared that no competing interests exist.
Author contributions. MR and LL performed the experiments. MR and EF conceived and designed the experiments, analyzed the data, and wrote the paper.
Citation: Rendl M, Lewis L, Fuchs E (2005) Molecular dissection of mesenchymal–epithelial interactions in the hair follicle. PLoS Biol 3(11): e331.
Abbreviations
Absantibodies
APalkaline phosphatase
DFdermal fraction
DPdermal papilla
ECMextracellular matrix
FACSfluorescence activated cell sorting
GFPgreen fluorescent protein
GOGene Ontology
IRSinner root sheath
K14keratin 14
K5keratin 5
Mcmelanocyte
Mxmatrix
ORSouter root sheath
P[number]postnatal day [number]
RFPred fluorescent protein
SCstem cell
SKPskin-derived precursor cell
==== Refs
References
Hardy MH The secret life of the hair follicle Trends Genet 1992 8 55 61 1566372
Schmidt-Ullrich R Paus R Molecular principles of hair follicle induction and morphogenesis Bioessays 2005 27 247 261 15714560
Morris RJ Liu Y Marles L Yang Z Trempus C Capturing and profiling adult hair follicle stem cells Nat Biotechnol 2004 22 411 417 15024388
Tumbar T Guasch G Greco V Blanpain C Lowry WE Defining the epithelial stem cell niche in skin Science 2004 303 359 363 14671312
Blanpain C Lowry WE Geoghegan A Polak L Fuchs E Self-renewal, multipotency, and the existence of two cell populations within an epithelial stem cell niche Cell 2004 118 635 648 15339667
Jahoda CA Horne KA Oliver RF Induction of hair growth by implantation of cultured dermal papilla cells Nature 1984 311 560 562 6482967
Lichti U Weinberg WC Goodman L Ledbetter S Dooley T In vivo regulation of murine hair growth: Insights from grafting defined cell populations onto nude mice J Invest Dermatol 1993 101 124S 129S 8326145
Inamatsu M Matsuzaki T Iwanari H Yoshizato K Establishment of rat dermal papilla cell lines that sustain the potency to induce hair follicles from afollicular skin J Invest Dermatol 1998 111 767 775 9804336
Kishimoto J Burgeson RE Morgan BA Wnt signaling maintains the hair-inducing activity of the dermal papilla Genes Dev 2000 14 1181 1185 10817753
O'Shaughnessy RF Christiano AM Jahoda CA The role of BMP signalling in the control of ID3 expression in the hair follicle Exp Dermatol 2004 13 621 629 15447722
O'Shaughnessy RF Yeo W Gautier J Jahoda CA Christiano AM The WNT signalling modulator, Wise, is expressed in an interaction-dependent manner during hair-follicle cycling J Invest Dermatol 2004 123 613 621 15373764
Sleeman MA Murison JG Strachan L Kumble K Glenn MP Gene expression in rat dermal papilla cells: Analysis of 2529 ESTs Genomics 2000 69 214 224 11031104
Fernandes KJ McKenzie IA Mill P Smith KM Akhavan M A dermal niche for multipotent adult skin-derived precursor cells Nat Cell Biol 2004 6 1082 1093 15517002
Le Douarin NM Dupin E Cell lineage analysis in neural crest ontogeny J Neurobiol 1993 24 146 161 8445384
Liu X Driskell RR Luo M Abbott D Filali M Characterization of Lef-1 promoter segments that facilitate inductive developmental expression in skin J Invest Dermatol 2004 123 264 274 15245424
Vasioukhin V Degenstein L Wise B Fuchs E The magical touch: Genome targeting in epidermal stem cells induced by tamoxifen application to mouse skin Proc Natl Acad Sci U S A 1999 96 8551 8556 10411913
Kopan R Fuchs E A new look into an old problem: Keratins as tools to investigate determination, morphogenesis, and differentiation in skin Genes Dev 1989 3 1 15 2468556
Zhou P Byrne C Jacobs J Fuchs E Lymphoid enhancer factor 1 directs hair follicle patterning and epithelial cell fate Genes Dev 1995 9 700 713 7537238
McElwee KJ Kissling S Wenzel E Huth A Hoffmann R Cultured peribulbar dermal sheath cells can induce hair follicle development and contribute to the dermal sheath and dermal papilla J Invest Dermatol 2003 121 1267 1275 14675169
Goding CR Mitf from neural crest to melanoma: Signal transduction and transcription in the melanocyte lineage Genes Dev 2000 14 1712 1728 10898786
Hong HK Noveroske JK Headon DJ Liu T Sy MS The winged helix/forkhead transcription factor Foxq1 regulates differentiation of hair in satin mice Genesis 2001 29 163 171 11309849
McGowan KM Tong X Colucci-Guyon E Langa F Babinet C Keratin 17 null mice exhibit age- and strain-dependent alopecia Genes Dev 2002 16 1412 1422 12050118
Furuse M Hata M Furuse K Yoshida Y Haratake A Claudin-based tight junctions are crucial for the mammalian epidermal barrier: A lesson from claudin-1-deficient mice J Cell Biol 2002 156 1099 1111 11889141
Runkel F Marquardt A Stoeger C Kochmann E Simon D The dominant alopecia phenotypes Bareskin, Rex-denuded, and Reduced Coat 2 are caused by mutations in gasdermin 3 Genomics 2004 84 824 835 15475261
Botchkarev VA Kishimoto J Molecular control of epithelial-mesenchymal interactions during hair follicle cycling J Investig Dermatol Symp Proc 2003 8 46 55
Momeni P Glockner G Schmidt O von Holtum D Albrecht B Mutations in a new gene, encoding a zinc-finger protein, cause tricho-rhino-phalangeal syndrome type I Nat Genet 2000 24 71 74 10615131
James K Hosking B Gardner J Muscat GE Koopman P Sox18 mutations in the ragged mouse alleles ragged-like and opossum Genesis 2003 36 1 6 12748961
Jhaveri S Erzurumlu RS Chiaia N Kumar TR Matzuk MM Defective whisker follicles and altered brainstem patterns in activin and follistatin knockout mice Mol Cell Neurosci 1998 12 206 219 9828086
Brown CW Houston-Hawkins DE Woodruff TK Matzuk MM Insertion of Inhbb into the Inhba locus rescues the Inhba-null phenotype and reveals new activin functions Nat Genet 2000 25 453 457 10932194
Ohuchi H Tao H Ohata K Itoh N Kato S Fibroblast growth factor 10 is required for proper development of the mouse whiskers Biochem Biophys Res Commun 2003 302 562 567 12615071
De Moerlooze L Spencer-Dene B Revest J Hajihosseini M Rosewell I An important role for the IIIb isoform of fibroblast growth factor receptor 2 (FGFR2) in mesenchymal-epithelial signalling during mouse organogenesis Development 2000 127 483 492 10631169
Guo L Degenstein L Fuchs E Keratinocyte growth factor is required for hair development but not for wound healing Genes Dev 1996 10 165 175 8566750
Chuang PT Kawcak T McMahon AP Feedback control of mammalian Hedgehog signaling by the Hedgehog-binding protein, Hip1, modulates Fgf signaling during branching morphogenesis of the lung Genes Dev 2003 17 342 347 12569124
Botchkarev VA Botchkareva NV Roth W Nakamura M Chen LH Noggin is a mesenchymally derived stimulator of hair-follicle induction Nat Cell Biol 1999 1 158 164 10559902
Rice R Spencer-Dene B Connor EC Gritli-Linde A McMahon AP Disruption of Fgf10/Fgfr2b-coordinated epithelial-mesenchymal interactions causes cleft palate J Clin Invest 2004 113 1692 1700 15199404
Reddy S Andl T Bagasra A Lu MM Epstein DJ Characterization of Wnt gene expression in developing and postnatal hair follicles and identification of Wnt5a as a target of Sonic hedgehog in hair follicle morphogenesis Mech Dev 2001 107 69 82 11520664
Botchkarev VA Bone morphogenetic proteins and their antagonists in skin and hair follicle biology J Invest Dermatol 2003 120 36 47 12535196
Kulessa H Turk G Hogan BL Inhibition of Bmp signaling affects growth and differentiation in the anagen hair follicle Embo J 2000 19 6664 6674 11118201
Toma JG Akhavan M Fernandes KJ Barnabe-Heider F Sadikot A Isolation of multipotent adult stem cells from the dermis of mammalian skin Nat Cell Biol 2001 3 778 784 11533656
Trainor PA Tam PP Cranial paraxial mesoderm and neural crest cells of the mouse embryo: Co-distribution in the craniofacial mesenchyme but distinct segregation in branchial arches Development 1995 121 2569 2582 7671820
Muller-Rover S Peters EJ Botchkarev VA Panteleyev A Paus R Distinct patterns of NCAM expression are associated with defined stages of murine hair follicle morphogenesis and regression J Histochem Cytochem 1998 46 1401 1410 9815282
Yu DW Yang T Sonoda T Gaffney K Jensen PJ Message of nexin 1, a serine protease inhibitor, is accumulated in the follicular papilla during anagen of the hair cycle J Cell Sci 1995 108 3867 3874 8719892
Huang X Saint-Jeannet JP Induction of the neural crest and the opportunities of life on the edge Dev Biol 2004 275 1 11 15464568
Meulemans D Bronner-Fraser M Gene-regulatory interactions in neural crest evolution and development Dev Cell 2004 7 291 299 15363405
Peters EM Botchkarev VA Muller-Rover S Moll I Rice FL Developmental timing of hair follicle and dorsal skin innervation in mice J Comp Neurol 2002 448 28 52 12012374
Dennis G Sherman BT Hosack DA Yang J Gao W DAVID: Database for Annotation, Visualization, and Integrated Discovery Genome Biol 2003 4 3
Bratka-Robia CB Mitteregger G Aichinger A Egerbacher M Helmreich M Primary cell culture and morphological characterization of canine dermal papilla cells and dermal fibroblasts Vet Dermatol 2002 13 1 6 11896964
DasGupta R Fuchs E Multiple roles for activated LEF/TCF transcription complexes during hair follicle development and differentiation Development 1999 126 4557 4568 10498690
Vaezi A Bauer C Vasioukhin V Fuchs E Actin cable dynamics and Rho/Rock orchestrate a polarized cytoskeletal architecture in the early steps of assembling a stratified epithelium Dev Cell 2002 3 367 381 12361600
Lynch MH O'Guin WM Hardy C Mak L Sun TT Acidic and basic hair/nail (“hard”) keratins: Their colocalization in upper cortical and cuticle cells of the human hair follicle and their relationship to “soft” keratins J Cell Biol 1986 103 2593 2606 2432071
O'Guin WM Sun TT Manabe M Interaction of trichohyalin with intermediate filaments: Three immunologically defined stages of trichohyalin maturation J Invest Dermatol 1992 98 24 32 1728637
|
16162033
|
PMC1216328
|
CC BY
|
2021-01-05 08:21:27
|
no
|
PLoS Biol. 2005 Nov 20; 3(11):e331
|
utf-8
|
PLoS Biol
| 2,005 |
10.1371/journal.pbio.0030331
|
oa_comm
|
==== Front
PLoS BiolPLoS BiolpbioplosbiolPLoS Biology1544-91731545-7885Public Library of Science San Francisco, USA 10.1371/journal.pbio.0030366SynopsisBioinformatics/Computational BiologyCell BiologyDevelopmentMolecular Biology/Structural BiologyDermatologyMus (Mouse)MammalsVertebratesAnimalsMolecular Signatures of the Developing Hair Follicle Synopsis11 2005 20 9 2005 20 9 2005 3 11 e366Copyright: © 2005 Public Library of Science.2005This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are properly credited.
Molecular Dissection of Mesenchymal-Epithelial Interactions in the Hair Follicle
An Ideal Society? Neighbors of Diverse Origins Interact to Create and Maintain Complex Mini-Organs in the Skin
==== Body
When a human baby enters the world, some 5 million hair follicles cover its body. For many, these follicles cease production with age, a fate that befalls both men and women, accounting for a multibillion-dollar hair growth industry, despite the products' limited success. Though the stages of hair development are understood, the molecular signals that guide the process are less clear.
Hair follicles develop from cells in the ectoderm, an embryonic epithelial layer that gives rise to the surface epidermis, and the underlying mesenchyme, connective tissue cells derived from the embryonic mesoderm. During fetal development, the spatial position and individual characteristics of the follicles (long versus short, for example) are determined. Once established, normal hair growth continues in cycles, with each follicle passing through three distinct stages before the hair is shed and the cycle begins anew. Both embryonic follicle formation and adult hair growth begin when mesenchymal dermal papilla (DP) cells, clustered beneath a single layer of epidermal cells, send the message to epithelial stem cells to make a hair follicle. Epithelial stem cells then send their progeny, epithelial matrix cells, to surround the DP cells and trigger the sequence of events that culminates in new hair.
To shed light on the molecular program that drives these mesenchymal–epithelial interactions, Michael Rendl, Lisa Lewis, and Elaine Fuchs used a cell-sorting technique that allowed them to isolate pure populations of DP cells along with populations of four neighboring cell types. Using microarrays to analyze the gene expression profiles of the different cell populations, the authors identified molecular signatures for the DP and its niche, including a group of little-studied genes linked to hair disorders. With these molecular signatures, researchers can begin to analyze each gene's role in hair development and growth.
To get the purified cell populations for their studies, the researchers used an innovative combination of existing techniques, and took advantage of the underlying biology of the hair follicle. In mice, hair begins to grow on the animal's back around 15 days in utero, and is fully formed roughly four days after birth. After the matrix cells envelop the DP, they proliferate, migrate upward, and differentiate into the hair shaft and the inner root sheath (IRS), which surrounds the hair. The IRS is covered by an outer root shaft (ORS), and the whole structure is enclosed by a membrane that separates the skin epithelium from the dermis and the DP. (Melanocytes, which determine hair color, sit just above this membrane.)
To extract cells from the DP along with its neighboring cells, Rendl et al. used transgenic mice and a cell-sorting technique called fluorescent-activated cell sorting, so they could retrieve the different cell types from the backs of the mice. Mice expressing red fluorescent protein (RFP) were bred with mice expressing green fluorescent protein (GFP). Because both reporters express the fluorescent proteins under known conditions—the promoter used to control RFP is typically expressed in DP cells, and the promoter used to control GFP is expressed in matrix and ORS cells—they serve as an initial filter in cell sorting.
Follicles were isolated from the backs of four-day-old mice, and five populations of cells—matrix, ORS, melanocytes, DP, or dermal fibroblasts (connective tissue cells)—were sorted based on whether or not they expressed GFP or RFP and whether they expressed known cell-surface markers. The authors were most interested in identifying the unique features of the cells that initiate the hair development program, the DP cells. They were sure the cells were from the DP because they expressed high levels of four known DP markers and triggered hair growth on mutant Nude mice, which only DP cells can do (along with ubiquitous keratinocytes).
Microarray analysis of the gene activity of the different cell types found that 4,000 genes were common to all the cell types, forming the “molecular backbone” of basic cellular functions. Just 150–300 genes showed elevated expression in any one cell type, forming the molecular signature for each population. Because the authors' DP signature genes differed substantially from those found in previous studies, they went on to verify protein expression in the cells, confirming that the signature accurately reflected the DP profile. Many of the signature genes are involved in genetic hair disorders, confirming the functional significance of the signatures, while others encode novel proteins involved in transcription, cell signaling, and cell adhesion. This subset provides a window into the special roles carried out by the different cells at the genomic level, and will help future studies explore their function in hair biology.
Having identified the molecular signatures of the dermal papilla and its neighboring cells, researchers can now study how these genes influence hair development and growth
Developmental biologists have made great progress in understanding the molecular agents of hair morphogenesis since Margaret Hardy called the hair follicle a “treasure waiting to be discovered” just over a decade ago. By identifying unique molecular profiles of the DP and its cohorts, this study has paved the way for revealing all the hair follicle's secrets, one signal at a time. For more on skin and hair follicle development, see the accompanying Primer by Sarah Millar (DOI: 10.1371/journal.pbio.0030372). —Liza Gross
|
0
|
PMC1216329
|
CC BY
|
2021-01-05 08:28:16
|
no
|
PLoS Biol. 2005 Nov 20; 3(11):e366
|
utf-8
|
PLoS Biol
| 2,005 |
10.1371/journal.pbio.0030366
|
oa_comm
|
==== Front
PLoS MedPLoS MedpmedplosmedPLoS Medicine1549-12771549-1676Public Library of Science San Francisco, USA 1616203410.1371/journal.pmed.0020265Research ArticleBioengineeringAllergy/ImmunologyDermatologyOncologyGeneral MedicineVaccinesBiotechnologyCancer BiologyCell BiologyImmunologyInfectious DiseasesMolecular Biology/Structural BiologySystems BiologyVirologyOncologyImmunology and AllergyDermatologyMarked Differences in Human Melanoma Antigen-Specific T Cell Responsiveness after Vaccination Using a Functional Microarray Human Tumor-Specific T Cell FunctionChen Daniel S
1
2
Soen Yoav
3
Stuge Tor B
4
Lee Peter P
4
Weber Jeffrey S
5
Brown Patrick O
2
3
Davis Mark M
2
6
*1Department of Internal Medicine/Division of Oncology, Stanford University, Stanford, California, United States of America,2Howard Hughes Medical Institute, Stanford University, Stanford, California, United States of America,3Department of Biochemistry, Stanford University, Stanford, California, United States of America,4Department of Medicine, Stanford University, Stanford, California, United States of America,5Norris Cancer Center, University of Southern California, Los Angeles, California, United States of America,6 Department of Microbiology and Immunology, Stanford University, Stanford, California, United States of AmericaRees Jonathan Academic EditorUniversity of EdinburghUnited Kingdom*To whom correspondence should be addressed. E-mail: [email protected]
Competing Interests: The authors have declared that no competing interests exist.
Author Contributions: DSC and YS conceived of the experiments and designed the study. DSC, YS, and TBS performed the experiments. DSC, YS, and TBS analyzed the data. JSW enrolled patients. DSC, YS, TBS, JSW, POB, and MMD contributed to writing the paper. POB and MMD suggested ideas and experiments.
10 2005 20 9 2005 2 10 e2657 4 2005 30 6 2005 Copyright: © 2005 Chen et al.2005This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are properly credited.
Cancer Vaccines: The Next Generation of Tools to Monitor the Anticancer Immune Response
Measuring the Immune Response to Tumor Vaccination
Background
In contrast to many animal model studies, immunotherapeutic trials in humans suffering from cancer invariably result in a broad range of outcomes, from long-lasting remissions to no discernable effect.
Methods and Findings
In order to study the T cell responses in patients undergoing a melanoma-associated peptide vaccine trial, we have developed a high-throughput method using arrays of peptide-major histocompatibility complexes (pMHC) together with antibodies against secreted factors. T cells were specifically immobilized and activated by binding to particular pMHCs. The antibodies, spotted together with the pMHC, specifically capture cytokines secreted by the T cells. This technique allows rapid, simultaneous isolation and multiparametric functional characterization of antigen-specific T cells present in clinical samples. Analysis of CD8+ lymphocytes from ten melanoma patients after peptide vaccination revealed a diverse set of patient- and antigen-specific profiles of cytokine secretion, indicating surprising differences in their responsiveness. Four out of four patients who showed moderate or greater secretion of both interferon-γ (IFNγ) and tumor necrosis factor-α (TNFα) in response to a gp100 antigen remained free of melanoma recurrence, whereas only two of six patients who showed discordant secretion of IFNγ and TNFα did so.
Conclusion
Such multiparametric analysis of T cell antigen specificity and function provides a valuable tool with which to dissect the molecular underpinnings of immune responsiveness and how this information correlates with clinical outcome.
Patients vaccinated with melanoma-associated peptides show a wide diversity of responses, analysis of which may help us understand the differing clinical responses to such vaccines.
==== Body
Introduction
Antigen-specific cellular immune responses are mediated by αβT cell receptor (TCR)-bearing T cells that recognize specific peptides bound to major histocompatibility complex (MHC) molecules on the surfaces of other cells. These T cells form a major part of the adaptive immune response. CD8+ T cells mediate direct lysis of infected or aberrant cells, whereas CD4+ T helper cells modulate antibody (B cell) responses and those of other cells. T cells may become activated following antigen recognition and respond by secreting soluble factors, which include mediators of target cell lysis, pleiotropic effector factors, growth factors, and inflammatory and regulatory cytokines (Table 1). This is a highly regulated and complex process. In many cases, antigen recognition by primed CD8+ T cells leads to the lysis of cellular targets and the release of inflammatory cytokines. Alternatively, this response may be partially or completely anergic.
Table 1 Factors Secreted by Lymphocytes or Other Cells of the Immune System
For many years, investigators have sought to direct T cell responses against tumors by vaccination [1]. These efforts have been greatly aided by the discovery of many peptide antigens that are displayed on MHC molecules on the surface of tumor cells and that have been shown to elicit T cell responses both in vitro and in vivo [2,3]. This discovery has given rise to a variety of strategies, including protein and peptide vaccination [4], adoptive cellular therapy [5], cytokine therapy (i.e., interleukin [IL]-2, granulocyte-macrophage colony-stimulating factor [GM-CSF], interferon [IFN] α) [6–8], and immune response modifiers such as anti-CTLA4 [9,10]. Despite intense efforts, the success of most of these protocols has been mixed. Although in many cases, specific T cell responses can be generated in patients (or expanded ex vivo and reintroduced intravenously), they are not usually effective against the tumor. A large part of the problem may be that most of these tumor-associated antigens are normal “self” peptides, and responses may be naturally suppressed. In this context, it is important to monitor the precise functional status of T cells that are elicited by a particular immunization protocol, and to determine what conditions result in T cells that are the most effective in bringing about clinically significant results. For this purpose, the ability to track antigen-specific T cells with peptide-MHC (pMHC) tetramers [11] has been an important tool in the identification and characterization of lymphocytes capable of recognizing specific tumor antigens. This technique, together with other assays (e.g., intracellular cytokine staining, CD107, ELISpot, killing assay) have been used to try to address T cell function [12–15]. However, these assays are labor intensive, require large quantities of clinical peripheral blood mononuclear cell (PBMC) specimens for a comprehensive analysis, have poor spatial resolution and/or low sensitivity for secreted responses, and do not address the growing need to track multiple T cell specificities for different functional events. To overcome these limitations, we previously reported on an array-based approach to capture and quantitate antigen-specific T cells based on their adherence to pMHC complexes [16]. Here, we report a further development of this technology, in which we combined the high-throughput capture and activation of antigen-specific T cells described previously with the simultaneous analysis of the secretion of a wide variety of factors with single-cell resolution. Using this technique, we assess antigen-specific T cells from different vaccine recipients and analyze different functional profiles following antigen recognition in an attempt to explore the variability of clinical outcomes that is characteristic of tumor vaccine trials.
Methods
Peptides and Cell Lines
The peptides gp100 209–2M (IMDQVPFSV), MART1 M26 (ELAGIGILTV), tyrosinase 370D (YMDGTMSQV), gp100 209 (ITDQVPFSV), MART1 27–35 (AAGIGILTV), CMV pp65 495–503 (NLVPMVATV), EBV BMLF1 280–288 (GLCTLVAML), and influenza MP 58–66 (GILGFVFTL) were produced at the Protein and Nucleic Acid Facility at Stanford University (Stanford, California, United States).
CD8+ T cell clones were derived and maintained as previously described [17]. Briefly, clones were derived from melanoma patients or healthy donors expressing the human leukocyte antigen (HLA)-A2.1 MHC molecule. Clone 132.2 specifically binds gp100 209 or gp100 209–2M/HLA-A2.1. Clones 461.30 and 461.24 specifically bind MART 27–35 or M26/HLA-A2.1. Clone CMV94.3 specifically binds CMV pp65 495–503/HLA-A2.1 and was derived by fluorescence-activated cell sorter (FACS) separation of individual tetramer-positive cells from PBMCs from a healthy donor. T cell clones were cultured in CTL medium (Iscove's modified Dulbecco's medium, with 10% fetal calf serum, 2% human AB sera, and standard cell-culture concentrations of penicillin, streptomycin, and L-glutamine) supplemented with 50 U/ml IL-2. The clones were expanded by stimulation with phytohemagglutinin (Invitrogen, Carlsbad, California, United States) at a 1:100 dilution, followed by 14 d of culture in CTL medium with irradiated feeder cells and 50 U/ml of IL-2. Following expansion, clones were either cryopreserved or maintained in culture with CTL medium supplemented with 50 U/ml IL-2 or 2 ng/ml IL-15, and used within 2 wk. Cryopreserved cells were thawed at least 2 d prior to assays, and were suspended in CTL medium with 100 U/ml IL-2. At 1 d prior to experiments, the clones were transferred to fresh CTL medium without interleukins.
Preparation of pMHC Class I
The pMHC tetramers were developed in the Davis lab and were prepared as previously described [11]. Alternatively, tetramers were purchased from Beckman Coulter (Allendale, New Jersey, United States). The pMHC dimers were purchased from BD Pharmingen (San Diego, California, United States) and prepared per manufacturer's protocol. All pMHC constructs were supplemented with glycerol prior to printing (2% final concentration).
Vaccination Protocol
A randomized phase II trial for patients with resected stages IIC/III and IV melanoma who were HLA A*0201-positive and expressed at least one of the following was conducted: HMB-45 (gp100), tyrosinase, or Melan-A (MART-1). Informed consent was obtained from all patients. A total of 60 patients were randomly allocated to receive three peptides at 1 mg each (gp100 209–2M, tyrosinase 370D, and MART-1 M26) emulsified with the adjuvant Montanide ISA 51 in a 1:1 ratio by volume with: (A) IL-12 at 30 ng per kilogram body weight, (B) IL-12 at 100 ng per kilogram, and (C) IL-12 at 30 ng per kilogram with GM-CSF at 83 μg per peptide emulsified. All patients had a CT scan of the chest/abdomen/pelvis and brain MRI required to show no evidence of disease within 28 d of initiation of vaccine treatment. Injections were given at weeks 0, 2, 4, 6, 10, 14, 18, and 26, then week 50 for a total of nine injections over 1 y. Leukopheresis was performed just prior to the first injection and within 2 wk after injection number 8 (week 28) for immune response assays. Smaller blood samples were obtained at 3, 9, and 12 mo. Clinical specimens were frozen and stored prior to use, as previously described [18,19]. Tetramer flow cytometry on pretreatment blood samples showed no detectable melanoma antigen-specific T cells.
Antibodies
Antibodies against human CD8 (HIT8a), HLA-A2 (BB7.2), IFNγ (NIB42, 4S.B3), TNFα (MAb1, MAb11), granzyme B (2CF/F5, GB11), GM-CSF (BVD2-23B6, BVD2-21C11), IL-2 (5344.111, B33–2), IL-4 (8D4–8, MP4-25D2), IL-5 (TRFK5, JES1-5A10), IL-10 (JES3-9D7, JES3-12G8), and IL-12p70 (20C2, C8.6) were purchased from BD Pharmingen (San Diego, California, United States); antibodies against IL-1a (4414.141, pAb BAF200), IL-1b (8516.311. pAb BAF201), IL-3 (4815.211, pAb BAF203), IL-6 (6708, pAb BAF206), IL-7 (7417, pAb BAF207), IL-13 (32116, pAb BAF213), IL-15 (34593, pAb BAF247), IL-17 (41809, pAb BAF317), lymphotactin (109001, pAb BAF695), IP-10/CXCL10 (33036.211, pAb BAF266), TGF-β1 (9005, 27240), TNFβ (5807, pAb BAF211), vascular endothelial growth factor (VEGF; pAb AF293NA, pAb BAF293), and VEGF-D (78902, 78923) were purchased from R&D Systems (Minneapolis, Minnesota, United States), and granzyme A (CLB-GA29, CLB-GA28) was purchased from Research Diagnostics Incorporated (Flanders, New Jersey, United States). Clone names are listed in parentheses, with biotinylated antibody clone names listed second, where appropriate. Polyclonal antibodies are noted with a pAb prefix.
Preparation of pMHC Functional Microarrays
Libraries of pMHC/antibody mixtures were prepared as follows: Each of the pMHC constructs was mixed with a panel of antibodies against potentially secreted factors (Table 1), such that each mixture contained a single pMHC construct (2.5 mg/ml final concentration) and a single antibody against a secreted factor (0.5–1mg/ml final concentration). Each of these pMHC-based mixtures was supplemented with 2% glycerol. A second library with 0.5 mg/ml of either anti-human CD8 or anti-human HLA-A2 (instead of pMHC) and an antibody against a secreted factor was prepared as a nonactivating control. A volume of 12 μl of each mixture was loaded into a 384-well plate (MJResearch, Waltham, Massachusetts, United States) and arrayed in triplicate onto three-dimensional substrates composed of microscope slides coated with a polyacrylamide gel (Perkin Elmer, Boston, Massachusetts, United States), preprocessed according to the manufacturer's instructions. Samples were dispensed using a non-contact piezoelectric arrayer (Perkin Elmer), such that each spot contained ten drops of approximately 0.45 nl each. Printed proteins were immobilized within the gel substrate by incubating the slides for 48 h at 4 °C in a humid chamber. Following the immobilization, the arrays were placed in a dry slide box, sealed with tape, and stored at 4 °C until use. Arrays were tested for specific capture of secreted factors using defined concentrations of recombinant human factors (Quantikines, R&D Systems) incubated on an unused array for 30 min at room temperature, followed by 12 h at 4 °C. Arrays were then washed, developed, and imaged, as described below.
Flow Cytometry Analysis
Patient PBMCs were analyzed for G209-2M-tetramer reactive cells by flow cytometry as described previously [13]. Briefly, cells were reacted with G209-2M-tetramer-PE (Beckman Coulter Immunomics Operations, San Diego, CA, United States) at 1:200 dilution for 20 min at room temperature, followed by anti-CD19 FITC (Caltag Laboratories, Burlingame, California, United States) and anti-CD8 PerCP-Cy5.5 (BD Biosciences, San Jose, California, United States) antibodies at final staining dilution of 1:40 and 1:20, respectively, for an additional 20 min. Cells were then washed and analyzed using a FACSCalibur flow cytometer (Beckton Dickinson, San Jose, California, United States). Approximately 105 events were acquired from each sample and analyzed using FlowJo software (TreeStar, San Carlos, California, United States). Plotted CD19−, CD8+, tetramer+ lymphocytes were calculated as percent of total CD8+ lymphocytes for each sample.
Binding and Secretion Assays
Binding and secretion assays were performed with either patient CD8+ T cells or cultured human CD8+ T cell clones. CD8+ T cells were isolated from 5 × 107 PBMCs from patients on the above-described vaccine protocol, using a CD8− isolation column (Miltenyi Biotec). The CD8+ T cells were then resuspended in 200 μl of incubation medium (RPMI supplemented with 5% FCS, glutamine, and standard concentrations of penicillin and streptomycin). Alternatively, 1 × 106 CD8+ T cell clones, as described above, were resuspended in 200 μl of incubation medium. For pMHC binding analysis, the single cell suspension was incubated on the pMHC array for 10–30 min at 20 °C. At the end of the incubation period, the array was washed in calcium- and magnesium-free PBS (CMF) to remove unbound cells and imaged as detailed below. To analyze cellular secretion, the cells were incubated on the array in 400 μl of incubation medium at 37 °C for 2 h (CD8+ T cell lines) or for 24 h (patient samples). To determine the secretion of factors, the arrays were washed in CMF and incubated in 200 μl of pooled biotinylated antibodies in staining medium (10% FCS in CMF) for 20 min at 20 °C. The biotinylated antibodies were each matched to a single, printed antibody specific against different epitopes of the same secreted factor. The final concentration of each biotinylated antibody was based upon concentrations recommended for ELISA or ELISpot, and titrated as necessary. After incubation with biotinylated antibodies, the array was washed twice in CMF and stained with 3.3 μg/ml streptavidin-phycoerythrin (BD Pharmingen) in 200 μl of staining medium for 20 min at 20 °C in the dark. The array was dip-washed twice again in CMF and then imaged as detailed below.
Image Acquisition and Analysis
Imaging was performed using a Zeiss Axiovert-200 microscope (Oberkochen, Germany) fitted with a high-speed piezo electric z-motor stage (Applied Scientific Instrumentation, Eugene, Oregon, United States), a 10× Zeiss Fluor objective, a CCD camera (Roper Scientific, Trenton, New Jersey, United States), and dual excitation and emission filter wheels (Sutter Instruments, Novato, California, United States). DIC and Cy3 images were collected from each spot on the array. Image acquisition was controlled by Metamorph (Universal Imaging, Downingtown, Pennsylvania, United States). Image analysis, feature extraction, and data analysis were performed using Metamorph, ImageXpress (Molecular Devices, Union City, California, United States), and Matlab Software (The MathWorks, Natick, Massachusetts, United States).
Analysis of Patient Data
Scoring of patient samples was performed in a blinded fashion. Coded samples were scored without prior information regarding patient age, sex, therapy, clinical, or immunological outcome. Scoring was based on a five-point scale (i.e., 0–4), with 0 representing background signal. A cell count score for IFNγ and TNFα secretion was based upon the number of responding cells per spot: 0, no response; 1, 1–5; 2, 6–10; 3, 11–20; and 4, more than 21 responding cells. A second score, for intensity, was based on average integrated pixel fluorescence over all replicate spots (after subtraction of the average integrated pixel fluorescence of control spots containing only pMHC). Each of the averaged intensities was normalized by a value greater than the highest intensity for that particular secreted factor, across all patients and expressed as a percentage of that value. The intensity score was assigned as follows: 0, 0%–5%; 1, 6%–25%; 2, 26%–50%; 3, 51%–76%; and 4, 77%–100%. A combined score for IFNγ and TNFα was obtained by adjusting the cell count score up or down by 1 if the intensity score was higher or lower than the cell count score. Scores for secreted factors lacking clear and consistent focal secretion across all patients (including granzyme B, IL-2, TGFβ, IL-1b, IL-6, GM-CSF, IL-1a, IL-3, IL-4, IL-5, IL-6, IL-7, IL-10, IL-12, IL-13, IL-15, IL-17, lymphotactin, IP-10, TNFβ, VEGF, VEGF-D, and granzyme A) received only intensity scores according to the following scale: 0, 0%–5%; 1, 6%–19%; 2, 20%–50%; 3, 51%–80%; and 4, 81%–100%. Six clinical specimens were initially tested using the pMHC functional array. Five additional specimens were tested to assess consistency of findings and expand the number of analyzed specimens.
Results
Specific Capture of Human Antigen-Specific T Cells Using a Readily Scalable pMHC Array
In our earlier work, we used pMHC tetramers spotted onto arrays to capture and quantitate specific T cells [16]. This has the disadvantage that the synthesis of pMHC tetramers is time consuming and not easily scalable to survey large numbers of different pMHC complexes. To address this problem, we used dimeric HLA-A2-immunoglobulin (Ig)-containing molecules (DimerX, BD Biosciences) [20] that lack peptide and are readily loaded with specific peptide antigens. In this way, dozens and potentially thousands of pMHCs can be made simultaneously. We tested the specificity of capture of these molecules using melanoma-specific antigens and two well-characterized human CD8+ T cell lines on pMHC microarrays printed on polyacrylamide-coated slides. The CD8+ T cell clones 132.2 and 461.30 were originally isolated from two patients vaccinated with gp100 209–2M peptide and MART1 M26 peptide [17] and expanded in vitro. Using flow cytometry, 132.2 stains exclusively with the gp100 209–2M/HLA-A2.1 tetramer while 461.30 stains with MART1 M26/HLA-A2.1, but not the gp100 209–2M/HLA-A2.1 tetramer. A microarray using pMHC constructs (gp100 209–2M/HLA-A2.1 tetramer, gp100 209–2M/HLA-A2.1 dimer-Ig, gp100 154/HLA-A2.1 dimer-Ig, MART1 M26/HLA-A2.1 tetramer, MART1 M26/HLA-A2.1 dimer-Ig and MART1 27/HLA-A2.1 dimer-Ig and monoclonal antibodies (anti-CD8a, and anti-HLA-A2) was constructed. For the assay, 132.2 (gp100-specific) and 461.30 (MART1-specific) cells were overlaid onto separate arrays. While both cells bound to the monoclonal antibody spots, 132.2 cells bound exclusively to the gp100 pMHC spots, while 461.30 cells bound only to the MART1 pMHC spots. Binding to specific tetramer and dimer spots was equivalent (Figure 1).
Figure 1 Peptide-MHC Cellular Microarrays
(A) A functional pMHC microarray diagram illustrating the array format. An array of spots containing different capture probes (pMHC molecules) cospotted with detector probes, which are antibodies against potentially secreted factors. Antigen-specific T cells are captured by recognition of printed pMHC, and may become activated. Subsequent secretion of specific factors is captured by the printed detector probes. The presence of those factors is detected by labeled secondary antibodies (developer probes).
(B) Specificity of pMHC T cell capture is peptide-specific. Human CD8+ lymphocyte clones 132.2 and 461.30 were incubated on duplicate microarrays, which included gp100-2M/HLA-A2.1 and MART-1 M26/HLA-A2.1 tetramer and dimer spots. 2M-specific 132.2 cells were exclusively captured by gp100 spots, and M26-specific 461.30 binding was restricted to MART-1 spots. Binding efficiency to pMHC tetramer or dimer of a given specificity was similar.
pMHC Arrays Are Sensitive to Low-Frequency T Cell Populations
We compared the sensitivity of array-based detection to pMHC tetramer staining and flow cytometry for gp100 209–2M-specific CD8+ T lymphocytes from human clinical samples. PBMC samples collected from a patient pre- and post-gp100 peptide vaccination were used in this comparison. FACS analysis using tetramer staining indicated that the pre-vaccine sample was negative for gp100 209–2M, while the postvaccine sample contained 0.19% positive CD8+ T cells. To test the limits of detection in both methods, we diluted postvaccine CD8+ T cells in the gp100 209–2M negative prevaccine sample. CD8+ T cells were isolated from both samples using a depletion column and mixed at post-vaccine:pre-vaccine ratios of 1:0, 1:2, 1:9, 1:29, and 0:1. Each mixture was analyzed separately, using a pMHC microarray and tetramer/FACS. T cells captured on the microarray spots were counted and averaged over five identical gp100 209–2M/HLA-A2.1 spots. Both methods were able to detect antigen-specific T cells at fractional abundances as low as one cell in 10,000, or 0.01% of the CD8+ population (Figure 2). Cellular microarray binding variability was minimized by using the average number of cells bound over the five replicate spots printed on the same array. Results of the tetramer/FACS varied by the selection gate for forward scatter/side scatter, CD19+ dump and CD8+/tetramer+ staining. However, both cellular microarray and tetramer/FACS produced antigen-specific T cell frequencies that correlated well with serial dilution.
Figure 2 Sensitivity of pMHC-Specific T-Cell Detection
The sensitivity of array-based T cell detection was compared to FACS-based detection. Patient PBMCs post-gp100 peptide vaccination were serially diluted in corresponding prevaccination sample. The dilution series was analyzed by both methods. Blue-labeled data correspond to the average number of bound cells per spot on the array (left graph) or the frequency of cells measured by FACS (right graph). The limit of detection in both methods was similar. Average number of bound cells per spot was obtained from five replicate spots on the same array. Error bars denote standard error of the mean across replicates. Red curves denote predicted results based on the values measured for the undiluted sample.
Functional Profiling of Secreted T Cell Factors following Antigen Recognition
To study the functional responses of T cells after antigen recognition, we combined cell capture molecules (capture probes) with molecules that bind secreted factors (detector probes). Mixtures of a capture probe and detector probe were printed in triplicate on individual spots on the functional microarray. Seven different pMHC molecules (gp100 209–2M/HLA-A2.1, MART1 M26/HLA-A2.1, CMV pp65 495/HLA-A2.1, gp100 209/HLA-A2.1, influenza MP 58/HLA-A2.1, EBV BMLF1 280, and tyrosinase 370D) and four different monoclonal antibodies (anti-HLA-A2, anti-CD8, anti-CD3/anti-CD28) were used as capture probes. These were combined with detector probes composed of 26 different monoclonal antibodies specific for secreted factors (IL-1a, IL-1b, IL-2, IL-3, IL-4, IL-5, IL-6, IL-7, IL-10, IL-12, IL-13, IL-15, IL-17, IFNγ, TNFα, TNFβ, GM-CSF, granzyme B, granzyme A, TGFβ, VEGF, VEGF-D, lymphotactin, and IP10). pMHC functional microarrays capture antigen-specific T cells, provide an activating signal, and capture specific secreted factor as they are released by a given T cell. The secretion profile is visualized using a sandwich assay. A mixture of matched biotinylated monoclonal antibodies (developer probes) are applied to the array, followed by streptavidin-phycoerythrin. The identity of each secreted cytokine is determined from its location on the array. The specific capture and detection of soluble factors was confirmed by directly incubating matched, quantified soluble factors on the array.
Four CD8 T cell lines (132.2, 461.24, 461.30, and CMV94.3) were tested for their functional profiles following binding to their respective cognate antigen spot, or binding to a nonactivating spot (e.g. anti-CD8, anti-HLA-A2). Individually immobilized T cells could be visualized by light microscopy, and correlated with secreted factor captured from specific cells (Figure 3A). All antigen-specific T cell lines exhibited similar secretion profiles following antigen recognition, characterized by strong IFNγ, TNFα, granzyme A (unpublished data), and granzyme B secretion, and weaker GM-CSF and IL-2 secretion (Figure 3B). Baseline secretion of granzyme B and GM-CSF were detectable from cells bound to anti-CD8 and anti-HLA-A2 monoclonal antibody spots, but were much weaker than antigen-stimulated secretion.
Figure 3 Profiling T Cell Function
Clonally derived MART-1/A2 specific human CD8+ T cells were incubated on a functional pMHC microarray on which individual spots contained pMHC or a control anti-CD8 monoclonal antibody (i.e., capture probes) cospotted with a panel of antibodies against potentially secreted factors (i.e., detector probes). MART-1-specific cells were immobilized on both MART-1/HLA-A2.1 and anti-CD8 containing spots. Bound cells were further incubated at 37 °C for 2 h. Secreted factors were captured by the coprinted antibodies at close vicinity to the secreting cells and detected using matched, biotinylated antibodies. Some of the initially bound cells detached during the staining procedure.
(A) Top and bottom rows display IFNγ and TNFα secretions, respectively, each detected at single-cell resolution. The fluorescence images (red) are overlaid onto the differential interference contrast light microscopy images in the rightmost two columns. Not all immobilized T cells secreted detectable factors. No T cell binding or fluorescence was detectable on irrelevant pMHC spots (unpublished data).
(B) Secretion profile for 17 different factors. Capture probes are either anti-CD8 antibody (left) or MART1 M26/A2 (right). Cospotted detector probes are indicated for each spot. Secretion signal is shown in pseudocolor, representing fluorescence intensity. Secreted factor-specific scaling has been applied to maximize resolution.
Detection of Heterogeneous Functional Profiles of Tumor-Associated Antigen-Specific T Cells Using a Functional pMHC Array
To gain insight into differences in patient responses to tumor vaccine therapy, we investigated the functional profiles of patient T cells on arrays of pMHC immobilized with different monoclonal antibodies directed against specific secreted factors. Eleven samples were taken from patients with resected stage IIC to IV malignant melanoma enrolled in a clinical trial involving gp100 209–2M, MART1 M26, and tyrosinase 370D peptide injection. These samples were taken by leukopheresis after eight injections of peptide and IL-12 adjuvant, 6 mo after the first injection (one sample was taken 12 mo after the first injection). CD8+ T cells were isolated from 5 × 107 PBMCs using a negative isolation column and incubated on a functional pMHC array at 37 °C and 5% CO2 for 24 h. Secretion profiles were detected as described above.
Specific T cell immobilization was visible within 10 min, as described previously [16]. Cytokine secretion was detectable in each clinical sample as a phycoerythrin fluorescent signal on spots where specific cytokine capture antibodies had been printed. Secretion from individual immobilized cells resulted in the highly localized capture of cytokine. Different cytokines had different typical appearances on the array (Figure 4). IFNγ capture resembled a “starburst” pattern emanating from specifically bound cells, whereas TNFα capture resembled a thin ring around the cell. Some cytokines produced a diffuse signal, which may have resulted either from saturation of capture reagent on a given spot, or secretion from unbound bystander cells rather than specifically captured and activated cells responding to a printed antigen. We characterized the secretion profiles for each patient sample (Figure 5). Quantitative data extracted from each image included average spot intensity, reflecting the total amount of a given cytokine captured on a spot; spot intensity standard deviation, which reflects granularity of the developed signal; and the number of responding cells. The gp100-specific CD8+ T cells from four samples from three patients, 10721, 10739, 10735, and 10794 (10735 and 10794 were isolated from the same patient, at 6 and 12 mo, respectively) gave strong IFNγ and TNFα secretion signals. All three patients remain disease free 25, 21, and 22 mo, respectively, after initiating vaccine therapy. The gp100-specific CD8+ T cells from patient 10722 also responded with similar IFNγ and TNFα secretion. At this writing, this patient remains without evidence of disease progression after 25 mo. In contrast, four of six patients in whom gp100-specific CD8+ T cells mounted a strong IFNγ response, or a strong TNFα response, but not both (Figure 6), experienced a relapse of disease (patient samples 10710, 10737, 10713, and 10757 at 8, 11, 6, and 2 mo following initiation of vaccine therapy, respectively). Three out of four patients in this study that had strong GM-CSF secretion also remain free of disease at this writing. IL-1b and IL-6 were both strongly secreted by three patients, of which only one patient, 10713, has experienced recurrent disease. Patient response profiles to different antigens did not appear to be global. While some patients had detectable IFNγ responses to a plethora of different antigens (i.e., 10794), or had no detectable IFNγ response to the tested antigens (i.e., 10742), several patients had detectable responses to some antigens, but not others. Patients 10713, 10770, and 10757 failed to generate IFNγ secretion in response to gp100 or MART-1, but were capable of excellent IFNγ secretion in response to viral antigens or a different melanoma antigen, tyrosinase (Figure 7). Patient 10713 had no IFNγ response to gp100 209–2M/HLA-A2.1, MART1 M26/HLA-A2.1, or tyrosinase 370D/HLA-A2.1, but a very strong IFNγ response to a common CMV antigen, pp65 495/HLA-A2.1. These five samples were also tested for secretion of IFNγ, TNFα, granzyme B, IL-2, IL-1b, and IL-6 in response to gp100 209/HLA-A2.1 (wild-type peptide sequence). All five tested samples showed functional responses to wild-type gp100 209/HLA-A2.1 that mirrored the measured response to gp100 209–2M/HLA-A2.1 (unpublished data). Cytokine detector spots were calibrated with recombinant protein, or cells that secrete those factors. Despite immobilization of antigen-specific T cells, spots containing detectors against other secreted factors are not shown, due to lack of detectable signal.
Figure 4 Anatomy of Cytokine Secretion
Secreted cytokine captured as it is released from activated lymphocytes immobilized on a pMHC cellular microarray shows cytokine-specific configurations. Select representative patient samples are shown for each labeled cytokine to illustrate the patterns of secretion for each individual cytokine.
Figure 5 Heterogeneity of Melanoma-Associated Antigen-Specific T Cell Responses following Peptide Vaccination
Eleven samples taken from patients enrolled in peptide vaccine trials were analyzed on pMHC functional microarrays. Patients received eight subcutaneous injections of peptides gp100 209–2M, MART1 M26, and tyrosinase 370D, along with adjuvant in a 6-mo period. Leukopheresis samples were collected after the eighth injection. Sample 10794 was collected from the same patient as 10735 after month 12. Functional profiles were generated by incubating patient CD8+ T lymphocytes on pMHC functional microarrays for 24 h at 37 °C and detecting the secreted factors with biotinylated secondary antibodies and streptavidin-phycoerythrin. Data were analyzed by automated fluorescence microscopy. Responses were scored on a five-point scale (0 to 4 bars), reflecting number of responders and overall fluorescent signal intensity per spot (Figure S1). Four bars indicate a strong response, and “0” indicates lack of a response. Each spot was printed in triplicate and analyzed individually. Patient clinical data are listed, including age and sex ( “ID”), stage of disease at enrollment (“Stage”), and outcome at follow-up (“Outcome”). Column labeled “IL12” specifies IL-12 adjuvant doses. Patient 10713 also received GM-CSF in addition to IL-12 adjuvant. Other secreted factors not shown include IL-4, IL-5, IL-10, IL-12p70, IL-1b, IL-3, IL-7, IL-13, IL-15, IL-17, lymphotactin, IP-10/CXCL10, TNFβ, VEGF, VEGF-D and granzyme A due to either lack of detectable secretion or limited analysis performed on only a fraction of the samples. In vitro restimulated cell lines directed against gp100 209 (132.2), MART1 M27 (461.30), or CMV pp65 495 (CMV94.3) were bound and secreted factors in response to gp100, MART1, and CMV (unpublished data), respectively.
Figure 6 Differences in Functional Profiles between Three Patients with Different Clinical Outcomes
gp100 209–2M spots co-spotted with anti-IFNγ, anti-TNFα, anti-IL1b, and anti-IL-6 are shown for three separate patient samples. Patient 10735, who remains free of disease at this writing, displays strong IFNγ, TNFα, IL-1b, and IL-6 secretion. Patients 10710 and 10737 show strong IFNγ secretion, but weak to no TNFα, IL-1b, and IL-6 secretion; both patients experienced disease recurrence soon after these samples were drawn. Note the diffuse pattern of IL-1b and IL-6 capture, which differs from the focal capture of IFNγ and TNFα.
Figure 7 Antigen-Specific Profiles within Individual Patients
CD8+ lymphocytes were isolated from PBMCs from patients 10794, 10713, 10770, 10757, and 10742, and incubated on a pMHC functional cellular microarray containing anti-IFNγ co-spotted with several different peptide-MHC (HLA-A2) complexes. These included melanoma-associated antigens gp100, MART-1, and tyrosinase, and viral antigens from cytomegalovirus, influenza virus, and Epstein-Barr virus (gp100 209–2M, MART1 M26, tyrosinase 370D, CMV pp65, influenza MP, and EBV BLF, respectively). The resulting IFNγ responses varied by antigen-specific T cell population.
Discussion
Cellular immune responses are complex and multifaceted events involving a multitude of cell types, secreted factors, microenvironments, cell states, and temporal factors. Under certain conditions, such as cellular immune responses to viral infection, the response is capable of eradicating specific target cells that express aberrant proteins and programming a multicellular response by secreting proinflammatory cytokines. Both processes are mediated by factors secreted by T cells. However, endogenous or postvaccination immune responses to tumor-associated antigens are less predictable than the response to viral infections, and are generally less effective at eliminating the offending cells. This response may be due to lower-affinity TCRs expressed on tumor-associated antigen-specific T cells [17], inefficient T cell priming of these cells [21], and/or the presence of regulatory cytokines, factors, or cells [22]. This diversity is reflected in the wide range of clinical responses to experimental cancer vaccines [23]. Here, we provide a possible explanation for this heterogeneity. Using pMHC functional microarrays to analyze viable patient T cells, we demonstrate a wide variation in the responses of tumor-associated antigen-specific CD8+ T cells following tumor peptide vaccination.
The mechanism that controls which specific factors are secreted in response to T cell activation and the impact of different functional profiles on the overall clearance of tumor has not been established, but the remarkable heterogeneity of these responses highlights the importance and the challenge of understanding these relationships. The differences in T cell behavior, as measured by multiple secreted factors, may stem from differences between melanoma cells from different patients, and the regulation of their T cell responses to melanoma antigens. Melanoma cells themselves may shape the behavior of tumor-associated antigen-specific T cells via secreted factors, or cell-contact [24,25]. Recognizing differences in functional responses between patients and between different antigen-specific T cells within a patient should help guide the development of cancer vaccines by providing causal relationships between treatment and clinical outcome, thereby accelerating the testing of different vaccine strategies.
Although this study is limited in scope, and does not allow us to link specific secretion response profiles to clinical outcomes, the results do suggest hypotheses that can be tested in expanded studies. One such possibility is that active secretion of both IFNγ and TNFα in response to tumor-associated antigen recognition may be necessary for effective tumor clearance. As one of its many functions, TNFα can mediate inflammation and promote T cell priming [26–28]. Similarly, IFNγ can mediate increased MHC class I expression on the cell surface and increase CD4+ T cell help by shifting toward a TH1 phenotype [29,30]. Without the involvement of both factors, it is possible that a threshold level of inflammation and effector activity is not reached. Another possibility is that dual IFNγ and TNFα secretion are associated with other secreted factors of critical importance. What is most important at this point, however, is that the data described here show that cytotoxic cells with identical specificity can have diverse functional response profiles. This heterogeneity is likely to have profound consequences for the functional specificity and clinical efficacy of cellular immune responses and may mirror the heterogeneity in clinical outcomes seen in essentially all of the immunotherapy trials performed to date [31,32]. With the methodology described here, we should be in an excellent position to determine what immune response profile correlates best with a positive clinical outcome.
The functional responses seen here do not seem to fit into easily categorized “good” or “bad” response profiles. Each individual patient appears to have a unique signature of functional responses. This is in contrast to preliminary analysis of anti-influenza T cell responses following vaccination (DSC and MMD, unpublished data). Furthermore, the variation in the responses is both patient-specific and independently antigen-specific. For example, individuals who responded to gp100 209–2M with strong IFNγ, but weak TNFα secretion (e.g., patient 10710), could respond to MART1 M26 with very strong IFNγ and TNFα secretion, in the same CD8+ T cell sample analyzed on the same array. This was also true of the variation in CD8+ T cell responses to viral antigens, in the absence of vaccination. As all patients underwent a similar melanoma vaccination protocol, these findings suggest that T cell populations with different antigen specificities are differentially regulated in the same patient at the same time, perhaps a major source of the variation in functional responses. Interestingly, the strong IFNγ and TNFα response to gp100 209–2M/HLA-A2.1 and MART1 M26/HLA-A2.1 seen in patient sample 10735 (after 6 mo) was still present in sample 10794 (same patient, after month 12).
The surprisingly multidimensional variation in the molecular specificity of the cellular immune response to a peptide vaccine raises the question of how and why these diverse variations arise. One possibility is that T cells acquire specific molecular programs based upon specific cues or signals accompanying or following the encounter with cognate antigen. Such signals could be mediated by secreted factors, or require cell-cell contact and might originate from antigen-presenting cells, CD4+ T cells, or a local complex inflammatory response. We refer to such a possible system as the “acquisition model,” in which incremental gains in their repertoire of functional responses (e.g., in the repertoire of effectors secreted in response to antigen stimulation) would change the CD8+ T cell's ability to effect killing and inflammation upon activation. An alternative possibility is that all T cells emerge from priming with the same level of functionality, capable of mediating a potent effector response. Following priming, their function could change with time, lack of stimulation, or the action of regulatory factors, causing them to slowly lose their ability to respond effectively to antigen recognition. We refer to this possibility as the “decay model” (Figure 8). Whether any stereotypical, ordered progression for either acquisition or decay of the response repertoire exists is unclear. However, our data do not support a single stereotyped progression, as T cells from some patients appear to respond with IFNγ, but not TNFα, while others respond with TNFα, but not IFNγ. One final possibility, an “intrinsic model,” is that the differences in functionality presented here are not evolving, but rather reflect the selection of T cells with specific predetermined molecular phenotypes. In this case, TCR affinity and other structural phenotypes might account for differences in function following activation. Changes in functional profiles in this model would reflect the emergence of new antigen-specific clones.
Figure 8 Two Models of T-Cell Function
Acquisition and decay models depict two possible mechanisms that account for variability in the factors secreted by activated CD8+ T lymphocytes in response to antigen recognition. Acquisition refers to independent, sequential increases in responsiveness to activation, triggered by both cellular and secreted signals. Decay accounts for maximally functional T cells immediately upon completion of priming, after which signals, or time, lead to diminished responsiveness to activation.
Isolating Individual Events in Complex Immune System Responses
The cellular arm of the adaptive immune response is based on specific recognition of target antigens presented on the surface of altered cells and subsequent triggering of a complex scenario of responses, collectively termed effector function. To investigate the cellular immune response to antigen recognition, we have developed a high-throughput multiparametric platform that simultaneously captures antigen-specific T cells and facilitates parallel induction and monitoring of distinct secreted factors from multiple T cell specificities [33,34]. A similar approach that has been used for the capture and analysis of antigen-specific T cell clones was recently reported by Stone, et al. [35]. However, the technique described here differs from that of Stone and colleagues in several important aspects, including the selection of a surface with lower cellular binding characteristics and greater protein loading capacity. We have found these features to be critical to the detection of rare populations of antigen-specific T cells, and their secreted proteins, from clinical specimens. Detection of antigen-specific T cell populations on this platform compares favorably with approaches such as pMHC tetramer staining followed by flow cytometry. We have noted a similar level of sensitivity for reproducible detection of rare cell populations. Analysis of a single clinical sample on the pMHC cellular microarray includes isolation, quantitation, and activation of antigen-specific T cells, followed by characterization of secreted proteins with single-cell resolution. This type of analysis is impractical or impossible to perform with more traditional approaches, such as pMHC tetramer staining, ELISpot [36], and cytokine flow cytometry [37]. Unlike ELISpot assays or cytokine capture arrays, the pMHC microarray immobilizes specific cells prior to functional characterization. In addition, due to a higher concentration of the coprinted detector probe antibodies, the secreted factors are captured and subsequently detected in close proximity to the secreting cells, with minimal dilution. The resultant signal is detectable in a physical pattern that may provide further clues to their physiologic roles and mechanisms of action.
In some cases, the presence of unresponsive (i.e., nonsecreting) cells can also be determined based on a characteristic signature of a nonfluorescent, cell-shaped region embedded within a brighter region (see Figure 3). By combining isolation with activation, antigen-specific T cells can be studied under controlled environments, where the influence of a specific factor or cell type can be ascertained. Cospotting of additional membrane-bound ligands (e.g., B7–1, ICAM), or even secreted factors (e.g., IFNγ, TGFβ, IL-2, IL-15) may further help to elucidate the complex network of interactions underlying T cell reactivity or lack thereof. The spatial resolution of secreted factor detection on a pMHC microarray is sufficiently high to distinguish between different factors based on the characteristic signature of secretion. For example, the IFNγ signature appears as a focal secretion characterized by an intense core, beyond which the signal decays sharply (Figure 7). In contrast, the appearance of captured TNFα is characterized by a clearly demarcated ring that appears outside the edge of bound or previously bound cells. These patterns suggest that IFNγ secretion is polarized toward the target cell, whereas TNFα is not detectable at the contact interface; thus, it may be broadcasting a signal rather than engaging in a dialog with the target cell (as also indicated by work in murine T cells; M. Huse, personal communication). The spatial resolution of detected cytokine secretion also reveals marked differences in the quantity of cytokine secreted by different T cells of the same antigen specificity. However, it is unclear what mechanisms control the quantity of cytokine secreted by a given responding T cell, or the significance of higher levels of secretion. One may speculate that higher numbers of tumor-associated antigen-specific T cells that secrete larger quantities of effector cytokines favor a more effective antitumor response.
In humans, the T cell component of the immune system comprises a tremendous number and diversity of T cells with different antigen specificities. Profiling a large and diverse range of T cell specificities on a single pMHC array platform can allow more thorough interrogation and understanding of ongoing responses from a single clinical sample. Here, we tested a strategy for constructing very large pMHC arrays by using a hybrid MHC (class I):Ig dimer construct that can be easily loaded with an arbitrary HLA- restricted peptide. The success of this approach suggests that a printable library of diverse pMHC constructs can be prepared simply by loading many different peptides in parallel.
The ability to generate functional profiles of cells present in clinical samples is not limited to characterization of antigen-specific T cell responses. This type of approach can be applied using a wide range of cell adhesion and signaling molecules to specifically capture cells in heterogeneous populations, and profile the molecules they secrete in response to specific signals, with single-cell resolution. As is true with responding CD8+ T cells, all cells use a diverse vocabulary of secreted proteins to communicate with other cells and modify their environment. Thus, the profiles that microarrays of this kind can provide may give us insight into what different populations of cells are capable of communicating and how they can manipulate their environment. The patterns that emerge from this type of systematic analysis could provide us an understanding of the language of molecular communication among cells, and how groups of cells orchestrate their behavior at the tissue level. This immediate improvement in the resolution with which we can profile the functional characteristics of living human cells is likely to find clinical applications beyond tumor immunology.
Supporting Information
Figure S1 Expanded gp100 and MART-1 Specific T Cell Functional Activity
The data used to generate Figure 5 are shown in this figure.
(A) Number of cells responding per spot to gp100 or MART1 with secretion of IFNγ or TNFα.
(B) Average intensity value gp100 spots and MART1 spots for each individual secreted factor. Spot fluorescence intensity was measured from each spot and averaged over all replicates. Intensity values are then expressed as a percentage of an arbitrary value specific for each secreted factor.
(25 KB PDF).
Click here for additional data file.
Patient Summary
Background
Malignant melanoma is a common skin cancer that is frequently fatal. One type of treatment being tested is vaccination with peptides (very short lengths of proteins) that are found on the surface of melanoma cells, in an attempt to produce an immune response to the tumor, which will then clear the tumor. However, the success of this treatment has been quite varied; in particular it has been very hard to predict which patients will respond to treatments and which will not.
Why Was This Study Done?
The authors wanted to understand why some people respond to vaccination and others do not. One way of measuring the response to vaccination is to look at the T cells (part of the body's immune response) that are specific to the melanoma proteins and which are produced after vaccination, and to measure how active they are in various ways.
What Did the Researchers Do and Find?
The researchers developed a way of catching individual T cells from the blood of patients onto a surface, stimulating the cells, and then measuring how the cells responded by measuring how much of various substances the cells produced. They tested the responses of T cells from ten patients who had been enrolled in a trial of vaccination against melanoma and found a wide variation in how much of various substances the patients' cells secreted; in patients whose tumors did not get bigger, it seemed necessary for the cells to secrete two particular compounds—interferon gamma and tumor necrosis factor alpha.
What Do These Findings Mean?
Unlike responses to vaccination for infectious diseases, the response to tumor vaccination is highly variable between patients. Studying the basis of these different responses may guide the future development of more effective vaccines.
Where Can I Get More Information Online?
Medline plus has links to many pages of information on melanoma:
http://www.nlm.nih.gov/medlineplus/melanoma.html
Cancer Bacup in the UK also has information for patients:
http://www.cancerbacup.org.uk/Cancertype/Melanoma/Resourcessupport/PatientInformationGuide
The National Cancer Institute has a page containing links to information on melanoma:
http://www.cancer.gov/cancertopics/types/melanoma
The National Institutes of Health has a searchable index of ongoing clinical trials for melanoma:
http://www.clinicaltrials.gov/ct/screen/SimpleSearch
We would like to thank Jonathan Fabian for assistance in reagent preparation, array preparation, and data analysis, M. S. Kuhns, M. Krogsgaard, B. F. Lillemeier, Y. Chien, P. J. Ebert, J. B. Huppa, Q. Li, and M. Huse for scientific discussions. This study was funded by grants from the Howard Hughes Medical Institute, the Human Frontier Science Program, and the National Institutes of Health. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Citation: Chen DS, Soen Y, Stuge TB, Lee PP, Weber JS, et al. (2005) Marked differences in human melanoma antigen-specific T cell responsiveness after vaccination using a functional microarray. PLoS Med 2(10): e265.
Abbreviations
CMFcalcium- and magnesium-free phosphate-buffered saline
DICdifferential interference contrast light microscopy
FACSfluorescence-activated cell sorter
FCSfetal calf serum
GM-CSFgranulocyte-macrophage colony-stimulating factor
HLAhuman leukocyte antigen
IFNinterferon
Igimmunoglobulin
MHCmajor histocompatibility complex
NEDno evidence of disease
PBMCperipheral blood mononuclear cell
pMHCpeptide-MHC
TCRT cell receptor
TNFtumor necrosis factor
VEGFvascular endothelial growth factor
==== Refs
References
Rosenberg SA Progress in human tumour immunology and immunotherapy Nature 2001 411 380 384 11357146
van der Bruggen P Traversari C Chomez P Lurquin C De Plaen E A gene encoding an antigen recognized by cytolytic T lymphocytes on a human melanoma Science 1991 254 1643 1647 1840703
Lee PP Yee C Savage PA Fong L Brockstedt D Characterization of circulating T cells specific for tumor-associated antigens in melanoma patients Nat Med 1999 5 677 685 10371507
Talebi T Weber JS Peptide vaccine trials for melanoma: Preclinical background and clinical results Semin Cancer Biol 2003 13 431 438 15001162
Kessels HW Wolkers MC Schumacher TN Adoptive transfer of T-cell immunity Trends Immunol 2002 23 264 269 12102748
Andersen MH Gehl J Reker S Pedersen LO Becker JC Dynamic changes of specific T cell responses to melanoma correlate with IL-2 administration Semin Cancer Biol 2003 13 449 459 15001164
Nasi ML Lieberman P Busam KJ Prieto V Panageas KS Intradermal injection of granulocyte-macrophage colony-stimulating factor (GM-CSF) in patients with metastatic melanoma recruits dendritic cells Cytokines Cell Mol Ther 1999 5 139 144 10641571
Kirkwood JM Manola J Ibrahim J Sondak V Ernstoff MS A pooled analysis of eastern cooperative oncology group and intergroup trials of adjuvant high-dose interferon for melanoma Clin Cancer Res 2004 10 1670 1677 15014018
Phan GQ Yang JC Sherry RM Hwu P Topalian SL Cancer regression and autoimmunity induced by cytotoxic T lymphocyte-associated antigen 4 blockade in patients with metastatic melanoma Proc Natl Acad Sci U S A 2003 100 8372 8377 12826605
Pardoll D Allison J Translation of cancer immunotherapies Nat Med 2004 10 1155 15516902
Altman JD Moss PA Goulder PJ Barouch DH McHeyzer-Williams MG Phenotypic analysis of antigen-specific T lymphocytes Science 1996 274 94 96 8810254
Maecker HT Dunn HS Suni MA Khatamzas E Pitcher CJ Use of overlapping peptide mixtures as antigens for cytokine flow cytometry J Immunol Methods 2001 255 27 40 11470284
Rubio V Stuge TB Singh N Betts MR Weber JS Ex vivo identification, isolation and analysis of tumor-cytolytic T cells Nat Med 2003 9 1377 1382 14528297
Speiser DE Pittet MJ Guillaume P Lubenow N Hoffman E Ex vivo analysis of human antigen-specific CD8+ T-cell responses: Quality assessment of fluorescent HLA-A2 multimer and interferon-gamma ELISPOT assays for patient immune monitoring J Immunother 2004 27 298 308 15235391
Johnson TR Massey RJ Deinhardt F Lymphocyte and antibody cytotoxicity to tumor cells measured by a micro-51 chromium release assay Immunol Commun 1972 1 247 261 4351536
Soen Y Chen DS Kraft DL Davis MM Brown PO Detection and characterization of cellular immune responses using peptide-MHC microarrays PLoS Biol 2003 1 e65 10.1371/journal.pbio.0000065 14691537
Stuge TB Holmes SP Saharan S Tuettenberg A Roederer M Diversity and recognition efficiency of T cell responses to cancer Plos Med 2004 1 e28 10.1371/journal.pmed.0010028 15578105
Wong R Lau R Chang J Kuus-Reichel T Brichard V Immune responses to a class II helper peptide epitope in patients with stage III/IV resected melanoma Clin Cancer Res 2004 10 5004 5013 15297401
Lee P Wang F Kuniyoshi J Rubio V Stuges T Effects of interleukin-12 on the immune response to a multipeptide vaccine for resected metastatic melanoma J Clin Oncol 2001 19 3836 3847 11559721
Dal Porto J Johansen TE Catipovic B Parfiit DJ Tuveson D A soluble divalent class I major histocompatibility complex molecule inhibits alloreactive T cells at nanomolar concentrations Proc Natl Acad Sci U S A 1993 90 6671 6675 8341685
Zimmermann VS Bondanza A Monno A Rovere-Querini P Corti A TNF-alpha coupled to membrane of apoptotic cells favors the cross-priming to melanoma antigens J Immunol 2004 172 2643 2650 14764739
Viguier M Lemaitre F Verola O Cho MS Gorochov G Foxp3 expressing CD4+CD25(high) regulatory T cells are overrepresented in human metastatic melanoma lymph nodes and inhibit the function of infiltrating T cells J Immunol 2004 173 1444 1453 15240741
Brichard VG Gerard C Melanoma vaccines: Achievements and perspectives Forum (Genova) 2003 13 144 154 14732881
Zha YY Blank C Gajewski TF Negative regulation of T-cell function by PD-1 Crit Rev Immunol 2004 24 229 238 15588223
Dummer W Bastian BC Ernst N Schanzle C Schwaaf A Interleukin-10 production in malignant melanoma: Preferential detection of IL-10-secreting tumor cells in metastatic lesions Int J Cancer 1996 66 607 610 8647620
Kwon B Kim BS Cho HR Park JE Kwon BS Involvement of tumor necrosis factor receptor superfamily(TNFRSF) members in the pathogenesis of inflammatory diseases Exp Mol Med 2003 35 8 16 12642898
Prevost-Blondel A Roth E Rosenthal FM Pircher H Crucial role of TNF-alpha in CD8 T cell-mediated elimination of 3LL-A9 Lewis lung carcinoma cells in vivo J Immunol 2000 164 3645 3651 10725721
Beutler BA The role of tumor necrosis factor in health and disease J Rheumatol Suppl 1999 57 16 21 10328138
Groettrup M Khan S Schwarz K Schmidtke G Interferon-gamma inducible exchanges of 20S proteasome active site subunits: Why? Biochimie 2001 83 367 372 11295499
Bernabei P Coccia EM Rigamonti L Bosticardo M Forni G Interferon-gamma receptor 2 expression as the deciding factor in human T, B, and myeloid cell proliferation or death J Leukoc Biol 2001 70 950 960 11739558
Chen DS Davis MM Cellular immunotherapy: Antigen recognition is just the beginning Springer Semin Immunopathol 2005 27 119 127 15834723
Antonia S Mule JJ Weber JS Current developments of immunotherapy in the clinic Curr Opin Immunol 2004 16 130 136 15023403
Chen DS Soen Y Davis MM Brown PO Functional and molecular profiling of heterogeneous tumor samples using a novel cellular microarray J Clin Oncol 2004 22 9507
Soen Y Chen DS Stuge TB Weber JS Lee PP A novel cellular microarray identifies functional deficiences in tumor-specific T cell responses J Clin Oncol 2004 22 2510 15197216
Stone JD Demkowicz WE Stern LJ HLA-restricted epitope identification and detection of functional T cell responses by using MHC-peptide and costimulatory microarrays Proc Natl Acad Sci U S A 2005 102 3744 3749 15728728
Miyahira Y Murata K Rodriguez D Rodriguez JR Esteban M Quantification of antigen specific CD8+ T cells using an ELISPOT assay J Immunol Methods 1995 181 45 54 7537312
Murali-Krishna K Altman JD Suresh M Sourdive D Zajac A In vivo dynamics of anti-viral CD8 T cell responses to different epitopes. An evaluation of bystander activation in primary and secondary responses to viral infection Adv Exp Med Biol 1998 452 123 142 9889966
Trapani JA Sutton VR Granzyme B: Pro-apoptotic, antiviral and antitumor functions Curr Opin Immunol 2003 15 533 543 14499262
Lieberman J Fan Z Nuclear war: The granzyme A-bomb Curr Opin Immunol 2003 15 553 559 14499264
Catalfamo M Henkart PA Perforin and the granule exocytosis cytotoxicity pathway Curr Opin Immunol 2003 15 522 527 14499260
Schroder K Hertzog PJ Ravasi T Hume DA Interferon-gamma: An overview of signals, mechanisms and functions J Leukoc Biol 2004 75 163 189 14525967
Aggarwal BB Signalling pathways of the TNF superfamily: A double-edged sword Nat Rev Immunol 2003 3 745 756 12949498
Lu L Zhu J Zheng Z Yan M Xu W Jak-STAT pathway is involved in the induction of TNF-beta gene during stimulation by IL-2 Eur J Immunol 1998 28 805 810 9541574
Kaijzel EL Bayley JP van Krugten MV Smith L van de Linde P Allele-specific quantification of tumor necrosis factor alpha (TNF) transcription and the role of promoter polymorphisms in rheumatoid arthritis patients and healthy individuals Genes Immun 2001 2 135 144 11426322
Rosenwasser LJ Biologic activities of IL-1 and its role in human disease J Allergy Clin Immunol 1998 102 344 350 9768571
Gronvik KO Andersson J The role of T cell growth stimulating factors in T cell triggering Immunol Rev 1980 51 35 59 6448799
Benczik M Gaffen SL The interleukin (IL)-2 family cytokines: Survival and proliferation signaling pathways in T lymphocytes Immunol Invest 2004 33 109 142 15195693
Heinrich PC Behrmann I Haan S Hermanns HM Muller-Newen G Principles of interleukin (IL)-6-type cytokine signalling and its regulation Biochem J 2003 374 1 20 12773095
Novelli F Casanova JL The role of IL-12, IL-23 and IFN-gamma in immunity to viruses Cytokine Growth Factor Rev 2004 15 367 377 15450252
Weng NP Liu K Catalfamo M Li Y Henkart PA IL-15 is a growth factor and an activator of CD8 memory T cells Ann N Y Acad Sci 2002 975 46 56 12538153
Akira S The role of IL-18 in innate immunity Curr Opin Immunol 2000 12 59 63 10679398
Hamilton JA GM-CSF in inflammation and autoimmunity Trends Immunol 2002 23 403 408 12133803
Widmer MB Morrissey PJ Goodwin RG Grabstein KH Park LS Lymphopoiesis and IL-7 Int J Cell Cloning 1990 8 Suppl 1 168 172 2182735
Farber JM Mig and IP-10: CXC chemokines that target lymphocytes J Leukoc Biol 1997 61 246 257 9060447
Lavender P Cousins D Lee T Regulation of Th2 cytokine gene transcription Chem Immunol 2000 78 16 29 12847715
Mocellin S Marincola F Rossi CR Nitti D Lise M The multifaceted relationship between IL-10 and adaptive immunity: Putting together the pieces of a puzzle Cytokine Growth Factor Rev 2004 15 61 76 14746814
Wahl SM Chen W TGF-beta: How tolerant can it be? Immunol Res 2003 28 167 179 14713712
Mukaida N Harada A Matsushima K Interleukin-8 (IL-8) and monocyte chemotactic and activating factor (MCAF/MCP-1), chemokines essentially involved in inflammatory and immune reactions Cytokine Growth Factor Rev 1998 9 9 23 9720753
Nishi J Maruyama I Increased expression of vascular endothelial growth factor (VEGF) in Castleman's disease: Proposed pathomechanism of vascular proliferation in the affected lymph node Leuk Lymphoma 2000 38 387 394 10830746
Li X Eriksson U Novel VEGF family members: VEGF-B, VEGF-C and VEGF-D Int J Biochem Cell Biol 2001 33 421 426 11312110
|
16162034
|
PMC1216330
|
CC BY
|
2021-01-05 10:40:26
|
no
|
PLoS Med. 2005 Oct 20; 2(10):e265
|
utf-8
|
PLoS Med
| 2,005 |
10.1371/journal.pmed.0020265
|
oa_comm
|
==== Front
PLoS MedPLoS MedpmedplosmedPLoS Medicine1549-12771549-1676Public Library of Science San Francisco, USA 1615930610.1371/journal.pmed.0020288Research ArticleNeuroscienceGeriatricsMedical ImagingMental HealthNeurology/NeurosurgeryDementiaGeriatric MedicineMedical ImagingNeurologyImpaired Cross-Modal Inhibition in Alzheimer Disease Cross-Modal Inhibition in Alzheimer DiseaseDrzezga Alexander
1
*Grimmer Timo
2
Peller Martin
1
3
Wermke Marc
1
Siebner Hartwig
3
Rauschecker Josef P
4
Schwaiger Markus
1
Kurz Alexander
2
1Department of Nuclear Medicine, Technische Universität München, Munich, Germany,2Department of Psychiatry and Psychotherapy, Technische Universität München, Munich, Germany3Department of Neurology, Christian-Albrechts-Universität, Kiel, Germany,4Department of Physiology and Biophysics, Georgetown University Medical Center, Washington, District of Columbia, United States of AmericaSmall Gary Academic EditorUniversity of California at Los AngelesUnited States of America*To whom correspondence should be addressed. E-mail: [email protected]
Competing Interests: The authors have declared that no competing interests exist.
Author Contributions: AD, MS, and AK designed the study. AD, TG, MP, MW, and HS analyzed the data. TG, MP, and MW enrolled patients. AD, TG, HS, JPR, and AK contributed to writing the paper.
10 2005 20 9 2005 2 10 e28818 10 2004 20 7 2005 Copyright: © 2005 Drzezga et al.2005This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are properly credited.
Alzheimer Disease: Failure to Tune Out Irrelevant Input?
Background
Successful cognitive performance depends not only on the activation of specific neuronal networks but also on selective suppression of task-irrelevant modalities, i.e., deactivation of non-required cerebral regions. This ability to suppress the activation of specific brain regions has, to our knowledge, never been systematically evaluated in patients with Alzheimer disease (AD). The aim of the current study was to evaluate both cerebral activation and deactivation in (1) healthy volunteers, (2) patients with mild cognitive impairment (MCI) who are at risk for AD, and (3) patients with moderate AD during active navigation, representing a cognitive task typically affected in AD.
Methods and Findings
Changes in regional cerebral blood flow (rCBF) were assessed with PET imaging during an active navigation task in a 3D virtual-reality environment. The task was based on visual cues exclusively; no auditory cues were provided. Age-matched groups of healthy individuals, patients with MCI, and patients with AD were examined. Specific differences in the activation patterns were observed in the three groups, with stronger activation of cerebellar portions and visual association cortex in controls and stronger activation of primary visual and frontal cortical areas in patients with MCI and AD. Highly significant bilateral decrease of rCBF in task-irrelevant auditory cortical regions was detected in healthy individuals during performance of the task. This rCBF decrease was interpreted as a cross-modal inhibitory effect. It was diminished in patients with MCI and completely absent in patients with AD. A regression analysis across all individuals revealed a clear positive relation between cognitive status (mini mental state examination score) and the extent of auditory cortical deactivation.
Conclusion
During active navigation, a high level of movement automation and an involvement of higher-order cerebral association functions were observed in healthy controls. Conversely, in patients with MCI and AD, increased cognitive effort and attention towards movement planning, as well as stronger involvement of lower-order cerebral systems, was found. Successful cognitive performance in healthy individuals is associated with deactivation of task-irrelevant cerebral regions, whereas the development of AD appears to be characterized by a progressive impairment of cross-modal cerebral deactivation functions. These changes may cause the generally decreased ability of patients with AD to direct attention primarily to the relevant cognitive modality.
When performing a navigation task with exclusively visual clues, healthy individuals de-activate areas of the brain involved in hearing. In individuals with Alzheimer Disease, no deactivation is seen.
==== Body
Introduction
It is a common hypothesis that information processing capacities of the brain are generally limited. Therefore, these capacities must be focused on the relevant sensory features and modalities. Modality-dependent selective attention mechanisms not only rely on the activation of specific neuronal networks, but also on selective suppression of task-irrelevant modalities, i.e., inhibition of less involved cortical areas [1]. This mechanism has been referred to as “cross-modal” inhibition. The underlying hypothesis is supported by several studies using neuroimaging tools. Particularly, cross-modal auditory/visual deactivation has been demonstrated repeatedly in healthy individuals [2–4].
In patients with Alzheimer disease (AD), multiple attention- and perception-related cognitive deficits are well-known. Functional neuroimaging studies offer a chance to evaluate the functional correlates of these deficits and to identify compensatory strategies. In some studies, changes in cerebral activation patterns have already been demonstrated in patients with AD [5,6]. However, to our knowledge, it has never been evaluated whether altered cerebral inhibitory processes are also involved in the cognitive deficits typically present in neurodegenerative disorders. Spatial navigation is among the first cognitive functions to be impaired in AD, leading to severe limitations in independent living. Perceptual and attention-related functions are particularly essential for this complex cognitive process, and previous neuroimaging studies have demonstrated extended activation of multi-modal cerebral systems in healthy individuals during navigational tasks [7]. However, information on functional cerebral changes during navigation in patients with AD is extremely limited.
In the present study, we examined changes of regional cerebral blood flow (rCBF) during active navigation in terms of increases (activation) and decreases (deactivation). We included healthy volunteers, patients with early AD, and patients with mild cognitive impairment (MCI). Patients with MCI were included because they represent a risk population for AD [8]. We selected a navigation task based on visual cues in a specially designed 3D virtual-reality (VR) environment, in order to simulate a real-life situation. All individuals had to navigate from a predefined starting point to a destination, and performance was measured in terms of time required to reach the end point. The purpose of the study was to identify differences between healthy individuals and the patient groups regarding cerebral activation and, particularly, cerebral deactivation during navigation.
Methods
Participants
Participants were recruited prospectively at the university outpatient clinic for memory research. Prior to starting the actual study protocol, all participants underwent an extensive evaluation including the following: (1) F18 FDG PET imaging (Siemens ECAT HR+ PET scanner, CTI, Knoxville, Tennessee, United States) and subsequent data analysis (NEUROSTAT, University of Michigan, Ann Arbor, Michigan, United States) to assess cerebral metabolic patterns [9,10]; (2) structural MRI (1.5 Siemens “Magnetom Symphony”) for exclusion of anatomic abnormalities, vascular pathology, or major atrophy; and (3) extended neuropsychological examination (CERAD-NP [Consortium to Establish a Registry for Alzheimer's Disease Neuropsychological battery], CDR [clinical dementia rating], GDS [geriatric depression scale], and ADL [activities of daily living]). Subsequently, participants were subdivided into three groups: (1) healthy volunteers, (2) patients with MCI, and (3) patients with moderate AD. This subdivision was performed according to established diagnostic criteria regarding neuropsychology, clinical examination and F-18 FDG PET results [9]. Based on the results of a previous study, patients with MCI but without any characteristic abnormalities in F-18 FDG PET and healthy participants who did show suspicious PET findings were excluded [9]. Using this approach, we attempted to enrich the MCI population with high-risk patients, in order to homogenize the groups and to exclude “healthy” controls with apparent brain pathology. Furthermore, participants with other neurological or psychiatric disorders or on medication with possible psychotropic effects were excluded. All participants were right-handed (assessed by the Edinburgh inventory; [11]). The study protocol was approved by the ethics committee of the Technische Universität München and the radiation protection authorities.
Materials
An IDL-based VR environment computer system was developed, operating on an SGI O2 workstation (Silicon Graphics, Mountain View, California, United States) for application in the PET scanner. A monitor was installed on a platform in front of the PET gantry, allowing direct view from the scanner. Three-dimensional stereoscopic perception was ensured using shutter glasses (StereoGraphics, San Rafael, California, United States). A SpaceMouse (SpaceBall, Labtec, Vancouver, Washington, United States) fixed laterally at the scanner allowed free movement within the system with minimal motion of the hand (Figure 1A). Based on this VR system, two different types of virtual environments were designed: (1) the control condition, constituted by a never-ending rectangular pathway without predetermined start or end points or turn-offs, and (2) the test condition, consisting of a trail across a sequence of four rooms in a complex spatial arrangement with additional exits, enabling false turn-offs into blind rooms (Figure 1B). Both conditions resembled the inside of a simple building. A plain design was selected and the general features of the environment (color, texture, and size) were kept identical to ensure an overall comparable visual impression and to avoid unnecessary distraction of the participants and contamination of the cognitive process of navigation. The entire system was evaluated and proved applicable in patients with AD in previous experiments [12].
Figure 1 Experimental Setup
(A) Experimental setup, showing a participant in the PET scanner during performance of a navigation task in the VR environment.
(B) Snapshot of the visual impression of the test condition in the virtual environment at the start point of the navigation task.
Procedure
Before the actual start of the scans, all participants were trained in the virtual environment outside of the scanner in order to familiarize them with the levels and handling of the system. Participants who were not able to perceive the virtual environment three-dimensionally or to operate the navigation system properly were excluded from the study. The performance of all participants was also observed during the entire PET examination by a study coordinator, to ensure correct transaction of every single scan. Participants performed the task in complete silence; the study coordinator was quietly watching the patient during the entire examination to report potential contamination with noise and to supervise the patient without the need of verbal interaction. Patients were instructed not to talk during the examination except in case of emergency. In the control condition, participants were instructed to steer along the never-ending pathway, thus no actual navigation was required but visual and motor demands were similar to the test condition. In the test condition, participants were asked to find their way from a start point to a predefined destination point. The two different conditions of the VR environment were presented in the scanner in a randomized order (4× control condition and 8× test condition). During the test conditions, performance (in terms of time required to reach the destination) was recorded.
O-15 Water PET Acquisition
A Siemens ECAT HR+ PET scanner was used for O-15 PET measurements (3D mode; total axial field of view, 15.5 cm). For each scan, 350 MBq of O-15 water was injected in a slow intravenous bolus after start of the VR paradigm using an infusion pump. Data acquisition was triggered by the peak of count rate. Each scan lasted 50 s for the measurement of rCBF. Twelve scans were performed in each participant. Attenuation-corrected data were reconstructed (63 slices; 128 × 128 pixel matrix; pixel size, 2.0 mm; plane separation, 2.42 mm).
Data Analysis
Statistical parametric mapping software (SPM 99, Wellcome Department of Cognitive Neurology, London, United Kingdom) was used for image realignment, transformation into standard stereotactic space, smoothing, and statistical analysis, resulting in 26 planes (pixel size, 2 × 2 mm; interplane distance, 4 mm), as previously described [13].
The effect of global differences in cerebral blood flow was removed by treating global activity as a confounding variable and using participant-specific scaling to a nominal grand mean global activity of 50 ml/100 g/min. Thus, data were adjusted for the global mean (rCBF), taking into account the repetitions across participants. The adjusted voxel values were then used for further statistical analysis [14].
The statistical analysis was performed according to the general linear model and the theory of Gaussian fields at each and every voxel using a mixed-effects model [13,15,16]. The resulting statistical parametric maps based on the t statistic were subsequently transformed into normally distributed statistical parametric (Z) maps [13].
A network of expected cortical activations was predefined based on previous publications on cerebral activation during visual spatial processing and active navigation [7,17–19]. This network included cerebral regions belonging to the ventral and dorsal streams of the visual system, such as primary and secondary visual cortex, inferotemporal cortex, ventrolateral prefrontal cortex, posterior parietal cortex, and dorsolateral prefrontal cortex. In addition, cerebral regions traditionally linked to spatial orientation were included in the network, such as the hippocampus (allocentric mapping, recognition of landmarks), posterior and superior parietal cortex (spatial attention, egocentric mapping, and optic flow processing; see above), and posterior cingulate cortex (spatial orientation). Finally, as the task required active movement, we expected activation of motor-associated systems such as premotor cortex, as well as supplementary and primary motor cortex (associated with movement planning and execution, respectively). Stereotactic coordinates of these anatomic regions were selected according to corresponding Brodmann areas (BAs) in the atlas of Talairach and Tournoux [20].
As mentioned previously, we expected to find cross-modal deactivation of auditory cortical areas. Therefore, as for cerebral activations, a network regarding deactivations was predefined based on previous data on the extent of auditory cortical areas [21,22] and on information about cross-modal deactivation of auditory areas during visually dominated tasks [2,23]. This network included BAs 41, 42, 21, and 22 in both hemispheres, representing primary auditory cortex (A1) and adjacent belt areas that have been linked to auditory associative functions [22,24]. Corresponding to former studies, analysis was mainly restricted to these areas [25]. Regarding the interindividual and interhemispheric variability of the location of the auditory cortical areas, we defined three-dimensional volumes independently for each hemisphere representing the probabilistic location of the primary auditory cortex according to Penhune et al. [21]. These volumes were transferred into the stereotactic reference system, to allow for correct anatomical assignment of the results. Within the predefined expected cortical activation and deactivation networks, all statistical results were based on a single-voxel z-threshold corresponding to p < 0.001, uncorrected for multiple comparisons. To our knowledge, no conclusive previous information on cerebral deactivation during visually guided navigation exists. Thus, only voxels surviving false discovery rate (FDR) correction for the entire volume at p < 0.05 were accepted in the statistical analysis of rCBF decrease between conditions within groups, in order to avoid false-positive results [26].
First the rCBF changes between the control and test conditions were examined in each of the different groups (control individuals, patients with MCI, and patients with AD). For the assessment of differences of rCBF changes between the groups, a “difference of differences” analysis was used. In all group analyses, the rating of navigation performance of participants (time required for passage from start to end point of the labyrinth) was used as a covariate of no interest, to diminish the possible effects due to individual differences of performance or subjective experience of the paradigm. Finally, a voxel-based linear regression analysis of cognitive performance (as measured with the mini mental state examination [MMSE]) and regional cerebral deactivation was carried out in the entire population, containing the participants from all three groups. We generated subtraction images of the individual patients (control condition) and put these images in relation to the MMSE scores of the patients, using a “covariate only” analysis. This analysis tested for voxels in the brain in which task-related deactivations showed a significant linear relationship with the MMSE scores. An a priori hypothesis for the location of the probable linear relation was defined for this analysis, based on the previously identified deactivation foci, and an uncorrected threshold of p < 0.001 was applied. We limited the linear regression analysis to clusters within the A1 (transverse gyrus of Heschl [TGH]). Additionally, a small volume correction was performed within a sphere of 20-mm radius, centered on the maximum deactivation in the TGH. The results of all analyses (maxima of the activation foci) were reported with the respective standard stereotactic coordinates according to Talairach and Tournoux [20].
Results
Based on the inclusion criteria, 32 participants were recruited and allocated to the three predefined age-matched groups: (1) 11 healthy volunteers (three female), (2) 11 patients with MCI (four female), and (3) ten patients with moderate AD (four female) (Table 1). No significant difference in age was detected (healthy versus MCI, p = 0.50 [95% confidence interval [CI], −5.75 to 11.39]; healthy versus AD, p = 0.43 [95% CI, −5.63 to 12.78]; MCI versus AD, p = 0.85 [95% CI, −7.54 to 9.07]). Regarding cognitive function, there was no significant difference in mean MMSE scores between control individuals and patients with MCI. However, patients with AD yielded significantly lower scores than controls (healthy versus MCI, p = 0.06 [95% CI, −2.97 to 0.06]; healthy versus AD, p < 0.001 [95% CI, −5.84 to −2.39]; MCI versus AD, p = 0.02 [95% CI, −4.87 to −0.46]).
Table 1 Participant Characteristics
Navigation Performance
Each of the 32 participants underwent a total of 12 O-15 water activation PET scans during active navigation and control conditions in the virtual environment, resulting in a total of 384 scans. All participants were able to understand the task and cope with the test paradigm requirements satisfactorily. The mean navigation time needed by the participants to reach the destination from the start point during test condition differed considerably between the groups (Table 1). In patients with MCI, performance was significantly impaired when compared to control individuals; still, they performed significantly better in the navigation task than patients with AD. The latter required significantly more time for accomplishment of the task, as compared to healthy control indivduals and patients with MCI (healthy versus MCI, p < 0.001 [95% CI, −87.44 to −57.54]; healthy versus AD, p < 0.001 [95% CI, −118.68 to −96.50]; MCI versus AD, p < 0.001 [95% CI, −47.64 to −22.55]).
Cerebral Activation
In the statistical group comparison between the control condition and navigation condition we found a significant increase in rCBF in a number of cortical regions associated with the predefined navigation network in all three groups (Table 2; see Figure 2). These increases included strong activation of posterior and superior parietal cortical areas, particularly the precuneus. Furthermore, all groups showed activation of visual areas. In the healthy group, activation of higher-order extrastriate visual areas (BA 19) was detected bilaterally. Less involvement of these areas in active navigation was found in patients with MCI (left hemisphere only), and none at all in patients with AD. Conversely, strong activation of primary visual and adjacent cortical areas (BA 17/18) was observed in patients with AD, to a lesser extent also in patients MCI, whereas no significant activation of these cerebral regions was detected in control indivduals. Motor activation included left sensorimotor cortex and supplementary motor cortex in the healthy control indivduals. Several cerebellar regions were activated during navigation in control individuals, whereas cerebellar activation was considerably less in patients with MCI and absent in patients with AD, and no activation of primary motor cortical areas was observed. However, activation of premotor and prefrontal areas (BA 8 and BA 10) was detected exclusively in patients with AD and MCI, respectively. Analysis of the difference of differences confirmed significantly stronger activation of cerebellar regions in healthy control individuals than in patients with MCI and AD, and stronger activation of extrastriatal visual areas than in patients with AD. In both the AD and MCI groups, the premotor and prefrontal cortical activations were found to be significantly stronger than in the healthy control group. Generally, the strongest differences were observed between patients with AD and control individuals, whereas patients with MCI showed specific similarities with both other groups. In none of the groups was any activation of auditory or auditory-associated cortical areas observed.
Figure 2 rCBF Changes in Control Individuals and Patients with MCI and AD during Navigation Task
Results are surface-rendered and superimposed on a standard MRI template. Green indicates significant increase of rCBF, and red indicates significant decrease of rCBF during active navigation (results are displayed at p < 0.001).
Table 2 Activation of Cortical Areas during Navigation
Cerebral Deactivation
In the statistical comparison between the control condition and navigation condition we found a significant decrease in rCBF in bilateral auditory cortical areas in the healthy control group during active navigation (Table 3; Figures 3 and 4A). This suppression of auditory-associated areas in healthy volunteers was demonstrated bilaterally, but with a predominance of the right hemisphere. A major portion of the deactivated area was located bilaterally within the predefined probability map of the A1, according to previously published criteria [21] (Figure 3). In addition, auditory cortex in the adjacent belt areas [22] was also deactivated. The strongest deactivation was found ventrolaterally to A1 in the right hemisphere, but deactivations were also observed in belt areas in the dorsal and caudal vicinity of A1 (Table 3; Figure 3).
Figure 3 Auditory Cortical Deactivation
Areas of significant deactivation in healthy volunteers (p < 0.001) during navigation are demonstrated in black on a glass-brain display. The probabilistic volume of A1 according to Penhune et al. [21] is outlined in green (right hemisphere) and red (left hemisphere). Aspects are (A) left lateral, (B) cranial, (C) right lateral, and (D) dorsal.
Figure 4 Deactivation in Control Individuals and Patients with MCI and AD
Results are superimposed on a standard MRI template. Yellow indicates significant cerebral deactivations during active navigation: (A) axial slices, cranial aspect; (B) coronal slices, dorsal aspect (results are displayed at p < 0.005, for illustration purposes).
Table 3 Deactivation of Cortical Areas during Navigation
Interestingly, only faint deactivation was observed in patients with MCI (Table 3; Figure 4B), and exclusively in auditory belt regions. These findings did not survive FDR correction for the entire volume. The difference-of-differences analysis confirmed that deactivation effects in right-hemispheric auditory cortical regions were significantly stronger in controls than in patients with MCI.
Moreover, patients with AD did not show any significant deactivation of task-irrelevant auditory cortical areas during visually based navigation (Table 3; Figure 4). Again, a difference-of-differences analysis confirmed that deactivation effects in right temporal auditory regions in controls were significantly stronger in control individuals than in patients with AD, whereas no significant difference between patients with MCI and AD was found (Table 3; Figure 4). Little extra-auditory deactivation was observed, and was seen exclusively in left superior prefrontal cortex (BA 9) of healthy participants and in left sensorimotor cortex (BA 3/4) of patients with MCI, with the latter effect not surviving the FDR correction.
Linear Regression Analysis
A voxel-based regression analysis of regional cerebral deactivation with cognitive performance (as measured with the MMSE score) was performed, in order to identify a possible connection between the degree of overall cognitive impairment and the deactivation capabilities. In this analysis, a clear linear relation of the MMSE scores with deactivation of left auditory cortex (BA 41) was indeed detected (Table 3; Figure 5), pointing to a direct association of cognitive performance with cortical deactivation. This cluster survived FDR correction with a p < 0.001 in the small volume correction. No additional significant linear relation was observed. Additionally, we performed a correlation analysis using the MMSE scores and the regional adjusted rCBF response at the location with Talairach coordinates x, −56; y, −14; and z, −2. This analysis revealed a coefficient of correlation r of −0.67 (p < 0.001 [95% CI, −0.83 to −0.42]).
Figure 5 Positive Linear Regression of Auditory Cortical Deactivation with Cognitive Performance (MMSE) in All Participants
Left: results are superimposed on a standard MRI template. Yellow indicates a significant relationship of cerebral deactivation with the MMSE: (A) axial slice, cranial aspect; (B) coronal slice, dorsal aspect (results are displayed at p < 0.005, for illustration purposes).
Right: Regression analysis (red indicates regression line) of the fitted and adjusted rCBF response (blue points, arbitrary units) to active navigation in relation to the MMSE score at the position of the significant cluster (Talairach coordinates x, −56; y, −14; z, −2; p < 0.001, uncorrected).
Discussion
Activation of Cortical Regions during Visual Navigation
In previous studies, extended cerebral networks associated with human navigation have been identified [7,17,18]. In the current study we observed a regional increase of rCBF, interpreted as cerebral activation in several elements of these predefined networks, in three groups of individuals (healthy control individuals, patients with MCI, and patients with AD) during active navigation in a VR environment. Consistent with previous studies, strong activation of parietal cortical areas was found during the navigation task. Generally, a role of the parietal cortex in visual spatial functions including spatial attention is widely accepted [1,27]. Parietal cortical areas are part of the dorsal stream of the visual cortical system and as such are involved in the processing of spatial and motion information [28].
In a previous study on navigation in VR, activation of medial parietal areas was identified and associated with optic flow induced by egocentric movement [7]. The medial parietal activations found in our study conform to such functions. In addition, recent studies suggest a role for medial parietal cortex (in particular the precuneus) in episodic memory retrieval, which may also be required for navigation [29].
In addition to parietal cortex, activation of visual cortical regions was observed in all three groups. In the healthy group, extended activation of extrastriate cortex within bilateral BA 19 was detected, i.e., in a cortical region that contains the visual association areas V3–V5, which have been linked to higher-order visual functions, such as perception of movement and shape and form of moving objects [30]. Less involvement of these areas in active navigation was found in patients with MCI and none at all in patients with AD. Conversely, strong activation of primary visual cortex (BA 17) and adjacent BA 18 was observed in patients with AD, and to a lesser extent also in MCI, whereas no significant activation of these early visual regions was detected in control individuals. This suggests that activation induced by the increased cognitive demand of the navigation task is restricted to lower-order systems of the visual hierarchy in patients with MCI and AD and does not lead to recruitment of higher-order visual association areas as in control individuals. This may be a consequence either of ongoing neurodegenerative changes, disconnection phenomena, or a shift in attentional priority and is also consistent with previous observations [31].
In healthy individuals, increased activation of left primary sensorimotor cortex and paramedian premotor cortex (supplementary motor cortex) was observed during active navigation, consistent with a more resolute motor execution (performed with the right hand) than in the control condition. No comparable activation was observed in patients with MCI or AD, consistent with a more hesitant performance of the task in these groups. Furthermore, stronger and more extended activation of cerebellar areas was detected in healthy volunteers than in patients with MCI and AD. Increasing evidence has been provided for a role of the cerebellum in cognitive functions, including navigation [32], which could explain the activation detected in the current study. Additionally, a major role of the cerebellum in automation of executive functions is widely accepted [33]. It appears plausible that healthy participants employed a more automated strategy to solve the navigation task, whereas patients with AD, and even patients with MCI, shifted to a less automated, more attentionally demanding approach. This hypothesis is supported by the exclusive activation of prefrontal cortical regions in patients with MCI (BA 8 and BA 10) and AD (BA 8). Prefrontal cortical areas have been associated with executive functions such as planning, problem solving, and reasoning but also with spatial attention and working memory [2,34,35]. In particular, the region in superior prefrontal cortex approximately anterior to the frontal eye field, which was activated in both patient groups, is thought to be specialized for spatial working memory [36,37].
A general limitation of the current study may be that an interaction between participant group and control versus test condition can not be completely excluded, as we could not measure performance of participants during control condition. However, the control condition was only used as a baseline (for correction of visual and sensorimotor effects) and no actual “navigation” was required.
Deactivation of Auditory Regions during Visual Navigation in Healthy Control Individuals
In the virtual navigation task used in our study, all external information regarding current localization was based on visual input; auditory cues were neither provided nor required. Therefore, we expected cross-modal inhibition of those cortical regions that are typically associated with cerebral processing of auditory information. Indeed, we found a striking decrease in rCBF in bilateral auditory cortical areas in the healthy control group during active navigation.
Corresponding to many previous studies we interpreted this regional decrease in cerebral rCBF as local cerebral deactivation [3,4,38]. In a recent study, regional decrease of fMRI BOLD signal was explicitly associated with true cortical deactivation, which supports our interpretation [39]. Even if the observed rCBF changes in the current study represented only regions spared from global activation effects, they would suggest cerebral inhibitory effects, justifying the term “deactivation.” The deactivation effect may also involve inhibitory neurons, but considering the small number of inhibitory cortical neurons in general (∼10%), a contribution to the observed results in our study is not probable [40].
Interference between visual and auditory information in the brain is well-documented [25], and cross-modal inhibition between the two modalities has been discussed in prior studies [23]. In particular, it must be taken into account that all task-relevant information in our study was derived from the visual stimuli in the virtual environment, whereas any auditory signals would have originated from the real local environment and, thus, would have been irrelevant or even misleading. Consequently, in our study, a suppression of auditory-associated areas in healthy volunteers was demonstrated predominantly in the right hemisphere, which is widely thought to be “specialized” for spatial attention [27,41], including in the auditory domain [42]. A major portion of the deactivated areas was located bilaterally within A1, as defined previously [21]. In addition, auditory cortex in the adjacent belt areas was deactivated [22,43]. The maximal deactivation was found anterolateral to A1, in a region that has been associated in animal and human studies with the identification of auditory objects [22,44,45]. However, deactivation was also observed posterior to A1 in regions that have been associated with the processing of auditory spatial information [22,42,46]. It appears, therefore, that in healthy individuals interference from conflicting auditory signals is minimized during processing of a visual navigation task by suppressing auditory activity at relatively early stages of cortical processing.
Absence of Auditory Deactivation in Patients with AD and MCI
The loss of navigation abilities represents one of the most disabling cognitive impairments in AD. The tendency to become lost can be found in almost all patients with AD in the course of the disease. Accordingly, in our study a significantly impaired performance in the navigation task was observed in patients with AD as compared to healthy control individuals and patients with MCI.
Interestingly, in contrast to healthy control individuals, patients with AD did not show any significant deactivation of task-irrelevant auditory cortical areas during visually based navigation. Often, the loss of orientation in AD has been associated with impaired memory function [47]. More recently, however, navigation impairment in AD has been attributed primarily to impaired perceptual abilities, such as optic flow discrimination, associated with constraints to attention and disturbed multi-sensory integration [48–50]. Perception of any type is inseparably related to attention. Multiple attentional deficits, concerning spatial and selective attention, as well as the inability to shift attention across levels of perceptual organization, are well-documented in patients with AD [51–54]. The principle of suppression of irrelevant input has been discussed as a mechanism essential for selective attention to a certain modality [1].
Therefore, the impaired visual navigation performance of patients with AD may be related to the striking absence of auditory deactivation observed in these individuals. Parasuraman et al. associated the abnormal disengagement/shifting abilities in patients with AD with a possible dysfunction of cortico-cortical networks [53]. This is of particular interest, as a dysfunction in cortical regions that are not primarily affected by the Alzheimer pathology has been detected in the current study [55].
Absence of Cross-Modal Deactivation as an Early Indicator of AD
Unfortunately, spatial disorientation is not a late symptom of AD, but has been shown to be one of the first signs of the disease. Impaired navigation abilities are observed in patients with MCI [56]. Like in patients with AD, perceptual deficits, such as impaired visual motion processing, have been associated with poorer performance in spatial navigation tests in patients with MCI [56]. Again, the affected perceptual functions are tightly associated with attention. Similarly to in patients with AD, attentional deficits have been documented in patients with MCI [48,57].
In our study, patients with MCI still performed significantly better than patients with AD in the navigation task; however, their performance was significantly impaired compared to that of control individuals. At the same time, only faint deactivation in auditory belt regions was observed in MCI patients. Thus, impaired attention-focusing properties may also be involved in disorientation in MCI. Considering that the examined patients with MCI did not yet fulfill criteria for dementia in neuropsychological assessment, this observation is particularly interesting, and it indicates that reduced deactivation of auditory areas in a visual navigation task could be used as an early indicator of AD.
An ongoing follow-up study using the same examination paradigm as in the current study confirms this hypothesis. A preliminary analysis of baseline data from patients with MCI (n = 15) revealed significant auditory deactivation (BA 41; Talairach coordinates x, 62; y, −4; and z, 4) exclusively in the subgroup of patients with MCI who remained cognitively stable (n = 8) within the observation period (2 y). In contrast, the group of patients who proceeded to AD (n = 7) during this time, showed no auditory deactivation at the initial evaluation. Differences in cerebral activation patterns have been previously considered to be useful for early diagnosis of AD [58]; the present results indicate that it may be even more effective to direct one's attention to changes in cerebral deactivation patterns.
Other Cortical Deactivations
Hardly any extra-auditory deactivation was observed in any of the groups. Exclusively in healthy participants a deactivation of a region located in the prefrontal cortex (BA 9) was observed in a region that has been associated with visuo-spatial attention, particulary to stimuli at selected peripheral locations [35]. Since the navigation task in the current study required attention directed to the center of gaze, a deactivation of this periphery-oriented attention area appears economical. The focal rCBF decrease in left sensorimotor cortex in patients with MCI did not survive FDR correction and may be the result of less resolute motor execution in the navigation condition than in the control condition.
Relation of Auditory Deactivation to Cognitive Performance
In addition to statistical group comparison, we performed a voxel-based regression analysis in the entire study population, to further evaluate a possible association of cognitive function with the deactivation properties. In this analysis, individual cognitive performance (MMSE score) across all participants was shown to be directly associated with the intensity of cerebral deactivation in the auditory cortex. This result is highly specific, considering that an association of cognitive function with rCBF changes was not observed in any other cerebral region. This implies that physiologic cerebral deactivation capabilities are indeed essential for normal brain function and appear to be progressively affected in ongoing neurodegeneration. Interestingly, the identified linear regression effect was only significant in the left hemisphere, i.e., the dominant side of the brain in our purely right-handed population. Thus, the changes in deactivation functions in relation to overall cognitive impairment may be most clearly expressed in the dominant hemisphere.
Impaired cortical inhibitory capabilities in patients with AD appear plausible for two reasons. First, a stronger vulnerability of the neocortical inhibitory system has often been suggested. Second, recent studies were able to demonstrate cortical disinhibition in AD and could relate it to a cholinergic deficit [59,60].
Conclusions
The results of this study indicate that altered deactivation patterns must be taken into account in activation studies comparing effects in patient groups versus control individuals, in order to avoid misinterpretation of differences. We have demonstrated that navigation based on visual cues induces a deactivation of auditory cortical areas in healthy individuals. These deactivation effects are impaired in patients with MCI and absent in patients with AD, pointing to a progressive inability to tune out irrelevant input and to focus attention on the task-relevant modalities. Thus, the orientation disability in the outside world seen in patients with AD may in fact be partially based on the inability to selectively orient spatial attention to task-relevant internal representations of perceptual stimuli.
Patient Summary
Background
Problems with finding one's way are one of the early signs of Alzheimer disease (AD). Researchers can measure how good people are at “spatial orientation” by asking them to solve navigation tasks on the computer. These are similar to virtual-reality video games in which one needs to find one's way based on a set of cues. Using brain imaging technology, scientists can then observe what is going on in the brains of people while they solve such navigation tasks.
Why Was This Study Done?
Scientists used to think of AD as mostly a disease of memory loss, and loss of orientation was considered to be mostly a memory problem. More recently, however, navigation problems have been linked to problems with perception and with paying attention. In this context, the researchers wanted to compare the brain activities in people with and without AD while they completed such navigation tasks. They were looking for which areas of the brain were activated, and also whether others (which had nothing to do with the navigation task) were deactivated. The idea was that maybe to do the task well, individuals need to both activate relevant areas of the brain and also deactivate others to be able to focus overall brain activity on processes that would help with the task at hand.
What Did the Researchers Do and Find?
They studied brain activity in three groups of people while the individuals solved a navigation task. The first group consisted of 11 healthy individuals, the second of ten individuals with mild cognitive impairment, and the third of 11 individuals with AD. The cues that guided them through the navigation task were exclusively visual, and all of the participants were able to understand the cues. They found that the activation patterns were different between the three groups. The healthy individuals showed more activity in parts of the brain thought to be involved in processing visual cues, whereas the other two groups showed stronger activation in the parts involved with visual perception. Moreover, the researchers found differences in the deactivation patterns, mostly in the area of the brain that deals with sound and hearing. In healthy individuals these areas shut down to some extent during the visual navigation task, but this effect was diminished in individuals with mild cognitive impairment, and individuals with AD showed no signs of deactivation in these “task-irrelevant” areas.
What Does This Mean?
This study suggests that in addition to difficulty integrating complex information, inability to focus activity in the relevant parts of the brain might be contributing to some of the orientation problems in patients with AD. It is not clear whether cognitive therapy could strengthen the ability to “focus one's brain,” but it seems an idea worth pursuing. The researchers also suggest that reduced deactivation of auditory areas in visual navigation tasks might be useful as an early indicator of AD, and they are currently doing studies to test this idea.
Additional Online Information
The following Web sites contain information on AD in general and describe some of the changes in the brains of patients with AD.
Alzheimer Research Forum:
http://www.alzforum.org/home.asp
US National Institute on Aging:
http://www.nia.nih.gov/
Alzheimer's Disease Education and Referral Center (search for “brain imaging”):
http://www.alzheimers.org/pubs/adfact.html
Alzheimer's Association (search for “brain imaging”):
http://www.alz.org/
US National Institute of Neurological Disorders and Stroke's Web page on AD:
http://www.ninds.nih.gov/disorders/alzheimersdisease/alzheimersdisease.htm
We thank Brigitte Dzewas and Coletta Kruschke for their technical assistance and the radiochemistry group for their reliable supply of radiopharmaceuticals. We are obliged to Denise Lee for the very careful review of the manuscript and to Wolfgang Kloiber for technical support. Furthermore, we thank Dr. Stephan Nekolla and Dr. Istvan Nagy for their essential contribution to the development of the VR environment computer system. This study has been supported in part by a grant for clinical research from the Kommission für klinische Forschung (KKF), Technische Universität München (AD), and by a Research Award from the Alexander-von-Humboldt Foundation (JPR). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Citation: Drzezga A, Grimmer T, Peller M, Wermke M, Siebner H, et al. (2005) Impaired cross-modal inhibition in Alzheimer disease. PLoS Med 2(10): e288.
Abbreviations
A1primary auditory cortex
ADAlzheimer disease
BABrodmann area
CIconfidence interval
FDRfalse discovery rate
MCImild cognitive impairment
MMSEmini mental state examination
rCBFregional cerebral blood flow
TGHtransverse gyrus of Heschl
VRvirtual reality
==== Refs
References
Posner MI Petersen SE The attention system of the human brain Annu Rev Neurosci 1990 13 25 42 2183676
Haxby JV Horwitz B Ungerleider LG Maisog JM Pietrini P The functional organization of human extrastriate cortex: A PET-rCBF study of selective attention to faces and locations J Neurosci 1994 14 6336 6353 7965040
Kawashima R O'Sullivan BT Roland PE Positron-emission tomography studies of cross-modality inhibition in selective attentional tasks: Closing the “mind's eye” Proc Natl Acad Sci U S A 1995 92 5969 5972 7597062
Laurienti PJ Burdette JH Wallace MT Yen YF Field AS Deactivation of sensory-specific cortex by cross-modal stimuli J Cogn Neurosci 2002 14 420 429 11970801
Bookheimer SY Strojwas MH Cohen MS Saunders AM Pericak-Vance MA Patterns of brain activation in people at risk for Alzheimer's disease N Engl J Med 2000 343 450 456 10944562
Pietrini P Furey ML Alexander GE Mentis MJ Dani A Association between brain functional failure and dementia severity in Alzheimer's disease: Resting versus stimulation PET study Am J Psychiatry 1999 156 470 473 10080567
Maguire EA Burgess N Donnett JG Frackowiak RS Frith CD Knowing where and getting there: A human navigation network Science 1998 280 921 924 9572740
Petersen RC Smith GE Waring SC Ivnik RJ Tangalos EG Mild cognitive impairment. Cinical characterization and outcome Arch Neurol 1999 56 303 308 10190820
Drzezga A Lautenschlager N Siebner H Riemenschneider M Willoch F Cerebral metabolic changes accompanying conversion of mild cognitive impairment into Alzheimer's disease: A PET follow-up study Eur J Nucl Med Mol Imaging 2003 30 1104 1113 12764551
Minoshima S Frey KA Koeppe RA Foster NL Kuhl DE A diagnostic approach in Alzheimer's disease using three-dimensional stereotactic surface projections of fluorine-18–FDG PET J Nucl Med 1995 36 1238 1248 7790950
Oldfield RC The assessment and analysis of handedness: The Edinburgh inventory Neuropsychologia 1971 9 97 113 5146491
Drzezga A Grimmer T Peller M Wermke M Nekolla S Functional reorganization of cerebral activation in patients with MCI and Alzheimer's disease during navigation in a virtual reality environment [abstract] Neuroimage 9th International Conference on Functional Mapping of the Human Brain 2003 June 19–22 New York, New York, United States CD-Rom
Friston KJ Holmes AP Worsley KJ Poline JP Frith CD Statistical parametric mapping in functional imaging: A general linear approach Hum Brain Mapp 1995 2 189 210
Friston KJ Frith CD Liddle PF Dolan RJ Lammertsma AA The relationship between global and local changes in PET scans J Cereb Blood Flow Metab 1990 10 458 466 2347879
Friston KJ Frith CD Liddle PF Frackowiak RS Comparing functional (PET) images: The assessment of significant change J Cereb Blood Flow Metab 1991 11 690 699 2050758
Worsley KJ Evans AC Marrett S Neelin P A three-dimensional statistical analysis for CBF activation studies in human brain J Cereb Blood Flow Metab 1992 12 900 918 1400644
Hartley T Maguire EA Spiers HJ Burgess N The well-worn route and the path less traveled: Distinct neural bases of route following and wayfinding in humans Neuron 2003 37 877 888 12628177
Maguire EA Burgess N O'Keefe J Human spatial navigation: Cognitive maps, sexual dimorphism, and neural substrates Curr Opin Neurobiol 1999 9 171 177 10322179
Kawashima R Imaizumi S Mori K Okada K Goto R Selective visual and auditory attention toward utterances—A PET study Neuroimage 1999 10 209 215 10417253
Talairach J Tournoux P Co-planar stereotaxic atlas of the human brain 1988 New York Thieme 132
Penhune VB Zatorre RJ MacDonald JD Evans AC Interhemispheric anatomical differences in human primary auditory cortex: Probabilistic mapping and volume measurement from magnetic resonance scans Cereb Cortex 1996 6 661 672 8921202
Rauschecker JP Tian B Mechanisms and streams for processing of “what” and “where” in auditory cortex Proc Natl Acad Sci U S A 2000 97 11800 11806 11050212
Shulman GL Corbetta M Buckner RL Raichle ME Fiez JA Top-down modulation of early sensory cortex Cereb Cortex 1997 7 193 206 9143441
Kaas JH Hackett TA Subdivisions of auditory cortex and processing streams in primates Proc Natl Acad Sci U S A 2000 97 11793 11799 11050211
Finney EM Fine I Dobkins KR Visual stimuli activate auditory cortex in the deaf Nat Neurosci 2001 4 1171 1173 11704763
Genovese CR Lazar NA Nichols T Thresholding of statistical maps in functional neuroimaging using the false discovery rate Neuroimage 2002 15 870 878 11906227
Mesulam MM Spatial attention and neglect: Parietal, frontal and cingulate contributions to the mental representation and attentional targeting of salient extrapersonal events Philos Trans R Soc Lond B Biol Sci 1999 354 1325 1346 10466154
Ungerleider LG Mishkin M Goodale MA Mansfield RJW Two cortical visual systems Analysis of visual behavior 1982 Cambridge MIT Press 549 586
Krause BJ Schmidt D Mottaghy FM Taylor J Halsband U Episodic retrieval activates the precuneus irrespective of the imagery content of word pair associates. A PET study Brain 1999 122 255 263 10071054
Zeki S Watson JD Lueck CJ Friston KJ Kennard C A direct demonstration of functional specialization in human visual cortex J Neurosci 1991 11 641 649 2002358
Rapoport SI Positron emission tomography in Alzheimer's disease in relation to disease pathogenesis: A critical review Cerebrovasc Brain Metab Rev 1991 3 297 335 1772739
Rondi-Reig L Burguiere E Is the cerebellum ready for navigation? Prog Brain Res 2004 148 199 212
Lang CE Bastian AJ Cerebellar damage impairs automaticity of a recently practiced movement J Neurophysiol 2002 87 1336 1347 11877508
Goldman-Rakic PS The prefrontal landscape: Implications of functional architecture for understanding human mentation and the central executive Philos Trans R Soc Lond B Biol Sci 1996 351 1445 1453 8941956
Corbetta M Miezin FM Shulman GL Petersen SE A PET study of visuospatial attention J Neurosci 1993 13 1202 1226 8441008
Paus T Location and function of the human frontal eye-field: A selective review Neuropsychologia 1996 34 475 483 8736560
Courtney SM Petit L Maisog JM Ungerleider LG Haxby JV An area specialized for spatial working memory in human frontal cortex Science 1998 279 1347 1351 9478894
Born AP Law I Lund TE Rostrup E Hanson LG Cortical deactivation induced by visual stimulation in human slow-wave sleep Neuroimage 2002 17 1325 1335 12414272
Czisch M Wehrle R Kaufmann C Wetter TC Holsboer F Functional MRI during sleep: BOLD signal decreases and their electrophysiological correlates Eur J Neurosci 2004 20 566 574 15233766
Peters A Kara DA Harriman KM The neuronal composition of area 17 of rat visual cortex. III. Numerical considerations J Comp Neurol 1985 238 263 274 4044915
Mapstone M Weintraub S Nowinski C Kaptanoglu G Gitelman DR Cerebral hemispheric specialization for spatial attention: Spatial distribution of search-related eye fixations in the absence of neglect Neuropsychologia 2003 41 1396 1409 12757911
Zatorre RJ Bouffard M Ahad P Belin P Where is ‘where' in the human auditory cortex? Nat Neurosci 2002 5 905 909 12195426
Rauschecker JP Tian B Hauser M Processing of complex sounds in the macaque nonprimary auditory cortex Science 1995 268 111 114 7701330
Zatorre RJ Bouffard M Belin P Sensitivity to auditory object features in human temporal neocortex J Neurosci 2004 24 3637 3642 15071112
Maeder PP Meuli RA Adriani M Bellmann A Fornari E Distinct pathways involved in sound recognition and localization: A human fMRI study Neuroimage 2001 14 802 816 11554799
Tian B Reser D Durham A Kustov A Rauschecker JP Functional specialization in rhesus monkey auditory cortex Science 2001 292 290 293 11303104
Henderson VW Mack W Williams BW Spatial disorientation in Alzheimer's disease Arch Neurol 1989 46 391 394 2705898
Rizzo M Anderson SW Dawson J Myers R Ball K Visual attention impairments in Alzheimer's disease Neurology 2000 54 1954 1959 10822436
Kavcic V Duffy CJ Attentional dynamics and visual perception: Mechanisms of spatial disorientation in Alzheimer's disease Brain 2003 126 1173 1181 12690056
Golob EJ Miranda GG Johnson JK Starr A Sensory cortical interactions in aging, mild cognitive impairment, and Alzheimer's disease Neurobiol Aging 2001 22 755 763 11705635
Filoteo JV Delis DC Massman PJ Demadura T Butters N Directed and divided attention in Alzheimer's disease: Impairment in shifting of attention to global and local stimuli J Clin Exp Neuropsychol 1992 14 871 883 1452635
Mohr E Cox C Williams J Chase TN Fedio P Impairment of central auditory function in Alzheimer's disease J Clin Exp Neuropsychol 1990 12 235 246 2341553
Parasuraman R Greenwood PM Haxby JV Grady CL Visuospatial attention in dementia of the Alzheimer type Brain 1992 115 711 733 1628198
Baddeley AD Baddeley HA Bucks RS Wilcock GK Attentional control in Alzheimer's disease Brain 2001 124 1492 1508 11459742
Kurylo DD Corkin S Allard T Zatorre RJ Growdon JH Auditory function in Alzheimer's disease Neurology 1993 43 1893 1899 8413944
Mapstone M Steffenella TM Duffy CJ A visuospatial variant of mild cognitive impairment: Getting lost between aging and AD Neurology 2003 60 802 808 12629237
Arnaiz E Almkvist O Neuropsychological features of mild cognitive impairment and preclinical Alzheimer's disease Acta Neurol Scand Suppl 2003 179 34 41 12603249
Wagner AD Early detection of Alzheimer's disease: An fMRI marker for people at risk? Nat Neurosci 2000 3 973 974 11017166
Luhmann HJ Heinemann U Hypoxia-induced functional alterations in adult rat neocortex J Neurophysiol 1992 67 798 811 1316953
Liepert J Bar KJ Meske U Weiller C Motor cortex disinhibition in Alzheimer's disease Clin Neurophysiol 2001 112 1436 1441 11459683
|
16159306
|
PMC1216331
|
CC BY
|
2021-01-05 10:39:14
|
no
|
PLoS Med. 2005 Oct 20; 2(10):e288
|
utf-8
|
PLoS Med
| 2,005 |
10.1371/journal.pmed.0020288
|
oa_comm
|
==== Front
PLoS MedPLoS MedpmedplosmedPLoS Medicine1549-12771549-1676Public Library of Science San Francisco, USA 10.1371/journal.pmed.0020356SynopsisNeuroscienceGeriatricsMedical ImagingMental HealthNeurology/NeurosurgeryDementiaMedical ImagingNeurologyGeriatric MedicineAlzheimer Disease: Failure to Tune Out Irrelevant Input? Synopsis10 2005 20 9 2005 2 10 e356Copyright: © 2005 Public Library of Science.2005This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are properly credited.
Impaired Cross-Modal Inhibition in Alzheimer Disease
==== Body
Some people close their eyes when they listen to music; others wear earplugs when they read in noisy buses or airplanes. Both serve to block out inappropriate sensory stimulation when we try to focus, something that happens automatically when we concentrate. Neuroimaging studies have shown that this focusing process is regulated within our brains; when people undergo tasks that depend exclusively on visual clues, they not only activate the areas of the brain involved in visual processing but also deactivate those areas that have to do with auditory processing. The reverse activation/deactivation pattern is seen when the clues are exclusively auditory. This phenomenon, called “cross-modal auditory/visual deactivation,” is thought to help focus the information processing capacities of the brain on the relevant areas
Alexander Drzezga and colleagues are trying to understand the cognitive changes associated with Alzheimer disease (AD). AD has historically been categorized as a disease of memory loss, but more recent results suggest that the disease is associated with more fundamental deficits in attention and perception. The researchers now report results from a study that examined the activation and deactivation patterns in the brains of patients with AD. The results suggest that patients with AD and patients with mild cognitive impairment (MCI) have problems focusing their brain activity: they show less activity than normal people in the “correct” brain areas but also more activity in the “wrong” areas.
To test whether cross-modal inhibition is affected in AD, the researchers recruited 32 participants, who fell into three age-matched groups: 11 healthy individuals, 11 individuals with MCI, and ten individuals with moderate AD. All participants were trained to perform a navigation task, based exclusively on visual clues, in a virtual reality environment. All participants understood the task and the clues. They then each completed 12 separate positron emission tomography scans, eight during active navigation (finding their way from a starting point to a specific destination) and four under control conditions (traveling along a never-ending pathway). The tasks were performed in complete silence.
The three groups differed in their performance—measured by the time needed to reach the destination—as would be expected. In addition, the researchers observed differences in their brain activation and deactivation patterns. Higher-order visual processing areas were activated to the greatest extent in healthy individuals, to a lesser extent in individuals with MCI, and not at all in individuals with AD. Similarly, cerebellar activation (suggesting movement automation) was absent in individuals with AD, present at low levels in individuals with MCI, and strongest in healthy individuals. Conversely, individuals with AD or MCI showed more activation in the (lower order) primary visual areas and in frontal cortical areas. This may indicate that sensory information gets “stuck” in lower levels of processing in AD. Regarding the inhibition of irrelevant input, strong bilateral deactivation in task-irrelevant auditory cortical regions was seen in healthy individuals. This was much less prominent in individuals with MCI, and individuals with AD showed no deactivation.
Cerebral activation and deactivation during spatial navigation
The researchers conclude that “successful cognitive performance in healthy individuals is associated with deactivation of task-irrelevant cerebral regions, whereas the development of AD appears to be characterized by a progressive impairment of cross-modal cerebral deactivation functions.” They go on to propose that “orientation disability in the outside world seen in patients with AD may in fact be partially based on the inability to selectively orient spatial attention to task-relevant internal representations of perceptual stimuli.” The researchers are now in the middle of a follow-up study that concentrates on individuals with MCI—some of whom remain stable for years, whereas others progress to AD—to see whether the extent of cross-modal inhibition correlates with progression to disease.
|
0
|
PMC1216332
|
CC BY
|
2021-01-05 10:39:16
|
no
|
PLoS Med. 2005 Oct 20; 2(10):e356
|
utf-8
|
PLoS Med
| 2,005 |
10.1371/journal.pmed.0020356
|
oa_comm
|
==== Front
PLoS MedPLoS MedpmedplosmedPLoS Medicine1549-12771549-1676Public Library of Science San Francisco, USA 10.1371/journal.pmed.0020360SynopsisBioengineeringAllergy/ImmunologyDermatologyOncologyGeneral MedicineVaccinesBiotechnologyCancer BiologyCell BiologyImmunologyInfectious DiseasesMolecular Biology/Structural BiologySystems BiologyVirologyDermatologyImmunology and AllergyOncologyMeasuring the Immune Response to Tumor Vaccination Synopsis10 2005 20 9 2005 2 10 e360Copyright: © 2005 Public Library of Science.2005This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are properly credited.
Marked Differences in Human Melanoma Antigen-Specific T Cell Responsiveness after Vaccination Using a Functional Microarray
==== Body
Tumor vaccination has perhaps been one of the most eagerly anticipated developments in cancer medicine. Research efforts in melanoma started with early unsuccessful attempts to induce a nonspecific immune response with Bacillé Calmette-Guerin (BCG). Subsequent vaccines combined BCG with autologous tumor cells or mixtures of allogeneic tumor cells, with varying responses. Most recently, however, work has focused on trying to produce a specific immune response using melanoma-derived peptide antigens. The antigens most commonly used are melanoma antigen recognized by T lymphocytes (MART-1)/Melan A, glycoprotein (gp) 100, and tyrosinase, all of which occur on both normal melanocyes and melanoma cells; randomized trials are currently under way on these vaccines.
However, the response of patients to these vaccines has been extremely variable and hard to predict. Ideally, researchers want to track antigen-specific T cells and measure their activation, but current assays are cumbersome and require large volumes of blood. Now, Mark Davis and colleagues from Stanford University and the University of Southern California present a high-throughput method for analyzing the response of patients to these vaccines, by means of an array that captures specific T cells, activates them, and then measures their response to activation.
Catching antigen-specific T cells on a biological chip
The researchers demonstrated the usefulness of this approach by studying ten patients from a phase II trial of 60 patients who had stage IIC/III and IV melanoma and who had been vaccinated with a combination of MART-1, gp100, and tyrosinase. They describe the use of peptide major histocompatibility complex arrays, which immobilizes CD8T cells, activates them, and then measures the degree of activation by the secretion of cytokines. The major new technical development is the method of measuring the amounts of cytokines released by means of labeled antibodies also present on the array.
What they found was a startling diversity of responses to vaccination. However, one predictor of good clinical response to vaccination was strong secretion of both interferon-γ and tumor necrosis factor-α; four of the four patients with this pattern of secretion remained free of melanoma recurrence, whereas only two of the six patients in whom there were marked differences in the secretion of these two cytokines were free of recurrence.
Where does this paper leave the field of research on tumor vaccination? It provides a detailed snapshot of T cell responses in individual patients, and if these patterns of responses are substantiated in larger numbers of patients, it may well allow doctors to begin to understand who is likely to respond to one of these vaccines. What it does not do is explain why vaccine responses vary so much among patients; this is a far more complex question.
|
0
|
PMC1216333
|
CC BY
|
2021-01-05 10:38:16
|
no
|
PLoS Med. 2005 Oct 20; 2(10):e360
|
utf-8
|
PLoS Med
| 2,005 |
10.1371/journal.pmed.0020360
|
oa_comm
|
==== Front
Ann Clin Microbiol AntimicrobAnnals of Clinical Microbiology and Antimicrobials1476-0711BioMed Central London 1476-0711-4-131615939910.1186/1476-0711-4-13ReviewNatural history and treatment of hepatitis B virus and hepatitis C virus coinfection Crockett Seth D [email protected] Emmet B [email protected] Department of Medicine, Stanford University School of Medicine, Stanford, California, USA2 Division of Gastroenterology and Hepatology, Stanford University School of Medicine, Stanford, California, USA2005 13 9 2005 4 13 13 30 8 2005 13 9 2005 Copyright © 2005 Crockett and Keeffe; licensee BioMed Central Ltd.2005Crockett and Keeffe; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Hepatitis B virus (HBV) and hepatitis C virus (HCV) coinfection is not uncommon as a result of similar routes of infection. Patients who are coinfected represent a unique group with diverse serologic profiles. Combined chronic hepatitis B and C leads to more severe liver disease and an increased risk of hepatocellular carcinoma. Furthermore, coinfected patients represent a treatment challenge. No standard recommendations exist for treatment of viral hepatitis due to dual HBV/HCV infection, and therefore treatment must be individualized based on patient variables such as serologic and virologic profiles, patient's prior exposure to antiviral treatment, and the presence of other parenterally transmitted viruses such as hepatitis D virus and human immunodeficiency virus. The natural history and treatment of patients with HBV and HCV coinfection is reviewed.
TreatmentHepatitis BHepatitis CHBV/HCVCoinfectionDual infectionInterferonRibavirinLamivudineTriple infectionHBV/HCV/HDVHBV/HCV/HIV
==== Body
Introduction
Hepatitis B and hepatitis C viruses are the most common causes of chronic liver disease worldwide. Acute infection with hepatitis B virus (HBV) or hepatitis C virus (HCV) may result in chronic infection, which occurs at a high rate in infants infected with HBV and the majority of individuals infected with HCV. Chronic HBV and/or HCV infection can progress to cirrhosis and be complicated by hepatocellular carcinoma (HCC). Coinfection with both viruses can occur because of shared routes of infection. Patients with dual HBV and HCV infection have more severe liver disease, and are at an increased risk for progression to HCC [1-4]. Coinfected patients represent a diverse group with various viral replication and immunity profiles. Because of their distinct clinical course and heterogeneity, identification of patients who are candidates for therapy and selection of the optimal antiviral therapy is a challenge for clinicians. Herein we review the natural history of HBV/HCV coinfection, current understanding of the interactions between these hepatotropic viruses, and the limited literature on treatment of coinfected patients.
Epidemiology
Approximately 350 million people are infected with HBV worldwide [5], and the World Health Organization estimates that approximately 170 million people are infected with HCV [6]. The exact number of patients infected with both HCV and HBV is unknown. One Eastern European study found a rate of dual infection in 0.68% of a randomly selected healthy population of over 2200 individuals [7]. In patients with chronic hepatitis B, estimates of the rates of HCV coinfection vary from 9% to 30%, depending on the geographic region [8]. One Italian study found that rates of dual infection increased with age, and was more common in patients over 50 years of age [9]. These numbers may underestimate the true number of patients with both viral infections because no large-scale studies have been performed, and there is a well-described phenomenon of "serologically silent" occult HBV infection (i.e. patients with negative hepatitis B surface antigen [HBsAg] but detectable serum HBV DNA) in patients with chronic hepatitis C [10].
Screening for Coinfection
Persons with a first episode of acute hepatitis should be screened for all viral causes including HBV and HCV. Some patients may be inoculated with both viruses simultaneously and will present with acute hepatitis due to both viruses. In addition, HBV superinfection in patients with chronic hepatitis C, and HCV superinfection in patients with chronic hepatitis B have both been reported [11-13]. Therefore, episodes of acute hepatitis in patients with known chronic HBV or HCV infection, especially those with ongoing risk behavior for infection with the alternative virus such as injection drug users, should raise suspicion and prompt screening for superinfection. In addition, as will be described below, silent or occult HBV infection in patients with chronic hepatitis C may alter patients' clinical course and response to therapy [14,15]. However, this phenomenon requires further study regarding its clinical significance before routine screening of HCV patients for HBV DNA can be recommended.
Interaction of Hepatitis Viruses
Several studies have shown that the HBV and HCV interact with each other and affect immune responses. HCV infection can suppress HBV replication, as demonstrated by studies showing that patients with chronic hepatitis B who are coinfected with HCV have lower HBV DNA levels, decreased activity of HBV DNA polymerase, and decreased expression of HBsAg and hepatitis B core antigen in the liver [16-18]. Furthermore, patients with chronic HBV infection who become superinfected with HCV can undergo seroconversion of hepatitis B e antigen (HBeAg) and HBsAg to respective antibodies [19-21]. Sheen et al. [22] conducted a longitudinal follow-up study of a large series of HBV infected patients and found that the annual incidence of HBsAg seroconversion was 2.08% in coinfected patients compared to 0.43% in patients with HBV monoinfection, and a subsequent study confirmed these results [23]. Several mechanisms of replicative interference of HBV by HCV have been proposed. Shih et al. implicated the hepatitis C core protein in suppression of HBV [24]. A subsequent study found that the hepatitis C core protein suppressed HBV enhancer activity, thereby affecting transcription [25]. This inhibitory effect appears to be more pronounced with HCV genotype 1 both in vitro and in vivo [25,26].
Several authors have reported that HBV can reciprocally inhibit HCV replication as well [27,28]. Specifically, HBV DNA replication has been shown to correlate with decreased HCV RNA levels in coinfected patients [29]. In one Italian study, coinfected patients had a rate of HCV RNA clearance of 71% compared to 14% with HCV monoinfection [26]. HBV replication in coinfected individuals may result in more liver inflammation, as demonstrated by studies in which HBV replication correlated with elevated ALT levels, while HCV replication did not [28,30,31]. Furthermore, coinfected patients have been demonstrated to have lower levels of both HBV DNA and HCV RNA than corresponding monoinfected controls, indicating that concurrent suppression of both viruses by the other virus can also occur. [32].
Overall, the available evidence demonstrates that both viruses can inhibit each other simultaneously; either virus can play a dominant role; both viruses have the ability to induce seroconversion of the other; the chronology of infection has a role in determining the dominant virus; and HBV and HCV can alternate their dominance [8]. However, the overall dominant effect appears to be HCV suppression of HBV [11,32,33].
Clinical Scenarios
Different scenarios of infection have been described with dual infection with HBV and HCV including acute dual viral hepatitis, occult HBV coinfection of chronic hepatitis C, and superinfection by either virus in patients with preexisting chronic hepatitis due to the alternative virus. In addition, coinfected patients are often found to have evidence of both HBV and HCV infection without a clear chronology of infection. In areas with high endemic rates of HBV infection due to vertical transmission, coinfection can generally be assumed to be due to HCV superinfection. In other geographic areas, the sequence of infections is less clear. Acute coinfection or superinfection with either virus can lead to fulminant hepatitis, chronic hepatitis, cirrhosis and HCC (Figure 1).
Figure 1 Infectious scenarios and treatment window for patients coinfected with hepatitis B and hepatitis C viruses. HBV = hepatitis B virus, HCV = hepatitis C virus.
Acute Hepatitis due to Dual Infection
Few studies have been reported on simultaneous acute hepatitis with HBV and HCV, but it appears that HBV and HCV interact in acute infections similar to their interactions in chronic infection. Simultaneous coinfection via an accidental needle stick that resulted in acute hepatitis was first described by Liaw in 1982 [34]. The diagnosis of HBV infection was delayed for 6 weeks, possibly related to HCV interference. Simultaneous coinfection resulting in acute post-transfusion hepatitis has also been described [35]. Mimms et al [36]. studied patients with concurrent acute infection with HBV and HCV compared to those with acute HBV alone, and reported decreased ALT levels, delayed appearance of HBsAg, and a shorter duration of hepatitis B surface antigenemia in coinfected patients compared to controls, suggesting HCV suppression of HBV activity.
Occult Hepatitis B Infection
Several reports have described the entity of occult or serologically silent HBV infection, particularly as it relates to treatment outcomes [10,14]. Occult HBV infection refers to patients who have low levels of circulating HBV DNA, but who lack both circulating antigens and their corresponding antibodies (i.e. HBsAg, HBeAg, hepatitis B surface antibody [anti-HBs] and hepatitis B e antibody [anti-HBe]). Such patients have been shown to have more severe liver disease [37]. One study found that 33% of patients with chronic hepatitis C and occult HBV infection had cirrhosis, compared to 19% of patients with chronic HCV infection without detectable HBV DNA [38]. Occult HBV may also be associated with more liver inflammation, greater histological activity of hepatitis, and higher ALT levels [14]. The effect of serologically silent infections on response to treatment will be discussed below.
Hepatitis C Superinfection
In areas of high prevalence of HBV infection such as Asian countries, the phenomenon of HCV superinfection is well described [11,13]. Superinfection with HCV can result in suppression of HBV replication, as well as termination of the HBsAg carriage [11]. There is one case report of acute HCV superinfection resulting in HBeAg seroconversion as well as clearance of HBsAg [20]. After superinfection and seroconversion of HBsAg, HCV infection may persist and result in continued chronic hepatitis [19]. Besides its viral interaction with HBV, HCV superinfection can result in more severe liver disease and an increased risk of fulminant hepatitis [39]. The mortality rate of HCV superinfection in chronic hepatitis B patients may be as high as 10% [11].
Hepatitis B Superinfection
Superinfection with HBV in patients with chronic hepatitis C is less common. Liaw et al. reported a series of 2 patients with chronic hepatitis C with documented HBV superinfection (positive IgM anti-HBc) [12]. One 82-year-old patient who was positive for both HBV DNA and HCV RNA on admission died of fulminant hepatic failure. The second patient was HBsAg positive, but HBV DNA negative. She subsequently recovered from acute hepatitis with normalization of ALT and seroconversion of HBsAg, as well as disappearance of anti-HCV antibody, indicating a possible suppressive role of HBV on HCV infection. It is difficult to draw conclusions from this limited report, but it is reasonable to assume that there may be an increased risk of fulminant hepatitis with HBV superinfection, and that those who recover are at risk for chronic coinfection. The clearance of anti-HCV in this case suggests that superinfection may allow HBV to play the dominant role in suppression of HCV.
Fulminant Hepatitis
Several studies have addressed the role of coinfection with HBV and HCV in fulminant hepatitis. Chu et al. [39] conducted a prospective study of patients admitted with acute hepatitis C in Taiwan. Eleven patients had fulminant hepatitis, and of these, 23% had underlying chronic HBV infection, compared to 2.9% for patients without fulminant hepatitis. This corresponded to an odds ratio (OR) for chronic HBV patients developing fulminant hepatitis with HCV superinfection of 10.2 (95% confidence interval 4.7–21.9, p < 0.01). A French study of 40 patients with fulminant and subfulminant hepatitis found that 5 of 40 patients (12.5%) had acute coinfection with HBV and HCV and 3 of 40 (7.5%) had superinfection of HCV [40]. Another Taiwanese study of 25 subfulminant and fulminant hepatitis cases found similar rates of coinfection (9.4%) and HCV superinfection of chronic hepatitis B (3.1%) [41]. These studies suggest an increased risk of fulminant hepatitis with HCV and HBV coinfection and superinfection.
Chronic Hepatitis
There are various immune profiles of dually infected patients with chronic hepatitis, and their immune profiles have a bearing on the choice of treatment. One possibility is dually active HBV and HCV, in which patients have detectable serum HBV DNA and HCV RNA. It stands to reason that these patients are at highest risk of progression to cirrhosis and decompensated liver disease, and therefore, should be considered for treatment. Another possibility is active HCV infection (positive HCV RNA) in the setting of an inactive HBsAg carrier. Such patients behave similar to patients with HCV monoinfection, and likely exhibit HCV viral suppression of HBV activity. Another possibility is active HBV infection in patients with inactive or prior HCV infection (HBV DNA positive/HBeAg positive/HCV RNA negative/anti-HCV positive). This immune profile is less common, and indicates HBV suppression of HCV. The various immune profiles have significance in regards to treatment options as will be discussed below (Table 3).
Table 3 Immune Profiles of Patients with Chronic Hepatitis due to Coinfection, and Suggested Treatments Based on Published Trials
Immune profile HBV DNA HBsAg HBeAg HCV RNA Anti-HCV Possible treatments [ref]
Dually active + +/- +/- + + IFN alone [54]
IFN + ribavirin [57-59]
IFN + lamivudine [61]
Active HCV in HBV carrier - + - + + IFN alone [23, 53, 55]
IFN + ribavirin [57-59]
Active HBV in chronic HCV + - - + + IFN alone [10]
IFN + lamivudine [61]
Silent HBV + - - +/- + IFN alone [10]
IFN + lamivudine [61]
Abbreviations: HCV = hepatitis C virus; HBV = hepatitis B virus; DNA = deoxyribonucleic acid; RNA = ribonucleic acid; HBsAg = Hepatitis B surface antigen; HBeAg = Hepatitis B e antigen; Anti-HCV = antibody to hepatitis C virus; IFN = interferon
Cirrhosis
Coinfected patients have higher rates of cirrhosis with decompensation. One cross-sectional study found higher rates of cirrhosis (44% vs. 21%) and decompensated liver disease (24% vs. 6%) in coinfected patients compared to patients with chronic HBV monoinfection [16]. Moreover, HBV replication in coinfected patients (detectable serum HBV DNA) has been correlated with higher rates of cirrhosis, Knodell score, piecemeal necrosis and fibrosis [29]. A Saudi Arabian study found that coinfected patients (compared to HCV monoinfected patients) had more decompensated liver disease with a higher proportion of cirrhosis (95% vs. 48.5%) and Child-Pugh class C (37% vs. 0%) [42].
Hepatocellular Carcinoma
Coinfection with HBV and HCV has been shown in many case-control studies to correlate with an increased risk of developing HCC [4,42]. One study found a rate of HCC in coinfected patients of 63% compared to 15% in HCV monoinfection [42]. Benvegnu et al. [1] conducted a prospective study of 290 cirrhotic patients and found that coinfection (detectable anti-HCV and HBsAg) was an independent predictor for development of HCC in a univariate and multivariate analysis (p < 0.02, p < 0.05 respectively). A subsequent longitudinal study confirmed these results, and reported a rate of incidence of HCC (per 100 person years) of 6.4 in dually infected patients, compared to 2.0 in HBV and 3.7 in HCV monoinfected patients, and a 45% cumulative risk of developing HCC at 10 years in coinfected patients, compared with 16% and 28% in HBV and HCV monoinfected controls [2]. Similarly, A South African study reported an 83-fold increased for developing HCC among coinfected patients compared to patients without hepatitis B or C [43]. Because of the likely increased risk of developing HCC, coinfected patients should receive regular 6-month and possible more frequent screening with ultrasound of the liver and serum alfa-fetoprotein levels.
Treatment
General Principles
Well established treatment guidelines exist for patients with chronic hepatitis B and chronic hepatitis C [44-49]. In regards to HBV infection, the Asian-Pacific Association for the Study of the Liver (APASL), the European Association for the Study of the Liver (EASL) and the American Association for the Study of Liver Diseases (AASLD) recommend treating patients who have moderate to severe chronic hepatitis as evidenced by >2-fold elevation of ALT levels or significant findings on liver biopsy associated with HBV DNA >105 copies/mL. Patients may have detectable HBeAg associated with wild type virus or HBeAg-negative from infection with the precore or core promoter mutant virus. Currently licensed drugs for chronic hepatitis B in the United States include interferon alfa-2b, lamivudine, adefovir, entecavir, and peginterferon alfa-2a. Patients with fulminant hepatitis and decompensated cirrhosis are not likely to respond to antiviral agents and are candidates for liver transplantation (Table 1). Patients with chronic hepatitis C with detectable serum HCV RNA are candidates for 24 to 48 weeks of antiviral therapy based on genotype. Currently standard treatment for hepatitis C is peginterferon alfa-2a or 2b plus ribavirin. As with chronic hepatitis B, patients with decompensated liver disease or fulminant hepatitis (rare with acute hepatitis C) are candidates for liver transplantation (Table 2).
Table 1 AASLD Recommendations for Treatment of Chronic Hepatitis B
HBeAg HBV DNA ALT Treatment Strategy
+ + ≤2 × ULN Low efficacy with current treatment.
Observe; consider treatment when ALT becomes elevated
+ + >2 × ULN IFN-α, LAM, or ADV may be used as initial therapy
End point of treatment = seroconversion from HbeAg to anti-Hbe Duration of therapy:
• IFN-α: 16 weeks
• LAM: minimum 1 year, continue for 3–6 months after HbeAg seroconversion
• ADV: minimum 1 year
IFN-α nonresponders/contraindications to IFN-α → LAM or ADV
LAM resistance → ADV
- + >2 × ULN IFN-α, LAM or ADV may be used as initial therapy, IFN-α or ADV is preferred
End point of treatment = sustained normalization of ALT and undetectable HBV DNA by PCR assay
Duration of therapy:
• IFN-α: 1 year
• LAM: > 1 year
• ADV: > 1 year
IFN-α nonresponders/contraindications to IFN-α → LAM or ADV
LAM resistance → ADV
- - ≤2 × ULN No treatment required
± + Cirrhosis Compensated: LAM or ADV
Decompensated: LAM (or ADV); Refer for liver transplant. IFN-α contraindicated.
± - Cirrhosis Compensated: Observe
Decompensated: Refer for liver transplant
*HBV DNA > 105. Abbreviations: HBeAg: hepatitis B e antigen; HBV: hepatitis B virus; ALT: alanine aminotransferase; ULN: upper limit of normal; IFN-α: interferon alfa; LAM: lamivudine; ADV: adefovir; PCR: polymerase chain reaction. Source: Lok AS, McMahon BJ. Chronic hepatitis B: update of recommendations. Hepatology 2004;39:857-61.
Table 2 AASLD Recommendations for Treatment of Chronic Hepatitis C
Therapy Widely Accepted Therapy Contraindicated Treatment Recommendations
• Detectable HCV RNA
• 18 years of age or older
• Elevated ALT
• Liver biopsy showing chronic hepatitis with significant fibrosis
• Compensated liver disease (total serum bilirubin <1.5 g/dL; INR <1.5; albumin >3.4 g/dL; platelet count >75,000 k/mm3; and no evidence of hepatic encephalopathy or ascites)
• Acceptable hematological and biochemical indices (hemoglobin >13 g/dL for men and >12 g/dL for women; neutrophil count >1.5 k/mm3; creatinine <1.5 mg/dL)
• Not treated previously for HCV infection
• History of depression but well controlled
• Patient willing to be treated and to conform to treatment requirements • Major, uncontrolled depression
• Renal, heart, or lung transplant recipient
• Autoimmune hepatitis or other condition known to be exacerbated by interferon and ribavirin
• Untreated hyperthyroidism
• Pregnant or unwilling/unable to comply with adequate contraception
• Severe concurrent disease such as severe hypertension, heart failure, significant coronary artery disease, poorly controlled diabetes, obstructive pulmonary disease
• Under 3 years of age
• Known hypersensitivity to drugs used to treat HCV Genotype 1 HCV infection:
• Peginterferon plus ribavirin (1000–1200 mg daily) for 48 weeks
• Treatment may be discontinued in patients who do not achieve an EVR at 12 weeks
• In patients who have negative HCV RNA at 48 weeks, retest HCV RNA at 72 weeks to confirm SVR
Genotype 2 or Genotype 3 infection:
• Peginterferon plus ribavirin (800 mg daily) for 24 weeks
• In patients who have negative HCV RNA at 24 weeks, retest HCV RNA at 48 weeks to confirm SVR
Abbreviations: INR: international normalized ratio; EVR: early virologic response; SVR: sustained virologic response. Source: Strader DB, Wright T, Thomas DL, Seeff LB. Diagnosis, management, and treatment of hepatitis C. Hepatology 2004;39:1147-71.
There is no currently established standard of care for patients who are coinfected with HBV and HCV. In general, the same treatment criteria should be applied to patients who are HBC/HCV dually infected as are applied to monoinfected patients (Tables 1 and 2). Initiation of treatment, as with both HBV and HCV, is recommended in patients with active chronic hepatitis or cirrhosis prior to decompensation (Figure 1). Given the complex interaction of HBV and HCV both with each other and with the immune system, care must be taken to select the most appropriate antiviral regimen based on serologic markers and levels of viremia. Because of its activity against both viruses, interferon therapy has been the most studied. The published data on the antiviral treatment of HBV and HCV coinfection is reviewed in the following discussion and summarized in Table 4.
Table 4 Medication Trials in Hepatitis B and Hepatitis C Coinfected Patients
Author [Ref] Patients # Treatment × Duration HCV SVR HBV DNA negative HBsAg loss HBeAg loss SBR
Interferon Trials
Gehenot [53] Anti-HCV+
HCV RNA+
HBsAg+
HBV DNA- 16 IFNα 3 MU TIW × 6 mo N/A N/A 12.5% N/A 19%
Weltman [3] Anti-HCV+
HBsAg+ 8 IFNα 3 MU TIW × 6 mo N/A N/A 12.5% N/A 12.5%
Guptan [54] Anti-HCV+
HCV RNA+
HBsAg+
HBV DNA+ 7 IFNα 6 MU TIW × 6 mo 29% 86% 28.6% 100% 0%
Villa [55] Anti-HCV+
HCV RNA+
HBsAg+ 30 IFNα 9 MU TIW × 6 mo or 6 MU TIW × 6 mo 16.7% (31%)* 66.7% (100%)* 3% N/A 20% (37.5%)*
Utili [23] Anti-HCV+
HBeAg ±
HCV RNA ± 16 IFNα 5 MU TIW × 12 mo 43.8% N/A N/A 15.4% 50%
Zignego [10] Anti-HCV+
HBV DNA+
HBsAg- 14 IFNα 3 MU TIW × 12 mo 0% 0% N/A N/A 0%
Liaw [56] Anti-HCV+
HBV DNA+
HBeAg+ 15 IFNα 9 MU TIW × 14 wk or 4–6 MU TIW × 12 wk 0% 6.7% 6.7% 6.7% 6.7%
Interferon plus ribavirin trials
Liu [57] Anti-HCV+
HCV RNA+
HBsAg+ 21 IFNα 6 MU TIW × 3 mo + 3 MU TIW × 3 mo + ribavirin × 6 mo 43% 35% 0% 100% 43%
Hung [58] Anti-HCV+
HCV RNA+
HBsAg+ 36 IFNα 3–5 MU TIW + ribavirin × 6 mo 69% 11% 0% 0% 56%
Chuang [59] Anti-HCV+
HCV RNA+
HBsAg+ 42 IFNα 6 MU TIW + ribavirin × 6 mo 69% 31.3% 14.3% 50% 54.8%
Interferon plus lamivudine trials
Marrone [61] HBeAg+
HBV DNA+
HCV RNA+ 8 IFN 5 MU TIW × 12 mo + lamivudine × 18 mo 50% 37.5% 0% 37.5% 50%
*9 MU arm. Abbreviations: # = number of patients; Ref = reference; SVR = sustained virologic response; SBR = sustained biochemical response; HCV = hepatitis C virus; HBV = hepatitis B virus; Anti-HCV = Antibody to hepatitis C virus; HBsAg = hepatitis B surface antigen; HBeAg = hepatitis B e antigen; mo = month; wk = week; TIW = thrice weekly; IFNα = interferon alfa
Interferon
Interferon is an immunomodulating medication with antiviral and antiproliferative effects and has been well studied in patients with chronic viral hepatitis. Interferon alpha (IFNα) is an approved treatment for both chronic hepatitis B and C. In appropriately selected patients with hepatitis C, interferon treatment led to sustained virological response (SVR), i.e. negative HCV RNA 6 months following completion of treatment, in approximately 10% of HCV patients. These results are improved when IFN is used with ribavirin (SVR up to 43%), and when peginterferon plus ribavirin is used (SVR up to 56%) [45]. In chronic HBV infection, IFN is indicated for patients with chronic hepatitis with elevated ALT and HBV DNA levels, and has the benefits of a lack of resistance and a durable response in those who respond to therapy. In studies of HBV patients, interferon is effective in roughly 35% of patients [50]. Peginterferon has also recently shown promise in treatment of hepatitis B [51], and peginterferon alfa-2a was recently licensed in the United States.
IFN has been the most studied agent in treatment of coinfected patients because of the wealth of experience with this agent in viral hepatitis and its proven activity against both viruses. One of the first case reports of successful treatment with IFN of a coinfected patient (resulting in HBeAg loss and a SVR of HCV infection) was published by Burt et al. in 1993 [52]. Several case series were subsequently reported. In 1995, Géhénot et al. [53] studied 16 patients with histologically proven chronic hepatitis with positive anti-HCV, HCV RNA, HBsAg, and anti-HBe (HBV DNA negative), compared to patients with chronic HCV infection alone. Treatment consisted of IFNα at 3 million units (MU) thrice weekly (TIW) for 6 months. A similar rate of sustained biochemical response (normal ALT at 6 months following treatment; SBR) was achieved in coinfected patients compared to controls (19 vs. 21%). Two patients treated with IFN seroconverted to HBsAg negative. This study demonstrated that comparable results could be obtained with interferon in coinfected patients who do not have evidence of HBV replication. A subsequent study by Weltman et al. [3] of 8 coinfected patients treated with the same regimen (IFN alfa-2b 3 MU TIW for 6 months) reported similar results, with 1 patient experiencing SBR (12.5%). The patient with SBR seroconverted at 6 months to HBsAg negative/anti-HBs positive. An Indian study of 7 dually infected patients (HBsAg, HBV DNA, anti-HCV, HCV RNA positive) used a higher dose of IFN alfa-2b (6 MU) for 6 months [54]. After 6 months follow-up, the authors reported 100% of patients lost HBV DNA, 100% of HBeAg-positive patients lost HBeAg (3/3), and 29% lost HCV RNA (i.e. SVR = 29%). Utili et al. [23] studied a cohort of 32 HBV/HCV coinfected patients, 16 of whom received IFN treatment (5 MU TIW for 12 months). They report an overall SVR rate of 43.8% for HCV infection, and this rate was increased in patients who were HBeAg negative (66.7%). Loss of HBeAg occurred in 2 of 13 patients (15.4%).
Villa et al. [55] conducted a larger prospective randomized trial of 30 patients with HBV/HCV coinfection (HBsAg-positive, Anti-HCV-positive, HCV RNA-positive), in which patients received either 6 or 9 MU IFN-α TIW for 6 months. This study found that higher dose IFN was more effective in inducing clearance of HCV RNA (31.2% vs. 0%, p = 0.045) and HBV DNA (100% vs. 0%), as well as inducing a SBR (37.5% vs. 0%, p = 0.019) compared to the lower dose. Histological scores were better in patients in the high-dose arm as well. This was the first study to suggest that higher doses of IFN may be indicated in coinfected patients.
Several studies have reported detectable HBV DNA in patients with chronic hepatitis C but negative HBsAg. This so-called "serologically silent" HBV infection or "inapparent coinfection" has been correlated with impaired response to IFN treatment. Zignego et al. [10] reported significantly worsened results in 14 chronically infected HCV patients with inapparent HBV coinfection (anti-HCV-positive, HBV DNA-positive, HBsAg-negative). Patients were treated with IFN alfa-2a TIW for 12 months. Four out of 14 patients had normal ALT levels at the end of therapy (28%), but all had relapsed within 6 month post-treatment, and thus none had a SVR. Fukuda et al. [14] also found that silent HBV infection was associated with higher ALT levels, greater histological activity scores and poor efficacy of IFN treatment. Some have proposed that the impaired response to IFN in such patients may be due to HBV-mediated down-regulation of intrahepatic IFN receptor gene expression [15].
A hepatitis C flare has been described in a coinfected patient who had HBeAg/HBV DNA clearance in response to IFN [56]. For this reason, some investigators have raised concern that IFN treatment of coinfected patients carries the risk of a severe hepatitis flare if the suppressive effect of one virus is removed, thereby allowing the other virus to become active.
Though there have been no published studies of the use of peginterferon (peg-IFN) in coinfected patients, the use of this agent will likely replace standard interferon in treatment of HBV/HCV coinfection, as peginterferon is now the standard of care for the treatment of chronic hepatitis C, and has recently been approved for the treatment of chronic hepatitis B.
Interferon plus Ribavirin
Several groups have published studies addressing treatment of coinfected patients with antiviral combination therapy with IFNα plus ribavirin. Liu et al. [57] treated 24 dually infected patients (HBsAg positive/anti-HCV positive) with IFN alfa-2a (6 MU TIW for 12 weeks followed by 3 MU TIW for 12 weeks), concurrently with ribavirin 1200 mg daily for 24 weeks. Seventeen patients were positive for both HBV DNA and HCV RNA. Results showed a SVR rate of 43% for clearance of HCV RNA, compared to 60% in similarly treated HCV monoinfected controls. Six of 17 patients with detectable HBV DNA at baseline had disappearance of HBV DNA at the end of treatment persisting to 24 weeks post-treatment (35%). A SBR rate of 43% was reported. These results demonstrate the effectiveness of combined IFN and ribavirin in coinfected patients, with a rate of SVR and SBR comparable to HCV monoinfected patients.
Hung et al. [58] treated 36 patients with HBV/HCV coinfection (HBsAg-positive/anti-HCV-positive/HCV RNA-positive) with combination IFN alfa-2b (3 or 5 MU TIW) and ribavirin (800–1200 mg/day) for 24 weeks. This study reported a SVR rate of 69% and SBR rate of 56%. Loss of HBV DNA was found in 2 of 18 patients with HBV DNA detected at baseline (11%). Interestingly, 53% of patients with negative HBV DNA at baseline had reactivation of HBV DNA at the 48 weeks of follow-up.
Chuang et al. [59] studied combination therapy with high-dose IFN alfa-2b (6 MU 3 times weekly for 24 weeks) and ribavirin (1000–1200 mg daily for 24 weeks) in 42 coinfected patients, compared to a control group of HCV monoinfected patients. The investigators found comparable rates of SVR in HCV coinfected (69.0%) and monoinfected (67.2%) patients (IFN naïve). Disappearance of HBV DNA occurred in 5 of 16 (31%) patients with positive HBV DNA at baseline. Viral interaction was evident, in that coinfected patients who achieved a SVR (compared HCV nonresponders) were less likely to achieve HBV DNA clearance (8.3% vs. 100%), and more likely to have reactivation of HBV (58.8% vs. 12.5%) or HBV flares (44.8% vs. 8.3%). Of the patients who cleared HBV DNA, 4 of 5 were patients who did not achieve a SVR with undetectable HCV RNA following combination therapy. Nonetheless, the high SVR rate for HCV infection in this population provides further evidence that IFN plus ribavirin combination therapy can be effective in coinfected patients, especially those who have active HCV replication.
As described in the above studies, successful treatment of chronic HCV infection may correlate to HBV reactivation and flaring. Yalcin et al. reported a severe hepatitis B flare in a patient with HBV/HCV coinfection (HBV DNA-negative) undergoing treatment with IFN and ribavirin [60]. This patient's hepatitis improved after discontinuation of therapy, but a relapse of HCV infection with rapid progression to cirrhosis occurred thereafter. Clinicians must exercise caution when treating coinfected patients with combination IFN plus ribavirin given this risk of HBV reactivation.
Interferon plus Lamivudine
One study of lamivudine therapy in addition to IFN for coinfected patients has been published by Marrone et al [61]. Eight patients with dually active HBV and HCV (HBeAg-positive/HBV DNA-positive/HCV RNA-positive) were treated with 5 MU of IFN and lamivudine 100 mg/day for 12 months followed by lamivudine alone for 6 months. Three patients had clearance of HBeAg, 3 had clearance of HBV DNA (37.5%), and post-treatment ALT levels normalized in 4 of 8 (50%) treated patients. Four patients (50%) also had clearance of HCV RNA that was persistent at 12 months post-treatment (i.e. HCV SVR = 50%). This initial study suggests that the addition of lamivudine to IFN may be effective in coinfected patients with chronic hepatitis C and active HBV replication. Further studies on this combination therapy in larger groups of patients must be performed in order to confirm these results.
Adefovir and Entecavir
There have been no published studies regarding treatment of coinfected patients with the newer agents adefovir and entecavir. However, these agents may be useful, particularly in patients with HBV-dominant disease. Studies need to be performed using these agents before they can be recommended for routine usage.
Transplantation
The United Network of Organ Sharing (UNOS) reported that 14 patients were transplanted for combined hepatitis B and C in the United States in 2004, and 434 patients have been transplanted for this indication since 1988 [62]. There are limited data on the post-transplant course of such patients. One small study by Huang et al. [63] reported results of 19 patients with coinfection (either acquired or persistent) following transplantation. This study found that survival of patients with HBV monoinfection was worse than that of dually infected patients, suggesting a possible beneficial role of HCV in the immunosuppressed post-transplant population. Further studies are needed to clarify the clinical course and optimal management of coinfected post-transplant patients.
Special Populations
Triple Infection with HBV, HCV and Hepatitis Delta Virus
Hepatitis delta virus (HDV) infects only patients with preexisting HBV infection. Triple hepatotropic viral infections with HBV, HCV and HDV can result in more severe hepatitis, and therefore compel the clinician to offer treatment [8]. Few studies have been published regarding treatment of patients with triple hepatitis virus infection. Weltman et al. [3] studied 7 patients with triple infection who received IFN therapy. One patient had a reported SBR, and 2 patients were withdrawn from treatment due to side effects. Interferon treatment is a reasonable recommendation despite the paucity of data to support its use. Further studies are needed on such patients to determine optimal therapy.
Triple Infection with HBV, HCV and Human Immunodeficiency Virus
Triple infection with HBV, HCV and the human immunodeficiency virus (HIV) is a complex clinical scenario, due to the interaction of HBV and HCV, and the impact of HIV on the immune system. In addition, a majority of patients with HIV/HCV coinfection are infected with HCV genotype 1, decreasing their response to interferon therapy and thereby rendering treatment more difficult [64]. Furthermore, HCV/HIV coinfection has been shown to result in more severe liver disease and an increased risk of liver disease-related death [65]. No standard of care exists for such patients, and treatment must be individualized and coordinated with an HIV specialist. Infection of HIV must be controlled before treatment of viral hepatitis can be considered [66]. Few studies have been performed on treatment of patients with triple infection, so treatment algorithms are often extrapolated from results of trials of patients with either HBV/HIV or HCV/HIV coinfection. IFN alone is associated with a SVR rate of approximately 17% in HCV/HIV coinfected patients, and this rate increases to 25% with the addition of ribavirin [67]. Lamivudine has been used in HBV/HIV coinfected patients, but is associated with a high rate of resistance [68]. There is promise for the use of newer agents such as adefovir, entecavir, tenofovir, and emtricitabine, but no studies have corroborated their utility in this population [66].
Summary of Treatment Recommendations
Thorough serologic and virologic testing is required in dually infected patients prior to consideration of therapy. Assessment of the "dominant" virus is helpful in determining a treatment strategy. Caution must be taken with treatment of coinfected individuals, as exacerbations of liver disease after initiation of therapy have been described, likely due to loss of viral suppression from the successfully treated dominant virus. Patients are candidates for therapy if they meet the inclusion criteria for standard treatment guidelines for either HBV or HCV monoinfection (Tables 1 and 2). In coinfected patients with HCV dominant disease, IFN plus ribavirin treatment has been well studied and has proven efficacy. In patients with HBV dominant disease, IFN with or without lamivudine is a reasonable option. Further studies of other HBV treatment agents such as adefovir and entecavir are needed before these agents can be routinely recommended, though they may be used on a case by case basis. In addition, future studies are needed to assess the effectiveness of peginterferon as well as triple therapy with lamivudine, IFN, and ribavirin in coinfected patients, though peginterferon should generally be used in place of standard interferon in coinfected patients given its proven efficacy in HBV and HCV monoinfected patients. Referral to a transplant center is indicated for patients with decompensated cirrhosis, fulminant hepatitis, or HCC in appropriate patients. In patients with triple HCV/HBV/HDV infections, few treatment studies have been published, but IFN is a reasonable treatment. Patients who have triple infection with HIV/HBV/HCV should have their care coordinated with an HIV specialist.
Conclusion
Coinfection with HBV and HCV is not uncommon, especially within areas of high prevalence of hepatitis B. Dual infections present unique management challenges given the complex interaction of HBV and HCV, and the propensity for developing more severe liver disease. Treatment options mostly include IFN with or without lamivudine or ribavirin. Treatment decisions should be made based upon the determination of the "dominant" hepatitis virus. Caution must be exercised in treating coinfected patients, as flares of the untreated virus may occur. No standard of care has been established for treatment of coinfected patients, and larger randomized, controlled trials are needed to clarify the optimal treatment for such patients and the role of newer antiviral agents.
==== Refs
Benvegnu L Fattovich G Noventa F Tremolada F Chemello L Cecchetto A Alberti A Concurrent hepatitis B and C virus infection and risk of hepatocellular carcinoma in cirrhosis. A prospective study Cancer 1994 74 2442 8 7922998
Chiaramonte M Stroffolini T Vian A Stazi MA Floreani A Lorenzoni U Lobello S Farinati F Naccarato R Rate of incidence of hepatocellular carcinoma in patients with compensated viral cirrhosis Cancer 1999 85 2132 7 10326690 10.1002/(SICI)1097-0142(19990515)85:10<2132::AID-CNCR6>3.0.CO;2-H
Weltman MD Brotodihardjo A Crewe EB Farrell GC Bilous M Grierson JM Liddle C Coinfection with hepatitis B and C or B, C and delta viruses results in severe chronic liver disease and responds poorly to interferon-alpha treatment J Viral Hepat 1995 2 39 45 7493293
Kaklamani E Trichopoulos D Tzonou A Zavitsanos X Koumantaki Y Hatzakis A Hsieh CC Hatziyannis S Hepatitis B and C viruses and their interaction in the origin of hepatocellular carcinoma JAMA 1991 265 1974 6 1848908 10.1001/jama.265.15.1974
Lee WM Hepatitis B virus infection N Engl J Med 1997 337 1733 45 9392700 10.1056/NEJM199712113372406
Hepatitis C–global prevalence (update) Wkly Epidemiol Rec 1999 74 425 7 10645164
Atanasova MV Haydouchka IA Zlatev SP Stoilova YD Iliev YT Mateva NG Prevalence of antibodies against hepatitis C virus and hepatitis B coinfection in healthy population in Bulgaria. A seroepidemiological study Minerva Gastroenterol Dietol 2004 50 89 96 15719010
Liaw YF Role of hepatitis C virus in dual and triple hepatitis virus infection Hepatology 1995 22 1101 8 7557857 10.1016/0270-9139(95)90615-0
Gaeta GB Stornaiuolo G Precone DF Lobello S Chiaramonte M Stroffolini T Colucci G Rizzetto M Epidemiological and clinical burden of chronic hepatitis B virus/hepatitis C virus infection. A multicenter Italian study J Hepatol 2003 39 1036 41 14642623 10.1016/S0168-8278(03)00470-7
Zignego AL Fontana R Puliti S Barbagli S Monti M Careccia G Giannelli F Giannini C Buzzelli G Brunetto MR Relevance of inapparent coinfection by hepatitis B virus in alpha interferon-treated patients with hepatitis C virus chronic hepatitis J Med Virol 1997 51 313 8 9093946 10.1002/(SICI)1096-9071(199704)51:4<313::AID-JMV9>3.0.CO;2-8
Liaw YF Hepatitis C virus superinfection in patients with chronic hepatitis B virus infection J Gastroenterol 2002 37 65 8 12109669
Liaw YF Yeh CT Tsai SL Impact of acute hepatitis B virus superinfection on chronic hepatitis C virus infection Am J Gastroenterol 2000 95 2978 80 11051381 10.1111/j.1572-0241.2000.02337.x
Liaw YF Chen YC Sheen IS Chien RN Yeh CT Chu CM Impact of acute hepatitis C virus superinfection in patients with chronic hepatitis B virus infection Gastroenterology 2004 126 1024 9 15057742 10.1053/j.gastro.2004.01.011
Fukuda R Ishimura N Niigaki M Hamamoto S Satoh S Tanaka S Kushiyama Y Uchida Y Ihihara S Akagi S Serologically silent hepatitis B virus coinfection in patients with hepatitis C virus-associated chronic liver disease: clinical and virological significance J Med Virol 1999 58 201 7 10447413 10.1002/(SICI)1096-9071(199907)58:3<201::AID-JMV3>3.0.CO;2-2
Fukuda R Ishimura N Hamamoto S Moritani M Uchida Y Ishihara S Akagi S Watanabe M Kinoshita Y Co-infection by serologically-silent hepatitis B virus may contribute to poor interferon response in patients with chronic hepatitis C by down-regulation of type-I interferon receptor gene expression in the liver J Med Virol 2001 63 220 7 11170061 10.1002/1096-9071(200103)63:3<220::AID-JMV1004>3.0.CO;2-3
Fong TL Di Bisceglie AM Waggoner JG Banks SM Hoofnagle JH The significance of antibody to hepatitis C virus in patients with chronic hepatitis B Hepatology 1991 14 64 7 1648540
Chu CM Yeh CT Liaw YF Low-level viremia and intracellular expression of hepatitis B surface antigen (HBsAg) in HBsAg carriers with concurrent hepatitis C virus infection J Clin Microbiol 1998 36 2084 6 9650968
Crespo J Lozano JL de la Cruz F Rodrigo L Rodriguez M San Miguel G Artinano E Pons-Romero F Prevalence and significance of hepatitis C viremia in chronic active hepatitis B Am J Gastroenterol 1994 89 1147 51 8053425
Liaw YF Tsai SL Chang JJ Sheen IS Chien RN Lin DY Chu CM Displacement of hepatitis B virus by hepatitis C virus as the cause of continuing chronic hepatitis Gastroenterology 1994 106 1048 53 8143971
Liaw YF Lin SM Sheen IS Chu CM Acute hepatitis C virus superinfection followed by spontaneous HBeAg seroconversion and HBsAg elimination Infection 1991 19 250 1 1917039 10.1007/BF01644957
Dai CY Yu ML Chuang WL Lin ZY Chen SC Hsieh MY Wang LY Tsai JF Chang WY Influence of hepatitis C virus on the profiles of patients with chronic hepatitis B virus infection J Gastroenterol Hepatol 2001 16 636 40 11422616 10.1046/j.1440-1746.2001.02494.x
Sheen IS Liaw YF Lin DY Chu CM Role of hepatitis C and delta viruses in the termination of chronic hepatitis B surface antigen carrier state: a multivariate analysis in a longitudinal follow-up study J Infect Dis 1994 170 358 61 7518488
Utili R Zampino R Bellopede P Marracino M Ragone E Adinolfi LE Ruggiero G Capasso M Indolfi P Casale F Dual or single hepatitis B and C virus infections in childhood cancer survivors: long-term follow-up and effect of interferon treatment Blood 1999 94 4046 52 10590048
Shih CM Lo SJ Miyamura T Chen SY Lee YH Suppression of hepatitis B virus expression and replication by hepatitis C virus core protein in HuH-7 cells J Virol 1993 67 5823 32 8396658
Schuttler CG Fiedler N Schmidt K Repp R Gerlich WH Schaefer S Suppression of hepatitis B virus enhancer 1 and 2 by hepatitis C virus core protein J Hepatol 2002 37 855 62 12445429 10.1016/S0168-8278(02)00296-9
Pontisso P Gerotto M Ruvoletto MG Fattovich G Chemello L Tisminetzky S Baralle F Alberti A Hepatitis C genotypes in patients with dual hepatitis B and C virus infection J Med Virol 1996 48 157 60 8835349 10.1002/(SICI)1096-9071(199602)48:2<157::AID-JMV7>3.0.CO;2-8
Sato S Fujiyama S Tanaka M Yamasaki K Kuramoto I Kawano S Sato T Mizuno K Nonaka S Coinfection of hepatitis C virus in patients with chronic hepatitis B infection J Hepatol 1994 21 159 66 7527435 10.1016/S0168-8278(05)80389-7
Pontisso P Ruvoletto MG Fattovich G Chemello L Gallorini A Ruol A Alberti A Clinical and virological profiles in patients with multiple hepatitis virus infections Gastroenterology 1993 105 1529 33 8224658
Zarski JP Bohn B Bastie A Pawlotsky JM Baud M Bost-Bezeaux F Tran van Nhieu J Seigneurin JM Buffet C Dhumeaux D Characteristics of patients with dual infection by hepatitis B and C viruses J Hepatol 1998 28 27 33 9537860 10.1016/S0168-8278(98)80198-0
Ohkawa K Hayashi N Yuki N Masuzawa M Kato M Yamamoto K Hosotsubo H Deguchi M Katayama K Kasahara A Long-term follow-up of hepatitis B virus and hepatitis C virus replicative levels in chronic hepatitis patients coinfected with both viruses J Med Virol 1995 46 258 64 7561800
Doi T Yamada G Endo H Nishimoto H Takahashi M Miyamoto R Fujiki S Shimomura H Mizuno M Tsuji T Hepatitis type C virus infection in patients with type B chronic liver disease Gastroenterol Jpn 1992 27 617 23 1330796
Jardi R Rodriguez F Buti M Costa X Cotrina M Galimany R Esteban R Guardia J Role of hepatitis B, C, and D viruses in dual and triple infection: influence of viral genotypes and hepatitis B precore and basal core promoter mutations on viral replicative interference Hepatology 2001 34 404 10 11481626 10.1053/jhep.2001.26511
Liaw YF Concurrent hepatitis B and C virus infection: Is hepatitis C virus stronger? J Gastroenterol Hepatol 2001 16 597 598 11422609 10.1046/j.1440-1746.2001.02523.x
Liaw YF Chu CM Chang-Chien CS Wui CS Simultaneous acute infections with hepatitis non-A, non-B, and B viruses Dig Dis Sci 1982 27 762 4 6807662 10.1007/BF01393772
Baginski I Chemin I Hantz O Pichoud C Jullien AM Chevre JC Li JS Vitvitski L Sninsky JJ Trepo C Transmission of serologically silent hepatitis B virus along with hepatitis C virus in two cases of posttransfusion hepatitis Transfusion 1992 32 215 20 1557801 10.1046/j.1537-2995.1992.32392213803.x
Mimms LT Mosley JW Hollinger FB Aach RD Stevens CE Cunningham M Vallari DV Barbosa LH Nemo GJ Effect of concurrent acute infection with hepatitis C virus on acute hepatitis B virus infection Bmj 1993 307 1095 7 8251805
Sagnelli E Coppola N Scolastico C Mogavero AR Filippini P Piccinino F HCV genotype and "silent" HBV coinfection: two main risk factors for a more severe liver disease J Med Virol 2001 64 350 5 11424125 10.1002/jmv.1057
Cacciola I Pollicino T Squadrito G Cerenzia G Orlando ME Raimondo G Occult hepatitis B virus infection in patients with chronic hepatitis C liver disease N Engl J Med 1999 341 22 6 10387938 10.1056/NEJM199907013410104
Chu CM Sheen IS Liaw YF The role of hepatitis C virus in fulminant viral hepatitis in an area with endemic hepatitis A and B Gastroenterology 1994 107 189 95 8020661
Feray C Gigou M Samuel D Reyes G Bernuau J Reynes M Bismuth H Brechot C Hepatitis C virus RNA and hepatitis B virus DNA in serum and liver of patients with fulminant hepatitis Gastroenterology 1993 104 549 55 8381098
Wu JC Chen CL Hou MC Chen TZ Lee SD Lo KJ Multiple viral infection as the most common cause of fulminant and subfulminant viral hepatitis in an area endemic for hepatitis B: application and limitations of the polymerase chain reaction Hepatology 1994 19 836 40 8138254 10.1016/0270-9139(94)90280-1
Mohamed Ael S al Karawi MA Mesa GA Dual infection with hepatitis C and B viruses: clinical and histological study in Saudi patients Hepatogastroenterology 1997 44 1404 6 9356863
Kew MC Yu MC Kedda MA Coppin A Sarkin A Hodkinson J The relative roles of hepatitis B and C viruses in the etiology of hepatocellular carcinoma in southern African blacks Gastroenterology 1997 112 184 7 8978357
Lok AS McMahon BJ Chronic hepatitis B: update of recommendations Hepatology 2004 39 857 61 14999707 10.1002/hep.20110
Strader DB Wright T Thomas DL Seeff LB Diagnosis, management, and treatment of hepatitis C Hepatology 2004 39 1147 71 15057920 10.1002/hep.20119
EASL International Consensus Conference on Hepatitis C. Paris, 26–28, February Consensus Statement. European Association for the Study of the Liver J Hepatol 1999 30 956 61 10365827 10.1016/S0168-8278(99)80154-8
de Franchis R Hadengue A Lau G Lavanchy D Lok A McIntyre N Mele A Paumgartner G Pietrangelo A Rodes J EASL International Consensus Conference on Hepatitis B. 13–14 September, 2002 Geneva, Switzerland. Consensus statement (long version) J Hepatol 2003 39 S3 25 14708673
Liaw YF Leung N Guan R Lau GK Merican I McCaughan G Gane E Kao JH Omata M Asian-Pacific consensus statement on the management of chronic hepatitis B: a 2005 update Liver Int 2005 25 472 89 15910483 10.1111/j.1478-3231.2005.01134.x
Keeffe EB Dieterich DT Han SH Jacobson IM Martin P Schiff ER Tobias H Wright TL A treatment algorithm for the management of chronic hepatitis B virus infection in the United States Clin Gastroenterol Hepatol 2004 2 87 106 15017613 10.1016/S1542-3565(03)00312-4
Wong DK Cheung AM O'Rourke K Naylor CD Detsky AS Heathcote J Effect of alpha-interferon treatment in patients with hepatitis B e antigen-positive chronic hepatitis B. A meta-analysis Ann Intern Med 1993 119 312 23 8328741
Chan HL Leung NW Hui AY Wong VW Liew CT Chim AM Chan FK Hung LC Lee YT Tam JS A randomized, controlled trial of combination therapy for chronic hepatitis B: comparing pegylated interferon-alpha2b and lamivudine with lamivudine alone Ann Intern Med 2005 142 240 50 15710957
Burt M Chapman BA Scrimshaw BJ Jennings LC George PM Concurrent hepatitis B and C infection treated successfully with alpha interferon Aust NZ J Med 1993 23 411 412
Gehenot M Marcellin P Colin JF Martinot M Benhamou JP Erlinger S Alpha Interferon Therapy in HBsAg Positive Patients with Chronic Hepatitis C Hepatology 1995 22 116A
Guptan RC Thakur V Raina V Sarin SK Alpha-interferon therapy in chronic hepatitis due to active dual infection with hepatitis B and C viruses J Gastroenterol Hepatol 1999 14 893 8 10535471 10.1046/j.1440-1746.1999.01952.x
Villa E Grottola A Buttafoco P Colantoni A Bagni A Ferretti I Cremonini C Bertani H Manenti F High doses of alpha-interferon are required in chronic hepatitis due to coinfection with hepatitis B virus and hepatitis C virus: long term results of a prospective randomized trial Am J Gastroenterol 2001 96 2973 7 11693335
Liaw YF Chien RN Lin SM Yeh CT Tsai SL Sheen IS Chu CM Response of patients with dual hepatitis B virus and C virus infection to interferon therapy J Interferon Cytokine Res 1997 17 449 52 9282824
Liu CJ Chen PJ Lai MY Kao JH Jeng YM Chen DS Ribavirin and interferon is effective for hepatitis C virus clearance in hepatitis B and C dually infected patients Hepatology 2003 37 568 76 12601355 10.1053/jhep.2003.50096
Hung CH Lee CM Lu SN Wang JH Tung HD Chen CH Changchien CS Combination therapy with interferon-alpha and ribavirin in patients with dual hepatitis B and hepatitis C virus infection J Gastroenterol Hepatol 2005 20 727 32 15853986 10.1111/j.1440-1746.2005.03791.x
Chuang WL Dai CY Chang WY Lee LP Lin ZY Chen SC Hsieh MY Wang LY Yu ML Viral interaction and responses in chronic hepatitis C and B coinfected patients with interferon-alpha plus ribavirin combination therapy Antivir Ther 2005 10 125 33 15751770
Yalcin K Degertekin H Yildiz F Kilinc N A severe hepatitis flare in an HBV-HCV coinfected patient during combination therapy with alpha-interferon and ribavirin J Gastroenterol 2003 38 796 800 14505137 10.1007/s00535-002-1149-5
Marrone A Zampino R D'Onofrio M Ricciotti R Ruggiero G Utili R Combined interferon plus lamivudine treatment in young patients with dual HBV (HBeAg positive) and HCV chronic infection J Hepatol 2004 41 1064 5 15582146 10.1016/j.jhep.2004.07.009
United Network of Organ Sharing Accessed August 5, 2005
Huang EJ Wright TL Lake JR Combs C Ferrell LD Hepatitis B and C coinfections and persistent hepatitis B infections: clinical outcome and liver pathology after transplantation Hepatology 1996 23 396 404 8617417
Rubio Caballero M Rubio Rivas C Nogues Biau A Manonelles Fernandez A [Epidemiology of chronic hepatitis C virus in patients infected by human immunodeficiency virus. Study of 767 patients] Med Clin (Barc) 2005 125 56 8 15970184 10.1157/13076466
Soriano V Martin-Carbonero L Garcia-Samaniego J Puoti M Mortality due to chronic viral liver disease among patients infected with human immunodeficiency virus Clin Infect Dis 2001 33 1793 1795 11641832 10.1086/323009
Sterling RK Triple infection with human immunodeficiency virus, hepatitis C virus, and hepatitis B virus: a clinical challenge Am J Gastroenterol 2003 98 2130 4 14572556 10.1111/j.1572-0241.2003.07720.x
Sterling RK Sulkowski MS Hepatitis C virus in the setting of HIV or hepatitis B virus coinfection Semin Liver Dis 2004 24 61 8 15346248 10.1055/s-2004-832930
Benhamou Y Bochet M Thibault V Long-term incidence of hepatitis B virus resistance to lamivudine in human immunodeficiency virus-infected patients Hepatology 1999 30 1302 1306 10534354 10.1002/hep.510300525
|
16159399
|
PMC1224851
|
CC BY
|
2021-01-04 16:38:20
|
no
|
Ann Clin Microbiol Antimicrob. 2005 Sep 13; 4:13
|
utf-8
|
Ann Clin Microbiol Antimicrob
| 2,005 |
10.1186/1476-0711-4-13
|
oa_comm
|
==== Front
Aust New Zealand Health PolicyAustralia and New Zealand Health Policy1743-8462BioMed Central London 1743-8462-2-201614454410.1186/1743-8462-2-20CommentaryRestructuring Primary Health Care Markets in New Zealand: from Welfare Benefits to Insurance Markets Howell Bronwyn [email protected] New Zealand Institute for the Study of Competition and Regulation; and Victoria Management School, Victoria University of Wellington, Rutherford House, 23 Lambton Quay, Wellington, New Zealand2005 6 9 2005 2 20 20 7 5 2005 6 9 2005 Copyright © 2005 Howell; licensee BioMed Central Ltd.2005Howell; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
New Zealand's Primary Health Care Strategy (NZPHCS) was introduced in 2002. Its features are substantial increases in government funding delivered as capitation payments, and newly-created service-purchasing agencies. The objectives are to reduce health disparities and to improve health outcomes.
Analysis
The NZPHCS changes New Zealand's publicly-funded primary health care payments from targeted welfare benefits to universal, risk-rated insurance premium subsidies. Patient contributions change from fee-for-service top-ups to insurance premium top-ups, and are collected by service providers who, depending upon their contracts with purchasers, may also be either insurance agents or risk-bearing insurance companies. The change invokes the tensions associated with allocating risk-bearing amongst providers, patients and insurance companies that accompany all insurance-based funding instruments. These include increases in existing incentives for over-consumption and new incentives for insurers to limit their exposure to variations in patient health states by engaging in active patient pool selection.
The New Zealand scheme is complex, but closely resembles United States insurance-based, risk-rated managed care schemes. The key difference is that unlike classic managed care models, where provider remuneration is determined by the insurer, the historic right for general practitioners to autonomously set patient charges alters the fiscal incentives normally available to managed care organisations. Consequently, the insurance role is being devolved to individual service providers with very small patient pools, who must recoup the premium top-ups from insured individuals. Premium top-ups are being collected only from those individuals consuming care, in proportion to the number of times care is sought. Co-payments thus constitute perfectly risk-rated premium levies set by inefficiently small insurers, raising questions about the efficiency and equity of a 'universal' insurance system pooling total population demands and costs. The efficacy of using financial incentives to constrain costs and encourage innovation when providers retain the right to arbitrarily recoup costs directly from patients, is also questioned.
Results
Initial evidence suggests that total costs are higher than initially expected, and prices to some patients have risen substantially under the NZPHCS. Limited competition and NZPHCS governance requirements mean current institutional arrangements are unlikely to facilitate efficiency improvements. System design changes therefore appear indicated.
==== Body
Background
New Zealand's state-funded primary health care system has undergone fundamental structural and financial change following the implementation of the Primary Health Care Strategy (NZPHCS), beginning in 2002 [1]. The principal stated objectives of the NZPHCS are to reduce health disparities, to improve health outcomes and to increase the share of government funding in primary care. In pursuit of these objectives, the NZPHCS utilises two financial instruments and a structural instrument.
The financial instruments are an increase in the quantum of taxpayer funds applied to primary health care (a 43% increase in the first three years [2], with $1.7 billion additional funds being applied over 6 years [3]), and universal capitation funding for all enrolled citizens, irrespective of individual consumption of primary health care services. Whilst all citizens are eligible to receive capitation funding, the rate paid varies depending upon patient characteristics such as age, gender, ethnicity and financial deprivation, and the characteristics of the entity to which the capitation subsidies are paid (see Table 1).
Table 1 PHO Types and Annual Capitation Subsidies, 2004–5 [67]
Capitaiton Subsidies
GMS/Nurse Services to Improve Access
PHO Type Interim Access All
HUHC HUHC Maori/Pacific Non Maori/Pacific
Age Group Gender CSC N Y N Y 1 thru 4 5 1 thru 4 5
00–04 F Y $308.12 $471.96 $315.73 $471.96 $63.15 $126.29 $0.00 $63.15
N $308.12 $471.96
M Y $327.88 $471.96 $332.42 $471.96 $66.48 $132.97 $0.00 $66.48
N $327.88 $471.96
05–14 F Y $79.33 $302.61 $99.94 $302.61 $19.99 $39.98 $0.00 $19.99
N $79.33 $302.61
M Y $75.18 $302.61 $93.54 $302.61 $18.71 $37.42 $0.00 $18.71
N $75.18 $302.61
15–24 F Y $78.90 $291.50 $92.22 $291.50 $18.44 $36.89 $0.00 $18.44
N $36.09 $291.50
M Y $42.38 $291.50 $50.75 $291.50 $10.15 $20.30 $0.00 $10.15
N $20.79 $291.50
25–44 F Y $72.61 $291.50 $81.04 $291.50 $16.21 $32.41 $0.00 $16.21
N $7.32 $291.50
M Y $43.16 $291.50 $52.38 $291.50 $10.48 $20.95 $0.00 $10.48
N $5.91 $291.50
45–64 F Y $88.74 $319.27 $110.99 $319.27 $22.20 $44.40 $0.00 $22.20
N $12.22 $319.27
M Y $67.96 $319.27 $82.90 $319.27 $16.58 $33.16 $0.00 $16.58
N $9.57 $319.27
65+ F Y $191.27 $342.40 $191.27 $342.40 $38.25 $76.51 $0.00 $38.25
N $191.27 $342.40
M Y $164.95 $342.40 $164.95 $342.40 $32.99 $65.98 $0.00 $32.99
N $164.95 $342.40
Per capita management fees are paid irrespective of Access or Interim status, and are based upon PHO size:
• $9.61 per individual up to 20,000 and $4.67 per individual thereafter, for PHOs with fewer than 75,000 registered individuals
• $6.41 for the first 20,000, $5.83 for individuals 20,001 to 75000, and $5.25 for all others, for PHOs with more than 75,000 registered individuals
Definitions
PHO Types:
• Interim PHOs: more than 50% of the registered population of Maori or Pacific Island Ethnicity, or living in areas determined to be in NZ Deprivation Index deciles 9 or 10.
• Access PHOs: the remainder
Individual Characteristics:
• HUHC: High User Health Card – individual with 12 or more GP consultations in 12 months
• CSC: Community Services Card – identifies low income or beneficiary status of registered individual – irrelevant for registered patients of Access practices
Capitation Subsidy Type:
• GMS/Nurse: subsidy for first contact services provided by General Practitioner or practice nurse – nominally based upon an effective consultation subsidy of $36.40 for children under 6 and $26 for all other population groups eligible for low or reduced cost access, thereby presuming 13 fully subsidised visits for a HUHC young child and 8.5 for others; 6.3 partially subsidised visits per annum for a 65+ man and 7.4 for a 65+ woman (3.3 and 3.8 fully subsidised visits respectively assuming a $50 cost per visit).
• Services to Improve Access: capitation to develop access initiatives for high-needs populations (paid in addition to GMS/Nurse capitation
The structural instrument is community-based nonprofit Primary Health Organisations (PHOs), created to register eligible individuals, receive associated capitation monies, co-ordinate delivery of care to registered individuals and manage the associated service delivery contracts. PHOs are also charged with responding to the health needs and preferences of their communities and developing innovative ways of providing services that people can afford [4]. There are no restrictions placed upon the nature (either for-profit or nonprofit) of the entities with whom the PHOs can enter into contracts, or the nature of the financial or operational risk-sharing that these contracts may entail. The only requirement is that PHOs be nonprofit entities openly accountable to the public for their decision-making, with contracted service providers and communities being represented in their governance and decision-making [5]. As District Health Boards (DHBs) oversee the contracts with PHOs, the NZPHCS requires that PHOs are formed within DHB boundaries.
The relationships between the government, its primary policy-making agency the Ministry of Health, the 21 geographically-determined government service purchasing and delivery agencies the DHBs, the PHOs, service providers and patients under the NZPHCS is illustrated in Figure 1.
Figure 1 The New Zealand Primary Health Care Sector.
This paper examines the NZPHCS by analysing the changes in the contractual relationships and institutions relative to the pre-NZPHCS contracts and institutions. Specifically, it examines the intertwined implications of changing the principal government-funding instrument from a fee-for-service payment to a capitation payment, and the substitution of PHOs for service providers as the 'other party' with taxpayers, patients and the government in the social contract for overseeing the purchase and delivery of subsidised health care.
Analysis of Contractual Changes
The change in the subsidy from a targeted welfare benefit upon the consumption of services to a universal subsidy independent of consumption alters the nature of the residual risk-bearing for variations in individual patient consumption of health services (that is, variations in individuals' underlying health states). Pre-NZPHCS, the government underwrote the risks of demand variation for a small number of subsidised individuals whilst the majority of individuals self-insured. Risk-sharing was required only for the subsidised proportion of the population. Under the NZPHCS, as over ninety percent of the population receives subsidies, the risks and costs associated with variations in the health states of the nearly the entire population are shared. In the first instance, PHOs are required to underwrite these risks, but can subcontract these risks onto service providers via service provision contracts. Via their ability to levy patient charges in addition to revenues received from PHOs, service providers may subcontract these risks further onto those who consume primary health care. Whilst the motivation for changes in locus of risk-bearing is to alter the behaviour of sector participants (principally service providers), the contracts that are emerging under the NZPHCS reveal that a reallocation of some of these risks is occurring in ways that are likely to lead to higher costs in total, an inequitable allocation of these additional costs amongst patients, and less efficient outcomes relative to the pre-NZPHCS contracts. The evolution of the entities engaging in contracting and the less-efficient contracts that they are entering into, appears to result from the acceptance under the NZPHCS that the pre-NZPHCS contractual arrangement allowing service providers to levy patient charges directly will continue under the new system. Whilst the strategy anticipates that structures and contracts will evolve over time in response to competitive pressures [6], as long as the practitioner right to levy patient charges independent of contractual constraints applied by the PHOs prevails, service providers continue to predominate in the governance of PHOs, and competitive pressures upon PHO-service provider co-operatives remains weak, it is likely that the current higher-cost arrangements will become entrenched.
Pre-NZPHCS Arrangements
The NZPHCS arrangements mark a fundamental change to the public-private partnership between government and privately-owned service providers (principally general practitioners) that characterised government subsidy of primary health care in New Zealand between the late 1930s and 2002. The public-private partnership arose out of a compromise between medical practitioners and the government in order to allow the Social Security Act 1938 and the subsequent Social Security Amendment Act 1941 to be passed [7]. Whilst the government at the time wished to offer 'free to the patient' taxation-funded health care, as in England's NHS, there was widespread opposition from members of the medical profession, who wished to retain their autonomy as private sector owners of the businesses delivering medical treatment. The outcome of the compromise was a bifurcation of the health sector into fully government-owned and funded public hospital, maternity and mental health services and privately-owned primary and specialist services. The government agreed to partially fund services provided by the private sector, whilst private practitioners retained the right to levy additional charges to patients [8].
Under the arrangements brokered in the 1930s, the 'social contract' resulted in private sector general medical practitioners being paid a fixed fee from taxation revenue for each treatment provided to eligible patients (termed a 'Section 88 Payment for General Medical Services', after the relevant section of the Act in which it was instituted). The payments were 'fee-for-service' in that they were paid for each consultation rendered to eligible patients by eligible practitioners. Whilst for administrative simplicity the payments were made direct to the servicing practitioner, in effect they constituted a 'welfare benefit' granted to an eligible patient, in that they were a part-payment of the debt incurred by the patient to the practitioner when medical treatment was sought by, and delivered to, the patient. The level of the subsidy was determined by individual patient characteristics and paid upon consumption of services. General practitioners were free to levy patients for the difference between the 'Section 88' payment (subsidy) and their actual costs of service delivery – the balance of the patient's debt, which was termed the 'patient co-payment'. As all registered general practitioners were eligible to receive 'Section 88' payments, there were no constraints placed upon patient choice of practitioner. These contractual arrangements are illustrated in Figure 2. Initially, the primary health care 'welfare benefit' was universal, in that all citizens were eligible to receive financial assistance, although the amount paid varied with patient characteristics (principally age). In the early 1990s, the subsidies became more tightly targeted, with family income, patient age and patient health state (principally the number of visits made in a twelve-month period) being the primary determinants of both eligibility for, and the size of, the subsidy.
Figure 2 Pre NZPHCS Primary Health Care Contracts.
Over time, the 'welfare benefit' share of the cost of general practitioner consultations fell substantially. When the NZPHCS was implemented, the government's contribution amounted to only 30% of general practitioner revenues [9]. The remaining 70% came from out-of-pocket expenses, or patient-funded private insurance contributions. Historically, private medical insurance has provided only a very small share of New Zealand primary health care costs, with fewer than 35% of the population having private medical insurance [10], and only 15% of total private medical insurance expenditure being applied to primary health care costs in the late 1990s [11]. As the range of available primary health care treatments increased, dissatisfaction grew with the restriction of subsidy payments solely to general practitioners and the 'medical intervention model' that it incentivised, potentially at the expense of models and practitioner types (e.g. nurses, dieticians, physiotherapists, social workers, educators) promoting the pursuit of wellness in the population.
During the 1990s, additional government funding was applied to a variety of contracts with a wider range of multi-disciplinary primary health providers (often, but not exclusively, nonprofit charitable trusts), typically serving high-need individuals or communities (often communities of specific geography, ethnicity, disease state or high health need). These contracts were typically either capitation-based (a fixed annual fee per individual eligible to receive the services offered) or price-and-volume-based (a fixed fee for a specific number of consultations). These relationships are also illustrated in Figure 2. In addition, a number of general practitioners moved voluntarily onto capitation-based payments as allowed under the Social Security Amendment Act 1941. Capitated practitioners retained the right to levy patients for the shortfall between capitation subsidies and actual costs. It is estimated that as many as 22% of general practitioners were receiving capitation funding in 2001 [12], although the number of patients involved is unknown. By 2001, government funded around 40% of the costs of all primary health care via both 'Section 88' and other contractual arrangements [13].
NZPHCS Arrangements
The NZPHCS fundamentally rewrites the 'social contract' between taxpayers, the government and general practitioners, with consequent changes in responsibilities and cash flows in the sector. The private-sector partners who receive government funding for primary health care under the new 'social contract' are no longer general practitioners and other service providers, but are the newly-created nonprofit Primary Health Organisations (PHOs). PHOs are charged with recruiting and registering patients for whom they will be responsible for purchasing and co-ordinating a range of primary health care services that will deliver upon the NZPHCS objectives of improved health outcomes and reduced disparities. PHOs are free to enter into contracts with any service providers (including, but not restricted to, general practitioners and providers receiving government contracts under the 1990s arrangements). Service providers who previously participated in 'Section 88' or other government-funded primary health care contracts must now enter into individual agreements with PHOs if they wish to receive income originating from government sources. Service providers are now no longer directly parties to the NZPHCS 'social contract': they participate in it only as subcontractors to a PHO. The terms and conditions of PHO-provider contracts appear to be freely negotiable between the parties concerned. In principle, this freedom of contract implies that PHOs need not bound by the requirements of the historic 'social contract' in their dealings with general practitioners or any other service providers.
Citizens who wish to receive government subsidies for primary health care under the NZPHCS 'social contract' are now required to have an explicit contractual relationship with a PHO. Unlike the pre-NZPHCS system, subsidies are universal, and are paid every quarter, irrespective of the actual quantity of health care a PHO-registered individual actually consumes in that period. Whilst for administrative convenience government payments are made direct to the PHO, it is the patient's contract with the PHO that determines government-sourced cash flows. In effect, the subsidy is a notional 'voucher' allocated to an individual and paid to the individual's chosen PHO upon production of evidence of the existence of a patient-PHO contractual relationship. A new contractual relationship, distinct from any other relationship that may exist between the individual as a patient and a service provider who may be acting as a PHO, is required because the basis of government funding has changed from the historic 'Section 88' instances of treatment (information historically supplied by service providers) to instances of PHO registration (a metric independent of any consumption of services and hence independent of service providers).
The patient's contractual relationship with the PHO grants the PHO both the right and the obligation to enter into contracts with service providers on behalf of the registered patient. The PHO-contracted service provider's obligation to provide subsidised care for a PHO-registered patient, either when the patient falls ill and requires treatment, or in respect of any well-patient services the PHO-provider contracts specify (e.g. preventative medicine, education), stems from the patient-PHO-service provider contractual nexus rather than from any other obligations or arrangements that may exist or may have existed historically between the patient and the service provider. Whilst capitated patients may nominate a preferred primary care provider, they are not restricted from seeking care from another provider. The NZPHCS allows for cash clawbacks by PHOs/practitioners delivering treatment for services provided to patients registered by other PHOs/practitioners. Clawbacks are made on a per-treatment basis using population-based averages of the number of visits made by patients of the relevant patient's class, and the quarterly capitation payment for that class of patient. Clawbacks can amount for up to 10% of quarterly general practitioner revenues [14].
The new contractual relationships outlined in the NZPHCS are illustrated in Figure 3.
Figure 3 NZPHCS Primary Health Care Contracts.
Transition Between Strategies
Whilst Figure 3 shows a direct relationship between the patient and the PHO, in practice there is rarely a direct, independent relationship. For administrative convenience, contracted service providers have been utilised as agents of the PHOs to facilitate the creation and maintenance of patient-PHO relationships: "existing lists of the patients who normally attend a practice or health clinic may form the starting point for enrolment" [15]. Indeed, a patient 'signals' a change in the PHO-patient relationship in most instances not on the basis of any direct transaction with the PHO, but by electing to change the preferred primary service provider. If the second provider has a contractual relationship with a different PHO from the first provider, the patient is deemed to have 'changed PHOs'. PHO revenue streams (and by extension, participation by their service provider-agents in government-subsidised service delivery contracts) are thus determined by the relationships between patients and the provider-agents. As PHO formation requires certainty of current and future revenue streams, and this certainty relies upon ongoing provider-patient relationships, most PHOs have been created based upon existing networks of service providers, such as general practitioner-governed Independent Practitioner Associations and community health organisations (including those based around communities of specific ethnicity) with strong provider networks.
The PHO creation arrangement also leads to PHOs binding service providers to exclusive contracts, denying them the right to enter into contracts simultaneously with any other PHO. Exclusive contracts are necessary as service providers with multiple PHO contracts cannot be relied upon to recruit and retain patients for a specific PHO. Consequently, PHOs compete with each other in order to sign up service providers rather than competing with each other to sign up patients. The PHO-service provider relationship 'crowds out' any incentives for PHOs to develop and manage direct relationships with patients independent of service providers, and results in effective competition in the sector being restricted to competing vertical alliances of PHOs and their exclusive providers [16]. The NZPHCS further requires that a PHO's contracted service providers (both those involved it its creation, and those who join subsequently) be active participants in the organisation's ongoing governance and management: "all providers and practitioners must be involved in the organisation's decision-making" [17]. Existing service providers are thus pivotal to the creation and operation of PHOs, with existing collectives advantaged by utilising existing relationships to 'become a PHO'. The operation of PHOs formed on this basis is likely to be strongly influenced by the providers around which the PHOs are formed.
The actual PHO registration, governance and management processes that have emerged raise the question of whether under the NZPHCS service providers are operating as contracted agents of PHOs, or whether PHOs organised around existing provider groups are in effect acting as service provider co-operatives (or even 'provider agents'). If the latter is an accurate representation of actual behaviour, then PHOs offer those providers (or groups of providers) who received subsidies pre-NZPHCS and who, through their collective bargaining ability, historically exhibited significant ability to influence the terms and conditions upon which government primary health care spending was applied, the ongoing ability to ensure access to, and to determine the contractual application of, government subsidies. If service providers control both PHO decision-making and PHO patient registration, it is unlikely that PHOs can freely enter into service provider contracts that are optimal for registered patients and contracting DHBs, and in the interests of the long-term financial viability and independence of the PHO itself, without jeopardising the relationships with providers upon which they are dependent for deriving their membership and hence their current and future revenue streams. If significant impediments to PHOs' independence and freedom to contract exist, then there may be substantial obstacles to the NZPHCS achieving its stated objectives efficiently using the PHO instrument as it is currently defined.
Registration Incentives
Outside of the NZPHCS, patients and service providers have complete freedom to enter into any contractual relationship of their choosing. The historic 'Section 88' payments will prevail at the 2001 levels. However, patients who opt for the pre-NZPHCS arrangements (and the PHOs who would otherwise register them) are denied access to the substantially more generous subsidies available under the NZPHCS. For example, a patient not qualifying for a 'Section 88' subsidy who consumes no services will generate revenue but no cost for a PHO under the NZPHCS. The same patient receives no funding at all if opting to remain under the 'Section 88' payment system, even if services are consumed, as there is no subsidy eligibility under this regime.
The availability of subsidies unrelated to actual consumption of health care is intended to provide strong incentives for PHOs to register individuals, consistent with the strategy's intention to increase the quantum of care provided to individuals who have not historically consumed health care and have poor health states. However, the PHO incentive to register is strongest in respect of those individuals who historically have not consumed large quantities of care due to their better-than-average underlying health states. These individuals bring revenues to the PHO based upon population averages, but the costs they incur are less, as they will likely consume fewer services. This exposes the NZPHCS to the risk of PHOs and their registering agents engaging in active patient pool selection ('cream-skimming') as it favours registering entities that can identify and register more profitable healthier-than-average individuals and exclude individuals who are less healthy than average and therefore less profitable. Yet it is the less-healthy-than-average individuals whose under-consumption of services under the pre-NZPHCS system was perceived to be contributing to substantial differences in patient health outcomes. Such pool selection incentives were not present under the 'Section 88' fee-for-service contracts, as service providers were fully compensated for all treatments provided, paid either by the individual or a combination of the individual and the state subsidy only in respect of treatments provided, irrespective of the patient's underlying health state.
The NZPHCS in Action
By December 2004, under the NZPHCS, capitation subsidies were being paid on behalf of 3.7 million New Zealanders (92.5% of the population) to 77 nonprofit PHOs [18], who enter into contracts with service providers for the delivery of primary health care to those individuals who fall ill. The subsidies vary depending upon the age, gender, income and historical health state of the registered individual, and upon PHO characteristics, determined by registered population size, and the ethnicity and deprivation levels of the registered population (see Table 1). Two types of PHO exist based upon these characteristics: Access and Interim. Higher-subsidised Access PHOs are required to have a registered patient base with at least 50% of individuals exhibiting specific ethnicity and financial deprivation characteristics [19].
General practitioners have taken a leading role in PHO formation. For example, thirteen of the South Island's seventeen PHOs are affiliated to Southlink IPA, and fourteenth, New Zealand's largest, is also based around an existing general practitioner-based primary health care collective. Consequently, over 92% of South Island PHO registrants belong to a general practitioner-formed PHO [20]. Practitioner influence over PHOs need not be confined to practitioner involvement in PHO formation. Wellington Independent Practice Association Limited (WIPA Ltd), a private company owned by fifty-nine general practitioners, has contracts to provide all management services to five PHOs across three DHBs in the lower North Island. The WIPA-managed PHOs have a combined market share of over 85%, and their contracts grant WIPA Ltd rights to act as if it were the PHO in respect of strategy, financial operation, service contracting and service development [21]. Nationwide, around 90% of the registered population is served by PHOs formed around general practitioner-dominated alliances.
The 77 PHOs in existence at December 2004 range in size between 3300 and 330,000 registered patients. They are also unevenly spread both geographically and by funding type. Higher-subsidised Access PHOs have on average 19,000 registered patients (median 11,200), whilst lower-subsidised Interim PHOs average 53,000 patients (median 31,200) [22]. South Island DHBs have on average 3.2 PHOs in their territories, whilst North Island DHBs have an average of 4.6 [23]. Only six percent South Island PHOs qualify for higher-subsidised Access funding, compared to sixty percent of North Island PHOs. Over 41% of individuals registered in the higher-subsidised Access PHOs in 2003 did not exhibit the population-based characteristics upon which the subsidy differentials are based [24]. If the population-based demographic characteristics upon which the subsidy differentials are based are good proxies for the actual costs of the differing underlying health needs of individuals, the high registration of non-Access individuals in Access PHOs suggests that active pool selection is occurring based upon subsidy differences.
Competition between the 77 PHOs is negligible, largely because of the geographic constraints that they be formed within DHB jurisdictions. Using as a benchmark for the presence of competition a three-firm concentration level of PHOs in each DHB area of less than 70%, and the largest PHO having a market share (measured as a PHO's share of the total PHO-registered population in a DHB area) of no more than 40%, there is little evidence of any meaningful competition occurring between PHOs [25]. On closer examination, most PHOs enjoy a local geographic monopoly. Even where patients have a notional choice of PHOs, effective choice does not appear to exist as PHOs are differentiated principally upon ethnicity and subsidy types, rendering the ability for the majority of patients to substitute practically non-existent. Higher-funded Access PHOs have no incentive to register individuals with non-Access characteristics if doing so would threaten the ability to claim higher subsidies for the entire registered base. Most Access PHOS are already close to the thresholds for non-Access registrants, making such substitutions unlikely to occur, even if there were no implicit ethnicity barriers discouraging substitution [26].
Discussion
The evidence to date suggests that despite the intention of the NZPHCS to change the identity of the 'other party' to the 'social contract' by establishing PHOs as new contracting entities, existing service providers, and in particular, general practitioners, have utilised the NZPHCS structures to replicate what appears to be similar relationships and responsibilities to those prevailing pre-NZPHCS. If general practitioners can dominate PHO decision-making, then under the guise of PHO governors they can continue to act as the de facto 'other party' to the social contract, as detailed in Figure 2. But whilst the NZPHCS structures may enable them to take these roles and in doing so preserve the outward form of previous relationships, the change in the funding instrument from targeted fee-for-service welfare benefits to universal insurance premium subsidies fundamentally alters the nature of the contracts which these relationships must fulfil. If the relationships prevail under the assumptions of the pre-NZPHCS funding contracts, but are required to serve the objectives intended under the NZPHCS arrangements, the potential exists for significant divergence between the intentions of the strategy and its outcomes. Intended outcomes may not be achievable, or if they are, it may be at substantially greater cost than if the NZPHCS relationships and contracts were built afresh. The balance of the paper discusses the key changes to contracts and relationships arising from the creation of a universal insurance instrument, and analyses the ways in which the relationships that have emerged under the NZPHCS institutions based upon pre-existing relationships are likely to affect the ability to deliver the intended objectives efficiently.
Subsidised Health Care: a Two-Sided Market
A fundamental distinction between markets for health care and markets for many other products is that patient-based demand for health care is a 'derived demand' arising from the occurrence of a stochastic event – illness. The unpredictability of falling ill, and the often-substantial costs associated with seeking treatment, creates demand uncertainty. In order to manage this uncertainty, individuals may prefer to 'pool' some of their risks and the associated costs of falling ill via private sector insurance schemes or centralised, taxpayer-funded institutions. These centralised entities then enter into contracts with service providers on behalf of the individuals to ensure both that treatments are provided when the patient falls ill, and that the service providers will receive payments for their services [27]. The central contracting entity, which may be an insurance company, a government agency or some other institution, becomes the 'hub' in a 'two-sided market' [28], entering into contracts on the one side with the individuals who seek certainty of access to and payment (or at least part-payment) for treatment, and on the other side with services providers who will deliver those treatments. The 'hub' entities maximise their profits (or optimise the health purchasing for their registered populations) by pooling the revenues received in respect of their registered populations and purchasing services only in respect of that subset of the pool that suffers an 'adverse event' (falls ill) and demands services. The two-sided market thus comprises a risk management (insurance) market on the one side, and a health service purchasing market on the other.
The Pre-NZPHCS 'Two-Sided Market'
Prior to the NZPHCS, government via the Ministry of Health operated the 'primary health care hub' in respect of the small proportion of individuals targeted for subsidised primary health care. Revenues in respect of targeted individuals were provided from taxation, comprising the 'risk management' component of the two-sided market, whilst the 'Section 88' agreement and payments, and agreements with other service providers, comprised the 'service purchasing' component. The government underwrote ('insured') the costs of any variations in the health states (and hence demands for health care) of the targeted individuals. The sicker they were, and the more treatments they sought, the higher the government's costs. Payments came from a single pool provided by all New Zealand taxpayers, and were made to providers when targeted individuals consumed care. The 'welfare benefit' payment was an output of the insurance 'hub' as it was a payment for services purchased in the service purchasing market.
As the majority of citizens were not eligible for targeted welfare benefits for primary health care, they had no need for the services of the taxpayer-funded 'primary health care insurance hub'. They were fully responsible for all of their costs of treatment. Costs from any variations in these individuals' health states were borne out-of-pocket, either directly in payments for services, or via private health insurance. These individuals were said to be 'self-insuring'. These patients paid the full costs of their treatment direct to service providers.
Under the fee-for-service payment arrangements pre-NZPHCS, service providers bore none of the insurance-related financial risks associated with variations in the health states of any of their patients, either self-insuring or targeted welfare beneficiaries. Each instance of treatment was fully paid for, either wholly by the patient, or by a combination of patient payment and taxpayer subsidy. Whilst some practitioners voluntarily waived charges for some payments, this is a normal commercial decision made by any business operator who chooses to offer selected discounts to specific clients, and does not impose any insurance-based risk management obligation in respect of that client, as is the case in the two-sided insurance market. Providers' engagement in the two-sided market was thus solely as suppliers of services.
As in the case of any subsidy where the patient does not pay the full cost of treatment out-of-pocket, subsidised patients will likely consume more care than is necessary (e.g. over-consumption by the 'worried well' and supplier-induced demand, where providers, knowing that the patient is not paying the full costs, recommends more expensive treatments than necessary or treats beyond the point where a 'cure' has been effected – known as moral hazard behaviour [29]). As only a small proportion of individuals received subsidies pre-NZPHCS (and most paid some form of co-payment out-of-pocket), moral hazard costs were likely to be relatively small. Any moral hazard costs that were incurred by subsidised patients were borne collectively by the entire pool of New Zealand taxpayers. Furthermore, as there was only one state-funded insurance 'hub' pooling risks for targeted individuals, and participation by eligible individuals in the pool could not be denied, the incentives that typically face competing insurers to manipulate the risk profile of the patient pool by 'cream-skimming' were absent.
Two-Sided Markets Under the NZPHCS
At the simplest level, the NZPHCS fundamentally changes both the size and the locus of residual risk bearing in respect of health state variations in the recipients of subsidised health care. Multiple PHOs, with the potential to compete against each other, rather than the single Ministry of Health, become the 'hubs' in the two-sided primary health care market, and all citizens who register with PHOs become parties to the insurance-based health-state risk pooling and service purchasing that the PHOs, rather than the Ministry of Health, are charged with undertaking. Government subsidies change from being an output of an insurance hub (i.e. service purchasing payments) to an input into one (i.e. insurance premiums). The structures and financing arrangements espoused in the NZPHCS model appear consistent with the trends emerging in many countries, where with the use of premium subsidies "government plays a major role in assuring that insurance coverage is universal and affordable, but with competition in the provision of insurance and of medical care, in order to stimulate efficiency and provider responsiveness to consumer preferences" [30]. Whilst questions have been raised about the extent of meaningful competition present in New Zealand between PHO-insurers, the fundamental shift in direction under the NZPHCS to a multi-insurer model nonetheless invokes all of the cost consequences that are associated with risk management under any insurance-based scheme.
Insurance Consequences: Universal Insurance and PHO-Insurers
Firstly, as all citizens who register with a PHO become eligible for subsidies, no registered citizen will face the full cost of health care treatments provided. Whilst increased subsidies and extensions from targeted to universal eligibility are intended to induce those who could not afford to pay for care previously to now seek it, these instruments will also lead to a rise in the levels of inefficient over-consumption relative to the pre-NZPHCS counterfactual as the entire population now faces the incentive to inefficiently over-consume. Such behaviour has been documented previously in New Zealand in 1996 in respect of targeted subsidies for under-six-year-olds becoming universal [31]. Furthermore, the new subsidies will 'crowd out' at least part of the private contributions that the newly-subsidised individuals would have previously paid. Total costs of the system will rise, in many cases for no additional health gain.
Secondly, although the quantum of government funding for primary health care has been increased, the change in the nature of funding away from fee-for-service to capitation shifts the costs of demand variation in subsidised patients' health states away from government into the private sector. The PHOs have become the insurance hubs. Capitation payments become insurance premium subsidies, and PHOs now bear the risks of patient demand variation. Once the capitation levels are set, the government faces a predetermined charge every quarter, independent of variations in the demand for health services based upon actual patient health states. The only variations government faces attend to the number of individuals for whom subsidies are paid and any movement of individuals between subsidy classes. In the first instance, responsibility for cost variations due to variations in the health states of both those individuals that the government used to insure, and a very much larger number of additional individuals who previously self-insured but are now party to risk-sharing agreements, has been placed upon the PHOs. However, where and how the costs of these variations are ultimately borne (and any surpluses that might accrue), and their sizes, will depend upon the decisions made by PHOs. Specifically, the ultimate size and allocation of costs and risks associated with patient demand variations will be determined by the contracts PHOs have with their service providers and their registered populations.
Relationships and Residual Risk-Bearing
The change in the funding instrument of the 'social contract' changes the relationships between all participants in the sector. PHOs become both health insurers and health service purchasers for their registered populations. Government is now simply a supplier of funding to the insurers and managers of service delivery contracts, rather than being the insurer and manager of those contracts itself. The subsidy is a partial contribution towards the insurance premiums of registered individuals. Government's sole role is to specify the size of the subsidies, and any contractual obligations associated with their application (e.g. minimum service ranges and qualities, reporting requirements). Patient out-of-pocket payments are now not a part- or full-payment of the costs of a service rendered by a service provider (a 'co-payment' in the classic fee-for-service subsidy environment), but a premium 'top-up' that brings the government premium subsidy up to the level of the actuarially-determined premiums that will be required to resource the PHO's contractual purchases of treatments. Even though for administrative convenience service providers collect patient payments, patient payment size will be determined ultimately by the costs of managing the insurance pool for, and of delivering treatments to, the PHO's entire registered patient pool. Patient payments must necessarily include factors that reflect the costs and risks of health state variations in the treatment-seeking behaviour of the PHO's entire registered patient base, and the risk management practices of the PHO and its subcontracted providers. These costs are incurred in addition to any shortfall between any government subsidy paid and the suppliers' actual costs of providing any specific contracted treatments.
The change in insurance responsibility means that any increase in moral hazard behaviour of patients alone makes it extremely unlikely that any increase in average subsidy under the NZPHCS will lead directly to a dollar-for-dollar decrease in the average patient 'co-payment' charged by providers, as it did under changes to the fee-for-service 'Section 88' subsidies. This occurs simply because the increases in cost from moral hazard actions previously borne collectively by the government and the taxpayer under Vote:Health are now reflected directly in the prices paid by patients, as it is the PHOs and their subcontractors who must now underwrite these risks. As long as the solitary instrument allowed under the NZPHCS for collecting any revenue shortfalls from demand variation is the patient payment to service providers, these payments must necessarily include a component to meet such costs. Thus it is not surprising that the Minister of Health has found that average patient co-payments in the "6–17 age group could not be seen to adequately reflect the increase in funding for that group in October 2003" [32] – indeed any other result would suggest an atypical consumer response to an increase in subsidy.
PHOs as Managed Care Organisations
As the NZPHCS institutions and relationships require PHOs to act as both insurers and service purchasers, it may be inferred that PHOs are acting as 'managed care organisations', balancing the costs of and demands for, primary health care within defined budgets (set by capitation subsidies or premiums) by matching the allocation of services purchased to the needs of the enrolled population. Managed care models have evolved around a set of fiscal and practice-based strategies, largely in response to the over-consumption of care, overly-high costs and weak incentives for providers to constrain costs that attend fee-for-service-subsidised health care systems [33,34]. Practice-based managed care strategies seek to "reduce variability in medical care by identifying 'best practices' and promoting adherence to guideline-based decision making. This includes evaluating the appropriateness of services rendered and the level of care necessary to provide the services" [35]. Practice-based managed care strategies appear appropriate for addressing the specific NZPHCS objectives of co-ordinating care across a range of providers, responding to community needs and continuously improving quality by better co-ordination, service innovation and use of information. Fiscal strategies in managed care typically involve some degree of financial risk-sharing between service purchasers and service providers, based upon the premise that it is more efficient for at least some of the risk associated with patient demand variations typically underwritten by insurers to be "borne by the individual physician for whom it is not a risk, but a controllable cost" [36]. Capitation payment for service delivery is one such fiscal strategy, along with preferred provider networks, price-and-volume contracts and other financial restrictions upon the freedom of practitioners to supply and commission treatments [37].
Managed Care Fiscal Strategies to Alter Provider Behaviour
Capitation payments for remuneration of service delivery are not axiomatic under the NZPHCS simply as a consequence of the insurance premium subsidy inputs paid to PHO-insurers being capitation payments. Any use of capitation payments by PHOs in respect of their service delivery contracts occur on the 'output' side of the 'insurance hub', and should be assessed in respect of their efficacy in assisting the managed care entity in achieving its balancing of demands and costs of care within its budgets. Each period, the insurer must pool the funds received in respect of all registered/insured individuals – the premiums – and from that income pay the costs of treatment for which the insurer is liable for that subset of the registered/insured population that requires services. The insurer thus balances the costs incurred by individuals against the revenues received in respect of the total registered base, and assumes responsibility for funding any shortfall (or keeps as a profit any surplus) that arises if the costs incurred are higher than the averages upon which the revenue is based because the registered base is less healthy than the average and consumes more services than average (that arises if the costs incurred are lower than revenue averages due to a more healthy than average registered base that consumes fewer services than the average). The insurer thus bears the financial variations associated with any variations in the patient health state from the averages upon which revenue is determined. Managed care entities are thus incentivised to look for ways to reduce costs in order to increase profits (or increase the quantum or quality of care provided within its fixed income budget). Even nonprofit managed care entities are incentivised to pursue higher profits, as the higher the profits (lower the costs) the more benefits the organisation is able to deliver to its beneficiaries, and the 'better' the nonprofit entity is deemed to perform [38].
Managed care entities can reduce the total cost of health care by using terms in the service contract to incentivise service providers to restrict their inefficient cost-causing behaviour. The contractual terms 'share' the managed care organisation's risks (and the associated costs) of variations in patient health states, and therefore demands for service, with the service providers. If providers can reduce costs (e.g. with preventative interventions and education), then financial risk-sharing will incentivise providers to alter their treatments in cost-conscious ways as they have to bear some of the costs associated with variations in patient health state, in the same manner as the insurance company. Financial risk-sharing contracts incentivise providers to reduce their costs by separating at least some of the determinants of providers' income streams (e.g. using capitation payments and discounted-price-and-volume payments) from the determinant of their costs (typically instances of care delivered to patients). With some of their revenues now 'fixed', providers can maintain their previous levels of profitability only by paying more attention to constraining their costs [39]. However, whilst provider costs can be constrained by reducing the extent of inefficient supplier-induced demand and other wasteful expenditure, and reducing the quality of service provided to an 'efficient' level, providers may also respond by reducing the quality provided below the 'efficient' level (e.g. overly-short consultations, queueing, rationing), further sharing the risks and costs with other parties if feasible (e.g. passing costs onto other contracting parties, shifting demand to other service providers), or actively engaging in 'selecting' a patient base that is lower-cost on average than the level at which the provider is remunerated, and thereby reducing the expected costs of the firm.
Managed Care Contracts and Incentive Intensity
The challenge for managed care organisations is to determine how much of the demand variation it is reasonable to share, given that not all cost-causing events are controllable by practitioners (e.g. a practitioner cannot influence the genetic predetermination or environmental circumstances that increase the health risk and hence demand of some individuals). If too much risk is shared, providers will pursue undesirable cost reduction activities at the expense of other objectives [40]; if too little is shared, then the costs of health care to the managed care insurer (and by extension the providers of its revenues) will be higher than necessary. Capitation funding provides the strongest cost-constraint incentives of all the fiscal strategies upon providers, with the strength of the incentive increasing with the proportion of revenue received from capitation [41]. Significant changes in practitioner behaviour have been observed in the United States under contracts with very low-powered financial incentives [42], suggesting that schemes with very high levels of capitation may over-incentivise provider cost-containment activities at the expense of other objectives, by placing more financial risk than is optimal upon providers [43].
Typically, managed care schemes also involve some compromises for patients. In order to constrain the prices paid by purchasers (via taxation, insurance premiums and out-of-pocket patient payments) in highly-subsidised systems and to ensure value-for-money spent, there must necessarily be some reductions in patient choice of practitioners, available treatments or treatment quality, and potentially even service rationing and queueing, relative to the counterfactual of an unrestrained fee-for-service payment regime [44]. Indeed, these are the very types of compromise that attend the provision of fully-capitated (via population-based budgets) state-funded public hospital services offered in New Zealand by the District Health Boards, which are in effect the 'care managers' in respect of these services for their designated geographic (equivalent to registered) populations, and the fully state-funded Primary Care Trusts in England's NHS [45]. Managed care schemes are therefore typically also associated with increased levels of overt monitoring relative to fee-for-service indemnity-type schemes in order to ensure that the service quality, range and availability do not fall below predetermined minimum acceptable levels [46]. Such monitoring adds substantial cost overheads to the managed care model, but as long as the cost reductions achieved exceed the additional costs within acceptable service qualities, managed care models can be more efficient than traditional indemnity-based fee-for-service insurance models [47].
Provider Charging and Managed Care Fiscal Strategies
In order to effectively design and manage contracts with service providers, so that the service providers are given sufficient incentives to constrain costs but still pursue other objectives, and are unable to shift onto other entities the risk that the managed care organisation deems providers should optimally bear, the managed care organisation typically uses contract terms to control all aspects of service provider remuneration in respect of treatments provided to registered patients. All service provider revenues relating to the contracts are typically provided from managed care funds (from premium/subsidy income or co-payments from insured individuals). Where co-payments are made direct to the service provider, these are generally in the form of a fixed deductible (excess) at a level determined by the managed care organisation rather than a fee set by the service provider to recover costs.
Indeed, any financial incentive effect upon a service provider in a contract with a managed care organisation will be 'undone' if the provider has the arbitrary ability to set patient co-payments to recoup any costs not provided by the managed care contract remuneration. A provider that has power to set co-payments has the power to shift onto patients that proportion of the risk that the managed care entity has deemed the provider should optimally bear. Furthermore, any financial incentives imposed by the managed care entity to alter provider behaviour (e.g. to induce increases in preventative interventions) are also reduced. Neither the cost savings nor service improvements sought from the financial risk-sharing contract are likely to be delivered to the extent anticipated by the managed care organisation when designing the contract. If any behaviour changes are educed in such a contractual arrangement, they will result from the practice-based strategies of the managed care organisation rather than the fiscal strategies.
If a provider with a financial incentive contract has the ability to set patient co-payments independently of the contract, then the provider can replicate the cash flows and absence of patient demand variation risk-bearing that attend a fee-for-service payment scheme. The provider fee-setting ability thus renders futile any attempt by the managed care organisation to influence provider cost-causing behaviour using fiscal strategies. The patient ends up paying the costs of the provider's share of the risks directly, in addition to the costs of the premium paid to the managed care entity. Any low-cost benefits that managed care models promise relative to fee-for-service insurance schemes are therefore also negated. The patient/consumer pays the extra costs in addition to facing the restrictions in choice, provider availability and service quality that attend the managed care model. Where the patient can balance the costs of premium, co-payment and service restrictions against the costs and benefits of alternative heath insurance models, if the managed care package is unfavourably costly, the patient will purchase an alternative. Overly costly managed care models where providers can arbitrarily levy patient charges will be unable to survive in the face of competition from other, less costly models such as provider-constrained managed care organisations and fee-for-service indemnity insurance arrangements.
If, however, the patient does not pay both the premium and the co-payment (e.g. the premium is paid from taxation revenue rather than directly from patient funds, so that the patient cannot 'internalise' the trade-off between the size of the premium and the size of the co-payment), any patient-based comparison with an alternative service is based solely upon the size of the co-payment that the patient faces. The higher the level of the premium subsidy, and the smaller the amount of the premium subsidy that can be transferred to an alternative insurer-provider model, the less favourable any competing alternative will appear to the patient [48], and the less likely it is that substitution to other models will occur. If the market for health care provision lacks the competitive discipline provided by alternative models, the overly-costly insurer-provider combination will persist, even though the alternative combinations would be less costly in total, resulting in persistently higher costs of primary health care than would be achievable in an environment where alternative models can freely compete.
PHO Fiscal Strategies and the 1938 Practitioner Charging 'Compromise'
The principal challenge facing New Zealand's managed care PHOs is that the very nature of the institutions and structures that have emerged under the NZPHCS pose some significant barriers to PHOs that restrict their ability to act as true managed care organisations. The stated intentions of the NZPHCS show contractual relationships between patients and PHOs (Figure 3), indicating that PHOs could charge registered individuals directly to recoup the difference between capitation subsidies and actuarially-calculated insurance premiums to fully-fund a managed care entity and utilise fiscal strategies to incentivise service providers to manage costs. In principle, such a funding arrangement would allow PHOs to develop a range of contractual relationships with a variety of providers that reflect both the extent of risk borne by each PHO given the health state of its patient base and the amount of risk desirable to be shared with specific providers in order to educe the desired behaviour changes, as in true managed care fiscal strategies, whilst simultaneously allowing patients to make the cost tradeoffs and select their desired PHO insurer on the basis of a bundle of cost and service characteristics. However, if the fundamental tenet of the 1938 compromise that allowed medical practitioners receiving income from government sources to levy additional charges on patients directly and independently of the PHO has been carried over into the NZPHCS, it renders impotent any attempt by PHOs to use fiscal strategies to constrain costs or incentivise desired provider behaviour.
Unconstrained Practitioner Charging Rights
Mandatory reporting by PHOs to their DHBs under the NZPHCS is based upon "the fees that their practices will be charging for standard consultations to the individuals in different groups" [49]. This suggests that practitioner charging of patients at the point of service consumption is the predominant, if not the only, way envisaged for PHOs and their subcontractors to recoup any additional costs of primary health care not covered by government subsidies. Even though alternative PHO charging models are theoretically feasible, 'grandfathering' of the historic right of medical practitioners to levy charges upon patients at service consumption has become the default method of premium top-up collection. Unless PHOs can contractually constrain the right of practitioners to charge patients, they will be able to operate as managed care entities only in respect of practice-based strategies.
There is little evidence of any constraints on general practitioners' rights to independently charge occurring in practice under the NZPHCS, at least in respect of the contracts between PHOs and the independent, for-profit private sector general practitioners that comprise the vast majority of the general practitioner workforce. The pro-forma 'back-to-back' contract offered on the Ministry of Health website as a model contract between general practitioners and PHOs (and is therefore likely to be similar to that between all private general practitioners and their PHOs) presumes that the PHO will simply 'pass on' capitation payments related to service provision directly to the service providers in respect of those patients that the service provider 'registers' with the PHO, suggesting that no such contractual constraint is being effected by the majority of PHOs. Whilst charged with managing risks as an insurance company on the one hand as a consequence of the shifting of responsibility for demand variation from the previous insurers (government and self-insuring individuals), PHOs are unable, due to the 'grandfathered' expectation of practitioners that they can still arbitrarily charge patients, from engaging effectively in using fiscal strategies to manage the financial risks that normally attend insurance-based managed care organisations.
Governance Arrangements and Contractual Content
Furthermore, as long as service providers dominate the governance of PHOs, it is extremely unlikely that PHOs will enter into contracts with their contractor-governors that share risks optimally by constraining the individual practice behaviour and profit-making potential of those selfsame contractor-governors. If the PHO cannot actively manage its financial risks using contracts, its optimal strategy is not to bear any financial risks at all. Thus, the default PHO contracting stance is likely to be one that simply shifts the capitation payments to service providers, who can then recoup any additional costs directly from patients.
As patients face barriers to exiting the subsidised system (they cannot take their state subsidies with them to a private insurer or any other practitioner operating outside the NZPHCS arrangements), and competition between PHOs is effectively non-existent, there are few competitive pressures acting to discourage PHOs from 'passing on' financial risks, or to constrain any additional costs that will arise from these actions. It is therefore extremely unlikely in the New Zealand context that capitation payment of service providers will be used in an actuarially-reasoned way to share a defined proportion of manageable risks between PHOs and service providers in order to exert controls upon service provider behaviour. Rather, the risk exists that as premium subsidies from government increase, and these are passed indiscriminately onto service providers, service providers may face overly-strong financial incentives. Whereas under constrained systems, these overly strong incentives may invoke quality reduction and under-servicing, in the NZPHCS where no such constraints apply, service providers will likely respond simply by passing on their increased share of financial risks via increased costs to patients. Such actions effectively 'undo' the benefits of patient risk pooling that in the first place provide the rationale for patients to join an insurance-based managed care scheme. All of the additional costs of an insurance scheme are incurred, but few of the cost-constraining tools of managed care are available. Indeed, such a scheme is likely more costly than a universal indemnity-based insurance scheme. Practitioners can recreate fee-for-service remuneration, so are left no worse off, but patients are worse off as they must endure restrictions in practitioner choice, service quality and service quantity, as well as paying the higher costs of treatment and institutional administration and monitoring than they would under the counterfactual.
Given the contractual limitations posed by the ability for general practitioners to charge patients directly, and strong general practitioner representation in the governance of PHOs, it is not surprising to find that most of New Zealand's seventy-seven PHOs as at December 2004 are undertaking no active insurance-based fiscal risk management activities. Financial risk management is not mentioned at all in the Ministry of Health-commissioned 2004 review [50] of PHO management services. An examination of the financial accounts of five PHOs affiliated to a general practitioner-owned management company in Wellington confirms that all capitation monies are being paid directly to either the practitioners (the component identified in Table 1 as GMS/Nurse Subsidies) or the management company (the payments identified in Table 1 as Services to Improve Access (SIA) Subsidies and per-capita management fees), with the PHOs undertaking no financial risk management activities [51]. Management fees and SIA payments are being used to fund the practice-based managed care strategies that PHOs are undertaking. PHOs are thus 'passing on' all the financial risks consequences of variations in patient demands under the NZPHCS arrangements are vested in the PHO-insurers to individual service providers by 'passing on' the capitation funding in its entirety.
Service Providers Become Insurance Companies
PHO 'passing on' contracts are in effect turning the service deliverers into the 'insurance hubs' of the NZPHCS. Even though it is the service providers who now receive the government-funded GMS/Nurse subsidies, the payments are still insurance premiums and must be managed as such. The subsidies cannot be considered in any way equivalent to the 'Section 88' payments. In any given period, service providers must now manage the demands of all insured individuals – that is, their entire registered patient base – and the costs that a subset of them will incur. This means that service providers should be undertaking the task of determining the actuarially-fair premiums to charge each registered individual and recouping the costs from them all. However, service providers are not trained or qualified in the operation of insurance companies, and they do not typically have mechanisms in place for collecting remuneration unrelated to service consumption. A service providers' interaction with registered individuals is typically confined to the episodic interaction that occurs when an individual consumes care. Given the pre-NZPHCS practice of charging co-payments only to those individuals who consume care, it is not surprising to find that, once again, for administrative convenience, the pre-NZPHCS business model of patient charging only upon consumption of services is being used to recoup premium top-ups in the NZPHCS.
Risk-Rated Insurance Premiums
Utilising pre-NZPHCS arrangements to collect premium top-ups invokes some significant distributional consequences. Whereas pre-NZPHCS, the demand variations and associated risk management costs for the small subset of the population who received subsidies were shared jointly by all taxpayers, and self-insuring individuals faced the costs of variation only in respect of their own individual demand, under the NZPHCS, as payments are collected in the form of patient charges at each consultation, the variations associated with the demands of all 3.7 million registered New Zealanders and the risk management costs associated with the very much larger insurance scheme are borne only by that subset of the registered population that consumes health care.
Furthermore, as the levy is collected upon each treatment consumed, the more treatments an individual consumes, the more of the additional costs intended to be borne by the provider and the PHO (and by extension the wider insured population), are paid by the individual patient seeking treatment. Rather than sharing risk management costs amongst the entire population as intended by an insurance scheme, the charging instruments of the NZPHCS, and specifically the 'grandfathered' 1938 charging agreement, ensure that these costs are paid only by the sick. Indeed, the result is that the frequently sick pay a perfectly risk-rated individual insurance premium contribution that is determined by their actual health state, as it is perfectly correlated with their instances of demand for care. Risk-rated insurance premiums vary the premium paid by an individual based upon the amount of risk an individual brings to the scheme. Sicker individuals bring more costs as they consume more care. In a community-rated system, all individuals pay the same premium, irrespective of their individual likelihood of consuming (based upon the 'community' or 'population-based' average likelihood of consuming). When an individual who is riskier pays a higher premium than a less risky individual, this is termed risk-rating, as the individual likely to cause more cost pays more of the costs of the pool. As a sicker individual consumes more treatments, more co-payments will be made. Premium top-ups are higher for sicker individuals, therefore the premium top-up under the NZPHCS is risk-rated.
Meanwhile, the infrequently sick enjoy lower charges for health care due to increased subsidies under the NZPHCS, so are more likely consume more care than previously, in the form of 'over-consumption by the worried well'. These increased costs are added to the charges levied by providers on consuming individuals, leading to even higher charges to the sick, which are borne disproportionately by the more frequently sick, who are less likely to inefficiently over-consume in response to increased subsidies as they are more likely to be genuinely ill.
Risk Redistribution and Equity
Whilst it is recognised that under the pre-NZPHCS, some individuals paid the full cost of their primary health care, and with constrained resources there will be a requirement for some patients to continue to make some contribution towards their health care costs, the equity of recouping the highest charges from the lowest-subsidised, sicker-than-average individuals, with the burden on these individuals rising substantially as more individuals of other subsidy classes receive higher subsidies over time, irrespective of any change to those individuals' health states, warrants consideration.
Pre-NZPHCS, the self-insuring majority faced no costs related to the consumption behaviour of services by other individuals except via taxation. Under the NZPHCS, they bear the financial consequences of changes to both the consumption behaviour of all other patients registered at their practitioner, and the financial consequences of the lack of behaviour-altering incentives resulting from capitation 'passing on' and patient charging, directly in their payments when consuming. Due to their financial status (i.e. aged 24–64 years and not living in areas identified as 9 and 10 in the New Zealand Deprivation Index), previously self-insuring individuals are also most likely to be paying the higher taxation required to fund higher premium subsidies. The burden of the additional risk management costs will be highest on the last group of previously self-insuring individuals to receive higher subsidies, and will be especially acute if the subsidy increases are accompanied by ill-informed regulatory restraint on the payments made by the newly-subsidised in order to keep constant the sum of their subsidy and their co-payment, without consideration of the additional costs associated with the subsidy increase. Under these circumstances, over time, previously self-insuring individuals can expect to pay substantially more to a subsidised provider than to a provider who eschews the subsidised system entirely.
To date there is very little evidence of providers eschewing the subsidised scheme. Therefore, there will be very little information available to either regulators or patients to determine what constitutes a 'fair' charge for services independent of the charges associated with financial risk-bearing. If provider collectives can utilise their professional membership status to limit the extent of competition from unsubsidised practitioners, then competition from unsubsidised practitioners is likely to be weak, and the likelihood of ill-informed regulatory restraint substantial. [Anti-competitive behaviour by medical practitioners has previously been found in respect of the members of the Ophthalmological Society of New Zealand found guilty under Section 27 of the Commerce Act 1986 for refusing to register foreign ophthalmologists to practice in New Zealand in order to carry out cataract surgery at lower remuneration from government contracts than the current incumbents collectively agreed to provide the surgery [52]. Moreover, in the absence of an unsubsidised sector with the ability to signal the costs of service provision independent from risk management, the government will have little information upon which to base the setting of its own subsidy contributions. This situation is very different from England, where the vibrant private sector stands as a benchmark against which government can assess both the size of subsidies offered to, and the quality and quantity of services provided by, NHS entities.
High Prices in Evidence
Evidence of higher patient co-payments under the NZPHCS, relative to its predecessor regime, is provided by the Consumers' Institute. Higher patient charges are reported in Interim PHOs than in Access PHOs, even for those individuals who generate identical subsidies under the two funding regimes [53]. This finding is consistent with the greater burden of additional risk management costs falling on practices with a larger number of lower-subsidised registered individuals. Rather than impose all the additional costs on the lower-subsidised individuals, these practices are opting to spread the higher costs amongst all patients. The Independent Practitioners Association chief executive has acknowledged that higher charges to patients are due to the increased financial risk that he claims practitioners are bearing. However, that patient charges have increased indicates that the practitioners are not absorbing the higher financial risks within their businesses, but are passing them onto patients.
That the burden of additional costs under the NZPHCS is real, significant and greatest upon frequently-ill low-subsidised individuals is in part reflected by the introduction, on 1 July 2004, of Care Plus. This additional capitated subsidy is designated to meet the additional costs incurred (initially to their practitioners, but ultimately passed on in higher patient charges) by frequently ill individuals who are not sufficiently sick to qualify for the higher subsidies available to high-use patients consuming twelve or more treatments in a twelve month period: "It's aimed at people who need to visit their family GP or nurse often because of significant chronic illnesses such as diabetes or heart disease, have acute medical or mental health needs, or a terminal illness" [54]. If the original capitation subsidies and population-based funding formulae had fairly allocated the costs of the system amongst individuals based upon patient need, and PHOs and practitioners had entered into contracts that minimised financial risk bearing costs and allocated the additional costs equitably across all patient classes, then arguably a Care Plus-type adjustment less than two years into the operation of the NZPHCS should not have been necessary. That the first substantive adjustment to the NZPHCS contracts addresses the costs of the category of patients that the foregoing analysis predicts will be most disadvantaged provides some strong circumstantial evidence supporting the analysis of this paper.
Patient Pool Size Cost Implications
The full extent of the inequitable allocation of additional costs under the NZPHCS model is not restricted solely to the costs of increased consumption associated with increased subsidies and the absence of effective cost-sharing to induce providers to constrain their cost-causing behaviour. The inability to effectively incentivise service providers that results in PHOs 'passing on' capitation contracts to service providers, results in the service providers becoming the effective insurance providers. Their insurance pools therefore have very small numbers, so will almost certainly lead to greater variation of profitability between providers than would occur with larger patient pools. Furthermore, patients with identical health states in different patient pools will incur substantially different premium payments, simply because of the wide distribution of patient health states between insurers with very small pools. Thus, the NZPHCS will be even more costly and even less equitable than both a standard managed care scheme and the pre-NZPHCS arrangements.
Large Numbers, Insurer Profitability and Pool Management
Insurance systems rely upon the 'law of large numbers' to reduce the variations in insurer profitability resulting from the unequal distribution throughout the population of the characteristics that cause the insurer to incur costs. In the case of primary health care, the cost-causing event is a patient developing a condition that causes the patient to seek primary health care treatment. Assuming the likelihood of an individual requiring treatment in any one period is random, any given pool of patients will have a pool 'average likelihood of requiring treatment' that is either higher or lower than the population average. The smaller the pool, the greater is the likelihood that the pool average will be substantially different from the population average. However, the larger the pool, the greater the likelihood that the pool average will be close to the population average. When insurer revenue is determined by population averages, but the costs are determined by pool averages, in any given period half the pools will have costs in excess of revenues (incur losses) and half will have costs less than revenues (incur profits). The smaller the pool, the greater the probability of making a profit or loss substantially different to that of a pool with the population average.
Typically, if demands between periods are random and unrelated, then over time each pool will incur a random number of profits and losses that cancel each other out. Where a pool makes a loss, it is covered either by profits retained from the past, or some other financing (e.g. owner underwriting or reinsurance). Managing cashflows between periods thus incur costs – known as risk management costs. Managers of insurance pools seek to maximise profits, which leads to incentives to minimise risk management costs. For random pools, merging pools to obtain a profit closer to the population average will reduce the costs associated with managing cash flows in the event of a loss. The optimum pool size may therefore be quite large. However, if the demands are correlated, either that in one period one pool has a known greater likelihood of incurring a profit or a loss, or that demands of a pool across time are linked to demands in past periods, then the allocation of costs will be correlated and the pool will be either habitually profitable or habitually loss-making. If an insurer knows the pool is habitually profitable, then profits can be routinely extracted. Such an insurer has no incentive to merge the pool with a pool of unknown costs, as profitability may be reduced. However, if the pool is habitually loss-making, then the insurer faces an incentive to merge with other pools in order to reduce the probability of making a loss. Merging pools in this manner leads to larger numbers and a reduction in the likelihood that any one pool is substantially different from the population average. However, if there is any possibility that an insurer with a known low-cost pool can resist merging with other pools, the 'average' profitability of the remaining pools will be less than the population average including the highly profitable pool [55].
The incidence of health costs is likely to be correlated, both within and between time periods. Specific individuals with similarly low costs and similarly high costs may patronise similar insurers (e.g. individuals within a community may all get sick simultaneously, or an entire family with genetically linked high health needs may seek cover from the same company), and there is substantial evidence that a very large proportion of health costs are incurred by a very small number of individuals, even in respect of primary care health care demands [56,57]. This suggests that the variation in health pools is significant. United States data suggests that even where there is information on an individual's past consumption of health care, only "an estimated 20% to 25% of total variation in health care expenditures on an individual basis is predictable, and the remainder is random" [58]. This suggests that the optimal size of a health insurance pool will be large in order to manage the risks of the insurer incurring losses, and to reduce the costs of risk management. The United States Health Care Financing Administration considers capitated primary health care physicians or physician groups to be at substantial financial risk if they have fewer than 25,000 patients, whilst "primary care physicians may find capitation disadvantageous even if they have only one or two patients who happen to require intensive medical care during a given year, or have a consistently sicker panel of patients relative to other primary care physicians" [59].
Pool Size of New Zealand Insurers
In New Zealand, private sector general practitioners are typically sole practitioners. Even though they may share some common overheads via a 'group practice', including clinic space and reception services, each practitioner usually maintains an individual, independent practice based upon a 'patient list' that contains typically between 1200 patients in a rural practice and 2000 patients in an urban practice. Whilst PHOs notionally provide insurance cover for patient bases of between 3000 and 300,000 individuals [60], the 'passing on' of capitation payments results in effective insurance pools being managed at the level of each individual general practitioner – that is, insurance pools of between 1200 and 2000 individuals. Given the United States evidence, these numbers are likely to be far too small to efficiently manage the demand variation that will occur even with a random distribution of health states amongst the population. Profits will be larger and losses larger than if pools were larger. The costs of risk management will therefore be greater than if the pools were larger.
Loss-making practitioner-insurers in New Zealand face few incentives to merge their pools to reduce risk management costs as would occur in a typical insurance market, as their first resort is to recoup losses simply by charging patients for the difference between costs and PHO subsidies. The system offers no other incentives to manage the pool efficiently, so the greater profit variations, and their associated costs to patients, will persist largely unchecked. As the pools are small and profitability variations great, there will therefore be substantial variations in patient co-payment prices amongst providers, simply because of the variations in the underlying patient health states of the patient list. That is, if the insurer-provider has a 'high cost' pool, patient co-payments will be higher, even for a low-demand individual, than those of a provider with a population average pool, simply to recover the additional costs of variation in the demand of the pool from the capitated population average. The patient's payments are therefore determined by the risk level of the provider's pool – a patient of a given health state will pay different prices at different practices simply because of the 'luck' that determines the practitioner's risk profile relative to the population average. In the presence of very small pools, variations between individual provider pools even within a single subsidy level may be substantially greater than the population-based variations upon which the differential capitation premium subsidies (Interim and Access) of the PHO are based, leaving individual providers subject to returning very large losses or very large profits. Profitable insurer-providers will be able to extract profits in excess of costs as dividends, and even charge co-payments similar to those of competing high-cost providers, not because of any effort on the insurer-provider's part, but simply because of the 'luck' in allocation of patients and health care demands.
Pool Management and Provider-Insurer Competition
Normally, a patient facing a high patient charge will seek to shift custom to a lower-charging provider. However, a profitable, 'low-cost' NZPHCS provider will face few incentives to register a patient who responds to the prices of the 'high-cost' pool by seeking to transfer to the 'low-cost' pool, simply because that patient's past patronage of a 'high-cost' pool suggests that the patient is more likely to have higher-than-average demand characteristics. A 'low-cost' provider would prefer to engage in screening behaviour to assess the health state of the patient to ascertain that the patient's health state is at or better than the provider's current average before making an offer to the patient to join ('cream-skimming'). High-cost pools, on the other hand, face opposite incentives. Any transferring patient may actually improve the pool average, so is less likely to be 'screened out'. Hence, where such pool management occurs, the system equilibrium tends towards a small number of very large high-cost pools and many small low-cost pools, as occurs in the United States managed care market with large state-funded Medicare and Medicaid pools, and many smaller privately-funded managed care entities. In the New Zealand environment, over time this could manifest as a large number of small private, for-profit pools, and a few very large pools more likely to be operated by private, nonprofit providers who are less motivated by the profitability of the pool than by the nonprofit's objective to serve individuals. The larger pools, however, will be the more costly ones. If premium subsidies are adjusted over time based upon the costs of the higher-cost pools, then the smaller lower-cost pools will become even more profitable.
Furthermore, efficient operation of the insurance market requires provider-insurers to know, in respect of the risk profile of their pool, whether a surplus in any one year is a profit, which can be extracted, or simply a surplus required to be held to underwrite losses in future years. Specialist actuarial knowledge is required to make such judgements: "since providers are exposed to exogenous risk, efficient risk pooling requires reinsurance for providers" [61]. Operating general practices under the pre-NZPHCS assumptions that any surplus in a given year is a profit that can be extracted under the NZPHCS conditions where general practices are insurers will inevitably result in the extraction of funds that should have been applied to risk management, leading to the likelihood of even greater losses in the future, simply because the surpluses that would ordinarily have been retained to manage the risks of future losses have been extracted as dividends by for-profit providers. Thus, contrary to expectations, the NZPHCS will not "guard against funds being diverted from health gain and health services to shareholder dividends" [62]. Rather, the structures and relationships that have emerged have actually made it more likely that such activities will occur, even if inadvertently, because general practitioners who do not have the skills or experience to act as insurers have, as a result of the changes in funding and evolution of contracts under the NZPHCS become charged with the insurance task.
Provider Information and Pool Management
Moreover, if there is any additional potential for providers to utilise information about patients to 'select' their patients to manage costs, the risk management overhead of the NZPHCS will be even greater than under the counterfactual of a standard, separate-insurer managed care model. Arguably, given that only around 25% of cost variation is predictable, such selection may be difficult to achieve. However, of the small amount of variation that is predictable, past consumption of health services provides the best indicator [63]. This suggests that if active pool selection can be undertaken by insurers, access to an individual's past consumption information provides the best potential for selection to occur. Provider-insurers are arguably the best-placed to practice selection on this basis as they have access, via medical records, to existing patients' past medical history.
Reported actions of NZPHCS health service providers declining to take people 'on the books' (i.e. as an insured individual) but providing care as a 'casual' patient [64] could be interpreted as either an act of screening in order to avoid taking on the risks of a patient with unknown health state until an assessment has been made, or an act of deliberately declining a patient of probable high demand based upon past registration. In either case, patients are being denied coverage by the insurer of their choice, even though that insurer, in the capacity as a service provider, is happy to provide casual treatment. Given that the risk to the provider from a casual treatment is less, by the foregoing reasoning, casual treatments should cost less to deliver than subsidised treatments, so the cost savings could be passed on to patients. However, few providers will be likely to offer such differentiated prices, as low-subsidised patients willing to self-insure would likely face lower costs under such a system, so would respond by eschewing the subsidised system and opting for casual treatment, thereby reducing both the possibility of the low-cost practitioner extracting profits in excess of those feasible at the population average, and the ability for providers to recoup risk-management costs disproportionately from the highest co-paying class of patients.
Conclusion
The preceding analysis indicates that the arrangements that have emerged under the NZPHCS, whilst in principle a managed care insurance system, are substantially less than optimal in respect of the requirements for a fully-functioning, efficient managed care insurance model. It appears that the use of institutions and relationships that prevailed under the pre-NZPHCS system in an 'evolutionary' move towards what appears to be a fully-fledged managed care model are likely to be counter-productive to the equity objectives of the strategy, substantially more costly than either the pre-NZPHCS system or the optimal managed care model, and as a consequence of the limited competition that exists, unlikely to respond to the normal competitive pressures to evolve into a more cost-efficient model. Thus, rather than the institutions and contracts being a 'first step' in an evolving strategy, the current higher-cost institutions and contracts are likely to become entrenched, to the long-term detriment of both taxpayers and patients.
Two instruments appear to be critical in the inability of the NZPHCS to achieve its full range of managed care objectives. The first is the presumption that general practitioners would retain their individual right to set patient co-payments independent of PHO contracts. This presumption has left PHOs with no meaningful ability to practice financial risk management. Whilst there is no policy statement about the retention of the right to charge, the unstated assumption that it exists has restricted the use of fiscal strategies associated with traditional managed care models. However, even if no such assumption existed, the second instrument, the requirement that service providers be part of PHO governance, means that providers have been granted sufficient power to ensure that the contracts under the NZPHCS do not leave them any worse off than pre-NZPHCS. In their capacity as PHO governors, general practitioners would be unlikely to be party to designing contracts that limit the professional autonomy that has been their non-negotiable bottom line in the development of New Zealand primary health care policy since 1938. As successive governments have been either unable or unwilling to restrain general practitioner charging autonomy using legislative powers, it is unlikely that PHOs, most of which are operating as general practitioner supplier-owned co-operatives, would be able to achieve such restraints using only mutually agreed contractual mechanisms.
Given that the outcomes of the NZPHCS are largely predictable from an analysis of risk-bearing in health insurance markets and the New Zealand history and institutions, the advisability of instituting a full, insurance-based managed care model with the associated requirements on the insurance companies to manage costs in a health care market that effectively denies to these managed care insurance organisations half of the tools normally available to such organisations to manage their costs must be questioned. The higher costs and inequitable distribution can be observed even at this early stage and will become more substantial over the next four years as premium subsidies increase. Indeed, the full extent of the higher costs at an individual level are likely being masked in the early stages by the sheer size of the additional government funding injected into the sector. As unrestricted practitioner charging negates the effects of patient demand pooling that normally accompany insurance schemes, any attempt to set up a managed care system based upon capitation payments of either the insurer or the service providers appears to be fundamentally flawed. This is not to say that practice-based strategies associated with PHO management and services to improve access cannot be legitimately and effectively funded on a capitation basis and provide measurable benefits. Indeed, benefits may have already accrued from these strategies. Rather it is a commentary on the wisdom of using capitation payments to fund the service delivery components of a universal insurance scheme in the presence of practitioner price-setting autonomy. Under current arrangements, the practice-based strategies would have to be extremely effective to outweigh the substantial additional costs of the imperfect insurance instruments that attend the NZPHCS in order to deliver a system that is of net greater benefit or offers better value-for-money than its predecessor.
In principle, leaving aside the power and professional autonomy of medical practitioners, the structures and intended contracts in the NZPHCS offered a potentially viable model of competing insurers and competing service providers that had the potential to deliver an efficient and effective insurance-based primary health care system for New Zealand that was capable of real innovation in both contracting and service delivery. However, to do so would have required a truly competitive environment, both in respect of insurance and service delivery markets, with fully transferable insurance premiums independent of insurer-service delivery contracts, and where insurers were governed, managed and operated as specialist insurance companies [65]. Such competition in the market for purchasers of services, where patients could freely exercise their insurance custom, would lead to genuine competition in the markets for both insurance customers and service provision contracts, where it would be harder for provider co-operatives to unilaterally determine the terms and conditions under which they would enter into contracts with insurers. This would have led to vibrant competition not just for contracts, but also in the models of insurance and care delivery that could move beyond the managed care model under which the system was established to alternative arrangements. If such a system allowed patient co-payments determined by the insurer, it would resemble the United States managed care model, but with the potential to evolve if models other than managed care proved more efficient. Such innovation is occurring in the United States as managed care proves less desirable for some patients and their insurers [66]. If patient co-payments were not allowed, then the New Zealand system it would resemble England's NHS. However, this model would require full funding from government sources.
In its present state however, the NZPHCS resembles neither of these models. It allows for all the additional costs of an insurance-based system, but none of the equity benefits of a fully state-funded system, and none of the fiscal benefits of a managed care system that constrains some of the excesses of insurance-based systems. Whilst there may be gains from the practice-based managed care strategies currently undertaken by PHOs, the costs of these gains will not necessarily be able to compensate for the substantial extra costs of the system as it is currently operating. If the costs and inequities of the NZPHCS escalate as predicted, any gains may be quickly eroded. Unless the deficiencies of the current insurance-based system are addressed soon, the very substantial proportion of the additional government funding committed to primary health care will likely amount to an ill-judged, overly-costly investment.
Competing interests
The author(s) declares no competing interests.
Acknowledgements
The author acknowledges the helpful comments provided by Mark Berry and Lewis Evans, and the support of Glenn Boyle and the New Zealand Institute for the Study of Competition and Regulation, in the preparation of this article.
==== Refs
King A The Primary Health Care Strategy Ministry of Health 2001
Hefford M Crampton P Foley J Reducing health disparities through primary care reform: the New Zealand experiment Health Policy 2005 72 9 23 15760695 10.1016/j.healthpol.2004.06.005
Consumers Institute
King op cit [1] at p 17 2001
King op cit [1] at p vii 2001
King op cit [1] at p 7 2001
Crampton P Sutton F Foley J Capitation funding of primary care services: principles and prospects Journal of the New Zealand Medical Association 2002 115 no. 1155. 07 June 2002
Ashton T Recent developments in the funding and organisation of the New Zealand health system Australia and New Zealand Health Policy 2005 2 9 at p 5 15679896
Austin C The public and private interface in New Zealand primary health care The Royal New Zealand College of General Practitioners 2004 at p 3
Good Returns
Davies P Health insurance in New Zealand Healthcare and Informatics 1999 3
Crampton Sutton Foley op cit [7] at pt 1 2002
King op cit [1] 2001
Perera R McDonald J Cumming J Goodhead A Primary Health Organisations: the first year (July 2003–June 2003) from the PHO perspective Victoria University of Wellington Health Services Research Centre 2003
King op cit [1] at p 7 2001
Howell B Restructuring primary health care markets in New Zealand: financial risk, competition, innovation and governance implications New Zealand Institute for the Study of Competition and Regulation 2005 Chapter Three
King op cit [1] at p viii 2001
Ministry of Health
Ministry of Health
Howell op cit [16] at p 172 2005
Howell op cit [16] Chapter Five 2005
Howell op cit [16] Appendix 4 2005
Howell op cit [16] at p 105 2005
Hefford Crampton Foley op cit [2] at p 17 2005
Howell op cit [16] at p 139 2005
Howell op cit [16] at p 140 2005
Arrow K Uncertainty and the economics of medical care American Economic Review 1963 53 941 973
Rochet J-C Tirole J Platform competition in two-sided markets Journal of the European Economic Association 2003 1 990 1029 10.1162/154247603322493212
Zeckhauser R Medical insurance: A case study of the tradeoff between risk spreading and appropriate incentives Journal of Economic Theory 1970 2 10 26 10.1016/0022-0531(70)90010-4
Danzon P Maclaine G Health care reform: an international perspective The Australian Economic Review 1994 81 88 Second Quarter at p 81
Barnett R Barnett P Primary health care in New Zealand: problems and policy approaches Social Policy Journal of New Zealand 2004 21 49 66 at p 60
King A Memorandum to Cabinet Primary Health Care Strategy- Achieving Low Cost Access Ministry of Health 2004 at p 2
Newhouse J Reimbursing health plans and health providers: efficiency in production versus selection Journal of Economic Literature 1996 34 1236 1263
Danzon Maclaine op cit [30] 1994
Rivers P Tsai K Managing costs and managing care International Journal of Health Care Quality Assurance 2003 14 302 307 at p 302
Danzon P Tort Liability: a minefield for managed care Journal of Legal Studies 1997 26 491 519 at p 500 10.1086/468006
Danzon op cit [36] 1997
James E Rose-Ackerman S The Nonprofit Enterprise in Market Economics Harwood 1986
Robinson J Theory and practice in the design of physician payment incentives The Milbank Quarterly 2001 79 149 77 11439463 10.1111/1468-0009.00202
Holmstrom B Milgrom P Multitask Principal-Agent Analyses: Incentive Contracts, Asset Ownership and Job Design Journal of Law and Economic Organisation 1991 7 24 52
Ma C-T A Riordan M Health insurance, moral hazard and managed care Journal of Economics and Management Strategy 2002 11 81 107 10.1162/105864002317247587
Ma C-T A McGuire T Network incentives in managed health care Journal of Economics and Management Strategy 2002 11 1 35 10.1162/105864002317247569
Newhouse op cit [33] at p 1236 1996
Robinson J Reinvention of health insurance in the consumer era Journal of the American Medical Association 2004 291 1880 1886 15100208 10.1001/jama.291.15.1880
Department of Health Delivering the NHS Plan: next steps on investment, next steps on reform United Kingdom Secretary of State for Health 2002 Document Cm 5503
Department of Health op cit [45] 2002
Danzon op cit [36] 1997 Newhouse (1996) op cit [33]
Gravelle H Capitation contracts: access and quality Journal of Health Economics 1999 18 315 340 10537898 10.1016/S0167-6296(98)00046-0
Ministry of Health
Jordan C McCardle J Norgrove K Review of Primary Health Organisation Management Services: Report to the Ministry of Health Auckland, New Zealand: Capital Strategy 2004
Howell op cit [16] Chapter Five 2005
Commerce Commission v Ophthalmological Society of New Zealand Incorporated, High Court, Wellington, CP 354/97 1 March 2004, Gendall J
Consumers Institute 2005
Ministry of Health 2005
Milgrom P Roberts J Economics, Organization and Management Englewood Cliffs 1992 New Jersey: Prentice-Hall 206 247
Anderson G Weller W Methods of reducing the financial risk of physicians under capitation Archives of Family Medicine 1999 8 149 155 at p 151 10101986 10.1001/archfami.8.2.149
Robinson op cit [44] at p 1883 2004
Anderson Weller op cit [56] at p 152 1999
Anderson Weller op cit [56] at p 152 1999
Howell op cit [16] Appendix Four 2005
Danzon Maclaine op cit [30] at p 81 1994
King op cit [1] at p 14 2001
Robinson op cit [46] 2004 Newhouse (1996) op cit [44]
Stevenson R Overworked GPs close books on new PHO patients Kapi Mana News 2005 29/03/05 p1
Gaynor M Haas-Wilson D Vogt W Are invisible hands good hands? Moral hazard, competition, and the second-best in health care markets Journal of Political Economy 108 992 1005 10.1086/317672
Robinson op cit [44] 2004
Ministry of Health op cit [19]
|
16144544
|
PMC1224852
|
CC BY
|
2021-01-04 16:38:28
|
no
|
Aust New Zealand Health Policy. 2005 Sep 6; 2:20
|
utf-8
|
Aust New Zealand Health Policy
| 2,005 |
10.1186/1743-8462-2-20
|
oa_comm
|
==== Front
BMC BioinformaticsBMC Bioinformatics1471-2105BioMed Central London 1471-2105-6-2051612021610.1186/1471-2105-6-205Research ArticleA comparative study of discriminating human heart failure etiology using gene expression profiles Huang Xiaohong [email protected] Wei [email protected] Suzanne [email protected] Xinqiang [email protected] Yingjie [email protected] Soon J [email protected] Leslie W [email protected] Jennifer [email protected] Division of Biostatistics, School of Public Health, University of Minnesota, Minneapolis, MN 55455, USA2 Cardiovascular Division, Department of Medicine, Medical School, University of Minnesota, Minneapolis, MN 55455, USA2005 24 8 2005 6 205 205 13 4 2005 24 8 2005 Copyright © 2005 Huang et al; licensee BioMed Central Ltd.2005Huang et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Human heart failure is a complex disease that manifests from multiple genetic and environmental factors. Although ischemic and non-ischemic heart disease present clinically with many similar decreases in ventricular function, emerging work suggests that they are distinct diseases with different responses to therapy. The ability to distinguish between ischemic and non-ischemic heart failure may be essential to guide appropriate therapy and determine prognosis for successful treatment. In this paper we consider discriminating the etiologies of heart failure using gene expression libraries from two separate institutions.
Results
We apply five new statistical methods, including partial least squares, penalized partial least squares, LASSO, nearest shrunken centroids and random forest, to two real datasets and compare their performance for multiclass classification. It is found that the five statistical methods perform similarly on each of the two datasets: it is difficult to correctly distinguish the etiologies of heart failure in one dataset whereas it is easy for the other one. In a simulation study, it is confirmed that the five methods tend to have close performance, though the random forest seems to have a slight edge.
Conclusions
For some gene expression data, several recently developed discriminant methods may perform similarly. More importantly, one must remain cautious when assessing the discriminating performance using gene expression profiles based on a small dataset; our analysis suggests the importance of utilizing multiple or larger datasets.
==== Body
Background
Human heart failure is a complex disease diagnosed in over 500,000 American people every year, causing more than 250,000 deaths annually. It may arise from coronary atherosclerosis, exposure to toxins, infection, inflammation, valvular disease leading to volume/pressure overload, or an underlying genetic or idiopathic event [1-3]. Emerging work suggests the heterogeneity of heart failure. For example, patients with ischemic heart failure have decreased survival compared to the non-ischemic heart failure [4,5] and respond differently to therapies [6-9]. Although benefits can be achieved using ischemic heart failure therapy for idiopathic heart failure (and vice versa), cure rates will be markedly diminished, and unwarranted toxicities problems will be encountered. It may be critical to distinguish these characteristically similar but clinically somewhat distinct heart failures, to better optimize therapy. The ability to distinguish between different etiologies of heart failure may be essential to guide appropriate therapy and determine prognosis for successful treatment.
A new approach to discriminating etiologies of heart failure is gene expression profiling using DNA microarray technology, which has been shown to be promising in the diagnosis of human diseases or subdiseases, especially in cancer [10-12]. Recent genomic studies by three separate groups have demonstrated a distinct etiology dependent genomic pattern in the failing heart [13-16]. These studies offer hope that the microarray gene expression analysis could potentially add to conventional laboratory approaches to diagnose different underlying etiologies of heart failure while simultaneously enhance prognostic criteria. It was hypothesized that heart failure arising from different underlying etiologies present with different gene expression patterns and that these differences could be used as a diagnostic tool. Here we test the hypothesis with two human heart failure datasets from different institutions.
Sample classification with gene expression data is statistically challenging due to the "small n, large p" problem [17]: the number of samples n is much smaller than the number of genes or predictors p. In our first dataset, we have n = 30 and p > 20000. Many new statistical methods have been developed or adapted to face the challenge. With more and more new methods emerging and existing methods being adapted, it becomes increasingly compelling for practitioners to compare and assess their performance, but there are few such comparative studies [18-20]. Huang and Pan [19] compared several methods, including partial least squares (PLS) [21], nearest shrunken centroid (SC) [12], and a penalized PLS (PPLS), for binary classification of gene expression data. They found that these methods are competitive. More recently, some authors [22,20] have shown the promising performance of least absolute shrinkage and selection operator (LASSO) [23] and random forest (RF) [24]. It is our main goal to evaluate and compare these methods using two human heart failure datasets. For this purpose, we also extend the PPLS, originally proposed for binary classification, to multiclass classification. We found that the above five statistical methods perform similarly. Furthermore, our analysis stresses the importance of utilizing multiple datasets for classification purposes.
Results
Minnesota data
Myocardial tissue samples from the left ventricular apex of patients with severe refractory heart failure were collected at the time of the left ventricular assist device (LVAD) placement at the University of Minnesota Medical School. A total of 30 tissue samples were processed for microarray analysis on the Affymetrix Human Genome U133A chip containing ~22,000 genes. The initial data analysis was completed using Affymetrix Microarray Suite (MAS 5.0). A more complete description on the data is provided in [25].
The heart failure patients are divided into three classes according to the underlying etiology. Patients with clinical ECHO and EKG evidence, history of previous myocardial infarctions, and direct observation of the heart for confirmation of infarction at the time of LVAD implantation are defined as ischemic. Patients with an ischemic etiology were further divided into two classes: patients with ischemia but without acute myocardial infarctions (ischemic class) and patients with ischemia that have had an acute myocardial infarctions within ten days of LVAD implant (IM class), the remaining patients were assigned to the idiopathic class. Among the total 30 samples, 10 of them are ischemic, 7 are IM and 13 are idiopathic.
PGA data
The PGA data were obtained in another heart failure study conducted at the Cardio-Genomics PGA (Programs for Genomic Applications) at the Harvard Medical School. Myocardial samples were collected from patients undergoing heart transplantation whose failure arises from different etiologies (e.g. idiopathic, ischemic, alcoholic, valvular, and hypertrophic) and from normal organ donors whose hearts were not used for transplants. The transcriptional profile of the mRNA in these samples was also measured with Affymetrix oligonucleatide microarray technology. HG-U133 plus 2 chips containing 54,675 probe sets were used and data were analyzed in MAS 5.0. In the PGA data set there were 11 normal samples, 11 ischemic samples and 14 idiopathic samples. The PGA dataset is publicly available at Genomics of Cardiovascular Development, Adaptation, and Remodelling. NHLBI Program for Genomic Applications, Harvard Medical School with URL: .
In order to make the results comparable to those based on the HG-U133A chips used in the Minnesota data, we matched the probe sets on a HG-U133 plus 2 chip with those on a HG-U133A chip. Only six probe sets on the HG-U133 plus 2 chip could not be found on a HG-U133A chip. Hence we used the remaining 22277 ( = 22283 - 6) probe sets in the following analyses with the PGA data.
Classification with Minnesota data: One-against-others approach
Table 1 reports the LOOCV misclassification errors of the five classification methods for the Minnesota data with models starting from different numbers of top ranked genes ranging from 50 to all genes. Four kinds of errors are reported for each method: three one-against-others two-class LOOCV misclassification errors (Ischemic vs. others, IM vs. others and Idiopathic vs. others) and one three-class LOOCV misclassification error. Note that throughout this article, three-class classification results for SC and RF were obtained by direct applications of SC and RF, rather than by combining multiple binary classifications.
Table 1 LOOCV errors for three-class classification with Minnesota data: all patients.
# of top genes Isch vs other IM vs other Idio vs other Overall
PPLS PLS SC LSO RF PPLS PLS SC LSO RF PPLS PLS SC LSO RF PPLS PLS SC LSO RF
50 7 9 8 11 12 6 8 6 7 9 8 6 5 8 6 14 11 11 12 14
100 9 6 8 9 11 6 7 6 7 8 10 9 5 11 8 11 9 11 13 13
200 9 7 12 9 11 7 9 7 7 9 6 7 5 6 7 12 11 12 8 14
400 11 11 13 10 12 8 9 8 7 8 10 8 8 8 8 16 15 14 11 15
800 8 11 12 13 12 9 9 8 8 8 8 8 7 10 8 15 17 15 14 16
1600 12 12 11 11 11 9 8 8 5 8 12 8 5 7 7 16 17 13 11 13
3200 10 12 9 15 10 9 9 7 6 7 9 7 5 10 8 14 14 14 14 13
6400 11 11 10 15 10 9 10 9 8 7 9 8 5 8 8 15 14 13 15 12
9600 10 11 10 15 10 10 8 7 9 7 7 8 4 9 10 14 12 15 16 15
12800 13 12 11 15 11 7 7 7 9 7 8 8 5 10 9 15 13 15 17 17
16000 13 12 13 15 10 8 7 10 9 7 8 8 5 10 9 15 13 14 17 16
19200 13 12 13 15 11 9 7 10 9 7 8 8 5 10 8 15 14 14 17 17
22283 13 13 11 15 10 7 6 11 9 7 7 7 5 10 10 14 14 14 18 16
Based on Table 1, we first note that the performances of any classifier is sensitive to the number of top ranked genes one starts with. For example, in the three-class classification, LASSO made 8 errors when only the top 200 genes are considered but made 18 errors when all 22283 genes are used. But there is no obvious relationship between the gene subset size and the five methods' performances.
In the two-class classification problem of ischemic vs. others, PPLS, PLS, SC, LASSO and RF yielded LOOCV errors range from 7 to 15, 6 to 13, 8 to 13, 9 to 15 and 10 to 12 respectively. The five methods obtained similar numbers of errors in all instances. In the two-class classification problems of IM vs. others and idiopathic vs. others, all five methods yielded similar numbers of errors. In the overall three-class classification, again all the five classification methods perform similarly. There is no clear evidence that one method is clearly superior to others.
In order to assess whether a gene expression profile is affected by gender, we classified the 23 samples from male patients only. Among the 23 samples for male patients, 9 of them are ischemic, 7 are IM and 7 are idiopathic.
Table 2 reports the LOOCV misclassification errors of these five methods for the Minnesota data with males only. Again, the models start from different numbers of top ranked genes. The three two-class LOOCV misclassification errors and one three-class LOOCV misclassification error are estimated for each method as described before.
Table 2 LOOCV errors for three-class classification with Minnesota data: males only.
# of top genes Isch vs other IM vs other Idio vs other Overall
PPLS PLS SC LSO RF PPLS PLS SC LSO RF PPLS PLS SC LSO RF PPLS PLS SC LSO RF
50 11 9 11 10 12 7 6 5 7 7 5 6 6 6 5 15 12 11 14 15
100 10 10 10 11 11 5 7 5 8 7 5 6 6 5 8 10 12 11 13 14
200 12 12 16 10 12 6 7 6 8 5 6 6 7 7 8 13 12 11 13 15
400 10 12 15 10 12 6 8 7 7 7 5 7 8 6 9 14 14 11 14 13
800 14 12 13 9 12 6 7 6 8 7 5 6 8 6 9 12 11 12 14 15
1600 15 15 12 7 10 6 4 7 8 6 5 5 8 8 8 13 12 13 13 17
3200 11 11 11 10 11 7 7 7 8 8 5 5 5 8 8 14 11 14 15 15
6400 11 12 13 10 11 6 8 6 11 7 7 6 4 8 8 11 14 15 19 14
9600 9 11 14 9 10 7 7 6 8 7 8 7 4 8 8 13 13 13 19 15
12800 9 10 15 10 10 5 8 5 8 8 8 8 5 8 7 12 14 15 20 15
16000 13 12 15 10 9 8 6 5 8 7 9 7 5 8 7 14 12 14 20 17
19200 13 10 11 10 9 6 6 5 8 7 7 6 6 8 8 14 12 13 20 17
22283 12 11 9 10 11 7 5 5 8 8 5 5 6 8 8 15 11 14 20 15
Based on Table 2, again we note that the classification performances of PPLS, PLS, SC, LASSO and RF can be quite sensitive to the number of top ranked genes one uses. For example, in the one-against-others two-class classification problem of ischemic vs. others, PPLS made 10 errors when the top 400 genes are considered but the number of errors suddenly increases to 14 when the top 800 genes are used. Again there is no obvious relationship between the gene subset size and the five methods' performances.
Comparing PPLS, PLS, SC, LASSO and RF, we find that the five methods perform very similarly in almost all instances. There is no evidence that any method is clearly the best. One thing we noticed about LASSO is that when the model contains many genes, say top 12800, then it gives more errors than PPLS, PLS, SC and RF in the three-class classification. This could happen by chance since we did not observe the same trend in the decomposed one-against-others binary classifications.
As in Table 1, the results in Table 2 suggest that discriminating ischemic group from the other two groups was less accurate than distinguishing IM from the other two groups and separating idiopathic group from the other two groups.
If we compare Table 1 (30 patients) with Table 2 (23 male patients), we can see that the misclassification error rates of all the five methods in Table 2 are much higher than those in Table 1. Reduced sample size is likely a factor.
To see whether the high prediction errors are due to the presence of the three classes, we considered a binary classification problem. We applied all the five methods to the 10 ischemic and 13 idiopathic samples. We also assessed the classification accuracy on the male patients with 9 ischemic and 7 idiopathic samples. The classification results were shown in Table 3.
Table 3 LOOCV errors for two-class classification with Minnesota data: ischemic vs idio-pathic.
# of top genes All (23 samples) Males (16 samples)
PPLS PLS SC LASSO RF PPLS PLS SC LASSO RF
50 10 6 5 6 6 10 9 11 8 11
100 7 7 5 7 6 10 9 11 11 12
200 9 9 7 11 6 9 9 11 11 9
400 6 8 7 11 8 9 9 8 12 9
800 6 9 8 4 7 10 10 9 12 10
1600 5 8 8 8 8 10 10 9 12 10
3200 9 10 9 8 10 8 8 10 12 9
6400 7 7 9 8 9 6 6 10 12 9
9600 9 8 8 7 8 6 6 10 12 8
12800 8 8 8 8 11 6 6 10 12 10
16000 8 7 7 8 9 7 6 10 12 11
19200 11 8 6 7 11 8 7 9 12 9
22283 9 10 5 7 9 8 8 9 12 7
From Table 3, we can see that all the five methods have very similar performances in classifying ischemic and idiopathic samples. If we compare the classification performances of these five methods with/without female patients, taking the sample size into consideration, we can see that the misclassification error rates with only male patients is much higher than those with all patients. This again is probably because the sample size (16) with only males is smaller. In particular, we noticed that LASSO is more sensitive to the small sample size.
Other models
In the previous classification problems, we only included linear terms of gene expression levels in a model. We also considered expanded models including squared terms of each gene's expression levels. The motivation is to possibly improve model fitting, for instance, to avoid masking in linear models [26]. In this way, the number of variables in the new data is doubled (the original variables plus their squared terms) and we have 44566 variables. We repeated all the previous classification procedures and found that the classification performance did not improve (results not shown).
Classification with Minnesota data: pair-wise approach
We repeated the three-class classification with PPLS via the pair-wise approach. The results are included in Table 4. We assessed the classification of PPLS by including all 30 patients and with 23 male only. By comparing Table 4 to Table 1 – 2, we can see that the PPLS via one-against-others approach gives much smaller errors. This suggests that for this specific problem, the one-against-others approach is probably better. Again, we see the classification with male patients gives much larger LOOCV misclassification error rates.
Table 4 LOOCV errors for three-class classification: PPLS with pair-wise approach.
#of top genes Minnesota data
All patients Male
50 18 16
100 13 18
200 16 18
400 17 17
800 18 14
1600 17 16
3200 16 14
6400 15 15
9600 17 15
12800 17 14
16000 19 14
19200 14 16
22283 19 16
Classification with PGA data: one-against-others approach
Table 5 reports the LOOCV misclassification errors of the five methods for the PGA data. Four kinds of errors are reported for each method: three one-against-others two-class LOOCV misclassification errors (Normal vs. others, Ischemic vs. others, and Idiopathic vs. others) and one three-class LOOCV misclassification error.
Table 5 LOOCV errors for three-class classification with PGA data: all patients.
# of top genes Normal vs other Isch vs other Idio vs other Overall
PPLS PLS sc LSO RF PPLS PLS SC LSO RF PPLS PLS SC LSO RF PPLS PLS SC LSO RF
50 0 0 0 0 0 3 1 2 2 1 5 1 5 1 2 2 1 1 1 1
100 0 0 0 0 0 4 2 3 1 2 3 1 3 1 1 1 1 1 1 1
200 0 0 0 1 0 5 1 2 1 2 2 2 2 2 1 1 1 1 2 1
400 0 0 0 0 0 3 1 2 4 2 2 1 2 2 2 2 1 1 2 1
800 0 0 0 0 0 3 1 2 1 2 2 1 2 2 2 2 1 1 1 1
1600 0 0 1 0 0 2 1 2 1 2 2 2 2 2 2 2 1 1 2 1
3200 0 0 2 0 0 3 2 2 2 2 2 3 3 2 2 2 2 1 2 1
6400 0 0 2 0 0 3 2 2 1 2 2 2 3 2 2 2 2 2 2 1
9600 0 0 0 0 0 2 1 2 1 2 3 2 2 2 2 2 1 1 2 1
12800 0 0 0 0 0 3 1 2 1 2 3 2 2 2 2 1 1 1 2 1
16000 0 0 0 0 0 2 1 2 1 2 3 2 2 2 2 1 1 1 2 1
19200 0 0 0 0 0 2 1 2 1 2 2 1 2 2 2 1 1 1 2 1
22277 0 0 0 0 0 2 1 2 1 2 2 1 2 2 2 1 1 1 2 1
Based on Table 5, we can see that the classification performances of PPLS, PLS, SC, LASSO and RF are quite stable with different numbers of top ranked genes one uses. In the two-class classification problem of normal vs. others, the five classification methods almost perform perfectly, where PPLS, PLS and RF have 0 errors in all circumstances, SC has 0 errors in all situations except for 3 cases and LASSO has 1 error in one case and 0 errors in all other cases. In the two-class classification problems of ischemic vs. others and idiopathic vs. others, PPLS, PLS, SC, LASSO and RF yielded 1–5 errors (mostly with 1–3 errors). That the problem of distinguishing normal from the others is the easiest confirms that the normal class is more separable from the other two classes. As for the three-class classification, the errors range from 1 to 2 and it is almost perfect.
It may be argued that a classification involving normal samples should be much easier because the normal class is very different from the other two classes. Correct diagnosis between ischemic and idiopathic would be much more challenging. Hence we conducted a two-class classification with 11 ischemic and 14 idiopathic samples. The results are shown in Table 6. The five methods perform almost perfectly: the misclassification errors range from 1 to 3 for all five methods in all models.
Table 6 LOOCV errors for two-class classification with PGA data: ischemic vs idiopathic.
# of top genes PPLS PLS SC LASSO RF
50 3 2 2 2 2
100 1 1 2 1 1
200 1 1 2 1 2
400 1 1 1 2 1
800 1 1 1 3 1
1600 1 1 1 2 1
3200 1 2 1 1 1
6400 1 3 1 1 1
9600 1 3 1 1 1
12800 1 3 1 1 1
16000 1 2 1 1 1
19200 1 1 1 1 1
22277 1 1 1 1 1
Genes identified
We consider genes remaining in a final model for each method. To save space, we restrict attention to models starting with all the genes for binary discrimination between ischemic and idiopathic samples. Briefly, LOOCV was first used to select any tuning parameters in a method (e.g. number of components in a PLS model), then a model with the selected parameters was fitted using all the samples. Except that all the genes are used in a final PLS model, for any of the other methods there may be fewer genes remaining in the final model. In particular, LASSO can select at most n genes with n the number of the samples. It turned out that random forest also used many of the genes.
Tables 7 and 8 lists the genes selected by at least four or three methods for the Minnesota data and PGA data respectively. It can be seen that there is no overlap at all between the two gene lists. Although the same genes are not identified from the two datasets, it is clear that the beta-adrenergic signalling pathway is likely a discriminatory pathway, given the inclusion of CREM in the Minnesota data and AKAP6 in the PGA data. Furthermore, the inclusion of metabolic-related genes, such as ATPase and GAPD, is not surprising given the class of ischemic tissue.
Table 7 Genes selected by ≥ 4 methods in two-class classification (ischemic vs idiopathic) with Minnesota data. The last row gives the total numbers of the genes in the final models.
Probe Set Gene Rank PPLS PLS SC LSO RF
215066_at PTPRF: protein tyrosine phosphatase, receptor type, F 1 X X X X
212008_at UBXD2: UBX domain containing 2 X X X X
217234_s_at VIL2: villin 2 (ezrin) 3 X X X X
202092_s_at BART1: binder of Arl Two 5 X X X X
218318_s_at NLK: nemo-like kinase 7 X X X X
212062_at ATP9A: ATPase, Class II, type 9A 9 X X X X
218208_at FLJ22378: hypothetical protein FLJ22378 13 X X X X
212093_s_at MTSG1: mitochondrial tumor suppressor gene 1 14 X X X X
214543_x_at QKI: quaking homolog, KH domain RNA binding (mouse) 24 X X X X
209487_at RBPMS: RNA binding protein with multiple splicing 26 X X X X X
202877_s_at C1QR1: complement component 1, q subcomponent, receptor 1 31 X X X X
64438_at FLJ22222: hypothetical protein FLJ22222 40 X X X X
221928_at LOC283445: hypothetical protein LOC283445 44 X X X X
212556_at SCRIB: scribble 48 X X X X
202641_at ARL3: ADP-ribosylation factor-like 3 50 X X X X
201559_s_at CLIC4: chloride intracellular channel 4 64 X X X X
216231_s_at Homo sapiens transcribed sequence with strong similarity to protein pdb:3HLA (H.sapiens) B Chain B, Human Class I Histocompatibility Antigen A2.1 (HLA-A2.1 Human Leucocyte Antigen) 73 X X X X X
220477_s_at C20orf30: chromosome 20 open reading frame 30 76 X X X X
208879_x_at C20orfl4: chromosome 20 open reading frame 14 88 X X X X
207630_s_at CREM: cAMP responsive element modulator 107 X X X X
212117_at ARHQ: ras homolog gene family, member Q 116 X X X X
212904_at KIAA1185: KIAA1185 protein 121 X X X X
M33197_5_at GAPD: glyceraldehyde-3-phosphate dehydrogenase 161 X X X X
213507_s_at KPNB1: karyopherin (importin) beta 1 163 X X X X
207627_s_at TFCP2: transcription factor CP2 208 X X X X
Total - - 275 22283 36 22 883
Table 8 Genes selected by ≥ 3 methods in two-class classification (ischemic vs idiopathic) with PGA data, The last row gives the total numbers of the genes in the final models.
Probe Set Gene Rank PPLS PLS SC LSO RF
206375_s_at HSPB3: heat shock 27kDa protein 3 1 X X X X
202430_s_at PLSCR1: phospholipid scramblase 1 2 X X X X
209948_at KCNMB1: potassium large conductance calcium-activated channel, subfamily M, beta member 1 3 X X X
AFFX-TrpnX-5_at 4 X X X X
212929_s_at KIAA0592: KIAA0592 protein 6 X X X X
221415_s_at MYCBP: c-myc binding protein 7 X X X
219099_at C12orf5: chromosome 12 open reading frame 5 8 X X X
219383_at FLJ14213: hypothetical protein FLJ14213 10 X X X X
208846_s_at VDAC3: voltage-dependent anion channel 3 11 X X X X
205359_at AKAP6: A kinase (PRKA) anchor protein 6 12 X X X
202324_s_at GOCAP1: golgi complex associated protein 1, 60 kDa 14 X X X X
208736_at ARPC3: actin related protein 2/3 complex, subunit 3, 21 kDa 15 X X X
215700_x_at CPNE6: copine VI (neuronal) 16 X X X
207600_at KCNC3: potassium voltage-gated channel, Shaw-related subfamily, member 3 17 X X X
217386_at 18 X X X
200961_at SPS2: selenophosphate synthetase 2 19 X X X X
211476_at MYOZ2: myozenin 2 21 X X X
209682_at CBLB: Cas-Br-M (murine) ecotropic retroviral transforming sequence b 22 X X X
208769_at EIF4EBP2: eukaryotic translation initiation factor 4E binding protein 2 23 X X X
208162_s_at FLJ10232: hypothetical protein FLJ10232 25 X X X
210500_at NICE-4: NICE-4 protein 26 X X X
216721_at LOC253512: hypothetical protein LOC253512 27 X X X
206475_x_at CSH1: chorionic somatomammotropin hormone 1 (placental lactogen) 28 X X X
210561_s_at WSB1: SOCS box-containing WD protein SWiP-1 29 X X X
206598_at INS: insulin 30 X X X
219293_s_at PTD004: hypothetical protein PTD004 75 X X X
207431_s_at DEGS: degenerative spermatocyte homolog, lipid desaturase (Drosophila) 139 X X X
205207_at IL6: interleukin 6 (interferon, beta 2) 272 X X X
221775_x_at RPL22: ribosomal protein L22 1043 X X X
200897_s_at KIAA0992: palladin 1175 X X X
total - - 32 22277 7 22 548
We also give the univariate ranks of the genes (based on the F-statistics for the two classes) in the above two tables. It shows that the two sets of genes (or more generally, the genes in a final model) may not include some genes ranked high in the univariate ranking while including some ranked low, highlighting a possible limitation of solely depending on univariate ranks to select important genes.
More numerical results
PLS plots
To further explore why the methods all work much better for the PGA data than for the Minnesota data, we drew some plots using the first and the second PLS components for binary classification of ischemic vs idiopathic groups. We found that, for both datasets, there was a clear separation between the two groups. However, in LOOCV, although again the two groups were separable for both datasets, the left-out sample was more likely to be closer to the other group than to its true group for the MInnesota data, leading to a high LOOCV error rate; Figure 1 gives two examples. In contrast, in the PGA data, a left-out sample tended to be close to its true group, resulting in a low LOOCV error rate. For more details see Supplemental Materials.
Figure 1 PLS plots for two cases in LOOCV for the Minnesota data comparing ischemic vs idiopathic. In both cases, the new sample labeled as "N" (i.e. left-out sample in LOOCV), belonging to class 1 and 2 respectively in the top and the bottom panels, is closer to the other group different from its true class, leading to misclassifications.
Permutation tests
Because of high misclassification error rates with the Minnesota data, it is of interest to investigate whether there is any signal at all in the data. This can be accomplished by a permutation test that compares misclassification errors resulting from using the original data with that from randomly permuted data; a P-value is defined as the proportion of permuted datasets with misclassification errors fewer than that of the original data. To generate a randomly permuted data, we randomly permute the class labels of the original data. Because all the methods have similar performance, we consider the nearest shrunken centroid method with the Minnesota data. Tables 9 and 10 summarize the results of misclassification errors for 50 randomly permuted datasets for three- and two-class classifications respectively. It can be seen that the misclassification errors based on the original data are fewer than that based on the permuted data, leading to small P-values. This implies that, although there are relatively high misclassification error rates with the Minnesota data, the methods perform significantly better than a random guess.
Table 9 LOOCV three-class misclassification errors with the original Minnesota data and the percentiles of LOOCV errors with 50 permuted datasets by SC.
# of top genes Original data Permutated data
LOOCV errors P-value 0% 25% 50% 75% 100%
50 11 .00 13 17 20 22.75 29
100 11 .00 12 17.25 20 22 30
400 14 .02 13 18 20.5 23 28
1600 13 .00 13 18 20 23 29
6400 13 .02 12 18 20 23 29
Table 10 LOOCV two-class misclassification errors with the original Minnesota data and the percentiles of LOOCV errors with 50 permuted datasets by SC: ischemic vs idiopathic.
# of top genes Original data Permutated data
LOOCV errors P-value 0% 25% 50% 75% 100%
50 5 .00 5 10 11 12.75 19
100 5 .00 5 10 11 12 19
400 7 .06 4 9 11 13 17
1600 8 .08 5 9.25 11.5 13 17
6400 9 .12 6 10 11 13 21
Simulations
We did a simulation study to further evaluate the performance of the five classification methods. To mimic the real data, simulated data were generated from either a fitted PPLS or a fitted LASSO model to the Minnesota data comparing ischemic vs idiopathic, each containing top 50 genes in the initial model. Specifically, suppose that is the fitted response value for sample i based on the original Minnesota data using PPLS or LASSO. Note that is a real number without being dichotomized yet. Suppose that Yi = 1 or -1 is the class label of sample i in the original data, and . To generate a simulated data, we independently draw from Normal distributions for i = 1, ..., 23 and b = 1, ..., 50. Then we apply each method to a simulated dataset , where Xi is the gene expression profile of sample i in the original Minnesota data, and obtain fitted values ; the resulting misclassification error number for dataset b is .
Table 11 summarizes the distributions of the misclassification errors of each method based on 50 simulated data with either PPLS or LASSO as the true model. It can be seen that in general all the methods perform similarly, though random forest seems to be most stable and has a slight edge, and the performance of LASSO and nearest shrunken centroid may deteriorate as the number of the genes included in a model is increased. We also did other simulations with the true models starting from various numbers of top genes and various noise levels, and observed similar phenomena: for details see Supplemental Materials.
Table 11 Percentiles of misclassification errors from 50 simulated datasets for two-class classification. The true model is either PPLS or LASSO fitted with top 400 genes to the Minnesota data to compare ischemic vs idiopathic.
Methods # of top genes True model: PPLS True model: LASSO
Mean 0% 25% 50% 75% 100% Mean 0% 25% 50% 75% 100%
PPLS 50 2.54 0 2 2 3 6 0 0 0 0 0 0
100 2.6 1 1.25 2 3 7 1 1 1 1 1 1
400 2.58 0 2 2 3 7 0 0 0 0 0 0
1600 2.62 0 2 2 3.75 7 0 0 0 0 0 0
6400 2.6 0 2 2 3 6 0 0 0 0 0 0
PLS 50 2.58 0 2 2 3 6 0 0 0 0 0 0
100 2.6 0 2 2 3 7 0 0 0 0 0 0
400 2.56 0 2 2 3 7 0 0 0 0 0 0
1600 2.52 0 2 2 3 7 0 0 0 0 0 0
6400 2.54 0 2 2 3 6 0 0 0 0 0 0
LASSO 50 2.48 0 2 2 3 6 0 0 0 0 0 0
100 2.48 0 2 2 3 6 0 0 0 0 0 0
400 2.84 0 2 2 3 10 0 0 0 0 0 0
1600 3.48 0 2 2 4.75 10 0 0 0 0 0 0
6400 3.48 0 2 2 4.75 10 0 0 0 0 0 0
SC 50 2.76 1 2 2 3 7 1 1 1 1 1 1
100 2.74 1 2 2 3 7 1 1 1 1 1 1
400 2.96 0 2 3 3.75 7 0.28 0 0 0 1 1
1600 3.58 0 2 3 5 8 0 0 0 0 0 0
6400 4.08 1 3 4 5 10 1 1 1 1 1 1
RF 50 2.48 0 2 2 3 6 0 0 0 0 0 0
100 2.48 0 2 2 3 6 0 0 0 0 0 0
400 2.48 0 2 2 3 6 0 0 0 0 0 0
1600 2.48 0 2 2 3 6 0 0 0 0 0 0
6400 2.48 0 2 2 3 6 0 0 0 0 0 0
Discussion
With more and more statistical methods being proposed for discriminant analysis for gene expression data, it has become increasing important to compare and evaluate their performances with real data, as it has been done in other contexts [27]. Comparing the five new methods with each other using the two real datasets, we did not find anyone uniformly better than the others. This may be disappointing to someone who wishes to find the best statistical method. However, in the current application, the similar performance of all the five methods on each of the two datasets provides reassurance on the interesting observation that it is not equally easy to distinguish the different etiologies of heart failure using expression profiles in the two datasets.
Both the one-against-others approach and the pair-wise approach have been widely used in extending a binary classifier to multi-class settings. Our result suggests that, at least for the two datasets used here, the one-against-others approach is better, which was found to be true with support vector machines but in general should also depend on which binary classifier is used [28]. We also have observed that any of the five methods may be sensitive to the number of genes being included. This is particularly relevant because, although all the five methods (and many other methods) can handle any large number of genes, this does not dismiss the potential importance of a user's preliminary ranking and screening of genes. Of course, all our observations here are based on the two datasets without consideration of statistical variability, further studies are needed to validate these points.
An interesting finding of this work is that it is difficult to discriminate the different etiologies of human heart failure using one gene expression dataset, and at the same time, it is quite easy for the other dataset. A possible explanation is the different types of the microarray chips used: Affymetrix HG-U133A chips were used in the Minnesota study while Affymetrix HG-U133 plus 2 chips were used in the PGA study. Because the HG-U133 plus 2 chips contain more genes (or ESTs), to minimize the effects of using different genes, we only used the genes present in the Minnesota data and still yielded much better performance for the PGA data. In fact, we used all the genes in the PGA data and obtained similar results for the PGA data. Although we can say that the performance difference in the two datasets is not caused by different genes contained on a chip, we do not know whether the more recent HG-U133 plus 2 chips provide more reliable measurements on gene expression. In addition, quality control criteria for the inclusion of a chip were nearly identical between the two datasets. We would suspect that the performance difference may be the result of different patient populations and different study protocols (e.g. lack of clearly pre-specified patient inclusion/exclusion criteria). As discussed in [29], a key to validating any prognostic and diagnostic biomarkers is the use of data that can reflect the full range of clinical variability. This highlights the importance of utilizing multiple datasets drawn from multiple subpopulations. Even for the purpose of prediction for one subpopulation, it is possible to improve the performance by borrowing information from other subpopulations [30]. It can be argued that the performance should be weighted on the complexity of the disease. Challenges with the current clinical discrimination of ischemic versus non-ischemic heart failure is indeed why defining potential gene expression biomarkers may be a helpful additional approach in this characterization. A recognized limitation of utilizing heart tissue to identify biomarkers is the difficulty of collecting tissue. In summary, the current and other studies stress the importance of collaborating efforts to share tissue/data to strengthen the search for applicable biomarkers.
Conclusions
Many studies have aimed to develop new statistical and machine learning methods for best sample discrimination. Our results suggest that, at least for some gene expression data, several existing methods may work almost equally well. More importantly, because of the quite different performances of the methods on the two datasets, one must remain cautious when assessing the performance of sample discrimination using a small gene expression dataset; it may be necessary to use larger or multiple datasets to draw a more reliable conclusion.
Methods
Binary classifiers: PLS, PPLS and LASSO
We first briefly review the three binary classifiers, which was first designed for regression and can be directly applied to two-class classification, even when the number of covariates (i.e. genes here) is much larger than the sample size.
We code the response variable (i.e. class label) as Y = 1 for class 1 and Y = -1 for class 2. Suppose that xi is the expression level of gene i, i = 1, ..., p with p as the total number of the genes, and that we have n samples in the training data. A challenge is that we have n <<p.
The main idea of partial least squares (PLS) [21] is to seek a few linear combinations of for j = 1, ..., m, then apply ordinary least squares (OLS) to regress Y on zj's to obtain
with β's as OLS estimates. The key of course is how to form linear components zj's. It turns out that
αj = argmaxαCorr2(y, Xα) Var(Xα)
with the constraints ||α|| = 1, for l = 1,..., j - 1, where y is the vector of observed Y's (in the training data), X is the design matrix (i.e. matrix of observed x's), and S is the sample covariance matrix of x's [31]. In practice, the number of linear components m has to be chosen, typically by a form of cross-validation, such as leave-one-out cross-validation (LOOCV), to minimize misclassification errors.
PPLS is a penalized regression method in the framework of PLS [19,32]. Suppose that we have built a PLS linear model, which can be rewritten as:
Then we shrink the PLS coefficients by soft-thresholding [33,34]
where sign(a) = 1 if a ≤ 0 and sign(-a) = -1 if a < 0, λ is a shrinkage parameter to be determined, and f+ = max(f, 0). It is common that the shrinkage leads to many , effectively eliminating gene i from the model, thus gene selection is automatically accomplished. Next we construct a linear component . Finally a PPLS model is built by regressing Y on z using OLS
which can be re-expressed as . The parameters involved in building a PPLS model, such as the shrinkage parameter λ and the number of PLS components, are estimated by LOOCV. The goal is to choose the largest shrinkage parameter and the smallest number of PLS components for which the LOOCV misclassification error estimate is minimized.
The LASSO estimates [23] in a linear model
are obtained by
subject to , where Yi is the observed response for sample i and is its LASSO estimate, i = 1, ..., n, and t can be chosen by LOOCV. The constraint can often force many , leading to gene selection.
Note that the class label (1/-1) for the response Y is binary, but in any of the above binary classifiers, the response Y is treated as a continuous variable and the estimate could be any real number. To predict the class of a new sample, we use sign(): if the estimated response is greater than or equal to 0, then we classify it into class 1; otherwise, class 2. In particular, this direct use of PLS for binary classification (as in [35]) is different from other approaches [36-40]; a distinct advantage of our approach is its simplicity, e.g., avoiding convergence problems when two classes are perfectly separable, which is common in microarray data with a small sample size and a large number of genes.
Multiclass classifiers: nearest shrunken centroids and random forest
Nearest shrunken centroids (SC) is built on a diagonalized linear discriminant analysis (DLDA) [26,41]. Suppose that we have K classes, is the mean expression level of gene i in class k of the training samples, is the pooled sample variance of gene i of the training samples, and πk is the prior probability of a new sample being in class k. The DLDA rule for a new sample is
SC is motivated from the observation that many of the genes will not be predictive of the class membership and should be eliminated from the above DLDA rule. Formally, define
where nk is the number of training samples in class k, and is the overall mean expression level of gene i in all the training samples. Note that by the definition, we have . Let
for all i and k, where Δ is the shrinkage parameter to be chosen by LOOCV. Then substituting in the DLDA rule by , we obtain a SC rule
The new sample x* is assigned to class k0 such that .
Note that, if , then and thus gene i plays no role in classifying for class k. Hence SC effectively accomplishes gene selection by shrinkage.
Random forest (RF) [24] is an ensemble of classification trees [42,43], which have been shown to be useful in tumor classification with microarray data [44]. It is designed to improve over a single classification tree. There are two random aspects that help generate multiple classification trees in RF. First, a bootstrap sample is repeatedly drawn from the original training data and then used to build a classification tree. Second, in building a classification tree, rather than using the best splitting variable (i.e. gene here) from all the available variables at each node, it chooses the best from a small random subset of all the variables. Each tree is grown to the maximum and no pruning is pursued. To predict the class for a new sample, the sample is applied to each tree and each tree votes by giving its prediction, then the majority vote is taken as the final prediction for the sample.
Extending a binary classifier to multiclass classification
Here we describe how a multi-class (K > 2) classification problem can be handled by a binary classifier, such as PLS, PPLS and LASSO. It is achieved by formulating a multi-class classification problem as multiple two-class classification problems. We consider two most popular approaches: one is to compare each class against all the others, and the other is to compare all possible pairs of classes. Applications of these two approaches can be found, among others, in [45-50]. In particular, some have considered the first approach for PLS [50].
The one-against-others approach is to reduce a K-class classification task to K two-class classification problems. Formally, a new response is defined in the kth binary problem as:
for k = 1, ..., K. Then we build K binary classifiers. To predict a new sample with gene expression profile x*, we apply x* to each binary classifier and yield . Finally, the class of the new sample C(x*) is predicted as
That is, the new sample is classified into the class maximizing .
The pair-wise approach reduces a K-class classification to K(K - 1)/2 two-class classification problems [45]. Specifically, for each of all possible pairs of classes, solve each of the two-class problems and then, for a new sample, combine all the pairwise decisions to form a K-class decision. Suppose that the new binary response in a pairwise comparison with classes k1 and k2 (with 1 ≤ k1 <k2 ≤ K) is defined as
We build a binary classifier with response using only samples belonging to class k1 or k2, and denote the fitted response value for a new sample with expression profile x* as . As described earlier, we classify the new sample into class k1 or k2 according to the sign of . After this is done for any 1 ≤ k1 <k2 ≤ K, the final decision is to assign the new sample to the class that wins the most pairwise comparisons. In the case when there are multiple winning classes, we randomly pick one of the winning classes to be the final winning class. Comparing to the one-against-others approach, the pair-wise approach is computationally more expensive if K ≥ 4.
Gene ranking
To explore the effect of the number of genes a model starts with on the classification performance, we have a preliminary gene ranking using a usual F-statistic. This univariate ranking is used throughout, and obviously is by no means to be optimal. For the purpose of the presentation in this section, we only need to consider a given gene. Suppose xik is the gene expression level of the gene in sample i that is in class k, i = 1, ..., nk, and k = 1, ..., K, where nk is the total number of samples in class k and K is the total number of classes. Let be the mean expression level of class-k samples, be the overall mean (across all the samples) and be the total number of samples. We can construct an F-statistic as the ratio of the mean sums of squares for between-class and within-class variations:
We can rank all the genes based on their corresponding F-statistics: a gene with a larger F-statistic indicates a stronger relationship between its expression levels and the class membership in the samples, and therefore has a higher rank as a potential predictor of the class. We started with various models by including different numbers of top ranked genes. We considered models starting from the top 50, 100, 200, 400, 800, 1600, 3200, 6400, 9200, 12800, 16000, 19200, and all (22283 and 22277 for the two datasets respectively) genes respectively.
It is an incorrect practice in microarray experiments to first select genes using all the samples and then perform cross-validation using the selected genes, which gives downward biased prediction error estimates [51,52]. Hence, it is essential to perform cross-validation on the entire model building process, including gene selection. In our study, we did honest cross-validation. In particular, we cross-validated gene selection (and other aspects of model building, such as parameter selection and estimation). Specifically, in LOOCV, we remove each sample from the data in turn (which is then treated as the test sample), carry out gene selection using F-statistic based on the remaining samples, build a classifier with the selected genes using the remaining samples, and then test the classifier on the left-out sample.
Data preprocessing
To facilitate the application of penalized regression (i.e. PPLS and LASSO) so that their regression coefficients are in the same unit and thus can be penalized using a global penalty parameter, the expression levels of each gene were scaled to have sample variance 1.
Evaluations
In addition to PLS/PPLS, we will consider the shrunken centroids (SC) method, the LASSO, and the random forest (RF). SC, LASSO and RF have been implemented in R [53], and are easy to use; we applied their R functions using default parameter settings. SC and RF are directly applicable to multiclass classification while LASSO, as PLS/PPLS, is itself a binary classifier. For multi-class classification with PLS and LASSO, we used the same approaches as described for PPLS.
We use the leave-one-out cross-validation (LOOCV) to estimate the prediction error for each of the methods. Within this first-level LOOCV, a second-level LOOCV is used to select tuning parameters for each method to minimize cross validation errors. Specifically, in PLS, the smallest number of PLS components is selected among the PLS models that give the minimum LOOCV error. In PPLS, among the models with minimum LOOCV error, we first pick the ones with the smallest number of PLS components, then pick the one with the largest shrinkage parameter. In the SC method, the largest shrinkage is selected among the models that minimize the LOOCV error. The number of candidate threshold values and the number of cross validation folds are both set to be default (i.e. 30 and the smallest class size respectively). In LASSO, the maximum fraction parameter of the models that minimize LOOCV error is selected while the number of the candidate fraction values is set to be 51 (equally spaced from 0 to 1) and the number of cross validation folds is set to be the total sample size. In RF, every parameter is set to be default. For example, the number of trees to grow is set to 500, and the number of candidate splitting variables considered at each split is set to by default, where p is the total number of variables (i.e. genes).
Due to the small sample size (about 10 in each class) in each dataset, it is quite challenging to estimate the prediction error well. Although it is straightforward to apply LOOCV or other cross-validation methods, their performance may not be optimal. After submitting this work, we became aware of the recent work by Fu et al [54], where a better method than LOOCV was proposed specifically for microarray data. This new method aims to reduce the variability of LOOCV. We reason that with the use of this new method, the main conclusions drawn in this work would not change.
Authors' contributions
XH downloaded the PGA data, did the analysis and simulations. WP conceived of and directed the study. SG cleaned and managed the Minnesota data. XH, YC, SJP, LWM and JH conducted the Minnesota study and generated the data. XH, WP and JH drafted the manuscript.
Supplementary Material
Additional File 1
PLS plots with the first two PLS components. PLS plot for the Minnesota data comparing is-chemic vs idiopathic groups.
Click here for file
Additional File 2
PLS plots with the first two PLS components. PLS plot for the PGA data comparing ischemic vs idiopathic groups.
Click here for file
Additional File 3
PLS plots with the first two PLS components. PLS plots for LOOCV for the Minnesota data comparing ischemic vs idiopathic groups.
Click here for file
Additional File 4
PLS plots with the first two PLS components. PLS plots for LOOCV for the PGA data comparing ischemic vs idiopathic groups.
Click here for file
Additional File 5
Simulation results with various simulation set-ups.
Click here for file
Acknowledgements
XH and WP were supported by NIH grant HL65462. JH was supported by an AHA grant, the Lillehei Heart Institute and the Minnesota Medical Foundation. The authors are grateful to two reviewers for helpful and constructive comments; in particular, section "Other numerical results" was added based on the reviewers' comments.
==== Refs
Levy D Larson MG Vasan RS Kannel WB Ho KK The progression from hypertension to congestive heart failure JAMA 1996 275 1557 1562 8622246 10.1001/jama.275.20.1557
Lloyd-Jones DM The risk of congestive heart failure: sobering lessons from the Framingham Heart Study Curr Cardiol Rep 2001 3 184 190 11305971
Nicol RL Frey N Olson EN From the sarcomere to the nucleus: role of genetics and signaling in structural heart disease Annu Rev Genomics Hum Genet 2000 1 179 223 11701629 10.1146/annurev.genom.1.1.179
Felker GM Thompson RE Hare JM Hruban RH Clemetson DE Howard DL Baughman KL Kasper EK Underlying causes and long-term survival in patients with initially unexplained cardiomyopathy New Engl J Med 2000 342 1077 1084 10760308 10.1056/NEJM200004133421502
Dries DL Sweitzer NK Drazner MH Stevenson LW Gersh BJ Prognostic impact of diabetes mellitus in patients with heart failure according to the etiology of left ventricular systolic dysfunction J Am Coll Cardiol 2001 38 421 428 11499733 10.1016/S0735-1097(01)01408-5
Kittleson M Hurwitz S Shah MR Nohria A Lewis E Givertz M Fang J Jarcho J Mudge G Stevenson LW Development of circulatory-renal limitations to angiotensin-converting enzyme inhibitors identifies patients with severe heart failure and early mortality J Am Coll Cardiol 2003 41 2029 2035 12798577 10.1016/S0735-1097(03)00417-0
Felker GM Benza RL Chandler AB Leimberger JD Cuffe MS Califf RM Gheorghiade M O'Connor CM Heart failure etiology and response to milrinone in decompensated heart failure: results from the OPTIME-CHF study J Am Coll Cardiol 2003 41 997 1003 12651048 10.1016/S0735-1097(02)02968-6
Doval HC Nul DR Grancelli HO Perrone SV Bortman GR Curiel R Randomised trial of low-dose amiodarone in severe congestive heart failure Lancet 1994 344 493 498 7914611 10.1016/S0140-6736(94)91895-3
Singh SN Fletcher RD Fisher SG Singh BN Lewis HD Deedwania PC Massie BM Colling C Lazzeri D Amiodarone in patients with congestive heart failure and asymptomatic ventricular arrhythmia New Engl J Med 1995 333 77 82 7539890 10.1056/NEJM199507133330201
Golub TR Slonim DK Tamayo P Huard C Gaasenbeek M Mesirov JP Coller H Loh ML Downing JR Caligiuri MA Molecular classification of cancer: Class discovery and class prediction by gene expression monitoring Science 1999 285 531 537 10447482 10.1126/science.286.5439.531
Hedenfalk I Duggan D Gene-expression profiles in hereditary breast cancer New England Journal of Medicine 2001 344 539 548 11207349 10.1056/NEJM200102223440801
Tibshirani R Hastie R Narasimhan B Chu G Diagnosis of multiple cancer types by shrunken centroids of gene expression PNAS 2002 99 6567 6572 12011421 10.1073/pnas.082099299
Aronow BJ Toyokawa T Canning A Haghighi K Delling U Kranias E Molkentin JD Dorn GW Divergent transcriptional responses to independent genetic causes of cardiac hypertrophy Physiol Genomics 2001 6 19 28 11395543
Tan FL Moravec CS Li J Apperson-Hansen C McCarthy PM Young JB Bond M The gene expression fingerprint of human heart failure Proc Natl Acad Sci 2002 99 11387 11392 12177426 10.1073/pnas.162370099
Hwang JJ Allen PD Tseng GC Lam CW Fananapazir L Dzau VJ Liew CC Microarray gene expression profiles in dilated and hypertrophic cardiomyopathic end-stage heart failure Physiol Genomics 2002 10 31 44 12118103
Kittleson M Ye SQ Irizarry RA Minhas KM Edness G Conte JV Parmigiani G Miller LW Chen Y Hall JL Garcia JGN Hare JM Identification of a gene expression profile that differentiates ischemic and nonischemic cardiomyopathy Circulation 2004 110 3444 51 15557369 10.1161/01.CIR.0000148178.19465.11
West M Bayesian Factor Regression Models in the "Large p, Small n" Paradigm Bayesian Statistics 2003 7 723 732
Dudoit S Fridlyand J Speed T Comparison of discrimination methods for the classification of tumors using gene expression data J Am Stat Assoc 2002 97 77 87 10.1198/016214502753479248
Huang X Pan W Linear regression and two-class classification with gene expression data Bioinformatics 2003 19 2072 2078 14594712 10.1093/bioinformatics/btg283
Wu BL Abbott T Fishman D McMurray W Mor G Stone K Ward D Williams K Zhao HY Comparison of statistical methods for classification of ovarian cancer using mass spectrometry data Bioinformatics 2003 19 1636 1643 12967959 10.1093/bioinformatics/btg210
Wold S Ruhe A Wold H Dunn WJ III The collinearity problem in linear regression. The partial least squares (PLS) approach to generalized inverses SIAM J of Scientific and Statistical Computing 1984 5 735 742 10.1137/0905052
Segal MR Dahlquist KD Conklin BR Regression approaches for microarray data analysis J Comp Biol 2003 10 961 980 14980020 10.1089/106652703322756177
Tibshirani R Regression shrinkage and selection via the lasso Journal Royal Statistical Society, Series B 1996 58 267 288
Breiman L Random forests Machine Learning 2001 45 5 32 10.1023/A:1010933404324
Hall JL Grindle S Han X Fermin D Park S Chen Y Bache RJ Mariash A Guan Z Ormaza S Thompson J Graziano J de Sam Lazaro SE Pan S Simari RD Miller LW Genomic profiling of the human heart before and after mechanical support with a ventricular assist device reveals alterations in vascular signaling networks Physiological Genomics 2004 17 283 291 14872006 10.1152/physiolgenomics.00004.2004
Hastie T Tibshirani R Friedman J The Elements of Statistical Learning. Data mining, Inference, and Prediction 2001 Springer
Titterington DM Murray GD Murray LS Spiegelhalter DJ Skene AM Habbema JDF Gelpke GJ Comparison of discrimination techniques applied to a complex data set of head injured patients (with discussion) Journal Royal Statistical Society, Series A 1981 144 145 175
Rifkin R Mukherjee S Tamayo P Ramaswamy S Yeang C-H Angelo M Reich M Poggio T Lander ES Golub TR Mesirov JP An analytical method for multi-class molecular cancer classification SIAM Review 2003 45 706 723 10.1137/S0036144502411986
Simon R When is a genomic classifier ready for prime time Nature Clinical Practice – Oncology 2004 1 4 5 15225835 10.1038/ncponc0006
Huang X Pan W Han X Chen Y Miller LW Hall J Borrowing information from relevant microarray studies for sample classification using weighted partial least squares Computational Biology and Chemistry 2005 29 204 211 15979040 10.1016/j.compbiolchem.2005.04.002
Frank IE Friedman JH A statistical view of some chemometrics regression tools (with discussion) Technometrics 1993 35 109 135
Huang X Pan W Park S Han X Miller LW Hall J Modeling the relationship between LVAD support time and gene expression changes in the human heart by penalized partial least squares Bioinformatics 2004 20 888 894 14751963 10.1093/bioinformatics/btg499
Donoho DL Johnstone IM Ideal spatial adaptation by wavelet shrinkage Biometrika 1994 81 425 455
Donoho DL De-Noising by Soft-Thresholding IEEE Transaction on Information Theory 1995 41 613 627 10.1109/18.382009
Hawkins DM Wolfinger RD Liu L Young SS Exploring blood spectra for signs of ovarian cancer Chance 2003 16 19 23
Nguyen DV Rocke DM Tumor classification by partial least squares using microarray gene expression data Bioinformatics 2002 18 39 50 11836210 10.1093/bioinformatics/18.1.39
Ghosh D Singular value decomposition regression models for classification of tumors from microarray experiments Pacific Sympo Biocomput 2002 18 29
Ghosh D Penalized discriminant methods for the classification of tumors from gene expression data Biometrics 2003 59 992 1000 14969478 10.1111/j.0006-341X.2003.00114.x
Ding B Gentleman R Classification using generalized partial least squares Technical Report 5, Bioconductor Project Working Papers 2004
Fort G Lambert-Lacroix S Classification using partial least squares with penalized logistic regression Bioinformatics 2005 21 1104 1111 15531609 10.1093/bioinformatics/bti114
McLachlan GJ Discriminant Analysis and Statistical Pattern Recognition 1992 New York: Wiley
Breiman L Friedman JH Olshen RA Stone CJ Classification and Regression Trees 1984 Belmont: Wadsworth
Zhang H Singer B Recursive Partitioning in the Health Sciences 1999 Springer-Verlag: New York
Zhang H Yu C-Y Singer B Xiong M Recursive partitioning for tumor classification with gene expression microarray data PNAS 2001 98 6730 6735 11381113 10.1073/pnas.111153698
Friedman J Another approach to polychotomous classification Technical report, Stanford University 1996
Hastie T Tibshirani R Classification by pairwise coupling Annals of Statistics 1998 26 451 471 10.1214/aos/1028144844
Allwein E Schapire R Singer Y Reducing multiclass to binary: A unifying approach for margin classifiers Journal of Machine Learning Research 2000 1 113 141 10.1162/15324430152733133
Ramaswamy S Tamayo P Rifkin R Mukherjee S Yeang CH Angelo M Ladd C Reich M Latulippe E Mesirov JP Poggio T Gerald W Loda M Lander ES Golub TR Multiclass cancer diagnosis using tumor gene expression signatures PNAS 2001 98 15149 15154 11742071 10.1073/pnas.211566398
Dettling M Buhlmann P Boosting for tumor classification with gene expression data Bioinformatics 2003 19 1063 1069 10.1093/bioinformatics/btf867
Tan Y Shi L Tong W Hwang GTG Wang C Multi-class tumor classification by discriminant partial least squares using microarray gene expression data and assessment of classification models Computational Biology and Chemistry 2004 28 235 243 15261154 10.1016/j.compbiolchem.2004.05.002
Ambroise C McLachlan GJ Selection bias in gene extraction on the basis of microarray gene-expression data PNAS 2002 99 6562 6566 11983868 10.1073/pnas.102102699
Simon R Radmacher MD Dobbin K McShane LM Pitfalls in the use of DNA microarray data for diagnostic and prognostic classification J Natl Cancer Inst 2003 95 14 18 12509396
Ihaka R Gentleman R R: A language for data analysis and graphics Journal of Computational and Graphical Statistics 1996 5 299 314
Fu WJ Carroll RJ Wang S Estimating misclassification error with small samples via bootstrap cross-validation Bioinformatics 2005 21 1979 1986 15691862 10.1093/bioinformatics/bti294
|
16120216
|
PMC1224853
|
CC BY
|
2021-01-04 16:27:27
|
no
|
BMC Bioinformatics. 2005 Aug 24; 6:205
|
utf-8
|
BMC Bioinformatics
| 2,005 |
10.1186/1471-2105-6-205
|
oa_comm
|
==== Front
BMC BioinformaticsBMC Bioinformatics1471-2105BioMed Central London 1471-2105-6-2081612237910.1186/1471-2105-6-208SoftwareAn edit script for taxonomic classifications Page Roderic DM [email protected] Gabriel [email protected] DEEB, IBLS, University of Glasgow, Glasgow G12 8QQ, UK2 Department of Software, Technical University of Catalonia, E-08034 Barcelona, Spain2005 25 8 2005 6 208 208 21 6 2005 25 8 2005 Copyright © 2005 Page and Valiente; licensee BioMed Central Ltd.2005Page and Valiente; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
The NCBI taxonomy provides one of the most powerful ways to navigate sequence data bases but currently users are forced to formulate queries according to a single taxonomic classification. Given that there is not universal agreement on the classification of organisms, providing a single classification places constraints on the questions biologists can ask. However, maintaining multiple classifications is burdensome in the face of a constantly growing NCBI classification.
Results
In this paper, we present a solution to the problem of generating modifications of the NCBI taxonomy, based on the computation of an edit script that summarises the differences between two classification trees. Our algorithms find the shortest possible edit script based on the identification of all shared subtrees, and only take time quasi linear in the size of the trees because classification trees have unique node labels.
Conclusion
These algorithms have been recently implemented, and the software is freely available for download from .
==== Body
Background
The NCBI Taxonomy [1] provides one of the most powerful ways to navigate the National Center for Biotechnology Information (NCBI) sequence data bases. Every sequence in GenBank is associated with a taxon (which, however, may be unidentified), and each taxon has a unique place in the NCBI taxonomy. Hence, not only can the user retrieve sequences for a given species (such as Homo sapiens), but also for a group of species, such as mammals (Mammalia) or animals (Animalia).
The NCBI provides a single classification, assembled from a variety of sources including published literature, a panel of expert advisors, and the taxonomy provided by users when they submit new sequences. Given that there is not universal agreement on the classification of organisms, providing a single classification places constraints on the questions biologists can ask.
To give a concrete example, Figure 1 shows a simplified classification of animals, based on the current NCBI taxonomy. In this classification, the Bilateria are split into three groups (Acoelomata, Pseudocoelomata, and Coelomata) based on the nature of the internal body cavity (coelom). The Coelomata are themselves split into two groups, the Protostomia and the Deuterostomia, characterised by the fate of the blastopore during development (in the Protostomia this becomes the mouth, in the Deuterostomia it becomes the anus).
Figure 1 Traditional view of animals. A "traditional" view of animal relationships, based on the NCBI classification.
An alternative view of animal classification is shown in Figure 2. The three-fold division based on body cavity disappears, leaving the fundamental split being between the Protostomia and the Deuterostomia. The Protostomia are divided into the Lophotrochozoa and the Ecdysozoa, the latter comprising arthropods, nematodes, and other moulting animals [2]. This classification has implications for comparative genomics. The best known animal genomes are Homo sapiens (human), Drosophila melanogaster (fly), and Caenorhabditis elegans (nematode). Under the classical classification (Fig. 1), the coelomates human and Drosophila are more closely related to either other than either is to the aceolomate C. elegans, suggesting it would be most productive to compare our genome with that of Drosophila, rather than the more distant nematode. However, in the alternative classification (Fig. 2) Drosphila and C. elegans are more closely related to each other than either is to humans, and we have no (phylogenetic) reason for choosing one over the other as a point of reference for interpreting the human genome. There is considerable debate about the merits of the two classifications [3-5]. However, because the NCBI provides only one classification users cannot, for example, easily query GenBank for all ecdysozoan sequences – this taxon simply does not exist in the NCBI database. Instead, users are forced to construct Boolean queries such as (Arthropoda AND Nematoda). While in this simplified example this is not a great hardship, as the trees get larger and the differences more profound, it becomes harder to pose a query that captures the taxa required.
Figure 2 An alternative view of animals. A alternative tree of animals reflecting the "new animal classification".
One solution is simply to download the NCBI taxonomy, edit it to reflect the desired alternative classification, then use that to obtain sequences from taxa such as Ecdysozoa. It is reasonably straightforward to store a tree in a relational database an query it using SQL [6]. However, the NCBI taxonomy is continually growing as new organisms are sequenced. Hence, a locally edited classification will quickly become obsolete. Having to download a fresh copy and then manually edit it would quickly become tedious.
Implementation
Taxonomic classifications
Although ideally classifications mirror phylogenetic relationships, it is important to distinguish between classifications and phylogenies. A taxonomic classification can be modelled as a rooted, labelled, unordered tree. Unlike classifications, internal nodes of phylogenetic trees need not be labeled, although the internal nodes of a phylogeny may be decorated with measures of support (such as bootstrap values or Bayesian posterior probabilities).
Subtree isomorphism
Our approach is to first find subtree isomorphisms between the two trees, T1 and T2. A subtree is a connected subgraph of a tree. We distinguish between top-down and bottom-up subtree isomorphism. A top-down node matching the parent of each node in the matching is itself in the matching (excluding the root which has no parent). In a bottom-up matching, all the children of a node in the matching are also in the matching (Fig. 3).
Figure 3 Connected subgraph and top-down and bottom-up subtrees. In the top-down subtree the parent of any node in the subtree is itself in the subtree. In the bottom-up matching, the children of any node in the matching are also in the matching. Modified from [10].
The algorithm first finds all subtrees, including bottom-up and top-down subtrees, that are common to T1 and T2. We find all kinds of subtree because, by themselves the subtrees found by each method can be small (Fig. 4).
Figure 4 Subtree isomorphisms. The top-down and bottom-up subtree isomorphisms between the animal classifications shown in Figs. 1 and 2. (ignoring the trivial bottom-up subtrees that comprise a single leaf).
Script
Having identified common subtrees, we then list the operations needed to transform T1 into T2. The first step is to delete nodes in T1 that are not in any of the shared subtrees. The deletion of a node entails deletion of all the edges incident with the deleted node. We then add nodes found only in T2, and the corresponding edges. The size of the script depends on the size of the shared subtrees, hence it is desirable to find the largest such subtrees.
Complexity
In general, computation of the least number of operations needed to transform T1 into T2 is an NP-hard problem [7], even for binary trees with a label alphabet of size two, as long as node and edge deletions, insertions, and label substitutions are allowed. However, in the case of trees with unique node labels, node label substitutions are forbidden because they may generate trees with non-unique node labels [8], and the least number of operations or edit distance becomes a function of the size of shared subtrees [9]. By identifying the largest common subtrees, we obtain the shortest possible edit script.
Computing an edit script
Taxonomic classifications are modelled as trees with unique node labels, and this fact makes it easier to deal with trees in terms of their sets of node labels and node label pairs, as done for graphs with unique node labels in [8].
Definition 1 Let T = (V,E) be a tree. The label representation of T, denoted by R(T), is given by R(T) = (L,C), where L = {ℓ(v) | v ∈ V} and C = {(ℓ(v),ℓ(w)) | (v,w) ∈ E}.
Thus, the label representation R(T) of a tree T defines the equivalence class of all those trees that are isomorphic to T. The use of label representations simplifies the notation, because isomorphic trees have exactly the same label representation.
The edit operations of node and edge deletion and insertion, allow one to transform any given tree into any other tree. Label substitutions are forbidden because they may generate trees with non-unique node labels [8].
Definition 2 Let T1 = (V1,E1) and T2 = (V2,E2) be trees, let R(T1) = (L1, C1), and let R(T2) = (L2,C2). Let also C = L1 ∪ L2 ∪ {λ}.
A node edit operation between T1 and T2 is a pair (a, b) ∈ C × C with a ≠ λ or b ≠ λ. A node edit operation of the form (a, λ) establishes deletion of the node v ∈ V1 with ℓ(v) = a together with the edge (parent(v), v), if v is not the root of T1, and deletion of edge (v,x) for each child x of v in T1. A node edit operation of the form (λ,b) establishes insertion of the node w ∈ V2 with ℓ(w) = b.
An edge edit operation between T1 and T2 is a triple (a, b, c) ∈ C × C × C with b ≠ λ and a ≠ λ or c ≠ λ. An edge edit operation of the form (a, b, λ) establishes deletion of the edge (v, x) ∈ E1 with ℓ(v) = a and ℓ(x) = b, and an edge edit operation of the form (λ, b, c) establishes insertion of the edge (w,y) ∈ E2 with ℓ(w) = b and ℓ(y) = c.
An edit operation is either a node edit operation or an edge edit operation.
An edit script between two trees is just a set of edit operations that, if applied in the right order (essentially, inserting an edge only after having inserted the nodes incident with the inserted edge), allow one to transform one tree into the other.
Definition 3 An edit script between two trees T1 = (V1,E1) and T2 = (V2,E2) is a set of edit operations that transform R(T1) into R(T2).
Given R(T1) = (L1, C1) and R(T2) = (L2, C2), an edit script between T1 and T2 can be easily obtained by sorting the label sets and computing set differences, as follows:
• Delete all nodes with labels in L1 \ L2
• Insert all nodes with labels in L2 \ L1
• Delete all edges with labels in C1 \ C2
• Insert all edges with labels in C2 \ C1
However, such a procedure does not, in general, lead to the shortest possible edit script, because some of the edge deletion operations may be redundant, given that deletion of a node entails deletion of all the edges incident with the deleted node. While any edit script would suffice to transform one tree into the other, the shortest edit script leads to a faster computation of the edited tree, given the script and the original tree.
The following, alternative procedure is based on the set of common node labels between the two trees, which can be easily obtained as the intersection of the sets of node labels in the label representation of the trees, that is, C = L1 ∩ L2 = {ℓ(v) | v ∈ V1} ∩ {ℓ(w) | w ∈ V2}. The procedure can be sketched as follows:
• Delete all nodes v ∈ V1 with ℓ(v) ∉ C.
• Insert all nodes w ∈ V2 with ℓ(w) ∉ C.
• Delete all edges (v,x) ∈ E1 with ℓ(v), ℓ(x) ∈ C and such that the node w ∈ V2 with ℓ(v) = ℓ(w) is not the parent in T2 of the node y ∈ V2 with ℓ(x) = ℓ(y).
• Insert all edges (w,y) ∈ E2 with ℓ(w), ℓ(y) ∈ C and such that the node v ∈ V1 with ℓ(v) = ℓ(w) is not the parent in T1 of the node x ∈ V1 with ℓ(x) = ℓ(y).
• Insert all edges (w, y) ∈ E2 such that ℓ(w) ∉ C or ℓ(y) ∉ C.
A detailed description of the algorithm is given in Fig. 5. Correctness of the edit script algorithm is easy to establish.
Figure 5 Algorithm for computing edit script. Let C be a set of common node labels of T1 and T2. A function call of the form edit script (T1, T2, C) returns a set E of elementary edit operations that transform T1 into T2.
Theorem 1 Let T1 and T2 be trees, let C ⊆ ℓ(V1) ⋂ ℓ(V2), let E = edit script (T1, T2, C), and let be the result of applying the set of edit operations in E to T1. Then, and T2 are isomorphic.
Proof It has to be shown that . Let R(T1) = (L1, C1) and R(T2) = (L2, C2). The edit script establishes the deletion of all nodes with labels in L1\ C and the insertion of all nodes with labels in L2 \ C. Thus, = L1 \ (L1 \ C) ∪ (L2 \ C) = C ∪ (L2 \ C) = L2.
The edit script also establishes the deletion of all edges with source and target labels in (C1 ∩ C × C) \ C2, the insertion of all edges with source and target labels in (C2 ∩ C × C) \ C1, and the insertion of all edges with source or target label in L2 \ C, that is, of all edges in C2 \ (C2 ∩ C × C). Furthermore, the deletion of all nodes with labels in L1 \ C entails the deletion of all edges with source or target label in L1 \ C, that is, of all edges in C1 \ (C1 ∩ C × C). (See Fig. 6.)
Figure 6 Illustration for the proof of Theorem 1. Given the label representation R(T1) = (L1, C1) and R(T2) = (L2, C2) of two trees, and a set of common node labels C ⊂ L1 ∩ L2, T1 can be transformed into T2 by deleting all nodes with labels in L1 \ (C, which implies deletion of all edges with source and target node labels in C1 \ (C1 ∩ C × C); inserting all nodes with labels in L2 \ C, deleting all edges with source and target node labels in (C1 ∩ C × C) \ C2; inserting all edges with source and target node labels in (C2 ∩ C × C) \ C1; and inserting all edges with source and target node labels in C2 \ (C2 ∩ C × C).
Now, C1 = (C1 \ C2) ∪ (C1 ∩ C2) = ((C1 ∩ C × C) \ C2) ∪ ((C1 \ (C1 ∩ C × C)) \ C2) ∪ (C1 ∩ C2). In a similar vein, C2 = ((C2 ∩ C × C) \ C1) ∪ ((C2 \ (C2 ∩ C × C)) \ C1) ∪ (C1 ∩ C2).
Thus,
= C1 \ ((C1 ∩ C × C) \ C2) \ ((C1 \ (C1 ∩ C × C)) \ C2) ∪ ((C2 ∩ C × C) \ C1) ∪ ((C2 \ (C2 ∩ C × C)) \ C1) = (C1 ∩ C2) ∪ ((C2 ∩ C × C) \ C1) ∪ ((C2 \ (C2 ∩ C × C)) \ C1) = C2 and therefore, = (L2, C2) = R(T2), that is, and T2 are isomorphic.
The edit script algorithm can be implemented to take time quasi linear in the size of the trees, by using any efficient dictionary data structure to represent the set of common node labels. The same dictionary data structure allows one to compute the set of common node labels within the same time bound and thus, the whole procedure can be implemented to take time quasi linear in the size of the trees. In particular, our C++ implementation uses the STL associative container set<string> as representation of the set of shared node labels.
Results
Here we suggest a solution based on the notion of an "edit script" that summarises the differences between two trees. Given two trees, T1 and T2, a script lists the operations required to convert T1 into T2. The script could be constructed manually, but it would be more efficient to generate it automatically. Hence, we imagine the following scenario. A user downloads the NCBI taxonomy tree (or that subtree relevant to their interests), then edits the tree to reflect their preferred classification. Using the algorithm we describe below, the user then computes the edit script that transforms the NCBI tree into their classification. When a new NCBI tree appears on the NCBI ftp site, the user downloads that tree and applies to edit script to regenerate their classification. In this way, the user need only edit the NCBI tree once.
As an example, given the two trees in Figs. 1 and 2, the edit script for these trees is:
delete node Pseudocoelemata
delete node Coelomata
delete node Protostomia
delete node Acoelomata
insert node Ecdysozoa
insert node Lophotrochozoa
insert node Protostomia
insert edge Bilateria -> Deuterostomia
insert edge Bilateria -> Protostomia
insert edge Ecdysozoa -> Nematoda
insert edge Ecdysozoa -> Arthropoda
insert edge Lophotrochozoa -> Annelida
insert edge Lophotrochozoa -> Brachiopoda
insert edge Lophotrochozoa -> Bryozoa
insert edge Lophotrochozoa -> Mollusca
insert edge Lophotrochozoa -> Nemertea
insert edge Lophotrochozoa -> Platyhelminthes
insert edge Protostomia -> Lophotrochozoa
insert edge Protostomia -> Ecdysozoa
Applying the script to the NCBI tree (Fig. 1) yields the tree shown in Fig. 7, which is identical to the tree shown in Fig. 2.
Figure 7 Result of applying the edit script. The result of applying the edit script to the tree in Fig. 1. This tree is the same as that shown in Fig. 2. Nodes which have been inserted into the tree are filled with light grey. A dashed line represents an edge that has been added to the original tree.
Discussion
The size of the edit script will be a function of the size of the input trees, and the degree to which they differ. At the time of writing, there are 83,802 metazoan taxa in GenBank. Given that the disagreement between the Coelomata and Ecdysozoa hypotheses concerns the deep level relationships, we can simplify the task by reducing the subtrees about which there is little or no disagreement to single nodes. For example, the 36,746 arthropod taxa can be represented by a single node. Hence, the tree shown in Fig. 1 is greatly simplified, compared to the complete NCBI tree.
One issue we don't directly address here is using the tree that results from applying the edit script to query GenBank. There are at least two approaches to doing this. The first is to store the tree in a local database and use a method such as visitation numbers [6] to generate queries involving higher taxa (such as listing all sequences from the Ecdysozoa).
Another approach would be to use the tree to rewrite queries in terms of the original GenBank taxonomy. For example, in our rather simplified example in Fig. 2, we could use the tree to automatically rewrite the query term "Ecdysozoa" as the sum of its children (Arthropoda and Nematoda) as both trees (Fig. 1 and Fig. 2) agree on the composition of these two taxa. One advantage of this approach is that we can continue to use tools such as BLAST, but in the context of a different taxonomic classification.
Conclusion
We present a solution to the problem of generating modifications of the NCBI taxonomy, based on the computation of an edit script that summarises the differences between two classification trees. Our algorithms find the shortest possible edit script based on the identification of all shared subtrees, and only take time quasi linear in the size of the trees because classification trees have unique node labels. We have implemented the edit function in a C++ program that makes use of the Graph Template Library (GTL) available from . The code has been compiled and tested with the GNU gcc compiler on Mac OS X and Linux machines, and is available from . The software comprises two programs, forest and script. The program forest takes two trees in GML format (the original tree and the edited tree) and computes an edit script. Given this script and the original tree, script generates the edited tree.
Availability and requirements
• Project name: Forest
• Project home page:
• Operating system(s): Unix/Linux, tested on Mac OS X and Red Hat 8.0
• Programming language: e.g. C++
• Other requirements: Graph Template Library (GTL) ()
• License: GNU GPL
• Any restrictions to use by non-academics: Forest depends on GTL, which can be downloaded free of charge for non-commercial use. Commercial use of GTL requires a licence from BRAINSYS – Informatiksysteme GmbH ()
Authors' contributions
RDMP posed the problem, and GV developed the algorithm. RDMP and GV jointly developed the software and wrote the manuscript.
Acknowledgements
This work was funded by BBSRC grant BB/C004310/1 to RDMP, by the Spanish CICYT, project GRAMMARS (TIN2004-07925-C03-01), and by the Japan Society for the Promotion of Science through Long-term Invitation Fellowship L05511 for visiting JAIST (Japan Advanced Institute of Science and Technology).
==== Refs
NCBI Taxonomy
Aguinaldo A Turbeville J Linford L Rivera M Garey J Raff R Lake J Evidence for a clade of nematodes, arthropods and other moulting animals Nature 1997 387 489 93 9168109 10.1038/387489a0
Wolf Y Rogozin I Koonin E Coelomata and not Ecdysozoa: evidence from genome-wide phylogenetic analysis Genome Res 2004 14 29 36 14707168 10.1101/gr.1347404
Philip GK Creevey CJ Mclnerney JO The Opisthokonta and the Ecdysozoa may not be Clades: stronger support for the grouping of plant and animal than for animal and fungi and stronger support for the Coelomata than Ecdysozoa Mol Biol Evol 2005 22 1175 1184 15703245 10.1093/molbev/msi102
Philippe H Lartillot N Brinkmann H Multigene analyses of bilaterian animals corroborate the monophyly of Ecdysozoa, Lophotrochozoa and Protostomia Mol Biol Evol 2005 22 1246 1253 15703236 10.1093/molbev/msi111
Celko J SQL for Smarties: Advanced SQL Programming 1999 San Francisco: Morgan Kaufmann
Zhang K Statman R Shasha D On the editing distance between unordered labeled trees Inform Process Lett 1992 42 133 139 10.1016/0020-0190(92)90136-J
Dickinson P Bunke H Dadej A Kraetzl M On graphs with unique node labels Proc 4 th IAPR Int Workshop Graph Based Representations in Pattern Recognition 2003 Springer-Verlag 13 23
Bunke H On a relation between graph edit distance and maximum common subgraph Pattern Recogn Lett 1997 18 689 694 10.1016/S0167-8655(97)00060-3
Valiente G Algorithms on Trees and Graphs 2002 Berlin: Springer-Verlag
|
16122379
|
PMC1224854
|
CC BY
|
2021-01-04 16:27:26
|
no
|
BMC Bioinformatics. 2005 Aug 25; 6:208
|
utf-8
|
BMC Bioinformatics
| 2,005 |
10.1186/1471-2105-6-208
|
oa_comm
|
==== Front
BMC BiolBMC Biology1741-7007BioMed Central London 1741-7007-3-191613892810.1186/1741-7007-3-19Research ArticleModulation of extracellular matrix genes reflects the magnitude of physiological adaptation to aerobic exercise training in humans Timmons James A [email protected] Eva [email protected] Helene [email protected] Thomas [email protected] Paul L [email protected] John [email protected] Jonathan [email protected] Carl Johan [email protected] Department of Physiology and Pharmacology, Karolinska Institutet, Stockholm, SE171 77, Sweden2 Centre for Genomics & Bioinformatics, Karolinska Institutet, Stockholm, SE171 77, Sweden3 Laboratory Medicine, Division of Clinical Physiology, Karolinska Institutet, Huddinge, 141 86, Sweden4 Centre for Integrated Systems Biology and Medicine, University Medical School, Nottingham, UK5 Department of Enabling Technologies, AstraZeneca, Alderly Park, UK6 OSI Prosidion Ltd, Oxfordshire, OX4 6LT, UK2005 2 9 2005 3 19 19 22 3 2005 2 9 2005 Copyright © 2005 Timmons et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Regular exercise reduces cardiovascular and metabolic disease partly through improved aerobic fitness. The determinants of exercise-induced gains in aerobic fitness in humans are not known. We have demonstrated that over 500 genes are activated in response to endurance-exercise training, including modulation of muscle extracellular matrix (ECM) genes. Real-time quantitative PCR, which is essential for the characterization of lower abundance genes, was used to examine 15 ECM genes potentially relevant for endurance-exercise adaptation. Twenty-four sedentary male subjects undertook six weeks of high-intensity aerobic cycle training with muscle biopsies being obtained both before and 24 h after training. Subjects were ranked based on improvement in aerobic fitness, and two cohorts were formed (n = 8 per group): the high-responder group (HRG; peak rate of oxygen consumption increased by +0.71 ± 0.1 L min-1; p < 0.0001) while the low-responder group (LRG; peak rate of oxygen consumption did not change, +0.17 ± 0.1 L min-1, ns). ECM genes profiled included the angiopoietin 1 and related genes (angiopoietin 2, tyrosine kinase with immunoglobulin-like and EGF-like domains 1 (TIE1) and 2 (TIE2), vascular endothelial growth factor (VEGF) and related receptors (VEGF receptor 1, VEGF receptor 2 and neuropilin-1), thrombospondin-4, α2-macroglobulin and transforming growth factor β2.
Results
neuropilin-1 (800%; p < 0.001) and VEGF receptor 2 (300%; p < 0.01) transcript abundance increased only in the HRG, whereas levels of VEGF receptor 1 mRNA actually declined in the LRG (p < 0.05). TIE1 and TIE2 mRNA levels were unaltered in the LRG, whereas transcription levels of both genes were increased by 2.5-fold in the HRG (p < 0.01). Levels of thrombospondin-4 (900%; p < 0.001) and α2-macroglobulin (300%, p < 0.05) mRNA increased substantially in the HRG. In contrast, the amount of transforming growth factor β2 transcript increased only in the HRG (330%; p < 0.01), whereas it remained unchanged in the LRG (-80%).
Conclusion
We demonstrate for the first time that aerobic training activates angiopoietin 1 and TIE2 genes in human muscle, but only when aerobic capacity adapts to exercise-training. The fourfold-greater increase in aerobic fitness and markedly differing gene expression profile in the HRG indicates that these ECM genes may be critical for physiological adaptation to exercise in humans. In addition, we show that, without careful demonstration of physiological adaptation, conclusions derived from gene expression profiling of human skeletal muscle following exercise may be of limited value. We propose that future studies should (a) investigate the mechanisms that underlie the apparent link between physiological adaptation and gene expression and (b) use the genes profiled in this paper as candidates for population genetic studies.
==== Body
Background
Regular exercise and high aerobic fitness both reduce the risk of cardiovascular- and metabolic-disease-related death for a multitude of potential reasons [1-5]. It is noteworthy that a very large intersubject variation exists when measuring the physiological adaptation to supervised exercise training [6-9]. While some subjects demonstrate a robust increase in aerobic capacity, others seem not to respond substantially at all [8,10,11]. This variation also applies to the improvement in insulin sensitivity seen after exercise [9]. Such observations may be important for future cardiovascular health, as an inherent lack of 'trainability' associates with increased cardiovascular risk factors [5]. The mRNA abundance for a huge number of genes (>500) have been shown to be increased many hours after exercise training in humans [8,12-18]. However, only very recently have gene expression changes been related to the magnitude of physiological adaptation [8,9]. Teran-Garica and others [9] observed a divergent mRNA response in subjects that increase their insulin sensitivity most following endurance training, whereas we have demonstrated that the expression of insulin-like growth factor related genes were increased with training and more markedly in those subjects that most enhanced their aerobic capacity [8]. Little else is known about the relationship between the extent of gene activation and the magnitude of physiological adaptation to exercise training in humans.
Increased aerobic capacity following a period of intense endurance training reflects both central and peripheral adaptations [19-23]. Activation of angiogenesis is presumably an important component of the response to endurance training [15,20,23-25], indicating that substantial remodeling of skeletal muscle extracellular matrix (ECM) is required. Using gene expression profiling, our knowledge of the many factors that regulate the extracellular environment and facilitate vascular remodeling following exercise has improved. Alterations in vascular endothelial growth factor (VEGF) and related receptor(s) transcript expression occur following acute exercise and endurance training [15,17,26-28], indicating that VEGF may be an important exercise factor. More recently, the Angiopoietin (ANG) signaling pathway has been shown to synergize with VEGF [29,30] and expression of both ANG1 and ANG2 is altered by intense aerobic training in rats [27]. There is currently no information on the physiological regulation of the ANG system in human skeletal muscle following aerobic training. Furthermore, the significance of changes in the abundance of transcripts from genes for various growth factors [15,26,27] has not been examined in the context of changes in aerobic capacity resulting from endurance training.
We therefore set out to establish if greater improvements in systemic cardiovascular adaptation would be associated with changes in muscle gene expression. We did so by examining the expression responses of a number of tissue remodeling genes in subjects that demonstrated either a substantial (high-responder group; HRG) or a modest/negligible (low-responder group; LRG) response to training [8]. We also aimed to establish if the inclusion of low responders in a gene analysis study could yield misleading information on genes that might genuinely contribute to physiological adaptation to exercise in humans.
Results
Physiological parameters
Table 1 presents the baseline physiological characteristics of the HRG and LRG. As can be observed, no differences in baseline demographics or physiological characteristics existed between LRG and HRG subjects prior to the training program [8]. The combined training response was significantly greater in the HRG (p < 0.001). The reduction in submaximal heart rate was more substantial (p < 0.01) in the HRG (-25.9 ± 3.6 beats min-1 (BPM)) vs. the LRG (-10.5 ± 3.5 BPM) during a 10 min fixed-workload, submaximal cycle. The increase in work performed during 15 min cycling after the six weeks of endurance exercise training was 60% greater (p < 0.05) in the HRG (+40.8 ± 4.3 kJ) than in the LRG (+24.9 ± 4.1 kJ). The increase in peak rate of oxygen consumption (peak VO2) was four times greater (p < 0.001) in the HRG (+0.71 ± 0.1 L min-1) than in the LRG (+0.17 ± 0.1 L min-1). Importantly, there was no correlation between baseline physiological characteristics and the magnitude of improvement noted; this observation is consistent with those from previous studies [31].
Growth factor-related genes
Subjects in both groups underwent six weeks of endurance training. Gene expression levels were then measured 24 h after the last training session. Levels of VEGF gene expression were not significantly altered in either group (Fig. 1). Likewise, levels of VEGF receptor 1 (VEGFR1) mRNA did not significantly increase in the HRG and actually declined in the LRG (p < 0.05). However, VEGF receptor 2 (VEGFR2) mRNA expression increased by threefold in the HRG (p < 0.01), whereas it did not significantly change in the LRG. Similarly, mRNA for the VEGFR2 coreceptor Neuropilin-1 (NP-1) increased in the HRG (p < 0.001) but not in the LRG (Fig. 1). Expression of the hypoxia-inducible factor 1α (HIF) gene did not significantly change after endurance training. Regulation by endurance training of ANG-related genes was only observed in the HRG (Fig. 2). Levels of mRNA coding for ANG1, an agonist for the Tyrosine kinase with immunoglobulin-like and EGF-like domains 2 (TIE2) receptor, increased significantly in the HRG (p < 0.05), although ANG2 levels did not change significantly in either group. Likewise, transcription of tyrosine kinase with immunoglobulin-like and EGF-like domains 1 (TIE1; 3.4 ± 1.0-fold increase; p < 0.01) and TIE2 increased in the HRG only (p < 0.01; Fig. 3).
Extracellular matrix growth factor binding genes
Using our microarray dataset [8], we selected ECM factors that (a) demonstrated evidence of being modulated by exercise training and (b) were relevant for ECM remodeling [32-34]. These genes included ones that encoded structural components of blood vessels (collagen type IIIα1) or known regulators of tissue angiogenesis (collagen type XVα1), factors known to influence ECM-derived growth factor activity (α2-macroglobulin (A2M) and thrombospondin-4 (THBS4)), transforming growth factor β2 (TGFB2; a potent regulator of tissue remodeling) and TGFB receptor II (TGFBR2). The level of fetal vascular collagen (i.e., collagen type IIIα1) gene expression was increased 14.7 ± 2.8-fold (p < 0.001) in the HRG and 8.5 ± 4.1 fold in the LRG (p < 0.001). The level of collagen type XVα1 expression increased in both groups: the HRG demonstrated a 7.6 ± 3.6-fold increase (p < 0.001) while the LRG demonstrated a 2.7 ± 0.5-fold increase (p < 0.01) in expression of the gene (Fig. 3). In contrast, while both A2M and THBS4 were significantly upregulated in the HRG (threefold (p < 0.01) and tenfold (p < 0.001) increases, respectively), in the LRG the THBS4 response was more modest (2.5-fold; p < 0.01) and A2M mRNA levels did not significantly change. TGFB2 was significantly upregulated in the HRG (p < 0.01), whereas, if anything, it tended to be downregulated in the LRG (Fig. 3). The level of TGFBR2 expression was unchanged in the LRG, but threefold increased in the HRG (p < 0.001; Fig. 3). Finally, three genes unrelated to ECM biology but previously described as being modulated by exercise [8], interleukin 17D, Rho-GTPase-activating-protein 1 and myristoylated alanine-rich protein kinase C substrate, did not significantly vary in their expression between HRG and LRG (data not shown).
Discussion
Classic alterations in skeletal muscle phenotype following physical training include improved fatigue resistance, enhanced aerobic capacity and greater insulin sensitivity [9,23,25,35,36]. The significance of an individual's ability to adapt to exercise training may ultimately influence multiple risk factors important for long term cardiovascular health [5]. In the present study we demonstrate for the first time that ANG genes are substantially modulated in humans following a six-week aerobic training program. Overall, many ECM-gene transcripts were only modulated in subjects that demonstrated a concurrent improvement in aerobic capacity. Our data suggest that activation of ECM genes may help determine the cardiovascular adaptation to aerobic exercise in humans. The present findings also indicate that, when carrying out expression studies of gene transcripts in humans, prior to any interpretation of muscle gene expression responses, attention must be afforded to the presence of physiological adaptation.
Modulation of genes that regulate extracellular matrix remodeling
We still have an incomplete understanding of the endogenous processes that regulate physiological adaptation to aerobic exercise. Vascular growth factors not only regulate tissue blood vessel density, but also enhances the expression of proteins that regulate oxygen levels in tissue [37]. Hence, it is naïve to think about the role of such growth factors only in terms of regulating tissue capillary density. Altered expression of HIF responsive genes (e.g. VEGF or VEGFR1) typically reflects the posttranslational stabilization of HIF1α protein [38], and consistent with this dogma, HIF mRNA was not significantly upregulated. VEGFR2, considered the major mediator of VEGF-A-related angiogenesis, was significantly upregulated in the HRG only. Furthermore, NP-1, a facilitator of VEGF165 action at VEGFR2, was markedly elevated in the HRG and unchanged in the LRG (Fig. 1). Upregulation of the VEGFR2 coreceptor transcript provides some evidence for greater VEGF activity in the HRG, as enhanced NP-1 gene expression can be mediated by VEGF signaling [39,40]. The significance of the downregulation of VEGFR1 in the LRG is unclear, but plausibly reflects an (unsuccessful) compensation for the general lack of VEGFR2/NP-1 gene activation in the LRG, as studies indicate VEGFR1 may oppose VEGF signaling via VEGFR2 in some situations [41].
The present study represents the first characterization of expression levels of the human ANG gene family in response to endurance-exercise training. Recently, the Terjung laboratory examined the impact of exercise on the angiopoietin system in rodents [27]. Alterations in TIE2, ANG1 and ANG2 transcript expression were profiled 2 h post exercise in various muscle groups taken from Sprague-Dawley rats after 1 to 24 days of intense aerobic running [27]. In rodents, activation of TIE2 and ANG2 was, broadly speaking, rather similar across each muscle tissue, especially when one considers that recruitment patterns would differ between the various muscle groups studied. The ANG1 response, however, differed both across muscle groups and when compared with our human data. In contrast with the upregulation of ANG1 seen in humans in the HRG (Fig. 2), in the rat ANG1 expression was slightly downregulated in oxidative muscle groups, while 'fast' gastrocnemius demonstrated only a modest increase in ANG1 expression [27]. It has recently been established that either pre- [30] or concurrent administration [29] of ANG1 synergizes with VEGF to promote hindlimb angiogenesis. Thus it makes sense that effective aerobic training might result in stable increases in ANG1 expression (as we have found). Therefore, differences between our human data and the rodent study by Lloyd et al [27] perhaps reflect differing muscle sampling times. Future human studies should ideally use multiple time points post exercise to clarify these issues. However, care should be taken to verify that the subjects studied are able to demonstrate a measurable aerobic training response otherwise such results may be unreliable.
It has been hypothesized that changes in the ratio between levels of ANG1 and ANG2 is of physiological importance [27,42] by contributing to the stabilization of the primary endothelial structures [43]. While ANG1's ability to antagonize ANG2 at the TIE2 receptor may be cell-type specific [42,44,45] and has yet to be proven to be important in vivo, it is interesting to note that ANG1 and ANG2 appear to have identical binding affinities for the TIE2 receptor, whereas a threefold molar excess of ANG2 is required to antagonize ANG1 activity [45]. If the gene expression patterns observed in the present study translate to greater Tie2 receptor signaling, then our data supports the idea that ANG1 cooperates with other growth factors in vivo to regulate and promote functional angiogenesis. This hypothesis clearly requires further investigation. ANG1 and TGFB signaling facilitate the maturation of VEGF-stimulated collateral vessel growth in adult tissue [29,30,46]. Notably, TGFB2 and TGFBR2 were substantially upregulated only in the HRG. This observation again suggests that the ECM-related gene response in the HRG may have allowed for greater tissue remodeling, which would have contributed to the fourfold greater increase in aerobic capacity.
To further examine the idea that the pattern of transcript expression in the HRG contributed to the enhanced aerobic adaptation, we profiled a second set of genes chosen from a gene-array study [8]. For example, upregulation of potent growth factors should be accompanied by upregulation of endogenous regulators, so that physiological control could be maintained. A2M is an ECM protein known to bind and regulate growth factor activity [47,48]. A2M was substantially upregulated in the HRG only (Fig. 3) suggesting that there was more active growth factor signaling within the ECM of the HRG. Thrombospondins also regulate ECM growth factors, including TGFB activity. In addition, a loss of function polymorphism in THBS4 is strongly associated with premature coronary artery disease [49]. In the present study a more substantial increase in THBS4 mRNA expression was noted in the HRG (Fig. 3). As dysfunction of THBS4 and lack of cardiorespiratory fitness are both risk factors for cardiovascular disease, it is plausible that THBS4 plays a role in the cardioprotective effects of exercise. Further analysis is required to establish if such a relationship exists within a larger human population.
Conclusion
There was unquestionably a differential physiological response between the groups, as the HRG increased their aerobic capacity by, on average, 0.71 L min-1, whereas the LRG did not significantly increase their aerobic capacity (+0.17 L min-1). At this time, our data cannot directly attribute cause-and-effect for the differentially responding genes. Although we profiled changes in skeletal muscle ECM-related gene expression changes, we are not implying that only the local (i.e., in skeletal muscle) role of these genes contributes to the magnitude of the training adaptation. It is entirely plausible that the observed difference between our cohorts reflects a genotype-dependant response, which would impact on gene expression in a range of tissues important for cardiovascular adaptation to endurance-exercise training. Importantly, histological analysis of muscle would not address this possibility. Instead, we would suggest that these differentially expressed genes represent reasonable candidates for future polymorphism studies in larger populations. Our data also demonstrates that direct evidence for physiological adaptability must be presented prior to concluding that gene transcript alterations may or may not occur during the physiological adaptation to exercise.
Methods
Human and physiological measurements
The study was approved by the ethics committee of the Karolinska Institutet, Stockholm, Sweden, and informed consent was obtained from each of the volunteers. Subjects abstained from strenuous exercise during the three weeks prior to obtaining pretraining muscle biopsies (taken from the vastus lateralis). Twenty-four subjects trained under full supervision on a cycle ergometer four times a week (45 min) at 75% of their pretraining peak VO2 for six weeks. Posttraining biopsies were taken 24 h after the last training session. Physiological measurements and muscle biopsies were performed as previously described [8,16,50]. All physiological parameters were derived from a minimum of two assessments, taken on separate days. Peak VO2 was determined using a cycle ergometer (Rodby, Sweden). An incremental protocol was combined with continuous analysis of respiratory gases using the Sensormedic ventilator (Sensormedic Co., USA). At peak VO2 the respiratory exchange ratio and heart rate exceeded 1.1 and 190 BPM, respectively, on all occasions. Total amount of work done in 15 min of cycling was determined using a self-paced protocol (Lode, Netherlands, test-re-test variability <5%). Submaximal physiological parameters were determined during two separate 10 min constant-load, submaximal cycling sessions (carried out at 75% of pretraining peak VO2).
Two groups were selected from a larger training cohort based on the magnitude of their training response (Table 1). The two groups represented the eight subjects that demonstrated the largest improvement in physiological capacity (HRG) and the eight subjects that demonstrated minimal changes in physiological capacity (LRG) after following an identical supervised training program. The average training-induced change in peak VO2 in the LRG equates to the lower ~10% of responses observed in the HERITAGE study [6], whereas the response in the HRG represents the top ~10% of responses. This information helps to establish a 'population perspective' on the responses observed in the present study. Subjects were assigned to each group after being ranked on the basis of the sum of changes in three main physiological parameters: (a) percent improvement in peak aerobic capacity, (b) percent reduction in submaximal heart rate during 10 min fixed-workload, submaximal cycling and (c) percent improvement in work done during a 15 min maximal cycling test. A combination of physiological markers of adaptation was used to ensure that no spurious individual physiological measurement would compromise the overall categorization of a subject [8]. Peak aerobic capacity, decreased exercise-induced-increases in heart rate and aerobic performance are all valid and accepted responses to aerobic exercise training. No molecular or biochemical analysis was carried out until the subjects were assigned to their particular group. This type of novel analysis strategy has previously been used by our group and also, more recently, by Bouchard et al [8,9]
Real-time quantitative PCR and statistics
Total RNA was prepared using the TRIzol method (Invitrogen, USA) and quantified using a spectrophotometer. Two μg of RNA was reverse transcribed by Superscript reverse transcriptase (Life Technologies, Sweden) using random hexamer primers (Roche Diagnostics GmbH, Mannheim, Germany) in a total volume of 20 μl. Detection of mRNA was performed using a ABI-PRISM® 7700 Sequence Detector (Perkin-Elmer Applied Biosystems Inc, Foster City, CA, USA). All reactions were performed in 96-well MicroAmp Optical plates (Perkin-Elmer Applied Biosystems Inc.). Amplification aliquots contained 5 μl of the sample cDNA, the TaqMan Universal PCR master mix (Perkin-Elmer Applied Biosystems Inc.) and an optimized concentration of each primer and probe, prepared according to the manufacturer's recommendation, in a final volume of 25 μl. 18S rRNA was selected as an endogenous control to correct for potential variations in RNA loading into the cDNA synthesis reaction. The 18S rRNA control was run in triplicate in separate wells (using 1:2000 dilution of the original cDNA). Thermal cycling conditions were, initially, 2 min at 50°C followed by 10 min at 95°C, and then, subsequently, 45 cycles of 15 s at 95°C and 1 min at 65°C.
Oligonucleotide primers and TaqMan probes were designed using Primer Express version 1.5 (Perkin-Elmer Applied Biosystems Inc.) and synthesized by Cybergene (Stockholm, Sweden) or ordered as a 'gene assay by demand' product (Perkin-Elmer Applied Biosystems Inc). The sequences or 'gene assay by demand' numbers can be provided on request. The probes were designed to cover exon-exon boundaries to avoid amplification of genomic DNA. As there are no predetermined low-abundance "house-keeping" genes for this experimental paradigm, the ΔΔCt method [51] was used to calculate relative changes in mRNA abundance. The threshold cycle (Ct) for 18S was subtracted from the Ct for the target gene to adjust for variations in mRNA/cDNA generation efficacy. This was carried out for both pre- and posttraining samples. The preexercise value reflects baseline gene expression levels and was subtracted from the postexercise value to calculate the increase or decrease in mRNA abundance.
A two-way mixed model ANOVA (GraphPad Prism 4.0) was used to establish whether a significant interaction between 'group' and 'extracellular gene responses' occurred (p < 0.0001) and to confirm appropriate baseline subject matching (p < 0.0001). Bonferoni post-hoc testing established which individual gene responses were significant for each subgroup, and this analysis was used for the basis of the discussion. For comparison between HRG and LRG training response data (e.g., peak VO2), a two-tailed unpaired t-test was utilized. Significance was accepted at the 5 % level. Data are mean ± SE.
List of abbreviations used
α2-Macroglobulin (AM2)
Angiopoietin (ANG)
Angiopoietin 1 (ANG 1)
Angiopoietin 2 (ANG2)
Beats per minute (BPM)
Collagen type IIIα1 (COL3A1)
Collagen type XVα1 (COL15A1)
Extracellular matrix (ECM)
High-responder group (HRG)
Hypoxia-inducible factor 1α (HIF)
Interleukin 17D (IL17D)
Low responder group (LRG)
Myristoylated alanine-rich protein kinase C substrate (MARKS)
Neuropilin (NP-1)
Rate of oxygen consumption (VO2)
Rho-GTPase-activating protein 1 (ARHGAP1)
Threshold cycle (Ct)
Thrombospondin-4 (THBS4)
Transforming growth factor β (TGFB)
Transforming growth factor β2 (TGFB2)
Transforming growth factor β receptor II (TGFBR2)
Tyrosine kinase with immunoglobulin-like and EGF-like domains 1 (TIE1)
Tyrosine kinase with immunoglobulin-like and EGF-like domains 2 (TIE2)
Vascular endothelial growth factor (VEGF)
Vascular endothelial growth factor related receptor 1 (VEGFR1)
Vascular endothelial growth factor related receptor 2 (VEGFR2)
Authors' contributions
EJ, JAT, PLG, JRa, JRi and CJS were responsible for the design and management of the endurance training study, subject recruitment and tissue sampling. JAT was responsible for generating the strategy for the gene expression analysis. JAT selected the biological theme and the majority of the genes to be profiled in this particular study, while TG selected some of the angiopoietin genes and ordered the majority of the primers. TG and HF were responsible for organizing the real time PCR analysis while JAT was responsible for the subsequent calculations and computations. JAT and CJS were responsible for drafting and revising the manuscript, respectively.
Acknowledgements
The authors would like to thank Laura Svensson for providing technical assistance and Helene Ameln for designing the HIF1α primers. This work was supported by the Swedish Heart-Lung Association, the Swedish Sports Council, the Thurings Foundation and the Swedish Medical Research Council (VR). JAT is supported by the Swedish Diabetes Association.
Figures and Tables
Figure 1 Change in Vascular endothelial growth factor-related gene expression following endurance training. Values are -fold changes in human skeletal muscle gene expression (mean ± SE) following six weeks of aerobic training. Gene expression was determined using real-time quantitative PCR. Following six weeks training (n = 24), the eight highest and eight lowest responders to exercise training were identified using the sum of (a) the percent improvement in maximal aerobic capacity, (b) the percent reduction in submaximal heart rate during 15 min fixed-workload, submaximal cycling and (c) the percent improvement in work done during a 15 min maximal cycling test. This ranking was carried out before any genomic analysis was carried out. The training responses were evaluated by two-way ANOVA and Bonferoni post-hoc tests. * indicates p < 0.05, ** indicates p < 0.01 and *** indicates p < 0.001.
Figure 2 Changes in Angiopoietin-related gene expression following endurance training. Values are -fold changes in human skeletal muscle gene expression (mean ± SE) following six weeks of aerobic training, as described in Fig. 1. * indicates p < 0.05, ** indicates p < 0.01 and *** indicates p < 0.001.
Figure 3 Change in extracellular-matrix-related gene expression following endurance training. Values are -fold changes in human skeletal muscle gene expression (mean ± SE) following six weeks of aerobic training, as described in Fig. 1. * indicates p < 0.05, ** indicates p < 0.01 and *** indicates p < 0.001.
Table 1 Baseline demographic and physiological parameters
LRG1 HRG1 p-value
Height (cm) 180 ± 3 183 ± 3 p = 0.53
Age (year) 23 ± 1 24 ± 1 p = 0.54
Mass (kg) 77 ± 3 77 ± 6 p = 0.97
Mean blood pressure (mm Hg) 92 ± 1 88 ± 4 p = 0.25
Peak VO2 (L min-1)2 3.7 ± 0.1 3.5 ± 0.3 p = 0.48
Resting heart rate (BMP) 71 ± 5 70 ± 6 p = 0.94
Submaximal heart rate3 170 ± 5 171 ± 5 p = 0.85
Respiratory exchange ratio4 1.0 ± 0.0 1.0 ± 0.0 p = 0.45
15 min work (kJ)5 220 ± 9 204 ± 16 p = 0.37
1Values are mean (± SEM) taken prior to training.
2Peak VO2 is the increase in 'maximal' oxygen uptake measured during an incremental maximal exercise protocol.
3Submaximal heart rate was measured during 10 min constant load cycling at 75% peak VO2.
4The respiratory exchange ratio was obtained during 10 min of submaximal exercise at 75% of pretraining peak VO2.
515 min work is the total work done in 15 min of self-paced cycling.
==== Refs
Kraus WE Houmard JA Duscha BD Knetzger KJ Wharton MB McCartney JS Bales CW Henes S Samsa GP Otvos JD Effects of the amount and intensity of exercise on plasma lipoproteins N Engl J Med 2002 347 1483 1492 12421890 10.1056/NEJMoa020194
Frederiksen H Christensen K The influence of genetic factors on physical functioning and exercise in second half of life Scand J Med Sci Sports 2003 13 9 18 12535312 10.1034/j.1600-0838.2003.20219.x
Manson JE Greenland P LaCroix AZ Stefanick ML Mouton CP Oberman A Perri MG Sheps DS Pettinger MB Siscovick DS Walking compared with vigorous exercise for the prevention of cardiovascular events in women N Engl J Med 2002 347 716 725 12213942 10.1056/NEJMoa021067
Wei M Gibbons LW Mitchell TL Kampert JB Lee CD Blair SN The association between cardiorespiratory fitness and impaired fasting glucose and type 2 diabetes mellitus in men Ann Intern Med 1999 130 89 96 10068380
Wisloff U Najjar SM Ellingsen O Haram PM Swoap S Al-Share Q Fernstrom M Rezaei K Lee SJ Koch LG Britton SL Cardiovascular risk factors emerge after artificial selection for low aerobic capacity Science 2005 307 418 420 15662013 10.1126/science.1108177
Bouchard C Rankinen T Individual differences in response to regular physical activity Med Sci Sports Exerc 2001 33 S446 451 discussion S452-443 11427769 10.1097/00005768-200105001-01273
Rico-Sanz J Rankinen T Rice T Leon AS Skinner JS Wilmore JH Rao DC Bouchard C Quantitative trait loci for maximal exercise capacity phenotypes and their responses to training in the HERITAGE Family Study Physiol Genomics 2004 16 256 260 14625375 10.1152/physiolgenomics.00035.2003
Timmons JA Larsson O Jansson E Fischer H Gustafsson T Greenhaff PL Ridden J Rachman J Peyrard-Janvid M Wahlestedt C Sundberg CJ Human muscle gene expression responses to endurance training provide a novel perspective on Duchenne muscular dystrophy Faseb J 2005 19 750 760 15857889 10.1096/fj.04-1980com
Teran-Garcia M Rankinen T Koza RA Rao DC Bouchard C Endurance training-induced changes in insulin sensitivity and gene expression Am J Physiol Endocrinol Metab 2005 288 E1168 E1178 15687108 10.1152/ajpendo.00467.2004
Lortie G Simoneau JA Hamel P Boulay MR Landry F Bouchard C Responses of maximal aerobic power and capacity to aerobic training Int J Sports Med 1984 5 232 236 6500788
Prud'homme D Bouchard C Leblanc C Landry F Fontaine E Sensitivity of maximal aerobic power to training is genotype-dependent Med Sci Sports Exerc 1984 16 489 493 6542620
Pilegaard H Saltin B Neufer PD Exercise induces transient transcriptional activation of the PGC-1alpha gene in human skeletal muscle J Physiol 2003 546 851 858 12563009 10.1113/jphysiol.2002.034850
Bengtsson J Gustafsson T Widegren U Jansson E Sundberg CJ Mitochondrial transcription factor A and respiratory complex IV increase in response to exercise training in humans Pflugers Arch 2001 443 61 66 11692267 10.1007/s004240100628
Hittel DS Kraus WE Tanner CJ Houmard JA Hoffman EP Exercise training increases electron and substrate shuttling proteins in muscle of overweight men and women with the metabolic syndrome J Appl Physiol 2005 98 168 179 15347626 10.1152/japplphysiol.00331.2004
Jensen L Pilegaard H Neufer PD Hellsten Y Effect of acute exercise and exercise training on VEGF splice variants in human skeletal muscle Am J Physiol Regul Integr Comp Physiol 2004 287 R397 402 15117722
Norrbom J Sundberg CJ Ameln H Kraus WE Jansson E Gustafsson T PGC-1alpha mRNA expression is influenced by metabolic perturbation in exercising human skeletal muscle J Appl Physiol 2004 96 189 194 12972445 10.1152/japplphysiol.00765.2003
Richardson RS Wagner H Mudaliar SR Saucedo E Henry R Wagner PD Exercise adaptation attenuates VEGF gene expression in human skeletal muscle Am J Physiol Heart Circ Physiol 2000 279 H772 778 10924077
Pedersen BK Steensberg A Fischer C Keller C Keller P Plomgaard P Febbraio M Saltin B Searching for the exercise factor: is IL-6 a candidate? J Muscle Res Cell Motil 2003 24 113 119 14609022 10.1023/A:1026070911202
Jensen L Bangsbo J Hellsten Y Effect of high intensity training on capillarization and presence of angiogenic factors in human skeletal muscle J Physioln 2004 557 571 582 10.1113/jphysiol.2003.057711
Hepple RT Mackinnon SL Goodman JM Thomas SG Plyley MJ Resistance and aerobic training in older men: effects on VO2peak and the capillary supply to skeletal muscle J Appl Physiol 1997 82 1305 1310 9104869 10.1063/1.365903
Klausen K Andersen LB Pelle I Adaptive-changes in work capacity, skeletal-muscle capillarization and enzyme levels during training and detraining Acta Physiologica Scandinavica 1981 113 9 16 7315443
Ingjer F Effects of endurance training on muscle-fiber ATP-ase activity, capillary supply and mitochondrial content in man J Physiol 1979 294 419 432 159945
Saltin B Gollnick P Peachey L, Adrian RH, Geiger SR Skeletal muscle adaptability: significance for metabolism and performance Handbook of Physiology Skeletal muscle 1983 10 Bethesda, Maryland: American Physiological Society
Sundberg CJ Exercise and training during graded leg ischaemia in healthy man with special reference to effects on skeletal muscle Acta Physiol Scand Suppl 1994 615 1 50 8140900
Blomqvist CG Saltin B Cardiovascular adaptations to physical-training Annual Review of Physiology 1983 45 169 189 6221687 10.1146/annurev.ph.45.030183.001125
Gustafsson T Puntschart A Kaijser L Jansson E Sundberg CJ Exercise-induced expression of angiogenesis-related transcription and growth factors in human skeletal muscle Am J Physiol 1999 276 H679 685 9950871
Lloyd PG Prior BM Yang HT Terjung RL Angiogenic growth factor expression in rat skeletal muscle in response to exercise training Am J Physiol Heart Circ Physiol 2003 284 H1668 1678 12543634
Gustafsson T Ameln H Fischer H Sundberg CJ Timmons JA Jansson E VEGF-A splice variants and related receptor expression in human skeletal muscle following submaximal exercise J Appl Physiol 2005 98 2137 2146 15661835 10.1152/japplphysiol.01402.2004
Shyu KG Chang H Isner JM Synergistic effect of angiopoietin-1 and vascular endothelial growth factor on neoangiogenesis in hypercholesterolemic rabbit model with acute hindlimb ischemia Life Sci 2003 73 563 579 12770612
Yamauchi A Ito Y Morikawa M Kobune M Huang J Sasaki K Takahashi K Nakamura K Dehari H Niitsu Y Pre-administration of angiopoietin-1 followed by VEGF induces functional and mature vascular formation in a rabbit ischemic model J Gene Med 2003 5 994 1004 14601137 10.1002/jgm.439
Kohrt WM Malley MT Coggan AR Spina RJ Ogawa T Ehsani AA Bourey RE Martin WH 3rdHolloszy JO Effects of gender, age, and fitness level on response of VO2max to training in 60–71 yr olds J Appl Physiol 1991 71 2004 2011 1761503
Webb DJ Weaver AM Atkins-Brady TL Gonias SL Proteinases are isoform-specific regulators of the binding of transforming growth factor beta to alpha 2-macroglobulin Biochem J 1996 320 551 555 8973565
Liu X Wu H Byrne M Krane S Jaenisch R Type III collagen is crucial for collagen I fibrillogenesis and for normal cardiovascular development Proc Natl Acad Sci U S A 1997 94 1852 1856 9050868 10.1073/pnas.94.5.1852
Sasaki T Larsson H Tisi D Claesson-Welsh L Hohenester E Timpl R Endostatins derived from collagens XV and XVIII differ in structural and binding properties, tissue distribution and anti-angiogenic activity J Mol Biol 2000 301 1179 1190 10966814 10.1006/jmbi.2000.3996
Hawley JA Exercise as a therapeutic intervention for the prevention and treatment of insulin resistance Diabetes Metab Res Rev 2004 20 383 393 15343584 10.1002/dmrr.505
Dela F Ploug T Handberg A Petersen LN Larsen JJ Mikines KJ Galbo H Physical training increases muscle GLUT4 protein and mRNA in patients with NIDDM Diabetes 1994 43 862 865 8013748
van Weel V Deckers MM Grimbergen JM van Leuven KJ Lardenoye JH Schlingemann RO van Nieuw Amerongen GP van Bockel JH van Hinsbergh VW Quax PH Vascular endothelial growth factor overexpression in ischemic skeletal muscle enhances myoglobin expression in vivo Circ Res 2004 95 58 66 15155530 10.1161/01.RES.0000133247.69803.c3
Semenza GL Regulation of mammalian O2 homeostasis by hypoxia-inducible factor 1 Annu Rev Cell Dev Biol 1999 15 551 578 10611972 10.1146/annurev.cellbio.15.1.551
Oh H Takagi H Otani A Koyama S Kemmochi S Uemura A Honda Y Selective induction of neuropilin-1 by vascular endothelial growth factor (VEGF): a mechanism contributing to VEGF-induced angiogenesis Proc Natl Acad Sci U S A 2002 99 383 388 11756651 10.1073/pnas.012074399
Soker S Miao HQ Nomi M Takashima S Klagsbrun M VEGF165 mediates formation of complexes containing VEGFR-2 and neuropilin-1 that enhance VEGF165-receptor binding J Cell Biochem 2002 85 357 368 11948691 10.1002/jcb.10140
Claesson-Welsh L Signal transduction by vascular endothelial growth factor receptors Biochem Soc Trans 2003 31 20 24 12546646
Holash J Maisonpierre PC Compton D Boland P Alexander CR Zagzag D Yancopoulos GD Wiegand SJ Vessel cooption, regression, and growth in tumors mediated by angiopoietins and VEGF Science 1999 284 1994 1998 10373119 10.1126/science.284.5422.1994
Lobov IB Brooks PC Lang RA Angiopoietin-2 displays VEGF-dependent modulation of capillary structure and endothelial cell survival in vivo Proc Natl Acad Sci U S A 2002 99 11205 11210 12163646 10.1073/pnas.172161899
Teichert-Kuliszewska K Maisonpierre PC Jones N Campbell AI Master Z Bendeck MP Alitalo K Dumont DJ Yancopoulos GD Stewart DJ Biological action of angiopoietin-2 in a fibrin matrix model of angiogenesis is associated with activation of Tie2 Cardiovasc Res 2001 49 659 670 11166279 10.1016/S0008-6363(00)00231-5
Maisonpierre PC Suri C Jones PF Bartunkova S Wiegand SJ Radziejewski C Compton D McClain J Aldrich TH Papadopoulos N Angiopoietin-2, a natural antagonist for Tie2 that disrupts in vivo angiogenesis Science 1997 277 55 60 9204896 10.1126/science.277.5322.55
Nishishita T Lin PC Angiopoietin 1, PDGF-B, and TGF-beta gene regulation in endothelial cell and smooth muscle cell interaction J Cell Biochem 2004 91 584 593 14755687 10.1002/jcb.10718
Mathew S Arandjelovic S Beyer WF Gonias SL Pizzo SV Characterization of the interaction between alpha2-macroglobulin and fibroblast growth factor-2: the role of hydrophobic interactions Biochem J 2003 374 123 129 12755687 10.1042/BJ20021655
Soker S Svahn CM Neufeld G Vascular endothelial growth factor is inactivated by binding to alpha 2-macroglobulin and the binding is inhibited by heparin J Biol Chem 1993 268 7685 7691 7681826
Stenina OI Byzova TV Adams JC McCarthy JJ Topol EJ Plow EF Coronary artery disease and the thrombospondin single nucleotide polymorphisms Int J Biochem Cell Biol 2004 36 1013 1030 15094117 10.1016/j.biocel.2004.01.005
Timmons JA Gustafsson T Sundberg CJ Jansson E Hultman E Kaijser L Chwalbinska-Moneta J Constantin-Teodosiu D Macdonald IA Greenhaff PL Substrate availability limits human skeletal muscle oxidative ATP regeneration at the onset of ischemic exercise J Clin Invest 1998 101 79 85 9421469
Winer J Jung CK Shackel I Williams PM Development and validation of real-time quantitative reverse transcriptase-polymerase chain reaction for monitoring gene expression in cardiac myocytes in vitro Anal Biochem 1999 270 41 49 10328763 10.1006/abio.1999.4085
|
16138928
|
PMC1224855
|
CC BY
|
2021-01-04 16:02:56
|
no
|
BMC Biol. 2005 Sep 2; 3:19
|
utf-8
|
BMC Biol
| 2,005 |
10.1186/1741-7007-3-19
|
oa_comm
|
==== Front
BMC NeurosciBMC Neuroscience1471-2202BioMed Central London 1471-2202-6-561613524310.1186/1471-2202-6-56Research ArticleReduction of the hand representation in the ipsilateral primary motor cortex following unilateral section of the corticospinal tract at cervical level in monkeys Schmidlin Eric [email protected] Thierry [email protected] Jocelyne [email protected] Abderraouf [email protected] Alexander F [email protected] Eric M [email protected] Unit of Physiology and Program in Neurosciences, Department of Medicine, Faculty of Sciences, University of Fribourg, Chemin du Musée 5, CH-1700 Fribourg, Switzerland2 Brain Research Institute, Department of Neuromorphology, University and ETH Zurich, Winterthurerstrasse 190, CH-8057 Zürich, Switzerland3 Department of Neurosurgery, Neurosurgery Clinic, University Hospital of Lausanne, Rue du Bugnon, CH-1011 Lausanne, Switzerland2005 31 8 2005 6 56 56 11 4 2005 31 8 2005 Copyright © 2005 Schmidlin et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
After sub-total hemi-section of cervical cord at level C7/C8 in monkeys, the ipsilesional hand exhibited a paralysis for a couple of weeks, followed by incomplete recovery of manual dexterity, reaching a plateau after 40–50 days. Recently, we demonstrated that the level of the plateau was related to the size of the lesion and that progressive plastic changes of the motor map in the contralesional motor cortex, particularly the hand representation, took place following a comparable time course. The goal of the present study was to assess, in three macaque monkeys, whether the hand representation in the ipsilesional primary motor cortex (M1) was also affected by the cervical hemi-section.
Results
Unexpectedly, based on the minor contribution of the ipsilesional hemisphere to the transected corticospinal (CS) tract, a considerable reduction of the hand representation was also observed in the ipsilesional M1. Mapping control experiments ruled out the possibility that changes of motor maps are due to variability of the intracortical microstimulation mapping technique. The extent of the size reduction of the hand area was nearly as large as in the contralesional hemisphere in two of the three monkeys. In the third monkey, it represented a reduction by a factor of half the change observed in the contralesional hemisphere. Although the hand representation was modified in the ipsilesional hemisphere, such changes were not correlated with a contribution of this hemisphere to the incomplete recovery of the manual dexterity for the hand affected by the lesion, as demonstrated by reversible inactivation experiments (in contrast to the contralesional hemisphere). Moreover, despite the size reduction of M1 hand area in the ipsilesional hemisphere, no deficit of manual dexterity for the hand opposite to the cervical hemi-section was detected.
Conclusion
After cervical hemi-section, the ipsilesional motor cortex exhibited substantial reduction of the hand representation, whose extent did not match the small number of axotomized CS neurons. We hypothesized that the paradoxical reduction of hand representation in the ipsilesional hemisphere is secondary to the changes taking place in the contralesional hemisphere, possibly corresponding to postural adjustments and/or re-establishing a balance between the two hemispheres.
==== Body
Background
Although voluntary dexterous movements of the hand are mainly under control of motor cortical areas in the opposite hemisphere, there is evidence that the ipsilateral motor cortex may also contribute, but to a lesser extent. For instance, the activity of single neurons in the primary (M1), supplementary (SMA), premotor (PM) and cingulate (CMA) motor cortical areas was found to be modulated when monkeys performed movements with the ipsilateral hand. [1-7]. Intracortical microstimulations in the primary motor cortex in monkeys were reported to evoke not only the expected movements of the contralateral digits [8] but also responses of the ipsilateral fingers [9]. In human subjects, there is increasing evidence that the motor cortex is involved in the control of ipsilateral hand movements (e.g. [10-16]). The possible contribution of the motor cortex in the control of the ipsilateral hand may be important for normal function, although its precise role has not been elucidated yet. Furthermore, it has been anticipated that the ipsilateral motor cortex may be crucial for recovery of motor function of a paretic hand after unilateral brain lesion, such as after stroke (e.g. [17-20]). However, the involvement of the intact hemisphere in the control of the ipsilateral paretic hand remains a matter for debate. [21-23]. Moreover, there is clear evidence, in both monkeys and human subjects, that a significant re-organization of motor maps takes place in the affected hemisphere after unilateral lesion of M1. [24-28], a re-arrangement crucially involved in the functional recovery of the paretic hand [25].
Regarding lesions of the spinal cord, several studies reported plastic changes of motor maps in the cerebral cortex in both monkeys [29-31] and human subjects [32-37]. In a recent report, we described in detail the changes of motor maps that occurred in the motor cortex contralateral to a unilateral section of the corticospinal (CS) tract at cervical level C7/C8 [31]. Several months after the lesion, the contralesional hemisphere showed a dramatic decrease of the hand representation, compared with before the lesion, ranging from 69 to 97 % depending on the site and extent of the lesion. The progressive changes in hand representation occurred during the 30–40 days post-lesion, in parallel to the functional recovery of the affected hand. Finally, we demonstrated that the re-arranged contralesional motor cortex with a diminished hand representation was crucial for the functional recovery of the affected hand, since its reversible inactivation abolished the recovered motor performance [31]. The goal of the present report was to address the issue of possible changes in the hemisphere ipsilateral to the unilateral cervical lesion, in other words the ipsilesional hemisphere. More specifically, the following questions were addressed:
- Does a unilateral section of the CS tract at cervical level affect the hand representation in the ipsilesional motor cortex? If yes, to what extent as compared to the contralesional hemisphere? Based on the small proportion of undecussated CS axons (about 5–10%; [38-40]), one would predict a very limited, if any, impact of a unilateral cervical lesion on the ipsilesional motor cortex.
- In the context of the recovery of motor control from unilateral section of the CS tract, does the ipsilesional hemisphere play a role, in addition to the substantial contribution of the contralesional hemisphere?
Results
Unilateral section of the CS tract at cervical level
Three monkeys (Mk1, Mk2 and Mk3) were subjected to a unilateral section of the CS tract at cervical level. The extent and precise position of the lesion were assessed from reconstructions of cervical cord histological sections cut in the paralongitudinal plane as explained previously [31,41]. The lesion was finally represented for Mks1-3 on a transverse reconstruction of the cervical cord (Fig. 1). For comparison, the left inset of Figure 1 shows the distribution of the CS axons at cervical level after injection of the anterograde tracer Biotinylated Dextran Amine (BDA) in the right motor cortex, as recently described [41]. The smallest lesion was in Mk1 (Fig. 1), which however interrupted most of the dorsolateral funiculus comprising the crossed CS axons originating from the contralesional hemisphere. A restricted part of the CS axons located in the most ventral and lateral part of the dorsolateral funiculus (about 5%) may have however been preserved in Mk1. In contrast, the lesion was larger in Mk2 and Mk3, for which the left dorsolateral funiculus was, respectively, completely or nearly completely transected (Fig. 1). Overall, as intended, the lesion affected substantially if not totally the left dorsolateral funiculus in Mks1-3, above the spinal segments where the motoneurons controlling hand muscles are located [42]. Regarding the uncrossed CS tracts originating from the ipsilesional hemisphere, the comparison with the inset of Figure 1 indicates that the lesion in Mks1-3 interrupted most (Mk1) or all (Mks2-3) of the uncrossed CS axons located in the dorsolateral funiculus, but not the few CS axons running in the ventral funiculus.
Mapping of ipsilesional M1 hand area before and after cervical cord lesion: ICMS data
The experimental paradigm is depicted in Figure 2. As a result of the left cervical lesion, the ipsilesional hand was paralyzed, but showed incomplete recovery in about 30–40 days [31]. In the present report, the ipsilesional hemisphere was investigated electrophysiologically, using intracortical microstimulation (ICMS) at two time points: before the lesion (pre-lesion motor map) and a few months after the lesion (post-lesion motor map) when the monkey has reached a plateau of manual dexterity recovery. The pre- and post-lesion cortical maps established in the ipsilesional hemisphere report on the movement elicited by ICMS on the contralesional (right) body side, with a focus on the hand not affected by the lesion (Fig. 2). In the present experiments, as a result of ICMS in the ipsilesional hemisphere, no movements were observed for the ipsilesional hand.
The motor cortical map established before the cervical lesion (left column in Fig. 3) exhibits the presence of a clear hand area, defined here as the outline of penetration points represented on the surface along which the lowest threshold ICMS effect was a movement of digits (D1- D5 symbols) of the opposite (right) hand. In the three monkeys, the hand area in M1 is surrounded by points where ICMS elicited movements of other body territories, such as face (F), wrist (W), elbow (E) or shoulder (S). In Mk2 and Mk3, a slightly more rostral position of the chronic recording chamber allowed us to investigate more anterior regions of the motor cortex. There we found a second, smaller hand area, characterized by slightly higher thresholds to elicit hand movements than the main hand area in M1 (Fig. 3, left panel of middle and bottom rows). To address the question of whether such somatotopic representations are influenced in the ipsilesional motor cortex as a result of the unilateral cervical lesion (as recently reported for the contralesional hemisphere, see [31]), the ipsilesional hemisphere was re-mapped a few months post-lesion (Fig. 3, right column). Surprisingly, there were substantial changes in motor map in the ipsilesional hemisphere as well. The area of the hand representation in M1 projected on the surface of the hemisphere was reduced by a factor of 52% in Mk1, 77% in Mk2 and 43% in Mk3 (Fig. 3). Clearly, some ICMS sites which elicited hand movements before the lesion were replaced at threshold by effects on other body territories post-lesion (gray symbols in Fig. 3) or became non microexcitable (X symbols within the outlined area in the right column of Fig. 3).
The ICMS maps shown in Figure 3 display only one point on the surface of the motor cortex for each electrode penetration and only the effect of the lowest threshold was indicated. As a result, other ICMS sites where digit movements were present (but for an intensity of stimulation higher than the lowest threshold for the corresponding electrode track) do not appear. For this reason, all electrode penetrations where digit movements have been observed (irrespective of the intensity of stimulation) are represented in Figure 4, in which only the ICMS data pertaining to digit movements are represented. As expected, the hand territories in Figure 4 (left column) spread on a larger zone of the motor cortex than in the map considering only ICMS effects at threshold (Fig. 3). Nevertheless, as a result of the unilateral cervical lesion, some of these additional hand ICMS sites also disappeared or a few of them exhibited a higher threshold after lesion as compared to the pre-lesion map (Fig. 4).
Instead of considering only one ICMS site (at lowest threshold) per electrode penetration as in Figure 3, or only one body territory (the hand) as in Figure 4 (also with only one ICMS site representing a given electrode penetration), one can also conduct an analysis taking into consideration all ICMS sites tested, pre- and post-lesion. In this way, one avoids (as in Figures 3 and 4) the underestimation of the hand territory lying in the rostral bank of the central sulcus, where a penetration all the way down the bank was represented by a single point on the surface map. In the post-lesion ICMS sessions, most electrode penetrations that were performed during the pre-lesion ICMS sessions were repeated. Table 1 lists the distribution, in absolute numbers as well as in percentage, of all ICMS sites tested in the three lesioned monkeys, separately for the pre- and post-lesion ICMS sessions. The total numbers of ICMS sites (rightmost column in Table 1) indicate that roughly comparable numbers of stimulation sites were tested pre- and post-lesion. For each monkey, the ICMS data points were divided into three groups based on the territories activated or absence of effect: i) digit movements (hand territory), ii) movements of "other" body territories (mainly wrist, elbow, shoulder or face), iii) ICMS sites which did not elicit any visible movement (corresponding to non-microexcitable sites). In all three lesioned monkeys, the comparison pre- and post-lesion of the distribution of ICMS effects observed at all sites tested clearly shows a decrease of the number of ICMS sites eliciting digit movements as a result of the lesion. In percentages, the number of digit ICMS sites was two to nearly three times lower post- than pre-lesion across monkeys, in line with the decrease of the surface of the hand area seen in Figures 3 and 4. The digit ICMS sites lost as a result of the lesion were replaced in most cases by non-microexcitable sites, as shown by the substantial increase (by a factor of 1.3 – 1.5 across monkeys) of such points comparing pre- and post-lesion data (Table 1). For the other body territories, the number of sites slightly increased post-lesion in Mk1 and Mk2 (suggesting replacement of a few original digit points), but decreased in Mk3. As shown in Figure 3, ICMS penetrations that belonged to the hand area before lesion (where ICMS elicited digit movements at lowest threshold) were replaced post-lesion mainly by penetrations which became part of the wrist, elbow, shoulder or even face representations.
Comparison of ICMS thresholds
Although both the surface of the hand representation and the number of digit sites decreased in the contralesional M1, it was observed that the ICMS thresholds in the hand area post-lesion were not significantly higher than the ICMS thresholds derived from the hand area pre-lesion [31]. This analysis demonstrated that, at least for the contralesional hemisphere, the hand area though decreased in size as a result of the unilateral cervical lesion did not change with respect to its excitability to address the motoneurons of hand muscles, as well as the muscles of other body territories (face, wrist, elbow, shoulder, trunk). The question here is whether the ICMS thresholds were also kept unchanged pre- versus post-lesion in the ipsilesional hemisphere? To address this question, we compared the ICMS thresholds in the ipsilesional hemisphere required to elicit movements from stimulation at the same stereotaxic points before and after the lesion, in the latter case when the manual dexterity score reached the plateau. In contrast to Figure 3 where only the best ICMS site along each electrode penetration was represented, all ICMS sites of stimulation were considered. In the three monkeys, there was no systematic and statistically significant difference between the ICMS thresholds obtained in the ipsilesional hemisphere, before and after the unilateral cervical cord lesion, neither for the hand nor for other body territories (wrist, elbow, shoulder and face).
Variability of the ICMS mapping method
One may argue that the size reduction of the hand representation post-lesion, as compared to the pre-lesion situation (Figs. 3 and 4), is due to the intrinsic variability of the ICMS method. In other words, what is the variability of ICMS sites elicited along two electrode penetrations performed at two distinct time points in an intact monkey? This question was addressed in an intact monkey (Mk4), in which six electrode penetrations taken from the left hemisphere were repeated at time intervals ranging between 12 and 140 days (Fig. 5). For instance, three electrode penetrations located in the hand area (tracks 3, 5 and 6), in which movements of the digits were observed at the lowest threshold site when the electrode was inserted the first time (arrows in the left columns), exhibited again a movement of the digits at the lowest effective current intensity 125, 26 and 138 days later. Interestingly, note that the ICMS thresholds were highly comparable at the two time points (Fig. 5). Along these three electrode tracks (3, 5 and 6), the sequence of territories activated at the consecutive ICMS sites (with a step of 1 mm along the penetration) remained generally comparable at the two time points.
Two other electrode penetrations were taken from the wrist representation (tracks 2 and 4) and repeated at two time points, separated by an interval of 21 and 140 days, respectively (Fig. 5). In both tracks, the first penetration yielded several ICMS sites at which the elbow ("E") articulation was activated, replaced in the second penetrations by wrist movements in most cases. However, the lowest efficient current intensity corresponded to a wrist ("W") movement in the two tracks, both during the first and the second penetrations (Fig. 5). Again, as for digits territories, the threshold obtained for these two wrist territories remained similar at the two time points tested. Along the same line, in the electrode track located in the face representation (track 1), at the two time points tested (12 days apart), ICMS at threshold elicited movements of face muscles (Fig. 5). In summary, from the six electrode tracks repeated at two time points (Fig. 5), one can conclude that, in spite of some variability at some ICMS sites, the territory assigned to each track as defined by the effect observed at threshold did not change, even when the time interval was as long as 140 days. These observations support the notion that the size reduction of the hand area observed here in Figures 3 and 4 cannot be explained by the intrinsic variability of the ICMS method and thus are indeed related to the cervical lesion. Along the same line, a few individual electrode penetrations repeated twice before the cervical lesion in Mks1-3 also showed reproducibility of motor map body territories assessed by ICMS [31].
Does the ipsilesional (reduced in size) hand area contribute to the post-lesion recovery of the affected hand?
To address this question, reversible inactivation sessions of M1 in the ipsilesional hemisphere were performed between 3 to 5 months post-lesion, when the manual dexterity score derived from the Brinkman board test had reached a plateau, indicative of the maximal level of recovery. In a typical inactivation session obtained by infusion of muscimol, the manual dexterity task was initially performed just before the muscimol infusion and repeated 40 – 45 minutes afterwards- a time point at which its effect is well established. A pre- and post-inactivation manual dexterity score was thus determined for each hand, corresponding to the bars "B" (= before) and "A" (= after) in Figure 6. Before infusion of muscimol ("B"), the manual dexterity score of the ipsilesional hand was that corresponding to the plateau of incomplete recovery (Fig. 6, top panel). In the three lesioned monkeys, a reversible inactivation of the ipsilesional M1 did not noticeably affect the recovered manual dexterity score of the ipsilesional hand (bars under "B" and "A" in the top panel of Fig. 6 are comparable for each animal). In other words, the level of the recovery plateau for the ipsilesional hand was not modified by inactivation of the ipsilesional M1 hand area. In sharp contrast, as recently reported [31], a reversible inactivation of the contralesional M1 led to a complete loss of the recovered performance of the ipsilesional hand (not shown here). The data presented in the bottom panel of Figure 6 for the contralesional hand demonstrate the efficacy of the muscimol reversible inactivation method since, as expected, the pharmacological lesion of the left M1 hand area dramatically suppressed the ability of the contralesional hand to perform the precision grip. Recovery from inactivation using muscimol is slow (several hours) and therefore could not be tested within the same inactivation session. However, the day after, the animal had fully recovered its manual dexterity from before the inactivation session.
Does the size reduction of the ipsilesional M1 hand area lead to a post-lesion motor deficit of the hand contralateral to the cervical hemi-section?
The manual dexterity of the contralesional hand was assessed during several weeks pre- and post-lesion using the "Brinkman board" task (Fig. 7). These data can determine whether a substantial reduction of the size of the M1 ipsilesional hand area is accompanied by a decrease of the manual dexterity of the contralesional hand. As shown by the behavioral plots, the manual dexterity scores for the three lesioned monkeys are very comparable pre- and post-lesion (Fig. 7). The manual dexterity of the contralesional hand is thus not affected by the unilateral cervical lesion or by the plastic changes taking place in the ipsilesional motor cortex, as measured by the Brinkman board test. In Mk1, there was a slight decrease of the manual dexterity score for a few days following the lesion, before returning to the pre-lesion level. Such a transient effect might reflect a modification of the posture of the animal following the cervical lesion, rather than a direct effect on the contralesional hand.
To further test the motor skill of the forelimbs of Mks 1–3, the three lesioned animals also performed the so-called "reach and grasp drawer" task (see methods). In addition to precision grip skill, the drawer task tests the ability of the monkey to develop force with one or the other forelimb. The analysis of these data demonstrated that there was a deficit (time intervals and their variability were increased) for the ipsilesional arm but not for the contralesional forelimb (not shown). In line with the "Brinkman board" task, the drawer task did not show any deficit of the contralesional hand in relation to the size reduction of the ipsilesional M1 hand area.
Discussion
The present results demonstrate a quite surprising and totally unexpected substantial reduction of the ipsilesional hand representation in M1, as assessed by ICMS, after unilateral section of the CS tract at cervical level. This observation appears robust since it was present in all of the three monkeys examined here. Indeed, the CS undecussated projection originating from the ipsilesional M1 and affected by the unilateral cervical lesion represents only 5–10% of the whole CS tract [38,40,41]. Moreover, the ICMS effects were assessed here only for the contralesional hand (Fig. 2) and thus the unilateral cervical lesion would impact only on the undecussated CS axons that cross the midline at cervical level, representing themselves only a small fraction (about 1/6) of the population of CS undecussated axons (as assessed elegantly by multiple tracing studies in monkeys subjected to cervical hemi-section at C3 level; [39]). In other words, based on these numbers (1/6 of 5–10% of CS axons), one would expect that the unilateral CS tract section would impact only marginally on the ICMS map in the ipsilesional hemisphere. In sharp contrast, the present data show a substantial reduction of the hand area projected on the surface of the ipsilesional hemisphere, amounting to 52%, 77% and 43% in Mk1, Mk2 and Mk3, respectively (Fig. 3), not far from the area reductions observed in the contralesional hemisphere (see [31]), amounting to 67%, 89% and 100% in Mk1, Mk2 and Mk3, respectively. In other words, the reduction of the hand area observed in the ipsilesional hemisphere was thus at least 50 times larger than expected, based on the very small contingent represented by the undecussated CS axons crossing the midline at cervical level.
It is important to stress that the reduction of the M1 hand area in the ipsilesional hemisphere as a result of unilateral cervical lesion in monkeys is an observation made at a specific time point, namely a few months post-lesion after the (incomplete) recovery of the affected hand had reached a plateau. The precise time course of such a reduction of the ipsilesional hand area during the few weeks post-lesion is unknown (in the absence of daily mapping) and one cannot exclude that the ipsilesional hand area was different than it appears after the recovery. Along this line, dynamic bi-hemispheric re-organization of motor networks during the recovery from hemi-paresis caused by corticospinal tract infarction has been observed [35]. Indeed, this study showed that the early recovery of the paretic hand was correlated to a predominant activity on the intact hemisphere but, in later phases of the recovery, the activity in the lesioned hemisphere increased. One cannot exclude such progressive changes of inter-hemispheric balance between the hand areas in the two M1 in our monkeys during the recovery period, leading to the final, fairly balanced size of hand areas between the two hemispheres after recovery reached its maximum.
Comparison with previous work
The present observation of substantial plastic changes of somatotopic maps as a result of a peripheral lesion (at the level of spinal cord, or peripheral nerve lesion or amputation) is in line with an abundant literature on this topic. However, most previous studies in the monkey addressed this issue in the contralesional hemisphere with respect to the spinal cord lesion, either in the somatosensory cortex [29,30,39] or in the motor cortex [31]. In human subjects too, although it is relatively rare to have a cervical cord lesion restricted to one side, most studies aimed at assessing the cortical motor re-organization after spinal cord lesion were focused on the contralesional hemisphere [36,37]. In spinal cord injured patients, the cortical motor map changes consisted mainly of a displacement of the centre of gravity of cortical activity when using the paretic hand after partial recovery, towards a more posterior region [32], which was interpreted as a possible role played by the somatosensory cortex in recovery. In the monkey, as a result of unilateral cervical lesion, in the contralesional hemisphere [31] and in the ipsilesional one (present study), the ICMS data showed a reduction of the hand representation, but no evidence for a posterior shift of the hand representation was found. However, this discrepancy may also be explained by the difficulty to compare directly motor maps based on ICMS in the monkey and on cortical territories activated when performing movements in human. The present observation of a considerable re-organization of the motor map after unilateral cervical cord lesion in the ipsilesional hemisphere is, to our knowledge, an original observation in monkeys. Our finding can only be, to some extent, compared to previous observations in human subjects of functional reorganization in the ipsilesional hemisphere with respect to a cord damage due to lower limb amputation [43].
Interpretation of the motor map changes in the ipsilesional hemisphere after unilateral cervical lesion
How to explain then a plastic change in M1 in the ipsilesional hemisphere, nearly as large as that in the contralesional hemisphere as a result of unilateral section of the CS tract? The reduction of the hand area in the contralesional hemisphere was correlated with anatomical changes such as a shrinkage of the soma of layer V pyramidal neurons, involving the 90% of the axotomized CS neurons [41]. In the ipsilesional hemisphere, we did not observe such shrinkage, as compared with intact animals [41], although this might have been difficult to detect since the unilateral cervical section would affect at most only 5–10% of the CS neurons, giving rise to the undecussated CS axons. In any case, the considerable plastic functional change observed for the hand representation in the ipsilesional hemisphere is not correlated with a major anatomical change (as far as the CS neurons are concerned), in contrast to the contralesional hemisphere. Consequently, the plastic change of motor map in the ipsilesional hemisphere most likely does not result from a direct impact of the axotomized CS tract. One may thus speculate that the size reduction of the hand area ipsilesionally is the result of more indirect (secondary) influences of the lesion. The present data support the notion that a reduction of the hand area in the contralesional hemisphere (which is expected) is accompanied by a nearly comparable reduction in the ipsilesional hemisphere. Although the M1 hand area is quantitatively less connected transcallosally than other body territories in M1 or other motor cortical areas such as SMA [44,45], one may still consider the possibility that the reduction of hand area in the contralesional hemisphere provokes a "secondary" plastic change in the ipsilesional hemisphere, via the callosal projection. The process of secondary change may also occur more indirectly via non-primary motor areas (premotor cortex, SMA), which are more densely connected via the corpus callosum, since lesions of the primary motor cortex induce modifications in the premotor cortex, for instance an extension of the hand area in the ventral premotor cortex [26]. Such "secondary" plastic change in the ipsilesional hemisphere may appear reminiscent to some extent of the transneuronal change observed in the brainstem and thalamus in adult monkeys subjected to long term dorsal rhizotomies [46]. However, a common mechanism is unlikely because, in the case of the rhizotomy there was an anterograde plastic change induced by a lesion, whereas here after cervical lesion the impact on the CS neurons is retrograde. Moreover, a transneuronal degeneration mechanism can be excluded because most axotomized CS neurons survived to the cervical lesion, although they shrank [41].
The extent of the ipsilesional reduction, nearly as large as in the contralesional hemisphere, suggests that such secondary adaptive plastic change may come about as a consequence of the re-balancing of activity in the two hand areas. A roughly balanced hand area in both hemispheres is perhaps more appropriate in the context of bimanual movements as well as in the context of the functional recovery of the ipsilesional hand. Indeed, after unilateral cervical lesion, the recovery of the ipsilesional hand strongly depends on the contralesional hemisphere [31] and not on the ipsilesional one (Fig. 6). If the hand area in the ipsilesional hemisphere had kept its original size after lesion, then there would be a bias in favor of the intact hand, which may be detrimental for mechanisms of recovery of the affected hand. Possibly, recovery may be more efficient if the cortical area responsible for it is not too much reduced in size as compared to its counterpart in the intact hemisphere. However, the hand area in the ipsilesional hemisphere should not be reduced too much either, because this may affect the performance of the contralesional hand. In the present study, as a result of unilateral section of the CS tract at cervical level, the reduction in size in the ipsilesional hemisphere did not affect the performance of the intact hand, at least as assessed by the modified Brinkman board test (Fig. 7) or the "drawer" task. One cannot exclude that a reduction of performance may appear for more challenging tasks, involving more complex synergies of the fingers. Along this line, one may speculate that the re-sizing in the ipsilesional hemisphere should be adjusted in order to reach an ideal compromise, favoring enough the recovery of the affected hand but preserving, as much as possible, the performance of the non-affected hand. The reduction of the hand area in the ipsilesional motor cortex may also be interpreted, at least in part, by postural adjustments as well as a facilitation of those movements not affected by the lesion, such as proximal movements and, but to a lesser extent, the wrist, taking place in the contralesional hemisphere. Such contralesional motor changes, for example comprising strategies of substitution recruited for the recovery, may secondarily induce changes in the ipsilesional hemisphere as well. Postural adjustments may also include the side of the body opposite to the unilateral cervical lesion, resulting in an increased engagement of more proximal muscles at the level of the wrist, elbow and shoulder in the ipsilesional hemisphere, at the expense of the hand representation.
Plastic changes of motor maps resulting from a cortical lesion have been shown to be dependent on the level of rehabilitative training [47,48]. It remains to be determined whether this would also be the case here in the contralesional hemisphere after cervical cord lesion and, if so, whether the same dependence on training would also be present in the ipsilesional hemisphere. In the present study, the monkeys did not undergo a particular and systematic rehabilitative training program, except for the standard behavioral tests they performed every day (essentially the modified Brinkman board test) to assess manual dexterity.
In the contralesional hemisphere, we demonstrated that rapid plastic changes of the hand motor map took place within the first few days post-lesion. During the 2 weeks after the unilateral cervical lesion, no ICMS digit sites were found [31]. Starting about 3 weeks after the lesion, ICMS digit sites progressively re-appeared, to form the stable, reduced hand area observed several months later. Unfortunately, such a repetitive ICMS investigation was not conducted in the ipsilesional hemisphere (because no change in the ipsilesional hemisphere was expected at the time of the experiments). Such a protocol is recommended for future experiments.
Relationship between cortical plasticity and a possible role played for post-lesional recovery
Regarding the mechanisms of the incomplete recovery of the hand affected by the cervical lesion, the present study confirms the notion previously put forward [31] that only the contralesional hemisphere contributes to the recovered performance of the manual dexterity, as assessed by the precision grip task (Brinkman board). Indeed, reversible inactivation of the ipsilesional hemisphere did not modify the recovered manual dexterity score of the affected hand (Fig. 6). Nevertheless, we observed a change of the motor map in the ipsilesional hemisphere. It can thus be concluded that the presence of plastic changes in a certain brain region after a lesion does not necessarily mean that this region contributes significantly to the recovery. In other words, in the debate about whether the intact hemisphere plays a role in the recovery following a unilateral cortical lesion in patients (see e.g. [21-23]), a change of motor map area in the intact hemisphere as compared to normal human subjects should thus not be systematically interpreted as a contribution of the intact hemisphere to the recovery. Clearly, the strategy of reversible inactivation applicable to monkeys, as illustrated in the present study (Fig. 6), remains a better proof than just the observation of motor map changes for the actual involvement of a given brain region in mechanisms of recovery. Another conclusion of the present study is that there is no straightforward relationship between the size of the hand area in M1 (in the ipsilesional hemisphere) and the manual dexterity of the hand controlled mainly by this hemisphere. Indeed, the manual dexterity score was the same pre-lesion with a large hand area and post-lesion with a reduced hand area (Fig. 7). Furthermore, regarding the generation of force, the drawer task did not show a difference of performance pre- and post-lesion. This conclusion, valid for the motor tests used in the present study (Brinkman board and drawer tasks), may not be true for other types, of most likely more complex finger movements, as the activity in the ipsilateral motor cortex is related to the complexity of unimanual hand movements [16].
Conclusion
As a result of unilateral section of the CS tract at cervical level, the hand representation in the contralesional motor cortex was as expected dramatically affected [31]. The present study demonstrates that a substantial post-lesional reduction of the hand representation also took place in the ipsilesional hemisphere, an original observation in the monkey. The considerable extent of the ipsilesional hand representation reduction cannot be explained by a direct effect of the lesion. Indeed, only a small number of transected CS axons originate from the ipsilesional hemisphere and could have contributed to the control of the hand opposite to the lesion by recrossing the midline below the lesion. We therefore propose that the paradoxical reduction of hand representation in the ipsilesional hemisphere is secondary to the changes taking place in the contralesional hemisphere, possibly corresponding to re-adjustments re-establishing a balance between the two hemispheres.
Methods
Overview of the experiments
The surgical procedures (anesthesia, physiological monitoring of the animal, implantation of chronic recording chamber above M1, post-operative care) were described in detail in previous reports from this laboratory [4,5,25,31,41,49]. The experiments were conducted in three young adult male (3–4 years old) Rhesus monkeys (Macaca mulatta), Mk1, Mk2 and Mk3, weighing around 4 kg, and subjected to a unilateral section of the CS tract at cervical level C7/C8. Control experiments to test the reproducibility of the intracortical microstimulation technique were conducted in a fourth, intact monkey (Mk4; Macaca fascicularis, weighing about 3 Kg). Mk1 and Mk2 are the same two animals included in the description of motor maps changes taking place in the contralesional hemisphere [31] and in the anatomical modifications in M1 resulting from the cervical lesion [41]. Mk3 underwent a comparable unilateral cervical lesion as Mk1 and Mk2, but was in addition treated during 4 weeks post-lesion with an antibody aimed at neutralizing the neurite growth inhibitor Nogo (see e.g. [50,51]). The antibody was delivered from an osmotic pump, placed in the back of the animal, using a small silastic tube positioned intrathecally 3–5 mm above the cervical lesion. The effect of the anti-Nogo treatment in Mk3 will be reported elsewhere. Mk1 and Mk2 were also implanted during 4 weeks with an osmotic pump, but delivering a control antibody. Surgical procedures and animal care were conducted in accordance with the Guide for the Care and Use of Laboratory Animals (ISBN 0-309-05377-3; 1996) and approved by local (Swiss) veterinary authorities.
Assessment of manual dexterity
The manual dexterity of each hand was assessed in Mk1, Mk2 and Mk3 using our modified "Brinkman board" task, as described in detail earlier [25,31,52], testing the ability to grasp a food pellet using the opposition of the thumb and the index finger (precision grip). The "Brinkman board" is a Perspex board (10 cm × 20 cm) with 50 randomly distributed holes (15 mm long, 8 mm wide and 6 mm deep) containing each a pellet; 25 holes were oriented horizontally and 25 vertically. The task was performed daily (lasting for 15 to 20 minutes) for several months before and several months after the spinal cord lesion. All sessions were recorded on a video tape and one to two weekly sessions were analyzed quantitatively. An attempt was considered as successful when the monkey grasped a pellet and transported it to the mouth. The manual dexterity was quantitatively measured as the number of slots successfully retrieved within 45 seconds. In addition to the precision grip tested with the "Brinkman board" task, Mks 1–3 were also examined using the so-called "reach and grasp drawer task" [4-6,53-55]. Using one forelimb (unimanual "drawer" task), the monkey had to grasp the knob of a drawer, generate enough force to pull the drawer and, finally, grasp a reward placed inside a well dug in the drawer. By means of different sensors, it was possible to measure several time intervals, separating different epochs of the task. The left and the right forelimbs were tested separately, in series of 20 trials for each forelimb (once a week).
Intracortical microstimulation experiments
The somatotopic organization in and around the hand area in M1 in both hemispheres was established based on daily ICMS sessions, as recently reported [31], using standard parameters of stimulation: 35 ms duration trains of 12 electric monophasic pulses (0.2 ms) presented once every 2 seconds, through a tungsten microelectrode (FHC, Maine, USA) with an impedance of 0.1 – 0.6 MOhms and a tip of about 20–30 μm. Electrode penetrations were oriented nearly perpendicular to the cortical surface, and ICMS was applied at 1 mm steps along the entire track, starting 2 mm below the dura and down to a depth of usually 8–10 mm, sometimes even deeper when the penetration went all the way down to the rostral bank of the central sulcus. On surface ICMS maps (see Figs. 3 and 4), each electrode penetration was represented by a single point corresponding to its position of entry in the brain. ICMS investigation was focused on the hand area with determination of the body territories represented a few mm around the representation of the fingers. The term "hand area" thus refers to the ensemble of ICMS sites in the motor cortex eliciting movements of the fingers observed on the contralateral hand.
The intact monkey Mk4 was included in the present study with the specific aim of assessing the variability of the ICMS method. The hand area of Mk4 was extensively mapped using the ICMS technique as described above. In the left hemisphere, six electrode penetrations selected among different body territories ("face", "wrist" and "digits") were repeated at two time points separated by a time interval ranging between 12 and 140 days.
Reversible inactivation experiments
Mks1-3 were subjected to a unilateral section of the CS tract at C7/C8 level, as described in detail recently [31,41]. Three to five months post-lesion, a time at which the incomplete recovery process had reached a plateau, sessions of reversible inactivation of M1 in either hemisphere using muscimol were conducted, as previously described in detail [25,31,54]. Two to four ICMS penetration sites were chosen in the pre-lesion hand area of M1 in one or the other hemisphere (see syringes in Fig. 3) and the GABA-agonist muscimol (1 μg in 1 μl saline) was infused at two depths along each penetration (separated from each other by 2–3 mm). The infusion of muscimol was performed at ICMS sites at which finger movements were elicited at low threshold and were separated from each other in order to cover the entire hand area, based on previous experiments [25,54]. In Mk1, three and five months post-lesion, two muscimol inactivation sessions were conducted on the left (ipsilesional) hemisphere, 7 weeks apart from each other. In each inactivation session, the total volume of muscimol injected was 18 μl, along three penetrations (Fig. 3, top left panel). In Mk2, only one muscimol inactivation session was conducted on the ipsilesional hemisphere five months post-lesion, in which a total volume of 12 μl of muscimol was infused along two penetrations (Fig. 3, middle left panel). Finally, in Mk3, one reversible inactivation session took place 4 months after the lesion, in which a total volume of 24 μl of muscimol was infused along four penetrations (Fig. 3, bottom left panel). For comparison, in the three monkeys, inactivation sessions were also conducted for the contralesional hemisphere in which muscimol infusion sites were also selected based on ICMS data (not shown; see however [31] for Mk1 and Mk2). The efficacy of such reversible inactivation protocol has been demonstrated previously, together with control experiments in which only saline was injected [54].
List of abbreviations
AR = arcuate sulcus
CE = central sulcus
CMA = cingulate motor area
CS = corticospinal
D = digit
E = elbow
F = face
ICMS = intracortical microstimulation
LH = left hemisphere
M1 = primary motor cortex
Mk = monkey
PM = premotor cortex
RH = right hemisphere
S = shoulder
SMA = supplementary motor area
W = wrist
Authors' contributions
EMR designed the study, helped with the reversible inactivation experiments, to the analysis of the behavioral and ICMS data and drafted the manuscript. ES and TW designed the study and carried out the behavioral, reversible inactivation, ICMS experiments and analyzed the corresponding data, including the histological assessment of the lesion. JB designed the study and performed the cervical lesions. ABS and AFW carried out the experiments related to the reproducibility of the ICMS methods. All authors read and approved the final manuscript.
Acknowledgements
The authors wish to thank the technical assistance of Véronique Moret, Françoise Tinguely and Christine Roulin (histology and behavioral evaluations), Josef Corpataux, Bernard Bapst and Bernard Morandi (animal house keeping), André Gaillard (mechanics), Bernard Aebischer (electronics), Laurent Monney (informatics). Thanks are due to Dr. C. Brown for valuable comments on the manuscript.
Grant Sponsors: Swiss National Science Foundation, grants No 31-43422.95, 4038-43918, 31-61857.00 (EMR); Novartis Foundation; The National Centre of Competence in Research (NCCR) on "Neural plasticity and repair".
Figures and Tables
Figure 1 Location and extent of unilateral cervical lesion in Mks 1–3. Location and extent of the unilateral lesion (grey area), performed at C7/C8 level in Mk1, Mk2 and Mk3, shown on a reconstructed frontal view of cervical cord (three rightmost panels). The inset on the left is a frontal section of the macaque monkey cervical cord showing the distribution of CS axons (black dots) as a result of injection of the anterograde tracer Biotinylated Dextran Amine (BDA) in the right motor cortex (as recently reported in [41]). The CS axons originating from the right hemisphere are distributed mainly in the left dorsolateral funiculus (90–95%), representing the decussated CS tract. The uncrossed CS tract (5–10%) comprises, on the right side, axons in the dorsolateral funiculus and in the ventral funiculus. Scale bars = 2 mm.
Figure 2 Experimental paradigm in Mks1-3. A cervical (left) sub-hemi-section at C7/C8 interrupted the main crossed corticospinal tract originating from the contralesional (right) hemisphere (RH) and the small, uncrossed, corticospinal tract, originating from the ipsilesional (left) hemisphere (LH). In order to assess the functional changes induced by the lesion in the ipsilesional hemisphere, electrode penetrations were performed in order to deliver intracortical microstimulations (ICMS) in the ipsilesional hemisphere. The effects of the ICMS were observed on the contralesional (right) side of the body, with emphasis on the contralesional hand, not affected by the lesion. In the present study, no effects were observed in the ipsilesional hand, affected by the lesion, for ICMS applied to the ipsilesional hemisphere. In other words, all ICMS effects reported in the present study are for the contralesional hand. MN = motoneuron
Figure 3 Intracortical stimulation mapping of the ipsilesional motor cortex. Somatotopic map in the motor cortex on the left (ipsilesional) hemisphere in the region of the hand area, before (left column) and after (right column) the cervical lesion for Mk1, Mk2 and Mk3 (from top to bottom). The data are presented on a surface view of the cerebral cortex, the portion accessible via the chronic recording chamber. The pre-lesion maps were established by daily ICMS sessions conducted during the 3 months preceding the lesion. The post-lesion maps were derived from daily ICMS sessions starting 2, 4.5 and 5 months after the lesion in Mk1, Mk2 and Mk3, respectively and lasting about 2 months each. Each symbol represents the location on the cortical surface of the penetration with an electrode for ICMS. The ICMS data given for each symbol is representative for the site of stimulation where the effect was observed at the lowest current intensity along the considered electrode track. The letter next to each symbol indicates the body territory activated at threshold for each track (see letter codes in the bottom right). The size of the symbols indicates the current intensity at threshold (in μA; see bottom left). Hand area(s), outlined by a solid line, was defined as a cortical region pre-lesion where ICMS at lowest threshold elicits movements of the fingers (electrode tracks represented by diamonds). Symbol X means that ICMS did not elicit any visible movement of muscles. The pointed lines outline the hand area, as defined post-lesion. Symbols in grey are for sites belonging to the hand area before lesion, which became part of other territories post-lesion. The grid in the background indicates steps of 1 mm. Syringes point to sites where muscimol was infused in order to inactivate M1 (see text). In order to make sure that the entire "post-lesion" hand representation was reversibly inactivated, the sites of infusion of muscimol were selected based on the "pre-lesion" map, exhibiting a larger hand representation than post-lesion. On the pre-lesion maps (left column), the approximate positions of the central (CE) and arcuate (AR) sulci are indicated by dashed lines.
Figure 4 Hand representation in the ipsilesional motor cortex. Somatotopic map in the ipsilesional hemisphere as in Fig. 3, but represented are only the electrode penetrations in which digit movements were observed. The gray zones correspond to the hand areas as defined in Figure 3. Outside these areas, other electrode tracks are represented, along which ICMS produced digit movements, but at an intensity that was higher than effects observed for other body territories. Same conventions as in Figure 3.
Figure 5 Reproducibility of the intracortical microstimulation method. To investigate the variability of the method, six ICMS electrode tracks (1 to 6) were each performed at two time points, separated by a time interval indicated in days (12, 21, 125, 140, 26, 138) below the horizontal bar. For each track, the penetration drawn on the left is the first one whereas the second, performed at the same location, is displayed on the right. For each penetration, the tics along the vertical line represent the sites of stimulation, at a distance of 1 mm. Usually, the first stimulation site is located 2 mm below the surface of the dura. The horizontal scale (in μAmps) indicates for each stimulation site the lowest current at which the effect was observed. The body territory activated by the ICMS is indicated as follows: "D" = digit; "E" = elbow; "F" = face; "S" = shoulder; "W" = wrist; "X" = non-microexcitable site. Along each penetration, the arrows (and bold letters) indicate the ICMS threshold for the entire penetration (criterion taken to define the body territory for the corresponding location on the surface map). The bottom panel is a surface representation of the left hemisphere of the intact Mk4, in which the six electrode penetrations represented above were performed twice (indicated by the numbers 1–6). The other circles with the same letter code as above represent the other electrode penetrations performed in the left hemisphere of Mk4. The size of the circles represents the ICMS threshold obtained for the corresponding electrode penetration. CE = central sulcus; AR = arcuate sulcus; rostral is to the left and medial towards top.
Figure 6 Effect of reversible inactivation of the ipsilesional motor cortex on manual dexterity. Effects of reversible inactivation, by infusion of muscimol, of the ipsilesional M1 hand area on the manual dexterity scores for the ipsilesional (left) hand and the contralesional (right) hand. The manual dexterity scores (ordinate) are the number of pellets successively grasped using the precision grip (opposition of thumb and index finger) in the modified Brinkman board task in 45 sec. All inactivation sessions took place post-lesion at time points indicated in the method section. The manual dexterity score was established before ("B") and after ("A") the infusion of muscimol, separately for the vertical and horizontal wells. Two inactivation sessions were conducted in Mk1 and only one in the two other monkeys. The stars are for scores equal to zero, in other words when muscimol infusion completely abolished the corresponding performance. The drop of all scores to zero after infusion ("A" bars) for the right hand demonstrates the efficacy of the inactivation. See text for more detailed description of the results. The star in the session "before muscimol" for Mk2 indicates that this animal did not recover any ability to grasp pellets in the horizontal slots.
Figure 7 Dexterity of the contralesional hand. Manual dexterity of the contralesional hand, assessed with the Brinkman-board task and given by the number of pellets retrieved during 45 sec. (ordinate) as a function of time (abscissa, day of the corresponding session). For each monkey, pre-lesion sessions were performed during the 2 months preceding the lesion (day 0; vertical line), as well as 2 months post-lesion. Squares are for the number of pellets retrieved from the horizontal slots and diamonds from the vertical slots. The triangles are for the sum of the horizontal and vertical slots.
Table 1 ICMS effects in the ipsilesional hemisphere. For each lesioned monkey, the total number of ICMS sites tested is given in the rightmost column, separately for the ICMS sessions pre- and post-lesion. These ICMS sites tested were then distributed in three groups, depending on whether no effect was observed ("non-microexcitable" sites) or elicited movements of the hand ("digit" sites) or movements of other body territories ("other territories"), such activation of wrist, elbow, shoulder or face muscles. Between parentheses, the number of ICMS sites in each group is given in %. For each line, the sum of the three groups is 100%.
Nb. of ICMS "digit" sites Nb. of ICMS "other territories" sites Nb. of "non-microexcitable" sites Total
Mk1
Pre-lesion 79 77 57 213
% (37.1) (36.2) (26.7) (100)
Post-lesion 39 78 82 199
% (19.6) (39.2) (41.2) (100)
Mk2
Pre-lesion 81 72 139 292
% (27.7) (24.7) (47.6) (100)
Post-lesion 29 70 158 257
% (11.3) (27.2) (61.5) (100)
Mk3
Pre-lesion 53 68 150 271
% (19.6) (25.1) (55.3) (100)
Post-lesion 20 37 207 264
(7.6) (14.0) (78.4) (100)
==== Refs
Tanji J Okano K Sato KC Neuronal activity in cortical motor areas related to ipsilateral, contralateral, and bilateral digit movements of the monkey. J Neurophysiol 1988 60 325 343 3404223
Donchin O Gribova A Steinberg O Bergman H Vaadia E Primary motor cortex is involved in bimanual coordination Nature 1998 395 274 278 9751054 10.1038/26220
Donchin O Gribova A Steinberg O Mitz AR Bergman H Vaadia E Single-unit activity related to bimanual arm movements in the primary and supplementary motor cortices J Neurophysiol 2002 88 3498 3517 12466464
Kermadi I Liu Y Tempini A Calciati E Rouiller EM Neuronal activity in the primate supplementary motor area and the primary motor cortex in relation to spatio-temporal bimanual coordination Somatosens Mot Res 1998 15 287 308 9875547 10.1080/08990229870709
Kermadi I Liu Y Rouiller EM Do bimanual motor actions involve the dorsal premotor (PMd), cingulate (CMA) and posterior parietal (PPC) cortices? Comparison with primary and supplementary motor cortical areas Somatosens Mot Res 2000 17 255 271 10994596 10.1080/08990220050117619
Kazennikov O Hyland B Corboz M Babalian A Rouiller EM Wiesendanger M Neural activity of supplementary and primary motor areas in monkeys and its relation to bimanual and unimanual movement sequences Neuroscience 1999 89 661 674 10199603 10.1016/S0306-4522(98)00348-0
Cisek P Crammond DJ Kalaska JF Neural activity in primary motor and dorsal premotor cortex in reaching tasks with the contralateral versus ipsilateral arm J Neurophysiol 2003 89 922 942 12574469
Sato KC Tanji J Digit-muscle responses evoked from multiple intracortical foci in monkey precentral motor cortex J Neurophysiol 1989 62 959 970 2681562
Aizawa H Mushiake H Inase M Tanji J An output zone of the monkey primary motor cortex specialized for bilateral hand movement Exp Brain Res 1990 82 219 221 2257909 10.1007/BF00230856
Kawashima R Roland PE O'Sullivan BT Activity in the human primary motor cortex related to ipsilateral hand movements Brain Res 1994 663 251 256 7874508 10.1016/0006-8993(94)91270-X
Kawashima R Matsumura M Sadato N Naito E Waki A Nakamura S Matsunami K Fukuda H Yonekura Y Regional cerebral blood flow changes in human brain related to ipsilateral and contralateral complex hand movements - a PET-study Eur J Neurosci 1998 10 2254 2260 9749754 10.1046/j.1460-9568.1998.00237.x
Wassermann EM Pascual-Leone A Hallett M Cortical motor representation of the ipsilateral hand and arm Exp Brain Res 1994 100 121 132 7813640 10.1007/BF00227284
Chen R Cohen LG Hallett M Role of the ipsilateral motor cortex in voluntary movement Can J Neurol Sci 1997 24 284 291 9398974
Chen R Gerloff C Hallett M Cohen LG Involvement of the ipsilateral motor cortex in finger movements of different complexities Ann Neurol 1997 41 247 254 9029074 10.1002/ana.410410216
Chen R Yung D Li JY Organization of ipsilateral excitatory and inhibitory pathways in the human motor cortex J Neurophysiol 2003 89 1256 1264 12611955
Tinazzi M Zanette G Modulation of ipsilateral motor cortex in man during unimanual finger movements of different complexities Neurosci Lett 1998 244 121 124 9593504 10.1016/S0304-3940(98)00150-5
Cramer SC Finklestein SP Schaechter JD Bush G Rosen BR Activation of distinct motor cortex regions during ipsilateral and contralateral finger movements J Neurophysiol 1999 81 383 387 9914297
Sunderland A Bowers MP Sluman SM Wilcock DJ Ardron ME Impaired dexterity of the ipsilateral hand after stroke and the relationship to cognitive deficit Stroke 1999 30 949 955 10229726
Sunderland A Recovery of Ipsilateral Dexterity After Stroke Stroke 2000 31 430 433 10657418
Muellbacher W Facchini S Boroojerdi B Hallett M Changes in motor cortex excitability during ipsilateral hand muscle activation in humans Clin Neurophysiol 2000 111 344 349 10680571 10.1016/S1388-2457(99)00243-6
Porro CA Cettolo V Francescato MP Baraldi P Ipsilateral involvement of primary motor cortex during motor imagery Eur J Neurosci 2000 12 3059 3063 10971647 10.1046/j.1460-9568.2000.00182.x
Zaaroor M Pratt H Starr A Influence of task-related ipsilateral hand movement on motor cortex excitability Clin Neurophysiol 2001 112 908 916 11336908 10.1016/S1388-2457(01)00516-8
Ziemann U Hallett M Hemispheric asymmetry of ipsilateral motor cortex activation during unimanual motor tasks: further evidence for motor dominance Clin Neurophysiol 2001 112 107 113 11137667 10.1016/S1388-2457(00)00502-2
Rau C Plewnia C Hummel F Gerloff C Event-related de synchronization and excitability of the ipsilateral motor cortex during simple self-paced finger movements Clin Neurophysiol 2003 114 1819 1826 14499743 10.1016/S1388-2457(03)00174-3
Sohn YH Jung HY Kaelin-Lang A Hallett M Excitability of the ipsilateral motor cortex during phasic voluntary hand movement Exp Brain Res 2003 148 176 185 12520405
Yarosh CA Hoffman DS Strick PL Deficits in movements of the wrist ipsilateral to a stroke in hemiparetic subjects J Neurophysiol 2004 92 3276 3285 15295013 10.1152/jn.00549.2004
Verstynen T Diedrichsen J Albert N Aparicio P Ivry RB Ipsilateral motor cortex activity during unimanual hand movements relates to task complexity J Neurophysiol 2005 93 1209 1222 15525809 10.1152/jn.00720.2004
Chollet F DiPiero V Wise RJS Brooks DJ Dolan RJ Frackowiak RSJ The functional anatomy of motor recovery after stroke in humans: A study with positron emission tomography Ann Neurol 1991 29 63 71 1996881 10.1002/ana.410290112
Weiller C Chollet F Friston KJ Wise RJS Frackowiak RSJ Functional reorganization of the brain in recovery from striatocapsular infarction in man Annals of Neurology 1992 31 463 472 1596081 10.1002/ana.410310502
Cramer SC Nelles G Benson RR Kaplan JD Parker RA Kwong KK Kennedy DN Finklestein SP Rosen BR A functional MRI study of subjects recovered from hemiparetic stroke Stroke 1997 28 2518 2527 9412643
Honda M Nagamine T Fukuyama H Yonekura Y Kimura J Shibasaki H Movement-related cortical potentials and regional cerebral blood flow change in patients with stroke after motor recovery J Neurol Sci 1997 146 117 126 9077507 10.1016/S0022-510X(96)00291-2
Cao Y D'Olhaberriague L Vikingstad EM Levine SR Welch KM Pilot study of functional MRI to assess cerebral activation of motor function after poststroke hemiparesis Stroke 1998 29 112 122 9445338
Johansen-Berg H Rushworth MFS Bogdanovic MD Kischka U Wimalaratna S Matthews PM The role of ipsilateral premotor cortex in hand movement after stroke Proceedings of the National Academy of Sciences of the United States of America 2002 99 14518 14523 12376621 10.1073/pnas.222536799
Lemon RN Cortical plasticity: Stroke recovery Curr Biol 1993 3 463 465 15335717 10.1016/0960-9822(93)90358-U
Turton A Wroe S Trepte N Fraser C Lemon RN Contralateral and ipsilateral EMC responses to transcranial magnetic stimulation during recovery of arm and hand function after stroke Electroencephalogr Clin Neurophysiol Electromyogr Motor Control 1996 101 316 328 10.1016/0924-980X(96)95560-5
Hallett M Plasticity of the human motor cortex and recovery from stroke Brain Res Rev 2001 36 169 174 11690613 10.1016/S0165-0173(01)00092-3
Seitz RJ Hoflich P Binkofski F Tellmann L Herzog H Freund HJ Role of the premotor cortex in recovery from middle cerebral artery infarction Arch Neurol 1998 55 1081 1088 9708958 10.1001/archneur.55.8.1081
Liu Y Rouiller EM Mechanisms of recovery of dexterity following unilateral lesion of the sensorimotor cortex in adult monkeys Exp Brain Res 1999 128 149 159 10473753 10.1007/s002210050830
Frost SB Barbay S Friel KM Plautz EJ Nudo RJ Reorganization of remote cortical regions after ischemic brain injury: A potential substrate for stroke recovery J Neurophysiol 2003 89 3205 3214 12783955
Werhahn KJ Conforto AB Kadom N Hallett M Cohen LG Contribution of the ipsilateral motor cortex to recovery after chronic stroke Ann Neurol 2003 54 464 472 14520658 10.1002/ana.10686
Fridman EA Hanakawa T Chung M Hummel F Leiguarda RC Cohen LG Reorganization of the human ipsilesional premotor cortex after stroke Brain 2004 127 747 758 14749291 10.1093/brain/awh082
Jain N Catania KC Kaas JH Deactivation and reactivation of somatosensory cortex after dorsal spinal cord injury Nature 1997 386 495 498 9087408 10.1038/386495a0
Jain N Florence SL Kaas JH Reorganization of somatosensory cortex after nerve and spinal cord injury News Physiol Sci 1998 13 143 149 11390778
Schmidlin E Wannier T Bloch J Rouiller EM Progressive plastic changes in the hand representation of the primary motor cortex parallel incomplete recovery from a unilateral section of the corticospinal tract at cervical level in monkeys Brain Research 2004 1017 172 183 15261113 10.1016/j.brainres.2004.05.036
Green JB Sora E Bialy Y Ricamato A Thatcher RW Cortical sensorimotor reorganization after spinal cord injury - An electroencephalographic study Neurology 1998 50 1115 1121 9566404
Green JB Sora E Bialy Y Ricamato A Thatcher RW Cortical motor reorganization after paraplegia - An EEG study Neurology 1999 53 736 743 10489034
Puri BK Smith HC Cox IJ Sargentoni J Savic G Maskill DW Frankel HL Ellaway PH Davey NJ The human motor cortex after incomplete spinal cord injury: an investigation using proton magnetic resonance spectroscopy J Neurol Neurosurg Psychiatry 1998 65 748 754 9810950
Lotze M Laubis-Herrmann U Topka H Erb M Grodd W Reorganization in the primary motor cortex after spinal cord injury - A functional Magnetic Resonance (fMRI) Study Restor Neurol Neurosci 1999 14 183 187 12671262
Marshall RS Perera GM Lazar RM Krakauer JW Constantine RC DeLaPaz RL Evolution of cortical activation during recovery from corticospinal tract infarction Stroke 2000 31 656 661 10700500
Mikulis DJ Jurkiewicz MT McIlroy WE Staines WR Rickards L Kalsi-Ryan S Crawley AP Fehlings MG Verrier MC Adaptation in the motor cortex following cervical spinal cord injury Neurology 2002 58 794 801 11889245
Rocca MA Mezzapesa DM Ghezzi A Falini A Agosta F Martinelli V Scotti G Comi G Filippi M Cord damage elicits brain functional reorganization after a single episode of myelitis Neurology 2003 61 1078 1085 14581668
Turner JA Lee JS Schandler SL Cohen MJ An fMRI investigation of hand representation in paraplegic humans Neurorehabil Neural repair 2003 17 37 47 12645444 10.1177/0888439002250443
Rouiller EM Moret V Tanné J Boussaoud D Evidence for direct connections between the hand region of the supplementary motor area and cervical motoneurons in the macaque monkey Eur J Neurosci 1996 8 1055 1059 8743753
Galea MP Darian-Smith I Corticospinal projection patterns following unilateral section of the cervical spinal cord in the newborn and juvenile macaque monkey J Comp Neurol 1997 381 282 306 9133569 10.1002/(SICI)1096-9861(19970512)381:3<282::AID-CNE3>3.0.CO;2-Z
Lacroix S Havton LA McKay H Yang H Brant A Roberts J Tuszynski MH Bilateral corticospinal projections arise from each motor cortex in the macaque monkey: A quantitative study J Comp Neurol 2004 473 147 161 15101086 10.1002/cne.20051
Wannier T Schmidlin E Bloch J Rouiller EM A unilateral section of the corticospinal tract at cervical level in primates does not lead to measurable cell loss in motor cortex J Neurotrauma 2005
Schwenkreis P Pleger B Cornelius B Weyen U Dertwinkel R Zenz M Malin JP Tegenthoff M Reorganization in the ipsilateral motor cortex of patients with lower limb amputation Neurosci Lett 2003 349 187 190 12951200 10.1016/S0304-3940(03)00838-3
Jenny AB Commissural projections of the cortical hand motor area in monkeys. J Comp Neurol 1979 188 137 146 115906 10.1002/cne.901880111
Rouiller EM Babalian A Kazennikov O Moret V Yu XH Wiesendanger M Transcallosal connections of the distal forelimb representations of the primary and supplementary motor cortical areas in macaque monkeys Exp Brain Res 1994 102 227 243 7705502 10.1007/BF00227511
Woods TM Cusick CG Pons TP Taub E Jones EG Progressive transneuronal changes in the brainstem and thalamus after long-term dorsal rhizotomies in adult macaque monkeys J Neurosci 2000 20 3884 3899 10804228
Nudo RJ Milliken GW Reorganization of movement representations in primary motor cortex following focal ischemic infarcts in adult squirrel monkeys J Neurophysiol 1996 75 2144 2149 8734610
Nudo RJ Wise BM SiFuentes F Milliken GW Neural substrates for the effects of rehabilitative training on motor recovery after ischemic infarct Science 1996 272 1791 1794 8650578
Liu J Morel A Wannier T Rouiller EM Origins of callosal projections to the supplementary motor area (SMA): A direct comparison between pre-SMA and SMA-proper in macaque monkeys J Comp Neurol 2002 443 71 85 11793348 10.1002/cne.10087
Schwab ME Repairing the injured spinal cord Science 2002 295 1029 1031 11834824 10.1126/science.1067840
Schwab ME Nogo and axon regeneration Curr Opin Neurobiol 2004 14 118 124 15018947 10.1016/j.conb.2004.01.004
Rouiller EM Yu XH Moret V Tempini A Wiesendanger M Liang F Dexterity in adult monkeys following early lesion of the motor cortical hand area: the role of cortex adjacent to the lesion Eur J Neurosci 1998 10 729 740 9749734 10.1046/j.1460-9568.1998.00075.x
Kazennikov O Wicki U Corboz M Hyland B Palmeri A Rouiller EM Wiesendanger M Temporal structure of a bimanual goal-directed movement sequence in monkeys Eur J Neurosci 1994 6 203 210 8167842
Kermadi I Liu Y Tempini A Rouiller EM Effects of reversible inactivation of the supplementary motor area (SMA) on unimanual grasp and bimanual pull and grasp performance in monkeys Somatosens Mot Res 1997 14 268 280 9443367 10.1080/08990229770980
Wannier T Liu J Morel A Jouffrais C Rouiller EM Neuronal activity in primate striatum and pallidum related to bimanual motor actions NeuroReport 2002 13 143 147 11924876 10.1097/00001756-200201210-00033
|
16135243
|
PMC1224856
|
CC BY
|
2021-01-04 16:03:48
|
no
|
BMC Neurosci. 2005 Aug 31; 6:56
|
utf-8
|
BMC Neurosci
| 2,005 |
10.1186/1471-2202-6-56
|
oa_comm
|
==== Front
Biomed Eng OnlineBioMedical Engineering OnLine1475-925XBioMed Central London 1475-925X-4-491611532410.1186/1475-925X-4-49ResearchValidity and reproducibility of arterial pulse wave velocity measurement using new device with oscillometric technique: A pilot study Naidu Madireddy Umamaheshwar Rao [email protected] Budda Muralidhar [email protected] Sridhar [email protected] Amar Narayana [email protected] Pingali Usha [email protected] Department of Clinical Pharmacology & Therapeutics, Nizam's Institute of Medical Sciences, Punjagutta, Hyderabad, 500082, India2 Department of Cardiology, Nizam's Institute of Medical Sciences, Punjagutta, Hyderabad, 500082, India2005 23 8 2005 4 49 49 3 6 2005 23 8 2005 Copyright © 2005 Naidu et al; licensee BioMed Central Ltd.2005Naidu et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Availability of a range of techniques and devices allow measurement of many variables related to the stiffness of large or medium sized arteries. There is good evidence that, pulse wave velocity is a relatively simple measurement and is a good indicator of changes in arterial properties. The pulse wave velocity calculated from pulse wave recording by other methods like doppler or tonometry is tedious, time-consuming and above all their reproducibility depends on the operator skills. It requires intensive resource involvement. For epidemiological studies these methods are not suitable. The aim of our study was to clinically evaluate the validity and reproducibility of a new automatic device for measurement of pulse wave velocity that can be used in such studies.
Methods
In 44 subjects including normal healthy control and patients with coronary artery disease, heart brachial, heart ankle, brachial ankle and carotid femoral pulse wave velocities were recorded by using a new oscillometric device. Lead I and II electrocardiogram and pressure curves were simultaneously recorded. Two observers recorded the pulse wave velocity for validation and one observer recorded the velocity on two occasions for reproducibility.
Results and Discussion
Pulse wave velocity and arterial stiffness index were recorded in 24 control and 20 coronary artery disease patients. All the velocities were significantly high in coronary artery disease patients. There was highly significant correlation between the values noted by the two observers with low standard deviation. The Pearson's correlation coefficient for various velocities ranged from (r = 0.88–0.90) with (p < 0.0001). The reproducibility was also very good as shown by Bland-Altman plot; most of the values were lying within 2 SD. The interperiod measurements of pulse wave velocity were also significantly correlated (r = 0.71 – 0.98) (P < 0.0001). Carotid-femoral pulse wave velocity was found to correlate significantly with heart brachial, heart ankle, brachial ankle pulse wave velocity and arterial stiffness index values. Reproducibility of our method was good with very low variability in both interobserver and interperiod analysis.
Conclusion
The new device "PeriScope" based on oscillometric technique has been found to be a simple, non-invasive and reproducible device for the assessment of pulse wave velocity and can be used to determine arterial stiffness in large population based studies.
Pulse wave velocityArterial stiffness indexValidityOscillometric device
==== Body
Background
Interest in, and measurement of the velocity of arterial wave propagation as an index of vascular stiffness and vascular health dates back to the early part of the last century. Many methodologies; both invasive and non-invasive have been applied to the assessment of arterial elasticity in vivo. The assessment of cardiovascular risk is one of the most important tasks. The predictive value of pulse wave velocity (PWV) is becoming increasingly recognized and it is one of the classical indices of arterial stiffness (AS). Several arterial assessments improve risk stratification. Three of the most cost-effective assessment parameters are pulse pressure, arterial stiffness and ankle brachial index. Arterial stiffness can be directly measured by non-invasive techniques like computerized oscillometry, tonometry and ultrasonography [1].
In one study [1] of 30 patients with and without cardiovascular disease (CVD), the diagnostic accuracies of these techniques were as follows: arterial stiffness 85%, pulse pressure 71% and endothelial function 58%. Pulse pressure is independently related to all cause mortality, but only marginally related to cardiovascular mortality, indicating that specific assessment of AS with PWV, may be of greater value in the evaluation of risk [1]. PWV measurements by ultrasonography are more time consuming, requires substantial training and dedicated staff.
A recent important observation, that AS is an independent predictor of cardiovascular mortality [2-4], has gained greater interest. Increased stiffness may precede the onset of clinically overt atheromatous disease [5]. Early identification of individuals at risk, by improved detection of changes in stiffness may help in providing beneficial intervention at an early stage [6]. Relevance and implications of possible therapy of arterial stiffness have led to wide spread interest in its measurement, resulting in large number of availability of commercial devices [7]. There is no gold standard to assess AS [8].
Recently non-invasive methods to measure AS have become available and are relatively easy to perform [9-12]. A simple device to measure PWV has been developed which measures brachial ankle pulse wave velocity (ba PWV) using an oscillometric method [13]. As it measures peripheral artery velocity, unlike aortic PWV the clinical significance may differ. Aortic PWV >13 m/s is a particularly strong predictor of cardiovascular mortality in hypertension [14]. Although ba PWV measurement is simple and non-invasive, the sensitivity and specificity in predicting coronary artery disease were only 62% and 29% [15]. A more accurate and integrated information of AS may therefore be obtained by using a combination of techniques based on different models [7]. There is a need to quantify the extent to which measures of AS can improve risk stratification and to determine whether its reduction is capable of predicting clinical benefit. In the present study, we have assessed a newly developed simple non-invasive device, which measures both peripheral and aortic PWV using an oscillometric method.
Methods
Total twenty-four male healthy subjects with mean age 24.7 ± 5.5 years, mean height 168.3 ± 5 cm and mean weight 52.6 ± 6.4 kgs were recruited for interperiod and interobserver reproducibility study. All the subjects gave their consent for the study, approved by the Institutional Ethics Committee, Nizam's Institute of Medical Sciences, Hyderabad. Before the inclusion in the study, subjects were thoroughly examined clinically to confirm the inclusion criteria. No subject was on any medication and had normal blood pressure, CVS, renal, hepatic functions, blood sugar, serum cholesterol and uric acid levels.
All subjects refrained from smoking and caffeine containing beverages at least 24 hours before the measurements. The entire test was performed in the morning, after 10 hours over night fasting condition in a quite room with controlled temperature. In each subject two sequences of measurements were performed, and their mean was considered for analysis. All procedures were repeated by two observers (observer 1 and observer 2) for analysis of the inter-observer reproducibility, between the two sequences of measurements of PWV, the BP cuffs were rewrapped at each measurement. For inter period reproducibility assessment, PWV was measured twice by the same observer with an interval of at least one day between the two measurements.
Additionally 20 patients (9 male & 11 female) with CAD having mean age 50 ± 13 years, mean height 157.9 ± 10 cm, mean weight 67 ± 16 kgs were also included in interobserver study from the OPD. History of diabetes mellitus, hypertension, smoking, and drug use were recorded. All patients had coronary angiographically confirmed CAD. Patients were allowed to take their prescribed medication during the study period. All patients gave their consent for study participation. Every time the PWV recording was carried out at least 10 minutes after resting.
The subjects were examined in supine position. Electrodes of electrocardiogram were placed on ventral surface of both wrists and medial side of ankles and BP cuffs were wrapped on both upper arm brachial artery and above tibial artery of ankles. The cuffs were connected to a plethysmographic sensor that determines volume pulse form and an oscillometric pressure sensor that measures blood pressure volume waveforms from the brachial and tibial arteries. All the pressure recordings were done for about 10 seconds and data was stored in the computer for analysis.
New device (PeriScope, developed by Genesis Medical Systems, Hyderabad, India) is an 8-channel real-time PC based simultaneous acquisition and analysis system. The acquisition rate is 200 samples/second, which is sufficient because the significant frequency content of the pressure as well as ECG waveform is not more than 40 Hz. According to Nyquist's criteria the minimum sampling rate should be 80 samples/second. Hence a sampling rate of 200 Hz per second is optimum. It supports a sophisticated digital signal-processing algorithm to calculate all the results. System has dedicated hardware module connected to 4 ECG electrodes and 4 Blood pressure measuring cuffs. It is very user friendly and fully automatic. Once started, the test recording completes itself by displaying results directly. The report contains 8-second traces of Lead I and II ECG, all Pressure Pulse Waveforms and all calculated results. Device has a built-in database that can be used to store patient folders for further referrals at any point of time. PeriScope is a PC based low cost instrument. When used with a laptop, it can be carried to remote locations. It uses of ECG as marker. It does not use phonocardiogram. PeriScope thus facilitates use in epidemiological studies.
Calculation of pulse wave velocity by oscillometric method
This method measures the blood pressure by detecting the pulsation of the artery, which is caused by the heart, as the pressure oscillation in the cuff. When the cuff around< the upper arm is fully inflated, blood flow stops but pulsation of the artery continues and causes oscillation of the pressure in the cuff. As the pressure in the cuff is decreased slowly, the amplitude of the pressure oscillation in the cuff gradually increases and eventually reaches to a peak. Further decrease of the cuff pressure causes the oscillation amplitude to decrease. Cuff pressure when the oscillation reaches a peak, is taken as the mean arterial pressure (MAP).
Figure 1 shows the oscillometric pressure pattern. As this maximum amplitude oscillation in each limb is detected, the pulse waveform along with ECG Lead I and II are stored simultaneously in the PC memory. These waveforms are used to detect various pulse wave velocities as described below.
Figure 1 Computerized oscillomertry. Atypical computerized Oscillometric pattern of pressure.
Pulse wave velocity is the speed at which the blood pressure pulse travels from the heart to the peripheral artery after blood rushes out during contraction. It is mainly used to evaluate stiffness of the artery wall. Pulse wave velocity increases with stiffness of the arteries.
With new device, the PTT (Pulse Transit Time) of each segment is calculated from the waveform taken from each sensor.
It calculates heart-brachial PWV of both upper limbs, heart-ankle PWV of both lower limbs, brachial-ankle PWV of both right and left limb pairs and effective estimated carotid-femoral PWV is calculated.
Where Lha = Distance between heart and respective ankle.
Lhb = Distance between heart and respective brachium.
Lba = Distance between respective brachium and ankle.
Distances were measured by direct superficial measurement with a measuring tape of 1 mm resolution as follows:
Heart to brachial distance = Heart to shoulder + shoulder to midpoint of brachial cuff.
Heart to ankle distance = Heart to midpoint of ankle cuff.
Brachial ankle distance = Heart to brachial distance + Heart to ankle distance
Pulse Transit Time (PTT) between heart and extremity (i.e. PTT hb and PTT ha) is calculated using R wave as pulse start maker and maximum pressure gradient as pressure wave arrival marker. This is due to the fact that the QRS complex designates highest-pressure deviation in the left ventricle where as maximum pressure gradient marks the highest-pressure deviation in the artery of the extremity (Fig. 2). A proprietary software algorithm is used to analyze the ECG and pressure waveforms. From this analysis the pre-ejection period is calculated and deducted from each "R-wave to max. pressure gradient" time. This gives the actual pulse transit time.
Figure 2 Pulse Wave Velocity calculations. Reference points of ECG and four pressure waveforms consider for calculation of pulse transit time and PWV.
Pulse Transit Time (PTT) between brachium and respective ankle is calculated as the time difference between the feet of respective pulse wave originated by the same QRS complex. The carotid femoral PWV (C-F PWV) is estimated from the composite brachial ankle pulse wave velocity (ba PWV) found out by averaging left and right ba PWV. The regression analysis between ba PWV and C-F PWV yields the following equation:
Estimated carotid femoral PWV = 0.8333 * (Avg. ba PWV) - 233.33.
This equation is arrived at by internal data collection and confirmed by studies conducted elsewhere [13].
Derivation of oscillometric envelopes
Oscillometric envelope is a graphical depiction of compressibility of the artery. It is derived from the oscillations in the artery when the BP cuffs are deflating while taking the BP reading. It is the graph of amplitude of oscillations verses the instantaneous pressure in the blood pressure cuff. In a normal arterial condition, the shape of the oscillometric envelope is like a bell, when arteries are stiffened or atherosclerosed the oscillometric envelope flattens out.
Calculation of arterial stiffness index
Arterial stiffness index is another measure of arterial stiffness. It quantifies the shape of the oscillometric envelope. As the arterial stiffness increases, it becomes harder to collapse the arteries by applying external pressure; hence the oscillometric envelope becomes flatter as the stiffness increases. The ASI value gives a clear indication of this flattening process (Fig. 3). The higher the stiffness, the higher the ASI values. It is calculated as:
Figure 3 Oscillometric Envelopes.
ASI = (Systolic side Value of cuff pressure at 80% of maximal oscillation amplitude of cuff) - (Diastolic side Value of cuff pressure at 80% of maximal oscillation amplitude of cuff).
The device acquires and computes the following parameters simultaneously: 2 channels of ECG, brachial BP of both limbs, ankle BP of both limbs, 4 channels of pulse pressure waveforms, mean arterial pressure (MAP), % MAP, ABI, pulse wave velocities, UT (Upstroke Time), arterial stiffness index.
Statistics
Data are expressed as mean ± SD. Pearson's correlation analysis, determination of coefficient of variation and Bland-Altman plotting were performed for the assessment of validity and reproducibility [21]. When two series of paired measurements were compared, the results were analyzed in two steps according to the recommendations of Bland and Altman. First, the correlation between measurement values (equation of the linear relationship, correlation coefficient r, and P value) was investigated. The first step was used to gauge the degree of agreement between two series of measurements. Second, the relative (positive or negative) differences between each pair of measures were plotted against the mean of the pair to make sure that no obvious relation appeared between the estimated value mean and difference. The mean difference and the SD of the differences estimated the lack of agreement between the two measurements. Unpaired student 't' test was used for comparisons among the two groups. Values of p < 0.05 was considered to indicate statistical significance. All the statistical analysis was performed using the Graph pad PRISM software version 4 (Graph pad software Inc. San Diego, California, USA).
Results
Total 44 subjects (24) healthy controls and (20) CAD patients participated in the present study. The demographic characteristic of all subjects is shown in the table-1. As our aim was to study the reproducibility and validity of a new method, we have selected both healthy controls and CAD patients. The control group had lower mean age than the CAD group; similarly the BMI was also high in CAD group. As compared to control group, all haemodynamic parameters HR, SBP, DBP, PP and MAP were also significantly higher in CAD group.
Table 1 Demographic characteristics of healthy subjects and patients with coronary artery disease (CAD) studied.
All Subjects Healthy subjects CAD patients
Number 44 24 20
Sex (male/female) 33/11 24/0 9/11
Age (yrs) 37.09 ± 16.69 24.67 ± 4.65 52.00 ± 13.22***
Weight (Kgs) 59.18 ± 13.73 52.67 ± 6.46 67.00 ± 16.04##
Height (cms) 163.5 ± 9.342 168.3 ± 5.18 157.9 ± 10.14***
BMI (Kg/m2) 22.26 ± 5.331 18.65 ± 2.46 26.59 ± 4.54***
Heart Rate (bpm) 71.13 ± 15.69 62.33 ± 7.56 81.80 ± 16.47***
Systolic BP (mmHg) 106.4 ± 13.46 110.3 ± 8.8 131.9 ± 21.95*
Diastolic BP (mmHg) 58.85 ± 12.99 63.08 ± 6.6 79.15 ± 11.38**
MAP (mmHg) 82.78 ± 9.68 85.1 ± 6.64 100.9 ± 15.8#
Pulse Pressure 46.64 ± 11.77 47.17 ± 5.59 52.7 ± 16.08***
Values are expressed as Mean ± SD.
*** p < 0.0001, ** p < 0.0003, * p < 0.01, # #p < 0.0002 and # p < 0.02 Vs Healthy subjects.
The mean PWV obtained in two groups is shown in table 2. In CAD group hb PWV, ha PWV, ba PWV and C-F PWV were found to be significantly higher than control. The mean ankle ASI was 60.6 ± 16.7 in CAD group and 48.2 ± 7.6 in control. This difference in ankle ASI was significant (p < 0.001).
Table 2 Pulse Wave Velocity and Arterial Stiffness Index.
Pulse Wave Velocity All Subjects Healthy subjects CAD patients
Number of observations 88 48 40
Heart Brachial PWV
Mean 279.0 262.4 298.9***
SD 45.10 17.03 58.67
SE 4.808 2.459 9.276
Heart Ankle PWV
Mean 482.8 425.1 551.9***
SD 85.97 17.28 84.50
SE 9.165 2.494 13.36
Brachial Ankle PWV
Mean 1392 1115 1725***
SD 397.3 103.8 362.1
SE 42.36 14.99 57.25
Carotid Femoral PWV
Mean 930.7 702.7 1204***
SD 329.2 90.54 301.8
SE 35.09 13.07 47.72
Brachial ASI
Mean 38.08 37.41 38.89
SD 10.23 4.373 14.46
SE 1.090 0.6312 2.287
Ankle ASI
Mean 53.84 48.22 60.59***
SD 13.97 7.601 16.73
SE 1.489 1.097 2.646
*** p < 0.0001 Vs Healthy subjects
The mean velocities obtained by the two separate observers in control and CAD group are shown in tables 3, 4 and 5. Values of all parameters obtained by two observers were found to be highly correlated with significant Pearson's correlation coefficients.
Table 3 Means and correlational analysis (Pearson r value) of arterial stiffness among healthy subjects (n = 24)
Pulse Wave Velocity Observer I Observer II Pearson (r) 95% CI P value
Heart Brachial PWV 262.4 ± 16.09 262.4 ± 18.28 0.88 0.75 to 0.95 p < 0.0001
Heart Ankle PWV 425.4 ± 17.96 424.9 ± 16.96 0.91 0.80 to 0.96 p < 0.0001
Brachial Ankle PWV 1124 ± 104 1106 ± 105.2 0.89 0.76 to 0.95 p < 0.0001
Carotid Femoral PWV 706.2 ± 90.36 699.2 ± 92.53 0.90 0.79 to 0.96 p < 0.0001
Brachial ASI 37.80 ± 3.93 37.01 ± 4.83 0.62 0.28 to 0.82 p < 0.001
Ankle ASI 48.39 ± 8.54 48.06 ± 6.71 0.85 0.69 to 0.94 p < 0.0001
Values are expressed as Mean ± SD
Table 4 Means and correlational analysis (Pearson r value) of arterial stiffness among CAD Patients (n = 20)
Pulse Wave Velocity Observer I Observer II Pearson (r) 95% CI P value
Heart Brachial PWV 298.9 ± 59.26 298.9 ± 59.61 0.99 0.97 to 0.99 p < 0.0001
Heart Ankle PWV 550.3 ± 83.60 553.5 ± 87.53 0.98 0.96 to 0.99 p < 0.0001
Brachial Ankle PWV 1719 ± 368.8 1731 ± 364.8 0.99 0.98 to 0.99 p < 0.0001
Carotid Femoral PWV 1199 ± 307.4 1210 ± 304 0.98 0.97 to 0.99 p < 0.0001
Brachial ASI 39.15 ± 14.21 38.63 ± 15.07 0.93 0.85 to 0.97 p < 0.0001
Ankle ASI 60.48 ± 17.77 60.70 ± 16.09 0.95 0.88 to 0.98 p < 0.0001
Values are expressed as Mean ± SD
Table 5 Means and correlational analysis (Pearson r value) of arterial stiffness among all subjects (Healthy + CAD) (n = 44)
Pulse Wave Velocity Observer I Observer II Pearson (r) 95% CI P value
Heart Brachial PWV 279.0 ± 45.05 279.0 ± 45.67 0.98 0.96 to 0.99 p < 0.0001
Heart Ankle PWV 482.2 ± 84.97 483.3 ± 87.94 0.99 0.98 to 0.99 p < 0.0001
Brachial Ankle PWV 1394 ± 394.7 1390 ± 404.5 0.99 0.98 to 0.99 p < 0.0001
Carotid Femoral PWV 940.7 ± 331.7 941 ± 338.3 0.99 0.98 to 0.99 p < 0.0001
Brachial ASI 38.41 ± 9.8 37.74 ± 10.65 0.91 0.85 to 0.95 p < 0.0001
Ankle ASI 53.88 ± 14.69 53.81 ± 13.38 0.94 0.90 to 0.96 p < 0.0001
Values are expressed as Mean ± SD
In the control group, the Pearson's correlation coefficient was 0.88, 0.91, 0.89 and 0.90 (p < 0.0001) for hb, ha, ba and C-F PWV respectively. The coefficient was 0.85 for ankle ASI with p < 0.0001. Similarly in CAD group, and in data from all subjects combined also the correlation was highly significant, suggesting high reproducibility of our method for PWV determination.
The relationship and Bland-Altman plot for hb, ha, ba and C-F PWV obtained by two observers is shown in figures 4, 5, 6 and 7 respectively. A very good correlation was observed between the two measurements (r = 0.98 – 0.99) in all these above parameters. There was no trend for the reproducibility of measurements to vary with the underlying mean values in any of the parameters analyzed. In the Bland-Altman Plots of interobserver measurements of PWV, there was no significant difference in the values for reproducibility reported between observers and most of the values ranged within a mean ± 2 SD.
Figure 4 Interobserver Heart Brachial PWV. Relationship between two independent measurements of hb PWV by two observers (Upper panel). Reproducibility of hb PWV. Bland-Altman plot showing observer difference in measurements. (Lower panel).
Figure 5 Interobserver Heart Ankle PWV. Relationship between two independent measurements of ha PWV by two observers (Upper panel). Reproducibility of ha PWV. Bland-Altman plot showing observer difference in measurements. (Lower panel).
Figure 6 Interobserver Brachial Ankle PWV. Relationship between two independent measurements of ba PWV by two observers (Upper panel). Reproducibility of ba PWV. Bland-Altman plot showing observer difference in measurements. (Lower panel).
Figure 7 Interobserver Carotid Femoral PWV. Relationship between two independent measurements of C-F PWV by two observers (Upper panel). Reproducibility of C-F PWV. Bland-Altman plot showing observer difference in measurements. (Lower panel).
The mean of hb, ha, ba, C-F PWV and ASI recorded during two-separate occasions in same subject was shown in the table 6. The relationship and Bland-Altman plot of two period measurements are shown in figure 8, 9, 10 and 11. The measurements of PWV had significant correlation with (r = 0.71 – 0.98). There was significantly less variation in reproducibility as clearly shown in Bland-Altman plot. Most of the values of PWV are lying within the mean ± 2 SD. The present study demonstrated, considerably high Pearson's correlation coefficients for both interobserver and interperiod reproducibility in PWV measurements with new device. Correlation analysis of carotid femoral PWV with other PWV parameters obtained is shown in table 7. Both brachial and ankle ASI, hb, ha, and ba PWV correlated significantly with C-F PWV. The Pearson's correlation coefficient was 0.99 for brachial ankle PWV and was 0.30 for brachial ASI.
Table 6 Means and correlational analysis (Pearson r value) of arterial stiffness among all subjects (Healthy + CAD) (n = 24)
Pulse Wave Velocity Period I Period II Pearson (r) 95% CI P value
Heart Brachial PWV 263.6 ± 15.38 263.6 ± 17.59 0.71 0.43 to 0.86 p < 0.0001
Heart Ankle PWV 424.4 ± 19.73 425.5 ± 15.74 0.86 0.70 to 0.93 p < 0.0001
Brachial Ankle PWV 1111 ± 105.7 1119 ± 104.1 0.98 0.96 to 0.99 p < 0.0001
Carotid Femoral PWV 698.6 ± 91.28 706.8 ± 91.58 0.94 0.96 to 0.97 p < 0.0001
Brachial ASI 37.08 ± 4.72 37.27 ± 3.95 0.84 0.67 to 0.93 p < 0.0001
Ankle ASI 48.88 ± 7.10 46.99 ± 7.85 0.81 0.65 to 0.91 p < 0.0001
Values are expressed as Mean ± SD
Figure 8 Interperiod Heart Brachial PWV. Relationship between measurements of hb PWV between two different periods (Upper panel). Reproducibility of hb PWV. Bland-Altman plot showing difference in measurement between two periods. (Lower panel).
Figure 9 Interperiod Heart Ankle PWV. Relationship between measurements of ha PWV between two different periods (Upper panel). Reproducibility of ha PWV. Bland-Altman plot showing difference in measurement between two periods. (Lower panel).
Figure 10 Interperiod Brachial Ankle PWV. Relationship between measurements of ba PWV between two different periods (Upper panel). Reproducibility of ba PWV. Bland-Altman plot showing difference in measurement between two periods. (Lower panel).
Figure 11 Interperiod Carotid Femoral PWV. Relationship between measurements of C-F PWV between two different periods (Upper panel). Reproducibility of C-F PWV. Bland-Altman plot showing difference in measurement between two periods. (Lower panel).
Table 7 The correlation between carotid-femoral PWV and other variables evaluated in all subjects (Healthy + CAD) (n = 88)
P-Value Pearson (r)
Brachial ASI <0.005 0.30
Ankle ASI <0.0001 0.67
Heart Brachial PWV <0.0004 0.37
Heart Ankle PWV <0.0001 0.83
Brachial Ankle PWV <0.0001 0.99
ASI, arterial stiffness index; PWV, pulse wave velocity
Our present study was not aimed to detect differences in the measured parameters between control group and CAD group; therefore no sub group analysis was performed. The CAD group was included to provide a wide range of PWV values, thus strengthening the reproducibility of method.
Discussion
In recent years with the development of readily available noninvasive assessment techniques investigation of arterial stiffness, especially of the large arteries has gathered pace. These include the measurement of PWV, the use of ultrasound to relate the change in diameter or area of an artery to distending pressure and analysis of arterial wave forms obtained by applanation tonometry. For the measurement of AS, several new techniques have been developed, but their association with aortic PWV, an established measure of central arterial stiffness has not been validated. Because of its size and elasticity, the aorta is the main determinant of systemic arterial compliance [16] and thus AS models incorporating wave reflection are influenced by both central and peripheral AS. Central stiffness is most commonly assessed using aortic PWV, which is a robust measurement and predictive of cardiovascular morbidity in hypertensive [17] and non-hypertensive subjects [18,19]. Aortic PWV can also be measured non-invasively by using MRI [20]. It has potential advantage of accurate determination of path length, however its use is very much limited due to prolonged time required to make recording, lack of availability in the clinical settings, high cost for measurement and difficulty in performing within a strong magnetic field.
Perhaps the best and most widely used technique to estimate the distensibility and stiffness of the aorta and proximal vessels is PWV. Although the properties of large arteries have been studied for several decades, the field of clinical arterial biomechanics remains in its infancy. The clinical value of any of the available techniques is yet to be proved convincingly and no single parameter of compliance or stiffness can ever be expected to describe all clinically relevant arterial wall properties. Measurement error can be substantial, including problems related to the measurement of both transit time and distance traveled by the pulse wave. There is a need for a simple, reliable, noninvasive method of detecting early disturbances at the time when therapeutic intervention can be most beneficial. Currently, none of the methodologies available are yet suitable for use in wide spread clinical practice. The results of our study indicate that a new device provides a simple non-invasive method for the assessment of arterial stiffness and PWV in clinical settings and operator does not require prolong period of training to achieve measurements.
We have determined the C-F PWV along with determination of other velocities like hb, ha, ba PWV. The present experiment included both healthy control and patients with CAD to provide a wide range of values. The correlation with C-F PWV was very high with other PWV estimated. Analysis of data was done by using Bland-Altman Plots and reproducibility was reported as the SD of the differences between the paired measurements. This method was used rather than only reporting coefficient of variance, as the latter is a less satisfactory method of assessing reproducibility and can some time misleading [21].
Arterial stiffness index (ASI) also well correlated with C-F PWV (r = 0.67, p < 0.0001). Similar correlation between ASI and C-F PWV has been documented earlier in elderly hypertensive patients [22]. Our results indicate that, non-invasive PWV measurements with new device were reproducible. Both within observer and between the observers, reproducibility was high with low SD for measurement differences. The lower value for between observer differences reflects the fact that, the means of two readings made by each observer are comparable, thus reducing variability. The method we used to measure PWV does not require any specialized technique and the examiner only has to wrap cuffs on all the limbs and place electrodes for recording lead II ECG. Our study demonstrated high Pearson's correlation coefficient for interobserver and interperiod reproducibility. In the Bland-Altman plot the deviation was greater at high PWV values. A similar deviation has also been reported in the measurements of C-F PWV [23]. It is known that, the variability in measurement of PWV increases when PWV are high due to confounding factors like high blood pressure, blood flow [24], and high sympathetic tone [25]. In our study though we have included CAD patients, the deviation in measurements ranged within a mean ± 2SD even with high PWV. Furthermore both interobserver and interperiod coefficient variations were less than 20 %. We made comparison of interobserver and interperiod measurements with both Bland-Altman and correlation analysis. Unlike correlation analysis, Bland-Altman Plots do not assume zero error for either of the two measurements under comparison and therefore is a better indicator of the true agreement between the measurements [26]. We noted significant association of correlations analysis between C-F PWV and other PWV and ASI, with low variability in Bland-Altman analysis.
There are a number of different ways to measure PWV, and these are generally simple to perform.
Conclusion
The results of this study indicate that, the measurement of arterial PWV by oscillometric technique using PeriScope is reproducible and simple. The validity and reproducibility of the measurement of PWV in interobserver and interperiod evaluation was high with a low SD for measurement difference. Derived carotid femoral velocity correlated well with hb, ha and ba PWV and arterial stiffness index. Simultaneous measurement of all velocities and stiffness index provide better over all assessment of arterial functions. New device may be useful in studying arterial stiffness in large population.
List of Abbreviations
ABI – Ankle Brachial Index
AS – Arterial Stiffness
ASI – Arterial Stiffness Index
ba – Brachial ankle
BP – Blood Pressure
CAD – Coronary Artery Disease
C-F – Carotid Femoral
CV – Coefficient of variation
CVD – Cardiovascular Disease
DBP – Diastolic blood pressure
DSP – Digital signal processing
ECG – Electrocardiogram
ha – Heart ankle
hb – Heart brachial
HR – Heart rate
MAP – Mean Arterial Pressure
OPD – Out Patient Department
PP – Pulse pressure
PTT – Pulse Transit Time
PWV – Pulse Wave Velocity
SBP – Systolic blood pressure
SD – Standard Deviation
SE – Standard Error
UT – Upstroke Time
Authors' contributions
MURN conceived and designed the study and drafted the manuscript. BMR and YS performed the study as two separate observers under the supervision of MURN. YS compiled and analyzed the data statistically. ANP performed and evaluated the angiogram of cardiac patients. PUR participated in the design and coordination of the study. All authors read and approved the final manuscript.
==== Refs
Oliver JJ Webb DJ Noninvasive assessment of arterial stiffness and risk of atherosclerotic events Arteriosclerosis Thrombosis Vascular Biology 2003 23 554 566 10.1161/01.ATV.0000060460.52916.D6
Blacher J Guerin AP Pannier B Marchais SJ Safar ME London GM Impact of aortic stiffness on survival in end-stage renal disease Circulation 1999 99 2434 2439 10318666
Shoji T Emoto M shinohara K Kakiya R Tsujimoto Y Kishimoto H Ishimura E Tabata T Nishizawa Y Diabetes mellitus, aortic stiffness, and cardiovascular mortality in end-stage renal disease J American Society of Nephrology 2001 12 2117 2124
Meaume S Benetos A Henry OF Rudnichi A Safar ME Aortic pulse wave velocity predict cardiovascular mortality in subjects >70 years of age Arteriosclerosis Thrombosis Vascular Biology 2001 21 2046 2050
Mc Veigh GE Brennan G Hayes R Cohn JN Finkelstein S Johnston D Vascular abnormalities in non-insulin dependent diabetes mellitus identified by arterial wave form analysis American Journal Medicine 1993 95 424 430 10.1016/0002-9343(93)90313-E
Cockcoroft JR Wilkinson IB Webb DJ Age, arterial stiffness and the endothelium Age Ageing 1997 26 53 60 9506434
Woodman RJ Kingwell BA Beilin LJ Hamilton SE Dart AM Watts GF Assessment of central and peripheral arterial stiffness Am J Hypertension 2005 18 249 260 10.1016/j.amjhyper.2004.08.038
O'Rourke M Staessen JA Vlachopoulos C Duprez D Plante GE Clinical applications of arterial stiffness; definitions and reference values American Journal of Hypertension 2002 15 426 444 12022246 10.1016/S0895-7061(01)02319-6
Asmar R Benetos A Topouchian J Laurent P Pannier B Brisac AM Target R Levy BI Assessment of arterial distensibility by automatic pulse wave velocity measurement Hypertension 1995 26 485 490 7649586
Hoeks AP Brands PJ Smeets FA Reneman RS Assessment of the distensibility of superficial arteries Ultrasound Med Biol 1990 16 121 128 2183458 10.1016/0301-5629(90)90139-4
Riley WA Barnes RW Evans GW Burke GL Ultrasonic measurement of the elastic modulus of the common carotid artery: the Atherosclerosis Risk in Communities (ARIC) Study Stroke 1992 23 952 956 1615543
Lehmann ED Elastic properties of the aorta Lancet 1993 342 1417 7901692 10.1016/0140-6736(93)92772-L
Yamashina A Tomiyama H Takeda K Tsuda H Arai T Hirose K Koji Y hori S Yamamoto Y Validity, reproducibility, and clinical significance of noninvasive brachial ankle pulse wave velocity measurement Hypertension Research 2002 25 359 364 12135313 10.1291/hypres.25.359
Blacher J Asmar R Djane S London GM Safar ME Aortic pulse wave velocity as a marker of cardiovascular risk in hypertensive patients Hypertension 1999 33 1111 1117 10334796
Su HM Lee KT Chu CS Tsai CL Lee MY Lai HM Sheu SH Lai WT The validity of brachial-ankle pulse wave velocity in predicting coronary artery disease Acta Cardio Sin 2003 19 237 242
Kelly RP Tunin R Kass DA Effect of reduced aortic compliance on cardiac efficiency and contractile function of in situ canine left ventricle Circulation research 1992 71 490 502 1386792
Laurent S Boutouyroe P Asmar R Gautier I Laloux B Guize L Ducimetiere P Benetos A Aortic stiffness is an independent predicator of all – cause and cardiovascular mortality in hypertensive patients Hypertension 2001 15 426 444
Blacher J Safar ME Guerin AP Pannier B Marchais SJ London GM Aortic pulse wave velocity index and mortality in end-stage renal disease Kidney International 2003 63 1852 1860 12675863 10.1046/j.1523-1755.2003.00932.x
Meaume S Benetos A Henry OF Rudnichi A Safar ME Aortic pulse wave velocity predicts cardiovascular mortality in subjects >70 years of age Arteriosclerosis Thrombosis Vascular Biology 2001 21 2046 2050
Mohiaddin RH Firmin DN Longmore DB Age-related changes of human aortic flow wave velocity measured noninvasively by magnetic resonance imaging J of applied physiology 1993 74 492 497
Bland JM Altman DG Statistical methods for assessing agreement between two methods of clinical measurement Lancet 1986 I 307 310 2868172
Kaibe M Ohishi M Komai N Ito N Katsuya T Rakugi H Ogihara T Arterial stiffness index: A new evaluation for arterial stiffness in elderly patients with essential hypertension Geriatrics and Gerontology International 2002 2 199 205 10.1046/j.1444-1586.2002.00045.x
Lehmann ED Hopkins KD Rawesh A Joseph RC Kongola K Coppack SW Gosling RG Relation between number of cardiovascular risk factors/events and noninvasive Doppler ultrasound assessment of aortic compliance Hypertension 1998 32 565 569 9740627
Asmar R Factor influencing pulse wave velocity, in arterial stiffness and pulse wave velocity 1999 Amsterdam, Elsevier 57 88
Lantelme P Milton H Gharib C Gayet C Fortrat JO White coat effect and reactivity to stress: cardiovascular and autonomic nervous system responses Hypertension 1998 31 1021 1029 9535430
Ludbrook J Comparing methods of measurements Clin Exp Pharmacology 1997 24 193 203
|
16115324
|
PMC1224857
|
CC BY
|
2021-01-04 16:37:35
|
no
|
Biomed Eng Online. 2005 Aug 23; 4:49
|
utf-8
|
Biomed Eng Online
| 2,005 |
10.1186/1475-925X-4-49
|
oa_comm
|
==== Front
Cancer Cell IntCancer Cell International1475-2867BioMed Central London 1475-2867-5-271611782910.1186/1475-2867-5-27Primary ResearchEstablishment of a doxycycline-regulated cell line with inducible, doubly-stable expression of the wild-type p53 gene from p53-deleted hepatocellular carcinoma cells Chi Tian-Yi [email protected] George G [email protected] Lok-Kee [email protected] Paul BS [email protected] Department of Surgery, Chinese University of Hong Kong, Prince of Wales Hospital, Shatin, New Territories, Hong Kong SAR, China2005 23 8 2005 5 27 27 20 2 2004 23 8 2005 Copyright © 2005 Chi et al; licensee BioMed Central Ltd.2005Chi et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
p53 is important in the development of hepatocellular carcinoma (HCC) and in therapeutic approaches, but the mechanism whereby it inhibits HCC growth is still unclear. The aim of the present study was to establish a HCC cell system in which p53 levels can be regulated. Full-length wild-type p53 cDNA obtained by PCR was cloned into a retroviral response vector controlled by the tetracycline responsive element (RevTRE-p53). The regulatory vectors RevTet-Off and RevTRE-p53 were transfected into a packaging cell line, PT67. Hep3B cells in which the p53 gene was deleted were infected with RevTet-Off viral particles from the PT67. Three G418-resistant cell clones with high luciferase expression and low background were infected with RevTRE-p53. By screening dozens of RevTRE-p53-infected clones with hygromycin we identified the one with the highest expression of p53 and the lowest background after doxycycline treatment. The results showed that p53 expression in this cell clone could be simply turned on or off by removing or adding doxycycline. Furthermore, it was found that the level of p53 protein was negatively and sensitively related to the doxycycline concentration. In conclusion, we have established a HCC cell line in which p53 expression can be switched on or off and regulated in a dose- and time-dependent manner.
hepatocellular carcinomap53tetracyclinedoxycyclinegene expression.
==== Body
Introduction
Hepatocellular carcinoma (HCC) is one of the most common malignancies worldwide and is estimated to cause over one million deaths every year [1]. Risk factors have been identified, including chronic infection with hepatitis B and C viruses and exposure to aflatoxin B1, but the exact molecular mechanisms of carcinogenesis have not been completely defined. Mutation or deletion of the p53 gene, which plays an important role in cell growth, division and apoptosis by acting as a transcription factor or by forming complexes with other proteins [2], are frequently detected in HCC [3]. In addition, loss of the functional p53 gene is associated with lower cellular differentiation and poor prognosis [4]. Hence, restoration of a wild-type p53 gene is an attractive approach to the treatment of HCC [5].
Tight quantitative and temporal control of gene expression is very useful for basic biological and medical research applications. Several inducible gene expression systems have been developed such as hormones, heavy metal ions and heat shock [6]. However, most of them are limited by low and nonspecific induction, toxic effects of the inducing agents, high basal levels of gene expression under non-induced conditions and pleiotropic effects [7]. The tetracycline (tet)-regulated gene expression system overcomes many of the problems of other inducible systems [8,9] and has been widely utilized in mammalian cell culture [10], transgenic mice [11] and other species [12]. Furthermore, the retrovirus-mediated tet-regulated gene expression system (RevTet-Off/On) combines the advantages of retroviral transfer with those of tet-regulation and allows faster and more efficient establishment of regulated gene expression in a wide variety of cell types [13]. In the RevTet-Off system, transcription of the gene of interest is turned off in the presence of tetracycline or doxycycline (a derivative of tetracycline). In contrast, transcription of this gene is turned on in the presence of tetracycline or doxycycline in the RevTet-On system.
Loss of p53 function by mutation not only results in the malignant transformation of cells but also enhances resistance to anticancer drugs and radiation [14]. Studies on human tumors carrying mutated or deleted p53 show that the introduction of exogenous wild-type (wt) p53 could lead to apoptotic death of the tumor cells and effectively inhibit their growth [15]. It is therefore important to gain better insight into p53 function in HCC. We therefore cloned the wt p53 gene from a human hepatoblastoma cell line, HepG2, into a human hepatocellular cell line, Hep3B, using a retrovirus-mediated tet-regulated gene expression system (RevTet-Off) and established a HCC cell line with tightly controllable, doubly-stable expression of the wt p53 gene.
Materials and methods
RevTet-Off System
The RevTet-Off System was purchased from Clontech Laboratories (Palo Alto, CA). It is a complete retroviral gene expression system including a retroviral regulatory tTA vector (pRevTet-Off), a retroviral response vector (pRevTRE), a control vector (pRevTRE-Luc) and a packaging cell line (PT67).
Cell lines and culture
Hep3B and HepG2 cells were purchased from the American Type Culture Collection. The Hep3B cells were derived from a human HCC deficient in p53 [16], the HepG2 cells from a human hepatoblastoma expressing wt p53 [17]. PT67 packaging cells were derived from an NH/3H3-based line expressing the 10A1 virus envelope [18]. All cells were grown in Dulbecco's modified Eagles's minimal essential medium with high glucose (Invitrogen, Carlsbad, California) supplemented with 10% fetal bovine serum (Clontech, Palo Alto, CA) and incubated at 37°C in a humidified atmosphere containing 95% air and 5% carbon dioxide.
Reagents
G418 (Clontech, Palo Alto, CA) was prepared as a 10 mg/ml stock solution. Hygromycin B was available from Clontech as a stock solution (50 mg/ml). Both were stored at 4°C. Aliquots of 1 mg/ml Doxycyclin (Clontech, Palo Alto, CA) were stored in the dark at -20°C.
Construction of the RevTRE-p53 plasmid
The full-length cDNA of wt p53 was obtained from the HepG2 cells by RT-PCR. The PCR primers were as follows: 5'-TTA AGC TTT TTG CGT TCG GGC TGG GAG C-3' and 5'-CGA TCG ATA TGG TGG CAT GAA CCT GTG G-3', which contained restriction sites for Hind III and Cal I respectively. The resulting PCR products were further purified with a QIAquick gel extraction kit (Qiagen, Valencia, CA) according to the manufacturer's instructions. The full-length p53 cDNA was cloned into the pRevTER vector using the restriction enzymes Hind III/Cla I.
Transformation, purification and identification of retrovirus plasmids
The retrovirus vectors provided in the RevTet-Off System and recombinant pRevTRE-p53 plasmid were transformed into E. coli (DH5α). Plasmid DNA was purified using a QIAprep mini kit (Qiagen, Valencia, CA) according to the manufacturer's instructions. The purified plasmids were digested with the restriction enzymes Hind III and Cla I (Invitrogen, Carlsbad, California) to confirm the insert and the vector.
Sequence analysis of the recombinant pRevTRE-p53 plasmid
The purified pRevTRE-p53 plasmid was sequenced with an ABI PRISM BigDye Terminator Cycle Sequencing kit (PE Applied Biosystems, Foster City, CA) according to the manufacturer's protocol. A pair of primers were designed for sequencing pRevTRE-p53: 5'-AAG CTT TTT GCG TTC GGG CTG GGA GC-3' and 5'-ATC GAT ATG GTG GCA TGA ACC TGT GG-3'. Recombinant pRevTRE-p53 plasmid was purified from E. coli with a QIAGEN Plasmid Midi kit (Qiagen, Valencia, CA).
Package of retroviral plasmids and recombinant pRevTRE-p53 plasmid
PT67 cells were transfected with the pRevTet-Off, pRevTRE-Luc or pRevTRE-p53 plasmids using Lipofectamine Reagent (Invitrogen, Carlsbad, California) according to the manufacturer's instructions. The transfected cells were selected in 600 μg/ml G418 or 400 μg/ml hygromycin for 2 weeks and the selective media were replaced every 4 days. Well-separated antibiotic-resistant clones were individually picked with cloning discs and transferred to 24-well plates in selective medium. The cells were then transferred to larger culture vessels before confluence, and aliquots of early passages of cells (1 × 106) were frozen in liquid nitrogen.
Determination of viral titer
The supernatant from virus-containing PT67 cells was filtered through a 0.45 μm filter and added to Hep3B cells in the presence of 4 μg/ml polybrene (Sigma Chemical Co. St. Louis, MO). To determine the viral titer, infection of six 10-fold serial dilutions was performed in 6-well plates. The medium was replaced 6 h after infection. Three serial infections were performed. Forty-eight hours after infection, Hep3B cells were subjected to selection by G418 (0.5 mg/ml) or hygromycin (0.05 mg/ml) for one week. The viral titer corresponded to the number of colonies developed at the highest dilution.
Infection of Hep3B cells by RevTet-Off virus
Hep3B cells were infected with supernatant from fresh PT67 cells containing RevTet-Off virus and selected with G418 as described above. The isolated G418-resistant clones were transferred to a 24-well plate with cloning cylinders. Before confluence, the cells were transferred into larger culture vessels.
Screening for inducible G418-resistant clones by RevTRE-Luc virus
The pRevTRE-Luc control vector was constructed by cloning the luciferase gene into the Hind III/Cal I sites in pRevTRE. G418-resistant clones infected with RevTRE-Luc virus were selected by screening luciferase expression. For each clone, two additional cultures in wells of a 6-well plate were deployed as a test plate for screening. Infection by the RevTRE-Luc virus was performed as described previously. Twenty-two hours after infection, 1 μg/ml doxycycline (Dox) was added to one of the two test wells. Luciferase activity was analyzed using a luciferase assay system kit (Promega Corporation, Madison) according to manufacturer's instructions.
Infection of G418-resistant clones by RevTRE-p53 virus
The selected G418-resistant clones were infected with supernatant from PT67 cells containing RevTRE-p53 virus. The clones were selected with hygromycin for two weeks as described previously. Healthy hygromycin-resistant clones were individually transferred to a 24-well plate with cloning cylinders and propagation was continued in selective medium. Aliquots of early passages of hygromycin-resistant cells (1 × 106) were frozen in liquid nitrogen.
Reverse transcription-polymerase chain reaction (RT-PCR) analysis
Total RNA was extracted from cells using a RNeasy Mini kit (Qiagen, Valencia, CA) according to the manufacturer's instructions and reverse-transcribed to cDNA with a First-Strand cDNA Systhesis Kit (Amersham Biosciences, Selangor Darul Ehsan, Malaysia). The neomycin gene and p53 cDNA were amplified by PCR, generating a 382-bp fragment of the neomycin gene (forward primer: 5'-TCC TGT CAT CTC ACC TTG CTC C-3' and reverse primer: 5'-GCA ATA TCA CGG GTA GCC AAC G-3'); a 282-bp cDNA fragment of p53 (5'-TAC ATG TGT TAA CAG TTC CTG CA-3' and 5'-TTC TGA CAA CGA TCG GAG GA-3'), or a 114-bp fragment of β-microglobulin gene (5'-ACC CCC ACT GAA AAA GAT GA-3' and 5'-ATC TTC AAA CCT CCA TGA TG-3') that served as an internal standard for RNA integrity and equal gel loading. Thermocycling was carried out at 94°C for 3 min and 30 amplification cycles of 94°C for 30 s, 62°C for 30 s and 72°C for 1 min, followed by 72°C for 7 min. The PCR products were electrophoresed in a 1.2% agarose gel and the cDNA fragments were stained with ethidium bromide and visualized under an Ultraviolet Transilluminator (UVP Inc.Upland, CA).
Western blot
Cells were lysed in extraction buffer and 20 μg/ml of protein were subjected to 12% SDS-polyacrylamide gel electrophoresis. The proteins were transferred electrophoretically on to nitrocellulose membranes (Amersham, Piscataway, NJ). The membranes were probed with a monoclonal mouse anti-human p53 antibody (1:1000) (Santa Cruz, CA) and β-actin (1:1500) (Santa Cruz, CA) and were then incubated with goat anti-mouse IgG-HRP antibody (1:2000) (Santa Cruz, CA). The signals were detected using an enhanced chemiluminescence system.
Results
Identification of recombinant RevTRE-p53 plasmids by restriction enzymes
The sizes of pRevTet-Off, pRevTRE, the recombinant pRevTRE-p53 plasmid and the cDNA fragment of p53 were 7.8 Kb, 6.5 Kb, 8.6 Kb and 2.1 Kb, respectively. The RevTRE-p53 plasmid was cleaved with Hind III and Cal I and fragments of 2.1 Kb and 6.5 Kb were generated. Restriction enzyme mapping showed that the full-length p53 cDNA had been successfully cloned into the Hind III/Cla I sites in the pRevTRE vector, and the sizes of the pRevTet-Off and pRevTRE-p53 fragments were consistent with expectation (Fig. 1).
Figure 1 Restriction mapping of recombinant RevTRE-p53 and RevTet-Off plasmids. The pRevTRE-p53 plasmid was cleaved with Hind III and Cla I. The fragments were as follows: Lane 1: 1 Kb marker, Lane 2: the fragments of the pRevTRE vector (6.5 Kb) and the p53 fragment (2.1 Kb). Lane 3: circled pRevTet-Off (7.8 Kb); Lane 4: circled pRevTRE-p53 (8.6 Kb).
Analysis of recombinant RevTRE-p53 plasmid sequence
To confirm the sequence and orientation of the inserted p53 cDNA in the pRevTRE vector, the purified recombinant pRevTRE-p53 plasmid was verified by sequencing. The p53 insert contained an ATG initiation codon and an entire open reading frame coding for the p53 protein. Sequencing results suggested that the inserted cDNA was identical to the human p53 gene published by GenBank (Accession number: X54156).
Selection of high titer packaging cell clones
After transfection with the purified RevTet-Off, RevTRE-p53 and RevTRE-Luc plasmids, PT67 cells were selected with G418 or hygromycin for two weeks. The antibiotic-resistant clones were individually isolated with cloning discs. The viral titers were determined by infecting Hep3B cells with 6 × 10-fold serial diluted virus-producing media from the PT67 clones and showed 5 × 104, 3 × 104, and 6 × 104 colony forming units for RevTet-Off, RevTRE-p53 and RevTRE-Luc, respectively. The highest titer clones were selected for further experiments.
Selection-stable expression of Hep3B/RevTet-Off clones through transient infection by RevTRE-Luc virus
In an effort to establish a tet-regulated inducible cell line, the RevTet-Off vector was first introduced into Hep3B cells (Hep3B/RevTet-Off) by retroviral infection. The G418-resistant clones were then infected transiently with RevTRE-Luc virus packaged by the PT67 cells. In 32 of the Hep3B/RevTet-Off clones the inductive efficacy of doxycycline-regulated luciferase expression varied from 2 to 70 fold. Six clones showed higher levels of luciferase expression in the absence of doxycycline but different backgrounds in the presence of 1 μg/ml Dox (Fig. 2). The results showed that doxycycline exerted tight control over the expression of this reporter gene in clones T13, T15 and T23. Thus, we selected these clones for further experiments.
Figure 2 Doxycycline-regulated luciferase activity in Hep3B/RevTet-Off clones. Hep3B/RevTet-Off clones were infected transiently using RevTRE-Luc retroviral particles. Cells were cultured in either the presence or absence of 1 μg/ml doxycycline. Clones were assayed by the luciferase assay system.
Identification of doubly-stable expression in Hep3B/RevTet-Off/TRE-p53 clones
The three selected Hep3B/RevTet-Off clones, T13, T15 and T23, were infected with supernatant from PT67 cells containing the recombinant RevTRE-p53 virus. Hygromycin-resistant Hep3B/RevTet-off/TRE-p53 (Hep3B/tet/p53) clones, 13T, 15T and 23T, were isolated and propagated. We used RT-PCR to establish whether these clones exhibited double-stable expression. Two sets of exploring primers were used to check if these clones are expressing both the neomycin resistance (Neor) gene and p53 gene. The RT-PCR results confirmed expectation (Fig. 3). Therefore, a cell line with doubly-stable gene expression, Hep3B/tet/p53, was established.
Figure 3 Establishment of a cell line with doubly-stable expression. Total RNA was extracted from the selected hygromycin-resistant clones in the absence of doxycycline. Two pairs of primers were used to probe the specific neomycin (NeoΓ) and p53 sequences in the tested cells. Lane 1:100 bp marker; Lane 2: HepG2 cells (carrying the wt p53 gene); Lane 3: Hep3B cells (deletion of p53 gene); Lanes 4, 5, and 6: the tested 13T, 15T and 23T clones.
Selection of a tightly doxycycline-regulated p53-expression clone
To obtain a high expression, low background clone, 28 different 13T, 15T and 23T clones were analyzed by Western blotting. Some, such as the 23T clone, displayed higher p53 expression in the absence of doxycycline and higher background in its presence, but others, such as the 15T clone, displayed the opposite results. Among the clones tested we found one (13T) that was optimal, exhibiting high p53 expression and low background controlled by doxycycline; this clone was named Hep3B/tet/p53-13 (or p53-13) (Fig. 4). Semi-quantitative RT-PCR demonstrated that the induction of p53 in the Hep3B/tet/p53-13 cells was regulated by doxycycline at the transcriptional level (Fig. 5).
Figure 4 Selection of the optimal clone under Dox control. For each of the hygromycin-resistant clones, cells were grown for 24 h in the presence (Dox+) or absence (Dox-) of 1 μg/ml Dox. The expression of p53 protein was analyzed by Western blotting. Lanes 1, 2: 13T (Dox+, Dox-); Lanes 3, 4: 15T (Dox+, Dox-); Lanes 5, 6: 23T (Dox+, Dox-).
Figure 5 Semi-quantitative RT-PCR analysis of expression of p53 regulated by Dox. Hep3B/tet/p53-13 (p53-13) cells were cultured for 24 h in the presence or absence of 1 μg/ml Dox and analyzed by semi-quantitative RT-PCR. Lane 1:100 bp marker; Lane 2: Hep3B cells; Lane 3: p53-13 + 1 μg/ml Dox; lane 4: p53-13 + 0 μg/ml Dox. β-microglobulin (MG) was used as an internal control.
Effect of Doxycycline on Hep3B/tet/p53 cell growth
The effect of doxycycline on the growth of Hep3B/tet/p53 cells was examined by trypan blue exclusion. The result showed 1 μg/ml doxycycline had no effect on the growth of Hep3B/tet/p53-13 cells when compared to parental Hep3B cells (result not shown).
Regulation of p53 expression by doxycycline
Hep3B/tet/p53-13 cells were grown in different concentrations of doxycycline and the expression of p53 protein was analyzed by Western blotting. The results showed that 1- 0.00001 μg/ml doxycycline correlated negatively with the level of p53 protein (Fig. 6). Therefore, the expression of p53 protein was regulated by doxycycline in a dose-dependent manner and was almost completely suppressed by a 1 μg/ml concentration. A time-course analysis of this regulation was performed by Western blotting. The results showed that p53 protein expression was gradually enhanced by the removal of doxycycline and reached maximum 6 h after removal of the antibiotic (Fig. 7). In contrast, the protein expression decreased slowly with the addition of doxycycline, becoming almost undetectable 12 h after treatment (Fig. 7). Obviously, doxycycline regulated p53 protein expression in a time-dependent manner.
Figure 6 Expression of p53 protein by Dox controlled quantitatively. After p53-13 cells were cultured in 0 and 1, 0.1, 0.01, 0.001, 0.0001, 0.00001 μg/ml Dox for 24 h, expression of p53 protein was examined by Western blotting. β-actin was served as an internal control.
Figure 7 Time-dependent Dox-regulated expression of p53 protein. In the removing group, medium was removed 24 h after p53-13 cells were cultured in 1 μg/ml Dox (Dox+), and then the cells were cultured continually in the absence of Dox (Dox-). In the adding group, cells were cultured in the absence of Dox for 24 h, and then the cells were incubated with 1 μg/ml of Dox. The expression of p53 protein was analyzed by Western blotting. Lane1: Dox+; Lanes 2, 3, 4: Dox- 1 h; 3 h; and 6 h; Lane 5: Dox-; Lanes 6, 7, 8: Dox+ 3 h; 6 h; and 12 h. β-actin was served as an internal control.
Discussion
The recent development of tetracycline (tet)-regulated gene expression has multiplied the tools available for quantitative and temporal control of exogenous genes in different areas of biology and medicine. The first tet-regulated gene expression system (tTA based) [8] consisted of two separate plasmids: the regulatory plasmid for the tet-transactor (tTA) and the response plasmid for the tet-operator minimal promoter driving the gene of interest. In practice, the regulatory plasmid must first be transfected into the target cells. However, conventional transfection methods are not always efficient enough for transfer and high expression of tTA is toxic to cells, probably because of transcriptional "squelching" by the VP16 transactivator domain in tTA [19]. Once cells that stably express tTA have been established, the response plasmid containing the gene of interest under the control of the tet-operator minimal promoter must be introduced. The establishment of a cell line with inducible gene expression can be tedious and time-consuming and modifications have been introduced to circumvent some of the limitations of this system [20-22]. One such modification involves utilizing retrovirus-mediated gene transfer instead of plasmid DNA transfection, thus ensuring more efficient transduction of mammalian cells in vitro. Retroviral vectors for regulated expression are powerful tools for gene transfer since the retroviruses generally integrate as single copies into the target cell genome. Moreover, they are stable, need minimal effort to prepare, and generate populations of cells that can regulate expression of the gene of interest within weeks.
To elucidate the functions of p53 in HCC we chose a retrovirus-mediated tet-regulated gene expression system, RevTet-Off, as a tool. In the present study a full-length cDNA of wt p53 from a human hepatoblastoma cell line, HepG2, which expresses small amounts of wt p53, was cloned into a retroviral response vector, pRevTRE. Subsequently, the target cells, Hep3B, which are derived from a human HCC and have a homozygous deletion of the p53 gene, were first infected by the pRevTet-Off regulatory virus and selected by an antibiotic, G418. Three of 32 tested Hep3B/RevTet-Off clones with the highest maximal luciferase activities were then infected with the recombinant pRevTRE-p53 response virus. After screening with hygromycin, the inducible, doubly-stable p53-expression clones, Hep3B/RevTet-Off/TRE-p53 (Hep3B/tet/p53), were identified by RT-PCR. In 28 other hygromycin-resistant clones we found that the expression of p53 protein was quite different. Clone 23T displayed high levels of p53 protein in the absence of doxycycline, but there was a high basal expression in the presence of 1 μg/ml doxycycline, indicating a high background. In contrast, clone 15T had a low level of p53 protein and low background. In contrast to clones 15T and 23T, clone 13T displayed a higher expression level of p53 and lower background, as demonstrated by Western blot analysis and semi-quantitative RT-PCR.
We speculate that such variations of p53 expression might result from the integration of the retrovirus. In general, the site of integration into the host cell genome is random [23]. When pRevTRE-p53 retroviruses integrate adjacent to an endogenous enhancer element in the Hep3B cell genome, this enhancer could activate p53 transcription even in the presence of 1 μg/ml doxycycline, resulting in a high background. At the same time, the tTA of pRevTet-Off retroviruses might co-operate with this enhancer in the absence of doxycycline to induce quite a high level of p53 expression (as in clone 23T). However, if pRevTRE-p53 retroviruses integrate into a site that is not influenced by an endogenous enhancer, the basal expression level of p53 could be very low as in clone 15T. No methods are currently available for ensuring that a retrovirus integrates at the most useful site. Therefore, it is important to screen the hygromycin-resistant clones for the one that is optimal in terms of high inducible expression and low background.
In this study, an HCC cell line, Hep3B/tet/p53, was successfully established with a retrovirus-mediated tet-regulated gene expression system, RevTet-Off, for inducible, doubly-stable expression of the wt-p53 gene in vitro. This cell line is totally different from other lines established through adenovirus-mediated transfer of the p53 gene [24,25]. To the best of our knowledge, such a p53 expression and regulation system has not been reported in a HCC cell line. The system of p53 expression we have established can be controlled by doxycycline not only in turn-on or turn-off manner, but also in a quantitative and temporal manner. The HCC line equipped with pRevTRE-p53 established here can be used not only to study biological effects of the p53 gene in vitro, but also in human gene therapy trials in the future.
It is important to investigate the translation product of the gene of interest for its functional role in cell biology or in therapeutic situations when this product is toxic or needs to be maintained at appropriate levels. An inappropriately high level of p53 may become toxic before p53 is delivered to target cells [26]. The pRevTRE-p53 system we have established may overcome this problem since the activation of p53 expression can be controlled quantitatively and temporally by administration of doxycycline.
==== Refs
Okuda K Hepatocellular carcinoma: recent progress Hepatology 1992 15 948 963 1314774
Ko LJ Prives C p53: puzzle and paradigm Genes Dev 1996 10 1054 1072 8654922
Hsu IC Metcalf RA Sun T Welsh JA Wang NJ Harris CC Mutational hotspot in the p53 gene in human hepatocellular carcinomas Nature 1991 350 427 428 1849234 10.1038/350427a0
Piao Z Kim H Jeon BK Lee WJ Park C Relationship between loss of heterozygosity of tumor suppressor genes and histologic differentiation in hepatocellular carcinoma Cancer 1997 80 865 872 9307185 10.1002/(SICI)1097-0142(19970901)80:5<865::AID-CNCR6>3.0.CO;2-K
Habib NA Ding SF El-Masry R Mitry RR Honda K Michail NE Dalla Serra G Izzi G Greco L Bassyouni M El-Toukhy M Abdel-Gaffar Y Preliminary report: the short-term effects of direct p53 DNA injection in primary hepatocellular carcinoma Cancer Detect Prev 1996 20 103 107 8706034
Yarranton GT Inducible vectors for expression in mammalian cells Curr Opin Biotech 1992 3 506 511 1368935 10.1016/0958-1669(92)90078-W
Lee SW Tsou AP Chan H Thomas J Petrie K Eugui EM Allison AC Glucocorticoids selectively inhibit the transcription of the interleukin 1 beta gene and decrease the stability of interleukin 1 beta mRNA Pro Natl Acad Sci USA 1988 85 1204 1208
Gossen M Bujard H Tight control of gene expression in mammalian cells by tetracycline-responsive promoters Proc Natl Acad Sci USA 1992 89 5547 5551 1319065
Gossen M Freundlieb S Bender G Muller G Hillen W Bujard H Transcriptional activation by tetracyclines in mammalian cells Science 1995 268 1766 1769 7792603
Howe JR Skryabin BV Belcher SM Zerillo CA Schmauss C The responsiveness of a tetracycline-sensitive expression system differs in different cell lines J Biol Chem 1995 270 14168 14174 7775477 10.1074/jbc.270.23.14168
Ewald D Li M Efrat S Auer G Wall RJ Furth PA Hennighausen L Time-sensitive reversal of hyperplasia in transgenic mice expression SV40 T antigen Science 1996 273 1384 1386 8703072
Thomas DD Donnelly CA Wood RJ Alphey LS Insect population control using a dominant, repressible, lethal genetic system Science 2000 287 2474 2476 10741964 10.1126/science.287.5462.2474
Iida A Chen ST Friedmann T Yee JK Inducible gene expression by retrovirus-mediated transfer of a modified tetracycline-regulated system J Virol 1996 70 6054 6059 8709228
Lowe SW Bodis A McClatchey A Remington L Ruley HE Fisher DE Housman DE Jacks T p53 status and the efficacy of cancer therapy Science 1994 266 807 810 7973635
Yang C Cirielli C Capogrossi MC Passaniti A Adenovirus-mediated wild-type p53 expression induces apoptosis and suppresses tumorigenesis of prostatic tumor cells Cancer Res 1995 55 4210 4213 7671222
Ponchel F Puisieux A Tabone E Michot JP Froschl G Morel AP Frebourg T Fontaniere B Oberhammer F Ozturk M Hepatocarcinoma-specific mutant p53-249ser induces mitotic activity but has no effect on transforming growth factor beta 1-mediated apoptosis Cancer Res 1994 54 2064 2068 8174105
Knowles BB Howecc Aden DP Human hepatocellular carcinoma cell lines secrete the major plasma proteins and hepatitis B surface antigen Science 1980 209 497 499 6248960
Miller AD Chen F Retrovirus packaging cells based on 10A1 murine leukemia virus for production of vectors that use multiple receptors for cells entry J Virol 1996 70 5564 5571 8764070
Shockett P Difilippantonio M Hellman N Schatz DG A modified tetracycline-regulated system provides autoregulatory, inducible gene expression in cultured cells and transgenic mice Proc Natl Acad Sci USA 1995 92 6522 6526 7604026
Hofmann A Nolan GP Blau HM Rapid retroviral delivery of tetracycline-inducible genes in a single autoregulatory cassette Proc Natl Acad Sci USA 1996 93 5185 5190 8643550 10.1073/pnas.93.11.5185
Mizuguchi H Hayakawa T Characteristics of adenovirus-mediated tetracycline-controllable expression system Biochim Biophys Acta 2001 1568 21 29 11731081
Paulus W Baur I Boyce FM Breakefield XO Reeves SA Self-contained, tetracycline-regulated retroviral vector system for gene delivery to mammalian cells J Virol 1996 70 62 67 8523582
Watsuji T Okamoto Y Emi N Katsuoka Y Hagiwara M Controlled gene expression with a reverse tetracycline-regulated retroviral vector (RTRV) system Biochem Biophys Res Commun 1997 234 769 773 9175791 10.1006/bbrc.1997.6705
Reiser M Neumann I Schmiegel W Wu PC Lau JY Induction of cell proliferation arrest and apoptosis in hepatoma cells through adenoviral-mediated transfer of p53 gene J Hepatol 2000 32 771 782 10845664 10.1016/S0168-8278(00)80246-9
Mitry RR Sarraf CE Havlik R Habib NA Detection of adenovirus and initiation of apoptosis in hepatocellular carcinoma cells after Ad-p53 treatment Hepatology 2000 31 885 889 10733544 10.1053/he.2000.5631
Ishimaru D Andrade LR Teixeira LS Quesado PA Maiolino LM Lopez PM Cordeiro Y Costa LT Heckl WM Weissmuller G Foguel D Silva JL Fibrillar aggregates of the tumor suppressor p53 core domain Biochemistry 2003 42 9022 9027 12885235 10.1021/bi034218k
|
16117829
|
PMC1224858
|
CC BY
|
2021-01-04 16:24:16
|
no
|
Cancer Cell Int. 2005 Aug 23; 5:27
|
utf-8
|
Cancer Cell Int
| 2,005 |
10.1186/1475-2867-5-27
|
oa_comm
|
==== Front
Cardiovasc UltrasoundCardiovascular Ultrasound1476-7120BioMed Central London 1476-7120-3-251613891910.1186/1476-7120-3-25Case ReportComplications during pharmacological stress echocardiography: a video-case series Varga Albert [email protected] Giuliano [email protected] Ferenc [email protected] Riccardo [email protected] Rafael [email protected] Eugenio [email protected] 2nd Department of Medicine, University of Sciences, Szeged, Hungary2 Institute of Clinical Physiology, Pisa, Italy3 Town Hospital, Orosháza, Hungary4 Cardiology, Department of Medicine and Surgery, University School of Medicine, S. Paolo Academic Hospital, Milan, Italy5 Research Center La Fe Hospital, Valencia, Spain2005 2 9 2005 3 25 25 30 5 2005 2 9 2005 Copyright © 2005 Varga et al; licensee BioMed Central Ltd.2005Varga et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Stress echocardiography is a cost-effective tool for the modern noninvasive diagnosis of coronary artery disease. Several physical and pharmacological stresses are used in combination with echocardiographic imaging, usually exercise, dobutamine and dipyridamole. The safety of a stress is (or should be) a major determinant in the choice of testing. Although large scale single center experiences and multicenter trial information are available for both dobutamine and dipyridamole stress echo testing, complications or side effects still can occur even in the most experienced laboratories with the most skilled operators.
Case presentation
We decided to present a case collection of severe complications during pharmacological stress echo testing, including a ventricular tachycardia, cardiogenic shock, transient ischemic attack, torsade de pointe, fatal ventricular fibrillation, and free wall rupture.
Conclusion
We believe that, in this field, every past complication described is a future complication avoided; what happens in your lab is more true of what you read in journals; and Good Clinical Practice is not "not having complications", but to describe the complications you had.
==== Body
Background
The safety of the stress test is a major issue in deciding its practicability and cost-effectiveness – yet, many major complications remain "unmentioned and unheard", for several reasons – mainly related to the "file drawer" bias, lack of time ("busy agenda bias") or unfamiliarity with the technicalities of scientific communication (editorial "black box bias").
Stress echocardiography is a cost-effective tool for the modern noninvasive diagnosis of coronary artery disease [1]. Several physical and pharmacological stresses are used in combination with echocardiographic imaging, usually exercise, dobutamine and dipyridamole. The safety of a stress is (or should be) a major determinant in the choice of testing. Although large scale single center experiences and multicenter trial information are available for both dobutamine and dipyridamole [2-6] stress echo testing, complications or side effects still can occur even in the most experienced laboratories with the most skilled operators. We believe that, in this field, every past complication described is a future complication avoided; what happens in your lab is more true than what you read in journals; and Good Clinical Practice is not "not having complications", but to describe the complications you had. Therefore, we decided to present an unusual case series, consisting in a collection of severe complications during pharmacological stress echo testing.
Case presentation
Case 1
A 73 year-old male patient, with a previous PTCA (percutaneous transluminal coronary angioplasty) of the left anterior descending artery and ramus intermedius, underwent a dipyridamole stress testing following a nondiagnostic exercise EKG (the exercise was terminated because of the occurrence of non sustained ventricular tachycardia). The baseline echo revealed an apical hypokinesis (additional file 1) which did not change during the test, however ventricular tachycardia developed again during dipyridamole echo (additional file 2). Lesson: it is useless to expose a patient with known coronary artery disease and a previously complicated test to another stressor. Indication must be appropriate.
Case 2
An 81 year-old female, with symptomatic and hemodynamically significant aortic stenosis and normal coronary angiogram underwent a high dose dipyridamole stress echo testing. The baseline wall motion was normal (additional file 3). The patient fell in cardiogen shock and had a transient ischemic attack of the brain following a negative test (additional file 4). Lesson: another dangerous experiment on a patient with already diagnosed normal coronary arteries. Indication must be always appropriate.
Case 3
A 57 year-old male patient with abdominal pain and claudicatio intermittens was studied with dobutamine echocardiography. Soon after the first (5 mcg/Kg/min) dose the patient had ventricular extrasystoles (additional file 5) and during the 20 mcg/Kg/min dose of dobutamine, Torsade de points ventricular tachycardia evolved (additional file 6). Lesson: in patients with arrhytmias in resting conditions, dobutamine can often provoke dangerous tachycardias. In this group of patients dipyridamole could be the first choice.
Case 4
A 55 year-old male patient with previous posterior myocardial infarction, quadruple by-pass, depressed left ventricular function and chest pain was sent to the echo lab for assessment of myocardial viability (additional file 7). Low dose dobutamine echo was performed, however, following the 10 mcg/Kg/min dose a fatal ventricular fibrillation developed (additional file 8). Lesson: there must always be an attending physician during pharmacological stress echo testing with all necessary equipment for reanimation. Dobutamine can provoke arrhytmias even in low doses.
Case 5
A 66 year-old male patient with a recent (12 days old) inferior infarction and inferior aneurysm underwent a high dose dobutamine stress test. A huge aneurysm of the inferior wall was present on the baseline echocardiogram (additional file 9). The patient died following an acute cardiac rupture (additional file 10). Lesson: indications for testing must always be first class, and in patients with recent infarction and aneurysm dipyridamole should be the first choice.
Conclusion
As stated in the American College of Cardiology/American Heart Association Clinical Competence Statement on Stress Testing, cognitive skills are required to attain competence in the direct supervision of stress echocardiographic tests, but not only the knowledge of the complications of different pharmacological agents but also the knowledge of their complication rate is important [7]. Therefore, both the patient and the physician, should be fully aware of the rate of complications during the application of all forms of stress. It is our stress policy, in the everyday echo lab activity, to strictly adopt the following criteria based on conventional wisdom and evidence-based medicine: 1) Avoid contraindications; 2) Never exceed standard dosages; 3) Perform the test after signed information consent has been obtained; 4) There must always be an attending physician; 5) Outpatients should be kept for 60' in the waiting room after testing; 6) Indications for testing must be class first class.
Supplementary Material
Additional File 1
The baseline echo (apical 4 chamber view) with apical hypokinesis.
Click here for file
Additional File 2
No change in wall motion, but ventricular tachycardia developed at peak stress.
Click here for file
Additional File 3
Resting parasternal short axis view and apical 4 chamber view with normal regional left ventricular wall motion.
Click here for file
Additional File 4
Following the dipyridamole administration cardiogenic shock occurred. Depressed global left ventricular function can be seen both from parasternal long axis view and apical 4 chamber view.
Click here for file
Additional File 5
Apical 4 chamber view during low dose dobutamine.
Click here for file
Additional File 6
Parasternal long axis view. The initiation of the torsade de pointe ventricular tachycardia.
Click here for file
Additional File 7
Apical 4 chamber view and apical long axis view. Apical and posterior akinesia on the resting images.
Click here for file
Additional File 8
Ventricular fibrillation following a low dose dobutamine.
Click here for file
Additional File 9
Quad-screen image of a patient with inferior aneurysm.
Click here for file
Additional File 10
The image of the heart following a cardiac rupture with huge pericardial effusion.
Click here for file
==== Refs
Cheitlin MD Alpert JS Armstrong WF Aurigemma GP Beller GA Bierman FZ Davidson TW Davis JL Douglas PS Gillam LD ACC/AHA Guidelines for the Clinical Application of Echocardiography. A report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (Committee on Clinical Application of Echocardiography). Developed in collaboration with the American Society of Echocardiography Circulation 1997 95 1686 744 9118558
Picano E Marini C Pirelli S Maffei S Bolognese L Chiriatti G Chiarella F Orlandini A Seveso G Colosso MQ Safety of intravenous high-dose dipyridamole echocardiography. The Echo-Persantine International Cooperative Study Group Am J Cardiol 1992 70 252 258 Please check if the page reference is correct 1626516 10.1016/0002-9149(92)91284-B
Picano E Mathias W JrPingitore A Bigi R Previtali M on behalf of the Echo Dobutamine International Cooperative study group Safety and tolerability of dobutamine-atropine stress echocardiography: a prospective multicentre study Lancet 1994 29 1190 1192 7934540 10.1016/S0140-6736(94)90508-8
Secknus MA Marwick TH Evolution of dobutamine echocardiography protocols and indications: safety and side effects in 3,011 studies over 5 years J Am Coll Cardiol 1997 29 1234 1240 9137218 10.1016/S0735-1097(97)00039-9
Mathias W JrArruda A Santos FC Arruda AL Mattos E Osorio A Campos O Gil M Andrade JL Carvalho AC Safety of dobutamine-atropine stress echocardiography: A prospective experience of 4033 consecutive studies J Am Soc Echocardiogr 1999 12 785 791 10511646
Lattanzi F Picano E Adamo E Varga A Dobutamine stress echocardiography: safety in diagnosing coronary artery disease Drug Saf 2000 22 251 262 10789822
Rodgers GP Ayanian JZ Balady G Beasley JW Brown KA Gervino EV Paridon S Quinones M Schlant RC Winters WL JrAchord JL Boone AW Hirshfeld JW JrLorell BH Rodgers GP Tracy CM Weitz HH American College of Cardiology/American Heart Association Clinical Competence statement on stress testing: a report of the American College of Cardiology/American Heart Association/American College of Physicians-American Society of Internal Medicine Task Force on Clinical Competence J Am Coll Cardiol 2000 36 1441 1453 11028516 10.1016/S0735-1097(00)01029-9
|
16138919
|
PMC1224859
|
CC BY
|
2021-01-04 16:38:31
|
no
|
Cardiovasc Ultrasound. 2005 Sep 2; 3:25
|
utf-8
|
Cardiovasc Ultrasound
| 2,005 |
10.1186/1476-7120-3-25
|
oa_comm
|
==== Front
Cardiovasc UltrasoundCardiovascular Ultrasound1476-7120BioMed Central London 1476-7120-3-281615940110.1186/1476-7120-3-28Case ReportPlatypnea-orthodeoxia associated with a fenestrated atrial septal aneurysm: Case Report van Gaal William J [email protected] Majo [email protected] Elizabeth [email protected] George [email protected] Mark [email protected] Department of Cardiology, John Radcliffe Hospital, Headington, Oxon, OX3 9DU, UK2 Department of Cardiology, Austin Health, Heidelberg, Victoria, 3084, Australia2005 13 9 2005 3 28 28 31 8 2005 13 9 2005 Copyright © 2005 van Gaal et al; licensee BioMed Central Ltd.2005van Gaal et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Platypnea-orthodeoxia describes the condition of combined dyspnea and hypoxia respectively, whilst in the upright position, which improves in the recumbent position.
Case Report
We present a case of platypnea-orthodeoxia due to a fenestrated atrial septal defect associated with an atrial septal aneurysm. Due to the fenestrated nature of the atrial septal defect, surgical rather than percutaneous correction was performed.
Conclusion
A high index of suspicion is required to diagnose the syndrome of platypnea-orthodeoxia. Careful echocardiographic evaluation is required to identify the syndrome, and to determine suitability for percutaneous repair.
==== Body
Background
Persistence of a patent foramen ovale, or existence of an atrial septal defect, are both common and usually asymptomatic in the absence of right-to-left shunting, which usually occurs in the presence of elevated pulmonary artery and right heart pressures. Platypnea-orthodeoxia describes the condition of combined dyspnea and hypoxia respectively, whilst in the upright position, which improves in the recumbent position. The syndrome can occur with pulmonary or intra-cardiac shunts, however why right-to-left intra-cardiac shunting should only occur in the upright position is not fully understood.
Case Report
A 75-year-old female presented with dyspnea. Her past medical history included hypertension and paroxysmal atrial fibrillation. Other than tachypnea, the cardiorespiratory examination was normal. The 12 lead electrocardiogram showed sinus rhythm with left atrial enlargement and non-specific T wave changes laterally. Arterial blood whilst breathing 12 litres of oxygen per minute via face mask was taken: pH 7.44, PaCO2 36 mmHg, PaO2 52 mmHg, O2 saturation 88%.
Full blood examination and D-Dimer were normal. A computed tomography pulmonary angiogram was normal. During admission there was fluctuating hypoxia, dyspnea and central cyanosis. An intra-cardiac shunt was suspected. A transesophageal echocardiogram (TEE) showed concentric left ventricular hypertrophy and normal right ventricular function. There was an atrial septal defect (ASD) with an aneurysmal interatrial septum (figure 1a), and mild left-to-right shunting only. Color flow Doppler across the atrial septum demonstrated multiple fenestrations (figure 1b). No right-to-left shunting was demonstrable with a saline contrast study, and a Swan Ganz catheter confirmed normal pulmonary artery pressures.
Figure 1 Aneurysmal atrial septal defect. (A) Transesophageal echocardiogram showing the left atrium (LA) and right atrium (RA). The atrial septal aneurysm can be seen bulging into the left atrium, and the ASD is shown (arrow). (B) Color flow Doppler across the atrial septum demonstrates multiple fenestrations with three distinct jets seen crossing the septum.
It was noted that her hypoxia was relieved in the recumbent position. On breathing room air whilst supine, the SpO2 was consistently above 90%, however in the sitting position, the SpO2 repeatedly dropped to as low as 70% with worsening dyspnea and development of central cyanosis. A repeat TEE was performed in both the supine and sitting position. At a 70 degree incline, there was significant right-to-left shunting. The aortic root appeared mildly dilated with distortion of the interatrial septum contributing to the shunt. Due to the fenestrated nature of the defect, surgical correction was performed (figures 2a and 2b). Twelve months later the patient is doing well without recurrence of symptoms.
Figure 2 Introperative findings. (A) Intra-operative photograph demonstrating the fenestrated nature of the atrial septal defect, and the final result following surgical repair (B).
Conclusion
Platypnea-orthodeoxia has been described to occur in pulmonary arteriovenous shunts, pulmonary parenchymal shunts (as in the hepatopulmonary syndrome), or with intra-cardiac right-to-left shunts [1]. One of the earliest reported series of seven patients was published in 1984 [2]. Of these, four had prior pneumonectomy, and two had pulmonary embolism, although it was noted in each case that pulmonary hypertension was absent, and right-sided haemodynamics were normal. Since that time there have been a number of reports describing platypnea-orthodeoxia without overt lung disease. A review of 31 such cases from 1949–1997 was performed by Faller et al in 2000, and included platypnea-orthodeoxia associated with diaphragmatic paralysis, right atrial myxoma, restrictive cardiac disease, kyphoscoliosis and elongation of the ascending aorta [3]. Rare reports of platypnea-orthodeoxia in association with atrial septal aneurysms have mostly been associated with patent foramen ovale [3-5]. In our case, the syndrome was associated with an atrial septal aneurysm and ASD with multiple fenestrations – a complicating feature that precluded the patient from percutaneous closure, and which was confirmed at the time of surgical correction. Although the majority of defects are amenable to percutaneous intervention [6,7], careful evaluation of the atrial septum including the use of color flow Doppler is required to ensure suitability for percutaneous repair.
Several mechanisms have been theorized to cause right-to-left shunting in patients with atrial communications and normal right heart pressures. Theories include compression of the right atrium in the upright posture, decreased compliance of the right ventricle, and development of abnormal anatomy between the vena cava and the atrial septum [3,8]. The latter most likely contributes to the syndrome in our patient, with the mildly dilated aorta and atrial septal aneurysm permitting right-to-left shunting in the face of normal right-sided pressures.
Our case highlights two important aspects in the management of patients with orthostatic hypoxia and dyspnea. First there is the difficulty in making the diagnosis of right-to-left shunting in the presence of normal right heart and pulmonary artery pressures. Many patients will undergo investigation for pulmonary embolus and other alternative diagnoses. A high index of suspicion is therefore required to make a prompt diagnosis. The second is the need for careful echocardiographic evaluation with the use of controlled tilt to identify the syndrome, and determine the suitability for percutaneous repair.
Physicians should be aware of the syndrome of platypnea-orthodeoxia which is now a well recognized syndrome, with nearly 50 reported cases in the literature. A high index of suspicion is required to make the diagnosis, with echocardiography using controlled tilting the most useful investigational method. Careful evaluation of patients with respect to suitability for percutaneous repair is needed, as patients with fenestrated defects or large aneurysmal components may be best served with surgical correction to achieve complete closure of the septum, and ensure no persistence of right-to-left shunting.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
WVG: Conceived of the study, participated in the study design and drafted the manuscript. EJ, MJ: Performed the echocardiography, provided echocardiography images, participated in the study design and helped drafting the manuscript. GM, MH: Performed invasive procedures, provided surgical images and participated in the study design.
==== Refs
Robin ED McCauley RF An analysis of platypnea-orthodeoxia syndrome including a "new" therapeutic approach Chest 1997 112 1449 1451 9404736
Seward JB Hayes DL Smith HC Williams DE Rosenow EC Reeder GS Piehler JM Tajik AJ Platypnea-orthodeoxia: clinical profile, diagnostic workup, management, and report of seven cases Mayo Clin Proc 1984 59 221 231 6708599
Faller M Kessler R Chaouat A Ehrhart M Petit H Weitzenblum E Platypnea-orthodeoxia syndrome related to an aortic aneurysm combined with an aneurysm of the atrial septum Chest 2000 118 553 557 10936158 10.1378/chest.118.2.553
Acharya SS Kartan R A case of orthodeoxia caused by an atrial septal aneurysm Chest 2000 118 871 874 10988220 10.1378/chest.118.3.871
Timmermans C Frans E Herregods C Decramer M Daenen W De Geest H Platypnea-orthodeoxia syndrome: a report of two cases Acta Cardiol 1993 48 583 590 8122482
Rao PS Palacios IF Bach RG Bitar SR Sideris EB Platypnea-orthodeoxia: management by transcatheter buttoned device implantation Catheter Cardiovasc Interv 2001 54 77 82 11553954 10.1002/ccd.1243
Rao PS Transcatheter management of platypnea-orthodeoxia syndrome J Invasive Cardiol 2004 16 583 584 15505356
Godart F Rey C Prat A Vincentelli A Chmait A Francart C Porte H Atrial right-to-left shunting causing severe hypoxaemia despite normal right-sided pressures. Report of 11 consecutive cases corrected by percutaneous closure Eur Heart J 2000 21 483 489 10681489 10.1053/euhj.1999.1944
|
16159401
|
PMC1224860
|
CC BY
|
2021-01-04 16:38:31
|
no
|
Cardiovasc Ultrasound. 2005 Sep 13; 3:28
|
utf-8
|
Cardiovasc Ultrasound
| 2,005 |
10.1186/1476-7120-3-28
|
oa_comm
|
==== Front
Curr Control Trials Cardiovasc MedCurrent Controlled Trials in Cardiovascular Medicine1468-67081468-6694BioMed Central 1468-6708-6-131612487810.1186/1468-6708-6-13ResearchCoronary artery bypass surgery in high-risk patients Kunt Alper Sami [email protected]ın Osman Tansel [email protected] Mehmet Halit [email protected] Department of Cardiovascular Surgery, Harran University Research Hospital, Sanlıurfa, Turkey2005 26 8 2005 6 1 13 13 4 3 2005 26 8 2005 Copyright © 2005 Kunt et al; licensee BioMed Central Ltd.2005Kunt et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
In high-risk coronary artery bypass patients; off-pump versus on-pump surgical strategies still remain a matter of debate, regarding which method results in a lower incidence of perioperative mortality and morbidity. We describe our experience in the treatment of high-risk coronary artery patients and compare patients assigned to on-pump and off-pump surgery.
Methods
From March 2002 to July 2004, 86 patients with EuroSCOREs > 5 underwent myocardial revascularization with or without cardiopulmonary bypass. Patients were assigned to off-pump surgery (40) or on-pump surgery (46) based on coronary anatomy coupled with the likelihood of achieving complete revascularization.
Results
Those patients undergoing off-pump surgery had significantly poorer left ventricular function than those undergoing on-pump surgery (28.6 ± 5.8% vs. 40.5 ± 7.4%, respectively, p < 0.05) and also had higher Euroscore values (7.26 ± 1.4 vs. 12.1 ± 1.8, respectively, p < 0.05). Differences between the two groups were nonsignificant with regard to number of grafts per patient, mean duration of surgery, anesthesia and operating room time, length of stay intensive care unit (ICU) and rate of postoperative atrial fibrillation
Conclusion
Utilization of off-pump coronary artery bypass graft (CABG) does not confer significant clinical advantages in all high-risk patients. This review suggest that off-pump coronary revascularization may represent an alternative approach for treatment of patients with Euroscore ≥ 10 and left ventricular function ≤ 30%.
Coronary arteryhigh riskoff-pumpon-pumpsurgery
==== Body
Introduction
Early mortality and morbidity represent clinical outcomes that have been used in many research models examining patients undergoing coronary artery bypass graft (CABG) surgery [1-6]. Studies utilizing these endpoints have provided valuable information for determining the indications for surgery, estimating the need for various resources and implementing quality control monitoring of surgeons and institutions. The European System for Cardiac Operative Risk Evaluation (EuroSCORE) used logistic regression analysis to identify and give appropriate weight to various risk factors related to in-hospital mortality in adult cardiac operations [7]. Standard EuroSCORE was first introduced in 1999 [8] as an additive system and has gained wide acceptance in Europe [9]. The logistic algorithm, which recently, became available [10], appears to be a better risk predictor of mortality and morbidity in CABG patients, especially among high-risk patients [11,12].
A EuroSCORE value >5 reflects a high level of risk in patients with coronary artery disease (CAD). Severe LV dysfunction represent another clinical outcome that has been reported to serve as an independent predictor of operative mortality in patiens undergoing CABG [13]. Coronary artery disease patients with reduced left ventricular function appear to benefit more from CABG than from medical therapy [14].
In high-risk CAD patients, surgical myocardial revascularization often produces poor results leading to significant mortality and morbidity [15]. Management of high-risk patients remains unclear. We describe our experience in the treatment of high-risk CAD patients undergoing on-pump versus off-pump surgery.
Materials and methods
Clinical data collection
The records of 86 consecutive high-risk patients who underwent primary isolated CABG at Harran University Research Hospital between January 2002 and December 2004 were reviewed retrospectively. The study was approved by the ethics committee of the Harran University Research Hospital, and informed consent was obtained from all patients. Patients were considered to be high-risk were included in the study if they had a preoperative Euroscore of ≥5 on admission to the hospital.
Preoperative and postoperative patient data were reviewed using registry databases, medical notes and charts. Forty patients underwent CABG using the off-pump technique while 46 patients were operated on using the conventional on-pump technique. Selection of either technique was done by the individual surgeon, and was based on his experience and preference. No randomization was involved in this cohort of patients.
Operative technique
Anesthesia
All routine cardiac medications were continued up until the morning of surgery. After premedication with 5 mg intramuscular Midazolam and 0.18 mg/kg of intrathecal morphine diluted in 4 ml of serum physiologic solution for postoperative analgesia, anesthesia was induced using 0.3 mg/kg of etomidate, μg/kg of remifentanil and 0.6 mg/kg of rocuronium intravenously. After endotracheal intubation, desflurane (3–10%) and remifentanil 0.25–1.0 μg/kg/min in air/oxygen and rocuronium were given to maintain anesthesia.
On-pump technique
After the standard median sternotomy, aorta-right atrial cannulation and cardiopulmonary bypass were performed in on-pump patients. During cardiopulmonary bypass (CPB), hematocrit, mean arterial pressure, and pump flow were maintained at 20–30%, 50–80 mmHg, and 2.2–2.5 l/m2, respectively. Adequacy of tissue perfusion was monitored, as well as arteriovenous partial carbon dioxide difference (Pv-a CO2), urine output, and base deficit. Patients were cooled to 32°C with moderate hypothermia. Desflurane-remifentanil anesthesia was administered during CPB. Revascularization procedures were performed with aortic cross-clamping. During myocardial ischemia antegrade cold hyperkalemic crystalloid cardioplegia was used (Plegisol®, Abbot Laboratories, IL, and USA). After completion of distal anastomosis, the proximal anastomosis was performed to the ascending aorta by using a side-biting clamp.
Off-pump technique
Left internal mammary artery and saphenous vein grafts were harvested for grafting for off-pump patients. To provide better access to lateral and posterior target vessels the pericardium was retracted by two or three deep sutures and two sponges were placed under the heart. Neither a heart stabilizer nor intraluminal shunts were used. Silicone snare sutures were placed proximal and distal to the anastomosis in order to provide a bloodless field. Remifentanil infusion and desflurane were discontinued at skin closure.
Statistical analysis
All clinical data were expressed as mean ± standard deviation. Data processing and statistical analysis were performed using SPSS statistical software package for Windows. The student's t-test and chi-square test were used. A p value < 0.05 was considered statistically significant.
Results
Preoperative characteristics
The preoperative characteristics of patients are listed in Table 1. Off-pump patients experienced significantly poorer left ventricular (LV) function (ejection fraction (EF) ≤ 30%) (p < 0.05) and significantly higher Euroscores 12.1 ± 1.8 (p ≤ 0.05). Respiratory problems included chronic obstructive pulmonary disease (COPD) that required active treatment at the time of surgery.
Table 1 Preoperative data of patients
Characteristics of the 86 patients studied Overall population (n = 86) Off-pump (n = 40) On-pump (n = 46) p
Mean age at operation (years) 61.5 ± 8.9 63 ± 12 60 ± 7 0.82
Female sex (%) 10.46 12.50 8.69 0.73
Smoker (%) 27.90 35.00 21.73 0.23
Diabetes (%) 23.25 27.50 19.56 0.38
Hypertension (%) 40.69 50.00 32.60 0.10
Chronic obstructive pulmonary disease (%) 6.97 12.50 2.10 0.06
Mean Left Ventricle
Ejection Fraction (±SD) 34.2 ± 9.1 28.6 ± 5.8 40.5 ± 7.9 0.032
Mean Euroscore (±SD) 9.7 ± 3.1 12.1 ± 1.8 7.26 ± 1.8 0.022
Operative characteristics
The operative characteristics of patients are presented in Table 2. There was no significant difference in the number of grafts between the off-pump and on-pump patients (2.03 ± 0.7 vs.1.99 ± 0.6 grafts per patient respectively, p = 0.15). Nor were there any significant differences between on-and off pump patients with regard to duration surgery (105 ± 22 vs. 80 ± 25, respectively, p = 0.26) or aortic cross-clamp time (21.5 ± 7.6 vs. 20.6 ± 7.5 min, respectively, p = 0.861).
Table 2 Intraoperative and postoperative variables.
Characteristics of the 86 patients studied (mean) Overall population (n = 86) Off-pump (n = 40) On-pump (n = 46) p
Distal anastomosis time (min) 22.75 ± 5.8 20.6 ± 7.5 21.5 ± 7.6 0.861
Duration of surgery (min) 92.50 ± 25 80 ± 25 105 ± 22 0.26
Duration of anesthesia (min) 118.5 ± 28.7 105 ± 19 132 ± 34 0.29
Operating room time (min) 134.5 ± 22.2 124 ± 15 145 ± 26 0.33
Number of grafts/patients 1.99 ± 0.6 1.93 ± 0.7 2.03 ± 0.7 0.15
Extubation time (min) 19.5 ± 10.25 15 ± 9 24 ± 11 0.33
Length of stay in ICU (h) 19 ± 5.2 18 ± 4 20 ± 7 0.69
Length of stay in hospital(days) 7.5 ± 1.5 8 ± 1 7 ± 2 0.48
Postoperative morbidity
Overall ten (11.62%) overall patients developed peri-operative atrial fibrillation, with no significant difference between the 2 groups. We could not show any significant difference between the off-pump and on-pump patients. Three (7.5%) off-pump patients developed low cardiac output syndrome (LCOS) in the postoperative period compared to 1 (2.1%) on-pump patients (p = 0.24). There was no statistically significant difference between the patients with regard to other complications (Table 3).
Table 3 Complications after coronary artery bypass grafting after 30 days
Complications of the 86 patients studied Overall population(%) (n = 86) Off-pump(%) (n = 40) On-pump(%) (n = 46) p
Atrial fibrillation 11.6 17.5 6.5 0.11
LCOS 4.6 7.5 2.1 0.24
Bleeding 3.4 2.5 4.3 0.64
Re-operaton 2.3 2.5 2.1 0.92
Re-ıntubation 3.4 2.5 4.3 0.64
Renal complications 5.8 7.5 4.3 0.53
Pulmonary complications 4.6 5 4.3 0.89
IABP 4.6 7.5 2.1 2.1
30-day mortality 3.4 7.5 0 0.06
LCOS: Low cardiac output syndrome
IABP: Intraaortic balloon pump
As noted in Table 2 the intensive care unit (ICU) stay for off-pump patients was 18 ± 4 h while for on-pump patients it was 20 ± 7 h (p = 0.69). The hospital stay was 8 ± 1 days for the off-pump patients and 7 ± 2 days for on-pump patients (p = 0.48).
Postoperative mortality
We defined postoperative mortality as death within the 30 days following the operation (Table 3). There were three (7.5%) deaths in the off-pump patients compared to no death in the on-pump patients (p = 0.06) within 30 days postoperatively. The three of off-pump deaths included two due to cardiac causes, one due to multi-organ failure (MOF).
Discussion
Based on our findings in this retrospective comparative study, use of the off-pump technique for myocardial revascularization in extreme preoperative high risk (Euroscore ≥ 10, EF < 30%) patients reduces the incidence of perioperative morbidity and mortality, ICU stay and other complications when compared to on-pump patients.
European and US institutional data demonstrate that patients undergoing CABG are progressively older and have a worse cardiac status and a higher incidence of systemic co-morbidities. It seems highly likely that this trend will increase and that high-risk patients will represent a greater proportion of patients treated by cardiac surgeons [16-21].
The initial application of the off-pump technique in the early nineties was mainly directed to highly selected and relatively low-risk surgical patients [22]. Since then there has been a growing body of evidence suggesting many potential advantages of the off-pump technique over the conventional CPB in different groups of high-risk patients [23,24].
In this setting the standard surgical strategy is often inappropriate and carries substantial operative risks. To date, however, to date few reports have focused on the results of off-pump versus on-pump conventional surgery in high-risk patients. In patients with acute or chronically ischemic myocardium and poorly functioning left ventricles, off-pump and on-pump surgical revascularization have been shown to improve survival, improve functional status or control ischemic symptoms, and diminish the prevalence of sudden cardiac deaths caused by arrhythmias [23-25]. Moreover, methodological issues and the heterogeneity of reported results have precluded any definitive conclusion on the possibility that off-pump surgery can reduce the operative risk of complex CABG patients [23-25].
Our study is a non-randomized comparative retrospective study of patients who underwent first-time isolated coronary bypass surgery on- or off-pump in our center. Preoperative variables in the overall patients showed little variation between the on-pump and off-pump patients except for EF (p = 0.032) and for Euroscore (p = 0.022) which were significantly lower in off-pump patients. The similar number of anastomoses performed in the on-pump and off-pump patients. New onset of atrial fibrillation was reduced in the on-pump patients in our series but not significantly (p = 0.11). There were no significant differences in the incidence of perioperative LCOS, renal complications, pulmonary complications and intraaortic ballon pump (IABP) using the off-pump vs. the on-pump technique. Similar fidings were also noted for intubation time, intensive care unit stay and hospital stay.
Data from the Euroscore project indicate that patients in the highest risk groups who undergo conventional surgery can have hospital mortality as high as 11.2% [8]. The present series describes our experience in the treatment of high-risk CABG patient and compares patients assigned to on-pump vs. off-pump revascularization. Overall mortality (3 of 86 patients, 3.4%) was one of the lowest reported in patients of this type. In contrast to other publications, the hospital mortality in our series was not significantly different between off-pump and on-pump patients. We agree with other authors that the improved results may be attributed to advances in myocardial protection, surgical technique, and perioperative care.
In conclusion, our data suggest that the adoption of off-pump CABG does not confer significant clinical advantages in all high-risk patients. This review supports the off-pump coronary revascularization, which may represent an alternative approach for treating patients with Euroscore ≥ 10 who have left ventricular function ≤ 30%.
==== Refs
Gurler S Gebhard A Godehardt E Boeken U Feindt P Gams E EuroSCORE as a predictor for complications and outcome Thorac Cardiovasc Surg 2003 51 73 77 12730814 10.1055/s-2003-38988
Huijskes RV Rosseel PM Tijssen JG Outcome prediction in coronary artery bypass grafting and valve surgery in the Netherlands: development of the Amphiascore and its comparison with the Euroscore Eur J Cardiothorac Surg 2003 24 741 749 14583307 10.1016/S1010-7940(03)00471-8
Immer F Habicht J Nessensohn K Bernet F Stulz P Kaufmann M Prospective evaluation of 3 risk stratification scores in cardiac surgery Thorac Cardiovasc Surg 2000 48 134 139 10903058 10.1055/s-2000-9638
Kurki TS Jarvinen O Kataja MJ Laurikka J Tarkka M Performance of three preoperative risk indices; CABDEAL, EuroSCORE and Cleveland models in a prospective coronary bypass database Eur J Cardiothorac Surg 2002 21 406 410 11888755 10.1016/S1010-7940(02)00007-6
Pitkanen O Niskanen M Rehnberg S Hippelainen M Hynynen M Intra-institutional prediction of outcome after cardiac surgery: comparison between a locally derived model and the EuroSCORE Eur J Cardiothorac Surg 2000 18 703 710 11113679 10.1016/S1010-7940(00)00579-0
Geissler HJ Holzl P Marohl S Kuhn-Regnier F Mehlhorn U Sudkamp M Risk stratification in heart surgery: comparison of six score systems Eur J Cardiothorac Surg 2000 17 400 406 10773562 10.1016/S1010-7940(00)00385-7
Roques F Nashef SA Michel P Gauducheau E Vincentiis C Baudet E Risk factors and outcome in European cardiac surgery: analysis of the EuroSCORE multinational database of 19030 patients Eur J Cardiothorac Surg 1999 15 816 822 10431864 10.1016/S1010-7940(99)00106-2
Nashef SA Roques F Michel P Gauducheau E Lemeshow S Salamon R European system for cardiac operative risk evaluation (EuroSCORE) EurJ Cardiothorac Surg 1999 16 9 13 10456395 10.1016/S1010-7940(99)00134-7
Roques F Nashef SA Michel P Pinna PP David M Baudet E The EuroSCORE study group. Does EuroSCORE work in individual European countries? Eur J Cardiothorac Surg 2000 18 27 30 10869937 10.1016/S1010-7940(00)00417-6
Roques F Michel P Goldstone AR Nashef SA The logistic EuroSCORE Eur Heart J 2003 24 881 882 12727160 10.1016/S0195-668X(02)00799-6
Michel P Roques F Nashef SA Logistic or additive EuroSCORE for high-risk patients? Eur J Cardiothorac Surg 2003 23 684 687 12754018 10.1016/S1010-7940(03)00074-5
Toumpoulis IK Anagnostopoulos CE Swistel DG DeRose JJ Does EuroSCORE predict length of stay and specific postoperative complications after cardiac surgery? Eur J Cardiothorac Surg 2005 27 1 128 133 15621484 10.1016/j.ejcts.2004.09.020
Ascione R Narayan P Rogers CA Kelvin H Lim H Capoun R Angelini GD Early and midterm clinical outcome in patients with severe left ventricular dysfunction undergoing coronary artery surgery Ann Thorac Surg 2003 76 3 793 9 12963202 10.1016/S0003-4975(03)00664-7
Arom KW Flavin TF Emery RW Kshettry VR Petersen RJ Janey PA Is low ejection fraction safe for off-pump coronary bypass operation? Ann Thorac Surg 2000 70 1021 1025 11016367 10.1016/S0003-4975(00)01761-6
Gaudino M Glieca F Alessandrini F Nasso G Pragliola C Luciani N Morelli M Possati G High risk coronary artery bypass patient: Incidence, surgical strategies, and results Ann Thorac Surg 2004 77 574 80 14759440 10.1016/S0003-4975(03)01534-0
Warner CD Weintraub WS Craver JM Effect of cardiac surgery patient characteristics on patients outcomes from 1981 through 1995 Circulation 1997 96 1575 9 9315549
Estefanaous FG Loop FD Higgins TL Increased risk and decreased morbidity of coronary artery bypass grafting between 1986 and 1994 Ann Thorac Surg 1998 65 383 9 9485233 10.1016/S0003-4975(97)01125-9
Hartz AJ Manley JC Walker JA Trends during 25 years of coronary artery bypass operation at St. Luke's Medical Center in Milwaukee, Wisconsin Ann Thorac Surg 2000 69 829 33 10750768 10.1016/S0003-4975(99)01418-6
Abramow D Tamariz MG Fremes SE Trends in coronary artery bypass surgery results: a recent, 9 year study Ann Thorac Surg 2000 70 84 90 10921687 10.1016/S0003-4975(00)01249-2
Stoica SC Sharples LD Ahmed I Preoperative risk prediction and intraoperative events in cardiac surgery Eur J Cardiothorac Surg 2002 21 41 6 11788254 10.1016/S1010-7940(01)01077-6
Ferguson TB JrHamil BG Peterson ED A decade of change: risk profiles and outcomes for isolated coronary artery bypass graftingprocedures, 1990–1999. A report from the STS National Dtabase Committee and Duke Research Instıtute: Society of Thoracic Surgeons Ann Thorac Surg 2002 73 480 9 11845863 10.1016/S0003-4975(01)03339-2
Benetti F Naselli G Wood M Geffner L Direct myocardial revascularization without extracorpereal circulation: Experience in 700 patients Chest 1991 100 312 316 1677884
Yokoyama T Baumgartner F Gheissari A Caponya F Panagiotides G Declusin R Off-pump versus on-pump coronary bypass in high risk subgroups Ann Thorac Surg 2000 70 1546 1550 11093485 10.1016/S0003-4975(00)01922-6
Arom K Flqvin T Emery R Kshettry V Janey P Petersen R Safety and efficacy of off-pump coronary artery bypass grafting Ann Thorac Surg 2000 69 704 710 10750747 10.1016/S0003-4975(99)01510-6
Bittner HB Savitt MA McKeown PP Off-pump coronary artery bypass grafting. Excellent results in a group of selected high risk patients J Cardiovasc Surg 2001 42 451 6 11455277
|
16124878
|
PMC1224861
|
CC BY
|
2021-01-04 16:47:32
|
no
|
Curr Control Trials Cardiovasc Med. 2005 Aug 26; 6(1):13
|
utf-8
|
Curr Control Trials Cardiovasc Med
| 2,005 |
10.1186/1468-6708-6-13
|
oa_comm
|
==== Front
J Circadian RhythmsJournal of Circadian Rhythms1740-3391BioMed Central London 1740-3391-3-111615330110.1186/1740-3391-3-11ResearchCircadian phase-shifting effects of a laboratory environment: a clinical trial with bright and dim light Youngstedt Shawn D [email protected] Daniel F [email protected] Jeffrey A [email protected] Katharine M [email protected] Department of Exercise Science, Norman J. Arnold School of Public Health, University of South Carolina, Columbia, SC 29208, USA2 Department of Psychiatry and Sam and Rose Stein Institute for Research on Aging, University of California, San Diego, USA2005 9 9 2005 3 11 11 30 8 2005 9 9 2005 Copyright © 2005 Youngstedt et al; licensee BioMed Central Ltd.2005Youngstedt et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Our aims were to examine the influence of different bright light schedules on mood, sleep, and circadian organization in older adults (n = 60, ages 60–79 years) with insomnia and/or depression, contrasting with responses of young, healthy controls (n = 30, ages 20–40 years).
Methods
Volunteers were assessed for one week in their home environments. Urine was collected over two 24-hour periods to establish baseline acrophase of 6-sulphatoxymelatonin (aMT6s) excretion. Immediately following home recording, volunteers spent five nights and four days in the laboratory. Sleep periods were fixed at eight hours in darkness, consistent with the volunteers' usual sleep periods. Volunteers were randomly assigned to one of three light treatments (four hours per day) within the wake period: (A) two hours of 3,000 lux at 1–3 hours and 13–15 hours after arising; (B) four hours of 3,000 lux at 6–10 hours after arising; (C) four hours of dim placebo light at 6–10 hours after arising. Lighting was 50 lux during the remainder of wakefulness. The resulting aMT6s acrophase was determined during the final 30 hours in the laboratory.
Results
Neither mood nor total melatonin excretion differed significantly by treatment. For the three light treatments, significant and similar phase-response plots were found, indicating that the shift in aMT6s acrophase was dependent upon the circadian time of treatment. The changes in circadian timing were not significantly correlated to changes in sleep or mood.
Conclusion
The trial failed to demonstrate photoperiodic effects. The results suggest that even low levels of illumination and/or fixed timing of behavior had significant phase-shifting effects.
==== Body
Introduction
Older adults have an altered synchronization of circadian rhythms compared to young adults [1-4]. Circadian misalignment of rhythms or malsynchronization might contribute to many age-related disorders of sleep or mood, as has been observed in conjunction with shift-work and jet lag.
It has been hypothesized that age-related circadian malsynchronization might be explained by reduced exposure to light and other zeitgebers in older adults. However, studies have found that, when compared to young adults, older adults are exposed to at least as much bright light (e.g., in San Diego [5,6]) and to environmental and social zeitgebers of even greater regularity [7]. Nonetheless, it is likely that retinohypothalamic neurotransmission of light to the SCN is compromised in older adults due to glaucoma, macular degeneration, senile miosis, and other eye problems [8,9]. Moreover, age-related neurodegeneration of the suprachiasmatic nuclei (SCN) [10] could make the SCN less responsive to light in older adults. Preliminary evidence suggests that aging subjects may display smaller phase shifts to light stimuli [11,12]. Thus, older adults might require increased exposure to light or other synchronizers for adequate circadian entrainment.
Although light exposure is apparently the most important circadian synchronizer, careful regulation of the sleep-wake schedule [13,14], as well as physical activity [15,16] and social interaction [17], can also influence circadian timing. These non-photic stimuli can produce effects added to those produced by light alone.
Appropriately timed exposure to bright light and other zeitgebers might help correct circadian malsynchronization and alleviate sleep and mood problems that are common in older adults. Evidence indicates that light can also have photoperiodic effects on the organization of the human circadian system [18]. The main aims of this study were: (1) to contrast the influences of different bright light schedules on circadian and photoperiodic organization in older adults with insomnia and/or depression and in young, healthy controls; (2) to examine whether circadian "phase correction", i.e., shifting the circadian system to more normal timing, could improve sleep or mood among the older adults. However, since differing light treatments produced similar results, this presentation will emphasize the phase changes produced.
Methods
Subjects
Volunteers were 72 older adults (49 women, 23 men) ages 60–79 years, who were selected for complaints of insomnia and/or depression. The volunteers reported that their symptoms were of sufficient severity to warrant treatment. Volunteers were screened for freedom from melatonin-altering medications (with few exceptions), and the absence of acute health problems. A control group of 30 (n = 15 women, 15 men) young, healthy volunteers ages 20–40 years was studied in parallel. Volunteers signed their consent to participate in the study as approved by the UCSD Institutional Review Board.
Home Baseline Procedures
Volunteers were first assessed for five-seven days in their home environments. Volunteers were asked to maintain their usual sleep and lifestyle habits during this time.
Baseline Urine Collection
Over two 24–30-hour periods (usually days three-four and six-seven at home), volunteers collected their urine samples approximately every two hours during wakefulness plus all voidings during the nighttime sleep period. Volunteers recorded the timing and volume of each collection, and stored 2-ml samples in their freezers. The samples were subsequently transferred to a laboratory -70°C freezer.
Baseline Sleep Assessment
An actigraph with minute-by-minute recordings of wrist activity and illumination was worn throughout home recording, except for short removals for bathing, etc. (Actillume I, Ambulatory Monitoring, Ardsley, New York). The nocturnal sleep periods were determined from actigraphic sleep and illumination recording combined with daily sleep diary data. Objective sleep was scored with a validated algorithm associating wrist movement with electroencephalographically-recorded sleep [19]. For each night, actigraphically-assessed sleep onset latency (SOL), total sleep time (TST), time spent awake after initial sleep onset (WASO), and sleep efficiency were determined. Each morning, subjective ratings of minutes of TST and WASO, and a 100 mm visual analogue rating of insomnia were also recorded. Mean baseline sleep levels were calculated, and have been reported previously [1].
Baseline Mood Assessment
The subjects' depressed moods were assessed on two days (usually days three and six) with the Center for Epidemiologic Studies-Depression (CES-D) questionnaire, which consists of 20 questions with four-point Likert responses (possible range: 0 to 60) [20]. The questionnaire was completed four hours after arising. Mean baseline CES-D was calculated. These data have been reported previously [1].
Laboratory Procedures
Immediately following home recording, volunteers spent five nights and four days in the Circadian Pacemaker Laboratory at UCSD, arriving on Sunday evening two hours before bedtime, and leaving on Friday morning after arising (Figure 1). Each volunteer stayed in an individual apartment equipped with a bed, comfortable chair, television, kitchen, and private shower and bathroom facilities. Volunteers were provided with food of their own choosing (except for alcohol and caffeine) and were free to prepare and eat food ad libitum during the wake periods.
Figure 1 Laboratory protocol. Arriving two hours before their usual bedtime, subjects spent five nights and four days in the laboratory. This figure displays the time of urine collections (shown in red), which began after the last voiding before morning (most participants urinated during the night) and continued through the final morning voiding after the next consecutive night, slightly more than 24 hours.
Sleep periods were fixed at eight hours, timed to correspond approximately with each volunteer's average home-recorded sleep schedule. Illumination levels were <0.5 lux during the sleep periods and ≤50 lux average in the horizontal direction at eye level during the 16-hour periods of wakefulness (except during the bright light treatments, described below). Sleep was discouraged during the wake periods with the aid of video monitoring, though some volunteers fell asleep for brief time periods (approximately two-five minutes). Volunteers were permitted one cup of coffee during the first four hours after arising. Vigorous exercise was not permitted, but light calisthenics and slow walking were allowed. Otherwise, volunteers were free to do what they wished during the wake periods, i.e., watch TV, receive visitors, read, etc. Because priority was placed on assuring that the laboratory experience did not trigger more severe depression, the laboratory staff made special efforts to help the volunteers feel comfortable and engaged in the laboratory experience. It was not uncommon for staff to spend several hours per day playing board games or chatting with a volunteer.
Light Treatments
Volunteers were randomly assigned to one of three light treatments, which were administered for four hours during each of the four days of the experiment (Figure 2). The light treatments were administered via overhead cool-white fluorescent lights, providing relatively even light levels at eye-level throughout the laboratory rooms.
Figure 2 Experimental Light Treatments. Volunteers were randomly assigned to three four-hour light treatments (detailed in this figure) administered on four consecutive days against a background of <0.5 lux during eight-hour sleep periods and 50 lux during 16-hour wake periods. Treatment A was two hours at 3,000 lux from 1–3 hours and 13–15 hours after arising. Treatment B was four hours at 3,000 lux from 6–10 hours after arising. Treatment C was four hours of dim red light placebo from 6–10 hours after arising. Note that the center of each treatment was eight hours after arising, and the abscissa was hours after usual wake time.
Treatment A consisted of two hours of bright light (3,000 lux in the direction of gaze) at one-three hours after arising, as well as at three to one hours before bedtime. The 3000 lux portion of Treatment A was designed to resemble a skeleton LD14:10-long photoperiod. An LD16:8 skeleton might have had greater photoperiodic effect, but there was concern that LD16:8 might disturb sleep excessively. Treatment B consisted of four hours of bright light (3,000 lux) in the middle of the wake period, i.e., six-ten hours after awakening. The bright portion of Treatment B might resemble a short LD4:20 photoperiod, to the extent that the surrounding 50 lux treatment was photoperiodically ineffective. Also, bright light in Treatment B would be expected to fall in a relatively insensitive zone of the light phase response curve. Treatment C, the control treatment, involved four hours of placebo dim red light (1 lux) given six-ten hours after arising. With no bright light, it was expected that Treatment C would have little circadian effect. All treatments were superimposed upon the 16th hour of background illumination of 50 lux. As observed in Figure 2, the center of timing of each light treatment was precisely eight hours after arising.
Volunteers were given standardized instructions designed to minimize potential differences in expectancy for beneficial effects of the treatments. After the volunteers were assigned to the treatments, expectancy for improvement in mood and sleep during the experiment was assessed via 100 mm visual analogue scales.
Urine Collection
As during home recording, urine was collected every two hours during wake and for any nighttime voidings. The collection time was over two periods of approximately 30 hours: from the last voiding during night one until wake-time on day two, and from the last voiding on night four until wake-time at the end of night five (see Figure 1).
Sleep Assessment
For each laboratory sleep period, measures of SOL, TST, WASO, and sleep efficiency were recorded and scored with standard polysomnographic procedures [21] as well as with actigraphy. In addition, subjective measures of TST, WASO, and insomnia were recorded each morning with diaries, as during home recording.
Mood Assessment
On day four, the subjects' depressed moods were assessed with the CES-D [20] four hours after arising. This represented the final CES-D score.
Assays
Urinary concentrations of 6-sulphatoxymelatonin (aMT6s), the primary metabolite of melatonin, were assayed with a highly specific RIA assay developed by Aldous and Arendt (ALPCO, Ltd., Windham, NH, USA) [22]. Sensitivity of the RIA technique was <0.2 ng/mL. Intra- and inter-assay coefficients of variation were 3.3% and 6.7%, respectively.
Data Analysis
Circadian Phase Assessment and Exclusion of Data
An investigator (JAE) used a four-point ranking system to rate the visual "quality" of the aMT6s excretion profiles: "excellent", "good", "poor", or "insufficient data". The ratings were based, for example, on whether the profiles had the expected patterns of transitions between daytime and nighttime levels, or whether higher or irregular baseline levels or abrupt spikes were observed. Artifacts might be attributable to many factors, including incomplete voiding of the bladder and inaccuracies in recording urinary volume or timing. Only data that were rated "excellent" or "good" were used for fitting 24-hour cosines to the aMT6s excretion data for estimation of aMT6s acrophase (24-hour fitted peak). Since estimation of circadian phase shift required assessment of both baseline and final circadian phases, the present analyses included only volunteers with aMT6s excretion profiles rated "excellent" or "good" for both baseline and final assessments. This reduced the number of older volunteers included in the analysis to 60, whereas aMT6s data could be analyzed in all 30 of the young volunteers.
Baseline aMT6s acrophase was estimated from data of the best quality. If both aMT6s profiles were of sufficient quality, a 24-hour cosine was fit across data for both days. If only one of the aMT6s profiles was of sufficient quality, the cosine was fit only for this day (e.g., Figure 3). If neither home profile was of sufficient quality, then baseline phase was defined by the aMT6s acrophase derived from day one in the laboratory (if this profile was of sufficient quality). In profiles of good quality, the home and first laboratory acrophases only differed, on average, by 0.03 hours. Baseline aMT6s acrophase was compared across treatment and age group via 3 × 2 ANOVA. Final aMT6 acrophase was determined from urinary data collected during the final 24–30 hours in the laboratory. The aMT6s parameters reflected the melatonin profile in the presence of light masking, both at home and in the laboratory.
Figure 3 Determining Urinary aMT6s Acrophase. An example of analyzing urinary aMT6s is shown. The yellow area shows that the excretion rate of aMT6s from one voiding to the next was associated with each interval between voidings. The red line shows that a best-fitting cosine curve was estimated. The salmon dotted line indicates the mesor (the mean level of the fitted cosine). The rose arrow shows that the amplitude of the rhythm is the level of the peak of the fitted cosine above the mesor. The lavender arrow shows that the acrophase is the time of the peak of the fitted cosine referenced to the prior midnight.
Treatment Phase-Shifting Effects
According to convention, circadian phase shifts following the light treatments were calculated by subtracting the final aMT6s acrophase from the baseline aMT6s acrophase. Thus, negative and positive shifts indicated phase delays and phase advances, respectively. Phase-response plots were derived by plotting resultant circadian phase shifts (y-axis) against the circadian timing of the light treatments (A, B, or C) relative to the subjects' baseline aMT6s acrophases. The phase reference used for all light treatments was the center of the four-hour treatment, which was also the center of the 16-hour period of background illumination and wakefulness. Within the restricted phase range, the phase response data was sufficiently linear for slopes and elevations of linear regression lines to be compared via ANOVA, using procedures described by Zar [23].
Circadian Abnormality and Phase Correction
Two measures of circadian misalignment were calculated. First, since our previous analysis of this group indicated that sleep was best when the aMT6s acrophase coincided approximately with mid-sleep (i.e., mid-point between lights out and final time of awakening) [1], circadian malsynchronization was defined as the absolute phase angle (h) between an individual's aMT6s acrophase and his/her mid-sleep. Second, circadian phase dispersion was defined as the absolute number of hours between an individual's aMT6s acrophase and the median aMT6s acrophase for his/her age group. These measures were calculated for both baseline and final aMT6s acrophase data. Baseline measures of circadian malsynchronization and phase dispersion were compared via 3 × 2 treatment-by-age group ANOVAs. Two measures of circadian phase correction were defined by the extent to which circadian malsynchronization and phase dispersion were decreased from baseline to final phase assessment. These changes were compared with 3 × 2 × 2 light treatment-by-age group (older vs. young)-by-time (baseline vs. final) ANOVAs.
Treatment Effects on Mood and Sleep
Mean home baseline CES-D data was compared with CES-D responses during the final day in the laboratory. The laboratory actigraphic data was regarded as the most relevant sleep data to compare to baseline, because actigraphic data allowed comparisons of objective home versus laboratory sleep. To assess changes in sleep associated with the treatments, mean home actigraphic data was compared with mean actigraphic data from the final two nights in the laboratory. Likewise, mean home sleep diary data was compared with the mean diary reports of the last two nights in the laboratory. Analysis of polysomonographic data compared data averaged across the first two nights in the laboratory to the last two nights in the laboratory. Changes in mood and sleep following the treatments were assessed via 3 × 2 × 2 treatment-by-age group-by-time ANOVAs.
Association of Phase Correction with Changes in Mood and Sleep
The association of changes in circadian malsynchronization and phase dispersion with changes in sleep and mood following treatment were assessed in two ways. First, Spearman rank-order correlations were calculated. Second, t-tests compared changes in sleep and mood between groups that had phase correction versus groups that had no phase correction (i.e., had no change or increases in malsynchronization and phase dispersion).
Results
Circadian Timing
As measured by Actillume in the week before entering the laboratory, the center of the sleep periods averaged 03:20 at home. In the laboratory, measurments mid-dark averaged 03:11 (a small but significant difference: p < 0.025). As measured by Actillume, the median mesor illumination (24-hour fitted mean) was 478 lux at home and 349 lux, 381 lux, and 30 lux respectively for treatments A, B, and C in the laboratory. However, the acrophases of 24-hour Actillume illumination measured in lux were 13:09 at home and 15:58 in the laboratory, reflecting the tendency of bright daylight exposures at home to occur before mid-wake.
Baseline and Final aMT6s Acrophase Data
The aMT6s data for the older and young volunteers are displayed in Table 1. No significant treatment, age group, or treatment-by-age group interaction effects were found for aMT6s acrophase data. Likewise, no significant treatment effects were observed for the final laboratory aMT6s mesor or the estimated duration of aMT6s secretion (data not shown).
Table 1 aMT6s acrophase and measures of circadian malsynchronization and phase dispersion in older (n = 60, ages 60–79) and young volunteers (n = 30, ages 20–40), mean and SE.
Age Group Baseline aMT6s Acrophase Final aMT6s Acrophase Baseline Circadian Malsynch. Final Circadian Malsynch. Baseline Circadian Dispersion Final Circadian Dispersion
Older 4.01 ± 0.25 4.68 ± 0.28 1.57 ± 0.16 2.19 ± 0.19 1.53 ± 0.15 1.60 ± 0.18
Young 4.14 ± 0.23 5.07 ± 0.28 0.69 ± 0.11 1.09 ± 0.23 1.04 ± 0.12 1.10 ± 0.20
Treatment Phase-Shifting Effects
Phase responses to the treatments are displayed in Figure 4. Significant linear regressions associating the circadian timing of the light treatments with shifts in aMT6s acrophase were found for each treatment. However, there was no significant difference between treatments in the slopes or in the origins of the regression lines. Across all treatments, there was a significant mean delay in aMT6s acrophase from baseline to final assessment (45 min ± 15 min SEM, t = 3.04, p = 0.003); however, there were no significant treatment-by-time or age group-by-time interaction effects.
Figure 4 Phase Response Plots for each Light Treatment. Shown are the shifts in aMT6s acrophase, which varied significantly for each treatment, as a function of the circadian timing of the light treatments, defined as the center of treatment (eight hours after arising) relative to the aMT6s acrophase at baseline.
Circadian Abnormality and Phase Correction
As compared to younger subjects, at baseline the older subjects had more circadian malsynchronization [t(1,88) = 4.57, p < 0.001] and greater circadian phase dispersion [t(1,88) = 2.50, p = 0.014]. However, there were no significant treatment or treatment-by-age group differences between these variables at baseline (before treatment). There was a significant increase in circadian malsynchronization from baseline to final assessment [F(1,88) = 8.5, p = 0.004] (Figure 5), indicating the delays in aMT6s acrophase. However, there was no significant treatment-by-time or age group-by-time interaction in this effect. Circadian phase dispersion showed no significant change over time (Figure 6), and no significant treatment-by-time or age group-by-time interaction.
Figure 5 Circadian malsynchronization at baseline and final assessment. Shown is circadian malsynchronization, defined as the absolute phase angle (mean ± SE hours) between aMT6 acrophase and mid sleep, determined at baseline and following the light treatments. A significant increase in malsynchronization was found.
Figure 6 Phase dispersion at baseline and final assessment. Shown is phase dispersion, defined as the absolute number of hours (mean ± SE) between aMT6s acrophase and the median aMT6s acrophase, determined at baseline and following the treatments.
Treatment Effects on Mood and Sleep
Volunteers reported equal expectancy for improvements in sleep and mood following each treatment. A significant reduction in the CES-D from baseline to final measurement was found [F(1,82) = 13.8, p < 0.001]. There was no treatment-by-time interaction for CES-D. A near-significant age-group-by-time effect was found for CES-D (F(1,82) = 3.7, p = 0.058]. CES-D was reduced from 15.4 ± 1.1 to 11.4 ± 1.0 in the older group and from 9.1 ± 1.5 to 8.0 ± 1.3 in the younger group.
In actigraphic data, significantly less TST, lower sleep efficiency, and greater WASO were found in the older volunteers as compared to the young volunteers. A significant age group-by-time effect for actigraphic TST was mediated by slight increases from baseline to final assessment in the older group (from 334.3 ± 8.8 min to 339.3 ± 7.6 min) but decreases in the young group (from 450.8 ± 9.4 min to 367.6 ± 7.7 min). No significant treatment or treatment-by-time effects for actigraphic sleep were found.
In sleep diary measures, significantly less TST and significantly greater WASO and insomnia (100 mm visual analogue) were found for the older volunteers in comparison to the young volunteers. A significant age group-by-time interaction for insomnia was mediated by decreases in the older group (from 44.7 ± 1.9 mm to 39.5 ± 2.8 mm) and increases in the young group (from 15.3 ± 2.6 mm to 19.8 ± 3.5 mm.) No significant treatment or treatment-by-time effects for these variables was found.
The older group had significantly less polysomnographic TST and more WASO compared with the young group. However, no significant age group contrasts by time, treatment, or treatment-by-time interaction were found for polysomnographic sleep.
Correlations of "Phase Correction" with Changes in Mood and Sleep
Changes in circadian malsynchronization and phase dispersion were not significantly correlated with changes in mood or sleep. Moreover, changes in mood and sleep were not different between individuals who experienced decreases in circadian malsynchronization or decreases in phase dispersion following treatment compared with those who experienced no phase correction.
Discussion
Surprisingly, no significant contrasts between the three light treatments were demonstrated for melatonin or actigraphic parameters, sleep, or mood, nor could changes in circadian misalignment be associated with disturbances of sleep or mood. Perhaps a longer period of randomized treatment would have resulted in greater contrasts, but significant effects of bright light treatment on mood have been demonstrated with treatment of one week or less, and it is not currently clear that treatment beyond one week produces any greater benefit [24]. Probably, 10,000 lux would have been slightly more effective than 3,000 lux. Urine collections of aMT6s are not the most precise method of estimating melatonin timing, but they were considered less burdensome on the depressed participants than alternative methods.
Significant phase-responses were evident after each laboratory treatment. The average delay in aMT6s acrophase in the laboratory was probably due to the dimmer and later illumination timing in the laboratory as compared to the home situation, because illumination peaked before mid-wake at home, mid-wake averaging well after noon. The similarity in phase responses between the two bright light treatments was consistent with previous studies, suggesting that phase-shifting effects may be estimated from the timing center of light treatments [25,26]. However, it might have been predicted that treatment A would fall on a more active portion of the phase-response curve than treatment B. Also, the remarkable similarity in the phase responses for the dim placebo treatment in comparison to the bright light treatments was unexpected. Dose-response studies have indicated that phase-shifting effects of light are related to cube-root [27] or logistic functions [28] of illumination, either of which would predict that the bright light treatments (3,000 lux) were at least two-fold stronger than the placebo treatment (50 lux). Nonetheless, it appears that the dim placebo and bright light treatments had similar phase-shifting potency in the present protocol.
Our phase shift responses could be explained by nonphotic zeitgebers, including the imposed sleep-wake cycle, social cues, and activity/rest. There is evidence that the sleep-wake cycle is a potent zeitgeber; this effect may be independent of the light-dark cycle [13,29,30]. Combining a fixed sleep-wake routine, with appropriately timed bright light, produces additive phase-shifting effects [13,29]. Both classic and recent research has shown that social interaction can be a significant zeitgeber [17,29,31,32].
Particular procedures used in the present study might have facilitated non-photic entrainment. For example, during the baseline week, volunteers in the present study were asked to maintain their usual sleep-wake and daily routines. These routines were often quite erratic, which might have rendered the circadian system more sensitive to the fixed high amplitude rest/activity schedule in the laboratory. In other studies, subjects have been required to maintain more rigid baseline sleep-wake schedules prior to experimental treatment [27,28].
The degree of social interaction between the volunteers and staff in the present study was greater than that which was permitted in many other studies. Laboratory social interaction was designed both to provide comfort and to help monitor the volunteers for safety. Social interaction may have significant independent zeitgeber effects, and can act synergistically with light exposure. Clinically, it is relevant to examine light effects in the presence of social interaction.
The equivalent reduction in depression following each treatment did not support the prediction of greater antidepressant effects with bright light. Possibly the influence of light could be attenuated by the kind care the volunteers received in the laboratory, the social interactions, and placebo effects. Moreover, the light treatment was for fewer days than that employed in the majority of clinical trials of bright light treatment [24], so an insufficient duration might explain the lack of significant effect. An insufficient duration or intensity of light treatment might also explain the failure to observe photoperiodic effects on the duration of aMT6s excretion.
Another unexpected finding was the significant increase in circadian malsynchronization following the bright light treatments. Phase dispersion also showed a non-significant increase. The phase-response plots indicated that the treatments resulted in "over-corrections" of circadian phase. Volunteers with the most advanced body clocks in reference to sleep at baseline (whose light treatment was therefore centered more than 12 hours after the aMT6s acrophase) demonstrated large phase delays as shown in Figure 4. Conversely, those most delayed in reference to sleep at baseline experienced large phase advances. The corrections were often greater than the amounts of initial phase abnormality, contrary to hypothesis. Also, reductions in circadian malsynchronization or phase dispersion (phase correction) were not correlated with improvements in sleep and mood. Chronic mood and sleep problems associated with circadian malsynchronization might be difficult to correct in such a short period of time, although we had expected to find measurable responses.
Conclusion
Consistent with previous studies, compared to young adults, older adults had significantly greater circadian malsynchronization and phase dispersion. Significant and remarkably similar phase-responses were found for each of the three light treatment schedules. The results suggest that low levels of illumination and/or fixed timing of behavior had significant circadian phase-shifting effects. The large phase-shifts resulted in a significant increase in circadian malsynchronization, rather than phase correction. Moreover, phase correction was not significantly associated with improvements in sleep or mood.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
SDY supervised the data collection, subject recruitment, data analysis, and drafting the manuscript.
DFK conceived of the study and was a principal investigator, screened the subjects, and assisted in data analysis and drafting of the manuscript.
JAE assisted in designing the study and drafting the manuscript and performed the aMT6s assays.
KMR assisted in designing the study, in laboratory data collection, and in drafting the manuscript.
Acknowledgements
This study was supported by AG12364, and HL71560. Raul S. Sepulveda, MD, Patricia Fahme, Yvonne C. Alcala, Julian Smith, MD, and Anthony C. Cress assisted with this study. The study was performed in Dr. Kripke's laboratory in the Department of Psychiatry and Sam and Rose Stein Institute for Research on Aging at UCSD.
==== Refs
Youngstedt SD Kripke DF Elliott JA Klauber MR Circadian abnormalities in older adults J Pineal Res 2001 31 264 272 11589762 10.1034/j.1600-079X.2001.310311.x
Lushington K Lack L Kennaway DJ Rogers N van den Heuvel C Dawson D 6-sulfatoxymelatonin excretion and self-reported sleep in good sleeping controls and 55–80-year-old insomniacs J Sleep Res 1998 7 75 83 9682178 10.1046/j.1365-2869.1998.00095.x
Weitzman ED Moline ML Czeisler CA Zimmerman JC Chronobiology of aging: temperature, sleep-wake rhythms and entrainment Neurobiol Aging 1982 3 299 309 7170047 10.1016/0197-4580(82)90018-5
Kripke DF Youngstedt SD Elliott JA Tuunainen A Rex KM Hauger RL Circadian phase in adults of contrasting ages Chronobiol Int
Youngstedt SD Kripke DF Elliott JA Baehr EK Sepulveda RS Holick MF, Jung EG Light exposure, sleep quality, and depression in older adults Biologic Effects of Light 1998 1999 Boston: Kluwer Academic Publishers 427 435
Jean-Louis G Kripke DF Ancoli-Israel S Klauber MR Sepulveda RS Sleep duration, illumination, and activity patterns in a population sample: effects of gender and ethnicity Biol Psychiat 2000 47 921 927 10807965 10.1016/S0006-3223(99)00169-9
Monk TH Reynolds CFI Kupfer DJ Hoch CC Carrier J Houck PR Differences over the life span in daily life-style regularity Chronobiol Intl 1997 14 295 306
Gaddy JR Rollag MD Brainard GC Pupil size regulation of threshold of light-induced melatonin suppression J Clin Endocrinol Metab 1993 77 1398 1401 8077340 10.1210/jc.77.5.1398
Brainard GC Gaddy JR Ruberg FM Barker FM Hanifin JP Rollag MD Moller M, Pevet P Ocular mechanisms that regulate the human pineal gland 1994 London: John Libby & Company Ltd 415 432
Swaab DF Fliers E Partiman TS The suprachiasmatic nucleus of the human brain in relation to sex, age and senile dementia Brain Res 1985 342 37 44 4041816 10.1016/0006-8993(85)91350-2
Zhang Y Kornhauser JM Zee PC Mayo KE Takahashi JS Turek FW Effects of aging on light-induced phase-shifting of circadian behavioral rhythms, Fos expression and CREB phosphorylation in the hamster suprachiasmatic nucleus Neurosci 1996 70 951 961 10.1016/0306-4522(95)00408-4
Klerman EB Duffy JF Dijk D-J Czeisler CA Circadian phase resetting in older people by ocular bright light exposure J Invest Med 2001 49 30 40
Eastman CI Martin SK How to use light and dark to produce circadian adaptation to night shift work Ann Med 1999 31 87 98 10344580
Horowitz TS Cade BE Wolfe JM Czeisler CA Efficacy of bright light and sleep/darkness scheduling in alleviating circadian maladaptation to night work Am J Physiol Endocrinol Metab 2001 281 E384 E391 11440916
Youngstedt SD Kripke DF Elliott JA Huegel GO Rex KM Exercise phase-response curves in young and older adults Program Abstracts Eighth Meeting Soc Res Biol Rhythms 2002 110
Buxton OM Lee CW L'Hermite-Baleriaux M Turek FW Van Cauter E Exercise elicits phase shifts and acute alterations of melatonin that vary with circadian phase Am J Physiol Reg Integ Comp Physiol 2003 284 R714 R724
Honma K Hashimoto S Nakao M Honma S Period and phase adjustments of human circadian rhythms in the real world J Biol Rhythms 2003 18 261 270 12828283 10.1177/0748730403018003008
Wehr TA Photoperiodism in humans and other primates: Evidence and implications J Biol Rhythms 2001 16 348 364 11506380 10.1177/074873001129002060
Jean-Louis G Kripke DF Cole RJ Assmus JD Langer RD Sleep detection with an accelerometer actigraph: comparisons with polysomnography Physiol Behav 2001 72 21 28 11239977 10.1016/S0031-9384(00)00355-3
Radloff LS The CES-D scale: A self-report depression scale for research in the general population Applied Psychol Measur 1977 1 385 401
Rechtschaffen A Kales A A Manual of Standardized Terminology, Techniques and Scoring System for Sleep Stages of Human Subjects 1968 Los Angeles: Brain Information Service, Brain Research Institute, UCLA
Aldous M Arendt J Radioimmunoassay of 6-sulphatoxy melatonin using an iodinated tracer Ann Clin Biochem 1988 25 298 303 3400987
Zar JH Biostatistical Analysis 1984 Englewood Cliffs, N.J.: Prentice-Hall
Tuunainen A Kripke DF Endo T Light Therapy for Non-Seasonal Depression (Cochrane Review) The Cochrane Library 2004
Czeisler CA Kronauer RE Allan JS Duffy JF Jewett ME Brown EN Bright light induction of strong (type 0) resetting of the human circadian pacemaker Science 1989 244 1328 1333 2734611
Daan S Colin Pittendrigh, Jurgen Aschoff, and the natural entrainment of circadian systems J Biol Rhythms 2000 15 195 207 10885874 10.1177/074873000129001396
Boivin DB Duffy JF Kronauer RE Czeisler CA Dose-response relationships for resetting of human circadian clock by light Nature 1996 379 540 542 8596632 10.1038/379540a0
Zeitzer JM Dijk D-J Kronauer RE Brown EN Czeisler CA Sensitivity of the human circadian pacemaker to nocturnal light: Melatonin phase resetting and suppression J Physiol (London) 2000 526 695 702 10922269
Wever RA The Circadian System of Man: Results of Experiments Under Temporal Isolation 1979 New York: Springer-Verlag
Eastman CI Miescke KJ Entrainment of circadian rhythms with 26-h bright light and sleep-wake schedules Am J Physiol Regulatory Integrative Comp Physiol 1990 259 R1189 R1197
Aschoff J Fatranska M Giedke H Doerr P Stamm D Wisser H Human circadian rhythms in continuous darkness: entrainment by social cues Science 1971 171 213 215 5538832
Honma K Honma S Nakamura K Sasaki M Endo T Takahashi T Differential effects of bright light and social cues on reentrainment of human circadian rhythms Am J Physiol 1995 268 R528 35 7864251
|
16153301
|
PMC1224862
|
CC BY
|
2021-01-04 16:39:13
|
no
|
J Circadian Rhythms. 2005 Sep 9; 3:11
|
utf-8
|
J Circadian Rhythms
| 2,005 |
10.1186/1740-3391-3-11
|
oa_comm
|
==== Front
J Neuroengineering RehabilJournal of NeuroEngineering and Rehabilitation1743-0003BioMed Central London 1743-0003-2-241607639410.1186/1743-0003-2-24ResearchFractional Langevin model of gait variability West Bruce J [email protected] Miroslaw [email protected] Mathematical and Informational Sciences Directorate US Army Research Office, P.O. Box 12211 Research Triangle Park, NC 27709, USA2 Physics Department Wroclaw University of Technology Wybrzeze Wyspianskiego 27, 50-370 Wroclaw, Poland2005 2 8 2005 2 24 24 12 4 2005 2 8 2005 Copyright © 2005 West and Latka; licensee BioMed Central Ltd.2005West and Latka; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
The stride interval in healthy human gait fluctuates from step to step in a random manner and scaling of the interstride interval time series motivated previous investigators to conclude that this time series is fractal. Early studies suggested that gait is a monofractal process, but more recent work indicates the time series is weakly multifractal. Herein we present additional evidence for the weakly multifractal nature of gait. We use the stride interval time series obtained from ten healthy adults walking at a normal relaxed pace for approximately fifteen minutes each as our data set. A fractional Langevin equation is constructed to model the underlying motor control system in which the order of the fractional derivative is itself a stochastic quantity. Using this model we find the fractal dimension for each of the ten data sets to be in agreement with earlier analyses. However, with the present model we are able to draw additional conclusions regarding the nature of the control system guiding walking. The analysis presented herein suggests that the observed scaling in interstride interval data may not be due to long-term memory alone, but may, in fact, be due partly to the statistics.
==== Body
Background
One strategy for understanding legged locomotion of animals is through the use of a Central Pattern Generator (CPG), an intraspinal network of neurons capable of producing a syncopated output [1]. The implicit assumption in such an interpretation is that a given limb moves in direct proportion to the voltage generated in a specific part of the CPG. As Collins and Richmond [1] point out, in spite of the studies establishing the existence of a CPG in the central nervous system of quadrupeds, such direct evidence does not exist for a vertebrate CPG for legged locomotion. Consequently, these and other authors have turned to the construction of models, based on the coupling of linear and nonlinear oscillators, to establish that the mathematical models are sufficiently robust to mimic the locomotion characteristics observed in the movements of segmented bipeds [2], as well as in quadrupeds [3]. These characteristics, such as the switching among multiple gait patterns, is shown to not depend on the detailed dynamics of the constituent nonlinear oscillators, nor on their inter-oscillator coupling strengths [1]. A nonlinear stochastic model of the dynamics of the human gait motor control system called the super CPG (SCPG) has been developed [4]. In the SCPG the stride interval time series is shown to be slightly multifractal, with a fractal dimension that is sensitive to physiologic stress. Herein we do not focus on the generation of each step during walking, but rather we examine the variation in successive steps and its underlying structure.
It has been known for over a century that there is a variation in the stride interval of humans during walking of approximately 3–4%. This random variability is so small that the biomechanical community has historically considered these fluctuations to be an uncorrelated random process, such as might be generated by a simple random walk. In practice this means that the fluctuations in gait were thought not to contain any useful information about the underlying motor control process. On the other hand, Hausdorff et al. [5,6] demonstrated that stride-interval time series exhibit long-time correlations, and suggested that the phenomenon of walking is a self-similar fractal activity. Subsequent studies by West and Griffin [7-9] support these conclusions using a completely different experimental protocol for generating the stride-interval time series data and very different methods of analysis. It was found that things are not quite that simple, however, and instead of the process having no characteristic time scale, as would be the case for a monofractal, there is a preference for a multiplicative time scale in the physiological control system [7].
Physiological time series invariably contain fluctuations so that when sampled N times the data set {Xj}, j = 1,..., N, appear to be a sequence of random points. Examples of such data are the interbeat intervals of the human heart [10,11], interstride intervals of human gait [5,9], brain wave data from EEGs [12] and interbreath intervals [13], to name a few. The analysis of the time series in each of these cases has made use of random walk concepts in both the analysis of the data and in the interpretation of the results. For example, the mean-square value of the dynamical variable in each of these cases (and many more) have the form 〈X(t)2〉∝ tδ, where δ ≠ 1 corresponds to "anomalous diffusion". A value of δ < 1 is often interpreted as an antipersistent process in which a step in one direction is preferentially followed by a step reversal. A value of δ > 1 is often interpreted as a persistent process in which a step in one direction is preferentially followed by another step in the same direction. A value of δ = 1 is, again, often interpreted as ordinary diffusion in which the steps are independent of one another. The initial analysis of each of these time series, using random walk concepts, suggested that they could be interpreted as monofractals. However, on further investigation the heart beat variability has been found to be multifractal [14], as were the interstride intervals [4].
A modeling approach complementary to random walks is the Langevin equation, a stochastic equation of motion for the dynamical variables in a physical system. This latter model has undergone a transformation similar to that of random walks since its introduction into physics by Langevin in 1908. The solution to the Langevin equation is a fluctuating trajectory for the particle of interest and an ensemble of such trajectories determines the statistical distribution function. In this way the Gaussian probability density for Brownian motion is obtained. The density can also be obtained by aggregating the steps to form a discrete trajectory using a random walk model [15,16]. These two kinds of models of the physical world, random walks and the Langevin equation, have long been thought to be equivalent. In fact, that equivalence has been used as the dynamical foundation of statistical mechanics and thermodynamics. This equivalence has also been used to interpret the monofractal statistical properties of physiological time series.
While the properties of monofractals are determined by the global scaling exponent, there exists a more general class of heterogenous signals known as multifractals which are made up of many interwoven subsets with different local scaling exponents h. The statistical properties of these subsets may be characterized by the distribution of fractal dimensions f(h). In order to describe the scaling properties of multifractal signals it is necessary to use many local Hölder exponents. Formally, the Hölder exponent h(t0) of a trajectory X(t) at t = t0 is defined as the largest exponent such that there exists a polynomial Pn(t) of order n that satisfies the following condition [17]:
for t in a neighborhood of t0 and the symbol O(ε) means a term no greater than ε. Thus the Hölder exponent measures the singularity of a trajectory at a given point. For example, h(t0) = 1.5 implies that the trajectory X is differentiable at t0 but its derivative is not. The singularity lies in the second derivative of X(t). The singularity spectrum f(h) of the signal may be defined as the function that for a fixed value of h yields the Hausdorff dimension of the set of points t. The singularity spectrum is used to determine whether or not the stride interval time series is multifractal.
A new kind of random walk has recently been developed, one having multifractal properties [18-21]. Herein we are guided by this earlier work, but use it to generalize the Langevin equation to describe a multifractal dynamical phenomenon. In Methods we review the multifractal formalism and apply the processing algorithm to the interstride interval time series. The mass exponent τ(q) is determined to be a nonlinear function of the moment q, and the singularity spectrum f(h) is found to be a convex function of local scaling exponent h. We also introduce a fractional Langevin equation and make the index of a fractional integral a random variable to show how this model can describe a multifractal process. The multifractal spectrum is shown to be a property of the solution to this fractional Langevin equation. In Results and Disscussion we apply the analytic expression for the singularity spectrum to the interstride interval data discussed in the Methods section. The agreement between the predictions of the fractional Langevin equation and experiment for human gait is remarkable. In Conclusions we explore some of the physiological implications of the fractional Langevin model including the suggestion that the observed scaling of the time series may not only be due to long-term memory but to the underlying statistics as well.
Methods
The distribution of Hölder exponents for a time series can be determined in a number of different ways. Herein we use the partition function. Let us cover the time axis with cells of size δ such that the time is given by t = Nδ and N > > 1. Following Falconer [17] we can define the partition function in terms of the moments, q, of a measure μ
where Bj is the jth box in the δ-coordinate mesh that intersect with the measure μ. We can construct the measure using the time series obtained from the interstride interval data. This measure is made by aggregating the observed interstride time intervals, tj, j = 1,2.., N,
such that T(n,δ) is interpreted as the random walk trajectory for a given data set. We use the random walk trajectory to construct the phenomenological measure in the partition function (2) as
where the integer n is the discrete time lag. For a monofractal random walk process the measure (4) is essentially uniform. For a multifractal, on the other hand, the theoretical scaling behavior of the partition function Sq(δ) in the limit of vanishing grid scale [17,24] is
Sq(δ) ≈ δ-τ(q) (5)
where τ(q) defines the mass exponent. We emphasize that (4) is a phenomenological measure with an undetermined lag time. The lag time is chosen in the present calculation to maximize the sensitivity of the partition function to the positive moments.
The mass exponent is related to the generalized dimension D(q) by the relation
τ(q) = (1 - q)D(q) (6)
where D(0) is the fractal or box-counting dimension, D(1) is the information dimension and D(2) is the correlation dimension [24]. The moment q therefore accentuates different aspects of the underlying dynamical process. For q > 0, the partition function Sq(δ) emphasizes large fluctuations and strong singularities through the generalized dimensions, whereas for q <0, the partition function stresses the small fluctuations and the weak singularities. This property of the partition function deserves a cautionary note because the negative moments can easily become unstable, introducing artifacts into the calculation. For this reason the interpretation of the trajectory approach must be judged with some caution for q < 0.
A monofractal time series can be characterized by a single fractal dimension. In general, time series have a local fractal exponent h that varies over the course of the trajectory. The function f(h), called the multifractal or singularity spectrum, describes how the local fractal exponents contribute to such time series. Here h and f are independent variables, as are q and τ. The general formalism of Legendre transform pairs interrelates these two sets of variables by the relation, using the sign convention in Feder [24],
f(h) = qh + τ(q). (7)
The local Hölder exponent h varies with the q-dependent mass exponent through the equality
so the singularity spectrum can be written as
f(h(q)) = - qτ'(q) + τ(q) (9)
where τ(q) is determined by data, that is, by the trajectory, as is its derivative τ'(q).
The multifractal behavior of time series can be modeled using a number of different formalisms. For example, a random walk [19,23], in which a multiplicative coefficient in the random walk is itself made random, becomes a multifractal process. This approach was developed long before the identification of fractals and multifractals and may be found in Feller's book [25] under the heading of subordination processes. The multifractal random walks have been used to model various physiological phenomena. Another method, one that involves an integral kernel with a random parameter, was used to model turbulent fluid flow [26]. Here we adopt a version of the integral kernel, but one adapted to time rather than space series. In order to accomplish this we review some of the history of the Langevin equation.
Fractional Langevin equation
A theoretical Langevin equation is generally constructed from a Hamiltonian model for a simple dynamical system coupled to the environment [27]. The equations of motion for the coupled system are manipulated so as to eliminate the degrees of freedom of the environment from the dynamical description of the system. Only the initial state of the environment (heat bath) remains in the Langevin description, where the random nature of the driving force is inserted through the choice of distribution of the initial states of the bath. The simplest Langevin equation for a dynamical system open to the environment has the form
where ξ(t) is a random process, λ is a dissipation parameter and there exists a fluctuation-dissipation relation [27] connecting the two. Of course, we cannot completely interpret (10) until we specify the statistical properties of the ξ-fluctuations and for this we need to know the environment of the system. The random driver is typically assumed to be a Wiener process, that is, to have Gaussian statistics and no memory.
When the system dynamics depends on what occurred earlier, that is, the environment has a memory, (10) is no longer adequate and the Langevin equation must be modified. The generalized Langevin equation takes this memory into account through an integral term of the form
where the memory kernel, K(t), replaces the dissipation parameter and there is a generalized fluctuation-dissipation relation [27]. Both these Langevin equations are monofractal if the fluctuations are monofractal, which is to say the time series given by the trajectory X(t) is a fractal random process, if ξ(t) is a fractal random process.
Now we come to the most recent generalization of the Langevin equation, one that incorporates memory into the system's dynamics through the use of fractional calculus. The simplest fractional Langevin equation has the form [28]
where is a Riemann-Liouville (RL) fractional derivative with 0 < β ≤ 1
and is related to the RL-fractional integral
Note that we have not included dissipation in this simple model, but the initial condition X0 = X(0) is incorporated into the dynamical equation in order to have a well-defined initial value problem. The formal solution to the fractional Langevin equation (12) is [28]
where the kernel in (15) is given by the weighting factor within the RL-fractional integral. As mentioned earlier, the form of this relation for multiplicative stochastic processes and its association with multifractals had been noted in the phenomenon of turbulent fluid flow [26], through a space, rather than time, integration kernel.
Multifractal time series
The random forcing term on the right-hand side of (15) is selected to be a zero-centered, Gaussian random variable and therefore to scale as [29]
ξ(λt) = λhξ(t) (16)
where the Hölder exponent is in the range 0 <h = 1. In a similar way the kernel in (15) is easily shown to scale as
Kβ(λt) = λβ-1Kβ(t) (17)
so that the solution to the fractional Langevin equation scales as
ΔX(λt) = λh+βΔX(t) (18)
where ΔX(t) = X(t) - X0. In order to make the solution to the fractional Langevin equation a multifractal we assume that the parameter β is random. To construct the traditional measures of multifractal stochastic processes we calculate the qth moment of the solution by averaging over both the random force ξ and the random parameter β to obtain, in an obvious notation,
Note that when ζ(q), the structure function exponent, is linear in q the underlying process is monofractal, whereas, when ζ (q) is nonlinear in q the proces is multifractal. This is the case because
ζ(q) = 1 - τ(q) (20)
relating the structure function exponent to the mass exponent [30].
To determine the structure function exponent we make an assumption about the statistics of the parameter β. We can always write the β-average as
where Z(s) is a random variable as well as a function of s. Note that in the present case the functionality is just one of linear proportionality. In this way the expression on the right-hand side of (21) is the Laplace transform of the probability density. We assume the random variable Z(s) is an α-stable Lévy process in which case the statistics of the multiplicative fluctuations are given by the distribution [15]
with 0 < α = 2. Inserting (22) into (21) to replace the averaging bracket and integrating over z yield the delta function δ(k+iq) which, integrating over k, results in
so that re-introducing s = lnλ into this equation we obtain
Consequently, from (20) we obtain for the moment correlation function
ζ(q) = qh - b|q|α (23)
Therefore the solution to the fractal Langevin equation corresponds to a monofractal process only in the case α = 1 and q > 0, otherwise the process is multifractal. We restrict the remaining discussion to q > 0.
Thus, we observe that when the memory kernel in the fractional Langevin equation is random, the solution consists of the product of two random quantities giving rise to a multifractal statistical process. This is analogous to Feller's subordination process. We observe that, for the statistics of the multiplicative exponent given by Lévy statistics, the singularity spectrum as a function of the positive moments, is
f(q) = 1 - (α - 1)bqα (25)
which is determined by substituting (24) into (9), through the relationships between exponents (20).
Results and Discussion
The data obtained, from individuals walking at a normal steady pace, consists of the time interval for each stride and the number of strides in a sequence of steps. The maximal extension of the right leg, the "stride interval" versus the stride number, plotted on a graph, has all the characteristics of a time series, cf. Figure 1. There were ten participants in the study (four males and six females), all in good health, with no acute injuries, ranging in age from 20 to 46 years old with a median age of 29 years. Normal steady walking was monitored for the ten participants, and an electrogoniometer was used to collect kinematic data on the angular extension of the right leg. The signal from the electrogoniometer was recorded at 100 Hz by a computer contained in a "fanny pank" attached to the walker. These data were downloaded to a PC after twelve to fifteen minutes and the interval between successive maximal extensions of the right leg in the analog signal was digitized and used as the time series data [7].
Figure 1 Typical interstride interval time series: The interstride interval time series for a person undergoing relaxed walking is depicted for 800 steps. This is taken from a 15 minute time series [7].
The signal shown in Figure 1 indicates a variation in the stride interval with a standard deviation of 0.12 seconds, and the resolution of the measurement is of the order 0.01 seconds. What can we learn from a time series that has such a potentially substantial error? Suppose our time series consists of the superposition of two independent processes. One process is determined by the dynamics of the system and the other by measurement error, so that the second moment of the time series after n intervals is given by
<X(t)2> = An + Bnδ (26)
The first process is, of course, that due to measurement error, modeled as a simple random walk, with strength A. For δ > 1 the second process is a persistent random walk and dominates for n > 1. In such a case we would expect for n sufficiently large, where the relative size of A and B determines what is meant by sufficiently large, to find the scaling
<X(λt)2> ≈ λδ <X(t)2>. (27)
This scaling was, in fact, observed for the data depicted in Figure 1, as well as for the other gait time series obtained in this study [7-9]. From the results of these earlier analyses we conclude that the level of statistical variation in the data, due to measurement error, will not change the conclusions drawn from subsequent analysis.
As mentioned above, a time series is monofractal when the mass exponent τ(q) is linear in q, otherwise the underlying process is multifractal. We apply the partition function measure and numerically evaluate
and the results are depicted in Figure 2a. Rigorously, the expression for the mass exponent requires δ → 0, but we cannot do that with the data, so there is some error in our results. The significance of that error is to be determined. In Figure 2a we only show the mass exponent for a typical walker from the ten subjects, since they individually do not look too different from the curve shown. It is clear from the figure that the mass exponent is not linear in the moment index q. In Table 1 we record the fitting coefficients for each of the ten time series using the fitting equation for the mass exponent suggested by the solution to the fractional Langevin equation,
Figure 2 Empirical mass exponent and singularity spectrum: (a) The mass exponent is determined using the partition function from (28) and given by the dots for a typical data set. The solid curve is the quadradic least-squares fit of (29) to the calculated points. (b) The singularity spectrum is determined from the mass exponent using (9).
Table 1 The fitting parameters for the mass exponent τ(q) are listed. The column-a1 is the fractal dimension for the time series. In each case the fractal dimension agrees with that obtained earlier using a different method [7]. The last two columns denote the Lévy index and the statistical significance of the comparison of the empirical and theoretical values is p = 0.01
Walker -a1 a2 Empirical Lévy index Theoretical Lévy index
1 1.26 0.13 1.57 1.52
2 1.41 0.19 1.57 1.82
3 1.32 0.09 1.83 1.64
4 1.26 0.24 1.54 1.52
5 1.12 0.28 1.47 1.24
6 1.07 0.07 1.84 1.14
7 1.17 0.07 1.69 1.34
8 1.29 0.27 1.39 1.58
9 1.14 0.12 1.63 1.28
10 1.17 0.12 1.64 1.34
Averages 1.21 ± 0.10 0.15 ± 0.07 1.61 ± 0.15 1.44 ± 0.21
τ(q) = 1 + a1q + a2|q|α. (29)
The fit to the data using (29) is indicated by the solid curve in Figure 2a.
The singularity spectrum can now be determined using the Legendre transformation by at least two different methods. One procedure is to use the fitting equation substituted into (9). We do not do this here, but we note in passing that if (29) is inserted into (8), the fractal dimension is determined by the q = 0 moment to be
The values of the parameter a1 listed in Table 1 agree with the fractal dimensions obtained earlier using a scaling argument for the same data [7].
A second method for determining the singularity spectrum, the one we use here, is to numerically determine both τ(q) and its derivative. In this way we calculate the multifractal spectrum directly from the data using (9). It is clear from Figure 2b that we obtain the canonical form of the spectrum, that is, f(h) is a convex function of the scaling parameter h. The peak of the spectrum is determined to be the fractal dimension, as it should. Here again we have an indication that the interstride interval time series describes a multifractal process, but we stress that we are only using the qualitative properties of the spectrum for q > 0, due to the sensitivity of the numerical method to weak singularities. This sensitivity is apparent from the asymmetry of the empirical singularity spectrum in Figure 2b. These results are in agreement with the weak multifractality found by Scafetta et al. [31] using a different interstride interval data set.
It is clear from Figure 3 that the singularity spectrum calculated from the data for positive q are well fit by the solution to the fractional Langevin equation with the parameter values α = 1.57 and a2 = 0.13, obtained through a mean-square fit of (25) to the data points. Note that this fit to the scaling exponent is denoted as the empirical Lévy index in Table 1. Adjacent to this column is the theoretical Lévy index obtained from the relation
Figure 3 Singularity spectum in terms of moments: The singularity spectrum is calculated as a function of the moment-order and denoted by the dots using (9) for a typical data set. The solid curve is the least-squares fit of (29) to the calculated points.
called the Lévy-walk diffusion relation [32] and which relates the scaling exponents when the underlying statistical process is an α-stable Lévy statistical process. Note that the Lévy probability density p(x, t) satisfies the scaling relation [32]
A comparison of the two columns for the Lévy index in Table 1, empirical and theoretical, using a statistical t-test, indicates statistical significance at the p = 0.01 level.
Conclusion
The nonlinear form of the mass exponent τ(q) in Figure 2a, the convex form of the singularity spectrum f(h) in Figure 2b and the fit to f(q) in Figure 3, are all indications that interstride interval time series are multifractal. This analysis is further supported by the fact that the maxima of the singularity spectra coincide with the fractal dimensions determined previously using the scaling properties of the time series without the construction of a random walk trajectory [7]. A complete discussion of the limitations associated with determining the multifractal nature of interstride intervals using the singularity spectrum with limited data is given by Scafetta et al [31]. Furthermore, the empirical values of the Lévy index in Table 1 are consistent with those predicted using Lévy-walk diffusion relation [32] at the 0.01 level of significance.
It has been suggested that the CPG for gait consists of a random walk among a number of neural centers, thereby giving rise to its fractal behavior [5,6]. This model gives rise to a process having Gauss statistics and a long-time memory determined by the scaling index. The present results, however, point in a different direction. Recall that anamolous diffusion (δ ≠ 1) can arise in two distinct ways. The more familiar is that of a random walk with memory, in which the statistics are Gaussian, but the frequency spectrum is given by P(ω) ∝ 1/ωδ-1. The second way anomalous diffusion can arise is through the scaling of the probability density as given by (32). If the statistics are Gaussian then the scaling indices are related by δ = 2 μ and for ordinary diffusion μ = 1/2 impling δ = 1; in addition, for μ ≠ 1/2 the process is that of fractional Brownian motion. However, when the statistics are Lévy stable the second moment diverges and special methods must be employed to obtain second-moment scaling.
Shlesinger et al.[33] showed that when the steps in a random walk can be arbitrarily long and the length of time required to take a step is accounted for in the walking process, one obtains a Lévy diffusion process with a finite second moment. The second moment in such a Lévy-walk has a scaling index given by (31) with δ = 1/α. Consequently, the quality of the fit of the Lévy index obtained using the Lévy-walk diffusion relation to that obtained from the singularity spectrum, given by the solution to the fractional Langevin equation, suggests that the scaling in the interstride interval data may not be due solely to long-term memory, as previous investigators have concluded. Instead the observed scaling in interstride interval time series might be due to both long-time memory and statistics.
We use the fractional Langevin equation to describe the motor control process rather than the random walks of previous authors because of the direct correspondence between the microscopic dynamics and the macroscopic fractional derivatives established by Grigolini et al. [34]. The latter authors demonstrate that the existence of a clear separation between microscopic and macroscopic time scales supports the use of random walks and traditional statistical mechanics to model the phenomena of interest. This separation of time scales would be consistent with the traditional random walk way of modeling memory in CPG. However, when the microscopic time scales diverge, such that they overlap with the macroscopic time scale, ordinary statistical mechanics breaks down and the non-differentiabiltiy of the microscopic dynamics is transmitted from the microscopic to the macroscopic level in the form of fractional derivatives. In the present context a manifestation of an inverse power-law distribution of neuron firing would be a fractional differential equation of motion for motor response.
Stated somewhat differently, Grigolini et al [34] showed that the fractional derivative in the fractional Langevin equation can be interpreted in terms of an inverse power-law waiting time distribution function using a Continuous Time Random Walk Model. Thus, not only is the frequency accessed by the control system selected randomly, but the length of time it spends at that particular frequency in SCPG is also random. This waiting time distribution function is inverse power law and directly proportional to the fractional integral kernel. The fractional Langevin equation implies this full dynamical picture and appears to be consistent with the human gait data.
We are cognizant of the fact that to establish that the scaling observed in interstride interval data is due to statistics and memory, rather than long-time memory alone, requires more than the limited analysis presented here. So we put this speculation in the form of a hypothesis which we are presently testing using extensive interstride interval data available from Physionet. The results of these tests will be presented elsewhere.
Supplementary Material
Additional File 1
List of symbols used.
Click here for file
Acknowledgements
The authors thank the U.S. Army Research Office for partial support of this research and Dr. L. Griffin for providing the data used in this analysis and for useful discussions.
==== Refs
Collins JJ Richmond SA Hard-wired central pattern generators for quadrupedal locomotion Biological-Cybernetics 1994 71 375 385 10.1007/s004220050098
Cohen AH Rossignol S Grillner S Neural control of rhythmic movements in vertebrates 1988 New York: Wiley
Collins JJ Stewart IN Coupled nonlinear oscillators and the the symmetries of animal gait J Nonlinear Sci 1993 3 349 392
West BJ Scafetta N Nonlinear dynamical model of human gait Phys Rev E 2003 67 051917 1 10.1103/PhysRevE.67.051917
Hausdorff JM Peng CK Ladin Z Wei JY Goldberger AL Is walking a random walk? Evidence for long-range correlations in stride interval of human gait J Appl Physiol 1995 78 349 358 7713836
Hausdorff JM Mitchell SL Firtion R Peng CK Cudkowicz ME Wei JY Goldberger AL Altered fractal dynamics of gait: reduced stride-interval correlations with aging and Huntington's disease J Appl Physiol 1997 82 262 269 9029225
Griffin L West DJ West BJ Random stride intervals with memory J Biol Phys 2000 26 185 202 10.1023/A:1010322406831
West BJ Griffin L Allometric control of human gait Fractals 1998 6 101 108
West BJ Griffin L Allometric control, inverse power laws and human gait Chaos, Solitons & Fractals 1999 10 1519 1527 10.1016/S0960-0779(98)00149-0
West BJ Zhang R Sanders AW Miniyar S Zuckerman JH Levine BD Fractal fluctuations in transcranial Doppler signals Physical Review E 1999 59 3492 3498 10.1103/PhysRevE.59.3492
Peng CK Mietus JE Hausdorff JM Havlin S Stanley HE Goldberger AL Long-range anticorrelations and non-Gaussian behavior of the heartdeat Phys Rev Lett 1993 70 1343 1346 10054352 10.1103/PhysRevLett.70.1343
West BJ Physiology, Promiscuity and Prophecy at the Millennium: A Tale of Tails 1999 World Scientific, Singapore
Szeto HH Cheng PY Decena JA Cheng Y Wu DL Dwyer G Fractal properties in fetal breathing dynamics Am J Physiol 1992 263 R141 R147 1636780
Ivanov PC Amaral LA Goldberger AL Havlin S Rosenblum MG Stanley HE Struzik ZR Multifractality in human heartbeat dynamics Nature 1999 399 461 465 10365957 10.1038/20924
Montroll EW West BJ Montroll EW, Lebowitz JL An Enriched Collection of Stochastic Processes Fluctuation phenomena, Pbk ed, updated edn 1987 Amsterdam: North-Holland 61 206
Weiss GH Aspects and applications of the random walk 1994 Amsterdam: North-Holland
Falconer KJ Fractal geometry mathematical foundations and applications 1990 Chichester: Wiley
Muzy JF Bacry E Arneodo A Wavelets and multifractal formalism for singular signals: Application to turbulence data Physical Review Letters 1991 67 3515 3518 10044755 10.1103/PhysRevLett.67.3515
Muzy JF Bacry E Arneodo A Multifractal formalism for fractal signals: The structure-function approach versus the wavelet-transform modulus-maxima method Phyical Review E 1993 47 875 884 10.1103/PhysRevE.47.875
Castaign B Gagne Y Hopfinger E Velocity probability density-functions of high Reynolds-number turbulence Physica D 1990 46 177 200
Mallat SG A Wavelet Tour of Signal Processing 1999 2 Cambridge: Academic Press
Ashkenazy Y Hausdorff JM Ivanov PCh Goldberger AL Stanley HE A stochastic model of human gait dynamics Physica A 2002 316 662 670
Schmitt FG Seuront L Multifractal random walk in Copepod behavior Physica A 2001 301 375 396
Feder J Fractals 1988 New York: Plenum Press
Feller W An introduction to probability theory and its applications 1967 3 New York: Wiley
Schertzer D Lovejoy S Schmit F Chiguirinskaya Y Marsan D Multifractal cascade dynamics and turbulent intermittency Fractals 2005 5 427 471
Lindenberg K West BJ The nonequilibrium statistical mechanics of open and closed systems 1990 New York: VCH Publishers
West BJ Grigolini P Bologna M Physics of Fractal Operators 2004 New York:Springer
Bassigthwaighte B Liebowitch LS West BJ Fractal Physiology 1994 Oxford: Oxford University press
Rajagopalan B Tarboton DG Understanding complexity in the structure of rainfall Fractals 1993 1 606 616
Scafetta N Griffin L West BJ Holder exponent spectra for human gait Physica A 2003 328 561 583
Scafetta N West BJ Mutiscaling comparative analysis on "earthquake conversations" in California: Phys Rev Lett 2004 92 138501-1 138501-4 15089646 10.1103/PhysRevLett.92.138501
Shlesinger MF West BJ Klafter J Lévy dynamics for enhanced diffusion: an application to turbulence Phys Rev Lett 1987 58 1100 1103 10034339 10.1103/PhysRevLett.58.1100
Grigolini P Rocco A West BJ Fractional calculus as a macroscopic manifestation of randomness Phys Rev E 1999 59 2603 2613 10.1103/PhysRevE.59.2603
|
16076394
|
PMC1224863
|
CC BY
|
2021-01-04 16:37:40
|
no
|
J Neuroengineering Rehabil. 2005 Aug 2; 2:24
|
utf-8
|
J Neuroeng Rehabil
| 2,005 |
10.1186/1743-0003-2-24
|
oa_comm
|
==== Front
J Neuroengineering RehabilJournal of NeuroEngineering and Rehabilitation1743-0003BioMed Central London 1743-0003-2-281613892210.1186/1743-0003-2-28ReviewHow useful is satellite positioning system (GPS) to track gait parameters? A review Terrier Philippe [email protected] Yves [email protected] Department of Physiology, University of Lausanne, Switzerland2005 2 9 2005 2 28 28 18 3 2005 2 9 2005 Copyright © 2005 Terrier and Schutz; licensee BioMed Central Ltd.2005Terrier and Schutz; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Over the last century, numerous techniques have been developed to analyze the movement of humans while walking and running. The combined use of kinematics and kinetics methods, mainly based on high speed video analysis and forceplate, have permitted a comprehensive description of locomotion process in terms of energetics and biomechanics. While the different phases of a single gait cycle are well understood, there is an increasing interest to know how the neuro-motor system controls gait form stride to stride. Indeed, it was observed that neurodegenerative diseases and aging could impact gait stability and gait parameters steadiness. From both clinical and fundamental research perspectives, there is therefore a need to develop techniques to accurately track gait parameters stride-by-stride over a long period with minimal constraints to patients. In this context, high accuracy satellite positioning can provide an alternative tool to monitor outdoor walking. Indeed, the high-end GPS receivers provide centimeter accuracy positioning with 5–20 Hz sampling rate: this allows the stride-by-stride assessment of a number of basic gait parameters – such as walking speed, step length and step frequency – that can be tracked over several thousand consecutive strides in free-living conditions. Furthermore, long-range correlations and fractal-like pattern was observed in those time series. As compared to other classical methods, GPS seems a promising technology in the field of gait variability analysis. However, relative high complexity and expensiveness – combined with a usability which requires further improvement – remain obstacles to the full development of the GPS technology in human applications.
==== Body
Analysis of the pattern in cyclic movements may be of great interest in neurosciences and behavioral sciences, since they rely on complex sensory-motor coordination requiring both automated and voluntary tasks [1]. Recent studies, based on non-linear analysis of time series, have shown the presence of complex temporal fluctuations in several biological repetitive processes, such as heart beats [2-4], respiration [5], or controlled finger movements [6].
Walking is the one of the most common repetitive movement that humans performed in real life. In addition to automatic rhythmic activation by Central Pattern Generators at the spinal level, the locomotor system is regulated by the cerebellum, the motor cortex and the basal ganglia, with feedback from proprioceptive, visual and vestibular sensors. Stride after stride, the final output of the control segment modulates the spatial (Step Length, SL), and temporal (Step Frequency SF or cadence) patterns of the gait in order to provide optimal movement in terms of mechanics and energetics [7-11].
Gait variability can be defined as the variation of gait parameters from stride to stride. It was reported that gait variability could by modified by different pathology (e.g. neuro-degenerative diseases), or to be related to the propensity to fall in elderly [12,13]. In addition, it has been shown that stride-to-stride variability diminished with the maturation of the gait in children [14].
Hausdorff's group has extensively studied long-term gait variability [12-21]. They reported [20] that the stride-to-stride variation of stride duration exhibited long-range, self-similar correlations. In other words, the fluctuation in the stride interval is characterized by an autocorrelation function that decays as a power law: the present value is statistically correlated not only with its most recent value but also with its long-term history in a scale invariant fractal manner [20,21]. They attempted to demonstrate the implication of basal ganglia in the control of the stability and the generation of the fractal pattern. In short, the underlying hypothesis is that fractal pattern is a marker for neural complexity: different factors (disease, aging, imposed stride frequency by metronome, called metronome walking) that affect this complexity lead to the loss of fractal patterns and to the emergence of random patterns [15].
For all these different experiments, Hausdorff et al. used a force-sensitive switch placed in shoes [17]. This sensor detects heel strike and therefore allows to obtain information about temporal pattern of the gait only. They addressed the issue as follows: "Additional information regarding the alterations of gait [...] might be provided [...] by obtaining stride-by-stride measures of stride length and gait speed" [18].
In this context, we propose the use of high-accuracy satellite positioning (Global Positioning System, GPS), as a alternative tool to obtain long time series of basic gait parameters, i.e. Walking Speed (WS), Step Length (SL) and Step Frequency (SF). The purpose of the present review article is to highlight the new GPS technique and compare it to other gait analysis methods. We present a thorough description of theoretical and practical aspects of GPS technology for high accuracy positioning. Next, we describe the underlying biomechanical assumptions necessary to obtain gait parameters from GPS positioning data. Finally, following a discussion of our recently published results about fluctuation analysis of gait parameters [22], we highlight the advantages and shortcomings of GPS techniques as compared to other methods.
Motion analysis: classical methods
Several gait analysis techniques have been developed over the last decades (fig. 1) [23]. A kinematic analysis of gait requires measurement of the displacement of the body segments during the walking cycle. Electrical, photographic, cinefilm and video or other electronic techniques have been used to calculate the position and orientation of each body segment to reconstruct the movements that took place. Measurement can be made in two or three dimensions. In order to understand how walking is accomplished, the forces acting on the human body must be also assessed (kinetics) [8,9,24,25]. By analyzing the moments and forces occurring at the joints to produce the motions of the limbs, an estimation can be made of the forces the muscles must produce. For a complete kinetic analysis of each body segment, kinematic data (displacements, velocity), anthropometric data (body segment parameters), and external force data (gravity, ground reaction force) are required. The ground reaction force is classically measured by a force plateform [25,10]. This device determines the magnitude and direction of the ground reaction force vector by measuring its three components (vertical, mediolateral and anteroposterior shear forces) and vectorally adding them. In parallel, in order to evaluate muscle activity, the depolarization of the muscles membrane by motor neuron activation can be tracked by using Electromyography (EMG).
Figure 1 Simplified scheme of the techniques available for gait analysis. Each method measure different parameters and have different advantages and shortcomings.
While a number of gait analysis systems have been developed over the years to allow an accurate and overall description of walking, most of them are impractical for fast-paced clinical settings. Furthermore, they are not designed to record long times series of gait parameters over numerous consecutive strides. Alternative techniques have been therefore used in order to analyze a reduced set of parameters with an increased practicability. Instrumented walkway [26] permits a rapid survey of several temporal and spatial gait parameters (step length, step width, stance/swing time, step duration, etc.); however, the distance is limited (typically 10 meters), and the subject must follow a straight trajectory.
The shortcoming of limited space in a laboratory environment can be partially overcome by using a treadmill. Video analysis or instrumented treadmill (force plateform [27] or kinematic arm [28-30]) allow investigators to analyze long duration walking or running. In theory, treadmill walking is supposed to be energetically and biomechanically identical to normal walking. However, treadmill walking alters the perception of motion by the participant and therefore may alter the gait parameters as compared to free walking. In addition, because of the narrow path offered by the treadmill, there is no freedom in the selection of the trajectory.
In parallel, other methods – based on portable sensors – have been developed to increase usability of gait analysis under free walking conditions. Accelerometers and gyroscopes have been used to retrieve several temporal and spatial gait parameters [31-37]. These techniques are very promising, however they rely on complex algorithms to convert raw measurements (acceleration, angular motions) into gait parameters (speed, step length, cadence). In addition, these algorithms are mostly calibrated to normal walking under standard conditions: there is no warranty that environmental changes (slope, quality of the terrain) or pathological gait (for instance claudication) are correctly taken into account. As a result, investigators must carefully select their devices and extensively test whether they obtain an output compatible with their experimental conditions. In our opinion, a less indirect methodology would offer more flexibility in the experimental design; by allowing a direct speed and position measurement, GPS is a good candidate for such an approach.
In 1995, Hausdorff and colleagues proposed a new footswitch method to analyze long term variability of the gait [17]. With a small portable sensor in the shoe, it is possible to retrieve stride duration stride by stride over very long periods (1 hour walking, [21].). However, it is not possible to assess spatial parameters (SL) by using this technique.
GPS in human applications: historical perspectives
Almost ten years ago, we proposed to utilize GPS for assessing physical activity in free living conditions, in particular walking and running [38]. Simple relatively cheap commercial instruments used for leisure navigation (e.g. sailing) was tested. Using this type of GPS receiver, it was concluded that the accuracy of speed was insufficient for research purpose and that it could be improved by using differential GPS (DGPS). In a subsequent study, it was shown that DGPS improved the speed accuracy by a factor of about 10 as compared to non-differential GPS (error below 0.1 km/h) [39]. However, the study was performed when the satellite signals was voluntarily degraded by the US Departement of Defense (Selective Availability), so that the improvement with DGPS is expected to be considerably greater than today (since SA was removed in 2000). Witte & Wilson [40] have shown, using non-differential GPS, that reasonable accuracy in straight trajectory could be observed, but the error increased in circular path especially with small radii of curvature where a tendency was observed to underestimate speed [40]. More recently, another group in Scandinavia used DGPS for assessing the performance of orienteering with DGPS, and suggested that it could be combined with complementary techniques (accelerometry, electromyography etc.) in the field of outdoor exercise physiology [41-43].
Standard GPS: principles
The Global Positioning System (GPS) is a satellite-based navigation system made up of a network of 24 satellites placed into orbit by the U.S. GPS works in any weather conditions, anywhere in the world, 24 hours a day. There are no subscription fees or setup charges to use GPS. GPS satellites circle the earth in a very precise orbit and transmit signal information. GPS receivers make use of triangulation to calculate the user's exact location. Essentially, the GPS receiver compares the time a signal was transmitted by a satellite with the time it was received. The time difference tells the GPS receiver how far away the satellite is. With distance measurements from a few more satellites, the receiver can determine the user's position.
GPS satellites transmit two low power radio signals, designated L1 and L2. The signals travel by line of sight, meaning they will pass through clouds, glass and plastic but will not go through most solid objects such as buildings and mountains.
A GPS signal contains three different bits of information – a pseudorandom code, ephemeris data and almanac data. The pseudorandom code is simply an I.D. code that identifies which satellite is transmitting information. Ephemeris data contains important information about the status of the satellite (healthy or unhealthy), current date and time. This part of the signal is essential for determining a position. The almanac data tells the GPS receiver where each GPS satellite should be at any time throughout the day. Each satellite transmits almanac data showing the orbital information for that satellite and for every other satellite in the system.
High accuracy GPS: principles
Assuming that two GPS receivers are close to each other (0–50 km), the different errors reducing the positioning accuracy (mainly atmospheric disturbance) affect both receivers the same way and with the same magnitude. If the exact location of one receiver is known (base receiver), this information can be used to calculate errors in the measurement and then report these errors (or correction values) to the other receiver with unknown position (rover receiver), so that it could compensate for them. This technique is called differential mode (DGPS, see fig. 2). This differential mode removes almost all errors except multipath (fake reflected signals) and receiver errors, because they are local to each receiver. The receiver error is typically about 10 cm for standard DGPS (differential code). If range errors are transmitted from the base receiver to the rover in real-time (radio link), then the system is called real-time DGPS. If real time results are not needed (typically in biomechanics), the measurement are time tagged and recorded in the base and rover receivers and later transferred to a computer to correct the data and calculate an accurate position of the rover at each instant (post processed DGPS).
Figure 2 Differential GPS principles. The satellites are viewed by both receivers, located closed to each other. Reference receiver 1 calculates signal errors for GPS satellites. The correction is used to enhance navigation accuracy of receiver 2.
Real Time Kinematics (RTK) is based on measuring distances to the satellites with carrier phase. As DGPS, this mode requires two receivers (base and rover), but the positioning does not rely on the pseudorandom code sent by satellites, which directly allows the estimation of the distance between the receiver and each satellite. Instead, the electromagnetic carrier of the signal is compared to a similar wave generated by the receiver (high accuracy oscillator). Doppler effect (frequency change due to relative speed between the satellite and the receiver) and phase shift (small time shift between the waves) are repeatedly measured (1–20 times per second). From this data, very small relative displacement between satellites and receiver can be tracked. However, there is a large ambiguity on the total distance (number of integer wave cycles). The solving of these ambiguities – i.e. to find the real number of wave cycles between each satellite and the receiver – is the major issue of RTK. However, by using code data and redundant information from at least 5 satellites, it is possible to lock position. In this case, the theoretical accuracy (given by the manufacturers) of each position computation is between 0.5 to 2 cm horizontal and 1 to 3 cm vertical (with a small baseline, i.e. the short distance between base and rover receivers). This method is very sensitive to sudden satellite loss due to obstructions (missing epochs). Actually, a new ambiguity solving process may be needed each time that there is missing data in the phase and Doppler measurements. Like DGPS, RTK can be performed in real-time or in post processing.
Validation of high accuracy GPS for gait analysis
Most applications of high-end GPS receivers in RTK-mode are static, i.e. implying the precise positioning of a fixed point on earth. Several studies report milimetric accuracy in this case [44], because it is possible to repeatedly measure the fix point and then calculate an average position with a greatly reduced error. Few applications need the kinematic use of RTK mode, i.e. the determination of a trajectory by repeatedly measuring a moving point with a high sampling frequency (10–20 Hz): therefore there are few validation studies in this research area.
In the field of wind engineering and industrial aerodynamics, Tamura and colleagues [45] recently demonstrated that GPS (RTK mode) was capable of an accurate assessment of small sinusoidal displacements (4–10 cm) in the 2–5 Hz frequency range by using a direct comparison with an electronic exciter. The sine-wave was correctly assessed, in terms of both amplitude and phase: the control and GPS curves were totally superimposed. In addition, 0.5 cm oscillation – an amplitude below the theoretical accuracy limits of GPS in RTK mode – was correctly tracked in terms of phase, but with small drift in amplitude in the +/- 1 cm range.
High accuracy GPS: usability and practicability
Strict quality standards are needed in order to reach the highest possible accuracy with GPS in RTK mode for analyzing walking biomechanics: 1) the use of high-quality professional GPS receivers tracking both L1-L2 frequencies is required, such as Topcon Javad or Leica. 2) The time of the measurement must be carefully selected: additional satellites above 5, add redundant information that increases accuracy. We found that optimal accuracy was obtained with at least 7 GPS satellites. 3) No satellite below 20 degrees of elevation above the horizon must be used to reduce multipath (fake satellite signals induced by unpredictable reflections). 4) The smallest possible baseline for the best atmospheric error reduction is mandatory (500 m maximum between the reference receiver and the moving receiver). 5) Special attention should be paid during the RTK post-processing of raw GPS data: the missing epochs, cycle slips and unsolved ambiguities must be carefully monitored and the whole trial should be rejected if too many errors are found: in practice one out of five trial may be subjected to voluntary rejection.
Under such experimental conditions, we assumed that the theoretical limit of 1 cm accuracy could be reached and even overcome: it became possible to calculate gait parameters stride-by-stride. The main drawback is that optimal satellite constellation occurs infrequently during the day (i.e. typically 2 to 3 hours window in the diurnal period). In addition, similar weather conditions should be a pre-requisite to standardize the experiment (this is the case for every outdoor experiment). As a result, it is not possible to efficiently measure a large group of individuals with the current GPS technology.
In practice, our lab uses GPS/GLONASS receivers (Legacy E GDD, Javad Navigation Systems, San Jose, CA, USA). These devices can simultaneously track both American (GPS) and Russian (GLONASS) positioning system, increasing the total number of satellites available. The rover receiver and its power supply (total weight: 0.9 kg) are put into a backpack worn by the subject; the flat antenna (weight: 0.33 kg, 14 × 14 × 3 cm) is rigidly fixed onto a cap. The receivers can acquire both code and carrier phase up to 20 times each second (20 Hz). The raw data are post-processed by using the Javad Pinnacle software and its kinematic engine: the subject's trajectory is assessed by the double-difference method after phase ambiguity resolution. The 3D positions are converted into the Swiss grid coordinate system which provides distance measurements in metric units. The 3D speed vector was also computed for each point of the trajectory. In short, the output file of the trajectory processing contains seven columns for each epoch: time of the measurement (20 Hz, GPS time, nanosecond accuracy), North, East, Altitude (m), Speed North, Speed East, Speed altitude (m/s).
From GPS positioning to gait parameters: the biomechanical assumptions
How can an antenna attached onto the top of the subject's head provide useful information about the stride by stride gait parameters? Beyond the question of positioning accuracy, 4 assumptions must be stated.
1) Average speed of the head over one gait cycle (two steps) is equal to the average body speed and hence average Walking Speed (WS). The head undergoes small rotations in different planes while walking [46]. However, there is no doubt that on average its speed is similar to the trunk and Center of Mass speed, because all body segments are interdependent. Therefore, the vector magnitude of 3D GPS speed vector can be averaged over one gait cycle to assess average walking speed.
2) The head vertically oscillates at the same frequency as the trunk and Center of Mass: the frequency of this oscillation can be defined as Step Frequency (SF). The vertical oscillation of the head has been found to oscillate at the same frequency as the trunk [46]. We have also observed that average SF measured by GPS was identical to average SF measured by an accelerometer attached to the low back [47]. We agree that the definition of SF based on the head trajectory may be different than others, such as the inverse of stride duration, i.e. the time between to heel strikes measured by force plate or footswitch. However, in our opinion, different body segment can be alternatively used to track the rhythmicity of walking with comparable efficiency.
3) One gait parameter can be computed by knowing the two others by the simple equation WS = SF × SL. Because of the repetitive pattern of walking, WS, SF and SL are strictly related. Indeed, walking can be seen as iterative gait cycles in both spatial and temporal dimensions. To the temporal repetition after one stride duration, it adds a spatial repetition after one stride length. The rate at which the spatial repetition occurs is precisely the speed (distance/duration). In practice, the length of step can obviously be defined as the distance traveled by the head over one gait cycle. However, an alternative rationale is that there is no need to measure the 3 gait parameters: it is sufficient to measure two of them and deduce the third. SL can be therefore defined as the ratio between WS and SF. Alternatively, SF can be computed from SL and WS (SF = WS/SL).
4) Accurate head trajectory can be assessed with a low sampling rate (10–20 Hz). The accurate assessment of head trajectory is the main requirement that make possible the computation of all gait parameters with GPS method. Indeed, the assumptions we have defined above (1–3) imply the recognition of a repetitive pattern in the raw trajectory signal in order to analyze each stride separately. In other words, the periodic return of a body segment to a similar state can be used to frame each gait cycle and hence to allow the measurement of the gait parameters stride by stride: the classical example is the repetition of heel strikes. In practice, we arbitrarily chose to detect the max altitude (peak) reached by the head on the vertical axis to define the beginning of each step (see fig. 3). The main obstacle to the detection of this point is that the head trajectory is not continuously tracked, but measured by the GPS receiver as successive discrete positions with a sampling rate ranging from 5 Hz [47-49] to 20 Hz [22]. We are convinced that such a sampling rate is sufficient to mathematically reconstruct the head trajectory with the required accuracy by interpolating extra-points between the GPS measurements. Indeed, there is a high correlation between successive points in the head trajectory, because of the inherent inertia and the low acceleration that are allowed by the system: a smooth trajectory is therefore expected. If the head would undergo small "erratic" unpredictable movements between two GPS points (1/20 s), this would imply a significant acceleration to the head (several g), and this is obviously not the case. In addition, multiple results in the literature clearly demonstrate that the body Center of Mass [24], the trunk [4], and the head [46] follow a sine-like, smooth, trajectory: the frequency of this sine-wave is precisely SF. From a digital signal processing point of view, it is obvious that a 10/20 Hz sampling rate is sufficient to perfectly describe a 1.5–2.5 Hz "sine-like" wave because of the Shannon's theorem. Fig. 3 illustrates the result of the interpolation process (spline interpolation) we apply to increase the temporal accuracy of head trajectory.
Figure 3 Raw GPS data and measurement of the length of step. One participant freely walked on the level ground. High precision GPS measured 3D positions of the moving participant with a centimeter accuracy at 20 Hz sampling rate (antenna fixed onto the head). The figure presents a small sample (3 m) of a 45 min. test. The top panel shows the sinusoidal variation of the vertical position (Z) as a function of the West-East (X) displacement. The bottom panel shows the South-North (Y) displacement as a function of West-East (X) displacement. The vertical lines indicate the beginning of each step. Dotted circles are raw 20 Hz GPS data. Small dots are 240 Hz interpolated positions.
High accuracy GPS and gait variability: the Lausanne results
In 1999 – in the field of physical activity assessment – we studied whether the combination of accelerometer with altimetry would lead to a major improvement of walking speed prediction in a variable slope environment [48]. The high accuracy RTK GPS with 5 Hz sampling rate was used as reference for speed and altitude measurement ("golden standard"). Because the trajectory assessment seemed very accurate, we tested the same instrument (Leica RTK GPS, 5 Hz sampling rate) to measure average walking parameters (WS, SL, SF) over 5 minutes steady state walking [47]. In addition, we measured vertical displacement and speed change stride-by-stride. We found that the average step duration measured with a portable accelerometer was statistically identical to GPS measurement. However, the parameters assessed stride by stride exhibited large variability. In a subsequent study, we attempted to assess average external power of walking [49]. However, the results were not totally in accordance with the results found in the literature, probably because of a poor recording of the phase shift between energy components [49]. More recently, we used a new device (10 Hz sampling rate) that allowed the recording of the basic gait parameters (walking speed, cadence, and step length) over several successive 5 sec periods [50]. We found that walking at low speed induced a different gait pattern compared to walking at preferred or high speed. In addition, slow walking exhibited higher variability of all gait parameters [50].
The most recently study was conducted by applying the method explained above (20 Hz, strict standards) [22]. We analyzed gait parameters stride-by-stride in 8 subjects under free and constrained (metronome) conditions. We obtained time series as illustrated in fig. 4. This allows the analysis of the fluctuation of the gait parameters (walking speed, cadence, and step length) both in terms of amplitude (Standard Deviation, Coefficent of Variation) and dynamics (long range correlation, fractal pattern). Under free walking conditions, DFA (Detrended Fluctuation Analysis [20,21,51-53]) and surrogate data tests showed that the fluctuation of WS, SL and SF exhibited a fractal pattern (i.e., scaling exponent α: 0.5 < α < 1) in a large majority of participants (7/8). Under constrained conditions (metronome), SF fluctuations became significantly anti-correlated (α < 0.5) in all participants. However, the scaling exponent of SL and WS was not modified. We conclude that, when the walking pace is controlled by an auditory signal, the feedback loop between the planned movement (at supraspinal level) and the sensory inputs induces a continual shifting of SF around the mean (persistent anti-correlation), but with no effect on the fluctuation dynamics of the other parameters (SL, WS) [22].
Figure 4 Times series of gait parameters for a walking man (preferred speed). The gait parameters were measured in a male volunteer stride by stride (1 stride = 2 steps) over ~32 min. by using the high accuracy GPS method. The intra-individual (stride to stride) variability is expressed as both Standard Deviation (SD) and Coefficient of Variation (CV = SD/mean × 100). Total distance, number of strides and duration are indicated below.
Advantages and drawbacks of GPS as compared to other methods
GPS technique falls under the category of methods that provide a limited set of biomechanical parameters with an increased practicability, such as, for example, portable accelerometers. The introduction of such a method will not displace high accuracy methods used in the "gait laboratories". However, it can provide useful alternative in the field of gait variability analysis, provided that the potential user is aware of the different constraints. In this context, table 1 summarizes the advantages and drawbacks of GPS.
Table 1 Potential advantages and shortcomings of the Global Positioning System (GPS) technique used for gait analysis
Advantages Shortcomings
Available anywhere on the earth in any weather conditions for outdoor measurements at no cost High cost of professional equipment
Tri-dimensional positioning with centimeter accuracy (Real Time Kinematics, RTK mode) Not fully validated for gait analysis yet
No space restriction: freedom in the path selection, including uphill/downhill locomotion. Limited time windows (2–4 h per day)
Free living conditions, i.e close to real life One body segment measured only (head): Because of mandatory constant satellite access, the antenna must not be obstructed by body parts.
Unlimited number of consecutive strides: limited only by the memory capacity of the receiver and the duration of the batteries. Outdoor analysis: difficult to standardize environmental conditions (weather, terrain).
Not fully miniaturized (cumbersome antenna).
Regarding the technical and organizational obstacles, it seems that the high-accuracy GPS technology is difficult to implement for biomedical applications. Some obstacles are inherent to satellite positioning technique (outdoor experiments, optimal satellite access). However, future developments will increase the usability of the technique. The receivers become smaller with a higher computation power: new 100 Hz GPS chips are already available. Concerning GPS satellites, a challenging modernization program will offer a third civilian frequency (L5) for better availability and accuracy. New additional Russian GLONASS satellites will be also launched in the next few years. The European GALILEO system is planned for the next decade: it will provide a third independent positioning system. Consequently, the accuracy, availability and usability of satellite positioning have a substantial potential for growth.
The development of GPS technique for gait analysis is still embryonic. When the investigators will realize the potential of this new technology, they may use it as a complementary tool to better track the gait parameters of human being in their own "natural" environment. Given the importance of intra-individual variability of these parameters, "exportation" of the laboratory to free-living conditions may be the unique solution to analyze them over prolonged periods of time.
Acknowledgements
The authors thank Mr. V. Turner and the technical staff of the Department of Physiology for their help. The development of GPS technique in human applications was financially supported by the Swiss National Science Foundation (Grant 3200-055928.98/1), by the foundation "Sport, Science et Société" and by the "Loterie Romande".
==== Refs
Delcomyn F Neural basis of rhythmic behavior in animals Science 1980 210 492 498 7423199
Ivanov PC Rosenblum MG Amaral LA Struzik ZR Havlin S Goldberger AL Stanley HE Multifractality in human heartbeat dynamics Nature 1999 399 461 465 10365957 10.1038/20924
Ivanov PC Rosenblum MG Peng CK Mietus J Havlin S Stanley HE Scaling behavior of heartbeat intervals obtained by wavelet-based time-series analysis Nature 1996 383 323 327 8848043 10.1038/383323a0
Kerrigan DC Viramontes BE Corcoran PJ LaRaia PJ Measured versus predicted vertical displacement of the sacrum during gait as a tool to measure biomechanical gait performance Am J Phys Med Rehabil 1995 74 3 8 7873111
Peng CK Mietus JE Liu Y Lee C Hausdorff JM Eugene Stanley H Goldberger AL Lipsitz LA Quantifying fractal dynamics of human respiration: Age and gender effects Ann Biomed Eng 2002 30 683 692 12108842 10.1114/1.1481053
Roberts S Eykholt R Thaut MH Analysis of correlations and search for evidence of deterministic chaos in rhythmic motor control by the human brain Phys Rev E Stat Phys Plasmas Fluids Relat Interdiscip Topics 2000 62 2597 2607 11088739
Alexander RM Energy-saving mechanisms in walking and running J Exp Biol 1991 160 55 69 1960518
Cavagna GA Kaneko M Mechanical work and efficiency in level walking and running J Physiol (Lond) 1977 268 647 681
Cavagna GA Saibene FP Margaria R External work in walking J Appl Physiol 1963 18 1 9 14019429
Saibene S Minetti AE Biomechanical and physiological aspects of legged locomotion in humans Eur J Appl Physiol 2003 88 297 316 12527959
Van de Crommert HW Mulder T Duysens J Neural control of locomotion: sensory control of the central pattern generator and its relation to treadmill training Gait posture 1998 7 251 260 10200392 10.1016/S0966-6362(98)00010-1
Hausdorff JM Nelson ME Kaliton D Layne JE Bernstein MJ Nuernberger A Singh MA Etiology and modification of gait instability in older adults: a randomized controlled trial of exercise J Appl Physiol 2001 90 2117 2129 11356774
Hausdorff JM Rios DA Edelberg HK Gait variability and fall risk in community-living older adults: a 1-year prospective study Arch Phys Med Rehabil 2001 82 1050 1056 11494184 10.1053/apmr.2001.24893
Hausdorff JM Zemany L Peng C Goldberger AL Maturation of gait dynamics: stride-to-stride variability and its temporal organization in children J Appl Physiol 1999 86 1040 1047 10066721
Goldberger AL Amaral LA Hausdorff JM Ivanov PC Peng CK Stanley HE Fractal dynamics in physiology: alterations with disease and aging Proc Natl Acad Sci U S A 2002 99 2466 2472 11875196 10.1073/pnas.012579499
Hausdorff JM Cudkowicz ME Firtion R Wei JY Goldberger AL Gait variability and basal ganglia disorders: stride-to-stride variations of gait cycle timing in Parkinson's disease and Huntington's disease Mov Disord 1998 13 428 437 9613733 10.1002/mds.870130310
Hausdorff JM Ladin Z Wei JY Footswitch system for measurement of the temporal parameters of gait J Biomech 1995 28 347 351 7730393 10.1016/0021-9290(94)00074-E
Hausdorff JM Lertratanakul A Cudkowicz ME Peterson AL Kaliton D Goldberger AL Dynamic markers of altered gait rhythm in amyotrophic lateral sclerosis J Appl Physiol 2000 88 2045 2053 10846017
Hausdorff JM Mitchell SL Firtion R Peng CK Cudkowicz ME Wei JY Goldberger AL Altered fractal dynamics of gait: reduced stride-interval correlations with aging and Huntington's disease J Appl Physiol 1997 82 262 269 9029225
Hausdorff JM Peng CK Ladin Z Wei JY Goldberger AL Is walking a random walk? Evidence for long-range correlations in stride interval of human gait J Appl Physiol 1995 78 349 358 7713836
Hausdorff JM Purdon PL Peng CK Ladin Z Wei JY Goldberger AL Fractal dynamics of human gait: stability of long-range correlations in stride interval fluctuations J Appl Physiol 1996 80 1448 1457 8727526
Terrier P Schutz Y GPS analysis of human locomotion: further evidence for long-range correlations in the stride-to-stride fluctuations of gait parameters Hum Mov Sci 2005
Rose J Gamble JG Human Walking 1994 Second Baltimore, Williams & Wilkins
Cavagna GA Thys H Zamboni A The sources of external work in level walking and running J Physiol (Lond) 1976 262 639 657 1011078
Cavagna GA Force platforms as ergometers J Appl Physiol 1975 39 174 179 1150585
Cutlip RG Mancinelli C Huber F DiPasquale J Evaluation of an instrumented walkway for measurement of the kinematic parameters of gait Gait Posture 2000 12 134 138 10998610 10.1016/S0966-6362(00)00062-X
Dierick F Penta M Renaut D Detrembleur C A force measuring treadmill in clinical gait analysis Gait Posture 2004 20 299 303 15531177 10.1016/j.gaitpost.2003.11.001
Belli A Avela J Komi PV Mechanical energy assessment with different methods during running Int J Sports Med 1993 14 252 256 8365832
Belli A Lacour JR Komi PV Candau R Denis C Mechanical step variability during treadmill running Eur J Appl Physiol Occup Physiol 1995 70 510 517 7556123 10.1007/BF00634380
Belli A Rey S Bonnefoy R Lacour JR A simple device for kinematic measurements of human movement Ergonomics 1992 35 177 186
Aminian K Najafi B Bula C Leyvraz PF Robert P Spatio-temporal parameters of gait measured by an ambulatory system using miniature gyroscopes J Biomech 2002 35 689 699 11955509 10.1016/S0021-9290(02)00008-8
Aminian K Robert P Jequier E Schutz Y Incline, speed, distance assessment during unconstrained walking Med Sci Sports Exerc 1995 27 226 234 7723646
Herren R Sparti A Aminian K Schutz Y The prediction of speed and incline in outdoor running in humans using accelerometry Med Sci Sports Exerc 1999 31 1053 1059 10416569 10.1097/00005768-199907000-00020
Mayagoitia RE Nene AV Veltink PH Accelerometer and rate gyroscope measurement of kinematics: an inexpensive alternative to optical motion analysis systems J Biomech 2002 35 537 542 11934425 10.1016/S0021-9290(01)00231-7
Paraschiv-Ionescu A Buchser EE Rutschmann B Najafi B Aminian K Ambulatory system for the quantitative and qualitative analysis of gait and posture in chronic pain patients treated with spinal cord stimulation Gait Posture 2004 20 113 125 15336280 10.1016/j.gaitpost.2003.07.005
Schutz Y Weinsier S Terrier P Durrer D A new accelerometric method to assess the daily walking practice Int J Obes Relat Metab Disord 2002 26 111 118 11791155 10.1038/sj.ijo.0801856
Tong K Granat MH A practical gait analysis system using gyroscopes Med Eng Phys 1999 21 87 94 10426508 10.1016/S1350-4533(99)00030-2
Schutz Y Chambaz A Could a satellite-based navigation system (GPS) be used to assess the physical activity of individuals on earth? Eur J Clin Nutr 1997 51 338 339 9152686 10.1038/sj.ejcn.1600403
Schutz Y Herren R Assessment of speed of human locomotion using a differential satellite global positioning system Med Sci Sports Exerc 2000 32 642 646 10731007 10.1097/00005768-200003000-00014
Witte TH Wilson AM Accuracy of non-differential GPS for the determination of speed over ground J Biomech 2004 37 1891 1898 15519597 10.1016/j.jbiomech.2004.02.031
Larsson P Burlin L Jakobsson E Henriksson-Larsen K Analysis of performance in orienteering with treadmill tests and physiological field tests using a differential global positioning system J Sports Sci 2002 20 529 535 12166879 10.1080/026404102760000035
Larsson P Global positioning system and sport-specific testing Sports Med 2003 33 1093 1101 14719979
Larsson P Henriksson-Larsen K The use of dGPS and simultaneous metabolic measurements during orienteering Med Sci Sports Exerc 2001 33 1919 24 11689744
Greenfeld JS How long must we wait? The duration of GPS observations for short baseline measurements Survey Land Inf Sci 2003 63 77 86
Tamura Y Matsui M Pagnini LC Yoshida A Measurement of wind-induced response of buildings using RTK-GPS J Wind Eng Indus Aerodyn 2002 90 1783 1793 10.1016/S0167-6105(02)00287-8
Hirasaki E Moore ST Raphan T Cohen B Effects of walking velocity on vertical head and body movements during locomotion Exp Brain Res 1999 127 117 130 10442403 10.1007/s002210050781
Terrier P Ladetto Q Merminod B Schutz Y High-precision satellite positioning system as a new tool to study the biomechanics of human locomotion J Biomech 2000 33 1717 1722 11006399 10.1016/S0021-9290(00)00133-0
Perrin O Terrier P Ladetto Q Merminod B Schutz Y Improvement of walking speed prediction by accelerometry and altimetry, validated by satellite positioning Med Biol Eng Comput 2000 38 164 168 10829408
Terrier P Ladetto Q Merminod B Schutz Y Measurement of the mechanical power of walking by satellite positioning system (GPS) Med Sci Sports Exerc 2001 33 1912 1918 11689743
Terrier P Schutz Y Variability of gait patterns during unconstrained walking assessed by satellite positioning (GPS) Eur J Appl Physiol 2003 90 554 561 12905048 10.1007/s00421-003-0906-3
Chen Z Ivanov PC Hu K Stanley HE Effect of nonstationarities on detrended fluctuation analysis Phys Rev E Stat Nonlin Soft Matter Phys 2002 65 041107 12005806
Hu K Ivanov PC Chen Z Carpena P Eugene Stanley H Effect of trends on detrended fluctuation analysis Phys Rev E Stat Nonlin Soft Matter Phys 2001 64 011114 11461232
Peng CK Buldyrev SV Havlin S Simons M Stanley HE Goldberger AL Mosaic organization of DNA nucleotides Phys Rev E Stat Phys Plasmas Fluids Relat Interdiscip Topics 1994 49 1685 1689 9961383
|
16138922
|
PMC1224864
|
CC BY
|
2021-01-04 16:37:40
|
no
|
J Neuroengineering Rehabil. 2005 Sep 2; 2:28
|
utf-8
|
J Neuroeng Rehabil
| 2,005 |
10.1186/1743-0003-2-28
|
oa_comm
|
==== Front
Malar JMalaria Journal1475-2875BioMed Central London 1475-2875-4-381609823210.1186/1475-2875-4-38ResearchA functional polymorphism in the IL1B gene promoter, IL1B -31C>T, is not associated with cerebral malaria in Thailand Ohashi Jun [email protected] Izumi [email protected] Akihiro [email protected] Jintana [email protected] Hathairad [email protected] Noppadon [email protected] Sornchai [email protected] Katsushi [email protected] Department of Human Genetics, Graduate School of Medicine, The University of Tokyo,7-3-1 Hongo, Bunkyo-ku, Tokyo 113-0033, Japan2 Faculty of Tropical Medicine, Mahidol University, 420/6 Rajvithi Road, Bangkok, Thailand2005 14 8 2005 4 38 38 19 5 2005 14 8 2005 Copyright © 2005 Ohashi et al; licensee BioMed Central Ltd.2005Ohashi et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
IL-1β and IL-1RA levels are higher in the serum of cerebral malaria patients than in patients with mild malaria. Recently, the level of IL1B expression was reported to be influenced by a polymorphism in the promoter of IL1, IL1B -31C>T.
Methods
To examine whether polymorphisms in IL1B and IL1RA influence the susceptibility to cerebral malaria, IL1B -31C>T, IL1B 3953C>T, and IL1RA variable number of tandem repeat (VNTR) were analysed in 312 Thai patients with malaria (109 cerebral malaria and 203 mild malaria patients).
Results
In this population, IL1B -31C>T and IL1RA VNTRwere detected, while IL1B 3953C>T (i.e., IL1B 3953T) was not observed in the polymorphism screening for 32 patients. Further analyses for IL1B -31C>T and IL1RA VNTR in 110 cerebral malaria and 206 mild malaria patients showed no significant association of these polymorphisms with cerebral malaria.
Conclusion
The present results suggest that IL1B -31C>T and IL1RA VNTR polymorphisms do not play a crucial role in susceptibility or resistance to cerebral malaria.
==== Body
Introduction
Interleukin 1 (IL-1) is a proinflammatory cytokine that has been suggested to play a crucial role in the pathogenesis of cerebral malaria [1]. Histopathology and immunohistologic studies detected IL-1β in the brain and liver from a case of fatal cerebral malaria [2]. Furthermore, expression of IL1B mRNA was detected in the spleen, cortex, cerebellum, and brain stem of most pediatric patients that died of cerebral malaria, but it was not detected in those of uninfected individuals [3]. In addition, the concentration of IL-1 receptor antagonist (IL-1RA), which competes for receptor binding with IL-1, was also found to be increased in Gambian children with cerebral malaria [4]. These observations raise the question of whether polymorphisms in the genes encoding IL-1β (IL1B) and IL-1RA (IL1RA) are associated with cerebral malaria.
IL1 family genes including IL1B and IL1RA form a cytokine gene cluster on human chromosome 2. Recently, the IL1B 3953C>T polymorphism, which lies in exon 5 of IL1B, was reported to be associated with cerebral malaria in Gambia [5]. Gyan et al. [6] reported that, in Ghanaian Children, levels of parasitemia were significantly higher in uncomplicated malaria patients possessing the IL1B 3953T allele (i.e., 3953T/T and 3953C/T genotypes) than those possessing only 3953C (i.e., 3953C/C genotype), although this polymorphism and the IL1RA variable number of tandem repeat (VNTR) polymorphism were not associated with cerebral malaria.
The functional significance of the IL1B 3953C>T polymorphism has been studied. Some results show an association of IL1B 3953C>T with the levels of IL-1 production [7,8], whereas others show no association [9,10]. Recently, Kimura et al. [11] employed allele-specific transcript quantification to examine the effects of IL1B -31C>T and 3953C>T polymorphisms on the level of IL-1B transcription. Their results revealed that expression of the -31T allele was significantly higher than that of the -31C allele, while IL1B 3953C>T did not affect the expression of IL1B. The increased level of -31T expression corresponds with the observation that the -31T allele constitutes a TATA sequence in the IL1B promoter and induces formation of the transcription initiation complex [12]. Therefore, IL1B -31C>T is an important candidate polymorphism that may influence the susceptibility or resistance to cerebral malaria through the upregulation of IL-1β. In addition, the IL1RA VNTR polymorphism influences the plasma IL-1RA levels [13], and IL-1RA modulates IL-1 production [14], although there was no significant association of IL1B -31C>T and IL1RA VNTR with cerebral malaria in Gambia [5]. In this study, IL1B -31C>T, IL1B 3953C>T, and IL1RA VNTR polymorphisms were analysed in Thai malaria patients to examine the possible association of these polymorphisms with cerebral malaria.
Materials and methods
Subjects
A total of 316 adult patients with P. falciparum malaria (110 cerebral malaria and 206 mild malaria patients) living in northwest Thailand were recruited in this study. All of them underwent treatment at the Hospital for Tropical Diseases, Faculty of Tropical Medicine, Mahidol University. Clinical manifestations of malaria have been described previously [15]. All individuals were 13 years of age or older, and the mean ages of the mild malaria and cerebral malaria patients were 25.5 and 28.6, respectively. This study was approved by the institute review board of the Faculty of Tropical Medicine, Mahidol University and the Faculty of Medicine, The University of Tokyo. Informed written consent was obtained from all patients.
Genotyping
Genomic DNA was extracted from peripheral blood leukocytes using a QIAamp Blood Kit (Qiagen, Hilden, Germany). Genotyping for the three polymorphisms was performed as previously described [11,16]. For the IL1B promoter polymorphism, IL1B -31C>T [11], PCR was carried out with primers IL1B5'F (5'-TAGTCCCCTCCCCTAAGAACG-3') and IL1Bint1R (5'-CCCAGAATATTTCCCGAGTCA-3') to amplify the region including IL1B -31C>T, and the PCR products were treated with AluI (New England Biolabs, Beverly, CA, USA), which digests the -31T allele. For the IL1B exon 5 polymorphism, IL1B 3953C>T [11], PCR was carried out with primers IL1Bint4F (5'-GCTCAGGTGTCCTCCAAGAAA-3') and IL1Bint5R (5'-GGCCAGTGCAATCAAATGTG-3'), and the PCR products were treated with TaqI (New England Biolabs), which digests the 3953C allele. For the IL-1RA VNTR polymorphism [16], the region including variable numbers of identical 86-bp tandem repeats was amplified by PCR using the following primers: 5'-CTCAGCCAACACTCCTAT-3' and 5'-TCCTGGTCTGCAGGTAA-3'. PCR products of 240(allele 2, two repeats), 325 (allele 3, three repeats), 410 (allele 4, four repeats), and 500 bp (allele 5, five repeats) were distinguished by agarose gel electrophoresis.
Statistical analysis
Deviation from Hardy-Weinberg equilibrium was examined with Arlequin software [17] with the default setting, where the exact P-value is calculated based on the Markov chain method [18]. Linkage disequilibrium (LD) between IL1B -31C>T and IL1RA VNTR was evaluated based on the expectation maximization (EM) algorithm using the Arlequin software with the default setting, where a permutation procedure is performed to calculate the P-value for the likelihood ratio test [19]. Fisher's exact test was performed to compare the genotype and allele frequency distributions between cerebral and mild malaria patients. The P-values for association of IL1B -31C>T with cerebral malaria were calculated based on 3 × 2 and 2 × 2 contingency tables for genotype and allele frequency distributions, respectively. The P-values for association of IL1RA VNTR with cerebral malaria were calculated based on 5 × 2 and 4 × 2 contingency tables for genotype and allele frequency distributions, respectively. Furthermore, Fisher's exact test was carried out for the comparison of each genotype frequency between mild and cerebral malaria patients based on 2 × 2 table for IL1B -31C>T and IL1RA VNTR. The frequencies of haplotype consisting of IL1B -31C>T and IL1RA VNTR were estimated based on EM algorithm [19] using Arlequin software [17]. The difference in haplotype frequency distribution between mild and cerebral malaria patiens was assessed by a likelihood ratio test. In the EM algorithm for the estimation of haplotype frequency, the logarithm of likelihood of the sample, ln L, was maximized in each subject group (i.e., lnLm for mild malaria subjects, lnLc for cerebral malaria subjects, and lnLall for the combined subjects). The likelihood-ratio statistic was defined by LR = 2 [lnLm + lnLc - lnLall], which had an approximate χ2 distribution with 7 degrees of freedom under the null hypothesis of no difference in haplotype frequency distribution. It should be noted here that there were 8 possible haplotypes, thus the degree of freedom was 7 in the analysis. The P-values for association of genotypes of IL1B -31C>T and TNFA -308G>A with cerebral malaria were calculated by Fisher's exact test based on 2 × 2 contingency table for each genotype.
Results
Because only the 3953C/C genotype was found in 16 patients with mild malaria and 16 patients with cerebral malaria in the polymorphism screening, a further genotyping was not performed. The genotype and allele frequencies of the IL1B -31C>T and IL1RA VNTR polymorphisms in Thai malaria patients are shown in Tables 1 and 2. It should be noted that some patients were not genotyped for both IL1B -31C>T and IL1RA VNTR because the target fragment was not amplified by PCR. In Thai malaria patients, the population frequencies of the -31C and -31T alleles were nearly equal, and allele 4 (with four repeats) was predominant in the IL1RN VNTR polymorphism. Genotype frequencies did not deviate from predictions of the Hardy-Weinberg equilibrium either in IL1B -31C>T (for mild malaria patients, P = 0.58; for cerebral malaria patients, P = 0.44) or in IL1RA VNTR (for mild malaria patients, P = 0.27; for cerebral malaria patients, P = 0.73). Significant differences in genotype and allele frequency distributions were not observed between mild and cerebral malaria patients for IL1B -31C>T or IL1RA VNTR (Tables 1 and 2). In addition, there was no difference in each genotype frequency between mild and cerebral malaria patients (data not shown).
Table 1 Genotype and allele frequencies of IL1B -31C/T polymorphism in Thai malaria patients.
Mild malaria (n = 192) Cerebral malaria (n = 106) Association P-valuea
Genotype
IL1B -31C>T
C/C 55 (0.286) 28 (0.264) 0.82
C/T 91 (0.474) 49 (0.462)
T/T 46 (0.240) 29 (0.274)
Allele
IL1B -31C>T
C 201 (0.523) 105 (0.495) 0.55
T 183 (0.477) 107 (0.505)
Frequencies are in parentheses.
aP-values were calculated by the Fisher's exact test based on 3 × 2 and 2 × 2 contingency tables for genotype and allele frequency, respectively.
Table 2 Genotype and allele frequencies of IL1RA VNTR polymorphism in Thai malaria patients.
Mild malaria (n = 203) Cerebral malaria (n = 109) Association P-valuea
Genotype
2/2b 2 (0.010) 2 (0.018) 0.88
2/4 42 (0.207) 24 (0.220)
3/4 3 (0.015) 1 (0.009)
4/4 154 (0.759) 82 (0.752)
4/5 2 (0.010) 0 (0.000)
Allele
IL1RA
2 46 (0.113) 28 (0.128) 0.84
3 3 (0.007) 1 (0.005)
4 355 (0.874) 189 (0.867)
5 2 (0.005) 0 (0.000)
Frequencies are in parentheses.
a P-values were calculated by the Fisher's exact test based on 5 × 2 and 4 × 2 contingency tables for genotype and allele frequency, respectively.
bIt should be noted here that allele with n repeats of VNTR is designated as "n" in this table.
The LD between IL1B -31C>T and IL1RA VNTR was not strong in the studied population (for mild malaria patients, P = 0.14 by likelihood ratio test based on a permutation precedure; for cerebral malaria patients, P = 0.50 by likelihood ratio test based on a permutation precedure). The estimated frequencies of haplotype consisting of IL1B -31C>T and IL1RA VNTR polymorphisms were shown in Table 3. No significant difference in the estimated haplotype frequency distribution was detected by a likelihood ratio test (P = 0.60).
Table 3 Estimated haplotype frequencies of IL1B -31C>T and IL1RA VNTR polymorphisms in Thai malaria patients.
Haplotypea Mild malaria Cerebral malaria Association P-valueb
IL1B – IL1RA
-31C - 2c 0.042 0.046 0.60
-31T - 2 0.072 0.084
-31C - 4 0.432 0.454
-31T - 4 0.443 0.414
aOnly haplotypes with the estimated haplotype frequency of more than 0.01 were presented.
bP-value was calculated by a likelihood ratio test.
cIt should be noted here that allele with n repeats of VNTR is designated as "n" in this table.
Because several association tests were performed for IL1B -31C>T, IL1RA VNTR, or the haplotype, the obtained P-values should be corrected based on the number of testings by using a proper method (e.g., Bonferroni's correction). However, none of raw P-values for the association tests were less than 0.05. Thus, the corrected P-values are not presented in this study.
Discussion
In this study, possible association of IL1B -31C>T and IL1RA VNTR polymorphisms with cerebral malaria was assessed in 316 Thai malaria patients. These polymorphisms were found not to be associated with cerebral malaria. If the increased expression of IL-1B observed in brain of cerebral malaria patients [2,3] is due to the increased transcription of IL1B caused by the IL1B -31T allele [11], the frequency of IL1B -31T should be higher in cerebral malaria patients than in mild malaria patients. However, the allele frequency of IL1B -31T was not significantly higher in the cerebral malaria patients. Because the allele frequency of IL1B -31T is approximately 0.5 in the studied population, the present study yields a statistical power high enough for the detection of an association between IL1B -31T and cerebral malaria unless the association is very weak. For example, according to the formurae for calculation of statistical power developed by Ohashi et al. [20], when the penetrances for the onset of cerebral malaria are 0.04 for -31T/-31T, 0.02 for -31C/-31T, and 0.01 for -31C/-31C, the estimated power of this study with the significance level of 0.05 exceeds 0.98.
The plasma level of IL1-RA was influenced by the number of repeats in the IL1RA VNTR [13], and the allele with two repeats (designated as allele 2 in the present study) has been reported to be associated with various diseases. Like IL1B -31T, our study yields a high statistical power for allele 2 of IL1RA VNTR if the association is not weak. However, neither IL1B -31C>T nor IL1RA VNTR showed any association with cerebral malaria. Furthermore, genotype combinations of these polymorphisms were not associated with cerebral malaria (data not shown). The present results are consistent with previous ones in Gambia, where these polymorphisms are not associated with cerebral malaria [5]. Taken together, we conclude that the IL1B -31C>T and IL1RA VNTR polymorphisms do not play a crucial role in susceptibility or resistance to cerebral malaria.
IL-1 is known to synergize with tumor necrosis factor-α (TNF-α), and their serum levels are thought to play a role in the severity of malaria [21]. The promoter allele of TNF-α gene (TNFA), TNFA -308A, was reported to be associated with susceptibility to cerebral malaria in Gambia [22]. Although TNFA -308A was not associated with cerebral malaria in our previous study for the same Thai malaria patients [23], it is interesting to examine the possible association of combined genotypes of IL1B -31C>T and TNFA -308G>A polymorphisms with cerebral malaria. Table 4 shows the frequencies of nine combined genotypes of IL1B -31C>T and TNFA -308G>A in mild and cerebral malaria patients. There was no significant difference in each genotype frequency between mild and cerebral malaria patients, suggests that there is no interactive effect of two polymorphisms on the risk of cerebral malaria.
Table 4 Genotypes of IL1B -31C>T and TNFA -308G>A in Thai malaria patients.
TNFA -308G>A
G/G G/A A/A
Mild malaria (n = 192)
IL1B -31C>T
C/Cb 48 (0.250) 7 (0.036) 0 (0.000)
C/T 85 (0.443) 6 (0.031) 0 (0.000)
T/T 43 (0.224) 3 (0.016) 0 (0.000)
Cerebralmalaria (n = 105)
IL1B -31C>T
C/C 25 (0.238) 2 (0.019) 0 (0.000)
C/T 41 (0.390) 8 (0.076) 0 (0.000)
T/T 25 (0.238) 3 (0.029) 1 (0.009)
Frequencies are in parentheses.
There was no significant difference in each genotype frequency between mild and cerebral malaria by the Fisher's exact test based on 2 × 2 contingency table where all the other genotypes are combined.
IL1B and IL1RA are located closely on human chromosome 2. Although IL1RA VNTR is in strong LD with IL1B polymorphisms, such as IL1B -511C>T, IL1B -31C>T, and IL1B 3953C>T in Caucasian populations [12,10,21], significant LD was not observed between IL1RA VNTR and IL1B -31C>T in the Thai population. In Indonesians, LD was not found even between IL1B -31C>T and 3953C>T [11]. These observations suggest that the structure of LD in the IL1 gene cluster differs among populations. Because significant LD was not observed between IL1B and IL1RA in the present study, we cannot exclude the possibility that polymorphisms in the other IL1 family genes located between IL1B and IL1RA, such as IL1 family member 9 (IL1F9) and IL1 family member 8 isoform 1 (IL1F8), are associated with cerebral malaria. It is currently unknown whether associations between IL1B 3953C>T and cerebral malaria [5] or parasitemia [6] in African populations result from a LD with other functional polymorphisms in this chromosomal region. Thus, further studies of various populations are required to clarify whether polymorphisms of IL1 family genes are involved in the pathogenesis of cerebral malaria.
Authors' contributions
S. Looareesuwan and N. Tangpukdee diagnosed malaria and collected blood samples. J. Patarapotikul and H. Hananantachai extracted DNA from the blood samples. J. Ohashi and J. Patarapotikul designed this study. I. Naka performed genotyping with A. Doi. J. Ohashi did statistical analyses and wrote the report. K. Tokunaga and S. Looareesuwan, as scientific coordinators of this project, received the financial support.
Acknowledgements
The authors sincerely thank the patients who participated in this study. We are indebted to Dr Ryosuke Kimura for introducing the PCR-RFLP methods for the IL1B polymorphisms. We thank three anonymous reviewers for valuable comments and suggestions. This study was supported by the Core University System Exchange Programme under the Japan Society for the Promotion of Science, coordinated by the University of Tokyo and Mahidol University; The National Research Council of Thailand; Mahidol University Grant and a Grant-in-Aid for Scientific Research from the Ministry of Education, Culture, Sports, Science, and Technology of Japan.
==== Refs
Clark IA Rockett KA The cytokine theory of human cerebral malaria Parasitology Today 1994 10 410 412 15275552 10.1016/0169-4758(94)90237-2
Udomsangpetch R Chivapat S Viriyavejakul P Riganti M Wilairatana P Pongponratin E Looareesuwan S Involvement of cytokines in the histopathology of cerebral malaria Am J Trop Med Hyg 1997 57 501 506 9392586
Brown H Turner G Rogerson S Tembo M Mwenechanya J Molyneux M Taylor T Cytokine expression in the brain in human cerebral malaria J Infect Dis 1999 180 1742 1746 10515846 10.1086/315078
Jakobsen PH McKay V Morris-Jones SD McGuire W van Hensbroek MB Meisner S Bendtzen K Schousboe I Bygbjerg IC Greenwood BM Increased concentrations of interleukin-6 and interleukin-1 receptor antagonist and decreased concentrations of beta-2-glycoprotein I in Gambian children with cerebral malaria Infect Immun 1994 62 4374 4379 7927698
Walley AJ Aucan C Kwiatkowski D Hill AV Interleukin-1 gene cluster polymorphisms and susceptibility to clinical malaria in a Gambian case-control study, Eur J Hum Genet 2004 12 132 138 14673470 10.1038/sj.ejhg.5201084
Gyan B Goka B Cvetkovic JT Perlmann H Lefvert A-K Akanmori B Troye-Blomberg M Polymorphisms in interleukin-1 and interleukin-1 receptor antagonist genes and malaria in Ghanaian children Scand J Immunol 2002 56 619 622 12472674 10.1046/j.1365-3083.2002.01161.x
Pociot F Molvig J Wogensen L Worsaae H Nerup J A TaqI polymorphism in the human interleukin-1 beta (IL-1 beta) gene correlates with IL-1 beta secretion in vitro Eur J Clin Invest 1992 22 396 402 1353022
Hernandez-Guerrero C Monzon-Bordonaba F Jimenez-Zamudio L Ahued-Ahued R Arechavaleta-Velasco F Strauss JF 3rdVadillo-Ortega F In vitro secretion of proinflammatory cytokines by human amniochorion carrying hyper-responsive gene polymorphisms of tumour necrosis factor-alpha and interleukin-1beta Mol Hum Reprod 2003 9 625 629 12970400 10.1093/molehr/gag076
Santtila S Savinainen K Hurme M Presence of the IL-1RA allele 2 (IL1RN*2) is associated with enhanced IL-1beta production in vitro Scand J Immunol 1998 47 195 198 9519856 10.1046/j.1365-3083.1998.00300.x
Dominici R Malferrari G Mariani C Grimaldi I Biunno I The Interleukin 1-beta exonic (+3953) polymorphism does not alter in vitro protein secretion Exp Mol Pathol 2002 73 139 41 12231216 10.1006/exmp.2002.2435
Kimura R Nishioka T Soemantri A Ishida T Cis-acting effect of the IL1B C-31T polymorphism on IL-1 mRNA expression Genes and Immunity 2004 5 572 575 15356674 10.1038/sj.gene.6364128
El-Omar EM Carrington M Chow WH McColl KE Bream JH Young HA Herrera J Lissowska J Yuan CC Rothman N Lanyon G Martin M Fraumeni F JrRabkin CS Interleukin-1 polymorphisms associated with increased risk of gastric cancer Nature 2002 404 398 402 10746728 10.1038/35006081
Hurme M Santtila S IL-1 receptor antagonist (IL-1Ra) plasma levels are co-ordinately regulated by both IL-1Ra and IL-1beta genes Eur J Immunol 1998 28 2598 2602 9710237 10.1002/(SICI)1521-4141(199808)28:08<2598::AID-IMMU2598>3.3.CO;2-B
Vamvakopoulos J Green C Metcalfe S Genetic control of IL-1beta bioactivity through differential regulation of the IL-1 receptor antagonist Eur J Immunol 2002 32 2988 2996 12355453 10.1002/1521-4141(2002010)32:10<2988::AID-IMMU2988>3.0.CO;2-9
Ohashi J Naka I Patarapotikul J Hananantachai H Looareesuwan S Tokunaga K Significant association of longer forms of CCTTT microsatellite repeat in inducible nitric oxide synthase (iNOS) promoter with severe malaria in Thailand J Infect Dis 2002 186 578 581 12195390 10.1086/341779
Tarlow JK Blakemore AI Lennard A Solari R Hughes HN Steinkasserer A Duff GW Polymorphism in human IL-1 receptor antagonist gene intron 2 is caused by variable numbers of an 86-bp tandem repeat Hum Genet 1993 91 403 404 8500797 10.1007/BF00217368
Schneider S Roessli D Excoffier L Arlequin ver. 2.000: A software for population genetic data analysis, Genetics and Biometry Laboratory 2000 University of Geneva, Switzerland
Guo SW Thompson EA Performing the exact test of Hardy-Weinberg proportions for multiple alleles Biometrics 1992 48 361 372 1637966
Excoffier L Slatkin M Maximum-likelihood estimation of molecular haplotype frequencies in a diploid population Mol Biol Evol 1995 12 921 927 7476138
Ohashi J Yamamoto S Tsuchiya N Hatta Y Komata T Matsushita M Tokunaga K Comparison of statistical power between 2 × 2 allele frequency and allele positivity tables in case-control studies of complex disease genes Ann Hum Genet 2001 65 197 206 11434330 10.1017/S000348000100851X
Rockett KA Awburn MM Rockett EJ Clark IA Tumor necrosis factor and interleukin-1 synergy in the context of malaria pathology Am J Trop Med Hyg 1994 50 735 4216 8024067
McGuire W Hill AV Allsopp CE Greenwood BM Kwiatkowski D Variation in the TNF-alpha promoter region associated with susceptibility to cerebral malaria Nature 1994 371 508 510 7935762 10.1038/371508a0
Hananantachai H Patarapotikul J Looareesuwan S Ohashi J Naka I Tokunaga K Lack of association of -308A/G TNFA promoter and 196R/M TNFR2 polymorphisms with disease severity in Thai adult malaria patients Am J Med Genet 2001 102 391 392 11503171 10.1002/ajmg.1486
Cox A Camp NJ Nicklin MJ di Giovine FS Duff GW An analysis of linkage disequilibrium in the interleukin-1 gene cluster, using a novel grouping method for multiallelic markers Am J Hum Genet 1998 62 1180 1188 9545388 10.1086/301817
|
16098232
|
PMC1224865
|
CC BY
|
2021-01-04 16:24:13
|
no
|
Malar J. 2005 Aug 14; 4:38
|
utf-8
|
Malar J
| 2,005 |
10.1186/1475-2875-4-38
|
oa_comm
|
==== Front
Microb Cell FactMicrobial Cell Factories1475-2859BioMed Central London 1475-2859-4-271615689310.1186/1475-2859-4-27ResearchAggregation as bacterial inclusion bodies does not imply inactivation of enzymes and fluorescent proteins García-Fruitós Elena [email protected]ález-Montalbán Nuria [email protected] Montse [email protected] Andrea [email protected] Rosa María [email protected]ís Anna [email protected] Salvador [email protected] Antonio [email protected] Institut de Biotecnologia i de Biomedicina, Universitat Autònoma de Barcelona, Bellaterra, 08193 Barcelona, Spain2 Departament de Genètica i de Microbiologia, Universitat Autònoma de Barcelona, Bellaterra, 08193 Barcelona, Spain3 Departament de Bioloquímica i de Biologia Molecular, Universitat Autònoma de Barcelona, Bellaterra, 08193 Barcelona, Spain2005 12 9 2005 4 27 27 11 8 2005 12 9 2005 Copyright © 2005 García-Fruitós et al; licensee BioMed Central Ltd.2005García-Fruitós et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Many enzymes of industrial interest are not in the market since they are bio-produced as bacterial inclusion bodies, believed to be biologically inert aggregates of insoluble protein.
Results
By using two structurally and functionally different model enzymes and two fluorescent proteins we show that physiological aggregation in bacteria might only result in a moderate loss of biological activity and that inclusion bodies can be used in reaction mixtures for efficient catalysis.
Conclusion
This observation offers promising possibilities for the exploration of inclusion bodies as catalysts for industrial purposes, without any previous protein-refolding step.
==== Body
Background
Protein misfolding is a common event during bacterial over-expression of recombinant genes [1]. The aggregation of insoluble polypeptide chains as inclusion bodies has seriously restricted the spectrum of proteins marketed by the biotechnology industry. Being widespreadly believed that inclusion body proteins are biologically inactive and therefore useless in bioprocesses, many aggregation-prone products have been disregarded for commercialisation. Protein solubility can be tailored by either process [2] or protein [3] engineering, although most efforts have been addressed to minimize inclusion body formation by co-production of folding modulators [4], or to refold purified inclusion body proteins by chemical denaturation followed by refolding procedures [5]. Both strategies need to be adapted to particular protein species and they render largely variable results regarding the final soluble protein yield.
Interestingly, independent reports have noted enzymatic activity associated to inclusion bodies formed by recombinant enzymes [6-8], but the extent of these side-observations has been never quantified and its biological and biotechnological relevance remained unexplored. In this work, we have quantitatively explored the biological activity of inclusion body recombinant proteins and their potential use for bioprocesses in the aggregated form.
Results
To determine the occurrence of active protein in inclusion bodies we analysed those formed upon overproduction of the wild-type human dihydrofolate reductase (hDHFR) and an engineered E. coli β-galactosidase fused to the aggregation-prone foot-and-mouth disease virus (FMDV) VP1 capsid protein (VP1LAC). In addition, we explored fluorescence emission of green and blue fluorescent proteins (GFP and BFP respectively) fused to different aggregating polypeptides, namely the FMDV VP1 and a point mutant of the human Aβ-amyloid peptide (Aβ(F19D)), by comparing specific fluorescence emission of protein in the soluble cell fraction and purified inclusion bodies. Upon overproduction, all these proteins form cytoplasmic inclusion bodies in E. coli, the fraction of the aggregated protein ranging between 28 and 88 % of the total recombinant production (Table 1). Surprisingly, both enzymatic activity and specific fluorescence of inclusion body proteins were unexpectedly high (Table 1), ranging from 6 to 166 % of that of their counterparts occurring in the soluble cell fraction. This fact indicates that protein inactivation mediated by in vivo aggregation is only moderate. In addition, it is shown that protein packaging as bacterial inclusion bodies into inter-molecular β-sheet architecture (characterized by the presence of a peak around 1620 cm-1 that dominates the FTIR spectrum in the amide I region) [9,10] in these model proteins (Figure 1) is compatible with the functionality of enzyme active sites and fluorophores. In this context, VP1GFP and Aβ42(F19D)-BFP inclusion bodies are noticeably fluorescent inside the producing cells (Figure 2).
Table 1 Enzymatic activity or fluorescence of inclusion bodies produced in E. coli
Construct name Reference Functional protein Fraction of inclusion body protein (range, %) a Aggregating domain or protein (all in the N-terminal position) Specific activity or emission b (enzymatic units/mg or fluorescence units/mg) Activity of the inclusion body fraction relative to that of soluble protein (%) c
Soluble protein Inclusion bodies
VP1LAC This work and [9] E. coli β-galactosidase 35.6–45.9 FMDV VP1 capsid protein 698.3 ± 153.0 1162.5 ± 256.0 166.4
hDHFR [25] Human dihydrofolate reductase 28.4–36.8 none 8.0 10-2 ± 2.6 10-2 4.7 10-3 ± 0.9 10-3 5.9
VP1GFP This work Green fluorescent protein 82.5–88.4 FMDV VP1 capsid protein 359.5 ± 66.0 70.4 ± 10.1 19.5
Aβ42(F19D)-BFP [26] Blue fluorescent protein 61.4–65.3 Aβ42(F19D) 118.1 ± 10.2 36.3 ± 2.2 30.7
a The percentage of protein found in inclusion bodies relative to the total intracellular amount of recombinant protein. Values were determined from different samples taken at 3 and 5 h after triggering recombinant gene expression.
b These values were determined in samples taken between 3 and 5 h after triggering recombinant gene expression.
c Specific activity or fluorescence emission of inclusion bodies relative to the values determined for the soluble counterpart fraction. Protein amounts were determined by Western blot analysis as described and enzymatic assays performed by conventional procedures. Excitation wavelengths were 450 nm for VP1GFP and 360 nm for Aβ42(F19D)-BFP.
Figure 1 FTIR spectra of inclusion bodies formed by either VP1LAC (black), hDHFR (green), VP1GFP (red) or Aβ42(F19D)-BFP (blue) in the amide I region [9]. The asterisk labels the peak indicative of extended inter-molecular β-sheet structures in bacterial inclusion bodies.
Figure 2 Optical micrographs of Aβ42(F19D)-BFP (top) and VP1GFP (bottom) inclusion bodies by phase contrast (left) and fluorescent microscopy (right).
We wondered if active inclusion bodies could be then used in suspension as efficient catalysts for bioprocesses. If so, the straightforward use of these particles, that in addition are easily removable from the reaction mixture once the reaction is completed by low speed centrifugation, would be a convenient alternative to in vitro protein refolding before use, a complex procedure for which efficiencies are highly variable but in general low [5]. The enzymatic activity of soluble and inclusion body versions of both VP1LAC and hDHFR was then monitored in reaction mixtures. As observed (Figure 3A and 3B), inclusion body-embedded enzymes performed very efficiently as catalysts of enzymatic reactions. Substrate hydrolysis mediated by the insoluble form of VP1LAC was significantly faster than that mediated by the same amount of the soluble version (Figure 3A), while substrate processing by hDHFR was slower when driven from inclusion bodies but still important (Figure 3B). These observations are nicely compatible with the specific activities displayed by both versions of these proteins (Table 1).
Figure 3 A) Product formed by soluble (black symbols) or inclusion body (red symbols) VP1LAC through ONPG hydrolysis as determined at 414 nm. Very coincident results have been obtained by using CPRG as alternative substrate (see the small panel), whose hydrolysis product was determined at 540 nm. B) Conversion of NADPH into NADP+ associated to tetrahydrofolate formation mediated by soluble (black symbols, left scale) and inclusion body (red symbols, right scale) hDHFR. Absorbance was determined at 340 nm.
Discussion
The quantitative similarity between protein activity in the soluble cell fraction and that of the aggregated forms of both enzymes and fluorescent proteins (Table 1) demonstrates that physiological aggregation as inclusion bodies does not necessarily split protein population into active and inactive fractions. Probably, protein solubility (observed as the occurrence in the soluble cell fraction) does not necessarily indicate the acquisition of a correctly folded and thus active structure. In this context, soluble micro-aggregates have been described [11] and recently characterized in detail [12]. The non complete coincidence between solubility and folding has been previously indicated by exhaustive mutational analysis of model proteins [13], showing that the genetic determinants of protein aggregation and misfolding are not coincident. In this way, natively unfolded proteins are unstructured but soluble [14]. Therefore, determinations of GFP-fusions solubility by using fluorescence as reporter [15] could have eventually been indicative of folding-misfolding extend rather than solubility-insolubility, since inclusion bodies formed by GFP fusions can be highly fluorescent (Figure 2). Furthermore, solubility does not appear to be an all-or-nothing attribute and polypeptide chains might exhibit a continuum of folding states in both soluble and insoluble cell fractions, between which they are dynamically transferred with the assistance of cellular folding modulators [16]. In this context, the occurrence and evolution of 'soluble' aggregates in bacteria (namely misfolded species occurring in the soluble cell fraction and presumably inactive) [12] could explain the variable specific activity observed in the soluble cell fraction of bacteria producing recombinant β-galactosidases [17].
Inversely, our results prove a major occurrence of native or native-like protein in inclusion bodies. In fact, deposition as inclusion bodies might even result in the enrichment of active species as suggested by the specific activity (166 % of that found in the soluble cell fraction; Table 1) and catalytic properties (Figure 3A) of VP1LAC inclusion bodies. This observation can be then again indirectly indicative of the presence of enzymatically inactive protein in the soluble cell fraction, since protein deposition is not expected to favour a correct folding.
Finally, although the existence of native-like structure in bacterial inclusion body proteins has been previously reported [18], here we demonstrate that this is not anecdotic but probably the architectonic nature of these kind of aggregates, as inclusion bodies formed by four structurally different proteins all display significantly high biological activity. Interestingly, the active and properly folded polypeptides in inclusion bodies coexist with a molecular β-sheet organization also manifest in all cases, although the extent of β-sheet structure and its coincidence with the biological activity of the aggregates cannot be quantitatively evaluated. Since is highly improbable that enzyme active sites involved in the intermolecular β-sheet structure could be themselves active, we suggest that enzymatic activity or fluorescence are supported by properly folded molecules or molecule segments. Aggregation, observed as protein deposition driven by intermolecular interactions between solvent-exposed hydrophobic patches [9] would not necessarily disturb the conformation of all protein domains, and the active site would be still functional if misfolded, aggregation-prone regions are located in a distant site of the polypeptide chain. Alternatively, properly folded and active molecules could coexist with β-sheet-enriched (inactive) versions of the same species, and both situations could in fact take place simultaneously in single aggregate units. Further structural and functional analysis would hopefully solve this issue.
From an applied point of view, inclusion bodies, being formed by sequence-specific interaction between homologous protein patches result in highly pure protein microparticles [9]. Since they are also porous and highly hydrated [19], efficient substrate diffusion would probably occur for most of the (or at least many) biotechnologically relevant aggregated enzymes, thus opening the possibility for a new industrial market of enzymatically active inclusion bodies.
Conclusion
Results presented here prove that aggregation of recombinant proteins as bacterial inclusion bodies does not necessarily inactivate them, despite the enriched intermolecular β-sheet structure observed in those formed by the tested model proteins. The extent of protein activity varies depending on the specific protein, but even the lowest functional values observed are still high enough to consider the use of inclusion body enzymes in bioprocesses, without any previous refolding step. The eventual incorporation of inclusion bodies in industrial catalysis could represent an important conceptual shift in the biotechnology market.
Methods
Strain, plasmids and culture conditions
E. coli MC4100 [20] was used for all the experiments. Plasmids encoding hDHFR and Aβ42(F19D)-BFP have been previously described and appropriate references can be found in Table 1. Briefly, in the Aβ42(F19D)-BFP vector (6.7 Kb) the DNA sequence encoding the 42-mer Alzheimer's amyloid peptide, (bearing a Phe19→Asp mutation to reduce its in vivo aggregation rate), is fused upstream of the BFP gene and under the control of the T7 promoter, in a pET-28 based vector. In the product, the two protein sequences were separated by 12-mer linker stretch to provide flexibility to the fusion protein and limit steric constraints between domains. pTVP1LAC was constructed by moving the SalI-NcoI VP1LAC fusion-encoding DNA segment (3.5 Kb) from pJVP1LAC (8.5 Kb) to the cloning vector pTRC99A [20]. The resulting pTVP1LAC construct (7.7 Kb) was used to direct the production of VP1LAC. The lacZ gene was further replaced there by an appropriate GFP-encoding DNA segment (0.7 Kb) through digestion with EcoRI and BamHI, rendering pTVP1GFP (5.5 Kb). All the production processes were performed in shaker-flask cultures growing at 37°C in LB rich medium [20] plus 100 μg/ml ampicillin for plasmid maintenance, and recombinant gene expression was induced when the OD550 reached 0.4, by adding 1 mM IPTG. Cell samples were taken at 3 and 5 h after induction of gene expression.
Analysis of enzymatic activity
Culture samples of 2.5 ml were jacketed in ice, disrupted by sonication for 5 min at 50 W under 0.5 s cycles [21] and centrifuged at 4°C for 15 min at 15000 g. The supernatant was directly used for the analysis as the soluble cell fraction. Inclusion bodies were purified by a detergent-washing protocol as described [19] and used in suspension for activity analysis. β-Galactosidase activity of both soluble cell fraction and inclusion bodies of VP1LAC was determined in microplates as described [7,22] under continuous stirring at 250 rpm. Kinetic analysis of VP1LAC enzymatic activity was monitored in 120 μl reaction mixtures with either 2 mM ONPG (pH 8.4) or 2 mM CPRG (pH 7.0). The hDHFR activity was determined by incubating 50 μl of the protein sample and 850 μl of the appropriate assay buffer (0.1 M K3PO4 pH 7.4, 1 mM DTT, 0.5 M KCl, 1 mM EDTA and 20 mM ascorbic acid) for 10 minutes at room temperature. Then, 50 μl of 2 mM 7,8-dihidrofolate and 50 μl of 2 mM NADPH were added and hDHFR activity was recorded every 15 seconds during 4 minutes at 340 nm. Protein concentration in all the assays was adjusted between 2 and 3 μg/ml.
Fluorescence (at 510 nm for GFP and 460 nm for BFP) was recorded in a Perkin-Elmer 650-40 fluorescence spectrophotometer by using excitation wavelengths of 450 nm and 360 nm for GPF and BFP respectively. Fluorescence was measured in 1 ml samples using dilutions when necessary. Both enzymatic activities and fluorescence were determined in triplicate.
Quantitative protein analysis
Samples of bacterial cultures (10 ml) were low-speed centrifuged (15 min at 12000 g) to harvest the cells. For protein quantification in soluble cell fractions, samples were resuspended in 400 μl of Z buffer without β-mercaptoethanol [23] with one tablet of protease inhibitor cocktail (Roche, ref. 1 836 170) per 10 ml buffer. Such mixtures, once jacketed in ice, were sonicated for 5 min (or longer when required to achieve a complete disruption) at 50 W under 0.5 s cycles as described [21], and centrifuged for 15 min at 12000 g. The supernatant was mixed with denaturing buffer at appropriate ratios [24]. For the determination of inclusion body protein, these structures were purified by repeated detergent washing as described [19] and resuspended in denaturing buffer [24]. After boiling for 20 min, appropriate sample volumes were loaded onto denaturing gels. For Western blot, polyclonal antibodies specific for each protein were used as previously described [17]. Dried blots were scanned at high resolution and bands quantified by using the Quantity One software from Bio Rad, by using appropriate protein dilutions of known concentration as controls. Determinations were always done within the linear range and they were used to calculate the specific activity values.
Conformational analysis by FTIR spectroscopy
For FTIR spectroscopy analysis, purified inclusion bodies were dried for two hours in a Seepd-Vac system before analysis to reduce water interference in the infrared spectra. The FTIR spectrum of the dry samples was analysed directly in a Bruker Tensor FTIR spectrometer. All processing procedures were carried out so as to optimise the quality of the spectrum in the amide I region, between 1600 cm-1 and 1700 cm-1. Second derivatives of the amide I band spectra were used to determine the frequencies at which the different spectral components were located. A general description of FTIR procedures can be found elsewhere [9,10].
Abbreviations
BFP blue fluorescent protein
CPRG phenol red β-D-galactopyranoside
FMDV foot-and-mouth disease virus
FTIR fourier transform infrared
GFP green fluorescent protein
HDHFR human dihydropholate reductase
IPTG isopropyl-β-D-thiogalactopyranoside
ONPG ortho-nitrophenyl β-D-galactopyranoside
Authors' contributions
EGF performed most of the experiments and prepared the final data and figures. NGM, A. Vera and AA analysed protein amounts by Western blot, RMF performed enzyme kinetics, MM performed part of optical microscopy analysis and SV part of FTIR analysis and data interpretation. A. Villaverde directed the work and prepared the manuscript.
Acknowledgements
This work has been supported by BIO2004-00700 from MEC, Spain and 2002SGR-0099 (AGAUR). EGF is recipient of a doctoral fellowship from MEC, Spain, and SV is supported by a "Ramón y Cajal" project awarded by the MCYT and co-financed by the Universitat Autònoma de Barcelona.
==== Refs
Baneyx F Mujacic M Recombinant protein folding and misfolding in Escherichia coli Nat Biotechnol 2004 22 1399 1408 15529165 10.1038/nbt1029
Strandberg L Enfors SO Factors influencing inclusion body formation in the production of a fused protein in Escherichia coli Appl Environ Microbiol 1991 57 1669 1674 1908208
Rinas U Tsai LB Lyons D Fox GM Stearns G Fieschko J Fenton D Bailey JE Cysteine to serine substitutions in basic fibroblast growth factor: effect on inclusion body formation and proteolytic susceptibility during in vitro refolding Biotechnology (N Y ) 1992 10 435 440 1368488 10.1038/nbt0492-435
Baneyx F Palumbo JL Improving heterologous protein folding via molecular chaperone and foldase co-expression Methods Mol Biol 2003 205 171 197 12491886
Vallejo LF Rinas U Strategies for the recovery of active proteins through refolding of bacterial inclusion body proteins Microb Cell Fact 2004 3 11 15345063 10.1186/1475-2859-3-11
Tokatlidis K Dhurjati P Millet J Beguin P Aubert JP High activity of inclusion bodies formed in Escherichia coli overproducing Clostridium thermocellum endoglucanase D FEBS Lett 1991 282 205 208 2026260 10.1016/0014-5793(91)80478-L
Garcia-Fruitos E Carrio MM Aris A Villaverde A Folding of a misfolding-prone beta-galactosidase in absence of DnaK Biotechnol Bioeng 2005 90 869 875 15849697 10.1002/bit.20496
Worrall DM Goss NH The formation of biologically active beta-galactosidase inclusion bodies in Escherichia coli Aust J Biotechnol 1989 3 28 32 2485330
Carrio M Gonzalez-Montalban N Vera A Villaverde A Ventura S Amyloid-like properties of bacterial inclusion bodies J Mol Biol 2005 347 1025 1037 15784261 10.1016/j.jmb.2005.02.030
Ami D Natalello A Gatti-Lafranconi P Lotti M Doglia SM Kinetics of inclusion body formation studied in intact cells by FT-IR spectroscopy FEBS Lett 2005 579 3433 3436 15949804 10.1016/j.febslet.2005.04.085
Sorensen HP Mortensen KK Soluble expression of recombinant proteins in the cytoplasm of Escherichia coli Microb Cell Fact 2005 4 1 15629064 10.1186/1475-2859-4-1
de Marco A Schroedel A Characterization of the aggregates formed during recombinant protein expression in bacteria BMC Biochem 2005 6 10 15927061 10.1186/1471-2091-6-10
Chiti F Taddei N Baroni F Capanni C Stefani M Ramponi G Dobson CM Kinetic partitioning of protein folding and aggregation Nat Struct Biol 2002 9 137 143 11799398 10.1038/nsb752
Wright PE Dyson HJ Intrinsically unstructured proteins: re-assessing the protein structure-function paradigm J Mol Biol 1999 293 321 331 10550212 10.1006/jmbi.1999.3110
Waldo GS Standish BM Berendzen J Terwilliger TC Rapid protein-folding assay using green fluorescent protein Nat Biotechnol 1999 17 691 695 10404163 10.1038/10904
Carrio MM Villaverde A Protein aggregation as bacterial inclusion bodies is reversible FEBS Lett 2001 489 29 33 11231008 10.1016/S0014-5793(01)02073-7
Cazorla D Feliu JX Villaverde A Variable specific activity of Escherichia coli beta-galactosidase in bacterial cells Biotechnol Bioeng 2001 72 255 260 11135194 10.1002/1097-0290(20010205)72:3<255::AID-BIT1>3.0.CO;2-#
Oberg K Chrunyk BA Wetzel R Fink AL Nativelike secondary structure in interleukin-1 beta inclusion bodies by attenuated total reflectance FTIR Biochemistry 1994 33 2628 2634 8117725 10.1021/bi00175a035
Carrio MM Cubarsi R Villaverde A Fine architecture of bacterial inclusion bodies FEBS Lett 2000 471 7 11 10760503 10.1016/S0014-5793(00)01357-0
Sambrook J Fritsch E Maniatis T Molecular Cloning, A Laboratory Manual, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY 1989
Feliu JX Cubarsi R Villaverde A Optimized release of recombinant proteins by ultrasonication of E. coli cells Biotechnol Bioeng 1998 58 536 540 10099290 10.1002/(SICI)1097-0290(19980605)58:5<536::AID-BIT10>3.0.CO;2-9
Ferraz RM Aris A Villaverde A Profiling the allosteric response of an engineered beta-galactosidase to its effector, anti-HIV antibody Biochem Biophys Res Commun 2004 314 854 860 14741715 10.1016/j.bbrc.2003.12.169
Miller JH Experiments in Molecular Genetics Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY 1972
Laemmli UK Cleavage of structural proteins during the assembly of the head of bacteriophage T4 Nature 1970 227 680 685 5432063 10.1038/227680a0
Davies JF Delcamp TJ Prendergast NJ Ashford VA Freisheim JH Kraut J Crystal structures of recombinant human dihydrofolate reductase complexed with folate and 5-deazafolate Biochemistry 1990 29 9467 9479 2248959 10.1021/bi00492a021
Sánchez de Groot N Avilés FX Vendrell J Ventura S Mutagenesis of the central hydrophobic cluster in Ab42 Alzheimer's peptide.
Simple rules to predict the aggregation propensities of polypeptides. submitted 2005
|
16156893
|
PMC1224866
|
CC BY
|
2021-01-04 16:26:55
|
no
|
Microb Cell Fact. 2005 Sep 12; 4:27
|
utf-8
|
Microb Cell Fact
| 2,005 |
10.1186/1475-2859-4-27
|
oa_comm
|
==== Front
Mol CancerMolecular Cancer1476-4598BioMed Central London 1476-4598-4-331615330210.1186/1476-4598-4-33ResearchTissue transglutaminase-induced alterations in extracellular matrix inhibit tumor invasion Mangala Lingegowda S [email protected] Banu [email protected] Aysegul A [email protected] Kapil [email protected] Department of Experimental Therapeutics, The University of Texas M. D. Anderson Cancer Center, Houston, Texas, USA2 Department of Breast Medical Oncology, The University of Texas M. D. Anderson Cancer Center, Houston, Texas, USA3 Department of Pathology, The University of Texas M. D. Anderson Cancer Center, Houston, Texas, USA4 Cancer Biology Program, Graduate School of Biomedical Sciences, The University of Texas, Houston, Texas, USA2005 9 9 2005 4 33 33 13 7 2005 9 9 2005 Copyright © 2005 Mangala et al; licensee BioMed Central Ltd.2005Mangala et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Alterations in the extracellular matrix (ECM) can affect host-tumor interactions and tumor growth and metastasis. Tissue transglutaminase (TG2, EC 2.3.2.13), a calcium-dependent enzyme that catalyzes covalent cross-linking of proteins, can render the ECM highly stable and resistant to proteolytic degradation. So we determined whether TG2 expression in a tumor or nontumor (stroma) environment could affect the process of metastasis. Two hundred archived samples from patients with breast cancer were studied for the TG2 expression. Also, in an in vitro model the invasive behavior of MDA-MB-231 cells in the presence or absence of exogenous TG2 was determined.
Results
Tumors associated with negative nodes showed significantly higher expression of TG2 in the stroma (P < 0.001). TG2 in the stroma was catalytically active, as revealed by the presence of isopeptide cross-links. Pretreatment of Matrigel with catalytically active TG2 resulted in strong inhibition of invasion of MDA-MB-231 cells through the Matrigel Transwell filters.
Conclusion
TG2-induced alterations in the ECM could effectively inhibit the process of metastasis. Therefore, selective induction of catalytically active TG2 at the site of tumor may offer promising approach for limiting the metastasis.
==== Body
Background
Despite significant advances in the treatment of primary breast cancer, predicting and preventing metastasis remains a daunting clinical challenge. To make progress in this area, it is imperative to understand the molecular mechanisms that regulate the progression from a primary tumor to metastatic disease.
Metastasis is a multistep process that involves intravasation, adhesion to a blood vessel wall, extravasation, infiltration, and the proliferation of cancer cells in the target tissue [1]. Many of these steps require interaction between tumor cells and the extracellular matrix (ECM). For example, the ECM can modulate tumor cell growth by binding to and storing cytokines, it can promote cell attachment and migration by providing a stable foundation, and it can support cell growth and survival by interacting with cell-surface receptors and activating appropriate signaling pathways [2,3].
Several lines of evidence have suggested that tissue transglutaminase (TG2, EC 2.3.2.13) plays an important role in stabilizing the ECM by cross-linking its component proteins and rendering it resistant to mechanical and proteolytic degradation [4-7]. TG2, a member of the Ca2+-dependent family of mammalian enzymes, catalyzes irreversible cross-linking of proteins by inserting highly stable ε(γ-glutamyl)lysine bonds between them [5,8,9]. Several ECM proteins, such as fibronectin, vitronectin, collagen, fibrin, laminin, osteonectin, and osteopontin, can serve as substrates in TG2-catalyzed cross-linking reactions [4,10-12]. Moreover, in various fibrotic disorders, such as pulmonary fibrosis, renal fibrosis, and atherosclerosis increased expression of TG2 has been observed, and its ability to cross-link ECM proteins has been implicated in facilitating the deposition of a new ECM and making it resistant to metalloproteinases [12-16]. In addition to its direct role in promoting the accumulation of the ECM, TG2 has been implicated in the storage and activation of transforming growth factor-beta (TGF-β) [17], a proinflammatory cytokine that is involved in the synthesis of various ECM proteins and inhibitors of metalloproteinases [18,19]. The ability of TG2 to affect the physicochemical properties of the ECM may influence the invasive properties of tumor cells by modulating cell-matrix interactions or by facilitating the assembly of the matrix and tissue remodeling.
In view of these facts and other observations that modification of the ECM can affect the growth of both normal and cancerous mammary epithelial cells and the processes of angiogenesis and tumor metastasis [20-22], we speculated that TG2 expression in the stroma of the host can affect breast cancer progression. To test this theory, we searched for such a correlation in tumor and stroma specimens in a total 200 samples from patients with early-stage breast cancer. Our findings suggested that TG2 expression in the stroma was associated with an absence of lymph node metastasis in patients with breast cancer. The results of our in vitro study further supported this link and suggested that TG2-mediated modification of the ECM could render it less susceptible to invasion by tumor cells. Taken together, these findings suggest that TG2 is a good candidate for therapeutic use to prevent progression from a primary tumor to metastatic disease in patients with breast cancer.
Results
Of the 200 samples studied, only 189 were evaluable (Table 1). Patients without lymph node metastasis (n = 95) were followed for a median of 4 years after diagnosis. Two of these patients experienced disease recurrence, and 4 died. Patients with lymph node metastasis (n = 94) were followed for a median of 3.2 years after diagnosis. Ten of these patients experienced disease recurrence, and 2 died.
Table 1 TG2 expression in tumor and stroma tissues of patients with breast cancer
TG2 in tumor TG2 in stroma
Total 0 1+ 2+ 3+ p* 0 1+ 2+ 3+ P*
Tumor samples 189 84 37 33 35 85 52 28 24
Node status
Negative 95 46(48) 17(18) 17(18) 15(16) 24(25) 33(35) 20(21) 18(19)
Positive 94 38(41) 20(21) 16(17) 20(21) 61(65) 19(20) 8(9) 6(6)
0.28 <0.001
Age (yrs)
<50 53 21(39) 11(21) 11(21) 10(19) 22(41) 11(21) 21(21) 9(17)
>=50 136 63(46) 26(19) 22(16) 25(19) 63(46) 41(30) 17(13) 15(11)
0.51 0.14
* Cochran-Armitage trend test
The numbers in parenthesis indicate the percentages of total number of patients in that row.
A typical pattern of TG2 expression in mammary tumor samples is shown in Fig. 1. Tumor samples showed either high TG2 expression that was predominantly restricted to the stroma surrounding the tumor (Fig. 1A &1B) or in some cases converse was true; tumors expressed high levels of TG2, but the stroma surrounding the tumor showed little or no TG2 expression (Fig. 1C and 1D). In few patient samples (approximately 12%) high TG2 expression was observed in both the tumor and the surrounding stroma (Fig. 1E and 1F). Similarly, some samples expressed little (≤1+) or no TG2 in the tumor or the stroma (Fig. 1G and 1H).
Figure 1 TG2 expression in tumor tissue samples from patients with breast cancer. Paraffin-embedded tissues from the primary tumors were retrieved from the Tumor Tissue Bank in the Department of Pathology at The University of Texas M. D. Anderson Cancer Center. The tissue sections were processed for immunostaining as described in Materials and Methods. One hundred samples each from patients with or without lymph node metastasis were studied and independently scored by a pathologist and a laboratory technician for TG2 expression. The figure shows the extent of TG2 expression in the primary tumor cells and the stroma surrounding the tumor from eight representative patients. a and b, tumor samples from representative patients (14% of patients) in which TG2 expression was mainly localized in the stroma; c and d, tumor samples from representative patients (23% of patients) in which TG2 expression was mainly localized in the tumor; e and f, TG2 expression in some patients' tumor samples (13% of patients) was observed both in the tumor and the stroma; g and h, half of all patients (50%) showed almost a complete lack of TG2 in the tumor and the stroma. The TG2-positive cells in g and h were endothelial cells that had high constitutive TG2 expression.
The association of TG2 expression in tumors and stroma with node-negative or -positive status is presented in Table 1. There was no evidence that TG2 expression in tumor cells differed between node-negative or node-positive samples. Nevertheless, there was strong evidence that tumors associated with negative nodes had a significantly higher TG2 expression in the stroma (P < 0.001). With regard to the association of TG2 expression in tumors and stroma with age, there was no evidence of a difference in younger (<50 yrs) or older (>50 yrs) patient populations.
To further delineate the importance of TG2 expression in the stroma and its possible implications in the modulation of metastatic disease, we used an in vitro Matrigel Transwell invasion assay with a human breast cancer cell line MDA-MB-231. Under our experimental conditions, MDA-MB-231 cells were highly invasive, as determined by the number of cells that invaded through the Matrigel Transwell filters (Fig. 2A). However, pretreatment of the Matrigel with catalytically active TG2 (in the presence of Ca2+) strongly inhibited the invasion of MDA-MB-231 cells through the Matrigel Transwell filters. Under identical conditions when the enzyme was rendered inactive by eliminating Ca2+ from the reaction mixture, the number of cells that invaded was not much different from that seen in the control wells (coated with Matrigel that had been preincubated with 5-mM Ca2+ but without TG2). TG2-mediated inhibition of tumor cell invasion was dose-dependent and could be observed by pretreating the Matrigel contents with as little as 1 μg of TG2. At 6 μg, TG2 inhibited the invasion of MDA-MB-231 cells by more than 80% (Fig. 2B). The inhibition of cell invasion through TG2-pretreated Matrigels was most likely due to the crosslinking of component proteins into high molecular weight scaffolds (Fig. 2C). The TG2-induced crosslinks are known to exhibit resistance to proteases and thus may pose a barrier against protease-induced invasion of cancer cells. Indeed, MDA-MB-231 cells produce active proteases, such as MMP-1 and MMP-2 as revealed by zymography and RT-PCR analysis (Fig. 2D). TG2 by itself, either in the presence or in absence of Ca2+, did not exert any noticeable effect on cell viability or growth of MDA-MB-231 cells (data not shown). These results suggest that active cross-linking of the Matrigel proteins by TG2 is effective in preventing the invasion of tumor cells.
Figure 2 TG2 inhibits the invasion of MDA-MD-231 cells through a Matrigel-coated Transwell membrane. A, Matrigel contents were incubated with buffer alone (control) or buffer containing the purified guinea pig liver TG2 protein (6 μg/0.75 ml) in the presence (TG2 + Ca2+) or absence (TG2 alone) of 5 mM Ca2+ at 37°C for 15 minutes before being coated onto the Transwell membranes. The MDA-MB-231 cells were compared for their ability to invade through the TG2-pretreated or untreated Matrigel-Transwell membranes. Representative fields with cells that migrated under the membrane were photographed. B, Matrigels (0.75 ml) containing increasing amounts of the purified TG2 protein were preincubated (37°C, 15 minutes) with 5 mM Ca2+. Another tube containing Matrigel plus 6 μg of TG2 protein was incubated in parallel, but without Ca2+, to serve as control. Matrigel contents pretreated with various amounts of TG2 were layered over the Transwell membranes and compared for their ability to support the invasion of MDA-MB-231 cells. Ten fields were counted randomly under the microscope for the number of cells that had migrated through the Matrigel and were plotted as an average number of cells ± SD per field. C, Equal volumes of Matrigel (0.75 ml) were incubated under conditions described in legend B with buffer alone or buffer containing varying concentrations of purified TG2 in the presence (+) or absence (-) of 5 mM Ca2+. Forty μl reactants, each were fractionated on SDS-PAGE. The gel was stained and viewed for TG2-induced changes in protein bands after destaining in methanol/acetic acid solution. The thin arrow indicates a prominent band that disappears in the presence of an enzymatically active TG2; whereas thick arrow indicates the appearance of a prominent band in the presence of active TG2. D, Basal levels of MMP-1 and MMP-2 in MDA-MB-231 cells as determined by zymogram performed on the supernatants collected from cultured cells (80% confluent in 6-well plate, incubated in 1 ml serum-free medium overnight) or by RT-PCR, using MMP-1 specific primers, as described in Materials and Methods.
Next, we sought to determine whether TG2 in the stroma of node-negative patients with breast cancer was catalytically active. To accomplish this, we studied the presence of ε(γ-glutamyl)lysine isopeptide, the product of TG2-catalyzed protein cross-linking reactions, in a few selected samples. The results (Fig. 3) confirmed strong positivity of the stroma for TG2 expression in selected node-negative patient samples (red fluorescence). Importantly, the expression of TG2 in stroma was closely associated with the presence of isopeptide bonds, as revealed by strong immunofluorescence staining with the anti-isopeptide antibody (green fluorescence). The expression of TG2 in the stroma, in general, closely resembled the isopeptide staining (Fig. 3), suggesting that the two antigens closely associate in the stroma surrounding the tumor. These results suggest that, in some patients, TG2 can localize extracellularly and effectively cross-link ECM proteins, thereby rendering the ECM resistant to invasion by tumor cells.
Figure 3 TG2 expression in the stroma is closely associated with the formation of isopeptide bonds. Tissue sections from 3 selected patients with breast cancer that showed high TG2 expression in the stroma were incubated simultaneously with rabbit anti-TG2 antibody and mouse anti-isopeptide monoclonal antibody. Anti-mouse IgG Alexa 488 and goat anti-rabbit IgG Alexa 546 were used as secondary antibodies to localize the presence of TG2 (A, C, E) and isopeptide (B, D, F), respectively. In a control experiment (G, H), a tissue sample lacking TG2 expression in the tumor and the stroma was used in a similar way (two primary and 2 secondary antibodies) to determine the specificity of the antigen-antibody reactions.
Discussion
Major findings to emerge from this study are: in some patients with breast cancer, tumor and nontumor (stroma) environments express increased levels of TG2; in the stroma, TG2 protein exists in a catalytically active configuration, leading to cross-linking of the ECM proteins; and TG2-catalyzed cross-linking of ECM proteins may mitigate the migration of cancer cells to distant sites.
TG2 is a multifunctional protein that can affect the ECM and its interaction with cells. Both the extracellular cross-linking and intracellular signaling activities of TG2 can considerably modulate cell-matrix interactions [7]. We previously observed that TG2 expression was up-regulated in drug-resistant and metastatic breast cancer cells and cell lines [23-28]. The increased expression of TG2 in drug-resistant and metastatic breast cancer cells was linked to their increased resistance to apoptosis, owing to the fact that TG2 in these cells closely associates with β integrins and thus may promote cell survival signaling [28-30]. The primary objective of this study was to determine whether higher levels of TG2 in breast cancer cells were associated with node involvement and development of aggressive tumors. We found no evidence that TG2 tumor staining differed between tumor samples associated with negative or positive node status. These results are in agreement with those of a previous study that revealed a negative correlation between TG2 transcript expression and nodal status in patients with breast cancer [31]. However, in another report, Jiang et al. observed that lymph node involvement and development of aggressive phenotype in breast cancer tumors was associated with increased expression of TG2 transcript and decreased levels of TG3 and TG7 transcript expression [32]. Similarly, another group has reported a marked increase in TG2 expression in the intraductal and invasive human breast cancer cells [33]. Taken together, these observations suggest that transglutaminases' repertoire is altered during breast cancer development and progression. Therefore, further studies are warranted with larger cohort of patients to determine the prognostic significance of transglutaminases in the progression and development of metastatic disease in patients with breast cancer.
More important, the results of our study suggest that TG2 expression in the stroma is strongly associated with node-negative status in patients with breast cancer. A similar increase in TG2 expression in the stroma surrounding the tumor was observed by Haroon et al. in subcutaneously implanted rat mammary tumors [34]. Treatment of these tumors with enzymatically active TG2 significantly delayed the tumor growth when compared with the tumors treated with catalytically inactive TG2 mutant. On the basis of these observations the authors concluded that TG2 might constitute a distinct part of the host response against growing tumor and by crosslinking the component proteins, it may stabilize the ECM and affect the tumor growth [34].
It is possible that local injury caused by the growing tumor may elicit host's response and induce cytokines production that in turn may promote wound healing and restrict the invasion of cells by producing new or stable ECM [35]. Indeed, TG2 is capable of cross-linking several constituent proteins in the ECM, which can render the ECM more resistant to proteases and mechanical disruptions [4-6]. TG2 can also enhance stability and strengthen the ECM by its ability to facilitate the activation of tumor growth factor-beta [17]. Thus, TG2-mediated alterations may render the ECM more amenable to tumor growth and resistant to proteases [36-38]. Indeed, our in vitro data shown in figure 2 clearly supported this contention and suggested that TG2-mediated crosslinking of component proteins rendered the matrigel resistant to invasion by MDA-MB-231 cells regardless of the fact that these cells actively produced matrix metalloproteinases (MMP) such as MMP-1 and MMP-2.
In conclusion, our findings suggest that increased expression of TG2 in the stroma represents a part of the host response to a growing tumor in an attempt to restrict tumor growth and prevent it from spreading to distant sites. However, the ability of tumor cells to produce proteases and other factors that could render TG2 inactive may overwhelm the ability of TG2 to inhibit growth and prevent metastasis. Indeed, TG2 has been shown to be a good substrate for MMP-2-like proteases that are abundantly produced by metastatic tumors [39,40]. Future studies to selectively enhance the production of active TG2 at the site of tumor growth may offer promising approaches to limiting tumor growth and metastasis.
Methods
Cell Lines and Tumor Tissues
Human breast cancer cell line MDA-MB-231 was purchased from American Type Culture Collection (Rockville, MD) and cultured by standard methods. Tumor samples from 200 patients with early-stage breast cancer diagnosed between 1979 and 2002 were obtained from the archives of the Pathology Department of The University of Texas M. D. Anderson Cancer Center after approval by the Institutional Review Board for the Welfare of Human Subjects. All the selected patients had tumors ranging from 1 cm to 2 cm in largest diameter. Samples were selected for inclusion to yield 100 tumor samples that were associated with lymph node invasion and 100 tumor samples that were not associated with lymph node invasion.
Invasion Assay
The invasive behavior of MDA-MB-231 cells in the presence or absence of exogenous TG2 was determined in vitro by counting how many cells invaded through Matrigel-coated Transwell polycarbonate membrane inserts, as described previously [41,42]. Briefly, Transwell inserts with a pore size of 12 μm were coated with 0.78 mg/ml Matrigel in cold, serum-free medium. In some cases, the Matrigel was preincubated (37°C for 15 minutes) with varying concentrations of purified TG2 protein (Sigma-Aldrich Chemical Co., St. Louis, MO) in the presence or absence of 5 mM Ca2+ before being coated onto the inserts. Also, Matrigel samples (40 μl aliquots) pretreated with varying concentrations of TG2 in the presence or absence of Ca2+ were subjected to 7% SDS-PAGE. The gel was stained with Coomassie brilliant blue and destained in a acetic acid and methanol solution before visualizing the changes in protein bands as a consequence of TG2-mediated crosslinking.
Cells were recovered by trypsinization and washed once with serum-free medium. The cell pellets were resuspended in serum-free medium, and 0.5 ml of the cell suspension (0.5 × 106 cells) was added to duplicate wells. After incubation for 48 hours, the cells that passed through the filter were stained using a Hema-3 stain kit (Fisher Scientific, Houston, TX), and the cells in 10 random fields were counted under a microscope.
Matrix Metalloproteinases (MMP)
Basal expression of MMP-1 and MMP-2 in MDA-MB-231 cells was determined by gelatin-substrate gel zymography as described [41]. Briefly, the supernatant from MDA-MB-231 cultured cells was subjected under nonreducing conditions to gelatin impregnated (0.1%; w/v) SDS-PAGE. The gel was washed several times to remove the SDS and incubated in buffer containing 5 mM CaCl2 plus 1 μM ZnCl2 for 24–48 h at 37°C. The gel was stained with Coomassie brilliant blue and destained. Proteolytic activity was visualized as clear bands (zones of gelatin degradation) against the blue background of stained gelatin.
For RT-PCR, total RNA was isolated from MDA-MB-231 cells in Trizol reagent (Invitrogen, Life technologies, CA) and cDNA was synthesized from 5 μg of total RNA using Superscript reverse transcriptase (Life Technologies, Inc) as per the manufacturer's instructions. cDNA was subjected to PCR (GeneAmp PCR System 9700, Applied Biosystems) using MMP-1 specific, 5'-CGACTCTAGAAACACAAGAGCAAGA-3' (sense) and 5'-AAGGTTAGCTTACTGTCACACGCTT-3'(antisense) primers. PCR consisted of 30 cycles was carried out at the following conditions: 95°C for 5 min, 30 cycles of 95°C for 30 sec, annealing at 58°C for 1 min and elongation at 72°C for 2 min. PCR products were analyzed on 1% agarose gel and visualized under UV light after staining with ethidium bromide.
Immunohistochemistry
Sections of formalin-fixed, paraffin-embedded tumor samples (5 μm thick) were heated to 60°C and dehydrated in xylene and graded alcohols. Antigen retrieval was performed with 0.01 M citrate buffer at pH 6.0 for 20 minutes in a 95°C steam bath. Slides were allowed to cool for 20 minutes at room temperature, followed by repeated rinsing with 0.1 M phosphate-buffered saline (PBS, pH 7.4) containing 0.1% Tween 20 (PBS-T). Endogenous peroxidase activity was quenched with 3% hydrogen peroxidase. Each incubation step was conducted at room temperature and was followed by three sequential washes (5–10 minutes each) in PBS-T. Sections were incubated with anti-TG2 monoclonal antibodies, CUB-7401 (Neomarkers, Fremont, CA) overnight at 4°C, followed by incubation for 30 minutes each with biotinylated secondary antibody and peroxidase-labeled streptavidin. Antigen-antibody reactions were detected by exposure to 3,3'diaminobenzidine and hydrogen peroxide chromogen substrate (Vector Labs, Burlingame, CA) for 3–5 minutes. Slides were counterstained with hematoxylin and mounted. The negative controls were incubated with nonimmune mouse immunoglobulin g (IgG) in place of the primary antibody. The immunostained slides were examined under the light microscope and scored (with values of 0, 1+, 2+, or 3+) independently by two researchers in the laboratory.
Confocal Microscopy
Paraffin-embedded tissue sections from selected tumor samples were deparaffinized, washed 3 times with PBS, and blocked with 5% normal goat serum in PBS for 1 hour. The sections were immunostained by an indirect method using primary rabbit anti-TG2 (Neomarkers, Fremont, CA) and anti-isopeptide monoclonal Ab422 (Abcam, Cambridge, MA) antibodies. The anti-mouse IgG Alexa 488 and goat anti-rabbit IgG Alexa 546 (both from Molecular Probes, Eugene, OR) were used as the secondary antibodies. The immunostained sections were mounted in 80% glycerol and 20% PBS and viewed under the Zeiss Laser Scanning Microscope 510 (Carl Zeiss MicroImaging, Inc., Thornwood, NY) for taking the digital images. Appropriate controls, including mouse and rabbit IgG in place of the primary antibodies and either primary antibody alone along with both the secondary antibodies were included to determine the specificity of the reaction.
Abbreviations
BCL2, B-cell leukemia/lymphoma 2; ECM, extracellular matrix; EGFR, epidermal growth factor receptor; FAK, focal adhesion kinase; Fn, fibronectin; MDR, multidrug resistance; TG2, tissue transglutaminase; TGF-β, transforming growth factor-beta.
Authors' contributions
SLM, conducted the experiments. BA, identified and retrieved the patients samples. AS, evaluated the slides for TG2 expression, KM. Participated in the study design, data interpretation, and manuscript preparation.
Acknowledgements
The authors wish to thank Michael Gilogli and Jansina Y Fok for technical assistance and Mr. David Galloway for editorial help.
==== Refs
Friedl P Wolf K Tumour-cell invasion and migration: diversity and escape mechanisms Nature Rev Cancer 2003 3 362 74 12724734 10.1038/nrc1075
Aznavoorian S Murphy AN Stetler-Stevenson WG Liotta LA Molecular aspects of tumor cell invasion and metastasis Cancer 1993 71 1368 83 8435813
Stetler-Stevenson WG Aznavoorian S Liotta LA Tumor cell interactions with the extracellular matrix during invasion and metastasis Ann Rev Cell Biol 1993 9 541 73 8280471
Aeschlimann D Thomazy V Protein crosslinking in assembly and remodelling of extracellular matrices: the role of transglutaminases Connect Tissue Res 2000 41 1 27 10826705
Fesus L Piacentini M Transglutaminase 2: an enigmatic enzyme with diverse functions Trends Biochem Sci 2002 27 534 9 12368090 10.1016/S0968-0004(02)02182-5
Prince CW Dickie D Krumdieck CL Osteopontin, a substrate for transglutaminase and factor XIII activity Biochem Biophys Res Commun 1991 177 1205 10 1676261 10.1016/0006-291X(91)90669-X
Stephens P Grenard P Aeschlimann P Langley M Blain E Errington R Kipling D Thomas D Aeschlimann D Crosslinking and G-protein functions of transglutaminase 2 contribute differentially fibroblast wound healing responses Journal of Cell Sci 2004 117 3389 3403 15199098 10.1242/jcs.01188
Chen JS Mehta K Tissue transglutaminase: an enzyme with a split personality Int J Biochem Cell Biol 1999 31 817 36 10481269 10.1016/S1357-2725(99)00045-X
Lorand L Graham RM Transglutaminases: crosslinking enzymes with pleiotropic functions Nature Rev Mol Cell Biol 2003 4 140 56 12563291 10.1038/nrm1014
Aeschlimann D Paulsson M Cross-linking of laminin-nidogen complexes by tissue transglutaminase. A novel mechanism for basement membrane stabilization J Biol Chem 1991 266 15308 17 1678389
Kaartinen MT Pirhonen A Linnala-Kankkunen A Maenpaa PH Transglutaminase-catalyzed cross-linking of osteopontin is inhibited by osteocalcin J Biol Chem 1997 272 22736 41 9278432 10.1074/jbc.272.36.22736
Kleman JP Kleman JP Aeschlimann D Paulsson M Vander Rest M Transglutaminase-catalyzed cross-linking of fibrils of collagen V/XI in A204 rhabdomyosarcoma cells Biochemistry 1995 34 13768 75 7577969 10.1021/bi00042a007
Griffin M Smith LL Wynne J Changes in transglutaminase activity in an experimental model of pulmonary fibrosis induce by paraquat Br J Exp Pathol 1979 60 653 61 44202
Johnson TS Griffin M Thomas GL Skill J Cox A Yang B Nicholas B Birckbichler PJ Muchaneta-Kubara C El Nahas AM The role of transglutaminase in the rat subtotal nephrectomy model of renal fibrosis J Clin Invest 1997 99 2950 60 9185519
Gross SR Balklava Z Griffin M Importance of tissue transglutaminase in repair of extracellular matrices and cell death of dermal fibroblasts after exposure to a solarium ultraviolet A source J Invest Dermatol 2003 121 412 23 12880435 10.1046/j.1523-1747.2003.12353.x
Johnson TS kill NJ El Nahas AM Oldroyd SD Thomas GL Douthwaite JA Haylor JL Griffin M Transglutaminase transcription and antigen translocation in experimental renal scarring J Am Soc Nephrol 1999 10 2146 57 10505691
Nunes I Gleizes PE Metz CN Rifkin DB Latent transforming growth factor-beta binding protein domains involved in activation and transglutaminase-dependent cross-linking of latent transforming growth factor-beta J Cell Biol 1997 136 1151 63 9060478 10.1083/jcb.136.5.1151
Clark DA Coker R Transforming growth factor-beta (TGF-beta) Int J Biochem Cell Biol 1998 30 293 8 9611771 10.1016/S1357-2725(97)00128-3
Taipale J Saharinen J Keski-Oja J Extracellular matrix-associated transforming growth factor-beta: role in cancer cell growth and invasion Adv Cancer Res 1998 75 87 134 9709808
Stracke ML Murata J Aznavoorian S Liotta LA The role of the extracellular matrix in tumor cell metastasis In Vivo 1994 8 49 58 8054511
Wang RN Zhu YB Xue JY The relation between integrin, type IV collagenase and extracellular matrix in invasion and metastasis of gastric carcinoma Chinese J Pathol 1994 23 278 81
Itano N Yamada Y Yoshida M Kimata K Cancer metastasis and extracellular matrix Japn J Cancer Chemother 1999 26 1663 8
Chen J Kanopleva M Multani A Pathak S Mehta K Drug resistant breast cancer MCF-7 cells are paradoxically sensitive to apoptosis J Cell Physiol 2004 200 223 34 15174092 10.1002/jcp.20014
Chen JS Agarwal N Mehta K Multidrug-resistant MCF-7 breast cancer cells contain deficient intracellular calcium pools Breast Cancer Res Treat 2002 71 237 47 12002342 10.1023/A:1014461832403
Devarajan E Chen J Multani AS Pathak S Sahin AA Mehta K Human breast cancer MCF-7 cell line contains inherently drug-resistant subclones with distinct genotypic and phenotypic features Int J Oncol 2002 20 913 20 11956583
Mehta K Devarajan E Chen J Multani A Pathak S Multidrug-resistant MCF-7 cells: an identity crisis? J Natl Cancer Inst 2002 94 1652 4 12419794
Mehta K High levels of transglutaminase expression in doxorubicin-resistant human breast carcinoma cells Int J Cancer 1994 58 400 6 7914183
Mangala LS Mehta K Mehta K, Eckert R Tissue transglutaminase in cancer biology. Transglutaminases; the family of enzymes with diverse functions Prog Exp Tumor Res 2005 38 Basel, Karger 125 138 15746533
Herman HF Mehta K Implications of increased tissue transglutaminase expression in drug-resistant breast cancer (MCF-7) cells Mol Cancer Res 2005
Mehta K Fok JY Mangala SL Tissue transglutaminase (TG2): from biological glue to signaling cues Front Biosci 2005
Grigoriev MY Suspitsin EN Togo AV Pozharisski KM Ivanova OA Nardacci R Falasca L Piacentini M Imyanitov EN Hanson KP Tissue transglutaminase expression in breast carcinomas J Exp Clin Cancer Res 2001 20 265 8 11484985
Jiang WG Ablin R Douglas-Jones A Mansel RE Expression of transglutaminases in human breast cancer and their possible clinical significance Oncol Rep 2003 10 2039 44 14534740
Hettasch JM Bandarenko N Burchette JL Lai TS Marks JR Haroon ZA Peters K Dewhirst MW Iglehart JD Greenberg CS Tissue transglutaminase expression in human breast cancer Lab Invest 1996 75 637 45 8941210
Haroon ZA Lai TS Hettash JM Lindberg RA Dewhirst MW Greenberg CS Tissue transglutaminase is expressed as a host response to tumor invasion and inhibits tumor growth Lab Invest 1999 79 1679 86 10616216
Mapara MY Sykes M Tolerance and cancer: mechanisms of tumor evasion and strategies for breaking tolerance J Clin Oncol 2004 22 1136 51 15020616 10.1200/JCO.2004.10.041
Lochter A Sternlicht MD Werb Z Bissell MJ The significance of matrix metalloproteinases during early stages of tumor progression Ann NY Acad Sci 1998 857 180 93 9917841
Dano K Romer J Nielsen BS Bjorn S Pyke C Rygaard J Lund LR Cancer invasion and tissue remodeling – cooperation of protease systems and cell types Acta Pathol Immunol Microbiol Scand 1999 107 120 7
Werb Z Vu TH Rinkenberger JL Coussens LM Matrix-degrading proteases and angiogenesis during development and tumor formation Acta Pathol Immunol Microbiol Scand 1999 107 11 8
Belkin AM Zemskov EA Hang J Akimov SS Sikora S Strongin AY Cell-surface-associated tissue transglutaminase is a target of MMP-2 proteolysis Biochemistry 2004 43 11760 69 15362860 10.1021/bi049266z
Belkin AM Akimov SS Zaritskaya LS Ratnikov BI Deryugina EI Strongin AY Matrix-dependent proteolysis of surface transglutaminase by membrane-type metalloproteinase regulates cancer cell adhesion and locomotion J Biol Chem 2001 276 18415 22 11278623 10.1074/jbc.M010135200
Koul D Parathasarathy R Shen R Davies MA Jasse SA Chintala SK Rao JS Sun Y Benvenisite EN Liu TJ Yung WKF Suppression of metalloproteinase-2 gene expression and invasion in human glioma cells by MMAC/PTEN Oncogene 2001 20 6669 78 11709701 10.1038/sj.onc.1204799
Mehta K Fok J Miller FR Koul D Sahin AA Prognostic Significance of tissue transglutaminase in drug resistant and metastatic breast cancer Clin Cancer Res 2004 10 8068 8076 15585642
|
16153302
|
PMC1224867
|
CC BY
|
2021-01-04 16:36:35
|
no
|
Mol Cancer. 2005 Sep 9; 4:33
|
utf-8
|
Mol Cancer
| 2,005 |
10.1186/1476-4598-4-33
|
oa_comm
|
==== Front
Mol PainMolecular Pain1744-8069BioMed Central London 1744-8069-1-261615939210.1186/1744-8069-1-26ResearchProtein kinases mediate increment of the phosphorylation of cyclic AMP -responsive element binding protein in spinal cord of rats following capsaicin injection Wu Jing [email protected] Guangxiao [email protected] Long [email protected] Xuan [email protected] Yongzhong [email protected] Junfa [email protected] Qing [email protected] Li [email protected] Division of Neurosurgery, Department of Surgery, The University of Texas Medical Branch, Galveston, TX 77555-0517; USA2 Department of Neurology, University of Texas Health Science Center, Houston, TX77030-1501, USA3 Department of Neuroscience and Cell Biology, The University of Texas Medical Branch, Galveston, TX 77555-1043, USA4 Institute for Biomedical Science of Pain, Department of Neurobiology, Capital University of Medical Sciences, Beijing 100054, China2005 13 9 2005 1 26 26 22 6 2005 13 9 2005 Copyright © 2005 Wu et al; licensee BioMed Central Ltd.2005Wu et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Strong noxious stimuli cause plastic changes in spinal nociceptive neurons. Intracellular signal transduction pathways from cellular membrane to nucleus, which may further regulate gene expression by critical transcription factors, convey peripheral stimulation. Cyclic AMP-responsive element binding protein (CREB) is a well-characterized stimulus-induced transcription factor whose activation requires phosphorylation of the Serine-133 residue. Phospho-CREB can further induce gene transcription and strengthen synaptic transmission by the activation of the protein kinase cascades. However, little is known about the mechanisms by which CREB phosphorylation is regulated by protein kinases during nociception. This study was designed to use Western blot analysis to investigate the role of mitogen-activated protein (MAP)/extracellular signal-regulated kinase (ERK) kinase (MEK 1/2), PKA and PKC in regulating the phosphorylation of CREB in the spinal cord of rats following intraplantar capsaicin injection.
Results
We found that capsaicin injection significantly increased the phosphorylation level of CREB in the ipsilateral side of the spinal cord. Pharmacological manipulation of MEK 1/2, PKA and PKC with their inhibitors (U0126, H89 and NPC 15473, respectively) significantly blocked this increment of CREB phosphorylation. However, the expression of CREB itself showed no change in any group.
Conclusion
These findings suggest that the activation of intracellular MAP kinase, PKA and PKC cascades may contribute to the regulation of phospho-CREB in central nociceptive neurons following peripheral painful stimuli.
Central sensitizationtranscription factorsprotein kinase cascadenociception
==== Body
Background
Peripheral tissue injuries lead to persistent pain which can be caused, in part, by the central sensitization of spinal sensory neurons. It reflects an amplified responsiveness of nociceptive neurons that involves long-lasting changes in synaptic transmission in the central nervous system [1-3]. The changed central synaptic plasticity may be initiated by activity-dependent intracellular biochemical pathways following noxious stimulation of the peripheral tissue [1-5].
It has been extensively reported that extracellular signals conveyed and transduced through the plasma membrane to the intracellular nucleus trigger the expression of several immediate early genes, such as c-fos, c-Jun, Egr 1 and CREB [6-13]. The activation of these critical transcription factors further initiates a cascade of biological changes in neural functioning through changes in gene expression. Cyclic AMP (cAMP)-responsive element binding protein (CREB) is one of the well-characterized stimulus-induced transcription factors. Stimulus-induced influx of calcium activates the second messenger systems which are believed to be important for the long-term alteration in CREB-related gene expression [6,8,14-17]. Furthermore, this event involves the binding of CREB to a specific sequence present in the promoter of many cAMP responsive genes. The transduction pathways are comprised of cascades of intracellular signaling protein kinases, such as the second messenger system, [4,5,9,18,19]. These kinases play an important role in the subsequent activation or phosphorylation of the transcription factor, CREB. Phospho-CREB-induced gene transcription and strengthened synaptic transmission were widely reported in neuronal processing of various extracellular signals [8,13,14,17,20-22]. For example, it was reported that an increased CREB phosphorylation of the Serine-133 residue is responsible for the induction of long-term potentiation (LTP) in CA1 in the hippocampus. Additionally, the coordinated phosphorylation of CREB was initiated by the rapid phosphorylation of MAP kinase, as well as other activated protein kinases, such as calcium/calmodulin protein-dependent kinase II (CaM kinase II), protein kinase A and C [6,9,17,22-24]. More recently, central sensitization of spinal nociceptive neurons to peripheral stimulus was extensively reported and the mechanism of central sensitization was found in a spinal cord form of LTP [2]. Earlier studies showed that intradermal capsaicin injection in experimental animals induces central sensitization, which resembles LTP in many respects [1,2,4,25,26].
Central sensitization involves several crucial events generated by the activation of protein kinases, such as CaM kinase II, MAP kinase, protein kinase A (PKA), protein kinase C (PKC) and protein kinase G (PKG) [2,4,5,9,15,18,19,27]. These important cellular events trigger further transactivator selective target substrates, such as CREB, to enhance gene regulation [26,28]. We have found increased phosphorylation of spinal CREB protein following capsaicin injection in rats. Although the findings reported that phosphorylation events can be regulated by nitric oxide, as well as CaM kinase II [6,13,29], it remains to be determined whether MAP kinase, PKA and PKC participate in the phosphorylation of CREB in spinal cord central sensitization.
The current study was designed to examine the contribution of MAP kinase, PKA and PKC in the regulation of CREB protein phosphorylation in the ipsilateral side of the cord during central sensitization triggered by intraplantar capsaicin injection in rats.
Results
The results from this study demonstrated that the CREB protein and its phosphorylated form, phospho-CREB, were detected by immunoblot analysis in each group of animals. The molecular weight of CREB and phospho-CREB were seen at the 45 KD band.
Following capsaicin injection in four different treated groups of rats (ACSF-, U0126-, H89-, NPC 15473- treatment), the expression of CREB protein itself showed no significant changes in rats treated with ACSF when compared to the vehicle-treated groups (Figures 1B, 2B, 3B). Meanwhile, a significant enhancement of Ser-133 CREB phosphorylation (45KD) was found in the spinal cord tissue in the group of rats treated with ACSF. As we reported previously, the infusion of ACSF made no significant difference in the phospho-CREB expression when compared to that of naïve rats [6].
Figure 1 Detection of CREB protein and phosphorylated CREB in spinal cord in rats following vehicle or capsaicin injection and the effect of MAP kinase inhibitor, U0126. Panel A. Immunoblot data of CREB and phospho-CREB protein expression. Panel B. Bar graph summarizing the density of the immunoblot band of CREB protein. (NS, no significance between vehicle treatment vs. capsaicin treatment group in either ACSF- or U0126- treatment). Panel C. Bar graph demonstrating the density of the Western blot bands of phospho-CREB protein. (*, p < 0.05; the value from vehicle treatment vs. capsaicin treatment; #, p < 0.05, the value of ACSF-treated vs. U0126-treated animals treated with capsaicin; NS, no significance between vehicle vs. capsaicin treatment in groups with U0126 administration; n = 5 in each group). Open bar, vehicle group; hatched bar, capsaicin group.
Figure 2 Expression of CREB protein and phospho-CREB protein in spinal cord in rats and the effect of intrathecal PKA inhibitor, H89, treatment. Panel A. Immunoblot bands of CREB and phospho-CREB protein expression. Panel B. Bar graph summarizing the density of the immunoblot CREB bands. (NS, no significance between vehicle treatment vs. capsaicin treatment group with either ACSF- or H89- treatment). Panel C. Bar graph showing the density of the immunoblotting bands of phospho-CREB protein. (*, p < 0.05; the value from vehicle treatment vs. capsaicin treatment; #, p < 0.05, the value of ACSF-treated vs. H89-treated animals after capsaicin injection; NS, no significance between vehicle vs. capsaicin treatment in groups with H89 intrathecal administration; n = 5 in each group). Open bar, vehicle group; hatched bar, capsaicin group.
Figure 3 Expression of CREB and phosphorylated-CREB protein in spinal cord and the effect of intrathecal PKC inhibitor, NPC 15473, administration on their expression. Panel A. Western blot bands of CREB and phospho-CREB protein expression. Panel B. Bar graph showing the density of the immunoblot band of CREB protein. (NS, no significance between vehicle treatment vs. capsaicin treatment group with either ACSF- or NPC 15473- treatment). Panel C. Bar graph indicating the density of the immunoblotting band of phospho-CREB protein. (*, p < 0.05; the value from vehicle treatment vs. capsaicin treatment; #, p < 0.05, the value of ACSF-treated vs. NPC 15473-treated animals following capsaicin injection; n = 6 in each group). Open bar, vehicle group; hatched bar, capsaicin group.
In our first experiment, we tested whether changes in CREB expression and CREB phosphorylation could be regulated by MAP kinase activity, The intrathecal application of U0126, a MEK 1/2 inhibitor (dissolved in 10% DMSO, 1 μg/ μl and diluted with PBS to a total volume of 10 μl), has been shown to reduce CREB phosphorylation in the spinal cord tissue, which is otherwise enhanced by capsaicin injection, when compared to the value from ACSF-treated animals (1.53 ± 0.29 vs. 2.91 ± 0.3, Figures 1A and 1C, p < 0.05). Within the group receiving U0126 for 30 minutes, the value of phospho-CREB expression in the capsaicin group demonstrated a higher level than that of the vehicle-injected group, which showed no significance (1.53 ± 0.29 vs. 1.34 ± 0.3, p > 0.05). However, ACSF treatment did not affect the increased phosphorylation of CREB protein in response to capsaicin injection (2.91 ± 0.3 vs. 1.31 ± 0.31, p < 0.05).
Next, we assessed the effect of protein kinase A (PKA) in mediating CREB phosphorylation since PKA is one of the important intracellular triggers for CREB activation. The intrathecal administration of the selective PKA inhibitor, H89 (10 μM for 30 minutes), was performed before inducing acute noxious stimulation with a capsaicin injection in rats. Results showed that changes in CREB protein expression were not found in either the ACSF-treated group or in H89-treated animals (Figure 2B). However, there was a significant increase in the phospho-CREB in ACSF-treated animals after capsaicin injection (1.29 ± 0.2 vs. 3.02 ± 0.25, p < 0.05). Moreover, this capsaicin-induced increase in CREB phosphorylation was inhibited by infusion of H89 when compared to that of ACSF-treated rats (1.49 ± 0.43 vs. 1.34 ± 0.37, p < 0.05, Figure 2C).
Finally, we demonstrated a role of protein kinase C (PKC) in regulating CREB phosphorylation in response to capsaicin injection. We infused the PKC inhibitor, NPC 15473, intrathecally before injection of capsaicin in rats. An enhancement of CREB phosphorylation was observed in capsaicin-treated animals when compared to the vehicle-treated group when ACSF was infused. However, the intrathecal infusion of NPC 15473 (20 mM for 30 minutes) significantly reduced the phosphorylation level of CREB following capsaicin injection. The decreased level showed a significant change when compared to that of rats receiving ACSF infusion (1.76 ± 0.2 vs. 2.89 ± 0.3, p < 0.05, Figure 3C). CREB protein itself did not respond to any treatment (Figure 3B).
Discussion
The major findings in the current study demonstrated that intradermal capsaicin injection causes an increase in Ser-133 CREB phosphorylation in the dorsal horn and that the enhancement of CREB phosphorylation is regulated by the activity of MAP kinase, PKA and PKC.
A number of studies have shown that enhanced phosphorylation of CREB transcription factor could account for the potential contribution to long-term changes in the spinal processing of nociceptive information [6,13]. As one of important transcription factors, CREB is phosphorylated in response to stimuli-triggered calcium influx into post-synaptic neurons [14,21,22]. Our previous findings reported the role of nitric oxide system and CaM kinase II in the CREB regulation [13,25]. In this investigation, we report that noxious stimulation with capsaicin injection induces phosphorylation of CREB protein (Ser-133) in the spinal cord. This finding is similar to what we reported previously [6] and studies of other investigators using different stimulation regimens. In addition, we assessed the regulating mechanism by which intracellular protein kinase mediates CREB phosphorylation. The data demonstrated that the intrathecal treatment with inhibitors of MAP kinase, PKA or PKC blocked capsaicin-triggered CREB phosphorylation (at the Serine 133 site), confirming that not only c-AMP-dependent protein kinase A (PKA), but also ERK/MAP kinase and PKC, play roles in the mediation of CREB phosphorylation during the central processing of nociception.
Peripheral noxious stimulation, such as capsaicin injection, causes an increased responsiveness in spinal nociceptive neurons that involve the activation of glutamate receptors. This event produces a large influx of calcium into the nociceptive neurons, activating multiple intracellular protein kinase cascades, such as CaM kinase II, PKC, as well as MAP kinase [2,4,6,18,30-34]. The PKA, PKG and nitric oxide synthase systems were also found to be activated following increased cAMP and cGMP in nociceptive neurons during painful stimulation [2,22,35]. Enhanced phosphorylation of CREB through the activation of glutamate receptors and the above kinase cascades during central sensitization suggest a connection between CREB phosphorylation and the molecular mechanisms governing stimuli-induced CREB activation by protein kinase pathways. Investigators from other laboratories also noted that increased CREB phosphorylation was found in animals following different kinds of noxious stimulation [7,8,11,17].
Like other transcription factors, such as AP-1 proteins, c-fos, and c-Jun, phosphorylated CREB may participate in the downstream signal transduction cascade in the longer lasting changes in synaptic plasticity. CREB phosphorylation was reported in the transcription regulation of nociception-related genes, such as dynorphin, enkephalin and opioid receptors, during the activation of nociceptive neurons. A CREB binding site is found in the promoter regions of the dynorphin, enkephalin and opioid receptor genes following painful stimulation [14,16,20,28,35]. CREB phosphorylation is required for prolonged synaptic plasticity strengthening during central sensitization [2,21,24].
Recently, mitogen-activated protein (MAP)/extracellular signal-regulated kinase (ERK) kinase (MEK 1/2) was reported to play a critical role in spinal dorsal horn neurons in response to various painful stimuli [5,18,30-34]. Following noxious peripheral stimuli or c-Fiber volleys, an increased phosphorylation of ERK was observed in nociceptive neurons in rats and an inhibitor of ERK reduced the formalin-injection-induced nociceptive behavior [5,36]. Peripheral nerve injuries or noxious stimulation with various inflammatory agents, such as carrageenan, capsaicin, Complete Freund's Adjuvent (CFA), mustard oil and melittin (a major toxic peptide of whole bee venom) were reported to active the MAP/MEK kinase pathway [5,9,18,20,31,33,34][37]. Furthermore, the activation of MAP kinase pathways was related to the activation of several types of glutamate receptors (NMDA, AMPA and mGlu-R) in dorsal horn neurons, as well as other intracellular protein kinase cascades in the pathogenesis of nociceptive sensitization [9,18]. The activation of MAP kinase signaling was also reported to be involved in the transcriptional regulation of prodynorphin and neurokinin-1 (NK-1) gene products [5,14,16,35]. In our present study, we found that U0126, a MAP kinase inhibitor, significantly blocked CREB phosphorylation at Ser-133 in response to capsaicin injection. Another investigator reported a similar effect of U0126 in a different spinal cord slice preparation following C-fiber stimulation [9]. In an investigation involving in vitro preparation of cultured striatal neurons, the NMDA- and AMPA/kainate-induced CREB phosphorylation was blocked by MAP kinase inhibitor, U0126 [23]. These findings provide supportive evidence that the MAP kinase pathway is one of the important intracellular elements and may contribute to the nociceptive plasticity in central sensitization.
It is well recognized that both the PKA and the PKC pathways are involved in the central neurotransmission of nociception. In experimental animals with capsaicin injection induced-central sensitization, activation of the PKA or the PKC pathways by their activators enhanced the response of spinothalamic (STT) cells to mechanical stimuli [15,19] and inhibitors of PKA or PKC blocked the hypersensitivity of STT neurons. Experiments from behavioral studies also support the roles of PKA and PKC in the generation of spinal cord central sensitization [2,15,19]. Our study further demonstrates the roles of PKA and PKC in molecular mechanism of regulation of gene transcription in response to capsaicin injection, in which they mediate the phosphorylation of the transcription factor, CREB. However, as important regulators, both PKA and PKC actively contribute to the activity-dependent hippocampal synaptic plasticity by regulating CREB-dependent target genes. The present study supports the results from our earlier work on central sensitization, which is believed to be a spinal cord form of LTP [2].
In summary, the findings from our investigation provide strong evidence that MAP kinase, PKA and PKC participate in the regulation of CREB phosphorylation status in the spinal cord following peripheral noxious stimulation. These results also suggest a possible explanation for how the activated intracellular kinase cascades convey extracellular signals into the nucleus for subsequent transcription of plasticity-associated genes in the spinal cord.
Methods
Male Sprague-Dawley rats (Harlan Sprague-Dawley, Houston, Texas) weighing 280–350 grams were used in this project. All experiments were performed in accordance with the ethical guidelines of the International Association for the Study of Pain and National Institutes of Health and were approved by the Institutional Animal Care and Use Committee of the University of Texas Medical Branch. Capsaicin (100 μg in 10 μl) was injected into the plantar surface of the left foot of rats following anesthesia with sodium pentobarbital (i.p., 50 mg/kg). For the intrathecal administration of agents, an intrathecal catheter (PE 32) was implanted in the spinal subarachnoid space. The other end was connected with PE10 and then PE20 tubing for a stepdown connection to a Hamilton syringe [4,6]. Briefly, a midline incision was made to expose the suboccipital region in the rats and the catheter was gently advanced ~7.4 cm into the spinal subarachnoid space. The tip of the catheter was placed at around the estimated level of the L4/5 spinal segments. The rats were returned to the facility for 5 days to allow them to recover from the surgery. On the day of the experiment, the catheter was connected to a 1 ml syringe, which was mounted on an infusion pump that kept a constant infusion rate of 1 μl/min for the delivery of inhibitors. U0126, a selective inhibitor of MEK 1/2 (1,4 -Diamino-2,3-dicyano-1,4-bis (2-aminophenylthio)-butadiene, Biosource, CA); H89, a protein kinase A inhibitor ((N-[2-((3-bromophenyl)-2-propenyl)amino)ethyl]-5-isoquinoline sulfonamide, HCL, Calbiochem., San Diego, CA); NPC15473, a protein kinase C inhibitor (2,6-diamino-N-([1-oxotridecyl)-2-piperidinyl]methyl)hexanamide, Alexis, San Diego, CA); and ACSF were infused. U0126 was dissolved in 10% DMSO (1 μg/μl) and diluted with PBS (total 10 μl). As we previously reported [25], the concentrations of NPC15473 and H89 were 20 mM and 10 μM, respectively. The specificity of the inhibitors was previously reported [5,19,25]. All of the agents were infused for 30 min. At 30-min post-infusion, the injection of capsaicin (100 μg in 10 μl) was performed. Spinal cord tissues were collected and Western blotting performed as previously reported [4,6,25]. Briefly, the animals were sacrificed at 30 min post-injection of capsaicin. The segments of the spinal cord tissues (L3–L6) were collected and immediately placed on a glass plate cooled with dry ice. The dorsal quadrants from the ipsilateral side were collected and immediately placed into liquid nitrogen. We were interested in detecting CREB phosphorylation in the ipsilateral side of the cord, since we previously reported that the biochemical changes in this side are more sensitive to capsaicin injection [4,6,26]. The tissue was homogenized with buffer containing 50 mol/l Tris buffer, pH 7.4(0.1 mol/l EGTA, 0.14 μl/ml β-mercapto-ethanol, 100 mol/l PMSF, and 0.2 mg/ml trypsin inhibitor). The protein concentration was measured by using a BCA kit (Pierce, Rockford, IL) and read on a microplate reader (Sun Bioscience).
Equal amounts of protein (60 μg) were loaded and size-fractionated by gel electrophoresis (SDS-PAGE) in a 4–20% Ready-Gel preparation and transferred onto a PVDF membrane (Bio-Rad, Hercules, CA). After incubating in blocking buffer, the membranes were incubated with primary polyclonal antibodies to CREB (1:800) or phospho-CREB (1:1000)at the Ser-133 site (Upstate Biotechnology, Lake Placid, NY) overnight at 4°C [17]. The anti-CREB is a polyclonal antibody from rabbits immunized with a synthetic peptide (C-SGAENQQSGDAAVTEAENQQ) corresponding to amino acids 5–24 of the human CREB. The antibody to phospho-CREB at the Ser-133 site was produced with another synthetic phospho-peptide corresponding to residues 123–136 (KRREILSRRPpSYRK). The blots were washed three times with PBS buffer and then incubated with HRP-conjugated anti-rabbit IgG (1: 5000; Santa Cruz, San Francisco, CA) in 5% (w/v) non-fat milk in PBS buffer. The membranes were enhanced with a chemiluminescence reagent (ECL kit, Amersham, Arlington Height, IL) and were exposed to the film. The expression of β-actin, as an internal control, was also measured with the same protocol in every group (monoclonal antibody against β-actin purchased from Sigma Co., St. Louis, MO). The density of the immunoreactive bands on the films was acquired by using a Doc-It Gel system and AlphaEase software. The densitometric units of bands of CREB or phospho-CREB were expressed relative to the values for β-actin. The significance of differences between groups treated with ACSF and protein inhibitors was calculated and compared with the Student's t-test [17]. P < 0.05 was considered significant. All data were expressed as mean ± S.E.M.
Acknowledgements
This work was funded by Sealy Grant 2951-02, NIH P30/DK 56338-02/05, DE 014814 (LF), NIH grants NS 11255 and NS 40723 (QL). J Wu is a JE Kempner Scholar. We thank Dr. WD Willis for critical reading with the manuscript.
==== Refs
Baumann TK Simone DA Shain CN LaMotte RH Neurogenic hyperalgesia: the search for the primary cutaneous afferent fibers that contribute to capsaicin-induced pain and hyperalgesia J Neurophysiol 1991 66 212 227 1919668
Willis WD Long-term potentiation in spinothalamic neurons Brain Res Brain Res Rev 2002 40 202 214 12589918 10.1016/S0165-0173(02)00202-3
Zhuo M NMDA receptor-dependent long term hyperalgesia after tail amputation in mice Eur J Pharmacol 1998 349 211 220 9671100 10.1016/S0014-2999(98)00197-6
Fang L Wu J Lin Q Willis WD Calcium-calmodulin-dependent protein kinase II contributes to spinal cord central sensitization J Neurosci 2002 22 4196 4204 12019337
Ji RR Befort K Brenner GJ Woolf CJ ERK MAP kinase activation in superficial spinal cord neurons induces prodynorphin and NK-1 upregulation and contributes to persistent inflammatory pain hypersensitivity J Neurosci 2002 22 478 485 11784793
Fang L Wu J Zhang X Lin Q Willis WD Calcium/calmodulin dependent protein kinase II regulates the phosphorylation of cyclic AMP-responsive element-binding protein of spinal cord in rats following noxious stimulation Neurosci Lett 2005 374 1 4 15631885 10.1016/j.neulet.2004.10.014
Hoeger-Bement MK Sluka KA Phosphorylation of CREB and mechanical hyperalgesia is reversed by blockade of the cAMP pathway in a time-dependent manner after repeated intramuscular acid injections J Neurosci 2003 23 5437 5445 12843242
Ji RR Rupp F Phosphorylation of transcription factor CREB in rat spinal cord after formalin-induced hyperalgesia: relationship to c-fos induction J Neurosci 1997 17 1776 1785 9030636
Kawasaki Y Kohno T Zhuang ZY Brenner GJ Wang H Van Der Meer C Befort K Woolf CJ Ji RR Ionotropic and metabotropic receptors, protein kinase A, protein kinase C, and Src contribute to C-fiber-induced ERK activation and cAMP response element-binding protein phosphorylation in dorsal horn neurons, leading to central sensitization J Neurosci 2004 24 8310 8321 15385614 10.1523/JNEUROSCI.2396-04.2004
Ko SW Vadakkan KI Ao H Gallitano-Mendel A Wei F Milbrandt J Zhuo M Selective contribution of Egr1 (zif/268) to persistent inflammatory pain J Pain 2005 6 12 20 15629414 10.1016/j.jpain.2004.10.001
Ma W Hatzis C Eisenach JC Intrathecal injection of cAMP response element binding protein (CREB) antisense oligonucleotide attenuates tactile allodynia caused by partial sciatic nerve ligation Brain Res 2003 988 97 104 14519530 10.1016/S0006-8993(03)03348-1
Wu J Fang L Lin Q Willis WD Fos expression is induced by increased nitric oxide release in rat spinal cord dorsal horn Neuroscience 2000 96 351 357 10683575 10.1016/S0306-4522(99)00534-5
Wu J Fang L Lin Q Willis WD The role of nitric oxide in the phosphorylation of cyclic adenosine monophosphate-responsive element-binding protein in the spinal cord after intradermal injection of capsaicin J Pain 2002 3 190 198 14622772 10.1054/jpai.2002.123653
Borsook D Konradi C Falkowski O Comb M Hyman SE Molecular mechanisms of stress-induced proenkephalin gene regulation: CREB interacts with the proenkephalin gene in the mouse hypothalamus and is phosphorylated in response to hyperosmolar stress Mol Endocrinol 1994 8 240 248 8170480 10.1210/me.8.2.240
Lin Q Wu J Willis WD Effects of protein kinase a activation on the responses of primate spinothalamic tract neurons to mechanical stimuli J Neurophysiol 2002 88 214 221 12091547
Mayr B Montminy M Transcriptional regulation by the phosphorylation-dependent factor CREB Nat Rev Mol Cell Biol 2001 2 599 609 11483993 10.1038/35085068
Miletic G Hanson EN Miletic V Brain-derived neurotrophic factor-elicited or sciatic ligation-associated phosphorylation of cyclic AMP response element binding protein in the rat spinal dorsal horn is reduced by block of tyrosine kinase receptors Neurosci Lett 2004 361 269 271 15135945 10.1016/j.neulet.2003.12.029
Adwanikar H Karim F Gereau RW Inflammation persistently enhances nocifensive behaviors mediated by spinal group I mGluRs through sustained ERK activation Pain 2004 111 125 135 15327816 10.1016/j.pain.2004.06.009
Sluka KA Rees H Chen PS Tsuruoka M Willis WD Inhibitors of G-proteins and protein kinases reduce the sensitization to mechanical stimulation and the desensitization to heat of spinothalamic tract neurons induced by intradermal injection of capsaicin in the primate Exp Brain Res 1997 115 15 24 9224830
Ginty DD Bading H Greenberg ME Trans-synaptic regulation of gene expression Curr Opin Neurobiol 1992 2 312 316 1322751 10.1016/0959-4388(92)90121-Z
Sheng M Greenberg ME The regulation and function of c-fos and other immediate early genes in the nervous system Neuron 1990 4 477 485 1969743 10.1016/0896-6273(90)90106-P
Sheng M Thompson MA Greenberg ME CREB: a Ca2+-regulated transcription factor phosphorylated by calmodulin-dependent kinases Science 1991 252 1427 1430 1646483
Mao L Tang Q Samdani S Liu Z Wang JQ Regulation of MAPK/ERK phosphorylation via ionotropic glutamate receptors in cultured rat striatal neurons Eur J Neurosci 2004 19 1207 1216 15016079 10.1111/j.1460-9568.2004.03223.x
Soderling TR CaM-kinases: modulators of synaptic plasticity Curr Opin Neurobiol 2000 10 375 380 10851169 10.1016/S0959-4388(00)00090-8
Fang L Wu J Lin Q Willis WD Protein kinases regulate the phosphorylation of the GluR1 subunit of AMPA receptors of spinal cord in rats following noxious stimulation Brain Res Mol Brain Res 2003 118 160 165 14559367 10.1016/j.molbrainres.2003.08.002
Fang L Wu J Zhang X Lin Q Willis WD Increased phosphorylation of the GluR1 subunit of spinal cord alpha-amino-3-hydroxy-5-methyl-4-isoxazole propionate receptor in rats following intradermal injection of capsaicin Neuroscience 2003 122 237 245 14596864 10.1016/S0306-4522(03)00526-8
Galan A Lopez-Garcia JA Cervero F Laird JM Activation of spinal extracellular signaling-regulated kinase-1 and -2 by intraplantar carrageenan in rodents Neurosci Lett 2002 322 37 40 11958838 10.1016/S0304-3940(02)00078-2
Sassone-Corsi P Visvader J Ferland L Mellon PL Verma IM Induction of proto-oncogene fos transcription through the adenylate cyclase pathway: characterization of a cAMP-responsive element Genes Dev 1988 2 1529 1538 2850967
Wu J Fang L Lin Q Willis WD Nitric oxide synthase in spinal cord central sensitization following intradermal injection of capsaicin Pain 2001 94 47 58 11576744 10.1016/S0304-3959(01)00340-2
Dai Y Iwata K Fukuoka T Kondo E Tokunaga A Yamanaka H Tachibana T Liu Y Noguchi K Phosphorylation of extracellular signal-regulated kinase in primary afferent neurons by noxious stimuli and its involvement in peripheral sensitization J Neurosci 2002 22 7737 7745 12196597
Galan A Cervero F Laird JM Extracellular signaling-regulated kinase-1 and -2 (ERK 1/2) mediate referred hyperalgesia in a murine model of visceral pain Brain Res Mol Brain Res 2003 116 126 134 12941468 10.1016/S0169-328X(03)00284-5
Ma W Quirion R Partial sciatic nerve ligation induces increase in the phosphorylation of extracellular signal-regulated kinase (ERK) and c-Jun N-terminal kinase (JNK) in astrocytes in the lumbar spinal dorsal horn and the gracile nucleus Pain 2002 99 175 184 12237195 10.1016/S0304-3959(02)00097-0
Obata K Yamanaka H Dai Y Mizushima T Fukuoka T Tokunaga A Noguchi K Activation of extracellular signal-regulated protein kinase in the dorsal root ganglion following inflammation near the nerve cell body Neuroscience 2004 126 1011 1021 15207334 10.1016/j.neuroscience.2004.04.036
Pezet S Malcangio M Lever IJ Perkinton MS Thompson SW Williams RJ McMahon SB Noxious stimulation induces Trk receptor and downstream ERK phosphorylation in spinal dorsal horn Mol Cell Neurosci 2002 21 684 695 12504600 10.1006/mcne.2002.1205
Lee PW Lee YM Transcriptional regulation of mu opioid receptor gene by cAMP pathway Mol Pharmacol 2003 64 1410 1418 14645671 10.1124/mol.64.6.1410
Yu YQ Chen J Activation of spinal extracellular signaling-regulated kinases by intraplantar melittin injection Neurosci Lett 2005 381 194 198 15882816 10.1016/j.neulet.2005.02.033
|
16159392
|
PMC1224868
|
CC BY
|
2021-01-04 16:40:08
|
no
|
Mol Pain. 2005 Sep 13; 1:26
|
utf-8
|
Mol Pain
| 2,005 |
10.1186/1744-8069-1-26
|
oa_comm
|
==== Front
Reprod Biol EndocrinolReproductive biology and endocrinology : RB&E1477-7827BioMed Central London 1477-7827-3-431613733510.1186/1477-7827-3-43ReviewFundamental roles of reactive oxygen species and protective mechanisms in the female reproductive system Fujii Junichi [email protected] Yoshihito [email protected] Futoshi [email protected] Department of Biomolecular Function, Yamagata University Graduate School of Medicine, 2-2-2 Iidanishi, Yamagata 990-9585, Japan2005 2 9 2005 3 43 43 28 7 2005 2 9 2005 Copyright © 2005 Fujii et al; licensee BioMed Central Ltd.2005Fujii et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Controlled oxidation, such as disulfide bond formation in sperm nuclei and during ovulation, plays a fundamental role in mammalian reproduction. Excess oxidation, however, causes oxidative stress, resulting in the dysfunction of the reproductive process. Antioxidation reactions that reduce the levels of reactive oxygen species are of prime importance in reproductive systems in maintaining the quality of gametes and support reproduction. While anti-oxidative enzymes, such as superoxide dismutase and peroxidase, play a central role in eliminating oxidative stress, reduction-oxidation (redox) systems, comprised of mainly glutathione and thioredoxin, function to reduce the levels of oxidized molecules. Aldo-keto reductase, using NADPH as an electron donor, detoxifies carbonyl compounds resulting from the oxidation of lipids and proteins. Thus, many antioxidative and redox enzyme genes are expressed and aggressively protect gametes and embryos in reproductive systems.
==== Body
Introduction
Since reproductive and developmental process accompany dynamic changes in metabolism and energy consumption, byproducts are also generated on an extraordinary scale. Among such byproducts, reactive oxygen species (ROS), which are inevitably generated during the physiological process of oxygen consumption, the levels of which are enhanced under some pathological conditions [1], are the most troublesome. Although ROS as well as nitric oxide (NO), which is produced in a limited amount in response to physiological stimuli, are considered to mediate inter- and intra-cellular signaling, generation of an excess them results in oxidative stress, which is determined by the balance between oxidants and antioxidants, and becomes the primary or secondary cause in deterioration, due to various diseases. Gametes are extremely sensitive to damage by ROS and must be protected against to maintain the species. Low molecular weight compounds, such as antioxidative vitamins (A, C, and E) and glutathione, react with ROS and convert them to harmless compounds. In addition, living organisms have evolved enzymatic systems that effectively suppress oxidative stress and minimize damage caused by ROS.
It is, however, not possible to eliminate all ROS and, hence, oxidative modification occurs in the building blocks of cells. Oxidative stress leads to disulfide bond formation of sulfoxidation in sulfhydryl residues in proteins. Unsaturated lipids are prone to oxidation and are converted to peroxides, a target of the peroxidase reaction, followed by the generation of degradation products with an aldehyde moiety. The resulting aldehyde compounds are more or less toxic and must be detoxified. Small molecules can be simply discarded into the urine, but the remainder must be reduced by a corresponding reduction-oxidation (redox) system. However, contents of carbonyl groups in oxidized proteins are generally increased and considered to be a marker of the oxidative modification of proteins and aging [2]. Thus, the reduction potential of NAD(P)H is of prime importance in the maintenance of the redox balance of cells. Nucleic acids suffer from oxidative modification, with the base moiety being the preferred target. However, in most cases, damaged DNA can be efficiently repaired by several systems according to the double stranded nature of DNA. Since unrepaired bases are mutagenic, cells carrying them must also be eliminated.
The physiological relevance of antioxidative/redox systems in male reproductive tract has been overviewed recently [3] and, hence, are only briefly mentioned here. The emphasis of this review is on systems in female reproductive organs under physiological and pathological conditions and viewpoint is extended to fertilization and early embryonal development.
Origin of oxidative, nitrosative, and carbonyl stress
ROS are generated via various reactions in the body. Some ROS are produced by non-enzymatic reactions, for example via the Fenton reaction in the presence of transient metal ions [1]. Biological reactions, such as electron transfer and oxygenase reactions that utilize oxygen molecules as the substrate, also generate large amounts of ROS. Since the mitochondrial respiratory chain is the main oxygen-consuming system in cells, the majority of ROS are produced from this system under physiological conditions. The generation of ROS becomes excessive under conditions of elevated metabolism and pathological conditions and is matter of special concern. The following are well-known enzymatic systems that generate ROS. Xanthine dehydrogenase, an enzyme involved in purine metabolism, is converted to xanthine oxidase that generates superoxide under ischemic conditions in cardiovascular systems [4]. Cyclooxygenase, which catalyzes the initial oxidation step in the conversion from arachidonate to prostanoids and is induced under inflammatory conditions, also generates ROS [5]. The generation of ROS by the P450 system is important during the metabolic process of steroid hormone synthesis from cholesterol in endocrine organs, such as the ovary and testis. Professional phagocytes, such as neutrophils, contain NADPH-oxidase that generates huge amounts of superoxide for microbicidal purposes.
Nitrogen oxide species (RNOS) are mainly derived from nitric oxide (NO) and also play various roles in reproductive organs [6]. Nitric oxide synthase (NOS), which is encoded by three different genes, NOS I, NOS II, and NOS III, catalyzes the formation of NO from arginine and oxygen using NADPH as an electron donor. Among NO derived from the three isoforms of NOS, NO from nNOS (NOS I) appears to function as a neurotransmitter. NO generated from endothelial NOS (NOS III) is involved in vascular relaxation. NOS III expression is increased by luteinizing hormone (LH) surge or human chorionic gonadotropin (hCG). NOS III may also be involved in oocyte maturation and the ovulatory process as described below. The expression of inducible NOS (NOS II) is induced by many stimuli, including inflammatory cytokines. Since a certain fraction of NO is converted to harmful RNOS, such as peroxynitrite by reaction with ROS, nitrosative stress occurs simultaneously with ROS generation. NOS II is also commonly a matter of concern in reproductive organs under pathological conditions because it produces large amounts of NO in response to inflammatory and other stimuli.
Malondialdehyde and 4-hydroxy-2-nonenal levels are increased under conditions of oxidative stress and are still toxic because they carry aldehyde group. Carbonyl compounds are produced under hyperglycemic conditions, in addition to oxidative and nitrosative stress, and increase during the aging process. Amino acid residues in proteins are also subject to oxidative modification and are converted into carbonyl containing compounds. Since carbonyl compounds are also reactive toward thiols and amino groups and cause carbonyl stress [7], detoxification by reduction constitutes another pivotal system.
Antioxide/redox system
Many low molecular weight antioxidants, such as antioxidative vitamins and polyphenols, are ordinarily present in nutrients. Although ROS are scavenged by these compounds, enzymatic detoxification is more efficient [1]. The following major antioxidative enzymes are present in our body.
Superoxide dismutase (SOD)
The superoxide anion is produced by a one-electron reduction of an oxygen molecule and initiates a radical chain reaction. It is believed that SOD, which dismutates the superoxide anion to hydrogen peroxide, plays a central part in antioxidative reactions. Three isozymes are produced by mammalians.
SOD1 encodes Cu,Zn-SOD that contains Cu and Zn as metal cofactors and is largely cytosolic, while SOD2-encoding Mn-SOD is a mitochondrial isoform containing Mn. SOD3, which encodes the extracellular form (EC-SOD), is structurally similar to CuZn-SOD, and also contains Cu and Zn as metal cofactors. Since a mutation in SOD1 causes amyotrophic lateral sclerosis, extensive studies have been carried out in neuronal cells. One of the striking phenotypes of SOD1-deficient mice is female infertility and this is discussed below.
Mn-SOD is a mitochondrial isoform but its gene, SOD2, is encoded by nuclear DNA. SOD2 is inducible under various oxidative stress and inflammatory conditions and, hence, the regulatory mechanism of the gene has been a subject of extensive study. Homozygous SOD2-deficient mice suffer severe cardiovascular damage and die soon after birth [8]. Although no abnormality in the genital tract has been reported for heterozygous mice, transgenic male mice that express higher levels of Mn-SOD are infertile but the mechanism for this is unknown [9].
EC-SOD is present at high levels in the epididymis as well as the lung [10]. EC-SOD is also localized in the nuclei in the seminiferous tubules of the testis [11]. Superoxide decreases levels of NO by converting it to peroxinitrite. Thus, scavenging superoxide in vasculature extends the half-life of nitric oxide (NO), which results in an increase in cGMP levels. It is probable that elevated levels of cGMP relax vascular smooth muscle and supports erectile responses. Erectile function is improved by transferring the SOD3 gene to the penis in aged rats [12]. However, no recognizable phenotype in the reproductive system has yet been reported in SOD3 knockout mice [13].
Peroxidases
Glutathione peroxidase (GPx) plays a central role in the detoxification of peroxides using the reduced form of glutathione (GSH) as an electron donor. Many enzymes that are classified into different family of proteins exhibit GSH-dependent peroxidase activity. Conventional GPx contains selenocysteine (Sec) at its active center and appears to play a pivotal role in detoxification of peroxides. At least four selenium-containing GPx isozymes are produced in mammalians. The cytosolic form, GPX1, is widely distributed in tissues and has been the most extensively investigated form. However, GPX1-knockout mice show no abnormality in phenotype including reproductive capability [14]. GPX2 encodes a gastrointestinal form, and no specific function for it is known in reproduction. GPX3 is present in plasma and epididymal fluid. GPX4 encodes an isoform that specifically detoxifies phospholipid hydroperoxide is thus referred to as PhGPx, and is expressed at high levels in the testis. A defect in GPX4 has been suspected as a cause of male infertility triggered by Se deficiency, although direct evidence for its requirement is missing [15]. GPX4 protein represents about 50% of the capsule material that embeds the helix of mitochondria in the midpiece of spermatozoa [16]. A correlation between male infertility and a GPX4 defect has actually been reported [17,18]. Thus, GPX4 may have some physiological role in the male reproductive system. A novel isoform is specifically present in sperm nuclei and is considered to act as a protamine thiol peroxidase. Since the molecule has a high sequence identity to GPX4, except for the N-terminal region [19], it is likely that they are products of the same gene and are generated by an alternate promoter and exon usage of GPX4 [20].
Catalase exclusively detoxifies hydrogen peroxide and has no requirement for an electron donor. It plays a role in organs such as the liver, but its specific function in the genital tract is largely unknown. The number of enzymes carrying peroxidase activity is still increasing. Peroxiredoxins are recently identified multifunctional redox proteins with peroxidase activity that require electrons from thioredoxin [21].
Glutathione redox system
The sulfhydryl residue forms various redox states, as illustrated in Figure 1. Mild sulfhydryl oxidation produces disulfides and sulfenic acids, which are easily converted to disulfides by reaction with an adjacent sulfhydryl residue. Sulfenic acid is further oxidized to sulfinic acid and then to sulfonic acid. Disulfides and sulfenic acids are reduced back to the sulfhydryl stage by thioredoxin (Trx), glutatredoxin (Grx), and other thiol reductases under high redox potential. Recent reports have shown that sulfinic acid also can be reduced back to the sulfhydryl stage although the reaction requires ATP and, hence, is not a simple reduction reaction [22]. Sulfonic acid is not reversibly reduced to a sulfhydryl under physiological conditions. It is not possible to evaluate their generation in cells accurately because they are highly reactive and in a dynamic equilibrium.
Figure 1 Reaction of sulfhydryl group in response to oxidative stress and interconversion among the oxidation products. Sulfhydryl residues form different oxidative states that largely depends on the source and extent of oxidative stress. Since the redox system generally relies on reactive sulfhydryls, whether they are reversible or not is of prime importance.
Glutathione is a tripeptidyl molecule and is present in either the reduced (GSH) or the oxidized state (GSSG) by forming a disulfide bond between two molecules. It has pleiotropic roles, which include the maintenance of cells in a reduced state and the formation of conjugates with some harmful endogenous and xenobiotic compounds [23]. In addition, GSH serves as an electron donor for glutathione peroxidase that reduces peroxide to the corresponding alcohol, as described above. GSH levels are maintained by de novo synthesis that is catalyzed by two enzymes, γ-glutamylcysteine synthetase (γ-GCS) and glutathione synthetase (GS). The rate-limiting step in glutathione synthesis is the first reaction, in which γ-glutamylcysteine is formed, catalyzed by -GCS. An increase in GSH levels in response to various stimuli is mainly attributed to the responsiveness of -GCS gene expression to the stimuli. Buthionine sulfoximine (BSO), a specific inhibitor for γ-GCS, is thus commonly used to deplete intracellular GSH. The reduction of GSSG is catalyzed by glutathione reductase (GR) using NADPH as an electron donor. Nitrosourea (BCNU), an anti-cancer drug, is used to inhibit GR [24]. One of anti-cancer functions of BCNU is, therefore, attributable to the inhibition of GR and the lowering of GSH levels. GR is also inhibited by compounds produced in response to nitrosative stress, such as nitrosoglutathione. In the female reproductive system, GSH is assumed to play a role in reducing oxidative stress either by interaction directly with ROS or by donating electron to GPx.
Thioredoxin (Trx) system
Trx, originally identified as an electron donor for ribonucleotide reductase, functions to regulate various enzymes and trans-activating factors of genes, and is intimately involved in cell growth, differentiation, and death [25]. Trx also functions as a protein disulfide isomerase that corrects disulfide bridges that are formed in error. Moreover, Trx directly donates electrons to peroxiredoxin and, hence, is directly linked to the peroxidase reaction [21]. After oxidation, an intramolecular disulfide bond is formed in Trx. Oxidized thioredoxin is reduced by thioredoxin reductase, a selenocystine-containing oxido-reductase, using NADPH as an electron donor. Since Trx-knockout mice are embryonically lethal [26], Trx appears to play essential roles in the reproductive system and/or fetal development. Among multiple roles of Trx, defect in electron donation to ribonucleotide reductase appears to be the main cause in Trx-knockout mice because DNA synthesis is essential for fetal development.
Aldo-keto reductase
Carbonyl compounds are produced by the oxidation of organic compounds, such as unsaturated fatty acids, and are highly reactive. They modify reactive sulfhydryl groups that are commonly present in redox-sensitive molecules, resulting in an impairment of the systems. Mammalians have several enzymatic systems that function to detoxify carbonyl compounds. The aldo-keto reductase family includes enzymes that reduce carbonyl groups to alcohol using NADPH as an electron donor. Among the members of this family, aldose reductase, the AKR1B gene product, has been the most extensively studied because it is intimately involved in diabetic complications [27]. An inhibitor of aldose reductase is one of proposed cures for diabetic complications. Aldehyde reductase, the AKR1A gene product, exhibits the highest similarity to aldose reductase among the family members [28] and appears to play a coordinate function [29]. Since steroid hormones and their derivatives contain carbonyl groups and can serve as substrates for aldo-keto reductase [30], enzymes that are highly expressed in tissues with steroid hormone production may have a role in their elimination. The detoxification of carbonyls is activated by the binding of GSH, which indicates crosstalk between the GSH redox system and aldo-keto reductase system [31].
Physiological relevance of ROS/RNOS and antioxide/redox enzymes in female reproductive system
ROS and antioxidative system
Ovary is a metabolically active organ and, hence, is under a variety of stresses continuously. ROS play a physiological role during ovulation that is similar in some respects to inflammation [32,33]. Ovulation is suppressed by agents that inhibit acute inflammatory reactions [34]. Since ROS is generated during inflammatory process, it is reasonably hypothesized that ROS is released in connection with follicle rupture and is involved in the process (Figure 2). The source of ROS appears to be inflammatory cells, such as macrophages and neutrophils, as they are present in ovary at ovulation [35-37] and produce tremendous amount of free radical. This notion is supported by the finding that the suppression of ROS by SOD and/or catalase in in-vitro perfused rabbit ovary preparations hinders ovulation [38]. Monooxygenase reaction, mediated by P450, is required for the steroidogenic process that inevitably produces ROS as byproducts. ROS levels in the corpus luteum actually increase during the regression phase [39-44]. The NADPH-dependent generation of superoxide in the mouse ovary increases during the early pre-ovulatory phase in cycling females and during the luteal phase in pregnant animals [45]. Ovarian as well as uterine NADPH-dependent superoxide production appears to be LH-inducible. ROS and related compounds may function as intracellular regulators of steroidogenesis and progesterone release in the corpus luteum [41,46-48].
Figure 2 Generation of ROS during ovulation and sterodogenesis in corpus luteum. Ovulation appears to be an inflammation-like process. ROS is locally produced during follicular rupture and may be involved in the ovulation process. ROS is also generated by the corpus luteum via the monooxygenase reaction as a byproduct during steroid hormone synthesis.
SOD is present in growing follicles, the membrane granulosa of Graafian follicles, ovulated follicles, and blood vessels. Cyclic changes in SOD levels during the reproductive cycle of rats and an inverse correlation between the levels of SOD and superoxide radical have been reported [49]. SOD may play a role in regulating follicular development, ovulation, and luteal functions [50]. In the gestational corpus luteum, theca and granulosa lutein cells show strong and moderate staining intensity, respectively [51]. SOD activity is also present in human pre-ovulatory ovarian follicular fluid at higher levels than in serum [52]. About a 7-fold higher level of SOD activity is present in porcine follicular fluid and appears to exert protection against oxidative damage in oocytes [53].
SOD levels are controlled by several humoral factors and vice versa. Gonadotropoin-mediated rat follicular development coincides with an enhanced expression of Mn-SOD and EC-SOD mRNA [54]. Mn-SOD expression is induced and suppresses apoptosis in the rabbit corpus luteum in vitro, suggesting that Mn-SOD is responsible for the gonadotropin-mediated inhibition of apoptosis [55]. Both CuZn-SOD and Mn-SOD mRNA level are increased in the rat corpus luteum by prolactin [56]. However, Cu,Zn-SOD and Mn-SOD are differently regulated by estrogen and progesterone in human endometrial stromal cells. The decrease in Cu,Zn-SOD after ovarian steroid withdrawal may be involved in endometrial breakdown [57]. In case of Mn-SOD, estrogen withdrawal led to an enhanced expression of TNF-α [58], which would increase Mn-SOD mRNA levels and Mn-SOD activity in a dose-dependent manner in human endometrial stromal cells [59]. The decrease in Cu,Zn-SOD expression and the increase in lipid peroxide in the decidua may be involved in the termination of spontaneous abortion, which is mediated via the increase in PGF2α synthesis. This suggests that Cu,Zn-SOD contributes to the maintenance of pregnancy by preventing the accumulation of superoxide radicals that causes PGF2α synthesis [60]. The stimulation of luteal Cu,Zn-SOD expression by HCG may be important in maintaining luteal cell integrity when pregnancy occurs [61]. In the process of decidualization, estradiol plus medroxyprogesterone acetate increases Mn-SOD expression via a cAMP-dependent pathway. Cu,Zn-SOD is also up-regulated by these compounds, but via a different pathway from that involving cAMP [62].
Roles of redox system
In addition to antioxidation, redox systems are also well developed and protect organs against damage by oxidative stress. GSH synthesis by cumulus cells occurs during in vitro oocyte maturation in cows [63,64] and during in vivo meiotic maturation in hamsters [65,66]. Oocytes, granulosa cells, and lutein cells all express high levels of GR [67]. Because the corpus luteum produces much of the progesterone in conjunction with the reaction of P450s by consuming molecular oxygen and, hence, produces ROS as a byproduct, damage could be inflicted by ROS. Cumulus cells participate in the enhancement of GSH content in oocytes and the protection of oocytes against oxidative stress-induced apoptosis [68]. The detoxification of the produced ROS by GSH in conjunction with antioxidative enzymes would be particularly important for the corpus luteum and surrounding cells.
GSH is present in oviductal fluids and may be involved in development of mouse embryos [69]. The high levels of GR in the epithelia of the oviducts would account for this finding [67]. The secreted GSH would protect oocytes against excessively produced ROS that occurs during the ovulation, thus maintaining fertilization potency. Many in vitro studies indicate significance of antioxidants for oocyte maturation and embryo development [e.g. [70,71]].
ROS and, in consequence, carbonyl compounds can be produced by activated metabolism. Thus, detoxification by aldo-keto reductase appears to contribute to the maintenance of the genital tract. In fact, granullosa cells and the epithelia of the genital tract produce high levels of aldose reductase and aldehyde reductase [72]. The separate role of these enzymes in maintaining reproductive function is a matter of concern. Aldose reductase is an enzyme that reduces carbonyls including steroid metabolites [31] to the corresponding alcohols. It is known that aldose reductase is hormonally regulated in rat ovary during the estrous cycle [73].
Roles of RNOS
RNOS also plays multiple roles in the ovary [6]. Of the three NOS isozymes, NOS II and NOS III are expressed in the ovary [74-77]. The expression of NOS III increases after a LH surge or hCG injection. The expression of NOS III in oocytes and the blockade of oocyte maturation by the oral administration of NOS inhibitors have been reported [78]. NO generated from NOS III stimulates the ovulatory process [79-84]. Oocyte meiotic maturation is arrested in NOS III knockout mice [83,85]. However, the source of NO as it relates to oocyte maturation is currently under debate. NOS II is mainly localized in granulosa cells and produces large amounts of NO. The decrease in nitrate/nitrite concentration in preovulatory follicles after a hCG injection is correlated mainly to a decreased NOS II expression in granulosa cells [86]. NO, generated from NOS that is present in human granulosa-luteal cells, appears to inhibit estradiol secretion by directly inhibiting aromatase [87]. In addition, excess NO inhibits progesterone production and causes apoptotic cell death in rat granulose cells [86,88]. An NO-donor, S-nitroso-N-acetyl-D,L-penicillamine (SNAP), dose-dependently inhibits germinal vesicle break down in denuded oocytes, and this effect of SNAP can be reversed by the addition of hemoglobin [74]. These data suggest that the NOS II-NO-(cGMP) system may play a role in oocyte meiotic maturation, but further studies will be required to ascertain the actual function in the ovary.
ROS and RNOS in fertilization and early development of embryo
The deteriorating effects of the oxidation reaction in sperm cells have been generally discussed and overviewed [89,90]. However, oxidation reactions, in conjunction with the appropriate redox system, also exert beneficial roles. One of the most striking functions is sulfoxidation in sperm nuclei during their maturation. While ROS easily damages DNA, the regulated oxidation of sulfhydryls to disulfide in protamines is required for sperm maturation in the epididymis (Figure 3) [91]. The regulated sulfoxidation plays a role in the correct packaging of the nucleus into the small sperm head and resistance to ROS during the fertilization process, GPX3 and GPX4 present in the epididymal fluid may be responsible for the reduction of coincidently produced peroxides [92]. After fertilization, a high redox potential is required for male pronuclear formation by reducing disulfide bonds in oocytes. The origin of the reducing power appears to be GSH in the nucleus [93] because GSH present at 9–10 mM is the major source of redox potential in the oocyte [94,95]. Oocytes are also rich in glutathione reductase [69] and support the view. Since GSH alone is not effective in the reduction of disulfide bonds, cross-reactions with Trx, which has protein disulfide isomerase activity, may occur. Glycolytic activity, which generates NADH, and the hexose monophosphate shunt, a regenerating system for NADPH, are enhanced during the penetration of spermatozoa into oocytes [96], and an elevated redox potential appears to be involved in fertilization [97]. Reducing equivalents generated during the conversion of 3α-androstanediol to 5α-dihydrotestosterone has been proposed to be an alternate source of NAD(P)H [98]. The coordinate activation of these redox and NAD(P)H generating systems would enable early embryonic development to proceed.
Figure 3 Redox regulation of spermatogenesis and fertilization. During the spermatogenic process, histones are converted to protamines via transition proteins in sperm nuclei. The maturation of spermatozoa proceeds in the epididymis. Oxidation mediated by sulfoxidase is involved in the packaging of chromatin into the small nucleus via disulfide bridge formation between protamines. After fertilization, the sperm head expands to the male pronucleus by reducing the disulfide bond in the oocyte.
Peri-hatching blastocysts generate a considerably large amount of ROS for an extremely short period of time when compared to unhatched and hatched blastocysts [99]. Despite the potential importance of SOD1, knockout mice are born and grow normally. The most striking phenotype is the infertility of the SOD1-deficient female [100,101]. In spite of a precise examination, the actual cause of embryonic lethality is unknown [100]. Although both homozygous and heterozygous embryos grow normally in heterozygous females, both embryos die in homo-knockout females. This suggests the cause can be attributed to a maternal factor. Other groups have reported defects in ovary function in homo-knockout mice [101]. Since SOD is involved in the elimination of superoxide that is generated during steroidogenesis, this may be related to steroidogenesis in the ovary. Among mice that are deficient in other antioxidative enzymes, GPX4-knockout mice show premature embryonal death in the uteri, but the direct cause of this is also not clear [102,103]. A comparative study of both knockout mice may provide a clue to understanding the mechanism.
The glutathione redox system is also deeply involved in embryogenesis. Preimplanted embryos are very sensitive to conditions that cause oxidative stress. Their glutathione status changes dramatically during development [104]. GSH in reproductive tract fluid may help protect preimplanted embryos from the adverse effects of toxicants [68]. Usefulness of glutathione in embryo production has been demonstrated in culture system [105]. Increased embryonic fragmentation and a slow cleavage rate may be partially attributed to the early exposure of embryos to high ROS levels in intracytoplasmic sperm injection cycles [106]. The presence of BSO decreases GSH levels to a greater extent in the blastocyst than in the two-cell embryo [107]. GSH synthesis and turnover increase between the two-cell and blastocyst stages. The increase in the ability of embryos to synthesize GSH on day 3 is dependent on protein synthesis [108]. Hence, the recycling of GSSG must play an important role in maintenance of intracellular GSH levels from the oocyte to the two-cell stage.
The placenta is rich in aldose reductase. The presence of abundant aldose reductase in uterine luminal fluids and term placenta has been detected by two-dimensional gel electrophoresis [109]. Although aldose reductase in conjunction with sorbitol dehydrogenase catalyzes the conversion of glucose to fructose, which can be the energy source for the spermatozoa [110], the quantity of aldose reductase appears to be in excess. Both enzymes are abundant in eggs and may participate in the production of fructose [72]. Thus, aldose reductase appears to have additional roles, beyond detoxification. The production of certain cytokines, such as IL-1 and TNF, are elevated and, hence, ROS levels are also elevated. The detoxification of carbonyls present in reproductive tract fluids would be advantageous to embryos at their early developmental stage in oviducts and the uterus. Steroid metabolites such as isocorticosteoids and progesterone [30] and lipid peroxidation products such as 4-hydroxynonenal and acrolein [111,112] are all aldose reductase substrates. The glutathione conjugate of 4-hydroxy-3-nonenal actually serves as a substrate for aldose redcutase [113]. Glutathione S-transferase is known to be present in the reproductive system [114], and, hence, the presence of GSH could facilitate the detoxification function of aldose reductase by producing glutathione conjugates. The production of carbonyl compounds is caused mainly by ROS, the level of which increases during repoduction processes such as cell proliferation, steoidgenesis, and ovulation.
Conclusion
Since the production of ROS is high in reproductive tissue due to active metabolism and steroidogenesis, the tissue is under continuous oxidative stress. ROS modifies susceptible molecules including DNA, lipids, and proteins. Carrying such damage in oocytes increases the risk of hereditable disease, and, hence, living organisms must eliminate such gametes to preserve the species. On the other hand, the reproductive system utilizes ROS in some processes that are essential for reproduction. To minimize the risk caused by ROS, antioxidative systems, such as SOD and GPX have been developed. When ROS levels exceed the scavenging capacity of the system, a redox system, under such situations, can repair oxidized and damaged molecules using NADPH as an original electron source. Thus, the maintenance of a high redox potential is prerequisite for maintaining the reproductive systems in a healthy state.
Acknowledgements
This work was supported, in part, by Grant-in-Aid for Scientific Research (C) (No. 16590238) and 21st Century COE Program from the Japan Society for the Promotion of Science (JSPS).
==== Refs
Halliwell B Gutteridge JMC Free Radicals in Biology and Medicine 1998 3 Oxford: Oxford Science
Stadtman ER Protein oxidation and aging Science 1992 257 1220 1224 1355616
Fujii J Iuchi Y Matsuki S Ishii T Cooperative function of antioxidant and redox systems against oxidative stress in male reproductive tissues Asian J Androl 2003 5 231 242 12937808
Harrison R Structure and function of xanthine oxidoreductase: where are we now? Free Radic Biol Med 2002 33 774 797 12208366 10.1016/S0891-5849(02)00956-5
Baud L Ardaillou R Reactive oxygen species: production and role in the kidney Am J Physiol 1986 251 F765 F776 3022602
Rosselli M Keller PJ Dubey RK Role of nitric oxide in the biology, physiology and pathophysiology of reproduction Hum Reprod Update 1998 4 3 24 9622410 10.1093/humupd/4.1.3
Baynes JW Thorpe SR Role of oxidative stress in diabetic complications: a new perspective on an old paradigm Diabetes 1999 48 1 9 9892215
Li Y Huang TT Carlson EJ Melov S Ursell PC Olson JL Noble LJ Yoshimura MP Berger C Chan PH Wallace DC Epstein CJ Dilated cardiomyopathy and neonatal lethality in mutant mice lacking manganese superoxide dismutase Nat Genet 1995 11 376 381 7493016 10.1038/ng1295-376
Raineri I Carlson EJ Gacayan R Carra S Oberley TD Huang TT Epstein CJ Strain-dependent high-level expression of a transgene for manganese superoxide dismutase is associated with growth retardation and decreased fertility Free Radic Biol Med 2001 31 1018 1030 11595386 10.1016/S0891-5849(01)00686-4
Mruk DD Silvestrini B Mo MY Cheng CY Antioxidant superoxide dismutase – a review: its function, regulation in the testis, and role in male fertility Contraception 2002 65 305 311 12020784 10.1016/S0010-7824(01)00320-1
Ookawara T Kizaki T Takayama E Imazeki N Matsubara O Ikeda Y Suzuki K Li Ji L Tadakuma T Taniguchi N Ohno H Nuclear translocation of extracellular superoxide dismutase Biochem Biophys Res Commun 2002 296 54 61 12147226 10.1016/S0006-291X(02)00804-5
Bivalacqua TJ Armstrong JS Biggerstaff J Abdel-Mageed AB Kadowitz PJ Hellstrom WJ Champion HC Gene transfer of extracellular SOD to the penis reduces O2-• and improves erectile function in aged rats Am J Physiol Heart Circ Physiol 2003 284 H1408 H1421 12505874
Carlsson LM Jonsson J Edlund T Marklund SL Mice lacking extracellular superoxide dismutase are more sensitive to hyperoxia Proc Natl Acad Sci USA 1995 92 6264 6268 7603981
Ho YS Magnenat JL Bronson RT Cao J Gargano M Sugawara M Funk CD Mice deficient in cellular glutathione peroxidase develop normally and show no increased sensitivity to hyperoxia J Biol Chem 1997 272 16644 16651 9195979 10.1074/jbc.272.26.16644
Hansen JC Deguchi Y Selenium and fertility in animals and man – a review Acta Vet Scand 1996 37 19 30 8659343
Ursini F Heim S Kiess M Maiorino M Roveri A Wissing J Flohe L Dual function of the selenoprotein PHGPx during sperm maturation Science 1999 285 1393 1396 10464096 10.1126/science.285.5432.1393
Imai H Suzuki K Ishizaka K Ichinose S Oshima H Okayasu I Emoto K Umeda M Nakagawa Y Failure of the expression of phospholipid hydroperoxide glutathione peroxidase in the spermatozoa of human infertile males Biol Reprod 2001 64 674 683 11159372
Maiorino M Bosello V Ursini F Foresta C Garolla A Scapin M Sztajer H Flohe L Genetic variations of gpx-4 and male infertility in humans Biol Reprod 2003 68 1134 1141 12606444 10.1095/biolreprod.102.007500
Pfeifer H Conrad M Roethlein D Kyriakopoulos A Brielmeier M Bornkamm GW Behne D Identification of a specific sperm nuclei selenoenzyme necessary for protamine thiol cross-linking during sperm maturation FASEB J 2001 15 1236 1238 11344099
Borchert A Savaskan NE Kuhn H Regulation of expression of the phospholipid hydroperoxide/sperm nucleus glutathione peroxidase gene. Tissue-specific expression pattern and identification of functional cis- and trans-regulatory elements J Biol Chem 2003 278 2571 2580 12427732 10.1074/jbc.M209064200
Fujii J Ikeda Y Advances in our understanding of peroxiredoxin, a multifunctional, mammalian redox protein Redox Rep 2002 7 123 130 12189041 10.1179/135100002125000352
Biteau B Labarre J Toledano MB ATP-dependent reduction of cysteine-sulphinic acid by S. cerevisiae sulphiredoxin Nature 2003 425 980 984 14586471 10.1038/nature02075
Meister A Glutathione biosynthesis and its inhibition Methods Enzymol 1995 252 26 30 7476360
Schallreuter KU Wood JM New aspects in the pathophysiology of cutaneous melanoma: a review of the role of thioproteins and the effect of nitrosoureas Melanoma Res 1991 1 159 167 1841712
Nakamura H Nakamura K Yodoi J Redox regulation of cellular activation Annu Rev Immunol 1997 15 351 369 9143692 10.1146/annurev.immunol.15.1.351
Matsui M Oshima M Oshima H Takaku K Maruyama T Yodoi J Taketo MM Early embryonic lethality caused by targeted disruption of the mouse thioredoxin gene Dev Biol 1996 178 179 185 8812119 10.1006/dbio.1996.0208
Yabe-Nishimura C Aldose reductase in glucose toxicity: a potential target for the prevention of diabetic complications Pharmacol Rev 1998 50 21 33 9549756
Jez JM Bennett MJ Schlegel BP Lewis M Penning TM Comparative anatomy of the aldo-keto reductase superfamily Biochem J 1997 326 625 636 9307009
Iuchi Y Kaneko T Matsuki S Ishii T Ikeda Y Uchida K Fujii J Carbonyl stress and detoxification ability in the male genital tract and testis of rats Histochem Cell Biol 2004 121 123 130 14673659 10.1007/s00418-003-0607-3
Wermuth B Monder C Aldose and aldehyde reductase exhibit isocorticosteroid reductase activity Eur J Biochem 1983 131 423 426 6403351 10.1111/j.1432-1033.1983.tb07280.x
Dixit BL Balendiran GK Watowich SJ Srivastava S Raman KV Petrash JM Bhatnagar A Srivastava SK Kinetic and structural characterization of the glutathione-binding site of aldose reductase J Biol Chem 2000 275 21587 1595 10764810 10.1074/jbc.M909235199
Bjersing L Cajander S Ovulation and the mechanism of follicle rupture. VI. Ultrastructure of theca interna and the inner vascular network surrounding rabbit graafian follicles prior to induced ovulation Cell Tissue Res 1974 153 31 44 4442080
Espey LL Ovulation as an inflammatory reaction – a hypothesis Biol Reprod 1980 22 73 106 6991013
Espey LL Stein VI Dumitrescu J Survey of antiinflammatory agents and related drugs as inhibitors of ovulation in the rabbit Fertil Steril 1982 38 238 247 6125417
Nakamura Y Smith M Krishna A Terranova PF Increased number of mast cells in the dominant follicle of the cow: relationships among luteal, stromal, and hilar regions Biol Reprod 1987 37 546 549 3676403
Adashi EY The potential relevance of cytokines to ovarian physiology: the emerging role of resident ovarian cells of the white blood cell series Endocr Rev 1990 11 454 464 2226351
Brannstrom M Mayrhofer G Robertson SA Localization of leukocyte subsets in the rat ovary during the periovulatory period Biol Reprod 1993 48 277 286 8439617
Miyazaki T Sueoka K Dharmarajan AM Atlas SJ Bulkley GB Wallach EE Effect of inhibition of oxygen free radical on ovulation and progesterone production by the in-vitro perfused rabbit ovary J Reprod Fertil 1991 91 207 212 1995849
Riley JC Behrman HR In vivo generation of hydrogen peroxide in the rat corpus luteum during luteolysis Endocrinology 1991 128 1749 1753 2004600
Sawada M Carlson JC Rapid plasma membrane changes in superoxide radical formation, fluidity, and phospholipase A2 activity in the corpus luteum of the rat during induction of luteolysis Endocrinology 1991 128 2992 2998 1645257
Sawada M Carlson JC Studies on the mechanism controlling generation of superoxide radical in luteinized rat ovaries during regression Endocrinology 1994 135 1645 1650 7925128 10.1210/en.135.4.1645
Sugino N Nakamura Y Takeda O Ishimatsu M Kato H Changes in activities of superoxide dismutase and lipid peroxide in corpus luteum during pregnancy in rats J Reprod Fertil 1993 97 347 351 8501704
Sugino N Telleria CM Gibori G Differential regulation of copper-zinc superoxide dismutase and manganese superoxide dismutase in the rat corpus luteum: induction of manganese superoxide dismutase messenger ribonucleic acid by inflammatory cytokines Biol Reprod 1998 59 208 215 9675014
Shimamura K Sugino N Yoshida Y Nakamura Y Ogino K Kato H Changes in lipid peroxide and antioxidant enzyme activities in corpora lutea during pseudopregnancy in rats J Reprod Fertil 1995 105 253 257 8568768
Jain S Saxena D Kumar GP Laloraya M NADPH dependent superoxide generation in the ovary and uterus of mice during estrous cycle and early pregnancy Life Sci 2000 66 1139 1146 10737364 10.1016/S0024-3205(00)00417-3
Carlson JC Wu XM Sawada M Oxygen radicals and the control of ovarian corpus luteum function Free Radic Biol Med 1993 14 79 84 8454226 10.1016/0891-5849(93)90511-R
Carlson JC Sawada M Boone DL Stauffer JM Stimulation of progesterone secretion in dispersed cells of rat corpora lutea by antioxidants Steroids 1995 60 272 276 7792831 10.1016/0039-128X(94)00053-F
Sawada M Carlson JC Intracellular regulation of progesterone secretion by the superoxide radical in the rat corpus luteum Endocrinology 1996 137 1580 1584 8612488 10.1210/en.137.5.1580
Laloraya M Pradeep KG Laloraya MM Changes in the levels of superoxide anion radical and superoxide dismutase during the estrous cycle of rattus norvegicus and induction of superoxide dismutase in rat ovary by lutropin Biochem Biophys Res Commun 1988 157 146 153 2848516 10.1016/S0006-291X(88)80025-1
Laloraya M Kumar GP Laloraya MM Histochemical study of superoxide dismutase in the ovary of the rat during the oestrous cycle J Reprod Fertil 1989 86 583 587 2760887
Shiotani M Noda Y Narimoto K Imai K Mori T Fujimoto K Ogawa K Immunohistochemical localization of superoxide dismutase in the human ovary Hum Reprod 1991 6 1349 1353 1770125
Sabatini L Wilson C Lower A Al-Shawaf T Grudzinskas JG Superoxide dismutase activity in human follicular fluid after controlled ovarian hyperstimulation in women undergoing in vitro fertilization Fertil Steril 1999 72 1027 1034 10593376 10.1016/S0015-0282(99)00411-2
Tatemoto H Muto N Sunagawa I Shinjo A Nakada T Protection of porcine oocytes against cell damage caused by oxidative stress during in vitro maturation: role of superoxide dismutase activity in porcine follicular fluid Biol Reprod 2004 71 1150 1157 15175235 10.1095/biolreprod.104.029264
Tilly JL Tilly KI Inhibitors of oxidative stress mimic the ability of follicle-stimulating hormone to suppress apoptosis in cultured rat ovarian follicles Endocrinology 1995 136 242 252 7828537 10.1210/en.136.1.242
Dharmarajan AM Hisheh S Singh B Parkinson S Tilly KI Tilly JL Antioxidants mimic the ability of chorionic gonadotropin to suppress apoptosis in the rabbit corpus luteum in vitro: a novel role for superoxide dismutase in regulating bax expression Endocrinology 1999 140 2555 2561 10342842 10.1210/en.140.6.2555
Sugino N Hirosawa-Takamori M Zhong L Telleria CM Shiota K Gibori G Hormonal regulation of copper-zinc superoxide dismutase and manganese superoxide dismutase messenger ribonucleic acid in the rat corpus luteum: induction by prolactin and placental lactogens Biol Reprod 1998 59 599 605 9716559
Sugino N Karube-Harada A Sakata A Takiguchi S Kato H Different mechanisms for the induction of copper-zinc superoxide dismutase and manganese superoxide dismutase by progesterone in human endometrial stromal cells Hum Reprod 2002 17 1709 1714 12093828 10.1093/humrep/17.7.1709
Tabibzadeh S Satyaswaroop PG von Wolff M Strowitzki T Regulation of TNF-alpha mRNA expression in endometrial cells by TNF-alpha and by oestrogen withdrawal Mol Hum Reprod 1999 5 1141 1149 10587369 10.1093/molehr/5.12.1141
Karube-Harada A Sugino N Kashida S Takiguchi S Takayama H Yamagata Y Nakamura Y Kato H Induction of manganese superoxide dismutase by tumour necrosis factor-alpha in human endometrial stromal cells Mol Hum Reprod 2001 7 1065 1072 11675473 10.1093/molehr/7.11.1065
Sugino N Nakata M Kashida S Karube A Takiguchi S Kato H Decreased superoxide dismutase expression and increased concentrations of lipid peroxide and prostaglandin F(2alpha) in the decidua of failed pregnancy Mol Hum Reprod 2000 6 642 647 10871652 10.1093/molehr/6.7.642
Sugino N Takiguchi S Kashida S Karube A Nakamura Y Kato H Superoxide dismutase expression in the human corpus luteum during the menstrual cycle and in early pregnancy Mol Hum Reprod 2000 6 19 25 10611256 10.1093/molehr/6.1.19
Sugino N Karube-Harada A Kashida S Takiguchi S Kato H Differential regulation of copper-zinc superoxide dismutase and manganese superoxide dismutase by progesterone withdrawal in human endometrial stromal cells Mol Hum Reprod 2002 8 68 74 11756571 10.1093/molehr/8.1.68
de Matos DG Furnus CC Moses DF Martinez AG Matkovic M Stimulation of glutathione synthesis of in vitro matured bovine oocytes and its effect on embryo development and freezability Mol Reprod Dev 1996 45 451 457 8956283 10.1002/(SICI)1098-2795(199612)45:4<451::AID-MRD7>3.0.CO;2-Q
Zuelke KA Jones DP Perreault SD Glutathione oxidation is associated with altered microtubule function and disrupted fertilization in mature hamster oocytes Biol Reprod 1997 57 1413 1419 9408248
de Matos DG Furnus CC Moses DF Glutathione synthesis during in vitro maturation of bovine oocytes: role of cumulus cells Biol Reprod 1997 57 1420 1425 9408249
Zuelke KA Jeffay SC Zucker RM Perreault SD Glutathione (GSH) concentrations vary with the cell cycle in maturing hamster oocytes, zygotes, and pre-implantation stage embryos Mol Reprod Dev 2003 64 106 112 12420305 10.1002/mrd.10214
Kaneko T Iuchi Y Kawachiya S Fujii T Saito H Kurachi H Fujii J Alteration of glutathione reductase expression in the female reproductive organs during the estrous cycle Biol Reprod 2001 65 1410 1416 11673257
Tatemoto H Sakurai N Muto N Protection of porcine oocytes against apoptotic cell death caused by oxidative stress during In vitro maturation: role of cumulus cells Biol Reprod 2000 63 805 810 10952924
Gardiner CS Salmen JJ Brandt CJ Stover SK Glutathione is present in reproductive tract secretions and improves development of mouse embryos after chemically induced glutathione depletion Biol Reprod 1998 59 431 436 9687318
Ikeda S Kitagawa M Imai H Yamada M The roles of vitamin A for cytoplasmic maturation of bovine oocytes J Reprod Dev 2005 51 23 35 15750294 10.1262/jrd.51.23
Dalvit G Llanes SP Descalzo A Insani M Beconi M Cetica P Effect of alpha-tocopherol and ascorbic acid on bovine oocyte in vitro maturation Reprod Domest Anim 2005 40 93 97 15819954 10.1111/j.1439-0531.2004.00522.x
Kaneko T Iuchi Y Takahashi M Fujii J Colocalization of polyol-metabolizing enzymes and immunological detection of fructated proteins in the female reproductive system of the rat Histochem Cell Biol 2003 119 309 315 12684816
Iwata N Hara S Nishimura C Takahashi M Mukai T Takayama M Endo T Hormonal regulation of aldose reductase in rat ovary during the estrous cycle Eur J Biochem 1996 235 444 448 8631365 10.1111/j.1432-1033.1996.00444.x
Nakamura Y Yamagata Y Sugino N Takayama H Kato H Nitric oxide inhibits oocyte meiotic maturation Biol Reprod 2002 67 1588 1592 12390892 10.1095/biolreprod.102.005264
Van Voorhis BJ Moore K Strijbos PJ Nelson S Baylis SA Grzybicki D Weiner CP Expression and localization of inducible and endothelial nitric oxide synthase in the rat ovary. Effects of gonadotropin stimulation in vivo J Clin Invest 1995 96 2719 2726 8675639
Jablonka-Shariff A Olson LM Hormonal regulation of nitric oxide synthases and their cell-specific expression during follicular development in the rat ovary Endocrinology 1997 138 460 468 8977436 10.1210/en.138.1.460
Zackrisson U Mikuni M Wallin A Delbro D Hedin L Brannstrom M Cell-specific localization of nitric oxide synthases (NOS) in the rat ovary during follicular development, ovulation and luteal formation Hum Reprod 1996 11 2667 2673 9021370
Jablonka-Shariff A Basuray R Olson LM Inhibitors of nitric oxide synthase influence oocyte maturation in rats J Soc Gynecol Invest 1999 6 95 101 10.1016/S1071-5576(98)00053-7
Shukovski L Tsafriri A The involvement of nitric oxide in the ovulatory process in the rat Endocrinology 1994 135 2287 2290 7525265 10.1210/en.135.5.2287
Bonello N McKie K Jasper M Andrew L Ross N Braybon E Brannstrom M Norman RJ Inhibition of nitric oxide: effects on interleukin-1 beta-enhanced ovulation rate, steroid hormones, and ovarian leukocyte distribution at ovulation in the rat Biol Reprod 1996 54 436 445 8788197
Yamauchi J Miyazaki T Iwasaki S Kishi I Kuroshima M Tei C Yoshimura Y Effects of nitric oxide on ovulation and ovarian steroidogenesis and prostaglandin production in the rabbit Endocrinology 1997 138 3630 3637 9275046 10.1210/en.138.9.3630
Hesla JS Preutthipan S Maguire MP Chang TS Wallach EE Dharmarajan AM Nitric oxide modulates human chorionic gonadotropin-induced ovulation in the rabbit Fertil Steril 1997 67 548 552 9091345 10.1016/S0015-0282(97)80084-2
Jablonka-Shariff A Olson LM The role of nitric oxide in oocyte meiotic maturation and ovulation: meiotic abnormalities of endothelial nitric oxide synthase knock-out mouse oocytes Endocrinology 1998 139 2944 2954 9607805 10.1210/en.139.6.2944
Jablonka-Shariff A Ravi S Beltsos AN Murphy LL Olson LM Abnormal estrous cyclicity after disruption of endothelial and inducible nitric oxide synthase in mice Biol Reprod 1999 61 171 177 10377046
Jablonka-Shariff A Olson LM Nitric oxide is essential for optimal meiotic maturation of murine cumulus-oocyte complexes in vitro Mol Reprod Dev 2000 55 412 421 10694749 10.1002/(SICI)1098-2795(200004)55:4<412::AID-MRD9>3.0.CO;2-W
Yamagata Y Nakamura Y Sugino N Harada A Takayama H Kashida S Kato H Alterations in nitrate/nitrite and nitric oxide synthase in preovulatory follicles in gonadotropin-primed immature rat Endocr J 2002 49 219 226 12081242
Van Voorhis BJ Dunn MS Snyder GD Weiner CP Nitric oxide: an autocrine regulator of human granulosa-luteal cell steroidogenesis Endocrinology 1994 135 1799 1806 7525252 10.1210/en.135.5.1799
Dave S Farrance DP Whitehead SA Evidence that nitric oxide inhibits steroidogenesis in cultured rat granulosa cells Clin Sci 1997 92 277 284 9093008
Agarwal A Saleh RA Bedaiwy MA Role of reactive oxygen species in the pathophysiology of human reproduction Fertil Steril 2003 79 829 843 12749418 10.1016/S0015-0282(02)04948-8
Sanocka D Kurpisz M Reactive oxygen species and sperm cells Reprod Biol Endocrinol 2004 2 12 18 15038829 10.1186/1477-7827-2-12
Perreault SD Naish SJ Zirkin BR The timing of hamster sperm nuclear decondensation and male pronucleus formation is related to sperm nuclear disulfide bond content Biol Reprod 1987 36 239 244 3567277
Fouchecourt S Metayer S Locatelli A Dacheux F Dacheux JL Stallion epididymal fluid proteome: qualitative and quantitative characterization; secretion and dynamic changes of major proteins Biol Reprod 2000 62 1790 1803 10819784
Yoshida M Ishigaki K Nagai T Chikyu M Pursel VG Glutathione concentration during maturation and after fertilization in pig oocytes: relevance to the ability of oocytes to form male pronucleus Biol Reprod 1993 49 89 94 8353194
Perreault SD Wolff RA Zirkin BR The role of disulfide bond reduction during mammalian sperm nuclear decondensation in vivo Dev Biol 1984 101 160 167 6692970 10.1016/0012-1606(84)90126-X
Perreault SD Barbee RR Slott VL Importance of glutathione in the acquisition and maintenance of sperm nuclear decondensing activity in maturing hamster oocytes Dev Biol 1988 125 181 186 3334716 10.1016/0012-1606(88)90070-X
Urner F Sakkas D Characterization of glycolysis and pentose phosphate pathway activity during sperm entry into the mouse oocyte Biol Reprod 1999 60 973 978 10084974
Urner F Sakkas D Involvement of the pentose phosphate pathway and redox regulation in fertilization in the mouse Mol Reprod Dev 2005 70 494 503 15685628 10.1002/mrd.20222
Chapman JC Michael SD Proposed mechanism for sperm chromatin condensation/decondensation in the male rat Reprod Biol Endocrinol 2003 1 20 26 12646056 10.1186/1477-7827-1-20
Thomas M Jain S Kumar GP Laloraya M A programmed oxyradical burst causes hatching of mouse blastocysts J Cell Sci 1997 110 1597 1602 9247193
Ho YS Gargano M Cao J Bronson RT Heimler I Hutz RJ Reduced fertility in female mice lacking copper-zinc superoxide dismutase J Biol Chem 1998 273 7765 7769 9516486 10.1074/jbc.273.13.7765
Matzuk MM Dionne L Guo Q Kumar TR Lebovitz RM Ovarian function in superoxide dismutase 1 and 2 knockout mice Endocrinology 1998 139 4008 4011 9724058 10.1210/en.139.9.4008
Imai H Hirao F Sakamoto T Sekine K Mizukura Y Saito M Kitamoto T Hayasaka M Hanaoka K Nakagawa Y Early embryonic lethality caused by targeted disruption of the mouse PHGPx gene Biochem Biophys Res Commun 2003 305 278 286 12745070 10.1016/S0006-291X(03)00734-4
Yant LJ Ran Q Rao L Van Remmen H Shibatani T Belter JG Motta L Richardson A Prolla TA The selenoprotein GPX4 is essential for mouse development and protects from radiation and oxidative damage insults Free Radic Biol Med 2003 34 496 502 12566075 10.1016/S0891-5849(02)01360-6
Gardiner CS Reed DJ Status of glutathione during oxidant-induced oxidative stress in the preimplantation mouse embryo Biol Reprod 1994 51 1307 1314 7888510
Luvoni GC Keskintepe L Brackett BG Improvement in bovine embryo production in vitro by glutathione-containing culture media Mol Reprod Dev 1996 43 437 443 9052934 10.1002/(SICI)1098-2795(199604)43:4<437::AID-MRD5>3.0.CO;2-Q
Bedaiwy MA Falcone T Mohamed MS Aleem AA Sharma RK Worley SE Thornton J Agarwal A Differential growth of human embryos in vitro: role of reactive oxygen species Fertil Steril 2004 82 593 600 15374701 10.1016/j.fertnstert.2004.02.121
Gardiner CS Reed DJ Glutathione redox cycle-driven recovery of reduced glutathione after oxidation by tertiary-butyl hydroperoxide in preimplantation mouse embryos Arch Biochem Biophys 1995 321 6 12 7639536 10.1006/abbi.1995.1361
Gardiner CS Reed DJ Synthesis of glutathione in the preimplantation mouse embryo Arch Biochem Biophys 1995 318 30 36 7726569 10.1006/abbi.1995.1200
Lee SR Kwon KS Kim SR Rhee SG Reversible inactivation of protein-tyrosine phosphatase 1B in A431 cells stimulated with epidermal growth factor J Biol Chem 1998 273 15366 15372 9624118 10.1074/jbc.273.25.15366
Leese HJ Astley NR Lambert D Glucose and fructose utilization by rat spermatozoa within the uterine lumen J Reprod Fertil 1981 61 435 437 7205789
Kolb NS Hunsaker LA Vander Jagt DL Aldose reductase-catalyzed reduction of acrolein: implications in cyclophosphamide toxicity Mol Pharmacol 1994 45 797 801 8183257
Vander Jagt DL Kolb NS Vander Jagt TJ Chino J Martinez FJ Hunsaker LA Royer RE Substrate specificity of human aldose reductase: identification of 4-hydroxynonenal as an endogenous substrate Biochim Biophys Acta 1995 1249 117 126 7599164
Srivastava S Chandra A Wang LF Seifert WE JrDaGue BB Ansari NH Srivastava SK Bhatnagar A Metabolism of the lipid peroxidation product, 4-hydroxy-trans-2-nonenal, in isolated perfused rat heart J Biol Chem 1998 273 10893 19000 9556565 10.1074/jbc.273.18.10893
Knapen MF Zusterzeel PL Peters WH Steegers EA Glutathione and glutathione-related enzymes in reproduction. A review Eur J Obstet Gynecol Reprod Biol 1999 82 171 184 10206412 10.1016/S0301-2115(98)00242-5
|
16137335
|
PMC1224869
|
CC BY
|
2021-01-04 16:37:12
|
no
|
Reprod Biol Endocrinol. 2005 Sep 2; 3:43
|
utf-8
|
Reprod Biol Endocrinol
| 2,005 |
10.1186/1477-7827-3-43
|
oa_comm
|
==== Front
RetrovirologyRetrovirology1742-4690BioMed Central London 1742-4690-2-501609296210.1186/1742-4690-2-50ResearchUse of Endogenous Retroviral Sequences (ERVs) and structural markers for retroviral phylogenetic inference and taxonomy Jern Patric [email protected] Göran O [email protected] Jonas [email protected] Section of Virology, Department of Medical Sciences, Uppsala University, Uppsala, Sweden2 Unit of Physiology, Department of Neuroscience, Uppsala University, Uppsala, Sweden2005 10 8 2005 2 50 50 14 7 2005 10 8 2005 Copyright © 2005 Jern et al; licensee BioMed Central Ltd.2005Jern et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Endogenous retroviral sequences (ERVs) are integral parts of most eukaryotic genomes and vastly outnumber exogenous retroviruses (XRVs). ERVs with a relatively complete structure were retrieved from the genetic archives of humans and chickens, diametrically opposite representatives of vertebrate retroviruses (over 3300 proviruses), and analyzed, using a bioinformatic program, RetroTector©, developed by us. This rich source of proviral information, accumulated in a local database, and a collection of XRV sequences from the literature, allowed the reconstruction of a Pol based phylogenetic tree, more extensive than previously possible. The aim was to find traits useful for classification and evolutionary studies of retroviruses. Some of these traits have been used by others, but they are here tested in a wider context than before.
Results
In the ERV collection we found sequences similar to the XRV-based genera: alpha-, beta-, gamma-, epsilon- and spumaretroviruses. However, the occurrence of intermediates between them indicated an evolutionary continuum and suggested that taxonomic changes eventually will be necessary. No delta or lentivirus representatives were found among ERVs. Classification based on Pol similarity is congruent with a number of structural traits. Acquisition of dUTPase occurred three times in retroviral evolution. Loss of one or two NC zinc fingers appears to have occurred several times during evolution. Nucleotide biases have been described earlier for lenti-, delta- and betaretroviruses and were here confirmed in a larger context.
Conclusion
Pol similarities and other structural traits contribute to a better understanding of retroviral phylogeny. "Global" genomic properties useful in phylogenies are i.) translational strategy, ii.) number of Gag NC zinc finger motifs, iii.) presence of Pro N-terminal dUTPase (dUTPasePro), iv.) presence of Pro C-terminal G-patch and v.) presence of a GPY/F motif in the Pol integrase (IN) C-terminal domain. "Local" retroviral genomic properties useful for delineation of lower level taxa are i.) host species range, ii.) nucleotide compositional bias and iii.) LTR lengths.
==== Body
Background
Retroviral and related endogenous retroviral sequences (ERVs) are integral parts of most eukaryotic genomes, sometimes constituting over 50% of them [1]. Their ability to transpose and transfer horizontally [2,3], confers genetic flexibility to complex genomes like those of humans [4], chimpanzees [5], other primates and vertebrates.
The origin of retroviruses is lost in a prebiotic mist. Assuming a 0.2% neutral substitution rate per million years [6] and a 50% divergence limit for nucleotide sequence recognition, retroviral sequences >250 Million years old cannot be found in current genomes. If any of their genes are selected for, they may stay recognizable longer. Thus, although the ERV record has limitations, the reconstruction of retrovirus evolution differs fundamentally from that of other viruses, due to the ERVs in the ever richer archive of genomic assemblies. According to the VIIth ICTV report [7], Retroviridae borders to Pararetroviridae (e.g. Hepatitis B), Metaviridae (Gypsy-like) and Pseudoviridae (Copia-like). Together with the even more more distant relatives Mal-R [8], DIRS [9] retrotransposons and chromoviruses [10], not included here, they show that retroviruses are parts of a vast retrotransposon sequence universe. In this work, we concentrated on retroviruses. An ancestral retrovirus likely had structural traits which at present are common denominators of the diverse related sequences. Although some structural traits may be absent in individual viruses, readily identifiable common denominators are 5'LTR, PBS, Gag (MA, CA and NC), Pro, Pol, Env, PPT and 3'LTR [11]. The most universal trait is the pol gene, with its reverse transcriptase (RT), RNAse H and integrase (IN). The use of other conserved but distinguishing traits in phylogenetic inference and retroviral classification discussed here are: nucleotide bias, number of zinc fingers, translational strategy, C-terminal Pro and Pol motifs, presence of dUTPase and accessory genes and LTR length. Env is an unreliable evolutionary marker, exemplified by the hybrid betaretroviral MPMV [11], but can be useful in narrow phylogenies to demarcate a specific group.
Retroviral taxonomy has traditionally been based on observed phenotypic qualities of exogenous retroviruses (XRVs) [7]. Classification using ERVs, with an almost complete lack of phenotypic information, necessitates a nucleotide sequence analytical approach. Seven retroviral genera have been described (alpha-, beta-, gamma-, delta-, epsilon-, lenti- and spuma-like retroviruses) using sequence similarities, mainly in the Pol RT region. Although much work remains before all ERVs are fully characterized, ERVs have also been divided into loosely defined classes, originally based on HERVs [12-14]. When analyzing the RT region, the gammaretroviruses cluster as class I and betaretroviruses as class II elements [12]. The spuma- and spumalike elements group within the class III [14]. Lenti- and deltaretroviruses have no known endogenous counterparts [15]. This was also the case in our computerized genomewide screenings (see below).
ERV classification and grouping originally was based on sequence similarity between the proviral PBS and the host tRNA [11]. This classification has proved useful for some ERVs, e.g. HERV-E [16] and mostly for HERV-H [17]. However, it is inconsistent for many other ERV groups that have alternative PBSes [18] e.g. HERV-H/F [17], ERV3 [16], and ERV9/HERV-W [19]. We did not extend these analyses here.
In several papers [[17,20] and Jern et al. submitted], we have used Pol similarity for ERV classification. Pol is highly conserved, and its large size (800–1100 aa) provides adequate information for a relatively detailed classification. This is facilitated by the program RetroTector© [Sperber G.O. et al. in preparation], which reconstructs probable Pol proteins ("puteins") from different reading frames in the often damaged gene candidates. The puteins are favored over nucleotide sequences since they are more conserved, easier to align and therefore allow phylogenetic inference and taxonomy over greater evolutionary distances. This is further discussed in the Methods and Results sections of this paper. A number of reliable distinguishing features must be defined to enable a durable retroviral taxonomy which can encompass the many new ERVs and XRVs, and to trace their evolution. In this study, we compared phylogenetic trees, based on Pol similarity, with distinct structural features of possible use as taxonomic and phylogenetic markers.
Results and Discussion
Genomic ERV collection
Using the program RetroTector© (see methods), we screened the human hg16 [4] and chicken gg01 [21] genomes for ERVs. We found them to encompass 3149 and 260 proviral sequences with a RetroTector© score of more than 300, respectively. A detailed account will be published separately [Blomberg J. et al. in preparation]. Based on experience from randomized data set scores (data not shown), this threshold separated false from true retroviral elements with a wide margin. We collected the sequences into an ERV databank, from which we extracted representative sequences for use in matching structural traits against sequence similarity based phylogenetic inference. Sequences scoring over 300 from the hg16 and gg01 genomes were analyzed for the presence of Pol. Those with a recognizable Pol were grouped into respective genera according to sequence similarity (Table 1). ERVs were found in all retroviral genera, except lenti- and deltaretroviruses. Our bioinformatic screening of a larger dataset thus confirmed the results of Herniou et al. [15]. As genomic assemblies from more species become available, analysis of upcoming retroviral sequences will increase the precision of phylogenetic inference and retroviral taxonomy.
Table 1 Detected ERV structural traits in the human and chicken genomes
Genome(s)1 Genus Class ERVs2 Gag-Pro f.s.3 Pro-Pol f.s.3 dUTPase4 C-term. motifs5
(ERV) -1 0 1 -1 0 1 (dUTPasePro) G-patch GPY/F
gg01 alpha 34 4 1 1 10 2 1 0 0 0
gg01 alpha-beta 67 5 2 0 5 3 1 21 0 0
gg01 and hg16 beta II 582 49 18 14 50 22 27 363 68 0
gg01 and hg16 gamma I 2069 14 55 13 32 64 43 0 0 264
gg01 and hg16 delta (-)6 (-) (-) (-) (-) (-) (-) (-) (-) (-)
gg01 and hg16 epsilon n.d.7 n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d.
gg01 and hg16 lenti (-) (-) (-) (-) (-) (-) (-) (-) (-) (-)
gg01 and hg16 spuma-like III 193 n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d.
1 Human genome version 16 (hg16) and Chicken genome (gg01). Alpha and alpha-beta retroviruses were only detected in the chicken genome
2 Detected ERVs from the databank collection, with RetroTector© score >300 and recognized Pol puteins.
3 Predicted frameshift (f.s.) translation strategy between the respective putein ORFs in elements with score >1000, and no f.s. in the near 3'-end.
4 Detected dUTPase N-terminal of protease domain (dUTPasePro) in elements with score >300.
5 Detected C-terminal Pro (G-patch) and Pol (GPY/F) motifs elements with score >300.
6 (-): no delta or lentiviral ERVs were detected in the genomes.
7 (n.d.): not determined. The scarcity of epsilon like elements [19] and the highly mutated nature of both epsilon and spuma-like elements, in the human genome did not provide sufficient information to conduct a thorough analysis.
Phylogenetic reconstruction based on Pol
Using the whole Pol proteins/puteins retrieved from the genetic archive, we reconstructed an unrooted retroviral neighbor joining (NJ) tree. We used the whole Pol and the principle of pairwise deletions in the alignment and distance matrix analyses to avoid problems with missing portions in RT as in e.g. the large HERV-H group [17]. To reconstruct a comprehensible condensed phylogeny, we chose to include only 12 representative ERVs from the human hg16 [4], chimpanzee pt01 [5] (collected and analyzed earlier [Jern et al. submitted]) and chicken gg01 [21] genome assemblies. The human and chicken sequences were chosen because they are diametrically opposite representatives of vertebrate retroviruses (including over 3300 proviruses). The representative ERVs from chimpanzee (one BaEV like and one PTERV1 like [3], not found in humans [Jern et al. submitted]) were included to provide a broad based phylogeny. The remaining representative ERVs, all with high RetroTector© scores, were selected to contribute with different aspects, e.g. intermediate positions in the tree, while still keeping the size of the Pol tree manageable. Annotated exogenous retroviruses retrieved from GenBank were added to form a tree backbone structure useful for taxonomic reference (Figure 1). The Pol NJ (500 bootstrap consensus) tree structure was confirmed using an array of maximum likelihood (ML) analyses (data not shown). The wealth of mutated ERV sequences sometimes makes delineation of genera and groups difficult. We have earlier used the Pol similarity (>80%) based clustering as a primary criterion for retroviral groups [16,17]. This corresponds to the finer branches of the genera (Figure 1), which themselves tend to have internal Pol similarities over 60% (Additional file 1). The phylogenetic Pol tree shows the seven retroviral genera, defined from clustering of ERVs next to the earlier classified XRVs (see bootstrap supports in NJ tree, additional file 1), and the three loosely defined ERV classes [12-14] (Figure 1). Further, the tree shows two major branches, ending in gamma- and betaretroviruses, respectively. They consistently have very high bootstrap supports (Additional file 1). The continuous influx of new data will eventually necessitate a revision of the retroviral genera. This was out of scope for the present study. An especially amorphous part of the tree is its center. In numerous phylogenetic analyses with a sequence set (not shown here), we found that the spuma-like group referred to here, includes both the exogenous spumaviruses and a diverse group of related endogenous retroviral sequences (primarily ERV-L). These and other centrally located elements often are highly mutated and difficult to analyze. Further, the tree (Figure 1) shows ERV and XRV sequences intermediate between the major genera. In the left major branch (the main "gamma" branch), Snakehead retrovirus (SnRV) is intermediate between epsilon and spumalike retroviruses. In the main "beta-branch", several chicken ERVs and the reptilian Python RV [22] are intermediate between the previously recognized delta, lenti, alpha and betaretroviruses, supporting a gradual evolution of betaretroviruses from delta/lenti and alpharetrovirus-like ancestors.
Figure 1 Representative unrooted Pol neighbor joining (NJ) dendrogram. Unrooted Pol neighbor joining (NJ) dendrogram (500 bootstraps consensus) of the seven retroviral genera: alpha-, beta-, gamma-, delta-, epsilon-, lenti- and spuma-like retroviruses. The somewhat more loosely defined (endogenous) retroviral classes are indicated in the periphery. The various host species are indicated with symbols next to each taxonomic unit. The novel sequences are named according to their chromosomal positions within respective genomes. (hg15 and 16: Human genome; gg01: Chicken genome and pt01: chimpanzee genome). The two pt01 sequences were unique to chimpanzee and not found in humans [Jern et al. submitted].
Host species
Although host species is not a structural feature, it is an easily definable trait, and is therefore discussed here. Retroviral classification using host species is at first sight appealing: Classical gammaretroviruses are murine, epsilon piscine, alpha avian and beta mammalian. However, as seen in figure 1, this order is not maintained when additional XRVs and ERVs are included. It has been shown that some avian retroviruses share similarity with human gammaretroviral (class I) HERV-I elements [23], and probably are the results of horizontal transfers [15]. In our screening, we confirmed these avian HERV-I like elements and also show a novel avian sequence extracted from the chicken genome that is similar to HERV-E (Figure 1). Further, it has been shown that piscine elements grouped together with some human elements [15]. In a recent bioinformatic study, we found human epsilon-like proviral elements [19]. One of them was included into the phylogeny (Figure 1). Transspecies transfers between vertebrates have been discussed repeatedly [15,22,23]. Indeed, the genomes of the two vertebrate species used here encompass ERVs clustering with five retroviral genera, indicating widespread cross-species transmissions (Figure 1). Several such horizontal transmission events have been described for gammaretroviruses [[3] and Jern et al. submitted] and lentiviruses [2]. Although co-evolution with the host (vertical transmission) is the dominant mode of retroviral transmission, occasional horizontal transmissions make the host species an often unreliable taxonomic marker.
Gag zinc fingers
In addition to Pol, the Gag is also suitable for structural analysis. It is relatively conserved and has well documented functional domains for retroviral RNA packaging, assembly and budding [24-30]. Analysis of the nucleocapsid (NC) from the different genera showed a difference in number of zinc finger motifs, involved in the retroviral RNA interaction [26,28]. Two zinc fingers were detectable in lenti-, alpha-, beta-, epsilon- and some gammaretroviruses (the HERV-H group), whereas the remaining gammaretroviruses had only one, and the spuma-like HERV-L and spumaviruses themselves had none (Figure 2A). The gammaretroviral MLV has a charged amino acid segment upstream of the zinc finger. Recently, we demonstrated that this feature appears to gradually have replaced the loss of the second NC zinc finger in the MLV like group [31]. In the extended data set used here, we could also see that the intermediate SnRV has only one zinc finger (Additional file 2), an indication of several zinc finger loss events. Spumaretroviruses and their related sequences, present in vertebrates and reptiles [15], stand out as structurally different. They have no zinc fingers. They have a separately spliced pro ORF and a relatively low Pol similarity (47.1–61.8%) to other retroviruses. Because most other retroviruses and related viruses (Gypsy and Copia) have NC zinc fingers, it is likely that the spuma-like elements lost theirs. The sequences of the main "beta" branch all have two NC zinc fingers. Aside from this "global" aspect, the uneven distribution and various numbers of NC zinc fingers in the comprehensive sequence collection (Figure 2A), makes the zinc finger trait useful for group delineation rather than for general taxonomy.
Figure 2 Structural traits projected onto the Pol dendrogram. The pol dendrograms in panels A to D are all derived from figure 1. A. The number of recognized Gag NC zinc finger motifs within respective genera. Detections of two NC zinc fingers are marked in light grey for all genera to the right from deltaretroviruses to betaretroviruses, and also for the gammaretroviral HERV-H. The remaining gammaretroviral elements (dark grey) had one NC zinc finger B. Presence of dUTPase is highlighted in grey. The non-primate lentiviral dUTPasePolA (dark grey) is found within Pol and the dUTPasePro (light grey) are found N-terminal of Pro. The dUTPasePro appears to occasionally have been lost, indicated by the two uncolored intermediate chicken and python ERVs. dUTPasePolB. of foamy viruses is not indicated. C. Nucleotide biases may be useful in demarcating retroviral groups locally and the most obvious found are here highlighted. For more detail see refs [31, 40]. D. Genera with detected Pol C-terminal GPY/F motifs are marked light grey and Pro C-terminal G-patch marked in dark grey (exclusively in betaretroviruses). Some betaretroviruses missed a G-Patch and are therefore unmarked.
Translational strategy
In order to produce differing amounts of the different retroviral proteins, the retrovirus may either use i.) ribosomal frameshifting, ii.) nonsense codon readthrough or iii.) splicing, as translational strategies. Well studied gamma and epsilonretroviruses have a distinct genomic structure where a gag-pol transcript with one ORF is produced [32]. The env transcript is a result of splicing activity, a general strategy for all retroviruses. However, the distantly related Errantivirus Cer1, which has all genes in a single ORF (Additional file 2), may possibly represent an original retroelement translational strategy without splicing. A single large polyprotein is also used by some other, even more, distantly related RNA viruses e.g. Picornaviruses. The difference in degree of Gag and Pol expression is regulated by a stop codon suppression readthrough after gag [11]. This genomic structure is shared with the closest related epsilon and even the intermediate epsilonlike SnRV. Mining in our collected ERV databank, we selected sequences with high RetroTector© scores and analyzed their "putein" reading frames. However, definition of the original proviral ORFs is difficult because of the gradual accumulation of postintegrational indel mutations. To minimize such errors, we excluded sequences with predicted frameshifts near the 3'-end of the respective gene and only included ERVs with RetroTector© scores over 1000, thus ensuring a relatively intact provirus. Results from the remaining 436 elements are shown in table 1. In the gammaretroviral genus (RetroTector© defined using motif similarities to known exogenous, and endogenous, gammaretroviral counterparts), we could detect ERVs with not only the predicted lack of frameshifts, "0 f.s.", but also "-1 f.s.", and "+1 f.s." in the Gag-Pro, and Pro-Pol boundaries (Table 1). However, "0 f.s." between Gag and Pro was detected in 67%, while "+1 f.s." and "-1 f.s." were detected in 16% and 17%, respectively. In the Pro-Pol boundary there were 46%, 31% and 23% for "0 f.s.", "+1 f.s." and "-1 f.s." respectively (Table 1).
Thus there is a propensity, however weaker in Pro-Pol, for gammaretroviral ERVs to enclose their Gag, Pro and Pol in the same reading frame. As a comparison, the analyses of exogenous gammaretroviral FLV and MLV genomic structures are also shown (Additional file 2 and [11]). They are known to use the stop codon suppression mechanism in a single gag/pro/pol ("0/0") frame. Although this analysis could not be performed for the few rather damaged epsilon-like HERVs [19], the epsilon retrovirus, WDSV, and the epsilon/spumalike intermediate also shared the single gag-pro-pol frame translational strategy with gammaretroviruses (Additional file 2 and [11]).
The betaretroviral ERVs have been described to have a different translational strategy [11]. There were 60% (-1 fs), 22% (0 fs) and 17% (+1 fs) in the Gag-Pro boundary. Between Pro and Pol there were 51% (-1 fs), 22% (0 fs) and 27% (+1 fs) (Table 1). Thus the betaretroviral ERV frame shift propensities, however weaker between Pro and Pol, agree with the predictions according to the related exogenous MMTV and JSRV (Additional file 2) with the Gag, Pro and Pol in different reading frames separated by "-1" frameshifts, a "-1/-1" pattern. This translational strategy is also recognized in the new intermediate betalike group of chicken and reptiles. We also found that the results ("-1/-1") for chicken alpha ERVs (Table 1) deviated somewhat from the expected "0/-1" pattern (see exogenous RSV in additional file 2 and [11]). The computer aided analysis of the exogenous delta and lentiretroviruses conformed with previous descriptions [11]. HIV had "-1/0", whereas HTLV had "-1/+1" in the Gag-Pro and Pro-Pol boundaries, respectively (Additional file 2). To summarize, we find support for similar translational strategies among ERVs and XRVs, although ERV sequences are harder to analyze due to postintegrational frameshifts. Further, two major directions in the Pol phylogeny could be noted (Figure 1). The viral sequences in the left main branch, the "gamma" branch, often have their gag, pro and pol within the same reading frame. Genera in the right main "beta-branch" (Figure 1), with gag, pro and pol separated in different ORFs, may use different forms of ribosomal frameshifting [11]. Despite the imprecision of reading frame predictions in ERVs (Table 1), we judge inferred translational strategy to be a "global" marker. It is especially suitable for distinction between the extremes of the major gamma and betaretroviral branches in figure 1.
Presence of dUTPase
A dUTPase that prevents incorporation of uracil into the retroviral DNA by dUTP degradation, can be advantageous for some retroviruses. A dUTPase was, in compliance with earlier results [33,34], detected by RetroTector© in both betaretroviruses and non-primate lentiviruses (Figure 2B). However, the localization of the dUTPase differs between the genera. Non-primate lentiviral dUTPase is located within the pol gene (here dubbed dUTPasePolA) [11], whereas the betaretroviral dUTPase is located N-terminal of Pro (here dubbed dUTPasePro). A third dUTPase acquisition event (Additional file 3. Here dubbed dUTPasePolB) in MuERV-L [35,36], is located C-terminal of IN. In the ERV dataset, dUTPasePro was detected in 363 betaretroviral ERVs and 21 intermediate beta-like(alpha-beta) chicken ERVs (Table 1). Thus, many of the chicken intermediate beta-like ERVs lacked detectable dUTPase. Neither could it be found in the intermediate Python ERV (Figure 2B). dUTPasePolB was not tested for.
To investigate the different retroviral acquisitions of dUTPases, we conducted a minimum evolution (ME) analysis, using 389 dUTPase sequences (Additional file 3). The ME tree shows that human betaretroviral dUTPasePro (HML1-10; [20,37] and Blikstad et al, in preparation) and chicken alphabetaretroviral dUTPasePro (GGERVAB1-14; Blomberg et al, in preparation) form one branch together with the more studied mammalian betaretroviral MMTV and MPMV dUTPasePro sequences. This indicates that dUTPasePro has a monophyletic origin and was acquired by an alpha-like retrovirus, earlier in evolution than previously suggested (see [38]), just before or during the formation of betaretroviruses, see figure 2B. The absence of dUTPase from the betaretrovirus like non-mammalian Python retroviruses [22] is in approximate accord with this interpretation. Judging from the ME tree, acquisition of dUTPasePolA (by non-primate lentiviruses) and dUTPasePolB (by the spumalike ERV-L) may also have been single events (Additional file 3). The validity of the detected dUTPases is illustrated by the consensus sequences of the conserved motifs, DSDYxGEIQ, IAQLilD and GGFGST (Additional file 4).
Nucleotide frequency bias
RNA editing, dependent on encapsidation of a host RNA editing enzyme, creates a combination of phenotypic and genotypic traits. In lentiviruses, the host enzyme APOBEC3G is responsible for G to A hypermutation, thus generating an A bias [39]. Although manifested in the retroviral genotype, the nucleotide bias can thus be the result of a phenotypic trait. Nucleotide biases were previously also demonstrated in delta- and betaretroviruses [40]. Using the ERV dataset and the additional XRVs, we confirmed this for lentiviruses, delta- and a subset of gammaretroviruses (Figure 2C), while the spuma-like genus did not show obvious biases [31,40]. Recently we described a group of human gammaretroviral ERVs, the HERV-H-like and adjacent HERV-H- like branching together close to the gammaretroviral root (Figure 2C), to have a uniquely strong G/C bias [31]. In analogy with the lentiviral bias, it is reasonable to assume that HERV-H-like sequences also met an innate antiretroviral defense involving a host RNA editing enzyme. However, the mechanism is unknown and must be different from the cytidine deamination caused by APOBEC3G. Mutational bias caused by the error-prone reverse transcriptase (for a review, see [41]) can also not be ruled out. Reverse transcriptase of different retroviruses has in vitro shown different mutational biases [42]. It has been discussed as a contributing factor for the observed skewed nucleotide composition [43].
C-terminal Protease G-patch domain
Several RNA-binding proteins include a glycine rich domain of about 48 amino acids called "G-patch". This was also present in a betaretroviral MPMV protease C-terminal domain [44]. In self-processing, this domain has been reported to be cleaved from the Pro as a separate protein [45]. The role of this small protein is not determined, but participation in the transport of unspliced retroviral mRNA (see [46]), was suggested [44]. Recently, G-patch was indeed shown to bind single stranded RNA [47]. Further, this G-patch has proved useful in phylogenetic studies, but has shown some inconsistency [48]. In order to extend the phylogenetic investigations and to determine if G-patches are present in other retroviral genera than the described mammalian betaretroviruses, we analyzed the ERV collection for detectable G-patch in the Pro C-terminal domains (Additional file 4). We found 68 positive ERVs (table 1), exclusively within the betaretroviral genus (Figure 2D). Irregularities [48] were also apparent in our Pol phylogeny (Figure 2D), where a G-patch was either degenerated or missing in three of the betaretroviruses, hence uncolored. The validity of the detected G-patch motifs is evident from the consensus sequence, GYx2GxGLGx4GxnG (Additional file 4). An interesting observation was that dUTPasePro occurs in avian beta-like intermediate ERVs (Figure 2B), but without the G-patch (Figure 2D). In fact, no chicken betaretrovirus had a detectable G-patch, while dUTPasePro was often readily detectable. From these data, and those of others [48], we conclude that G-patch entered the genus betaretrovirus after dUTPasePro and that presence of G-patch may be a useful marker for mammalian betaretroviruses, independent of dUTPasePro.
C-terminal polymerase IN motif
The C-terminal end of retroviral Pol integrases (IN) has interesting features. Its terminal position allows for addition of functional modules without disturbing the basic integrase functions, represented by the HHCC zinc finger and the DD35E catalytic domains. Alterations in this C-terminal IN domain may alter the specificity of the integration [49]. The C-terminal domain sometimes contains the motif GPY/F (Additional file 4 and [50]). To this domain, another "chromo" (chromatin-binding) -domain is sometimes appended [50], which interacts with chromatin via DNA-binding proteins [49,51]. Recently, we showed that HERV-H and ERV3 have GPY/F-domains [16,31]. Here we used our ERV collection to extend the analysis. We found 264 ERVs with GPY/F motifs (Table 1). A larger portion had a similar mutated, but still detectable, C-terminal IN region (data not shown). An extended consensus GPY/F motif of the ERVs was computed, WxnGPyxV (Additional file 4). Its typical sequence demonstrates the validity of the detected GPY/F motifs. All these ERVs were gammaretroviral. No betaretroviral element was detected with this domain (Figure 2D). Further, GPY/F motifs were found in epsilon, delta, lenti and errantiviruses (Figure 2D). Thus, in figure 2D, we can demarcate a line where GPY/F, or mutated remnant motifs, can be detected to the left, from the lentiviral branch towards gammaretroviruses, in analogy to how the translational strategies (see above) separated the Pol tree into two major branches.
Accessory genes
The presence of accessory genes in complex retroviruses can also be used for evolutionary inference (Figure 3). Recognition of unknown accessory genes is a difficult bioinformatic problem and absence of accessory genes is hard to ascertain. The analysis therefore rests on demonstrable ones. The delta and lenti genera have several accessory genes with similar functions as integral parts of their replication strategy. They can to some extent replace each other; rex and rev, tax and tat [11]. The sometimes drastic influences of these trans-activating gene products on cellular functions may have kept these viruses out of the germline. Recently, the betaretroviral HERV-K(HML2) was shown to have the accessory genes, rec and/or np9 [52,53], and is thus a complex retrovirus. rec is at least functionally related to rev and rex [54]. Also the epsilon (WDSV) and spumaretroviruses have accessory ORFs, Orf1, 2 and 3, and Bel etc., respectively [11]. The phylogeny of accessory genes (see [55]) is a separate issue, which we do not study further here. From the available information, the accessory genes mainly contribute to rather local properties in the retroviral tree.
Figure 3 Structural traits summary. Simplified view of the different genotypic traits suggested for retroviral phylogeny inference. The branch for Gypsy and Copia represent an imagined midpoint reference in the tree. The number of NC zinc fingers, presence of dUTPase (dUTPasePolB is not indicated), known accessory genes, C-terminal Pro (G-patch) and Pol (GPY/F) motifs are shown. Nucleotide bias was defined to 25 ± 5 %. (↑) shifted upwards; (↓) shifted downwards; (≈) uncertain bias. Exploration of the LTR lengths of the different groups as detected by RetroTector© are shown as boxplots. In addition, the translational strategy may be used in the phylogeny to separate the gammaretroviruses (including class I ERVs) from spuma-like elements (class III ERVs), deltaretroviruses, lentiviruses, alpharetroviruses and the betaretroviruses (class II ERVs) with respective intermediate groups. The Gypsy and Copia are not included in the translational strategy analysis.
LTR lengths
As a final point in the conceptual use of structural traits in phylogenies, a brief exploration of LTR lengths showed a significant difference between the most distantly related gamma and beta genera, where gammaretroviral LTRs are short and betaretroviral LTRs significantly longer (Figure 3). LTR length is therefore a useful additional property for the distinction of these genera.
Conclusion
Inferring phenotypic traits and phylogenies from interpreted genotypic (sequence) ERV properties is similar to the use of fossilized remains for similar purposes in paleontology. The analysis will gather strength with increasing numbers of analyzed host genomes. Pol similarities and structural traits like the ones discussed here, contribute to a better understanding of the retroviral phylogeny. There are at least two major retroviral branches. One contains the gammaretroviruses (including class I ERVs) together with the epsilonretroviruses, and another which includes betaretroviruses (including class II ERVs) together with delta, lenti and alpharetroviruses with their respective intermediate groups. In between, closer to an imaginary root of the retroviral evolutionary tree, we find the older spuma and spuma-like (class III ERVs) retroviruses. The two major branches, schematized in figure 3, differ in "global" genomic properties as i.) translational strategy, ii.) number of Gag NC zinc finger motifs, iii.) presence of dUTPase, iv.) presence of Pro C-terminal G-patch and v.) presence of GPY/F motifs in the IN C-terminal domain. "Local" retroviral properties useful for more narrow delineation of taxa are i.) host species, ii.) nucleotide compositional bias and iii.) LTR lengths.
Methods
Data collection
Genomic data were downloaded from the UCSC genome browser , and annotated retroviral reference sequences included in the phylogenies were extracted from GenBank .
GenBank accession numbers or chromosomal positions in Homo sapiens (version hg16 and 15) for reference sequences in the main phylogenetic tree were as follows: ALV [NC001408], RSV [NC001407], MMTV [NC001503], MPMV [NC001550], JSRV [M80216], HML1 (Chr19-21849393), HML2 (Chr11-101600013), HML3 (Chr1-48344461), HML4 (Chr8-75679221), HML5 [AC004536], HML6 (consensus), HML7 (Chr6-121300220), HML8 (Chr3-131452286), HML9 (Chr9-62700428), HML10 (Chr6-32017925), HERV-H (consensus), HERV-H/RGH2 [D11078], HERV-H/RTVLH2 [M18048], HERV-Fc1 [AL354685], HERV-Fc2 [AC019088], HERV-W (Chr7-9105739), ERV9 [AC073410], ERV3 (Chr7-63865366), HERV-E [M10976], MLV [NC001501], MoLV [AF033811], BaEV [D10032], GaLV [M26927], HERV-ADP [AC005741], HERV-FRD [AC004022], HERV-I (Chr16-72821350), HERV-T (Chr14-104635791), HERV-S [AC004385], FLV [NC001940], PERV [AJ293656], WDSV [NC001867], Xen1 [AJ506107], SnRV [NC001724], BLV [NC001414], HTLV-1 [NC001436], and HTLV-2 [NC001488], Gypsy [AJ000387], HERV-L (RepBase), HSRV [AF033816], HFV [NC001736], MER4like (Chr13-54208300), HERV-L66 (RepBase), HERV-L74 (RepBase), HERV-L40 (RepBase) and Python molurus [AAN77283].
Endogenous retroviral sequences
We used the bioinformatic program RetroTector©, developed by us, to screen the downloaded genomic sequences for proviral integrations. Briefly, the program recognizes conserved retroviral consensus motifs and constructs putative proteins ("puteins") from the different reading frames in the gene candidates. Codon statistics, frequency of stop codons and alignment to known retroviral proteins are used to approximate an original ORF. Finally the puteins are validated and classified using alignments of earlier described proteins from the literature. The validity of the puteins used for alignment and phylogenetic inference, can be confirmed by inspection of excised parts of RT and IN from the full Pol alignment (Additional file 5). The program yields a preliminary genus classification based on motif usage. In several papers, the computerized motif based preliminary retroviral classification was shown to be consistent and robust with reference to other means of classification [16,17,19]. Using a RetroTector© cutoff score of more than 300, we found 3149 proviral sequences in the human genome version hg16 [4] and 260 proviral sequences in the chicken gg01 [21], which were included into our ERV databank. From this databank, we could extract representative proviral sequences for later analyses. The extracted representative sequences had high RetroTector© scores and were selected for their contribution to phylogenetic reconstruction, with preference for intermediates between previously recognized retroviral genera (see figure. 1)
Data analysis
Multiple alignments were conducted using ClustalX (1.83) [56]. A consensus NJ was produced in MEGA2.1 [57] using the pairwise deletion option, Poisson amino acid correction and 500 bootstraps. A set of maximum likelihood analyses using the PHYLIP program package [58] were used to verify the tree topologies. Consensus analysis of C-terminal Pro (G-patch) and Pol (GPY/F) motifs were conducted using WebLogo at , with default settings
Statistics were extracted from the ERV databank collected through the RetroTector© analysis of the different genomes.
The Pol FASTA sequences are included into the additional files (Additional file 6).
List of Abbreviations used
aa amino acids
CA Capsid
dUTPase deoxyuridine triphosphatase
Env Envelope
ERVs Endogenous retroviral sequences
Gag Group specific antigen
HERVs Human endogenous retroviral sequences
IN Integrase
LTR Long terminal repeat
MA Matrix
NC Nucleocapsid
PBS Primer binding site
Pol Polymerase
PPT Polypurine tract
Pro Protease
RNAse H Ribonuclease H enzyme
RT Reverse transcriptase
XRV Exogenous retrovirus
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
PJ conceived the study, participated in its design and coordination, carried out the molecular genetic studies and drafted the manuscript. GOS developed algorithms and carried out the programming needed for the analyses. JB conceived the study, and participated in its design and coordination, participated in programming and helped to draft the manuscript. All authors read and approved the final manuscript.
Supplementary Material
Additional File 1
Pol similarity matrix. Pol NJ cladogram (1000 bootstraps and pairwise deletions) aligned to a similarity matrix based on PAM250.
Click here for file
Additional File 2
Retroviral genomic structures. RetroTector© output of selected retroviral genomic structures described in the text.
Click here for file
Additional File 3
dUTPase phylogenetic tree. Retroviral dUTPase acquisitions. Minimum Evolution (ME) tree (100 bootstraps).
Click here for file
Additional File 4
Motif consensus validations. WebLogo consensus of the partial dUTPasePro, C-terminal Pro (G-patch) and Pol (GPY/F) motifs.
Click here for file
Additional File 5
Putein validations. Validation of puteins from Pol alignment. Excised parts of RT and IN are shown.
Click here for file
Additional File 6
Pol FASTA sequences. Pol FASTA sequences.
Click here for file
Acknowledgements
This work was supported by the Swedish research council (grant #: K2004-32X-14252-03A) and Stanley Foundation (grant #: 03R-584).
==== Refs
Kumar A Bennetzen JL Plant retrotransposons Annu Rev Genet 1999 33 479 532 10690416 10.1146/annurev.genet.33.1.479
Gao F Bailes E Robertson DL Chen Y Rodenburg CM Michael SF Cummins LB Arthur LO Peeters M Shaw GM Sharp PM Hahn BH Origin of HIV-1 in the chimpanzee Pan troglodytes troglodytes Nature 1999 397 436 441 9989410 10.1038/17130
Yohn CT Jiang Z McGrath SD Hayden KE Khaitovich P Johnson ME Eichler MY McPherson JD Zhao S Paabo S Eichler EE Lineage-Specific Expansions of Retroviral Insertions within the Genomes of African Great Apes but Not Humans and Orangutans PLoS Biol 2005 3 1 11 10.1371/journal.pbio.0030110
International Human Genome Sequencing Consortium Initial sequencing and analysis of the human genome Nature 2001 409 860 921 11237011 10.1038/35057062
Unpublished data from the Washington University and MIT/Broad Institute Sequencing Centers (directed by Rick Wilson and Eric Lander, respectively) are publicly released at [http://genome.ucsc.edu/]. See also NHGRI Press Release [www.nih.gov/news/pr/dec2003/nhgri-10.htm].
Li WH Molecular Evolution 1997 Sunderland, MA, USA, Sinnauer Associates, Inc., Publishers
van Regenmortel MHV Fauquet CM Bishop DHL Cartens EB Estes MK Lemon SM Maniloff J Mayo MA McGeoch DJ Pringle CR Wickner RB Seventh ICTV report 2000 San Diego, CA, USA, Academic Press 1024
Smit AF Identification of a new, abundant superfamily of mammalian LTR-transposons Nucleic Acids Res 1993 21 1863 1872 8388099
Cappello J Handelsman K Lodish HF Sequence of Dictyostelium DIRS-1: an apparent retrotransposon with inverted terminal repeats and an internal circle junction sequence Cell 1985 43 105 115 2416457 10.1016/0092-8674(85)90016-9
Gorinsek B Gubensek F Kordis D Evolutionary genomics of chromoviruses in eukaryotes Mol Biol Evol 2004 21 781 798 14739248 10.1093/molbev/msh057
Coffin JM Hughes SH Varmus HE Retroviruses 1997 New York, Cold Spring Harbor Laboratory Press
Wilkinson DA Mager DL Leong JC Levy J Endogenous human retroviruses The Retroviridae 1994 New York, Plenum Press 465 535
Lindeskog M Transcription, splicing and genetic structure within the human endogenous retroviral HERV-H family. Infectious diseases and Medical microbiology 1999 Lund, Lund University 142
Mager DL Medstrand P Cooper D Retroviral Repeat Sequences Nature Encyclopedia of the Human Genome 2003 , Nature Publishing Group, London, UK
Herniou E Martin J Miller K Cook J Wilkinson M Tristem M Retroviral diversity and distribution in vertebrates J Virol 1998 72 5955 5966 9621058
Andersson AC Yun Z Sperber GO Larsson E Blomberg J ERV3 and Related Sequences in Humans: Structure and RNA Expression J Virol 2005 79 9270 9284 15994821 10.1128/JVI.79.14.9270-9284.2005
Jern P Sperber GO Blomberg J Definition and variation of human endogenous retrovirus H Virology 2004 327 93 110 15327901 10.1016/j.virol.2004.06.023
Tristem M Identification and characterization of novel human endogenous retrovirus families by phylogenetic screening of the human genome mapping project database J Virol 2000 74 3715 3730 10729147 10.1128/JVI.74.8.3715-3730.2000
Oja M Sperber GO Blomberg J Kaski S Self-organizing map-based discovery and visualization of human endogenous retroviral sequence groups. International Journal of Neural Systems
Andersson ML Lindeskog M Medstrand P Westley B May F Blomberg J Diversity of human endogenous retrovirus class II-like sequences J Gen Virol 1999 80 ( Pt 1) 255 260 9934709
International Chicken Genome Sequencing Consortium Sequence and comparative analysis of the chicken genome provide unique perspectives on vertebrate evolution Nature 2004 432 695 716 15592404 10.1038/nature03154
Huder JB Boni J Hatt JM Soldati G Lutz H Schupbach J Identification and characterization of two closely related unclassifiable endogenous retroviruses in pythons (Python molurus and Python curtus) J Virol 2002 76 7607 7615 12097574 10.1128/JVI.76.15.7607-7615.2002
Martin J Herniou E Cook J Waugh O'Neill R Tristem M Human endogenous retrovirus type I-related viruses have an apparently widespread distribution within vertebrates J Virol 1997 71 437 443 8985368
Wills JW Cameron CE Wilson CB Xiang Y Bennett RP Leis J An assembly domain of the Rous sarcoma virus Gag protein required late in budding J Virol 1994 68 6605 6618 8083996
Chance MR Sagi I Wirt MD Frisbie SM Scheuring E Chen E Bess JWJ Henderson LE Arthur LO South TL Extended x-ray absorption fine structure studies of a retrovirus: equine infectious anemia virus cysteine arrays are coordinated to zinc Proc Natl Acad Sci U S A 1992 89 10041 10045 1332027
Bowzard JB Bennett RP Krishna NK Ernst SM Rein A Wills JW Importance of basic residues in the nucleocapsid sequence for retrovirus Gag assembly and complementation rescue J Virol 1998 72 9034 9044 9765448
Freed EO Viral late domains J Virol 2002 76 4679 4687 11967285 10.1128/JVI.76.10.4679-4687.2002
Cheslock SR Poon DTK Fu W Rhodes TD Henderson LE Nagashima K McGrath CF Hu WS Charged Assembly Helix Motif in Murine Leukemia Virus Capsid: an Important Region for Virus Assembly and Particle Size Determination J Virol 2003 77 7058 7066 12768025 10.1128/JVI.77.12.7058-7066.2003
Mortuza GB Haire LF Stevens A Smerdon SJ Stoye JP Taylor IA High-resolution structure of a retroviral capsid hexameric amino-terminal domain Nature 2004 431 481 485 15386017 10.1038/nature02915
Kikonyogo A Bouamr F Vana ML Xiang Y Aiyar A Carter C Leis J Proteins related to the Nedd4 family of ubiquitin protein ligases interact with the L domain of Rous sarcoma virus and are required for gag budding from cells Proc Natl Acad Sci U S A 2001 98 11199 11204 11562473 10.1073/pnas.201268998
Jern P Sperber GO Ahlsen G Blomberg J Sequence variability, gene structure, and expression of full-length human endogenous retrovirus h J Virol 2005 79 6325 6337 15858016 10.1128/JVI.79.10.6325-6337.2005
Petropoulos C Coffin JM, Hughes SH and Varmus HE Retroviral Taxonomy, Protein Structures, Seqiences, and Genetic Maps Retroviruses 1997 New York, NY, USA, Cold Spring Harbor Laboratory Press 757 805
Baldo AM McClure MA Evolution and horizontal transfer of dUTPase-encoding genes in viruses and their hosts J Virol 1999 73 7710 7721 10438861
Mayer J Meese EU Presence of dUTPase in the various human endogenous retrovirus K (HERV-K) families J Mol Evol 2003 57 642 649 14745533 10.1007/s00239-003-2514-6
Benit L De Parseval N Casella JF Callebaut I Cordonnier A Heidmann T Cloning of a new murine endogenous retrovirus, MuERV-L, with strong similarity to the human HERV-L element and with a gag coding sequence closely related to the Fv1 restriction gene J Virol 1997 71 5652 5657 9188643
Benit L Lallemand JB Casella JF Philippe H Heidmann T ERV-L elements: a family of endogenous retrovirus-like elements active throughout the evolution of mammals J Virol 1999 73 3301 3308 10074184
Medstrand P Lindeskog M Blomberg J Expression of human endogenous retroviral sequences in peripheral blood mononuclear cells of healthy individuals J Gen Virol 1992 73 ( Pt 9) 2463 2466 1402820
Vogt VM Coffin JM, Hughes SH and Varmus HE Retroviral Virions and Genomes Retroviruses 1997 New York, NY, USA, Cold Spring Harbor Laboratory Press 27 70
Mangeat B Turelli P Caron G Friedli M Perrin L Trono D Broad antiretroviral defence by human APOBEC3G through lethal editing of nascent reverse transcripts Nature 2003 424 99 103 12808466 10.1038/nature01709
Berkhout B Grigoriev A Bakker M Lukashov VV Codon and amino acid usage in retroviral genomes is consistent with virus-specific nucleotide pressure AIDS Res Hum Retroviruses 2002 18 133 141 11839146 10.1089/08892220252779674
Katz RA Skalka AM Generation of diversity in retroviruses Annu Rev Genet 1990 24 409 445 1708222 10.1146/annurev.ge.24.120190.002205
Vartanian JP Meyerhans A Sala M Wain-Hobson S G-->A hypermutation of the human immunodeficiency virus type 1 genome: evidence for dCTP pool imbalance during reverse transcription Proc Natl Acad Sci U S A 1994 91 3092 3096 7512722
Berkhout B Das AT Beerens N HIV-1 RNA editing, hypermutation, and error-prone reverse transcription Science 2001 292 7 11292864 10.1126/science.292.5514.7a
Aravind L Koonin EV G-patch: a new conserved domain in eukaryotic RNA-processing proteins and type D retroviral polyproteins Trends Biochem Sci 1999 24 342 344 10470032 10.1016/S0968-0004(99)01437-1
Hruskova-Heidingsfeldova O Andreansky M Fabry M Blaha I Strop P Hunter E Cloning, bacterial expression, and characterization of the Mason-Pfizer monkey virus proteinase J Biol Chem 1995 270 15053 15058 7797487 10.1074/jbc.270.25.15053
Ernst RK Bray M Rekosh D Hammarskjold ML A structured retroviral RNA element that mediates nucleocytoplasmic export of intron-containing RNA Mol Cell Biol 1997 17 135 144 8972193
Svec M Bauerova H Pichova I Konvalinka J Strisovsky K Proteinases of betaretroviruses bind single-stranded nucleic acids through a novel interaction module, the G-patch FEBS Lett 2004 576 271 276 15474050 10.1016/j.febslet.2004.09.010
Gifford R Kabat P Martin J Lynch C Tristem M Evolution and distribution of class II-related endogenous retroviruses J Virol 2005 79 6478 6486 15858031 10.1128/JVI.79.10.6478-6486.2005
Singleton TL Levin HL A long terminal repeat retrotransposon of fission yeast has strong preferences for specific sites of insertion Eukaryot Cell 2002 1 44 55 12455970 10.1128/EC.01.1.44-55.2002
Malik HS Eickbush TH Modular evolution of the integrase domain in the Ty3/Gypsy class of LTR retrotransposons J Virol 1999 73 5186 5190 10233986
Sandmeyer S Integration by design Proc Natl Acad Sci U S A 2003 100 5586 5588 12732725 10.1073/pnas.1031802100
Lower R Lower J Kurth R The viruses in all of us: characteristics and biological significance of human endogenous retrovirus sequences Proc Natl Acad Sci U S A 1996 93 5177 5184 8643549 10.1073/pnas.93.11.5177
Armbruester V Sauter M Roemer K Best B Hahn S Nty A Schmid A Philipp S Mueller A Mueller-Lantzsch N Np9 protein of human endogenous retrovirus k interacts with ligand of numb protein x J Virol 2004 78 10310 10319 15367597 10.1128/JVI.78.19.10310-10319.2004
Magin C Lower R Lower J cORF and RcRE, the Rev/Rex and RRE/RxRE homologues of the human endogenous retrovirus family HTDV/HERV-K J Virol 1999 73 9496 9507 10516058
Tristem M Marshall C Karpas A Hill F Evolution of the primate lentiviruses: evidence from vpx and vpr Embo J 1992 11 3405 3412 1324171
Thompson JD Gibson TJ Plewniak F Jeanmougin F Higgins DG The CLUSTAL_X windows interface: flexible strategies for multiple sequence alignment aided by quality analysis tools Nucleic Acids Res 1997 25 4876 4882 9396791 10.1093/nar/25.24.4876
Kumar S Tamura K Jakobsen IB Nei M MEGA2: molecular evolutionary genetics analysis software Bioinformatics 2001 17 1244 1245 11751241 10.1093/bioinformatics/17.12.1244
Felsenstein J PHYLIP (Phylogeny Inference Package) 2004 version 3.6. Distributed by the author. Department of Genome Sciences, University of Washington, Seattle.
|
16092962
|
PMC1224870
|
CC BY
|
2021-01-04 16:36:39
|
no
|
Retrovirology. 2005 Aug 10; 2:50
|
utf-8
|
Retrovirology
| 2,005 |
10.1186/1742-4690-2-50
|
oa_comm
|
==== Front
RetrovirologyRetrovirology1742-4690BioMed Central London 1742-4690-2-521610722310.1186/1742-4690-2-52ReviewRetroviral superinfection resistance Nethe Micha [email protected] Ben [email protected] der Kuyl Antoinette C [email protected] Dept. of Human Retrovirology, Academic Medical Centre, University of Amsterdam, Meibergdreef 15, 1105AZ Amsterdam, The Netherlands2005 18 8 2005 2 52 52 18 4 2005 18 8 2005 Copyright © 2005 Nethe et al; licensee BioMed Central Ltd.2005Nethe et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
The retroviral phenomenon of superinfection resistance (SIR) defines an interference mechanism that is established after primary infection, preventing the infected cell from being superinfected by a similar type of virus. This review describes our present understanding of the underlying mechanisms of SIR established by three characteristic retroviruses: Murine Leukaemia Virus (MuLV), Foamy Virus (FV), and Human Immunodeficiency Virus (HIV). In addition, SIR is discussed with respect to HIV superinfection of humans.
MuLV resistant mice exhibit two genetic resistance traits related to SIR. The cellular Fv4 gene expresses an Env related protein that establishes resistance against MuLV infection. Another mouse gene (Fv1) mediates MuLV resistance by expression of a sequence that is distantly related to Gag and that blocks the viral infection after the reverse transcription step. FVs induce two distinct mechanisms of superinfection resistance. First, expression of the Env protein results in SIR, probably by occupancy of the cellular receptors for FV entry. Second, an increase in the concentration of the viral Bet (Between-env-and-LTR-1-and-2) protein reduces proviral FV gene expression by inhibition of the transcriptional activator protein Tas (Transactivator of spumaviruses). In contrast to SIR in FV and MuLV infection, the underlying mechanism of SIR in HIV-infected cells is poorly understood. CD4 receptor down-modulation, a major characteristic of HIV-infected cells, has been proposed to be the main mechanism of SIR against HIV, but data have been contradictory. Several recent studies report the occurrence of HIV superinfection in humans; an event associated with the generation of recombinant HIV strains and possibly with increased disease progression. The role of SIR in protecting patients from HIV superinfection has not been studied so far.
The phenomenon of SIR may also be important in the protection of primates that are vaccinated with live attenuated simian immunodeficiency virus (SIV) against pathogenic SIV variants. As primate models of SIV infection closely resemble HIV infection, a better knowledge of SIR-induced mechanisms could contribute to the development of an HIV vaccine or other antiviral strategies.
==== Body
Introduction
Viral entry and replication is a complex process that involves multiple viral and host proteins. Many host gene products can interfere with virus infection at the cellular level (for a review, see: [1]). These proteins are encoded by variants of essential genes (that can not support viral infection), or represent true anti-viral factors (gene products whose main role it is to protect the cell from a productive virus infection). A special form of virus resistance is the capacity of cells to prevent a second infection by a virus that is closely related to the virus that has already established an infection. In most cases, virus-encoded proteins are responsible for this phenomenon, which is termed superinfection resistance (SIR) or viral interference. A simple form of SIR is receptor occupancy by viral Env proteins, preventing the binding of a second virus, but many additional mechanisms have been described. Although SIR is not restricted to retroviruses, it has been studied in depth for this class of viruses. This review deals with the molecular mechanisms of SIR at the cellular level in three retrovirus classes: simple retroviruses (here MuLV), spumaretroviruses (FV), and lentiviruses (HIV). The mechanisms and clinical consequences of HIV-1 superinfection in patients, which is defined as the reinfection of an individual with a second heterologous strain of HIV-1 [2], will also be discussed.
Murine leukaemia virus
In the early 1950's, Gross identified a virus that could induce leukaemia in mice [3]. This discovery was quickly followed by the identification of additional leukaemia-inducing viruses, which led to the definition of the class of Murine Leukaemia Viruses (MuLVs). Although the list of MuLV related viruses is still expanding, most MuLVs can be divided into four classes: ecotropic, amphotropic, polytropic (sometimes called MCF viruses), and xenotropic. This classification is based on the type of host cell that is infected, based on the fact that the 4 classes use 3 different receptors. Ecotropic MuLVs can only infect murine cells, whereas polytropic MuLVs infect a broad host range of mammalian species including mice, albeit with variable efficiencies. Xenotropic MuLVs can infect many species, e.g. mink, rabbit, duck and human, but not cells of laboratory mice (reviewed in [4]). The polytropic and xenotropic viruses use the same receptor, Xpr1, also called Syg1. Polymorphisms in the Xpr1 protein determine the exact host range of the polytropic and xenotropic MuLVs. Ecotropic viruses use the amino acid transporter mCAT1 as their receptor, while the receptor for amphotropic MuLVs is the sodium-dependent phosphate transporter Pit2.
Cellular factors associated with MuLV restriction, have been studied extensively, whereby polymorphisms in the MuLV receptor genes were found to play a major role. Different cell lines were found to express functional variants of the ecotropic-, polytropic-, and xenotropic-MuLV receptors, which block infection by certain MuLV strains [5-7]. Proviral endogenous genes, like the mouse Fv1 and Fv4 gene products, can mediate restriction of MuLV replication by SIR associated mechanisms [8,9].
Fv1 mediated resistance to MuLV infection
In 1967 the Fv1 gene was reported to be an important determinant of cell susceptibility towards MuLV infection [10]. Two common alleles for the Fv1 gene (Fv1b and Fv1n)present in prototypical mouse strains of BALB/c and NIH/Swiss, were found to interfere with certain classes of MuLVs (reviewed by [1,9]). Cells from NIH/Swiss mice, which carry the Fv1n allele were resistant to infection with the so-called B-tropic MuLVs. BALB/c mouse cells, which carry the Fv1b allele, were resistant to N-tropic MuLVs. In addition, a third class of MuLVs, the so-called NB tropic MuLVs, defined strains that can infect Fv1n as well as Fv1b expressing cells (all reviewed in [9]).
Substitution of defined regions within the N-tropic and B-tropic MuLV genomes by recombinant DNA cloning revealed that the Gag gene encoding the capsid protein CA determines the cell tropism. In particular, a single amino acid within the CA protein was identified to determine N or B tropism [11]. Fv1 mediated restriction occurs post-penetration and at or before integration of the proviral DNA genome [12], reviewed in [13].
Cloning and sequencing of the Fv1 gene [8] showed that the Fv1 sequence is similar to the presumptive Gag gene of human endogenous retrovirus HERV-L (60% identity over a stretch of 1.3 kb). The Fv1n and Fv1b alleles differ by a few mutations, and in addition have a length difference of 19 amino acids at the C-terminal end. Gag proteins are known to interact tightly with each other, which is essential during virion assembly [1,14]. Possibly, interactions between the Fv1 Gag-like protein and viral Gag derived CA are involved in the Fv1 mechanism of resistance (for reviews, see [1,15,16]). CA has been suggested to act as a transport signal for the pre-integration complex (PIC) to facilitate import into the nucleus. The subcellular localization of the Fv1 product suggests it may affect virions on their way to the nucleus [17]. The most straightforward explanation of Fv1 mediated interference would be binding of Fv1 to CA in an allele specific way manner that alters CA binding to the PIC (fig. 1). The PIC could remain captured in the cytoplasmic compartment and thus not be able to migrate into the nucleus. However, a direct interaction between Fv1 and CA has never been demonstrated, although crystallographic studies recently suggested that a potential Fv1 binding domain exists in the MuLV CA [18]. Finally, a direct interaction of Fv1 with the PIC cannot be excluded, changing its conformation in such a way that it becomes non-functional (fig. 1). However, all mechanisms presented here to explain Fv1 restriction lack solid experimental evidence, and it should be noted that in the mouse genome there are hundreds of retroviral elements more closely related to MuLV than Fv1, and none of these restricts MuLV replication. A protein named TRIM5alpha has recently been characterized to restrict HIV-1 by an Fv1-like mechanism in primate cells. Restriction capabilities of TRIM5alpha vary amongst primates, so that rhesus monkey TRIM5alpha restricts N-tropic MuLV and HIV-1, but not B-tropic MuLV, while human TRIM5alpha restricts N-tropic MuLV, but not B-tropic MuLV or HIV-1 (reviewed in [19]). The ability to restrict HIV-1 is determined by a single amino acid in the C-terminal SPRY domain of TRIM5alpha [19]. As for Fv1 restriction of MuLV, TRIM5alpha targets the HIV-1 CA protein. Several mechanisms have been proposed for TRIM5alpha restriction, including binding and trapping of incoming virus, interference with uncoating, inhibition of SUMOylation (and thereby interfere with intracellular trafficking of the PIC), and targeting the incoming particle for proteasomal degradation whereby TRIM5alpha transfers the ubiquitin molecules to CA (reviewed in [1,9]). Elucidating the way by which TRIM5alpha restricts retroviruses might also shed light upon the mechanism of Fv1 restriction.
Fv4 mediated resistance to MuLV infection
In 1975, Suzuki described the discovery of a new resistance gene, Fv4, in the G strain of laboratory mice [20,21]. The Fv4 gene was also identified in Asian wild mouse species [22]. Genetic mapping studies located the Fv4 gene on chromosome 12 [23]. There are two alleles at the Fv4 locus: the Fv4r resistance allele is dominant [20,21]. A first clue about the nature of the Fv4 gene came with the identification of MuLV Env related proteins in Fv4 resistant cell lines [24], which suggested an Env-like sequence for the Fv4 gene. Using an Env-specific probe, a 5.2 kb fragment of the Fv4r was cloned that contained part of the Pol gene, the entire MuLV Env region and the 3' long-terminal repeat (LTR) of an ecotropic MuLV [25,26]. Sequence analysis revealed that Fv4 Env encodes a surface (SU) and transmembrane (TM) Env domain that closely resembled (>90%) the homologous Env sequences in the unusual ecotropic MuLVs found in Asian wild mice [22]. Transgenic mice carrying the Fv4 gene showed complete resistance to ecotropic MuLV infection [27]. Moreover, transplantation of a certain percentage of Fv4 resistant bone marrow cells into the bone marrow of Fv4 susceptible mice strains induced full resistance against MuLV infection [28]. Although Fv4 mediated resistance has been demonstrated in different experimental systems, the underlying molecular mechanism remains unclear. As described earlier, Env-receptor interactions mediate retroviral entry into the target cell. Therefore, Fv4r mediated resistance has been suggested to rely on Fv4 Env binding to the MuLV receptor, which prevents exogenous MuLV infection. Substitution of the complete Fv4r Env gene in MuLV clones abrogated viral entry, indicating that the protein is defective [29]. The defect was attributed to a single amino acid substitution in the fusion peptide of the Fv4r Env protein, which when artificially introduced into an MuLV clone led to an Env protein that was able to bind to the cellular receptor, and was incorporated into virus particles at normal levels, but was incapable of promoting fusion and viral entry [30].
Fv4 and mCAT1 interactions
Sequence analysis of the ecotropic MuLV receptor showed it to be a cationic type 1 amino acid transporter (mCAT1) [31,32]. Comparable expression patterns of mCAT1 mRNA have been described for different tissues of Fv4r congenic MuLV-resistant (C4W = BALB/c-Fv-4Wr) and -susceptible (C3H/HeMsNrs and C56BL/6) mice strains. However, recombinant F-SU/GFP, consisting of the SU domain of Friend MuLV and the GFP protein, was unable to stain most mCAT1 mRNA expressing tissues of the C4W (Fv4r) mice strain [33], suggesting that either an intracellular downregulation of the receptor has occurred, or that the receptor is blocked at the cell surface by the Fv4 gene product (fig. 2).
Altogether, these data strongly suggest that Fv4r interferes with MuLV infection by masking of the MuLV receptor through binding of Fv4 Env. Two other mouse interference genes, named Rmcf1 and Rmcf2, also cause MuLV resistance by Fv4-related interference mechanisms [34,35]. Crosses between an Rmcf1 resistant mouse strain and an Rmcf1 lacking mouse strain revealed that inheritance of Rmcf1 resistance correlated with the inheritance of an endogenous MCF virus Env gene. The Rmcf2 gene also encodes an Env glycoprotein, and its expression blocks infection by polytropic MuLVs [35].
Foamy viruses
In 1950 a new type of retrovirus was isolated from cell cultures derived from monkey kidneys. Foamy viruses (FV) were named after the characteristic foam-like effect they induce in cell culture. FVs are considered to be harmless in experimentally infected animals. The various unique features of FVs concerning their replication led in 2002 to the establishment of a new, distinct retroviral subfamily: the spumaretrovirinae (reviewed in [36]).
The genomic structure of FVs indicates that these viruses belong to the more complex retroviruses. The FV genome transcribes, besides Gag, Pol and Env, two major mRNA's from an internal promoter near the 3'end of the genome (reviewed in [36]): a DNA binding protein called Transactivator of spumaviruses (Tas), and the 60 kDA Bet protein. Tas is involved in the switch from latent to lytic virus replication, while Bet has a negative regulatory effect upon the internal promoter [37,38]. Furthermore, Bet can inhibit the APOBEC3 family of antiretroviral proteins [39], and mediates SIR [40].
FV Env mediated SIR
As retrovirus entry depends on the interaction of the SU domain of Env with the target receptor, down-regulation of such a receptor would be a plausible mechanism for SIR. To date no receptor has been identified for FV. It has been proposed that a pH-dependent fusion process mediates foamy virus entry [41]. To investigate FV superinfection, Moebes and colleagues [42] tested whether overexpression of the FV Env protein induced SIR by downregulation of the putative receptor. Indeed, BHK-21 cell lines containing a stably transfected Env gene were completely resistant to infection with FV vectors that use FV Env for entry.
Deletion analysis of the FV Env protein showed that several properties of Env are needed to induce SIR: membrane anchorage of Env extracellular domains, efficient cell surface transport of the Env protein, and correct processing of the Env subunits [43]. So, in contrast to MuLV Env, secretion of FV Env is not sufficient to induce SIR.
A recombinant FV SU-Ig protein and FV Env expressing cell lines were constructed to study FV Env binding to the surface of target cells [44]. The receptor for FV is still undetermined, and it is possible that general features on the membrane surface, like for example glycolipids, mediate FV entry. This would explain the broad infection range of FV on mammalian and non-mammalian cells [45]. However, the binding experiments suggested that SIR by FV Env is similar to SIR by other retroviruses, whereby high expression of FV Env in stably transfected cell lines led to a complete resistance to FV SU-Ig binding and FV permissiveness, and low expression of FV Env led to a decreased susceptibility to infection and a lowered FV SU-Ig binding [44].
Concluding, the expression of FV Env proteins establishes resistance against FV superinfection. Moreover, FV Env proteins induce SIR at the cell surface, which suggests down-regulation of cell surface FV entry mediators. However, the exact underlying mechanism of SIR remains unclear.
Bet mediated resistance to FV superinfection
Chronically infected FV cells, which are characterized by reduced production of Tas, are found to express predominately ΔHFV, a distinct proviral form of FV [46,47]. A persistent but latent infection is common in FV infected animals (reviewed in [48]). ΔHFV contains a 301-bp deletion in the Tas gene, which is spliced out from the pregenomic RNA [46]. Interestingly, ΔHFV seemed to interfere with FV infection [38]. This interference strongly correlated with the number of integrated ΔHFV copies [38]. ΔHFV constructs with a defective Bet gene were unable to interfere with FV infection [38], suggesting that Bet is involved in SIR. Normally, ΔHFV transfected cells contain stable levels of Bet mRNA and protein, and Bet is the major viral protein expressed in chronically infected cells [38].
The establishment of a Bet-expressing cell line confirmed a Bet-mediated induction of SIR [40]. Interestingly, Bet-induced SIR is unlikely to be mediated by Env-directed down-regulation of the FV receptor, as no Env mRNA or proteins were detected during the early phase of ΔHFV interference with FV infection [38]. In addition, Bet+ cells did not prevent infection by a GFP-MuLV vector containing a ΔHFV envelope construct whereby the cytoplasmic tail of the transmembrane part is derived from MuLV [40]. As this vector contains the HFV envelope surface and TM domains, it must use the FV receptor to gain access to the Bet+ cells.
Infection of Bet+ and Bet- cells by FV resulted in 3–4 fold lower titres in the Bet+ cells [40]. As proviral DNA was able to integrate into the host genome, Bet possibly interferes with FV replication during transcription of the provirus, although the lower levels of FV in Bet+ cells could suggest an additional effect upon viral entry. Foamy viruses contain an internal promoter that drives transcription of Bet and Tas mRNA (reviewed in [36]). The transactivator Tas activates both the LTR and internal promoters by direct binding [37]. Bet and Tas are produced from overlapping reading frames and mediate opposite effects on FV replication (fig. 3). Cell lines chronically infected with FV contain abundant levels of the negative regulator Bet, low levels of structural proteins and of the transactivator Tas, and a high ΔFV load [37,38]. Increasing the level of Tas by transfecting latently infected cells with a Tas expression vector triggered FV replication and cell lysis [37]. Thus, Bet reduces FV replication by inhibition of Tas expression, which in turn reduces internal promoter activity. The exact mechanism by which Bet inhibits Tas expression is not clear. Bet protein could stimulate splicing of its own mRNA, which consequently would alter Tas RNA levels. Other possibilities are Bet-mediated inhibition of Tas RNA transport or decreased stability of Tas RNA. It seems unlikely that Bet prevents Tas expression by stimulation of promyelocytic leukaemia protein (PML), the only known inhibitor of Tas [49], as significant amounts of PML were unable to prevent FV replication [50].
HIV superinfection resistance
To date an estimated 40 million people worldwide are infected with the Human Immunodeficiency Virus (HIV), classified as a lentivirus within the class of retroviruses. HIV is associated with the development of Acquired Immune Deficiency Syndrome (AIDS). Two main virus types exist, HIV-1 and HIV-2, of which HIV-1 infection is the most important cause of AIDS.
Like all other retroviruses, the HIV virion contains two copies of an RNA genome that is encapsulated by CA and Env proteins. The Env glycoproteins gp120 and gp41 mediate viral entry by interacting with CD4 molecules on susceptible cells. The CD4 receptor is a type 1 transmembrane glycoprotein and is mainly found on primary T lymphocytes, dendritic cells and macrophages. Interaction of gp120 with CD4 induces conformational changes in the Env protein structure, which enables Env to interact with a coreceptor, such as the CCR5 or CXCR4 chemokine receptor, which leads to HIV entry into the target cell (reviewed in [51]). Several host factors have been identified that interfere with early steps during entry or replication of HIV-1, e.g. APOBEC3G/CEM15, Lv1, Lv2, and TRIM5alpha (for a review, see: [16]). Additional mechanisms by which an initial virus can inhibit entry or replication of a second virus will be discussed below.
Since the identification of the AIDS virus, various strategies have been proposed to prevent the spread of HIV infection. The underlying mechanisms of SIR in HIV-infected cells are of particular interest for the development of novel antiviral approaches related to SIR. However, as a caveat, we note that several studies describe the occurrence of HIV superinfection in patients. The next sections will describe the current understanding of the underlying mechanisms of SIR by HIV-1.
CD4-mediated resistance to HIV superinfection
One of the major characteristics of HIV-infected cells is down-modulation of the CD4 receptor [52-54]. To date three viral HIV proteins; Vpu, Env, and Nef have been identified that mediate CD4 down-regulation by distinct mechanisms (reviewed in [55,56]), indicating the importance of CD4 down-regulation for HIV infection. As receptor down-modulation is a simple way of preventing a second viral infection, and a method that is successfully used by other retroviruses, CD4 down-modulation was initially assumed to be the main SIR mechanism in HIV infection.
All primate lentiviruses, HIV-1, HIV-2 and Simian Immunodeficiency Virus (SIV), encode the Nef protein (reviewed in [55]). Nef binds directly to a di-leucine-like motif in the cytoplasmic domain of CD4. Nef is able to bind different members of the adaptor proteins (AP-1, AP-2, AP-3 and AP-4), which contain distinct transport signals. Simultaneous binding of Nef to CD4 and AP-2 at the cell surface induces endocytosis of CD4. In addition, Nef binding of AP-1 and AP-3 in the trans-Golgi network may mediate trafficking of newly synthesized CD4 directly to lysosomes. Stable transfection of the SIV Nef gene in a CD4+ T cell line reduced cell surface-expression of CD4, and rendered the cells resistant to subsequent HIV-1 infection [57]. As HIV-1 transcription was not inhibited in these cells, the authors speculate that the inhibition of superinfection in this model system is due to Nef-induced CD4 down-modulation. Besides, a clonal HIV-1 containing T cell line with down- regulated CD4 expression is also resistant to HIV-2 superinfection [54]. HIV-2 infected cells do not seem to resist subsequent HIV-1 infection, which may be explained by the inability of HIV-2 to induce CD4 down-modulation.
In contrast to Nef, Env and Vpu mediate CD4 down-modulation by preventing the intracellular transport of newly synthesised CD4 molecules (reviewed in [56]). Binding of CD4 by the Env precursor protein gp160 in the endoplasmatic reticulum (ER) triggers the formation of aggregates, which block further CD4 transport to the cell surface. In addition, Vpu mediates CD4 down-modulation by directing newly synthesised CD4 to proteosomes for degradation. Among the immunodeficiency viruses, Vpu is encoded nearly exclusively by HIV-1. Vpu has been suggested to redirect CD4 trafficking by acting as an adaptor between CD4 and the h-βTrCP protein that is a key connector in the ubiquitin-mediated proteolysis machinery. Restriction of Vpu mediated CD4 down-modulation either by inhibition of the proteosome activity or mutation of putative ubiquitination sites in the CD4 cytoplasmic domain supports this hypothesis.
The most important physiological purpose of CD4 down-modulation is likely not to resist superinfection, but rather to increase viral replication and to promote the release of progeny virions [57,58]. Reduction of CD4 cell surface-expression results in particles with less CD4 and more Env molecules, which probably eases their release from the cell. When using HIV-1 variants with different coreceptor usage obtained from patients, it was found that down-modulation of CD4 was not associated with CCR5-using viruses that are present early in infection, but were characteristic of CXCR4- or CXCR4/CCR5-using viruses that are mostly seen later in infection during the onset of AIDS [59]. In line with this, Lusso et al. [60] found that a macrophage-tropic, non-cytopathic strain of HIV-1 that did not down-regulate CD4, did also not resist subsequent superinfection with a cytopathic HIV-1 strain in a CD4+ T-cell clone (PM1) susceptible to a wide variety of HIV isolates. Furthermore, Nef-genes from AIDS patients were far more efficient in down-regulating CD4 than Nef-alleles from asymptomatic patients [58]. Together, these results raised a question whether CD4 down-modulation in vivo is a significant cause of SIR in HIV-1 infection.
Additional questions about the relevance of CD4 down-regulation come from analysis of the kinetics of CD4 down-modulation in HIV-infected T cells. CD4 down-regulation starts two days after infection and just a few hours before the cells are committed to die (reviewed in [56]). This leaves only a small time span in which CD4 down-modulation of infected transformed T cell lines may interfere with HIV superinfection. Moreover, down-modulation of CD4 in primary T-lymphocytes occurs even later. The half-life of HIV-infected cells in patients has been estimated at 1 to 2 days. Volsky and colleagues [61] demonstrated SIR to be established relatively early between 4 and 24 hrs after primary HIV-infection. Thus, the kinetics of CD4 down-modulation would imply that the established resistance to HIV-1 superinfection is not mediated by CD4 down-modulation. Indeed, HIV-1 SIR has been demonstrated to occur independently of CD4 down-modulation as will be discussed hereafter.
Co-receptor down-regulation could be an alternative SIR mechanism. However, down-regulation of CXCR4 was not observed in culture, and although chronic infection with CCR5-using viruses abrogated CCR5 expression, the effect on superinfection was not tested [59]. A single study suggested that CCR5 down-modulation in an HIV-2 infected cohort of Senegalese women protected them from HIV-1 superinfection [62].
CD4-independent mechanisms contributing to HIV SIR
A few studies have shown cellular resistance to HIV superinfection by mechanisms unrelated to CD4 (reviewed in [63,64]). Volsky and colleagues [61] demonstrated SIR in HIV-1-infected T cells that still expressed substantial levels of CD4. Moreover, non-functional HIV-1 mutants and HIV-1 mutants that could only bind CD4, but not enter the T-cells, did not restrict superinfection of HIV-1 in these cells. The mechanism was HIV-1 specific, as the cells could be infected by other (retro)viruses, indicating that the results could not be explained by a general block of virus replication. HIV-1 mutants that encode inactive Vpu, Vpr and Nef genes were fully active in SIR, ruling out these genes as contributing to HIV SIR.
Another study demonstrated CD4 independent SIR mechanisms in cells infected with a non-producer HIV mutant [65]. CD4 down-modulation in these F12-HIV-infected cells did not change their susceptibility to a challenge HIV strain. However, SIR was established by inhibiting the replication of the superinfecting HIV strain. An additional study evaluated SIR in cells transfected with distinct vectors containing a particular HIV protein [66]. The F12-HIV genes Gag, Vif and Nef were all found to alter replication of the superinfecting HIV-1 strain. Moreover, expression of Nef established complete resistance against the challenge by inhibiting HIV-1 replication at a late stage. Nef mediated inhibition of viral replication has been associated with interference of Gag processing by preventing the cleavage of the p41 Gag precursor protein into p17 (MA) and p24 (CA) [67]. Moreover, the disturbed processing of Gag has been correlated with an altered sub-cellular distribution of F12-Nef compared to the wild-type Nef protein.
CD8 T-cells and HIV superinfection resistance
In animals, antiviral effects, either to the initial or to a second viral infection, are in large mediated by the immune system, making superinfections of animals greatly different from SIR in cells. A 100% effective SIR mechanism could prevent superinfection of a given cell in an animal, but a second virus could infect another, non-infected cell, leading to superinfection of the animal, but not to superinfection of the already infected cell. Neutralizing antibodies restrict re-infection of cells from seropositive donors in culture [68], and cytokines induced by the first viral infection can have a negative effect on subsequent infections [69]. An important immune-mediated inhibition of viral replication is exerted by non-cytotoxic CD8+ T-cells. These cells belong to the innate immune system and were found to suppress HIV-1 replication in CD4+ T-cells by a non-cytotoxic mechanism mediated by a soluble antiviral factor, provisionally named CAF [70] (for reviews see: [71-73]). Until now, the identity of CAF, short for CD8+-cell antiviral factor, has not been resolved, but it suppresses transcription of viral RNA [74,75], is found in both healthy persons and in asymptomatic HIV-1 infected patients [76], can be inhibited by protease inhibitors [77], and strongly suppresses HIV-1/HIV-2 superinfection in culture [78], and as such is included in this review. The mechanism is not virus or species specific, and is also operational in vivo. It has been found in HIV-2 infected baboons [79], and in FIV (feline immunodeficiency virus)-infected cats [80]. HIV-2 infected PBMC from pig-tailed macaques, however, can be superinfected with another strain of HIV-2 in vitro in the presence of CD8+ T-cells [81]. Furthermore, 80–100% of chimpanzees experimentally infected with HIV-1 could be superinfected after 8 to 64 months with a same or different viral subtype despite a fully functional immune system (reviewed in [82]).
Besides CAF as soluble factor, the studies by Locher et al. [79] and Chun et al. [83] suggest that contact between CD4+ and CD8+ cells is important for inhibiting viral replication, including HIV-1 superinfection. During disease progression, the anti-HIV effect of the CD8 T-cells is gradually lost [76,84], as is their ability to suppress superinfection [78], which is probably due to a functional impairment of the (HIV-specific) CD8+ cells in the AIDS phase [85].
HIV superinfection in vivo
HIV-1 can be classified into three distinct groups based on genome sequences; M (major group), O (outlier group) and N (non-M/non-O group), which can be further subdivided into different subtypes (reviewed in [86]). The M group represent the major HIV-1 strains responsible for the worldwide spread of AIDS, and encompasses at least 10 distinct subtypes. The O group represents a minority of the HIV-1 strains and is found in approximately 2–5% of HIV-1-infected individuals in West and Central Africa.
In the late 80's and early 90's of the last century, a number of primate models demonstrated the possibility of HIV-1 superinfection in vivo, a phenomenon that was later also described in humans (reviewed in [2]). Several papers report HIV-1 dual infections as co-infections and not superinfections, as successive infection with two different viruses is often difficult to prove due to limited sampling. It is likely that in a patient, a second virus infects cells that are not infected by the resident virus. Superinfection of HIV-1 in humans can be classified as intra-subtype-, inter-subtype- or inter-strain (M/O)- superinfection. Three studies reported HIV-1 group B-infected individuals to be infected by a distinct subtype B virus [87-89]. HIV-1 subtype B superinfection occurred in two cases in the absence of any antiviral drugs, and in one case during treatment interruptions. A multiple drug-resistant virus was the initial infecting clade B virus in two patients. In all cases, the appearance of the second virus resulted in a decline in CD4+ T-cell counts and an increase in HIV-1 plasma levels. Three cases of HIV-1 superinfection with different subtypes of HIV-1 group M, all with subtype B and CRF01_AE, have been reported so far (reviewed in [2]). A HIV-1 triple infection was recently reported in a Dutch patient practising unsafe sex [90]. One year after the original infection with a subtype B strain, this patient was superinfected with a second subtype B strain, and again a year later another superinfection occurred, this time with subtype CRF01_AE. Only the second superinfection resulted in an increase in viral plasma load and a decrease in CD4+- cell counts and was accompanied by flu-like symptoms. Another triply infected individual, this time from Tanzania, was infected with a subtype C strain and two divergent subtype A strains [91]. However, in this patient it was not clear whether the triple infection was the result of superinfection or of simultaneous infection.
Thus, different HIV-1 group M subtypes are able to establish superinfection resulting in all cases in increased disease progression. Several studies have identified individuals who are dually infected with two distinct HIV-1 strains. In 1999, a dual M/O infected Cameroonian patient was identified [92], followed by five additional M/O dually infected individuals [93,94], and one O/M superinfected patient [95]. Apart from these anecdotal reports, several studies have attempted to study superinfection rates in cohorts of highly exposed individuals. In two cohorts, and in one study involving 14 HIV-seroconcordant couples, no evidence for superinfection was found [96-98], but several other studies report significant rates of superinfection in recently infected individuals. Three cohorts of intravenous drug users showed a 2.5–5% incidence of HIV-1 superinfection [99-101]. A 19% incidence was scored in a cohort of female sex workers within three months of primary infection [102]. However, the latter superinfections were transient, and no evidence of dual infection was seen after 24 months of follow-up [102]. A transient subtype B superinfection was also apparent in one of the intravenous drug users [99]. The incidence of HIV-1 superinfection is probably increasing as more people become infected, as this enhances the chance of meeting an already infected partner.
Viral recombinants, which are an indicator of superinfection on a cellular level, have been reported from the beginning of the epidemic. It has been suggested that recombination is an important viral evolutionary strategy for HIV, and may be considered a key aspect of viral reproduction, so-called "viral sex" [103]. Recombinants provide strong evidence that cellular SIR is not absolute, i.e. in patients some cells are superinfected at some frequency.
The occurrence of HIV-1 superinfection in humans raises questions about the possibility of developing an effective HIV-1 vaccine, both because of the obvious lack of protection of an already infected individual to a second infection (reviewed in: [82,104]). Nowadays, at least 15 circulating recombinant forms have been recognized within HIV-1 group M [2], and many more exist in individual patients. Fang and colleagues [105] described an A/C recombinant HIV-1 virus that was formed in a female sex worker, who was superinfected with HIV-1 subtype C after primary infection with HIV-1 subtype A. The development of new HIV-1 recombinants could also quickly alter various properties of HIV-1, such as cell tropism, viral pathogenicity, antiretroviral drug susceptibility and disease progression.
The studies described here clearly demonstrated HIV-1 superinfection in humans both with different HIV-1 strains and with closely related HIV-1 subtypes. In an asymptomatic patient, a large reservoir of uninfected cells is available for infection by a second virus, as only 1:2,500 to 1:100,000 CD4 cells are estimated to be infected by HIV [106,107]. During disease progression, substantially more virus is produced, and more CD4+ cells become infected [107]. Thus, in the later stages of HIV infection, when both viruses produce a significant amount of progeny, uninfected cells can become dually infected with both virus strains, enabling the formation of recombinant forms. In other cells of the same organism, SIR might be operational and prevent a second infection. Indeed, in splenocytes of two HIV patients, an average of three to four HIV-1 proviruses was found [108]. Sequence analysis showed that proviruses in a single cell were often genetically distinct, and gave rise to recombinants [108].
Discussion
Are mechanisms of SIR comparable among retroviruses?
Apart from the immune response, other cellular mechanisms are operational to prevent superinfection of cells by a second, related virus. SIR mechanisms from three retroviruses, from simple to complex, have been reviewed here. Are there any general lessons to be learned from these studies? For MuLV, a simple retrovirus that contains no accessory genes, SIR mechanisms have been deduced for two viral genes, Gag and Env that were captured by the host. Expression of these genes prevents infection of the cells by MuLV, probably by interfering with viral entry and reverse transcription. For FV, a more complex virus, the accessory gene-encoded protein Bet induces SIR, as did expression of the Env protein. The situation with HIV, the most complex retrovirus of the three, is less clear. Receptor down-modulation occurs late in infection, induced either by the Env, Vpu or Nef proteins, but this does not seem to be the principal SIR mechanism. It may instead be more important for efficient production of virions. HIV-specific, CD4-independent superinfection resistance has been described that occurs early after initial infection, but the proteins involved have not been identified conclusively. One study ruled out Vpu, Vpr and Nef, while another study showed that expression of Nef induced complete resistance against a challenging HIV strain, possibly by interfering with Gag processing. In the latter study, Gag and Vif expression was also found to interfere with viral replication.
Thus, no general picture regarding SIR mechanisms emerges from the study of these retroviruses. Although Env expression is often found to interfere with infection, simply by occupying the viral receptor, the accessory proteins play a more prominent role in complex retroviruses. Especially for HIV, the mechanism is far from clear, and multiple viral proteins may be involved. In no instances have specific host factors been identified.
SIR and clinical HIV superinfection
The most important questions regarding HIV superinfections in a clinical sense are how often do they occur, and what are the consequences? Also, is in vitro research into SIR translatable into clinical practice?
Concerning the first question, if recombination is a valid viral evolutionary strategy, more HIV superinfections may occur than we detect. Two papers report transient superinfections, where after a short time of proven double infection in asymptomatic patients, only a single virus is detected later on [99,102]. If transient superinfections are common, they add to our underestimation of the phenomenon.
Ideally, superinfections should be prevented by the phenomenon of SIR as well as by the immune system. However, SIR cannot protect every target cell in an organism, as only infected cells can display SIR. Neutralizing antibodies and/or virus-specific CD8+ cell response against the first virus do, unfortunately, not seem to prevent HIV superinfection [2].
In studies of vaccinated macaques, a window period for superinfection was found. Monkeys challenged with a second SIV strain later than 10 days [109] or 4 weeks [81] after the first SIV infection, resisted superinfection, whereas all earlier challenges resulted in superinfection. In studies where the animals were challenged much later (15–122 weeks), all monkeys except one in the 15-week challenge group were resistant to superinfection, irrespective of the infection route [110-113]. However, in humans it is questionable whether a such a window period is operational, as for example in the study of Ramos et al. [114], one subject became infected 3–6 weeks after the initial infection, while a second patient was superinfected 5–9 months after seroconversion. Also, in the study by Yerly et al. [99], superinfections occurred years after the initial infection. In the chronic phase of infection, only a small fraction of susceptible cells are infected and many remain available to host a second virus. During disease progression, as CD4+ cells and the CD4 levels of infected cells decline, the patient should become less susceptible to superinfection, also because pathologic symptoms decrease the risk of re-exposure. Possibly, every patient is susceptible to HIV superinfection at some time, with the risk of re-exposure being the main limiting factor. It could also be that superinfected patients have some molecular defect that allows them to establish a second productive infection, or that the primary HIV strain is defective in SIR induction. That would imply that most HIV-infected individuals possess some resistance mechanism, and that the few identified HIV superinfected individuals among the large groups of HIV-infected participants are exceptions. However, in chronically infected patients, a productive HIV superinfection could be regarded as an opportunistic infection that warrants the diagnosis of AIDS. Here it is important to note that HIV-1 superinfection is associated with an increased viral load, a decrease in CD4+ T cell count, and increased disease progression in most cases. A shorter time to death was seen in HIV-2 dually infected monkeys compared to animals that resisted superinfection [81]. So, a productive HIV superinfection should be considered as a marker of disease progression and the start of the AIDS phase.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
ACvdK designed the review, MN and ACvdK drafted the manuscript, and BB critically revised the manuscript.
Figure 1 Schematic of three possible interference mechanisms mediated by Fv1 expression. Although the mechanism of Fv1 interference is still poorly understood and firm experimental evidence is lacking, several likely routes can be envisaged. Route A depicts the binding of Fv1 to CA, thereby restricting CA participation in the integration of the pre-integration complex (PIC) of MuLV DNA. In favour of this model, crystallographic studies recently suggested that a potential Fv1 binding domain exists in the MuLV CA [18]. Alternatively, if yet undetermined CA helper factors (CAhf) are needed during CA mediated integration of the PIC; binding of Fv1 to CAhf would prevent CAhf to assist CA during integration of the PIC (route B). A third possible route would involve direct binding of Fv1 to the PIC, thereby changing its conformation, and restricting it from further processing during CA-mediated integration (route C).
Figure 2 Schematic of Fv4 mediated interference of MuLV infection. Fv4 expression results in sustained levels of Fv4 Env proteins in the cytoplasm. Binding of the mCAT1 receptor by Fv4 Env proteins, either in the cytoplasm or at the cell surface (the exact location of interaction is unresolved, which is represented by question marks), prevents MuLV Env to interact with mCAT1, as either the receptor is already occupied by Fv4, or it cannot reach the cell surface when bound to Fv4 in the cytoplasm.
Figure 3 Expression of Bet and Tas in FV susceptible and restricted cell lines. FV susceptible cell lines containing abundant concentrations of Bet and low concentrations of Tas are still able to enhance the LTR and internal promoter (IP) (panel A). Restricted FV cell lines are associated with reduced LTR and IP activity (panel B). Bet-mediated inhibition of IP activation results in reduced concentrations of Tas and consequently further inhibition of IP activity. The underlying mechanism of Bet-mediated inhibition of IP could be a negative control on transcription or translation of the Tas gene as indicated by a question mark.
Acknowledgements
The authors thank two anonymous reviewers for improving the manuscript.
==== Refs
Goff SP Genetic control of retrovirus susceptibility in mammalian cells Annu Rev Genet 2004 38 61 85 15568972 10.1146/annurev.genet.38.072902.094136
Allen TM Altfeld M HIV-1 superinfection J Allergy Clin Immunol 2003 112 829 835 14610465 10.1016/j.jaci.2003.08.037
Gross L "Spontaneous" leukemia developing in C3H mice following inoculation in infancy, with AK-leukemic extracts, or AK-embrvos Proc Soc Exp Biol Med 1951 76 27 32 14816382
Tailor CS Lavillette D Marin M Kabat D Cell surface receptors for gammaretroviruses Curr Top Microbiol Immunol 2003 281 29 106 12932075
Eiden MV Farrell K Warsowe J Mahan LC Wilson CA Characterization of a naturally occurring ecotropic receptor that does not facilitate entry of all ecotropic murine retroviruses J Virol 1993 67 4056 4061 8510216
Lyu MS Kozak CA Genetic basis for resistance to polytropic murine leukemia viruses in the wild mouse species Mus castaneus J Virol 1996 70 830 833 8551621
Marin M Tailor CS Nouri A Kozak SL Kabat D Polymorphisms of the cell surface receptor control mouse susceptibilities to xenotropic and polytropic leukemia viruses J Virol 1999 73 9362 9368 10516044
Best S Le Tissier P Towers G Stoye JP Positional cloning of the mouse retrovirus restriction gene Fv1 Nature 1996 382 826 829 8752279 10.1038/382826a0
Goff SP Retrovirus restriction factors Mol Cell 2004 16 849 859 15610729 10.1016/j.molcel.2004.12.001
Lilly F Susceptibility to two strains of Friend leukemia virus in mice Science 1967 155 461 462 6015694
Kozak CA Chakraborti A Single amino acid changes in the murine leukemia virus capsid protein gene define the target of Fv1 resistance Virology 1996 225 300 305 8918916 10.1006/viro.1996.0604
Pryciak PM Varmus HE Fv-1 restriction and its effects on murine leukemia virus integration in vivo and in vitro J Virol 1992 66 5959 5966 1326652
Goff SP Operating under a Gag order: a block against incoming virus by the Fv1 gene Cell 1996 86 691 693 8797814 10.1016/S0092-8674(00)80141-5
Craven RC Parent LJ Dynamic interactions of the Gag polyprotein Curr Top Microbiol Immunol 1996 214 65 94 8791725
Stoye JP Fv1, the mouse retrovirus resistance gene Rev Sci Tech 1998 17 269 277 9638816
Nisole S Saib A Early steps of retrovirus replicative cycle Retrovirology 2004 1 9 15169567 10.1186/1742-4690-1-9
Yap MW Stoye JP Intracellular localisation of Fv1 Virology 2003 307 76 89 12667816 10.1016/S0042-6822(02)00053-3
Taylor WR Stoye JP Consensus structural models for the amino terminal domain of the retrovirus restriction gene Fv1 and the murine leukaemia virus capsid proteins BMC Struct Biol 2004 4 1 14715087 10.1186/1472-6807-4-1
Yap MW Nisole S Stoye JP A single amino acid change in the SPRY domain of human Trim5alpha leads to HIV-1 restriction Curr Biol 2005 15 73 78 15649369 10.1016/j.cub.2004.12.042
Suzuki S FV-4: a new gene affecting the splenomegaly induction by Friend leukemia virus Jpn J Exp Med 1975 45 473 478 1232464
Kai K Ikeda H Yuasa Y Suzuki S Odaka T Mouse strain resistant to N-, B-, and NB-tropic murine leukemia viruses J Virol 1976 20 436 440 185426
Inaguma Y Miyashita N Moriwaki K Huai WC Jin ML He XQ Ikeda H Acquisition of two endogenous ecotropic murine leukemia viruses in distinct Asian wild mouse populations J Virol 1991 65 1796 1802 1672163
Ikeda H Sato H Odaka T Mapping of the Fv-4 mouse gene controlling resistance to murine leukemia viruses Int J Cancer 1981 28 237 240 6797957
Ikeda H Odaka T Cellular expression of murine leukemia virus gp70-related antigen on thymocytes of uninfected mice correlates with Fv-4 gene-controlled resistance to Friend leukemia virus infection Virology 1983 128 127 139 6192586 10.1016/0042-6822(83)90324-0
Ikeda H Laigret F Martin MA Repaske R Characterization of a molecularly cloned retroviral sequence associated with Fv-4 resistance J Virol 1985 55 768 777 2991595
Ikeda H Sugimura H Fv-4 resistance gene: a truncated endogenous murine leukemia virus with ecotropic interference properties J Virol 1989 63 5405 5412 2555565
Limjoco TI Dickie P Ikeda H Silver J Transgenic Fv-4 mice resistant to Friend virus J Virol 1993 67 4163 4168 8510219
Kitagawa M Aizawa S Kamisaku H Ikeda H Hirokawa K Sado T Cell-free transmission of Fv-4 resistance gene product controlling Friend leukemia virus-induced leukemogenesis: a unique mechanism for interference with viral infection Blood 1995 86 1557 1563 7632964
Masuda M Yoshikura H Construction and characterization of the recombinant Moloney murine leukemia viruses bearing the mouse Fv-4 env gene J Virol 1990 64 1033 1043 2304138
Taylor GM Gao Y Sanders DA Fv-4: identification of the defect in Env and the mechanism of resistance to ecotropic murine leukemia virus J Virol 2001 75 11244 11248 11602766 10.1128/JVI.75.22.11244-11248.2001
Kim JW Closs EI Albritton LM Cunningham JM Transport of cationic amino acids by the mouse ecotropic retrovirus receptor Nature 1991 352 725 728 1652100 10.1038/352725a0
Wang H Kavanaugh MP North RA Kabat D Cell-surface receptor for ecotropic murine retroviruses is a basic amino-acid transporter Nature 1991 352 729 731 1908564 10.1038/352729a0
Yamaguchi S Hasegawa M Suzuki T Ikeda H Aizawa S Hirokawa K Kitagawa M In vivo distribution of receptor for ecotropic murine leukemia virus and binding of envelope protein of Friend Murine leukemia virus Arch Virol 2003 148 1175 1184 12756622 10.1007/s00705-003-0017-9
Jung YT Lyu MS Buckler-White A Kozak CA Characterization of a polytropic murine leukemia virus proviral sequence associated with the virus resistance gene Rmcf of DBA/2 mice J Virol 2002 76 8218 8224 12134027 10.1128/JVI.76.16.8218-8224.2002
Wu T Yan Y Kozak CA Rmcf2, a Xenotropic Provirus in the Asian Mouse Species Mus castaneus, Blocks Infection by Polytropic Mouse Gammaretroviruses J Virol 2005 79 9677 9684 16014929 10.1128/JVI.79.15.9677-9684.2005
Delelis O Lehmann-Che J Saib A Foamy viruses--a world apart Curr Opin Microbiol 2004 7 400 406 15358259 10.1016/j.mib.2004.06.009
Meiering CD Linial ML Reactivation of a complex retrovirus is controlled by a molecular switch and is inhibited by a viral protein Proc Natl Acad Sci U S A 2002 99 15130 15135 12415120 10.1073/pnas.242491999
Saib A Koken MH van der SP Peries J de The H Involvement of a spliced and defective human foamy virus in the establishment of chronic infection J Virol 1995 69 5261 5268 7636968
Russell RA Wiegand HL Moore MD Schafer A McClure MO Cullen BR Foamy virus Bet proteins function as novel inhibitors of the APOBEC3 family of innate antiretroviral defense factors J Virol 2005 79 8724 8731 15994766 10.1128/JVI.79.14.8724-8731.2005
Bock M Heinkelein M Lindemann D Rethwilm A Cells expressing the human foamy virus (HFV) accessory Bet protein are resistant to productive HFV superinfection Virology 1998 250 194 204 9770433 10.1006/viro.1998.9362
Picard-Maureau M Jarmy G Berg A Rethwilm A Lindemann D Foamy virus envelope glycoprotein-mediated entry involves a pH-dependent fusion process J Virol 2003 77 4722 4730 12663779 10.1128/JVI.77.8.4722-4730.2003
Moebes A Enssle J Bieniasz PD Heinkelein M Lindemann D Bock M McClure MO Rethwilm A Human foamy virus reverse transcription that occurs late in the viral replication cycle J Virol 1997 71 7305 7311 9311807
Berg A Pietschmann T Rethwilm A Lindemann D Determinants of foamy virus envelope glycoprotein mediated resistance to superinfection Virology 2003 314 243 252 14517077 10.1016/S0042-6822(03)00401-X
Herchenroder O Moosmayer D Bock M Pietschmann T Rethwilm A Bieniasz PD McClure MO Weis R Schneider J Specific binding of recombinant foamy virus envelope protein to host cells correlates with susceptibility to infection Virology 1999 255 228 236 10069948 10.1006/viro.1998.9570
Hill CL Bieniasz PD McClure MO Properties of human foamy virus relevant to its development as a vector for gene therapy J Gen Virol 1999 80 ( Pt 8) 2003 2009 10466797
Saib A Peries J de The H A defective human foamy provirus generated by pregenome splicing EMBO J 1993 12 4439 4444 8223453
Saib A Neves M Giron ML Guillemin MC Valla J Peries J Canivet M Long-term persistent infection of domestic rabbits by the human foamy virus Virology 1997 228 263 268 9123833 10.1006/viro.1996.8383
Meiering CD Linial ML Historical perspective of foamy virus epidemiology and infection Clin Microbiol Rev 2001 14 165 176 11148008 10.1128/CMR.14.1.165-176.2001
Regad T Saib A Lallemand-Breitenbach V Pandolfi PP de The H Chelbi-Alix MK PML mediates the interferon-induced antiviral state against a complex retrovirus via its association with the viral transactivator EMBO J 2001 20 3495 3505 11432836 10.1093/emboj/20.13.3495
Meiering CD Linial ML The promyelocytic leukemia protein does not mediate foamy virus latency in vitro J Virol 2003 77 2207 2213 12525655 10.1128/JVI.77.3.2207-2213.2003
Doms RW Trono D The plasma membrane as a combat zone in the HIV battlefield Genes Dev 2000 14 2677 2688 11069884 10.1101/gad.833300
Hoxie JA Alpers JD Rackowski JL Huebner K Haggarty BS Cedarbaum AJ Reed JC Alterations in T4 (CD4) protein and mRNA synthesis in cells infected with HIV Science 1986 234 1123 1127 3095925
Salmon P Olivier R Riviere Y Brisson E Gluckman JC Kieny MP Montagnier L Klatzmann D Loss of CD4 membrane expression and CD4 mRNA during acute human immunodeficiency virus replication J Exp Med 1988 168 1953 1969 3264318 10.1084/jem.168.6.1953
Le Guern M Levy JA Human immunodeficiency virus (HIV) type 1 can superinfect HIV-2-infected cells: pseudotype virions produced with expanded cellular host range Proc Natl Acad Sci U S A 1992 89 363 367 1346069
Oldridge J Marsh M Nef--an adaptor adaptor? Trends Cell Biol 1998 8 302 305 9704405 10.1016/S0962-8924(98)01318-X
Lama J The physiological relevance of CD4 receptor down-modulation during HIV infection Curr HIV Res 2003 1 167 184 15043201 10.2174/1570162033485276
Benson RE Sanfridson A Ottinger JS Doyle C Cullen BR Downregulation of cell-surface CD4 expression by simian immunodeficiency virus Nef prevents viral super infection J Exp Med 1993 177 1561 1566 8098729 10.1084/jem.177.6.1561
Arganaraz ER Schindler M Kirchhoff F Cortes MJ Lama J Enhanced CD4 down-modulation by late stage HIV-1 nef alleles is associated with increased Env incorporation and viral replication J Biol Chem 2003 278 33912 33919 12816953 10.1074/jbc.M303679200
Chenine AL Sattentau Q Moulard M Selective HIV-1-induced downmodulation of CD4 and coreceptors Arch Virol 2000 145 455 471 10795515 10.1007/s007050050039
Lusso P Cocchi F Balotta C Markham PD Louie A Farci P Pal R Gallo RC Reitz MSJ Growth of macrophage-tropic and primary human immunodeficiency virus type 1 (HIV-1) isolates in a unique CD4+ T-cell clone (PM1): failure to downregulate CD4 and to interfere with cell-line-tropic HIV-1 J Virol 1995 69 3712 3720 7745720
Volsky DJ Simm M Shahabuddin M Li G Chao W Potash MJ Interference to human immunodeficiency virus type 1 infection in the absence of downmodulation of the principal virus receptor, CD4 J Virol 1996 70 3823 3833 8648718
Shea A Sarr DA Jones N Penning L Eisen G Gueye-Ndiaye A Mboup S Kanki P Cao H CCR5 receptor expression is down-regulated in HIV type 2 infection: implication for viral control and protection AIDS Res Hum Retroviruses 2004 20 630 635 15242539 10.1089/0889222041217383
Potash MJ Volsky DJ Viral interference in HIV-1 infected cells Rev Med Virol 1998 8 203 211 10398509 10.1002/(SICI)1099-1654(1998100)8:4<203::AID-RMV224>3.0.CO;2-#
Saha K Volsky DJ Matczak E Resistance against syncytium-inducing human immunodeficiency virus type 1 (HIV-1) in selected CD4(+) T cells from an HIV-1-infected nonprogressor: evidence of a novel pathway of resistance mediated by a soluble factor(s) that acts after virus entry J Virol 1999 73 7891 7898 10438887
Federico M Nappi F Bona R D'Aloja P Verani P Rossi GB Full expression of transfected nonproducer interfering HIV-1 proviral DNA abrogates susceptibility of human He-La CD4+ cells to HIV Virology 1995 206 76 84 7831838 10.1016/S0042-6822(95)80021-2
D'Aloja P Olivetta E Bona R Nappi F Pedacchia D Pugliese K Ferrari G Verani P Federico M gag, vif, and nef genes contribute to the homologous viral interference induced by a nonproducer human immunodeficiency virus type 1 (HIV-1) variant: identification of novel HIV-1-inhibiting viral protein mutants J Virol 1998 72 4308 4319 9557721
Fackler OT D'Aloja P Baur AS Federico M Peterlin BM Nef from human immunodeficiency virus type 1(F12) inhibits viral production and infectivity J Virol 2001 75 6601 6608 11413327 10.1128/JVI.75.14.6601-6608.2001
Tremblay M Numazaki K Li XG Gornitsky M Hiscott J Wainberg MA Resistance to infection by HIV-1 of peripheral blood mononuclear cells from HIV-1-infected patients is probably mediated by neutralizing antibodies J Immunol 1990 145 2896 2901 1698866
Shirazi Y Pitha PM Alpha interferon inhibits early stages of the human immunodeficiency virus type 1 replication cycle J Virol 1992 66 1321 1328 1738192
Walker CM Erickson AL Hsueh FC Levy JA Inhibition of human immunodeficiency virus replication in acutely infected CD4+ cells by CD8+ cells involves a noncytotoxic mechanism J Virol 1991 65 5921 5927 1920621
Mackewicz C Levy JA CD8+ cell anti-HIV activity: nonlytic suppression of virus replication AIDS Res Hum Retroviruses 1992 8 1039 1050 1354466
Vella C Daniels RS CD8+ T-cell-mediated non-cytolytic suppression of human immuno-deficiency viruses Curr Drug Targets Infect Disord 2003 3 97 113 12769788 10.2174/1568005033481196
Ahmed RK Biberfeld G Thorstensson R Innate immunity in experimental SIV infection and vaccination Mol Immunol 2005 42 251 258 15488612 10.1016/j.molimm.2004.06.027
Mackewicz CE Blackbourn DJ Levy JA CD8+ T cells suppress human immunodeficiency virus replication by inhibiting viral transcription Proc Natl Acad Sci U S A 1995 92 2308 2312 7534418
Mackewicz CE Patterson BK Lee SA Levy JA CD8(+) cell noncytotoxic anti-human immunodeficiency virus response inhibits expression of viral RNA but not reverse transcription or provirus integration J Gen Virol 2000 81 1261 1264 10769068
Mackewicz CE Ortega HW Levy JA CD8+ cell anti-HIV activity correlates with the clinical state of the infected individual J Clin Invest 1991 87 1462 1466 1707063
Mackewicz CE Craik CS Levy JA The CD8+ cell noncytotoxic anti-HIV response can be blocked by protease inhibitors Proc Natl Acad Sci U S A 2003 100 3433 3438 12624186 10.1073/pnas.0630379100
Barker E Bossart KN Locher CP Patterson BK Levy JA CD8+ cells from asymptomatic human immunodeficiency virus-infected individuals suppress superinfection of their peripheral blood mononuclear cells J Gen Virol 1996 77 ( Pt 12) 2953 2962 9000085
Locher CP Blackbourn DJ Barnett SW Murthy KK Cobb EK Rouse S Greco G Reyes-Teran G Brasky KM Carey KD Levy JA Superinfection with human immunodeficiency virus type 2 can reactivate virus production in baboons but is contained by a CD8 T cell antiviral response J Infect Dis 1997 176 948 959 9333153
Hohdatsu T Sasagawa T Yamazaki A Motokawa K Kusuhara H Kaneshima T Koyama H CD8+ T cells from feline immunodeficiency virus (FIV) infected cats suppress exogenous FIV replication of their peripheral blood mononuclear cells in vitro Arch Virol 2002 147 1517 1529 12181672 10.1007/s00705-002-0827-1
Otten RA Ellenberger DL Adams DR Fridlund CA Jackson E Pieniazek D Rayfield MA Identification of a window period for susceptibility to dual infection with two distinct human immunodeficiency virus type 2 isolates in a Macaca nemestrina (pig-tailed macaque) model J Infect Dis 1999 180 673 684 10438354 10.1086/314968
Fultz PN HIV-1 superinfections: omens for vaccine efficacy? AIDS 2004 18 115 119 15090837 10.1097/00002030-200401020-00014
Chun TW Justement JS Moir S Hallahan CW Ehler LA Liu S McLaughlin M Dybul M Mican JM Fauci AS Suppression of HIV replication in the resting CD4+ T cell reservoir by autologous CD8+ T cells: implications for the development of therapeutic strategies Proc Natl Acad Sci U S A 2001 98 253 258 11136258 10.1073/pnas.98.1.253
Blackbourn DJ Mackewicz CE Barker E Hunt TK Herndier B Haase AT Levy JA Suppression of HIV replication by lymphoid tissue CD8+ cells correlates with the clinical state of HIV-infected individuals Proc Natl Acad Sci U S A 1996 93 13125 13130 8917555 10.1073/pnas.93.23.13125
Kostense S Vandenberghe K Joling J Van Baarle D Nanlohy N Manting E Miedema F Persistent numbers of tetramer+ CD8(+) T cells, but loss of interferon-gamma+ HIV-specific T cells during progression to AIDS Blood 2002 99 2505 2511 11895786 10.1182/blood.V99.7.2505
Tatt ID Barlow KL Nicoll A Clewley JP The public health significance of HIV-1 subtypes AIDS 2001 15 Suppl 5 S59 S71 11816176 10.1097/00002030-200100005-00009
Altfeld M Allen TM Yu XG Johnston MN Agrawal D Korber BT Montefiori DC O'Connor DH Davis BT Lee PK Maier EL Harlow J Goulder PJ Brander C Rosenberg ES Walker BD HIV-1 superinfection despite broad CD8+ T-cell responses containing replication of the primary virus Nature 2002 420 434 439 12459786 10.1038/nature01200
Koelsch KK Smith DM Little SJ Ignacio CC Macaranas TR Brown AJ Petropoulos CJ Richman DD Wong JK Clade B HIV-1 superinfection with wild-type virus after primary infection with drug-resistant clade B virus AIDS 2003 17 F11 F16 12700477 10.1097/00002030-200305020-00001
Yang OO Daar ES Jamieson BD Balamurugan A Smith DM Pitt JA Petropoulos CJ Richman DD Little SJ Brown AJ Human immunodeficiency virus type 1 clade B superinfection: evidence for differential immune containment of distinct clade B strains J Virol 2005 79 860 868 15613314 10.1128/JVI.79.2.860-868.2005
van der Kuyl AC Kozaczynska K Van den Burg R Zorgdrager F Back N Jurriaans S Berkhout B Reiss P Cornelissen M Triple HIV-1 infection New England Journal of Medicine 2005 352 2557 2559 15958817 10.1056/NEJM200506163522420
Gerhardt M Mloka D Tovanabutra S Sanders-Buell E Hoffmann O Maboko L Mmbando D Birx DL McCutchan FE Hoelscher M In-depth, longitudinal analysis of viral quasispecies from an individual triply infected with late-stage human immunodeficiency virus type 1, using a multiple PCR primer approach Journal of Virology 2005 79 8249 8261 15956571 10.1128/JVI.79.13.8249-8261.2005
Peeters M Liegeois F Torimiro N Bourgeois A Mpoudi E Vergne L Saman E Delaporte E Saragosti S Characterization of a highly replicative intergroup M/O human immunodeficiency virus type 1 recombinant isolated from a Cameroonian patient J Virol 1999 73 7368 7375 10438826
Brand D Beby-Defaux A Mace M Brunet S Moreau A Godet C Jais X Cazein F Semaille C Barin F First identification of HIV-1 groups M and O dual infections in Europe AIDS 2004 18 2425 2428 15622319
Yamaguchi J Bodelle P Vallari AS Coffey R McArthur CP Schochetman G Devare SG Brennan CA HIV infections in northwestern Cameroon: identification of HIV type 1 group O and dual HIV type 1 group M and group O infections AIDS Res Hum Retroviruses 2004 20 944 957 15585082 10.1089/aid.2004.20.944
Plantier JC Lemee V Dorval I Gueudin M Braun J Hutin P Ruffault A Simon F HIV-1 group M superinfection in an HIV-1 group O-infected patient AIDS 2004 18 2444 2446 15622327
Gonzales MJ Delwart E Rhee SY Tsui R Zolopa AR Taylor J Shafer RW Lack of detectable human immunodeficiency virus type 1 superinfection during 1072 person-years of observation J Infect Dis 2003 188 397 405 12870121 10.1086/376534
Tsui R Herring BL Barbour JD Grant RM Bacchetti P Kral A Edlin BR Delwart EL Human immunodeficiency virus type 1 superinfection was not detected following 215 years of injection drug user exposure J Virol 2004 78 94 103 14671091 10.1128/JVI.78.1.94-103.2004
Chakraborty B Valer L De Mendoza C Soriano V Quinones-Mateu ME Failure to detect human immunodeficiency virus type 1 superinfection in 28 HIV-seroconcordant individuals with high risk of reexposure to the virus AIDS Res Hum Retroviruses 2004 20 1026 1031 15585092 10.1089/aid.2004.20.1026
Yerly S Jost S Monnat M Telenti A Cavassini M Chave JP Kaiser L Burgisser P Perrin L HIV-1 co/super-infection in intravenous drug users AIDS 2004 18 1413 1421 15199317 10.1097/01.aids.0000131330.28762.0c
Smith DM Wong JK Hightower GK Ignacio CC Koelsch KK Daar ES Richman DD Little SJ Incidence of HIV superinfection following primary infection JAMA 2004 292 1177 1178 15353529 10.1001/jama.292.10.1177
Hu DJ Subbarao S Vanichseni S Mock PA Ramos A Nguyen L Chaowanachan T Griensven F Choopanya K Mastro TD Tappero JW Frequency of HIV-1 dual subtype infections, including intersubtype superinfections, among injection drug users in Bangkok, Thailand AIDS 2005 19 303 308 15718841
Grobler J Gray CM Rademeyer C Seoighe C Ramjee G Karim SA Morris L Williamson C Incidence of HIV-1 dual infection and its association with increased viral load set point in a cohort of HIV-1 subtype C-infected female sex workers J Infect Dis 2004 190 1355 1359 15346349 10.1086/423940
Burke DS Recombination in HIV: an important viral evolutionary strategy Emerg Infect Dis 1997 3 253 259 9284369
Blackard JT Cohen DE Mayer KH Human immunodeficiency virus superinfection and recombination: current state of knowledge and potential clinical consequences Clin Infect Dis 2002 34 1108 1114 11915000 10.1086/339547
Fang G Weiser B Kuiken C Philpott SM Rowland-Jones S Plummer F Kimani J Shi B Kaul R Bwayo J Anzala O Burger H Recombination following superinfection by HIV-1 AIDS 2004 18 153 159 15075531 10.1097/00002030-200401230-00003
Brinchmann JE Albert J Vartdal F Few infected CD4+ T cells but a high proportion of replication-competent provirus copies in asymptomatic human immunodeficiency virus type 1 infection J Virol 1991 65 2019 2023 1672165
Yerly S Chamot E Hirschel B Perrin LH Quantitation of human immunodeficiency virus provirus and circulating virus: relationship with immunologic parameters J Infect Dis 1992 166 269 276 1353103
Jung A Maier R Vartanian JP Bocharov G Jung V Fischer U Meese E Wain-Hobson S Meyerhans A Multiply infected spleen cells in HIV patients Nature 2002 418 144 12110879 10.1038/418144a
Stebbings R Berry N Stott J Hull R Walker B Lines J Elsley W Brown S Wade-Evans A Davis G Cowie J Sethi M Almond N Vaccination with live attenuated simian immunodeficiency virus for 21 days protects against superinfection Virology 2004 330 249 260 15527850 10.1016/j.virol.2004.09.026
Titti F Sernicola L Geraci A Panzini G Di Fabio S Belli R Monardo F Borsetti A Maggiorella MT Koanga-Mogtomo M Corrias F Zamarchi R Amadori A Chieco-Bianchi L Verani P Live attenuated simian immunodeficiency virus prevents super-infection by cloned SIVmac251 in cynomolgus monkeys J Gen Virol 1997 78 ( Pt 10) 2529 2539 9349474
Cranage MP Sharpe SA Whatmore AM Polyanskaya N Norley S Cook N Leech S Dennis MJ Hall GA In vivo resistance to simian immunodeficiency virus superinfection depends on attenuated virus dose J Gen Virol 1998 79 ( Pt 8) 1935 1944 9714241
Sernicola L Corrias F Koanga-Mogtomo ML Baroncelli S Di Fabio S Maggiorella MT Belli R Michelini Z Macchia I Cesolini A Cioe L Verani P Titti F Long-lasting protection by live attenuated simian immunodeficiency virus in cynomolgus monkeys: no detection of reactivation after stimulation with a recall antigen Virology 1999 256 291 302 10191194 10.1006/viro.1999.9652
Sharpe SA Cope A Dowall S Berry N Ham C Heeney JL Hopkins D Easterbrook L Dennis M Almond N Cranage M Macaques infected long-term with attenuated simian immunodeficiency virus (SIVmac) remain resistant to wild-type challenge, despite declining cytotoxic T lymphocyte responses to an immunodominant epitope J Gen Virol 2004 85 2591 2602 15302953 10.1099/vir.0.80050-0
Ramos A Hu DJ Nguyen L Phan KO Vanichseni S Promadej N Choopanya K Callahan M Young NL McNicholl J Mastro TD Folks TM Subbarao S Intersubtype human immunodeficiency virus type 1 superinfection following seroconversion to primary infection in two injection drug users J Virol 2002 76 7444 7452 12097556 10.1128/JVI.76.15.7444-7452.2002
|
16107223
|
PMC1224871
|
CC BY
|
2021-01-04 16:36:39
|
no
|
Retrovirology. 2005 Aug 18; 2:52
|
utf-8
|
Retrovirology
| 2,005 |
10.1186/1742-4690-2-52
|
oa_comm
|
==== Front
Respir ResRespiratory Research1465-99211465-993XBioMed Central London 1465-9921-6-801604277410.1186/1465-9921-6-80ResearchExpression profiles of hydrophobic surfactant proteins in children with diffuse chronic lung disease Griese Matthias [email protected] Silja [email protected] Mohammed [email protected] Manuela [email protected] Annika [email protected] Susan [email protected] Michael F [email protected] Michel [email protected] Kinderklinik and Poliklinik, Dr. von Haunersches Kinderspital, Ludwig-Maximilians University, Munich, Germany2 Service de Biochimie et Biologie Moléculaire, Hôpital d'Enfants Armand-Trousseau (AP-HP), Paris, France3 Division of Neonatology, Childrens' Hospital of Philadelphia, University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania 19104-4318, USA4 Pulmonary and Critical Care Division, University of Pennsylvania School of Medicine Philadelphia, Pennsylvania 19104-6160, USA2005 22 7 2005 6 1 80 80 1 3 2005 22 7 2005 Copyright © 2005 Griese et al; licensee BioMed Central Ltd.2005Griese et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Abnormalities of the intracellular metabolism of the hydrophobic surfactant proteins SP-B and SP-C and their precursors may be causally linked to chronic childhood diffuse lung diseases. The profile of these proteins in the alveolar space is unknown in such subjects.
Methods
We analyzed bronchoalveolar lavage fluid by Western blotting for SP-B, SP-C and their proforms in children with pulmonary alveolar proteinosis (PAP, n = 15), children with no SP-B (n = 6), children with chronic respiratory distress of unknown cause (cRD, n = 7), in comparison to children without lung disease (n = 15) or chronic obstructive bronchitis (n = 19).
Results
Pro-SP-B of 25–26 kD was commonly abundant in all groups of subjects, suggesting that their presence is not of diagnostic value for processing defects. In contrast, pro-SP-B peptides cleaved off during intracellular processing of SP-B and smaller than 19–21 kD, were exclusively found in PAP and cRD. In 4 of 6 children with no SP-B, mutations of SFTPB or SPTPC genes were found. Pro-SP-C forms were identified at very low frequency. Their presence was clearly, but not exclusively associated with mutations of the SFTPB and SPTPC genes, impeding their usage as candidates for diagnostic screening.
Conclusion
Immuno-analysis of the hydrophobic surfactant proteins and their precursor forms in bronchoalveolar lavage is minimally invasive and can give valuable clues for the involvement of processing abnormalities in pediatric pulmonary disorders.
SFTPBSFTPCSP-B deficiencySP-Cpro-SP-Cprocessingpulmonary alveolar proteinosis (PAP)unexplained respiratory distressinterstitial lung diseasechildreninfantneonate
==== Body
Introduction
Pulmonary surfactant is a highly surface active complex of lipids and specific proteins, including surfactant proteins (SP-) A, B, C and D [1]. The maintenance of the patency of the airspaces at end-expiration is heavily dependent on the phospholipid components and their interaction with SP-B and SP-C [2]. SP-B is encoded by a single gene (SFTPB) [3] and translated in the alveolar type II cells into a preproprotein (~40 kDa). Post-translational processing of pro-SP-B to yield mature SP-B is a multistep entirely intracellular process involving multiple sites and enzymes [4-7]. SP-C is encoded by the SFTPC gene on chromosome 8 [8] and the SP-C proprotein processing [9-11] is integrally linked to the metabolism of SP-B in that infants and mice with genetic SP-B deficiency exhibit incompletely processed pro-SP-C peptides of 6–14 kDa in intra- and extracellular surfactant [12,13]. In lung homogenates of most infants with SFTPB mutations, aberrant pro-SP-C forms (Mr 6–12 kD) are observed [14]. Similarly, pro-SP-B forms of variable sizes have been detected in lung homogenates from some children with chronic lung disease but were predominantly absent in patients with SFTPB mutations [14].
Bronchoalveolar lavage (BAL) is a commonly used first line diagnostic tool to sample the alveolar space content and this technique is much less invasive than open lung biopsy. Thus the profiles of SP-B, SP-C and their propeptide precursors present in the extracellular, intraalveolar space represent a potential diagnostic tool for assessment of neonatal and childhood lung disease.
Neonates with respiratory distress of unknown cause are likely candidates for abnormalities of SP-B and SP-C metabolism. Similarly, but much less appreciated, SP-B and SP-C abnormalities might play a role in infants or older children with chronic respiratory distress developing beyond the neonatal period. Pediatric pulmonary alveolar proteinosis (PAP) is a rare abnormality of the surfactant metabolism, characterized by the accumulation of large amounts of surfactant in the alveolar space, leading to gas exchange abnormalities [15,16]. In contrast to the adult form of acquired PAP where GM-CSF autoantibodies appear to play a pathogenic role, the causes of pediatric PAP are as yet unresolved. In particular the characteristics of SP-B and SP-C peptides and their precursors in the alveolar space of pediatric patients with lung disease have not been described.
Using defined pediatric patient populations, Western blotting of BAL identified several distinct banding profiles for the hydrophobic surfactant proteins and their precursors. These data support the feasibility of using immunoanalyses of BAL fluid to evaluate chronic pediatric pulmonary disorders in more detail.
Patients, Materials and methods
Patients
The lavage effluents from 15 children without lung disease and 19 children with chronic obstructive bronchitis were used as controls or disease controls, for comparison with the lavage effluents that were available from our previously described cohort of neonatal, pediatric or juvenile patients with respiratory distress of unknown cause. These children were seen in western European medical hospital centers (mainly from France and Germany) and were analyzed for a genetic defect leading to deficiency in SP-B and SP-C [17,18].
The lavage effluents from the children without lung disease were aliquots obtained previously in a study that assessed inflammation in children with chronic tracheostoma in comparison to these controls [19]. The lavage effluents were obtained during anesthesia for elective surgery for minor conditions. The usage of this material and that of the children with chronic bronchitis for this study was approved by the ethics committee at the University of Munich. Written informed consent was obtained from the patients where appropriate from age and from the caregivers.
Children with chronic obstructive bronchitis in whom anomalies of the airways, cystic fibrosis, primary ciliary dyskinesia, gastro-esophageal reflux, immuno deficiencies, allergic asthma and passive smoke exposure were excluded as causes and in whom a lavage was performed during the diagnostic work up, were used as a disease control group. The obstruction was determined by chest auscultation during the course of the disease. Details of these patients are given in table 1.
Table 1 Patient characteristics, lavage protein content and apparent molecular weight of SP-B and SP-C
Children N (males) Age (y) Time of follow up (years), outcome Protein (μg/ml) SP-B Mr of band (kDa) SP-C Mr of band (kDa)
without lung disease 15 (8) 5.4 (0.5–12) not applicable 62 (21–275) 7.1 (5.9–11.6) 4.8 (4.3–5.8)
with chronic obstructive bronchitis 19 (13) 5.3 (1–15) 4 (0.3–10) years, 14/19 better, 3/19 same, 1/19 worse, 1/19 unknown 76 (17–207) 11.0 (8–13.5) 5.2 (3.9–5.6)
with no SP-B 6 (3) neonates 5 pts [2–6] died at 0.3 (0.1–0.4) years, pt [1] alive with corticosteroids 318* (131–2048) no SP-B bands in any pt 5.6 (3.6–6.5) pt [4] no SP-C
with pulmonary alveolar proteinosis 15 (9) 1.4 (0.6–4) Pts [6,10,14,15] died at ages 1.3 and 1.7 years and at 4 and 5 months of age. 11/15 alive with repetitive whole lung lavages and oxygen-dependence 495** (87–2099) 10.5 (8.8–12.5) 4.8 (3.6–5.4)
with chronic respiratory distress of unknown cause 7 (7) neonates, one subject 4 months 4 died at age 8 days to 4 months, 3 [3,6,7] lost on follow up 449* (184–474) 9.7 (6.3–11.2) 5.6 (4.3–7)
All data are medians and range, n.d. = not determined. Significantly higher compared to children without lung disease or children with obstructive bronchitis, which did not differ *p < 0.01, **p < 0.001 by Kruskal-Wallis-Analysis followed by Dunn's multiple comparisons test
From the cases with SP-B deficiency we initially described, sufficient BAL material for analysis was available from 6 neonates (URD 6-II.1, 2-II.1, 7-II.1, 4-II.1, 3-II.1, 9-II.4), now labeled no-SP-B 1–6. All these babies had respiratory distress, and alveolar infiltrates with various degrees of interstitial involvement. A congenital heart disease or a lung disease due to mycoplasma, chlamydia, and viruses had also been ruled out. Details on the subjects are given in table 1. All but 2 subjects had mutations of SP-B as the cause for the SP-B deficiency.
From the cases with pulmonary alveolar proteinosis, sufficient BAL material for analysis was available from 15 children (URD 10-II.1, 11-II.3, 17-II.2, 25-II.3, 19-II.1, 20-II.2, 21-II.1, 16-II.2, 27-II.3, 22-II.1, 26-II.1, 23-II.3, 13-II.1, 13-II.2, 18-II.2), now labeled PAP 1–15. Most of these cases were less severely affected, had dyspnea and progressive cough, sometimes accompanied by cyanosis, finger clubbing, failure to thrive in the younger ones, and asthenia or weight loss in the others. Chest x-ray showed typical alveolar as well as interstitial infiltrates (table 1). In all these patients mutations of SP-B were excluded, 3 patients (PAP 04, PAP 10 and PAP 12) had heterozygous mutations in SFTPC. None of these children was investigated for ABCA3 mutations. All known secondary causes of PAP were excluded.
In addition, 7 subjects with chronic respiratory distress of unknown cause, in the absence of SP-B deficiency or alveolar proteinosis were investigated. BAL was available from 6 (URD 31-II.3, 40-II.1, 36-II.2, 30-II.1, 39-II.1, 37-II.2) of the initial 15 patients and from another infant born at 36 wks of gestation, with acute respiratory distress and development of chronic respiratory distress of unknown cause, after exclusion of SP-B, SP-C deficiency, and pulmonary alveolar proteinosis. None of these children was investigated for ABCA3 mutations. The children were labeled cRD 1–7 and their outcomes are given in table 1.
Bronchoalveolar lavage
Routinely, the fluid recovered from BAL (4 × 1 ml 0.9% NaCl/kg body weight, b.w.) was pooled and the cells separated before analysis. Alternatively, in very sick neonates, repetitive tracheal aspirates after the instillation of 1 ml 0.9%NaCl/kg b.w. were collected over time periods of several hours up to a week, pooled and used for biochemical analyses.
Antisera
All antisera used were polyclonal and raised in rabbits. The antibodies against SP-B (c329) and SP-C (22/96) were gifts from Dr W. Steinhilber, Altana AG, Konstanz, FRG and were used at a dilution of 1:10,000 [20]. The antisera against pro-SP-B were raised against peptides of pro-SP-B, which were also used to determine the specificity of the signals on the immunoblots in all cases. The abbreviations and location of these peptides in the pro-SP-B sequence is indicated in figure 1. NFPROX was raised against SRQPEPEQEPGMSDPL, NFLANK against QARPGPHTQDLSEQQ, both were used at 1:2000 dilution. CFLANK was raised against GPRSPTGEWLPRDSECHLCMS, used at 1:1000 dilution and CTERMB was raised against LDREKCKQFVEQHTPQLLTL, used at 1:5000 dilution. Pro-SP-C was detected by anti-serum used at 1:5.000 dilution and raised previously against ESPPDYSTGPRSQ, i.e. Glu10–Gln23 of the amino acid sequence in pro-SP-C. The characteristics of all these antibodies has been described previously in detail [21-23].
Figure 1 Schematic diagram of pro-SP-B and its processing to SP-B. Upper panel: Indicated are the antibodies used, the symbols for their identification, the amino acid stretches against which the antibodies were developed, and a diagram of the structure of pro-SP-B. Lower panel: The molecular weight and the reactivity of the antibodies (in the absence, but not in the presence of the competing peptides) during Western blotting is indicated. The sizing of the letters used for indication of the molecular weights is proportional to the frequency at which the bands were observed (biggest: common >75% of subjects, 2nd biggest: frequent, in <75 but >50% of the subjects, 3rd biggest: sporadic, in <50 but >25% of the subjects, smallest: rare, in <25% of the subjects). The sequence of SP-B within the pro-SP-B sequence is indicated in pink. All bands were analyzed under reducing conditions.
Surfactant protein characterization
Total protein content of the samples was determined with the Biorad Protein Assay Kit (Biorad, Richmond, CA), which is based on the method by Bradford [24]. Ten to twenty-five μg of total protein were separated under reducing conditions on NuPage10% Bis-Tris gels using a NOVEX X-cell II Mini-Cell system (Novex, San Diego, CA). At least two sets of gels were prepared in parallel for each patient. Following electrophoresis the gels were either silver stained [25], or subjected to Western transfer. For immunodetection, the proteins in the gels were transfered onto a PVDF membrane (ImmobilonP, Millipore, Bedford, MA) with a NuPage Blot module (Novex, San Diego, CA) according to the manufacturers recommendations.
Surfactant proteins and their pro-forms were detected on the PVDF membrane by immunoblot using the polyclonal rabbit antisera described in detail above, and the enhanced chemiluminescence assay (Amersham Biosciences, Buckinghamshire, UK) with horseradish peroxidase conjugated goat anti-rabbit polyclonal anti-IgG (1:10,000; Dianova, Hamburg, FRG).
To verify the specificity of the antibodies used to probe the pro-forms of SP-B and SP-C, a duplicate blot was prepared in each case and probed with an antibody solution containing 1 μM of the peptide, against which the antibody was raised. Antigen specific bands on the blot disappeared under these conditions. The blots were developed by exposure of X-ray film (Hyperfilm ECL, Amersham Biosciences, Buckinghamshire, UK) to the blots.
In the group of controls blots were first incubated with antibody against CTERMB and after that with the SP-B antibody respectively first with antibody against SP-C and after that with the pro-SP-C antibody, with and without competing peptide. In the other groups there were separate blots for each incubation with antibodies against SP-B and SP-C and their proforms, with and without competing peptide. Under these conditions the assay could detect about 2.5 ng of SP-B or SP-C per lane. In several experiments, aliquots of a patient with pro-SP-C forms were run in parallel as a positive control for pro-SP-C forms.
Immunoblots and silver stained gels were scanned with the Fluor-S MultiImager (Biorad, Richmond, CA) gel documentation system, and the resulting images were analyzed with the Software MultiAnalyst (Biorad, Richmond, CA).
Deglycosylation
To determine if the proteins that reacted with the CTERMB antibody on the immunoblots were glycosylated, the samples were deglycosylated before applying them on the gel (4). In brief, 1 unit of recombinant N-glycosidase F (Roche Molecular Biochemicals, Mannheim) was added to 500 μl incubation buffer (100 mM Na-phosphate, 25 mM EDTA, 1% β-mercaptoethanol, 0.5% Triton X-100, 0,1% SDS, pH 7.2). The vacuum dried sample was resuspended in 20 μl of this solution and incubated for 15 h at 37°C. The sample was then vacuum dried and analyzed by Western immunoblot.
Genetic analysis
For SFTPB mutation screening, first the 121ins2 frame-shift mutation was searched using the restriction enzyme cleavage SfuI endonuclease by PCR. In 121ins2-negative patients, SFTPB exons 1–11 and the promoter region were PCR-amplified and the purified PCR products served as templates in the sequencing reaction using Ready Reaction Dye Terminator Cycle Sequencing Kit With AmpliTaq® DNA Polymerase, FS (PEBiosystems, Foster City, CA) with forward and reverse PCR oligonucleotides used as extension primers. Extension products were analyzed using the ABIPRISM™ 310 Genetic analysis System (PEBiosystems), as previously reported in detail [18]. Similarly, SFTPC exons 1–6 were analysed [17].
Statistical analysis
Statistical calculations were performed with the Software GraphPad Prism 4.0 (GraphPad Software, San Diego, CA). Differences in nonparametric values were calculated with the Kruskal-Wallis test. For pair wise comparisons of groups we used Dunn's test (2). Differences in frequencies were calculated with the Fisher exact test. Correlation coefficients were determined according to Pearson. Results with a p ≤ 0.05 were considered significant.
Results
Children without lung disease and children with chronic bronchitis
The children with chronic obstructive bronchitis had a slight increase in neutrophils (3% (2; 15)(data are median and (25.; 75. percentile)) compared to children without lung disease (1% (1; 2); p = 0.035) and a somewhat lower viability (80% (70; 90) and 90% (80; 97) in children without lung disease; p = 0,035). The other variables did not differ and were within the normal range, i.e. children with chronic obstructive bronchitis: total cell count 150/μl (82; 275), macrophages 80% (69; 90) of total cells, lymphocytes 10% (4/14), eosinophils 0% (0; 2) and recovery was 54% (39; 70) and the children without lung disease: total cell count 115/μl (82; 180), macrophages 87% (82; 92) of total cells, lymphocytes 11.5% (7; 14.5), eosinophils 0% (0; 0.5) and recovery was 48% (42; 62).
Mature SP-B was regularly detected in all lavages from normal children and from those with chronic bronchitis at a median molecular weight of 7 kDa (Tab. 1, Fig. 2). Similarly, pro-SP-B forms with a molecular weight of 25–26 kDa were commonly observed using an antibody against the C-terminal flanking propeptide of pro-SP-B (Tab. 2, Fig. 2). Those bands never reacted with NFPROX, but showed reactivity with NFLANK, demonstrating that this was a processing intermediate generated by removal of the proximal N-terminal amino acids. A similar, but somewhat more truncated, 19–21 kDa pro-SP-B fragment was detected sporadically in these children (Tab. 2, Fig. 2). The pro-SP-B forms at 25–26 and 19–21 kDa were glycosylated as treatment with N-glycosidase F resulted in a significant drop in size for both peptides (not shown). A 40–42 kDa form and a 34–36 kDa form of pro-SP-B were rarely detected. Except for a single case when a 9 kDa C-terminal cleavage fragment was observed, in these children no other cleavage products of pro-SP-B processing were identified. Mature SP-C with Mr of 5.0 kDa was present in both controls and children with chronic bronchitis, whereas pro-SP-C forms were never detected in BAL (Tabs. 1 and 2, Fig. 2).
Figure 2 Children with chronic bronchitis. Representative Western blotting pattern of BAL from child with chronic bronchitis (patient control 03). After SDS-PAGE and transfer, the membranes were probed with different antibodies directed against SP-B, certain sequences of the pro-SP-B, in the absences (-) and presence (+) of excess of the peptides, used to raise the antibodies, SP-C and against pro-SP-C, in the absence (-) and presence (+) of excess of the N-terminal peptide, used to raise these antibodies. The numbers next to the lanes indicate the molecular weight in kDa. The arrow heads indicate bands of interest, as described in the text. All bands were analyzed under reducing conditions.
Table 2 Pro-SP-B and pro-SP-C in the comparison groups, i.e. children without lung disease and in children with chronic bronchitis.
pro-SP-B pro-SP-C
Detecting antibody CTERMB CFLANK NFLANK NFPROX
Mr of band 40–42 34–36 25–26 19–21 9 25–26
Children without lung disease (n = 15) 7% [8] 0% 80% [1,3,4,6–8,10–15] 7% [3] nd nd nd no bands
Chronic obstructive bronchitis (n = 19) 5% [14] 26% [12–14,18] 100% [1–19] 37% [4–6,10,13,15,18] 5% [15] 21% [4–6,9] no bands no bands
Percent of subjects with bands and identification numbers of those subjects in whom bands reacting with the anti-pro-SP-B-antibodies CTERMB, NFLANK, CFLANK and NFPROX displaced by the CTERMB, CFLANK, NFLANK or NFPROX peptides, or the anti-pro-SP-C-antibody NPRO-SP-C-C2 and displaced by the respective peptide, were identified. The identification numbers of the patients are given in square brackets []. Numbers in bold refers to bands not identified by the CTERMB antibodies. Due to shortage of lavage material in the normal controls (no lung disease), not all 4 antibodies were tested in this group (nd = not done).
Children with no SP-B
6 of all children investigated did not have SP-B in their lavages. Of these, 4 had lethal mutations of the SFTPB gene, i.e. SP-B deficiency (Tab. 3). Pro-SP-B processing products were not found in patient 5, having a 457delC/121ins2 compound heterozygote mutation (Fig. 3, Tab. 3). Unexpectedly, patients 3 and 6, homozygous for 121ins2, and patient 2 homozygous for 496delG had small but specific (competitive) pro-SP-B bands at about 19–21, 25–26 or 34–36 kDa (Tab. 3). Aberrant pro-SP-C bands previously thought to be diagnostic of SP-B mutations were only detected in 121ins2-mutations but not with 457delG [17,26] or with 496delG mutations.
Table 3 Pro-SP-B and pro-SP-C in children with no SP-B
pro-SP-B pro-SP-C
Detecting antibody CTERMB NFLANK NPROSP-C-C2
Mr of band (kDa) 34–36 25–26 19–21 25–26
Subject Genetic analysis of SFTPB
no SP-B 01 no SFTPB mutation; marker exclusion - ++ - - -
no SP-B 02 496delG homozygote + + + - -
no SP-B 03 121ins2 homozygote - + - - 6 and 7.9 kDa
no SP-B 04 no SFTPB mutation - ++ - - -
no SP-B 05 457delC/121ins2 compound heterozygotes - - - - -
no SP-B 06 121ins2 homozygote - - - + 6.6. and 9 kDa
Bands reacting with the anti-pro-SP-B-antibodies CTERMB, NFLANK, CFLANK and NFPROX displaced by the CTERMB, NFLANK, CFLANK or NFPROX peptides are indicated by "+", or the anti-pro-SP-C-antibody NPRO-SP-C-C2 and displaced by the respective peptide are indicated by the molecular weight directly.
Figure 3 SP-B deficiency. Western blotting of a lavage from patient SP-B 06 homozygous for the 121ins2 SFTPB mutation. After SDS-PAGE and transfer, the membranes were probed with the antibodies indicated. The pro-forms were probed in the absence (-) and presence (+) of an excess of the peptide used to raise this antibody. Note that bands that are not displaced by the competing peptide were not considered as specific bands (marked by an asterisk). The numbers next to the lanes indicate the molecular weight in kDa. The closed arrowheads indicate the absence of SP-B and of proforms of SP-B. Arrows show the presence of SP-C (open arrow) and of abberant pro-SP-C (closed arrows). Some aberrant pro-SP-C can also be seen on the SP-C blot, above the SP-C band, which is indicated by an open arrowhead. All bands were analyzed under reducing conditions.
In the other two infants with no SP-B in the lavages, SFTPB and SFTPC mutations were excluded [17,18]. These patients had significant amounts of pro-SP-B at 25–26 kDa, similar to that observed in the comparison groups. They also did not have pro-SP-C forms in their lavages, providing additional indirect evidence against SP-B processing defects. However, one of these two patients, i.e. patient 4 (Tab. 1), also lacked mature SP-C. This infant died at the age of 1 month from respiratory failure. This case suggests the presence of SP-B and SP-C processing defects arising by means other than from mutations of these genes, i.e. alterations in the protein processing machinery or in the lipid transporters, like ABCA3, as recently shown [27]. The other child (patient 1, Tabs. 1 and 3) is still alive with corticosteroids.
Children with PAP
In all subjects with PAP, except patient 5, antibodies against GM-CSF in their sera or lavages were excluded in the pathogenesis of their disease. Although SP-B was abundantly present and mutations of SFTPB were excluded [18], alterations of SP-B processing from other causes have not been excluded. In general, the same pro-SP-B processing products were observed as in the control and the chronic bronchitis group, however, the 25–26 kDa band was stained by NFLANK at increased frequency (Tab. 4, Fig. 4, lanes 4 and 5). In addition, 15 kDa and 13 kDa bands were present that were only stained by NFPROX. These peptides represent the N-terminal cleaved processing fragments, which were detected only in these patients and not in the respective control group (Tab. 4, Fig. 4, lanes 6 and 7). Three of the PAP patients (PAP 14, PAP 05 and PAP 10) had bands reacting merely with CFLANK or NFLANK. These bands were at 8, 9, 11 and 12 kDa. These may represent imprecisely processed SP-B, still having not completely removed small N- or C-terminal peptide stretches (Figs. 1, Tab. 4).
Table 4 Pro-SP-B and pro-SP-C in 15 children with pulmonary alveolar proteinosis
pro-SP-B pro-SP-C
Detecting antibody CTERMB CFLANK NFLANK NFPROX NPROSP-C-C2
Mr of bands(kDa)
40–42 7% [5] - - - -
34–36 7% [5] - - - -
25–26 93% [1–5,7–15] 20% [3,5,15] 87%+ [1–5,7–12,14,15] - -
19–21 87%* [1–5,7–13,15] 7% [5] 20% [4,5,9] 7% [4] 7% [8]
15 7% [8] - 7% [8] 33%+ [2,8,9,11,12] 7% [8]
13 - - - 20% [3,8,9] -
12 - - 7% [14] - -
11 - 7% [14] - - 7% [8]
9 - 7% [10] 14% [5,10] - -
6 - - - - 7% [4]
Percent of subjects with bands and identification numbers of those subjects in whom bands reacting with the anti-pro-SP-B-antibodies CTERMB, NFLANK, CFLANK, NFPROX and displaced by the CTERMB, NFLANK, CFLANK, NFPROX peptides, or the anti-pro-SP-C-antibody NPRO-SP-C-C2 and displaced by the respective peptide, were identified. Differences in the frequency of bands of all the disease groups were evaluated by the Fisher exact test and those with a P ≤ 0.05 were indicated by an * for comparison with the healthy control group or by a + for comparison with the disease control group, bronchitis (see table 2). The identification numbers of the patients are given in square brackets []. Numbers in bold indicate bands not identified by the CTERMB antibodies.
Figure 4 Children with pulmonary alveolar proteinosis. Western blotting of a lavage from patient PAP 12 (only NFPROX bands) and PAP 04 (all other bands) to demonstrate the most frequent abnormalities. After SDS-PAGE and transfer, the membranes were probed with the antibodies indicated. The pro-forms were probed in the absence (-) and presence (+) of an excess of the peptide used to raise this antibody. Note that bands that are not displaced by the competing peptide were not considered as specific bands and they are marked by an asterisk. The numbers next to the lanes indicate the molecular weights in kDa. The arrowheads indicate the abundance of SP-B, the bands at 19–21 and 25–26 kDa using CTERMB which also react with NFLANK, and some of the break-down fragments reacting with NFPROX which are more frequently seen in this condition and in cRD as compared to the other lung diseases (see figure 5). All bands were analyzed under reducing conditions.
Among the PAP patients, only 2 had consistent pro-SP-C bands (Tab. 4). Subject PAP 08, a patient with a heterozygous SFTPC mutation and previously described in detail, had 3 bands, and subject PAP 04, in whom no SP-C mutation was detected, had one band at 6 kD [17]. Those 2 patients with the SFTPC mutation g.2125G>A [17] had no pro-SP-C bands with this antibody.
Infants with chronic respiratory distress of unknown cause
The infants with chronic respiratory distress of unknown cause had no mutations of SFTPB or SFTPC, and normal SP-B and SP-C in their lavages (Tab. 1). Nevertheless, aberrant pro-SP-C was detected in one of these infants at 9 kDa (Tab. 5). Concerning the processing of pro-SP-B significant deviations from the pattern observed in the control groups were observed in some of these children with cRD. Indeed a pro-SP-B precursor at 40–42 kDa was observed more frequently in these patients (Fig. 5, Tab. 5). Similarly, as in PAP, bands reacting with NFPROX, representing fragments of the cleaved N-terminus, were detected (Tab. 5, Figs. 1 and 5).
Table 5 Pro-SP-B and pro-SP-C in 7 children with chronic respiratory distress of unknown cause (cRD)
pro-SP-B pro-SP-C
Detecting antibody CTERMB CFLANK NFLANK NFPROX NPROSP-C-C2
Mr of bands (kDa)
40–42 57%* [2,5–7] 14% [7] - - -
25–26 71% [2,3,5–7] 38%+ [4,6,7] 57% [2,4,5,6] - -
19–21 - - 14% [6] 14%§ [6] -
15 - - - 29%§ [2,7] -
9 14% [6] 14% [7] - - 14% [6]
3.6 - - - 14%§ [3] -
Percent of subjects with bands and identification numbers of those subjects in whom bands reacting with the anti-pro-SP-B-antibodies CTERMB, NFLANK, CFLANK, NFPROX and displaced by the CTERMB, NFLANK, CFLANK, NFPROX peptides, or the anti-pro-SP-C-antibody NPRO-SP-C-C2 and displaced by the respective peptide, were identified. Differences in the frequency of bands of all the disease groups were evaluated by the Fisher exact test and those with a P < 0.05 were indicated by an * for comparison with the healthy control group or by a + for comparison with the disease control group, bronchitis (see table 2). §indicates a significant difference to the disease control group, bronchitis, when all NFPROX reactive bands were combined (P < 0.01). The identification numbers of the patients are given in square brackets []. Numbers in bold indicate bands not identified by the CTERMB antibodies.
Figure 5 Children with chronic respiratory distress of unknown cause (cRD). Western blotting of a lavage from patient cRD 06 (NFLANK) and from patient cRD 07 (all other blots), performed as described in detail in the legend to figure 4. An asterisk marks non-specific bands, i.e. bands not displaced by the competing peptide. The arrowheads indicate the bands reacting with CTERMB at 40–42 kDa which are more frequently observed in these conditions than in the others. Similarly, with CFLANK, bands are seen at 40–42, 25–26, and 19–21 kDa. Cut off fragments likely generated during protein processing react with NFLANK or NFPROX. All bands were analyzed under reducing conditions.
Discussion
In this study we defined the presence and characteristics of SP-B, SP-C and their processing forms in bronchoalveolar lavages from children with severe chronic respiratory distress and in comparison groups of normal children and children with chronic obstructive bronchitis (Fig. 1). The major findings are the presence of mature SP-B and SP-C in all children, except those with SP-B deficiency, supporting the view that analysis of BAL for these surfactant proteins may aid in the diagnostic work up of children with severe respiratory distress. Overall pro-SP-C forms were rarely detected, and their presence was specific, but not pathognomonic for a SP-B deficiency due to SFTPB mutations. In addition, using epitope specific antisera, we identified unique pro-SP-B forms containing residues 145–160 of proSP-B (i.e. the "NFPROX" epitope) exclusively in BAL from patients with alveolar proteinosis and chronic respiratory distress. Taken together, the data suggest that immunobiochemical analysis of BAL can detect abnormalities in surfactant biosynthesis and metabolism associated with a variety of parenchymal lung diseases.
Of the 6 patients with SP-B deficiency defined as a lack of mature SP-B on Western blotting, 4 had mutations in SFTPB (Tab. 3). Based on our results, the biochemical analysis of BAL fluid for mature SP-B, previously thought to be diagnostic for SP-B deficiency, is not 100% specific, as there are additional cause(s) leading to a lack of SP-B. Possible mechanisms include mutations or secondary changes in regulatory elements or other defects in the synthesis and secretion of surfactant, as recently shown for the ABCA3 transporter [27].
An important finding of this study is the regular detection of certain pro-SP-B peptides in BAL from children without bronchoalveolar disease. Most prominent was a 25–26 kDa band, detected in almost all patients. This protein corresponds to removal of N'-terminal peptides from pro-SP-B, liberating 13–15 kDa fragments. SP-B is synthesized as a proprotein by alveolar type II epithelial cells and non-ciliated bronchiolar (Clara) cells; however, complete processing of the precursor to the biologically active, mature peptide occurs only in type II cells. Clara cells merely generate the 25 and 42 kDa precursors [28]. Thus, this intermediate represents a normal pro-SP-B processing intermediate of SP-B biosynthesis and could result from either constitutive secretion of this form by type II cells or from the physiologic release of 25 kD pro-SP-B into the airways by Clara cells. The 25–26 kD bands of pro-SP-B have previously been described in amniotic fluid from a 24-week-old human fetus, in lung tissue from an infant with severe bronchopulmonary dysplasia at the time of lung transplantation, as well as in normal adult lung tissue and lavages and plasma [21,29]. Here we show that these peptides are released into the bronchoalveolar space in normal patients. Since lamellar bodies do not contain pro-SP-B, this likely occurs via constitutive, non-regulated secretory pathways.
In children with pulmonary alveolar proteinosis we discovered increased amounts of a 19–21 kD intermediate which reacted against C-terminal pro-SP-B antisera and with the NFLANK SP-B antibody. This finding of a complex pro-SP-B intermediate containing both the C-terminal propeptide and a vestigial N-terminal propeptide (approximate residues 186–201) extends the work of Brasch and colleagues who also noted the presence of pro-SP-B forms containing C-terminal propeptide epitopes [30]. Consistent with our data, this group also found that, in contrast to patients with congenital respiratory distress due to SP-B deficiency, the appearance of pro-SP-C forms in these PAP patients was a rare occurrence. Thus, despite similar chest x-rays and histopathological findings, the BAL profile for SP-B, SP-C and their proforms appears useful in distinguishing PAP from SP-B deficiency of any etiology.
Children with chronic respiratory distress of unknown cause (cRD) exhibited the 40–42 kD proprotein with increased frequency. The N'-terminal peptides liberated from pro-SP-B pre-protein during intracellular processing, i.e. 13–15 kDa peptides or smaller fragments and reacting with NFPROX, were found exclusively in both cRD and PAP (Fig. 1, Tab. 4, Tab. 5). As such they may give diagnostic hints for the involvement of processing defects in, especially in pediatric PAP.
Other peptides reacted with the antibodies directed to the flanking aminoacids next to the SP-B core (NFLANK and CFLANK). The presence of these relatively rarely observed bands at 11 to 15 kDa was not related to specific clinical features of the subjects, i.e. more pronounced lung injury, high protein to phospholipids ratio or high abundance of SP-B. Both, a 9 kDa intermediate, reactive to NFLANK [21] and a 9 kDa band reacting with antibodies directed to the C'-terminal flanking of pro-SP-B, have previously been observed in human isolated type II cells and fetal lung. Such bands were indeed detected in the lavages we investigated, although very rarely.
Pro-SP-C peptides were never detected in the control groups. This is in agreement with an earlier observation on a limited number of samples [31]. However, we found pro-SP-C forms that were clearly, but not exclusively, associated with SP-B deficiency or SFTPC mutation. On the other hand, not all infants with SFTPB (496delC) or SFTPC (R167Q) mutations had pro-SP-C in their lavages. Thus the presence of pro-SP-C in lavages may give strong, but surely not definitive, diagnostic evidence for SP-B and SP-C processing defects.
The aberrant pro-SP-C species observed in patients with SP-B deficiency carrying the 121ins2 mutation consists of a N-terminal extension of SP-C by the N-flanking 12 aminoacids of pro-SP-C [13]. The pro-SP-C forms observed in patients not bearing a SFTPB mutation clearly differed in molecular weights from those detected in SP-B deficiency, suggesting that several processing defects may result in aberrant pro-SP-C in the alveolar space.
Conclusion
Here we defined the presence and characteristics of SP-B, SP-C and their processing forms in bronchoalveolar lavage fluids from children with severe chronic respiratory distress and in comparison groups of normal children and children with chronic obstructive bronchitis. Pro-SP-B of 25–26 kD was commonly detected in all groups, suggesting that this form currently does not appear to be of great diagnostic value for processing defects. In contrast, pro-SP-B of 19–21 kD was increased in children with alveolar proteinosis while the cleaved flanking propeptides liberated during intracellular processing of pro-SP-B were exclusively found in these children and in chronic respirator distress of unknown cause. Furthermore, although identified at low frequency, pro-SP-C forms when present in the BAL suggest the presence of one of the parenchymal diseases studied in this report. Though often associated with mutations in SFTPB and SFTPC genes, this was not an exclusive finding limiting the usage of pro-SP-C as a surrogate for SFTP/SFTPC diagnostic screening procedures. Taken together, our results demonstrate that significant perturbations in the metabolism of these hydrophobic surfactant proteins occur in a variety of chronic lung diseases.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
MG designed the study, categorized and organized the subjects, wrote initial drafts of the manuscript, SS performed the blots, MT and MB determined the genotype of the patients, MG, SS, MS, AB, MT and MB collected the case histories, reviewed the subjects data and clinical courses, SG and MFB participated in the design for the methods to blot for the surfactant proteins, helped to organize the data and the results, and to prepare the manuscript. All authors read and approved the final manuscript.
Acknowledgements
The authors are grateful to Andrea Schams and Yvonne Wüst from Ludwig-Maximilians Universität, Munich, for expert technical assistance. We thank Dr. Wolfram Steinhilber, ALTANA Pharma AG, Konstanz, Germany for donating antibodies to the surfactant proteins B and C. Supported by: DFG Gr 970/7-1 (MG), HL 076064 (MFB), and P50-HL56401 (MFB).
==== Refs
Griese M Pulmonary surfactant in health and human lung diseases: state of the art Eur Respir J 1999 13 1455 1476 10445627
Weaver TE Conkright JJ Functions of surfactant proteins B and C Annu Rev Physiol 2001 63 555 578 11181967 10.1146/annurev.physiol.63.1.555
Weaver TE Synthesis, processing and secretion of surfactant proteins B and C Biochimica et Biophysica Acta-Molecular Basis of Disease 1998 1408 173 179 10.1016/S0925-4439(98)00066-0
Brasch F Ochs M Kahne T Guttentag S Schauer-Vukasinovic V Derrick M Involvement of napsin A in the C- and N-terminal processing of surfactant protein B in type-II pneumocytes of the human lung J Biol Chem 2003 278 49006 49014 13129928 10.1074/jbc.M306844200
Ueno T Linder S Na CL Rice WR Johansson J Weaver TE Processing of Pulmonary Surfactant Protein B by Napsin and Cathepsin H J Biol Chem 2004 279 16178 16184 14766755 10.1074/jbc.M312029200
Brasch F Johnen G Winn-Brasch A Guttentag SH Schmiedl A Kapp N Surfactant Protein B in Type II Pneumocytes and Intra-Alveolar Surfactant Forms of Human Lungs Am J Respir Cell Mol Biol 2004 30 449 458 12972403 10.1165/rcmb.2003-0262OC
Guttentag S Robinson L Zhang P Brasch F Buhling F Beers M Cysteine protease activity is required for surfactant protein B processing and lamellar body genesis Am J Respir Cell Mol Biol 2003 28 69 79 12495934 10.1165/rcmb.2002-0111OC
Wood S Yaremko ML Schertzer M Kelemen PR Minna J Westbrook CA Mapping of the Pulmonary Surfactant SP5 (SFTP2) Locus to 8p21 and Characterization of a Microsatellite Repeat Marker That Shows Frequent Loss of Heterozygosity in Human Carcinomas Genomics 1994 24 597 600 7713515 10.1006/geno.1994.1673
Glasser SW Korfhagen TR Perme CM Pilot-Matias TJ Kister S Whitsett JA Two SP-C genes encoding human pulmonary surfactant proteolipid J Biol Chem 1988 263 10326 10331 2839484
Qanbar R Cheng S Possmayer F Schurch S Role of the palmitoylation of surfactant-associated protein C in surfactant film formation and stability Am J Physiol (Lung Cell Mol Physiol) 1996 271 L572 L580
Stults JT Griffin PR Lesikar DD Naidu A Moffat B Benson BJ Lung surfactant protein SP-C from human, bovine, and canine sources contains palmityl cysteine thioester linkages Am J Physiol (Lung Cell Mol Physiol) 1991 261 L118 L125
Nogee LM de Mello DE Dehner LP Colten HR Brief-report: deficiency of pulmonary surfactant protein B in congenital alveolar proteinosis N Engl J Med 1993 328 406 410 8421459 10.1056/NEJM199302113280606
Li J Ikegami M Na CL Hamvas A Espinassous Q Chaby R N-terminally extended surfactant protein (SP) C isolated from SP-B-deficient children has reduced surface activity and inhibited lipopolysaccharide binding Biochemistry 2004 43 3891 3898 15049696 10.1021/bi036218q
Nogee LM Wert SE Proffit SA Hull WM Whitsett JA Allelic heterogeneity in hereditary surfactant protein B (SP-B) deficiency Am J Respir Crit Care Med 2000 161 973 981 10712351
Mahut B Delcourt C Scheinmann P de Blic J Mani T Fournet J Pulmonary alveolar proteinosis: Experience with eight pediatric cases and a review Pediatrics 1996 97 117 122 8545205
Seymour JF Presneill JJ Pulmonary alveolar proteinosis: progress in the first 44 years Am J Resp Crit Care Med 2002 166 215 235 12119235 10.1164/rccm.2109105
Tredano M Griese M Brasch F Schumacher S de Blic J Marque S Mutation of SFTPC in infantile pulmonary alveolar proteinosis with or without fibrosing lung disease Am J Med Genet 2004 126A 18 26 10.1002/ajmg.a.20670
Tredano M Griese M de Blic J Lorant T Houdayer C Schumacher S Analysis of 40 sporadic or familial neonatal and pediatric cases with severe unexplained respiratory distress: Relationship to SFTPB Am J Med Genet 2003 119A 324 339 10.1002/ajmg.a.20058
Griese M Felber J Reiter K Strong P Reid K Belohradsky BH Airway inflammation in children with tracheostomy Pediatr Pulmonol 2004 37 356 361 15022133 10.1002/ppul.10432
Schmidt R Steinhilber W Ruppert C Grimminger F Seeger W Günther A An ELISA technique for quantification of surfactant apoprotein (SP)-C in bronchoalveolar lavage fluid Am J Respir Crit Care Med 2002 165 470 474 11850338
Guttentag S Beers MF Bieler BM Ballard PL Surfactant protein B processing in human fetal lung Am Phys Soc 1998 L559 L566
Beers MF Kim CY Dodia C Fisher AB Localization, synthesis, and processing of surfactant protein SP-C in rat lung analyzed by epitope-specific antipeptide antibodies J Biol Chem 1994 269 20318 20328 7519606
Johnson AL Braidotti P Pietra GG Russo SJ Kabore A Wang WJ Posttranslational processing of surfactant protein C proprotein. Targeting motifs in the NH2-terminal flanking domain are cleaved in late compartments Am J Respir Cell Mol Biol 2001 24 253 263 11245624
Bradford MM A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding Anal Biochem 1976 72 248 254 942051
Heukeshoven J Dernick R Improved silver staining procedure for fast staining in PhastSystem Development Unit. I. Staining of sodium dodecyl sulfate gels Electrophoresis 1988 9 28 32 2466645 10.1002/elps.1150090106
Tredano M van Elburg RM Kaspers AG Zimmermann LJ Houdayer C Aymard P Compound SFTPB 1549C-GAA (121ins2) and 457delC heterozygosity in severe congenital lung disease and surfactant protein B (SP-B) deficiency Hum Mutat 1999 14 502 509 10571948 10.1002/(SICI)1098-1004(199912)14:6<502::AID-HUMU9>3.0.CO;2-C
Shulenin S Nogee LM Annilo T Wert SE Whitsett JA Dean M ABCA3 Gene Mutations in Newborns with Fatal Surfactant Deficiency N Engl J Med 2004 350 1296 1303 15044640 10.1056/NEJMoa032178
Lin S Na CL Akinbi HT Apsley KS Whitsett JA Weaver TE Surfactant protein B (SP-B) -/- mice are rescued by restoration of SP-B expression in alveolar type II cells but not Clara cells J Biol Chem 1999 274 19168 19174 10383422 10.1074/jbc.274.27.19168
Doyle IR Bersten AD Nicholas TE Surfactant proteins-A and -B are elevated in plasma of patients with acute respiratory failure Am J Respir Crit Care Med 1997 156 1217 1229 9351625
Brasch F Birzele J Ochs M Guttentag S Schoch OD Boehler A Surfactant proteins in pulmonary alveolar proteinosis in adults Eur Respir J 2004 24 426 435 15358702 10.1183/09031936.04.00076403
Vorbroker DK Profitt SA Nogee LM Whitsett JA Aberrant processing of surfactant protein C in hereditary SP-B deficiency Am J Physiol 1995 268 L647 L656 7537464
|
16042774
|
PMC1224872
|
CC BY
|
2021-01-04 16:23:24
|
no
|
Respir Res. 2005 Jul 22; 6(1):80
|
utf-8
|
Respir Res
| 2,005 |
10.1186/1465-9921-6-80
|
oa_comm
|
==== Front
Respir ResRespiratory Research1465-99211465-993XBioMed Central London 1465-9921-6-981612022710.1186/1465-9921-6-98ResearchMortality in GOLD stages of COPD and its dependence on symptoms of chronic bronchitis Ekberg-Aronsson Marie [email protected] Kerstin [email protected] Jan-Åke [email protected] Peter M [email protected]öfdahl Claes-Göran [email protected] Department of Respiratory Medicine and Allergology, University of Lund, S-221 85 Lund, Sweden2 Department of Medicine, University of Lund, University hospital, S-205 02 Malmö, Sweden2005 25 8 2005 6 1 98 98 25 4 2005 25 8 2005 Copyright © 2005 Ekberg-Aronsson et al; licensee BioMed Central Ltd.2005Ekberg-Aronsson et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
The GOLD classification of COPD severity introduces a stage 0 (at risk) comprising individuals with productive cough and normal lung function. The aims of this study were to investigate total mortality risks in GOLD stages 0–4 with special focus on stage 0, and furthermore to assess the influence of symptoms of chronic bronchitis on mortality risks in GOLD stages 1–4.
Method
Between 1974 and 1992, a total of 22 044 middle-aged individuals participated in a health screening, which included a spirometry as well as recording of respiratory symptoms and smoking habits. Individuals with comorbidity at baseline (diabetes, stroke, cancer, angina pectoris, or heart infarction) were excluded from the analyses. Hazard ratios (HR 95% CI) of total mortality were analyzed in GOLD stages 0–4 with individuals with normal lung function and without symptoms of chronic bronchitis as a reference group. HR:s in smoking individuals with symptoms of chronic bronchitis within the stages 1–4 were calculated with individuals with the same GOLD stage but without symptoms of chronic bronchitis as reference.
Results
The number of deaths was 3674 for men and 832 for women based on 352 324 and 150 050 person-years respectively. The proportion of smokers among men was 50% and among women 40%. Self reported comorbidity was present in 4.6% of the men and 6.6% of the women. Among smoking men, Stage 0 was associated with an increased mortality risk, HR; 1.65 (1.32–2.08), of similar magnitude as in stage 2, HR; 1.41 (1.31–1.70). The hazard ratio in stage 0 was significantly higher than in stage 1 HR; 1.13 (0.98–1.29). Among male smokers with stage 1; HR: 2.04 (1.34–3.11), and among female smokers with stage 2 disease; HR: 3.16 (1.38–7.23), increased HR:s were found in individuals with symptoms of chronic bronchitis as compared to those without symptoms of chronic bronchitis.
Conclusion
Symptoms fulfilling the definition of chronic bronchitis were associated with an increased mortality risk among male smokers with normal pulmonary function (stage 0) and also with an increased risk of death among smoking individuals with mild to moderate COPD (stage 1 and 2).
chronic bronchitisCOPDepidemiologyFEV1GOLDmortalityrespiratory symptomssmoking
==== Body
Background
Chronic obstructive pulmonary disease (COPD) is a major cause of increased morbidity and mortality [1]. The Global Initiative for Chronic Obstructive Lung Diseases (GOLD) guidelines were published in 2001 [2], and revised in 2004 [3] with the aim of increasing awareness of COPD and of decreasing morbidity and mortality from the disease. The GOLD staging system for COPD introduces a stage 0 (at risk), defined as the presence of chronic respiratory symptoms such as productive cough in individuals with preserved normal pulmonary function. The importance of the "at risk" stage, especially the effect of productive cough on morbidity and mortality, is an issue under debate [4]. To assess prognosis in patients with established COPD, the BODE-index for prediction of mortality (based on degree of airflow limitation, dyspnea, 6 min walking distance and body mass index) has been developed [5]. The BODE index is probably less helpful in the early stages of COPD when individuals may have no symptoms or symptoms of chronic bronchitis only ("smoker's cough"), regarded by most individuals (and physicians) as innocent symptoms. Contrary to this view, some previous studies have shown an increased mortality risk associated with symptoms of chronic bronchitis [6,7].
We hypothesized that chronic productive cough (symptoms of chronic bronchitis) could predict an increased mortality risk. The aims of this study were firstly to investigate all-cause mortality risks in relation to GOLD stages 0–4 with special focus on stage 0. Secondly, the effect of symptoms of chronic bronchitis on mortality risks in GOLD stages 1–4 was also assessed.
Study Population and Methods
Study population
With a population of 250000 inhabitants, Malmö is Sweden's third largest city. The Malmö Preventive Program (MPP), a preventive, case-finding programme for cardiovascular risk factors and alcohol abuse, was created in 1974 at the Department of Preventive Medicine of Malmö University Hospital. The aim of this programme was to screen large strata of the adult population, mostly middle-aged men and women born in prespecified years, in order to find high-risk individuals for preventive intervention [8,9]. Between 1974 and 1992, a total of 22,444 men (mean age 44 years range 27 – 61) and 10 902 women (mean age 50 years, range 28 – 58) attended the programme, with an overall attendance rate of 70% (range 64–78) [8]. The present study was based on prospectively collected data including detailed data on smoking habits obtained from 22,044 individuals with complete data on smoking habits and lung function among the individuals who took part in the MPP. The study population is described in fig 1.
Figure 1 Flow chart of the study population.
Screening procedure and intervention
All individuals who participated in the MPP took part in a comprehensive health screening, which included a physical examination, spirometry and blood tests. Additionally, lifestyle-related risk factors were assessed by means of a self-administered questionnaire including questions on smoking habits. Various interventions (lifestyle modification, drug therapy) were offered to nearly 25% of the screened individuals for shorter or longer periods [8,10]. Intervention against smoking was however only instituted if another cardiovascular risk factor was present and consisted of advice given by a nurse to stop smoking, sometimes supported by measurements of carboxyhemoglobin (COHb) and feedback information to the individual with a history of smoking [11].
Recording of tobacco smoking
All individuals completed a self-administered questionnaire regarding their tobacco consumption and inhalation habits. Tobacco consumption was calculated by equating one cigarette to 1 gram, one cheroot to 3 grams, and one cigar to 5 grams of tobacco. Individuals were categorized according to their self-reported smoking habits. Individuals who gave negative answers to all smoking-related questions regarding current and previous smoking habits were classified as never smokers. Individuals who stated that they had stopped smoking and who gave negative answers to all other smoking-related questions were classified as former smokers. Individuals who stated that they were currently smoking were classified as smokers. Those who gave contradictory answers were excluded. In order to adjust for differences in the amount of tobacco smoked a categorical variable on tobacco consumption was used separating individuals who smoked 20 grams or more tobacco per day from those who smoked less than that amount.
Assessment of pulmonary function
Pulmonary function was assessed by a screening spirometry. A Spirotron® apparatus (Drägerwerk AG, Lűbeck, Germany) was used with the individual in an upright standing position without a nose clip. Specially trained nurses performed the tests. Forced expiratory volume in one second (FEV1) and forced vital capacity (FVC) were recorded. One acceptable manoeuvre with respect to the individual's cooperation and performance was required [12]. FEV1 was analysed as percent of predicted (pred) values. Predicted values of FEV1 were obtained from an internally derived linear regression based on height and age in a subgroup of 3467 male and 2961 female never-smokers.
Men: Pred. FEV1 (L): 4.422 × height (m) - 0.0381 × age (year) - 2.483, SD 0.63
Women: Pred. FEV1 (L): 3.615 × height (m) - 0.0217 × age - 2.134, SD 0.45
Assessment of respiratory symptoms
Respiratory symptoms were assessed by a self-administered questionnaire. We used the classical definition of chronic bronchitis. The questions regarding chronic productive cough varied somewhat during the screening period. In the period from screening start on the 1st September 1974 until the 1st March 1978 the following question was used:
1."During any period of your life, have you had daily cough lasting more than three months in more than two years?" ("Ever symptoms of symptoms of chronic bronchitis")
In the period from 2nd of March 1978 until the end of the screening the following two questions were used :
2. "During the last two years have you had daily cough lasting for more than 3 months", and
3. "During the last two years, have you had phlegm coming up from your chest daily for more than three months? ("Recent symptoms of chronic bronchitis")
Individuals who answered affirmatively to the first question ("ever symptoms of chronic bronchitis" or both the second and third question ("recent symptoms of chronic bronchitis") were regarded as having symptoms of chronic bronchitis corresponding to the definition of GOLD stage 0.
Classification of pulmonary function
GOLD Stages 0–4 were defined as follows:
stage 0 FEV1/FVC ≥ 0.70 and FEV1 ≥ 80% pred and symptoms of chronic bronchitis,
stage 1 (mild): FEV1/FVC < 0.70 and FEV1 ≥ 80% pred,
stage 2 (moderate): FEV1/FVC < 0.70 and FEV1 < 80% pred and FEV1 ≥ 50% pred,
stage 3 (severe): FEV1/FVC < 0.70 and FEV1 < 50% pred and FEV1 ≥ 30% pred,
stage 4 (very severe) : FEV1/FVC < 0.70 and FEV1 < 30% pred,
Individuals without symptoms of chronic bronchitis and with FEV/FVC ≥ 70 and FEV1 ≥ 80% predicted were classified as normal and used as a reference group throughout the study except for the analyses of the effect of symptoms of chronic bronchitis within stage 1–3 where individuals with the same GOLD stage, but without symptoms of chronic bronchitis were used as reference.
Assessment of comorbidity
Baseline comorbidity was assessed by a self-administered questionnaire. Individuals who answered affirmatively to any of the following four questions were regarded as having significant comorbidity and were excluded from further analyses.
1. "Have you previously had a stroke?"
2. "Have you a physician's diagnosis of angina pectoris?"
3. "Do you have diabetes?"
4. "Have you been hospitalised due to a heart infarction?"
5. "Have you previously or currently received a diagnosis of cancer?"
Register follow-up analyses
All individuals were followed up in national registers for total mortality until 31st December 2003. Cause-specific mortality was assessed in male smokers of GOLD stage 0. Notably, data on cause-specific mortality was only available until 31st December 2002. The Swedish Board of Health and Welfare provided data from national registers on death certificates and cancer diagnoses. The overall autopsy rate was 44% during the study period. The cases were coded according to the International Classification of Diseases ICD 8–10. Mortality for all causes of death, cardiovascular disease (ICD 8–9: 401–414, 424, 426–429, 431–444; ICD 10: I10–25, I34–37, I44–49, I50–52, I61–71), all cancer except for lung cancer (ICD 8–9: 140–208; ICD 10: C00–C97), lung cancer (ICD 8–9: 162–165; ICD10: C34, 38, 39, 45) and respiratory diseases (ICD 8–9: 460–466, 480–487, 490–496, 500–503, 510–519; ICD10: J00–99) were recorded. The vital status at December 31st 2003 was unknown in 412 individuals who had left the country by that time. The accumulated person-years until they left the country were used in the analyses.
Data analysis
The computer-based analysis programme, Statistical Programme for Social Sciences (SPSS) version 10.0, was used for all calculations. The Cox Proportional Hazards Regression Model [13] was used to calculate relative total mortality rates in GOLD stages 0–4 with adjustments for age only in never-smokers and former smokers. Cause specific mortality was analysed only in smoking men of stage 0. In smokers, adjustments were also made for tobacco consumption (<20 grams/≥20 grams per day) and inhalation habits (yes/no). The reference group in all analyses consisted of individuals with normal pulmonary function and without symptoms of chronic bronchitis. In order to address possible differences in mortality risk between "ever symptoms of chronic bronchitis and "recent symptoms of chronic bronchitis" we calculated the mortality risk in GOLD stage 0 among smoking men using both "ever symptoms of chronic bronchitis" and "recent symptoms of chronic bronchitis" for the definition of stage 0, with the same adjustments and reference group as described above. Among smokers, the influence of symptoms of chronic bronchitis was assessed comparing individuals with and without symptoms of chronic bronchitis within the GOLD stages 1–4 after adjustment for age, tobacco consumption and inhalation habits. All variables were included in the analyses as categorical (yes/no) except for age. All individuals with self-reported comorbidity at baseline were excluded from the analyses. All analyses were performed separately for men and women. A p-value of less than 0.05 was considered statistically significant.
Results
Characteristics of the study population
The total number of deaths was 3674 among men and 832 in women based on 352 324 and 150 050 person-years, respectively. Among men of GOLD stage 0, 217 individuals had "ever symptoms of chronic bronchitis" and 101 individuals had "recent symptoms of chronic bronchitis". The corresponding figures among women were 33 and 142.
The proportion of current smokers was high at the time of screening and slightly higher among men (50%) than among women (40%). Former smokers were also more common among men (27%) than among women (16%). Among men, 24% of the individuals and among women, 40% stated that they had never smoked (never smokers). Self-reported comorbidity was present in 4.6% of the men and 6.6% of the women. The prevalence of COPD in the study population was for men in GOLD stage 0–4: 2.2%, 9.9%, 6.9%, 1.0%, and 0.3%. The corresponding figures among women were: 2.4%, 4.5%, 4.3%, 0.6%, and 0.2% respectively, (Table 1).
Table 1 Characteristics of the study population*
Men Women
Number (n) 14 630 7414
Smoker 7221 (49.4) 2961 (39.9)
Exsmoker 3942 (26.9) 1169 (15.8)
Never-smoker 3467 (23.7) 2961 (39.9)
Age 46.4 (5.7) 47.5 (7.8)
Follow-up time 22.2 (5.7) 20.2 (4.7)
Person-years 352 324 150 050
Deaths 3674 (25.1) 832 (11.2)
Reference** 10215 (69.8) 5706 (77)
Not classif*** 1453 (9.9) 821 (11.0)
Gold 0 318 (2.2) 175 (2.4)
Gold 1 1443 (9.9) 336 (4.5)
Gold 2 1010 (6.9) 316 (4.3)
Gold 3 149 (1.0) 48 (0.6)
Gold 4 42 (0.3) 12 (0.2)
Comorbidity**** 676 (4.6) 489 (6.6)
* Data are given as numbers (%), but age and follow-up time are given as years (mean, standard deviation). ** Normal pulmonary function, without symptoms of chronic bronchitis
*** Not-classif: individuals not classified as normal or categorized according to the GOLD stages.
**** Self-reported previous/current stroke, angina pectoris diagnosed by a physician, diabetes, hospitalization due to a heart infarction or a diagnosis of cancer at baseline.
Mortality risks in smokers
Among smoking men the hazard ratio (HR) increased stepwise from stage 1 to stage 4 (p for trend < 0.0001). Stage 1 showed a slightly increased HR (1.13) of borderline significance and the HR increased further in stage 2 (1.41). In the later stages a pronounced increase in mortality risk was seen. In stages 3 and 4 the risks increased two- to tenfold, respectively, suggesting a possible threshold with respect to an increased mortality risk at a FEV1 equal to or lower than 50% predicted. Among smoking men, stage 0 was associated with significantly elevated mortality risk, HR: 1.65 (1.32–2.08), of similar magnitude as in stage 2, HR: 1.41 (1.31–1.70), suggesting an adverse influence of symptoms of chronic bronchitis on mortality risks in these symptomatic smokers. The relative risk in stage 0 was significantly higher than in stage 1; HR: 1.13 (0.98–1.29). No difference in mortality risk was seen among smoking men of GOLD stage 0 when we used "ever symptoms of chronic bronchitis" (217 individuals), HR: 1.62 (1.25–2.12 p-value: <0.0001) or "recent symptoms of chronic bronchitis" (101 individuals) HR: 1.74 (1.13–2.68 p-value: 0.012) to define stage 0. Among smoking women, as for men, the HR:s in stage 1–4 increased stepwise (p for trend <0.0001). There was an overlap in the confidence intervals in stages 1 and 2 with no significant difference between the stages. As for men, GOLD 3 and 4 conveyed an increase in mortality risk that was significantly higher than in the other GOLD stages, again suggesting a possible "threshold" at a FEV1 of 50% predicted. The HR was slightly increased in stage 0, with a similar magnitude as in smoking men but of borderline significance, probably due to low statistical power or differences in the influence of symptoms of chronic bronchitis between genders, (Table 2).
Table 2 Mortality risks in GOLD stages 0–4 stratified for smoking status and gender*
Smoker*** Former smoker*** Never smoker***
Men N* Deaths RR** CI P-value N* Deaths RR CI p-value N* Deaths RR CI p-value
Ref 4430 1205 2993 517 2792 404
Gold 0 215 92 1.65 1.32–2.08 <0.0001 54 17 1.75 1.05–2.43 0.033 49 6 1.03 0.46–2.31 0.95
Gold 1 792 269 1.13 0.98–1.29 0.087 405 68 0.80 0.61–1.05 0.10 246 51 1.31 0.97–1.76 0.080
Gold 2 710 309 1.41 1.31–1.70 <0.0001 173 44 1.16 0.81–1.66 0.43 127 32 1.73 1.18–2.54 0.005
Gold 3 101 63 2.42 1.84–3.18 <0.0001 33 18 2.42 1.44–4.08 0.001 15 7 3.93 1.86–8.30 <0.0001
Gold 4 27 21 3.57 2.23–5.71 <0.0001 8 4 2.59 0.83–8.07 0.10 7 2 1.04 0.15–7.39 0.97
Women
Ref 2174 256 1 993 87 1 2539 168 1
Gold 0 108 17 1.64 0.98–2.73 0.060 25 1 0.53 0.07–3.79 0.53 42 7 3.12 1.46–6.67 0.003
Gold 1 184 39 1.75 1.22–2.51 0.002 46 2 0.23 0.03–1.65 0.14 106 8 0.83 0.37–1.87 0.64
Gold 2 234 52 1.75 1.27–2.42 0.001 32 5 1.48 0.59–3.67 0.40 50 7 2.11 0.99–4.52 0.054
Gold 3 36 19 5.11 3.09–8.45 <0.0001 5 1 NA 7 2 3.91 0.96–15.8 0.056
Gold 4 9 6 10.26 4.53–23.25 <0.0001 0 0 NA 3 1 18.10 2.53–129.6 0.0004
* Individuals with self-reported baseline comorbidity (current/previous cancer, stroke, angina pectoris, heart infarction or diabetes) were excluded. GOLD stage 4 were not analysed due to low number of cases. The reference group consists of individuals without chronic bronchitis and with normal pulmonary function.
** HR denotes hazard ratio with CI 95% confidence interval.
*** In smokers adjustments were made for age, inhalation habits (yes/no) and tobacco consumption (>=20 g/<20 g tobacco). In former and never-smokers adjustments were made for age only.
Cause specific mortality among male smokers with GOLD stage 0
Among male smokers HR for cardiovascular mortality was (16 deaths): 0.88 (0.53–1.44), p-value 0.60, mortality in all cancer mortality except lung cancer: (22 deaths): 2.03 (1.31–3.15) p-value 0.001, mortality in lung cancer: (12 deaths) 2.22 (1.23–4.02), p-value: 0.008, other causes: (24 deaths) 2.05 (1.45–2.88) p-value: <0.0001. Respiratory mortality was not analysed due to a lower number of deaths, (only one death in respiratory diseases was recorded in stage 0).
Mortality risks among former smokers and never smokers
Among male former smokers the numbers of deaths were generally low except for stages 1 and 2, which showed no increased mortality risk. Stage 0 was associated with an increased risk, however the number of deaths was low and the results should therefore be interpreted with caution. This was also the case among male never smokers except for stage 2 (32 deaths) that showed a significantly increased HR of 1.73 (1.18–2.54), suggesting an effect of low FEV1 also among never-smokers. Among former and never-smoking women, the numbers of deaths were very low in all GOLD stages and the results should therefore be interpreted with caution, (Table 2).
The influence of symptoms of chronic bronchitis on mortality risks in stages 1 and 2
We compared subjects with and without self-reported symptoms of chronic bronchitis at baseline and found that the HR was significantly increased among male smokers with symptoms of chronic bronchitis and mild COPD (stage 1, with airflow limitation only) as compared to those without symptoms of chronic bronchitis. Among smoking women this was also the case among individuals with moderate COPD (stage 2) suggesting an effect of symptoms of chronic bronchitis on mortality risks in the early stages of COPD, (Table 3).
Table 3 Mortality risks among smokers within GOLD stages 1–4 in individuals with symptoms of chronic bronchitis, with individuals from the same GOLD stage without symptoms of chronic bronchitis as reference*
Gold stage Chr Br (%) Deaths HR** 95%CI p-value
Men
GOLD 1 44 (5.8) 25 2.04 1.34–3.11 0.001
GOLD 2 55 (8.3) 26 1.20 0.80–1.80 0.38
GOLD 3 17 (18.7) 9 0.71 0.34–1.51 0.38
GOLD 4 5 (20.8) 2 0.41 0.09–1.87 0.25
Women
GOLD 1 4 (2.4) 1 1.58 0.22–11.79 0.66
GOLD 2 17 (8.1) 7 3.16 1.38–7.23 0.006
GOLD 3 7 (21.2) 3 2.19 0.27–18.02 0.47
GOLD 4 3 (37.5) 2 0.58 0.08–4.11 0.59
* HR denotes hazard ratio, with CI: 95% confidence interval, Chr Br: symptoms of chronic bronchitis.
**Adjusted for age, tobacco consumption (>=20 g/<20 g tobacco) and inhalation habits.
Discussion
We investigated mortality risks in relation to the GOLD classification of obstructive pulmonary disease in a large population-based study of middle-aged Swedish men and women. The main findings were firstly that among smokers, symptoms of chronic bronchitis were associated with increased total mortality risk among individuals with normal pulmonary function (corresponding to GOLD stage 0) and among those with mild to moderate COPD (GOLD stage 1 and 2). Secondly, among male smokers of stage 0, cause-specific mortality showed increased HR for all cancer (except lung cancer), lung cancer and other diseases. Thirdly, the total mortality hazard ratio in stages 1–4 increased stepwise with a pronounced increase in risk at a level of FEV1 below 50% predicted, suggesting a threshold effect in mortality risk at this level. Individuals with comorbidity of cardiovascular disease, diabetes and cancer at baseline were excluded and we have adjusted for age and smoking habits as completely as possible including tobacco consumption and inhalation habits.
Comparison with other studies
Prevalence of stage 0
The prevalence of individuals with stage 0 in our study was 2.2% in men and 2.4% in women. In the Copenhagen City Heart study the prevalence was slightly higher; criteria for GOLD Stage 0 were met in 5.8% of the total adult population and in 7.2% of smokers [4]. In the National Health and Nutrition Examination Study (NHANES) study [14] the prevalence of respiratory symptoms among individuals with normal pulmonary function was substantially higher (approximately 13%). This is probably mostly dependent on different selection criteria due to variation in the definition of stage 0. Vestbo and coworkers limited the selection of individuals to those with symptoms of chronic bronchitis but included all individuals without airflow limitation irrespective of the level of FEV1. On the other hand the NHANES study had higher prevalence rates probably due to a broader definition of the respiratory symptoms [14]. In our study a specific definition of symptoms of chronic bronchitis was used and we only selected individuals with normal pulmonary function, both with respect to FEV1 and the FEV1/FVC ratio. Consequently the prevalence of stage 0 was lower in our study than in the NHANES [14] and the Copenhagen City Heart study [4].
Mortality risks associated with symptoms of chronic bronchitis
Among male smokers of our study, the HR of total mortality in GOLD stage 0 was strongly significant irrespective of whether we used "ever" or "recent" symptoms of chronic bronchitis" to define stage 0. The HR:s were of a similar magnitude as for stage 2. Among men with GOLD stage 1 in men and among women with stage 2 in women, increased HR:s were noticed in smokers who reported symptoms of chronic bronchitis, implying that respiratory symptoms might be important for an adverse prognosis. However, in GOLD stage 3 we were not able to show an additive risk associated with symptoms of chronic bronchitis possibly due to low statistical power.
Similar results have previously been demonstrated in some studies. In the Copenhagen City Heart study, chronic mucus hypersecretion was associated with an increased risk of all cause mortality statistically significant in men only (relative risk: 1.3 in men and 1.1 in women) [7]. There was no interaction between the effect of chronic mucus hypersecretion and FEV1 percent predicted with regard to total mortality which is in agreement with our results. However, also in the same study, regarding obstructive lung disease mortality, the effect of chronic mucus hyper-secretion varied with the level of ventilatory function, being weak in individuals with normal ventilatory function but more pronounced in individuals with reduced ventilatory function. This interaction was statistically significant [7]. Furthermore, another study using the Copenhagen City Heart study population showed that chronic mucus hyper-secretion was associated with later hospitalization due to COPD, relative risks were 2.4 (1.3 to 4.5) for men and 2.6 (1.2 to 5.3) for women [16].
In contrast, in the National Health and Nutrition Examination Study (NHANES) the presence of respiratory symptoms was not associated with an increased mortality risk in subjects with normal lung function [14]. The relative risk (1.2) was elevated but not statistically significant probably due to the fact that a much wider definition of respiratory symptoms was used which included as well individuals with wheeze and a diagnosis of asthma. In our study based on a similar number of deaths, but also on a more specific definition of respiratory symptoms (symptoms of chronic bronchitis) a HR of 1.6 was statistically significant among smoking men.
Occupational studies have shown results in agreement with our study. In a population of 1061 men working in the Paris area and followed for 22 years, chronic phlegm production was significantly associated with mortality: Relative risk 1.35; (p less than 0.01) [6]. In a group of gold miners, chronic mucus hyper-secretion was not related to COPD mortality, but related to mortality in ischemic heart disease and other causes after adjustment for tobacco and dust exposure [17].
On the contrary, in another occupational cohort of 2,718 British men examined between 1954 and 1961, and followed for 20 to 25 years, the risk of death from COPD was correlated with the initial degree of air-flow obstruction. Among men with similar initial air-flow obstruction, age-specific COPD death rates were not significantly related to initial mucus hypersecretion [18].
The explanation for the relation between chronic productive cough and future mortality risk is still unclear. Previously, in the classic study by Fletcher and coworkers, conducted in a cohort of working men in London, chronic phlegm production was unrelated to the development of airway obstruction and regarded as a less important condition [19]. Furthermore, the presence of stage 0 at baseline was not a meaningful predictor for the subsequent development of COPD in later stages after 5 and 15 years follow-up in a Danish prospective population based study [4].
Thus it seems that the symptoms of chronic bronchitis do not affect the FEV1 decline and development of COPD to the same extent that it affects all-cause and obstructive lung disease mortality and risk of COPD hospitalization. An explanation could be that symptoms of chronic bronchitis may be an independent pulmonary disorder associated with inflammatory and/or infectious exacerbations [20] and possibly also with the development of other diseases such as lung cancer [15] and thus not specifically associated with COPD development and progression. We were able to analyse cause specific mortality in GOLD stage 0 in male smokers. Our findings of significantly increased HR for lung cancer, all cancers, and other diseases in individuals of stage 0 support this view.
Mortality risks in stages 1–4
GOLD stage 1 in our study, as well as in the NHANES study [14], conveyed only a slight increase in total mortality risk of borderline significance, even lower than found for GOLD stage 0. Among former smokers there was a tendency towards a decreased risk. Thus among individuals who quit smoking when still in GOLD stage 1, the future mortality risk will be normal. In GOLD stages 2 to 4 our findings of increased mortality HR are in agreement with other previously conducted studies, as they actually reflect a successive reduction in FEV1 [21-23]. Mortality risks were also increased among never smokers, and these results are entirely in line with other studies [24]. The NHANES study showed increased mortality risks associated with moderate to severe COPD, among smokers and former smokers of approximately the same magnitude as we report [14].
Study limitations
A methodological limitation of these findings is the measurement of pulmonary function. The protocol of the lung function tests did not meet the standards of current recommendations [12]. The values of the internal regression equation based on height and age and the standard deviation of the equation is similar to that of the Working party of the European Community for Steel and Coal [25]. Hence, we believe that the methodological error is similar to the one achieved after optimal spirometric technique. However there remains the possibility of misclassification due to decreased specificity in the spirometries which could have resulted in an underestimation of the results rather than the reverse.
Another possible limitation of this study is dependent on the fact that smoking prevalence in Sweden has decreased since the 1970's [26]. Some smokers in this study are likely to have stopped smoking during the follow-up time. Smoking habits were assessed at one point in time and not repeated thereafter. Hence, we have not been able to calculate accumulated smoking exposure in terms of individual pack-years. However, even the use of pack-years has its limitations, especially with respect to recall biases in subjects with a long history of smoking [27].
Furthermore, we have no follow-up data on symptoms of chronic bronchitis and thus we were not able to determine whether the symptoms of chronic bronchitis were stable or not. According to previous studies symptoms of chronic bronchitis may disappear in some individuals [4]. In those individuals, symptoms of chronic bronchitis may have a limited influence on mortality risks and could cause an underestimation of the effect of persistent symptoms of chronic bronchitis in this study. In addition, we have no spirometric follow-up data and hence we are not able to determine whether individuals with GOLD stage 0 at baseline have developed COPD in the later GOLD stages during the follow up time.
Another limitation of the study was the lower statistical power in the analyses of women. Our findings in women, especially in former smokers and never smokers should therefore be interpreted with caution.
Implications of our findings
We believe that symptoms of chronic bronchitis should be regarded as a marker for an increased all-cause mortality risk. Our data show that among middle-aged smokers, symptoms of chronic bronchitis might represent an independent risk factor for an early mortality equal to that of having moderate COPD.
It is however uncertain from previous studies whether symptoms of chronic bronchitis really are related to the development and progression of COPD [4]. An increased risk of COPD mortality associated with chronic mucus hyper secretion among individuals with established COPD has been shown previously [7]. In our study, symptoms of chronic bronchitis increased the risk of all cause mortality and mortality in cancer, lung cancer and other causes in smoking individuals with normal lung function (stage 0) and the risk of total mortality in individuals with GOLD stage 1 and 2. Our findings indicate that symptoms of chronic bronchitis might be an independent pulmonary disorder that may or may not be associated with COPD. It is therefore questionable whether GOLD stage 0 should be used as an "at risk" stage for COPD development in future COPD classifications.
Finally, GOLD stage 0 indicates a group of individuals with a very unfavourable prognosis similar to that of individuals with GOLD stage 2. Among smokers, symptoms of chronic bronchitis contributed to an adverse prognosis not only in individuals "at risk", but also in those with early stages of COPD with GOLD 1 and 2.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
Marie Ekberg-Aronsson: Taking part in the conception and design of the study. Performed the statistical analyses and interpretation of data, drafting and revising the article, giving final approval to the version to be published.
Kerstin Pehrsson: Taking part in the conception and design of the study. Interpretation of data and revising the article, giving final approval to the version to be published.
Jan-Åke Nilsson: Taking part in the conception and design of the study. Statistical analysis and interpretation of data, giving final approval to the version to be published.
Peter Nilsson: Taking part in the conception and design of the study, acquisition and interpretation of data and revising the article, giving final approval to the version to be published.
Claes-Göran Löfdahl: Taking part in the conception and design of the study, interpretation of data and revising the article, giving final approval to the version to be published.
Acknowledgements
This study was supported by grants from the City of Malmö, The Swedish Medical Research Council, and the Swedish Heart and Lung Foundation
==== Refs
Hurd S The impact of COPD on lung health worldwide: epidemiology and incidence Chest 2000 117 1S 4S 10673465 10.1378/chest.117.2_suppl.1S
Pauwels RA Buist AS Calverley PM Jenkins CR Hurd SS Global strategy for the diagnosis, management, and prevention of chronic obstructive pulmonary disease. NHLBI/WHO Global Initiative for Chronic Obstructive Lung Disease (GOLD) Workshop summary Am J Respir Crit Care Med 2001 163 1256 76 11316667
Global Initiative For Chronic Obstructive Lung Disease™ GOLD workshop report update 2004
Vestbo J Lange P Can GOLD Stage 0 provide information of prognostic value in chronic obstructive pulmonary disease? Am J Respir Crit Care Med 2002 166 329 32 12153965 10.1164/rccm.2112048
Celli BR Cote CG Marin JM Casanova C Montes de Oca M Mendez RA Pinto Plata V Cabral HJ The body-mass index, airflow obstruction, dyspnea, and exercise capacity index in chronic obstructive pulmonary disease N Engl J Med 2004 350 1005 12 14999112 10.1056/NEJMoa021322
Annesi I Kauffmann F Is respiratory mucus hypersecretion really an innocent disorder? A 22-year mortality survey of 1,061 working men Am Rev Respir Dis 1986 134 688 93 3767125
Lange P Nyboe J Appleyard M Jensen G Schnohr P Relation of ventilatory impairment and of chronic mucus hypersecretion to mortality from obstructive lung disease and from all causes Thorax 1990 45 579 85 2402719
Trell E Community-based preventive medical department for individual risk factor assessment and intervention in an urban population Prev Med 1983 12 397 402 6878198 10.1016/0091-7435(83)90248-7
Berglund G Eriksson KF Israelsson B Kjellstrom T Lindgarde F Mattiasson I Nilsson JA Stavenow L Cardiovascular risk groups and mortality in an urban swedish male population: the Malmo Preventive Project J Intern Med 1996 239 489 97 8656142 10.1046/j.1365-2796.1996.483819000.x
Berglund G Nilsson P Eriksson KF Nilsson JA Hedblad B Kristenson H Lindgarde F Long-term outcome of the Malmo preventive project: mortality and cardiovascular morbidity J Intern Med 2000 247 19 29 10672127 10.1046/j.1365-2796.2000.00568.x
Janzon L Lindell SE Trell E Larme P Smoking habits and carboxyhaemoglobin. A cross-sectional study of an urban population of middle-aged men J Epidemiol Community Health 1981 35 271 3 7200119
Engstrom G Hedblad B Nilsson P Wollmer P Berglund G Janzon L Lung function, insulin resistance and incidence of cardiovascular disease: a longitudinal cohort study J Intern Med 2003 253 574 81 12702035 10.1046/j.1365-2796.2003.01138.x
Cox DR Regression models and life-tables (with discussion) Journal of the Royal Statistical Society 1972 Serie B 187 220
Mannino DM Buist AS Petty TL Enright PL Redd SC Lung function and mortality in the United States: data from the First National Health and Nutrition Examination Survey follow up study Thorax 2003 58 388 93 12728157 10.1136/thorax.58.5.388
Lange P Nyboe J Appleyard M Jensen G Schnohr P Ventilatory function and chronic mucus hypersecretion as predictors of death from lung cancer Am Rev Respir Dis 1990 141 613 7 2310094
Vestbo J Prescott E Lange P Association of chronic mucus hypersecretion with FEV1 decline and chronic obstructive pulmonary disease morbidity. Copenhagen City Heart Study Group Am J Respir Crit Care Med 1996 153 1530 5 8630597
Wiles FJ Hnizdo E Relevance of airflow obstruction and mucus hypersecretion to mortality Respir Med 1991 85 27 35 2014356
Peto R Speizer FE Cochrane AL Moore F Fletcher CM Tinker CM Higgins IT Gray RG Richards SM Gilliland J The relevance in adults of air-flow obstruction, but not of mucus hypersecretion, to mortality from chronic lung disease. Results from 20 years of prospective observation Am Rev Respir Dis 1983 128 491 500 6614643
Fletcher C Peto R The natural history of chronic airflow obstruction Br Med J 1977 1 1645 8 871704
Prescott E Lange P Vestbo J Chronic mucus hypersecretion in COPD and death from pulmonary infection Eur Respir J 1995 8 1333 8 7489800 10.1183/09031936.95.08081333
Schunemann HJ Dorn J Grant BJ Winkelstein W JrM Trevisan Pulmonary function is a long-term predictor of mortality in the general population: 29-year follow-up of the Buffalo Health Study Chest 2000 118 656 64 10988186 10.1378/chest.118.3.656
Vollmer WM McCamant LE Johnson LR Buist AS Respiratory symptoms, lung function, and mortality in a screening center cohort Am J Epidemiol 1989 129 1157 69 2729254
Hole DJ Watt GC Davey-Smith G Hart CL Gillis CR Hawthorne VM Impaired lung function and mortality risk in men and women: findings from the Renfrew and Paisley prospective population study Bmj 1996 313 711 5 discussion 715–6 8819439
Strachan DP Ventilatory function, height, and mortality among lifelong non-smokers J Epidemiol Community Health 1992 46 66 70 1573363
Quanjer PH Tammeling GJ Cotes JE Fabbri LM Matthys H Pedersen OF Peslin R Roca J Sterk PJ Ulmer WT Symbols, abbreviations and units. Working Party Standardization of Lung Function Tests, European Community for Steel and Coal Eur Respir J Suppl 1993 16 85 100 8499056
Tillgren P Haglund BJ Lundberg M Romelsjo A The sociodemographic pattern of tobacco cessation in the 1980s: results from a panel study of living condition surveys in Sweden J Epidemiol Community Health 1996 50 625 30 9039380
Bernaards CM Twisk JW Snel J Van Mechelen W Kemper HC Is calculating pack-years retrospectively a valid method to estimate life-time tobacco smoking? A comparison between prospectively calculated pack-years and retrospectively calculated pack-years Addiction 2001 96 1653 61 11784461 10.1046/j.1360-0443.2001.9611165311.x
|
16120227
|
PMC1224873
|
CC BY
|
2021-01-04 16:23:25
|
no
|
Respir Res. 2005 Aug 25; 6(1):98
|
utf-8
|
Respir Res
| 2,005 |
10.1186/1465-9921-6-98
|
oa_comm
|
==== Front
Respir ResRespiratory Research1465-99211465-993XBioMed Central London 1465-9921-6-981612022710.1186/1465-9921-6-98ResearchMortality in GOLD stages of COPD and its dependence on symptoms of chronic bronchitis Ekberg-Aronsson Marie [email protected] Kerstin [email protected] Jan-Åke [email protected] Peter M [email protected]öfdahl Claes-Göran [email protected] Department of Respiratory Medicine and Allergology, University of Lund, S-221 85 Lund, Sweden2 Department of Medicine, University of Lund, University hospital, S-205 02 Malmö, Sweden2005 25 8 2005 6 1 98 98 25 4 2005 25 8 2005 Copyright © 2005 Ekberg-Aronsson et al; licensee BioMed Central Ltd.2005Ekberg-Aronsson et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
The GOLD classification of COPD severity introduces a stage 0 (at risk) comprising individuals with productive cough and normal lung function. The aims of this study were to investigate total mortality risks in GOLD stages 0–4 with special focus on stage 0, and furthermore to assess the influence of symptoms of chronic bronchitis on mortality risks in GOLD stages 1–4.
Method
Between 1974 and 1992, a total of 22 044 middle-aged individuals participated in a health screening, which included a spirometry as well as recording of respiratory symptoms and smoking habits. Individuals with comorbidity at baseline (diabetes, stroke, cancer, angina pectoris, or heart infarction) were excluded from the analyses. Hazard ratios (HR 95% CI) of total mortality were analyzed in GOLD stages 0–4 with individuals with normal lung function and without symptoms of chronic bronchitis as a reference group. HR:s in smoking individuals with symptoms of chronic bronchitis within the stages 1–4 were calculated with individuals with the same GOLD stage but without symptoms of chronic bronchitis as reference.
Results
The number of deaths was 3674 for men and 832 for women based on 352 324 and 150 050 person-years respectively. The proportion of smokers among men was 50% and among women 40%. Self reported comorbidity was present in 4.6% of the men and 6.6% of the women. Among smoking men, Stage 0 was associated with an increased mortality risk, HR; 1.65 (1.32–2.08), of similar magnitude as in stage 2, HR; 1.41 (1.31–1.70). The hazard ratio in stage 0 was significantly higher than in stage 1 HR; 1.13 (0.98–1.29). Among male smokers with stage 1; HR: 2.04 (1.34–3.11), and among female smokers with stage 2 disease; HR: 3.16 (1.38–7.23), increased HR:s were found in individuals with symptoms of chronic bronchitis as compared to those without symptoms of chronic bronchitis.
Conclusion
Symptoms fulfilling the definition of chronic bronchitis were associated with an increased mortality risk among male smokers with normal pulmonary function (stage 0) and also with an increased risk of death among smoking individuals with mild to moderate COPD (stage 1 and 2).
chronic bronchitisCOPDepidemiologyFEV1GOLDmortalityrespiratory symptomssmoking
==== Body
Background
Chronic obstructive pulmonary disease (COPD) is a major cause of increased morbidity and mortality [1]. The Global Initiative for Chronic Obstructive Lung Diseases (GOLD) guidelines were published in 2001 [2], and revised in 2004 [3] with the aim of increasing awareness of COPD and of decreasing morbidity and mortality from the disease. The GOLD staging system for COPD introduces a stage 0 (at risk), defined as the presence of chronic respiratory symptoms such as productive cough in individuals with preserved normal pulmonary function. The importance of the "at risk" stage, especially the effect of productive cough on morbidity and mortality, is an issue under debate [4]. To assess prognosis in patients with established COPD, the BODE-index for prediction of mortality (based on degree of airflow limitation, dyspnea, 6 min walking distance and body mass index) has been developed [5]. The BODE index is probably less helpful in the early stages of COPD when individuals may have no symptoms or symptoms of chronic bronchitis only ("smoker's cough"), regarded by most individuals (and physicians) as innocent symptoms. Contrary to this view, some previous studies have shown an increased mortality risk associated with symptoms of chronic bronchitis [6,7].
We hypothesized that chronic productive cough (symptoms of chronic bronchitis) could predict an increased mortality risk. The aims of this study were firstly to investigate all-cause mortality risks in relation to GOLD stages 0–4 with special focus on stage 0. Secondly, the effect of symptoms of chronic bronchitis on mortality risks in GOLD stages 1–4 was also assessed.
Study Population and Methods
Study population
With a population of 250000 inhabitants, Malmö is Sweden's third largest city. The Malmö Preventive Program (MPP), a preventive, case-finding programme for cardiovascular risk factors and alcohol abuse, was created in 1974 at the Department of Preventive Medicine of Malmö University Hospital. The aim of this programme was to screen large strata of the adult population, mostly middle-aged men and women born in prespecified years, in order to find high-risk individuals for preventive intervention [8,9]. Between 1974 and 1992, a total of 22,444 men (mean age 44 years range 27 – 61) and 10 902 women (mean age 50 years, range 28 – 58) attended the programme, with an overall attendance rate of 70% (range 64–78) [8]. The present study was based on prospectively collected data including detailed data on smoking habits obtained from 22,044 individuals with complete data on smoking habits and lung function among the individuals who took part in the MPP. The study population is described in fig 1.
Figure 1 Flow chart of the study population.
Screening procedure and intervention
All individuals who participated in the MPP took part in a comprehensive health screening, which included a physical examination, spirometry and blood tests. Additionally, lifestyle-related risk factors were assessed by means of a self-administered questionnaire including questions on smoking habits. Various interventions (lifestyle modification, drug therapy) were offered to nearly 25% of the screened individuals for shorter or longer periods [8,10]. Intervention against smoking was however only instituted if another cardiovascular risk factor was present and consisted of advice given by a nurse to stop smoking, sometimes supported by measurements of carboxyhemoglobin (COHb) and feedback information to the individual with a history of smoking [11].
Recording of tobacco smoking
All individuals completed a self-administered questionnaire regarding their tobacco consumption and inhalation habits. Tobacco consumption was calculated by equating one cigarette to 1 gram, one cheroot to 3 grams, and one cigar to 5 grams of tobacco. Individuals were categorized according to their self-reported smoking habits. Individuals who gave negative answers to all smoking-related questions regarding current and previous smoking habits were classified as never smokers. Individuals who stated that they had stopped smoking and who gave negative answers to all other smoking-related questions were classified as former smokers. Individuals who stated that they were currently smoking were classified as smokers. Those who gave contradictory answers were excluded. In order to adjust for differences in the amount of tobacco smoked a categorical variable on tobacco consumption was used separating individuals who smoked 20 grams or more tobacco per day from those who smoked less than that amount.
Assessment of pulmonary function
Pulmonary function was assessed by a screening spirometry. A Spirotron® apparatus (Drägerwerk AG, Lűbeck, Germany) was used with the individual in an upright standing position without a nose clip. Specially trained nurses performed the tests. Forced expiratory volume in one second (FEV1) and forced vital capacity (FVC) were recorded. One acceptable manoeuvre with respect to the individual's cooperation and performance was required [12]. FEV1 was analysed as percent of predicted (pred) values. Predicted values of FEV1 were obtained from an internally derived linear regression based on height and age in a subgroup of 3467 male and 2961 female never-smokers.
Men: Pred. FEV1 (L): 4.422 × height (m) - 0.0381 × age (year) - 2.483, SD 0.63
Women: Pred. FEV1 (L): 3.615 × height (m) - 0.0217 × age - 2.134, SD 0.45
Assessment of respiratory symptoms
Respiratory symptoms were assessed by a self-administered questionnaire. We used the classical definition of chronic bronchitis. The questions regarding chronic productive cough varied somewhat during the screening period. In the period from screening start on the 1st September 1974 until the 1st March 1978 the following question was used:
1."During any period of your life, have you had daily cough lasting more than three months in more than two years?" ("Ever symptoms of symptoms of chronic bronchitis")
In the period from 2nd of March 1978 until the end of the screening the following two questions were used :
2. "During the last two years have you had daily cough lasting for more than 3 months", and
3. "During the last two years, have you had phlegm coming up from your chest daily for more than three months? ("Recent symptoms of chronic bronchitis")
Individuals who answered affirmatively to the first question ("ever symptoms of chronic bronchitis" or both the second and third question ("recent symptoms of chronic bronchitis") were regarded as having symptoms of chronic bronchitis corresponding to the definition of GOLD stage 0.
Classification of pulmonary function
GOLD Stages 0–4 were defined as follows:
stage 0 FEV1/FVC ≥ 0.70 and FEV1 ≥ 80% pred and symptoms of chronic bronchitis,
stage 1 (mild): FEV1/FVC < 0.70 and FEV1 ≥ 80% pred,
stage 2 (moderate): FEV1/FVC < 0.70 and FEV1 < 80% pred and FEV1 ≥ 50% pred,
stage 3 (severe): FEV1/FVC < 0.70 and FEV1 < 50% pred and FEV1 ≥ 30% pred,
stage 4 (very severe) : FEV1/FVC < 0.70 and FEV1 < 30% pred,
Individuals without symptoms of chronic bronchitis and with FEV/FVC ≥ 70 and FEV1 ≥ 80% predicted were classified as normal and used as a reference group throughout the study except for the analyses of the effect of symptoms of chronic bronchitis within stage 1–3 where individuals with the same GOLD stage, but without symptoms of chronic bronchitis were used as reference.
Assessment of comorbidity
Baseline comorbidity was assessed by a self-administered questionnaire. Individuals who answered affirmatively to any of the following four questions were regarded as having significant comorbidity and were excluded from further analyses.
1. "Have you previously had a stroke?"
2. "Have you a physician's diagnosis of angina pectoris?"
3. "Do you have diabetes?"
4. "Have you been hospitalised due to a heart infarction?"
5. "Have you previously or currently received a diagnosis of cancer?"
Register follow-up analyses
All individuals were followed up in national registers for total mortality until 31st December 2003. Cause-specific mortality was assessed in male smokers of GOLD stage 0. Notably, data on cause-specific mortality was only available until 31st December 2002. The Swedish Board of Health and Welfare provided data from national registers on death certificates and cancer diagnoses. The overall autopsy rate was 44% during the study period. The cases were coded according to the International Classification of Diseases ICD 8–10. Mortality for all causes of death, cardiovascular disease (ICD 8–9: 401–414, 424, 426–429, 431–444; ICD 10: I10–25, I34–37, I44–49, I50–52, I61–71), all cancer except for lung cancer (ICD 8–9: 140–208; ICD 10: C00–C97), lung cancer (ICD 8–9: 162–165; ICD10: C34, 38, 39, 45) and respiratory diseases (ICD 8–9: 460–466, 480–487, 490–496, 500–503, 510–519; ICD10: J00–99) were recorded. The vital status at December 31st 2003 was unknown in 412 individuals who had left the country by that time. The accumulated person-years until they left the country were used in the analyses.
Data analysis
The computer-based analysis programme, Statistical Programme for Social Sciences (SPSS) version 10.0, was used for all calculations. The Cox Proportional Hazards Regression Model [13] was used to calculate relative total mortality rates in GOLD stages 0–4 with adjustments for age only in never-smokers and former smokers. Cause specific mortality was analysed only in smoking men of stage 0. In smokers, adjustments were also made for tobacco consumption (<20 grams/≥20 grams per day) and inhalation habits (yes/no). The reference group in all analyses consisted of individuals with normal pulmonary function and without symptoms of chronic bronchitis. In order to address possible differences in mortality risk between "ever symptoms of chronic bronchitis and "recent symptoms of chronic bronchitis" we calculated the mortality risk in GOLD stage 0 among smoking men using both "ever symptoms of chronic bronchitis" and "recent symptoms of chronic bronchitis" for the definition of stage 0, with the same adjustments and reference group as described above. Among smokers, the influence of symptoms of chronic bronchitis was assessed comparing individuals with and without symptoms of chronic bronchitis within the GOLD stages 1–4 after adjustment for age, tobacco consumption and inhalation habits. All variables were included in the analyses as categorical (yes/no) except for age. All individuals with self-reported comorbidity at baseline were excluded from the analyses. All analyses were performed separately for men and women. A p-value of less than 0.05 was considered statistically significant.
Results
Characteristics of the study population
The total number of deaths was 3674 among men and 832 in women based on 352 324 and 150 050 person-years, respectively. Among men of GOLD stage 0, 217 individuals had "ever symptoms of chronic bronchitis" and 101 individuals had "recent symptoms of chronic bronchitis". The corresponding figures among women were 33 and 142.
The proportion of current smokers was high at the time of screening and slightly higher among men (50%) than among women (40%). Former smokers were also more common among men (27%) than among women (16%). Among men, 24% of the individuals and among women, 40% stated that they had never smoked (never smokers). Self-reported comorbidity was present in 4.6% of the men and 6.6% of the women. The prevalence of COPD in the study population was for men in GOLD stage 0–4: 2.2%, 9.9%, 6.9%, 1.0%, and 0.3%. The corresponding figures among women were: 2.4%, 4.5%, 4.3%, 0.6%, and 0.2% respectively, (Table 1).
Table 1 Characteristics of the study population*
Men Women
Number (n) 14 630 7414
Smoker 7221 (49.4) 2961 (39.9)
Exsmoker 3942 (26.9) 1169 (15.8)
Never-smoker 3467 (23.7) 2961 (39.9)
Age 46.4 (5.7) 47.5 (7.8)
Follow-up time 22.2 (5.7) 20.2 (4.7)
Person-years 352 324 150 050
Deaths 3674 (25.1) 832 (11.2)
Reference** 10215 (69.8) 5706 (77)
Not classif*** 1453 (9.9) 821 (11.0)
Gold 0 318 (2.2) 175 (2.4)
Gold 1 1443 (9.9) 336 (4.5)
Gold 2 1010 (6.9) 316 (4.3)
Gold 3 149 (1.0) 48 (0.6)
Gold 4 42 (0.3) 12 (0.2)
Comorbidity**** 676 (4.6) 489 (6.6)
* Data are given as numbers (%), but age and follow-up time are given as years (mean, standard deviation). ** Normal pulmonary function, without symptoms of chronic bronchitis
*** Not-classif: individuals not classified as normal or categorized according to the GOLD stages.
**** Self-reported previous/current stroke, angina pectoris diagnosed by a physician, diabetes, hospitalization due to a heart infarction or a diagnosis of cancer at baseline.
Mortality risks in smokers
Among smoking men the hazard ratio (HR) increased stepwise from stage 1 to stage 4 (p for trend < 0.0001). Stage 1 showed a slightly increased HR (1.13) of borderline significance and the HR increased further in stage 2 (1.41). In the later stages a pronounced increase in mortality risk was seen. In stages 3 and 4 the risks increased two- to tenfold, respectively, suggesting a possible threshold with respect to an increased mortality risk at a FEV1 equal to or lower than 50% predicted. Among smoking men, stage 0 was associated with significantly elevated mortality risk, HR: 1.65 (1.32–2.08), of similar magnitude as in stage 2, HR: 1.41 (1.31–1.70), suggesting an adverse influence of symptoms of chronic bronchitis on mortality risks in these symptomatic smokers. The relative risk in stage 0 was significantly higher than in stage 1; HR: 1.13 (0.98–1.29). No difference in mortality risk was seen among smoking men of GOLD stage 0 when we used "ever symptoms of chronic bronchitis" (217 individuals), HR: 1.62 (1.25–2.12 p-value: <0.0001) or "recent symptoms of chronic bronchitis" (101 individuals) HR: 1.74 (1.13–2.68 p-value: 0.012) to define stage 0. Among smoking women, as for men, the HR:s in stage 1–4 increased stepwise (p for trend <0.0001). There was an overlap in the confidence intervals in stages 1 and 2 with no significant difference between the stages. As for men, GOLD 3 and 4 conveyed an increase in mortality risk that was significantly higher than in the other GOLD stages, again suggesting a possible "threshold" at a FEV1 of 50% predicted. The HR was slightly increased in stage 0, with a similar magnitude as in smoking men but of borderline significance, probably due to low statistical power or differences in the influence of symptoms of chronic bronchitis between genders, (Table 2).
Table 2 Mortality risks in GOLD stages 0–4 stratified for smoking status and gender*
Smoker*** Former smoker*** Never smoker***
Men N* Deaths RR** CI P-value N* Deaths RR CI p-value N* Deaths RR CI p-value
Ref 4430 1205 2993 517 2792 404
Gold 0 215 92 1.65 1.32–2.08 <0.0001 54 17 1.75 1.05–2.43 0.033 49 6 1.03 0.46–2.31 0.95
Gold 1 792 269 1.13 0.98–1.29 0.087 405 68 0.80 0.61–1.05 0.10 246 51 1.31 0.97–1.76 0.080
Gold 2 710 309 1.41 1.31–1.70 <0.0001 173 44 1.16 0.81–1.66 0.43 127 32 1.73 1.18–2.54 0.005
Gold 3 101 63 2.42 1.84–3.18 <0.0001 33 18 2.42 1.44–4.08 0.001 15 7 3.93 1.86–8.30 <0.0001
Gold 4 27 21 3.57 2.23–5.71 <0.0001 8 4 2.59 0.83–8.07 0.10 7 2 1.04 0.15–7.39 0.97
Women
Ref 2174 256 1 993 87 1 2539 168 1
Gold 0 108 17 1.64 0.98–2.73 0.060 25 1 0.53 0.07–3.79 0.53 42 7 3.12 1.46–6.67 0.003
Gold 1 184 39 1.75 1.22–2.51 0.002 46 2 0.23 0.03–1.65 0.14 106 8 0.83 0.37–1.87 0.64
Gold 2 234 52 1.75 1.27–2.42 0.001 32 5 1.48 0.59–3.67 0.40 50 7 2.11 0.99–4.52 0.054
Gold 3 36 19 5.11 3.09–8.45 <0.0001 5 1 NA 7 2 3.91 0.96–15.8 0.056
Gold 4 9 6 10.26 4.53–23.25 <0.0001 0 0 NA 3 1 18.10 2.53–129.6 0.0004
* Individuals with self-reported baseline comorbidity (current/previous cancer, stroke, angina pectoris, heart infarction or diabetes) were excluded. GOLD stage 4 were not analysed due to low number of cases. The reference group consists of individuals without chronic bronchitis and with normal pulmonary function.
** HR denotes hazard ratio with CI 95% confidence interval.
*** In smokers adjustments were made for age, inhalation habits (yes/no) and tobacco consumption (>=20 g/<20 g tobacco). In former and never-smokers adjustments were made for age only.
Cause specific mortality among male smokers with GOLD stage 0
Among male smokers HR for cardiovascular mortality was (16 deaths): 0.88 (0.53–1.44), p-value 0.60, mortality in all cancer mortality except lung cancer: (22 deaths): 2.03 (1.31–3.15) p-value 0.001, mortality in lung cancer: (12 deaths) 2.22 (1.23–4.02), p-value: 0.008, other causes: (24 deaths) 2.05 (1.45–2.88) p-value: <0.0001. Respiratory mortality was not analysed due to a lower number of deaths, (only one death in respiratory diseases was recorded in stage 0).
Mortality risks among former smokers and never smokers
Among male former smokers the numbers of deaths were generally low except for stages 1 and 2, which showed no increased mortality risk. Stage 0 was associated with an increased risk, however the number of deaths was low and the results should therefore be interpreted with caution. This was also the case among male never smokers except for stage 2 (32 deaths) that showed a significantly increased HR of 1.73 (1.18–2.54), suggesting an effect of low FEV1 also among never-smokers. Among former and never-smoking women, the numbers of deaths were very low in all GOLD stages and the results should therefore be interpreted with caution, (Table 2).
The influence of symptoms of chronic bronchitis on mortality risks in stages 1 and 2
We compared subjects with and without self-reported symptoms of chronic bronchitis at baseline and found that the HR was significantly increased among male smokers with symptoms of chronic bronchitis and mild COPD (stage 1, with airflow limitation only) as compared to those without symptoms of chronic bronchitis. Among smoking women this was also the case among individuals with moderate COPD (stage 2) suggesting an effect of symptoms of chronic bronchitis on mortality risks in the early stages of COPD, (Table 3).
Table 3 Mortality risks among smokers within GOLD stages 1–4 in individuals with symptoms of chronic bronchitis, with individuals from the same GOLD stage without symptoms of chronic bronchitis as reference*
Gold stage Chr Br (%) Deaths HR** 95%CI p-value
Men
GOLD 1 44 (5.8) 25 2.04 1.34–3.11 0.001
GOLD 2 55 (8.3) 26 1.20 0.80–1.80 0.38
GOLD 3 17 (18.7) 9 0.71 0.34–1.51 0.38
GOLD 4 5 (20.8) 2 0.41 0.09–1.87 0.25
Women
GOLD 1 4 (2.4) 1 1.58 0.22–11.79 0.66
GOLD 2 17 (8.1) 7 3.16 1.38–7.23 0.006
GOLD 3 7 (21.2) 3 2.19 0.27–18.02 0.47
GOLD 4 3 (37.5) 2 0.58 0.08–4.11 0.59
* HR denotes hazard ratio, with CI: 95% confidence interval, Chr Br: symptoms of chronic bronchitis.
**Adjusted for age, tobacco consumption (>=20 g/<20 g tobacco) and inhalation habits.
Discussion
We investigated mortality risks in relation to the GOLD classification of obstructive pulmonary disease in a large population-based study of middle-aged Swedish men and women. The main findings were firstly that among smokers, symptoms of chronic bronchitis were associated with increased total mortality risk among individuals with normal pulmonary function (corresponding to GOLD stage 0) and among those with mild to moderate COPD (GOLD stage 1 and 2). Secondly, among male smokers of stage 0, cause-specific mortality showed increased HR for all cancer (except lung cancer), lung cancer and other diseases. Thirdly, the total mortality hazard ratio in stages 1–4 increased stepwise with a pronounced increase in risk at a level of FEV1 below 50% predicted, suggesting a threshold effect in mortality risk at this level. Individuals with comorbidity of cardiovascular disease, diabetes and cancer at baseline were excluded and we have adjusted for age and smoking habits as completely as possible including tobacco consumption and inhalation habits.
Comparison with other studies
Prevalence of stage 0
The prevalence of individuals with stage 0 in our study was 2.2% in men and 2.4% in women. In the Copenhagen City Heart study the prevalence was slightly higher; criteria for GOLD Stage 0 were met in 5.8% of the total adult population and in 7.2% of smokers [4]. In the National Health and Nutrition Examination Study (NHANES) study [14] the prevalence of respiratory symptoms among individuals with normal pulmonary function was substantially higher (approximately 13%). This is probably mostly dependent on different selection criteria due to variation in the definition of stage 0. Vestbo and coworkers limited the selection of individuals to those with symptoms of chronic bronchitis but included all individuals without airflow limitation irrespective of the level of FEV1. On the other hand the NHANES study had higher prevalence rates probably due to a broader definition of the respiratory symptoms [14]. In our study a specific definition of symptoms of chronic bronchitis was used and we only selected individuals with normal pulmonary function, both with respect to FEV1 and the FEV1/FVC ratio. Consequently the prevalence of stage 0 was lower in our study than in the NHANES [14] and the Copenhagen City Heart study [4].
Mortality risks associated with symptoms of chronic bronchitis
Among male smokers of our study, the HR of total mortality in GOLD stage 0 was strongly significant irrespective of whether we used "ever" or "recent" symptoms of chronic bronchitis" to define stage 0. The HR:s were of a similar magnitude as for stage 2. Among men with GOLD stage 1 in men and among women with stage 2 in women, increased HR:s were noticed in smokers who reported symptoms of chronic bronchitis, implying that respiratory symptoms might be important for an adverse prognosis. However, in GOLD stage 3 we were not able to show an additive risk associated with symptoms of chronic bronchitis possibly due to low statistical power.
Similar results have previously been demonstrated in some studies. In the Copenhagen City Heart study, chronic mucus hypersecretion was associated with an increased risk of all cause mortality statistically significant in men only (relative risk: 1.3 in men and 1.1 in women) [7]. There was no interaction between the effect of chronic mucus hypersecretion and FEV1 percent predicted with regard to total mortality which is in agreement with our results. However, also in the same study, regarding obstructive lung disease mortality, the effect of chronic mucus hyper-secretion varied with the level of ventilatory function, being weak in individuals with normal ventilatory function but more pronounced in individuals with reduced ventilatory function. This interaction was statistically significant [7]. Furthermore, another study using the Copenhagen City Heart study population showed that chronic mucus hyper-secretion was associated with later hospitalization due to COPD, relative risks were 2.4 (1.3 to 4.5) for men and 2.6 (1.2 to 5.3) for women [16].
In contrast, in the National Health and Nutrition Examination Study (NHANES) the presence of respiratory symptoms was not associated with an increased mortality risk in subjects with normal lung function [14]. The relative risk (1.2) was elevated but not statistically significant probably due to the fact that a much wider definition of respiratory symptoms was used which included as well individuals with wheeze and a diagnosis of asthma. In our study based on a similar number of deaths, but also on a more specific definition of respiratory symptoms (symptoms of chronic bronchitis) a HR of 1.6 was statistically significant among smoking men.
Occupational studies have shown results in agreement with our study. In a population of 1061 men working in the Paris area and followed for 22 years, chronic phlegm production was significantly associated with mortality: Relative risk 1.35; (p less than 0.01) [6]. In a group of gold miners, chronic mucus hyper-secretion was not related to COPD mortality, but related to mortality in ischemic heart disease and other causes after adjustment for tobacco and dust exposure [17].
On the contrary, in another occupational cohort of 2,718 British men examined between 1954 and 1961, and followed for 20 to 25 years, the risk of death from COPD was correlated with the initial degree of air-flow obstruction. Among men with similar initial air-flow obstruction, age-specific COPD death rates were not significantly related to initial mucus hypersecretion [18].
The explanation for the relation between chronic productive cough and future mortality risk is still unclear. Previously, in the classic study by Fletcher and coworkers, conducted in a cohort of working men in London, chronic phlegm production was unrelated to the development of airway obstruction and regarded as a less important condition [19]. Furthermore, the presence of stage 0 at baseline was not a meaningful predictor for the subsequent development of COPD in later stages after 5 and 15 years follow-up in a Danish prospective population based study [4].
Thus it seems that the symptoms of chronic bronchitis do not affect the FEV1 decline and development of COPD to the same extent that it affects all-cause and obstructive lung disease mortality and risk of COPD hospitalization. An explanation could be that symptoms of chronic bronchitis may be an independent pulmonary disorder associated with inflammatory and/or infectious exacerbations [20] and possibly also with the development of other diseases such as lung cancer [15] and thus not specifically associated with COPD development and progression. We were able to analyse cause specific mortality in GOLD stage 0 in male smokers. Our findings of significantly increased HR for lung cancer, all cancers, and other diseases in individuals of stage 0 support this view.
Mortality risks in stages 1–4
GOLD stage 1 in our study, as well as in the NHANES study [14], conveyed only a slight increase in total mortality risk of borderline significance, even lower than found for GOLD stage 0. Among former smokers there was a tendency towards a decreased risk. Thus among individuals who quit smoking when still in GOLD stage 1, the future mortality risk will be normal. In GOLD stages 2 to 4 our findings of increased mortality HR are in agreement with other previously conducted studies, as they actually reflect a successive reduction in FEV1 [21-23]. Mortality risks were also increased among never smokers, and these results are entirely in line with other studies [24]. The NHANES study showed increased mortality risks associated with moderate to severe COPD, among smokers and former smokers of approximately the same magnitude as we report [14].
Study limitations
A methodological limitation of these findings is the measurement of pulmonary function. The protocol of the lung function tests did not meet the standards of current recommendations [12]. The values of the internal regression equation based on height and age and the standard deviation of the equation is similar to that of the Working party of the European Community for Steel and Coal [25]. Hence, we believe that the methodological error is similar to the one achieved after optimal spirometric technique. However there remains the possibility of misclassification due to decreased specificity in the spirometries which could have resulted in an underestimation of the results rather than the reverse.
Another possible limitation of this study is dependent on the fact that smoking prevalence in Sweden has decreased since the 1970's [26]. Some smokers in this study are likely to have stopped smoking during the follow-up time. Smoking habits were assessed at one point in time and not repeated thereafter. Hence, we have not been able to calculate accumulated smoking exposure in terms of individual pack-years. However, even the use of pack-years has its limitations, especially with respect to recall biases in subjects with a long history of smoking [27].
Furthermore, we have no follow-up data on symptoms of chronic bronchitis and thus we were not able to determine whether the symptoms of chronic bronchitis were stable or not. According to previous studies symptoms of chronic bronchitis may disappear in some individuals [4]. In those individuals, symptoms of chronic bronchitis may have a limited influence on mortality risks and could cause an underestimation of the effect of persistent symptoms of chronic bronchitis in this study. In addition, we have no spirometric follow-up data and hence we are not able to determine whether individuals with GOLD stage 0 at baseline have developed COPD in the later GOLD stages during the follow up time.
Another limitation of the study was the lower statistical power in the analyses of women. Our findings in women, especially in former smokers and never smokers should therefore be interpreted with caution.
Implications of our findings
We believe that symptoms of chronic bronchitis should be regarded as a marker for an increased all-cause mortality risk. Our data show that among middle-aged smokers, symptoms of chronic bronchitis might represent an independent risk factor for an early mortality equal to that of having moderate COPD.
It is however uncertain from previous studies whether symptoms of chronic bronchitis really are related to the development and progression of COPD [4]. An increased risk of COPD mortality associated with chronic mucus hyper secretion among individuals with established COPD has been shown previously [7]. In our study, symptoms of chronic bronchitis increased the risk of all cause mortality and mortality in cancer, lung cancer and other causes in smoking individuals with normal lung function (stage 0) and the risk of total mortality in individuals with GOLD stage 1 and 2. Our findings indicate that symptoms of chronic bronchitis might be an independent pulmonary disorder that may or may not be associated with COPD. It is therefore questionable whether GOLD stage 0 should be used as an "at risk" stage for COPD development in future COPD classifications.
Finally, GOLD stage 0 indicates a group of individuals with a very unfavourable prognosis similar to that of individuals with GOLD stage 2. Among smokers, symptoms of chronic bronchitis contributed to an adverse prognosis not only in individuals "at risk", but also in those with early stages of COPD with GOLD 1 and 2.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
Marie Ekberg-Aronsson: Taking part in the conception and design of the study. Performed the statistical analyses and interpretation of data, drafting and revising the article, giving final approval to the version to be published.
Kerstin Pehrsson: Taking part in the conception and design of the study. Interpretation of data and revising the article, giving final approval to the version to be published.
Jan-Åke Nilsson: Taking part in the conception and design of the study. Statistical analysis and interpretation of data, giving final approval to the version to be published.
Peter Nilsson: Taking part in the conception and design of the study, acquisition and interpretation of data and revising the article, giving final approval to the version to be published.
Claes-Göran Löfdahl: Taking part in the conception and design of the study, interpretation of data and revising the article, giving final approval to the version to be published.
Acknowledgements
This study was supported by grants from the City of Malmö, The Swedish Medical Research Council, and the Swedish Heart and Lung Foundation
==== Refs
Hurd S The impact of COPD on lung health worldwide: epidemiology and incidence Chest 2000 117 1S 4S 10673465 10.1378/chest.117.2_suppl.1S
Pauwels RA Buist AS Calverley PM Jenkins CR Hurd SS Global strategy for the diagnosis, management, and prevention of chronic obstructive pulmonary disease. NHLBI/WHO Global Initiative for Chronic Obstructive Lung Disease (GOLD) Workshop summary Am J Respir Crit Care Med 2001 163 1256 76 11316667
Global Initiative For Chronic Obstructive Lung Disease™ GOLD workshop report update 2004
Vestbo J Lange P Can GOLD Stage 0 provide information of prognostic value in chronic obstructive pulmonary disease? Am J Respir Crit Care Med 2002 166 329 32 12153965 10.1164/rccm.2112048
Celli BR Cote CG Marin JM Casanova C Montes de Oca M Mendez RA Pinto Plata V Cabral HJ The body-mass index, airflow obstruction, dyspnea, and exercise capacity index in chronic obstructive pulmonary disease N Engl J Med 2004 350 1005 12 14999112 10.1056/NEJMoa021322
Annesi I Kauffmann F Is respiratory mucus hypersecretion really an innocent disorder? A 22-year mortality survey of 1,061 working men Am Rev Respir Dis 1986 134 688 93 3767125
Lange P Nyboe J Appleyard M Jensen G Schnohr P Relation of ventilatory impairment and of chronic mucus hypersecretion to mortality from obstructive lung disease and from all causes Thorax 1990 45 579 85 2402719
Trell E Community-based preventive medical department for individual risk factor assessment and intervention in an urban population Prev Med 1983 12 397 402 6878198 10.1016/0091-7435(83)90248-7
Berglund G Eriksson KF Israelsson B Kjellstrom T Lindgarde F Mattiasson I Nilsson JA Stavenow L Cardiovascular risk groups and mortality in an urban swedish male population: the Malmo Preventive Project J Intern Med 1996 239 489 97 8656142 10.1046/j.1365-2796.1996.483819000.x
Berglund G Nilsson P Eriksson KF Nilsson JA Hedblad B Kristenson H Lindgarde F Long-term outcome of the Malmo preventive project: mortality and cardiovascular morbidity J Intern Med 2000 247 19 29 10672127 10.1046/j.1365-2796.2000.00568.x
Janzon L Lindell SE Trell E Larme P Smoking habits and carboxyhaemoglobin. A cross-sectional study of an urban population of middle-aged men J Epidemiol Community Health 1981 35 271 3 7200119
Engstrom G Hedblad B Nilsson P Wollmer P Berglund G Janzon L Lung function, insulin resistance and incidence of cardiovascular disease: a longitudinal cohort study J Intern Med 2003 253 574 81 12702035 10.1046/j.1365-2796.2003.01138.x
Cox DR Regression models and life-tables (with discussion) Journal of the Royal Statistical Society 1972 Serie B 187 220
Mannino DM Buist AS Petty TL Enright PL Redd SC Lung function and mortality in the United States: data from the First National Health and Nutrition Examination Survey follow up study Thorax 2003 58 388 93 12728157 10.1136/thorax.58.5.388
Lange P Nyboe J Appleyard M Jensen G Schnohr P Ventilatory function and chronic mucus hypersecretion as predictors of death from lung cancer Am Rev Respir Dis 1990 141 613 7 2310094
Vestbo J Prescott E Lange P Association of chronic mucus hypersecretion with FEV1 decline and chronic obstructive pulmonary disease morbidity. Copenhagen City Heart Study Group Am J Respir Crit Care Med 1996 153 1530 5 8630597
Wiles FJ Hnizdo E Relevance of airflow obstruction and mucus hypersecretion to mortality Respir Med 1991 85 27 35 2014356
Peto R Speizer FE Cochrane AL Moore F Fletcher CM Tinker CM Higgins IT Gray RG Richards SM Gilliland J The relevance in adults of air-flow obstruction, but not of mucus hypersecretion, to mortality from chronic lung disease. Results from 20 years of prospective observation Am Rev Respir Dis 1983 128 491 500 6614643
Fletcher C Peto R The natural history of chronic airflow obstruction Br Med J 1977 1 1645 8 871704
Prescott E Lange P Vestbo J Chronic mucus hypersecretion in COPD and death from pulmonary infection Eur Respir J 1995 8 1333 8 7489800 10.1183/09031936.95.08081333
Schunemann HJ Dorn J Grant BJ Winkelstein W JrM Trevisan Pulmonary function is a long-term predictor of mortality in the general population: 29-year follow-up of the Buffalo Health Study Chest 2000 118 656 64 10988186 10.1378/chest.118.3.656
Vollmer WM McCamant LE Johnson LR Buist AS Respiratory symptoms, lung function, and mortality in a screening center cohort Am J Epidemiol 1989 129 1157 69 2729254
Hole DJ Watt GC Davey-Smith G Hart CL Gillis CR Hawthorne VM Impaired lung function and mortality risk in men and women: findings from the Renfrew and Paisley prospective population study Bmj 1996 313 711 5 discussion 715–6 8819439
Strachan DP Ventilatory function, height, and mortality among lifelong non-smokers J Epidemiol Community Health 1992 46 66 70 1573363
Quanjer PH Tammeling GJ Cotes JE Fabbri LM Matthys H Pedersen OF Peslin R Roca J Sterk PJ Ulmer WT Symbols, abbreviations and units. Working Party Standardization of Lung Function Tests, European Community for Steel and Coal Eur Respir J Suppl 1993 16 85 100 8499056
Tillgren P Haglund BJ Lundberg M Romelsjo A The sociodemographic pattern of tobacco cessation in the 1980s: results from a panel study of living condition surveys in Sweden J Epidemiol Community Health 1996 50 625 30 9039380
Bernaards CM Twisk JW Snel J Van Mechelen W Kemper HC Is calculating pack-years retrospectively a valid method to estimate life-time tobacco smoking? A comparison between prospectively calculated pack-years and retrospectively calculated pack-years Addiction 2001 96 1653 61 11784461 10.1046/j.1360-0443.2001.9611165311.x
|
16176592
|
PMC1224874
|
CC BY
|
2022-09-08 23:19:43
|
no
|
Saline Syst. 2005 Jul 4; 1:1
|
latin-1
|
Saline Syst
| 2,005 |
10.1186/1746-1448-1-1
|
oa_comm
|
==== Front
Saline Syst
Saline Syst
Saline Systems
1746-1448
BioMed Central
1746-1448-1-2
16176593
10.1186/1746-1448-1-2
Review
A hundred years of Dunaliella research: 1905–2005
Oren Aharon [email protected]
1 The Institute of Life Sciences, The Hebrew University of Jerusalem, 91904 Jerusalem, Israel
2 the Moshe Shilo Minerva Center for Marine Biogeochemistry, The Hebrew University of Jerusalem, 91904 Jerusalem, Israel
2005
4 7 2005
1 22
3 4 2005
4 7 2005
Copyright © 2005 Oren; licensee BioMed Central Ltd.
2005
Oren; licensee BioMed Central Ltd.
https://creativecommons.org/licenses/by/2.0/ This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
A hundred years have passed since the description of the genus Dunaliella, the unicellular green alga which is responsible for most of the primary production in hypersaline environments worldwide. The present paper provides an historical survey of research on Dunaliella, from the early work in the 19th century to the thorough taxonomic studies by Teodoresco, Hamburger, Lerche and others from the beginnig of the 20th century onwards. It attempts to trace the origin of some of the most important breakthroughs that have contributed to our present understanding of this alga that plays such a key role in many hypersaline environments.
==== Body
pmc1. Introduction
A hundred years have passed since the description of the genus Dunaliella, the unicellular green alga which is responsible for most of the primary production in hypersaline environments worldwide. First sighted in 1838 in saltern evaporation ponds in the south of France by Michel Felix Dunal [1], it was named after its discoverer by Teodoresco in 1905 [2].
In the century that has elapsed since its formal description, Dunaliella has become a convenient model organism for the study of salt adaptation in algae. The establishment of the concept of organic compatible solutes to provide osmotic balance was largely based on the study of Dunaliella species. Moreover, the massive accumulation of β-carotene by some strains under suitable growth conditions has led to interesting biotechnological applications.
The present paper provides an historical survey of research on Dunaliella, from the early work in the 19th century to the thorough taxonomic studies by Teodoresco, Hamburger, Lerche and others from the beginning of the 20th century onwards. It attempts to trace – often through quotations from the original articles – the origin of some of the most important breakthroughs that have contributed to our present understanding of this alga that plays such a key role in many hypersaline environments.
Extensive additional information on the alga can be found in a review by Ginzburg [3], in the multi-author review edited by Avron and Ben-Amotz [4], and in my monograph on halophilic microorganisms and their environments [5].
2. Reports on Dunaliella Prior to 1905
The first description of a unicellular biflagellate red-colored alga living in concentrated brines (Fig. 1) was given in 1838 by Dunal [1], who reported occurrence of the organism we know today as Dunaliella salina in the salterns of Montpellier, on the Mediterranean coast of France. He named the organisms observed Haematococcus salinus and Protococcus. The discovery of these algae was made in the framework of an investigation, invited by the Académie des Sciences, Paris, of the cause of the red coloration of saltern brines. At the time it was widely assumed that chemical and physical parameters are responsible for the coloration of these brines. Dunal refuted an earlier claim that the color is due to the brine shrimp Artemia salina. The Académie then appointed a committee to reexamine the matter, and this committee confirmed Dunal's finding [6]; see also [7]. Another idea brought forward during that period was that Artemia contributes to the color due to the partially digested and decaying red flagellates ("Monas Dunalii") present in its intestine [8]. Nowadays it is clear that, although β-carotene-rich Dunaliella salina are indeed present in the saltern ponds, most of the coloration of the crystallizer brine is caused not by the algae but by red halophilic Archaea instead [9,10].
Figure 1 Dunaliella salina cells from the crystallizer brine of the salterns in Eilat, at the Red Sea coast of Israel.
In the course of the 19th century, Dunal's red flagellate algae has been observed by other biologists as well in salt lakes and other hypersaline sites in Crimea [11,12], Algeria [13], Lorraine, France [14] and Romania [15]. Different names were attached to the organism by each investigator [1,8,11,13,15-20] (Table 1).
Table 1 Names attached to the red-colored unicellular flagellate algae observed in hypersaline brines, 1838–1906.
Name Author
Haematococcus salinus Dunal [1]
Protococcus salinus Dunal [1]; Geleznow [11]
Monas Dunalii Joly [8]; Blanchard [16]; Butschinsky [12]
Diselmis Dunalii Dujardin [17]
Chlamydomonas Dunalii Cohn [18]; Blanchard [13]; Bujor [15]
Sphaerella lacustris var. Dunalii Hansgirg [19]
Dunaliella salina Teodoresco [2,20]
3. The Description of the Genus Dunaliella
The year 1905 saw the publication of two papers presenting in-depth descriptions of Dunaliella as a new genus, one by E.C. Teodoresco from Bucharest [2] and the second written by Clara Hamburger from Heidelberg [7]. Teodoresco's publication preceded that by Hamburger, who only learned about the Teodoresco paper when finalizing the writing of her own article [7]:
"Anfang März wollte ich an die Ausarbeitung meiner Notizen gehen, als ich am 10. März von Herrn Prof. Lauterborn eine Arbeit von Teodoresco mit dem Titel: "Organisation et développement du Dunaliella, nouveau genre de Volvocaceae – Polyblepharidée erhielt, welche as Separatdruck aus dem botanischen Centralblatt soeben versendet war. Dunaliella ist der von mir untersuchte Organismus, den ich schon als Vertreter einer neuen Gattung erkannt hatte. Unsere Resultaten stimmten in vielen Punkten überein, in anderen müssen meiner Ansicht nach erst weitere Untersuchungen die entgültige Entscheidung bringen. Da jedoch meine Studien, besonders bezüglich des innern Baues eingehender sind (Teodoresco hat nur lebendes Material untersucht) und ich auch einige noch bestehende Lücken ausfüllen kann; da ferner alle meine Resultate unbeinflußt von denen Teodoresco's erhalten wurden, so möchte ich sie dennoch veröffentlichen."
[In the beginning of March [1905] I wanted to start to work out my notes, when on March 10 I received from Prof. Lauterborn a paper by Teodoresco entitled: "Organization and development of Dunaliella, a new genus of the Volvocida – Polyblepharidae", which was just sent as offprint from the Botanisches Centralblatt. Dunaliella is the organism that I had been investigating, and that I had already recognized as representative of a new genus. Our results corresponded in many respects, while in other respects I am of the opinion that further investigations will have to decide. However, because my studies, especially with respect to the internal structure, are more thorough (Teodoresco had studied only live material) and I also can fill in certain still existing gaps in the knowledge, and also because my results were obtained independently of those of Teodoresco, I still would like to publish them.]
Teodoresco studied material collected from a Romanian salt lake, while Hamburger worked with samples sent to her from the salterns of Cagliari, Sardinia. Both authors presented detailed drawings of the organisms (Fig. 2 and 3) and provided extensive information on its morphology, cell structure, reproduction, behavior and ecology. A formal description of the genus Dunaliella, named in honor of Dunal who had first seen these organisms in salterns in France almost seventy years earlier, and of the first two species within the genus, D. salina and D. viridis, was published in 1906 [20].
Figure 2 Drawings by Hamburger (1905) of red cells (Dunaliella salina) (1–4) and green cells (D. viridis) (5–8), diverse shapes observed in a drop that becomes more concentrated by evaporation (9–29), spherical forms obtained upon dilution (30–31), and initiation of cell division (32–34). From [2].
Figure 3 Dunaliella salina (1–17) preserved with different fixation techniques; (3), pigment crystals; (8), granules (starch?); (10–13), division stages; (14,15), aplanospores; (16–17), and green cells (D. viridis?), as drawn by Hamburger (1905) [7].
The papers by Teodoresco and Hamburger were soon followed by others. Noteworthy studies in the early years of Dunaliella research are articles by Cavara [21], who extended the study of the organism in the Cagliari, Sardinia salterns, a study of the algae in the Salton Sea, California [22], a series of ecological papers by Labbé based on observations made in the salterns of Le Croisic on the Atlantic coast of France [23-25], articles by Baas Becking and coworkers, who collected specimens from all over the world [26-28], and the taxonomic studies by Hamel [29] and Lerche [30].
4. The Taxonomy of the Genus Dunaliella
Dunaliella is a genus of unicellar algae belonging to the family Polyblepharidaceae. Its cells lack a rigid cell wall, and they reproduce by longitudinal division of the motile cell or by fusion of two motile cells to form a zygote.
Teodoresco [2,20] described two species: D. salina and D. viridis. D. salina has somewhat larger cells, and under suitable conditions it synthesizes massive amounts of carotenoid pigments, coloring the cells brightly red. D. viridis never produces such red cells. It is interesting to note that in the early years there have been extensive discussions whether indeed two species are present or whether the red and the green cells represent different forms of the same species. For example, Blanchard [13] and Hamburger [7] considered the green cells as juvenile stages of the red ones. Labbé [23,24] was of the opinion that differences in salt concentration of the environment are responsible for the different colors of the cells. Upon transferring of saltern brine samples to a lower salinity he grew a form of Dunaliella adapted to fresh water and lacking the brown-red pigment. His statements that:
"En ce qui concerne les facteurs de la transformation, l'hypothèse simpliste de Teodoresco ne peut être conservée et it ne s'agit pas là de deux espèces distinctes (D. salina et D. viridis). Il s'agit d'une alternance de formes due aux changements de milieu."
[Concerning the factors of the transformation, the simplistic hypothesis of Teodoresco cannot be maintained, and we do not have here two distinct species (D. salina and D. viridis). We deal with an alternation of forms due to environmental changes.]
and:
"L'organisme qui colore en rouge les marais salants et à qui nous pouvons conserver le nom de Dunaliella salina n'est que la phase ultime de l'évolution d'un flagellé chlorophyllien voisin de Volvocinées, très eurihyalin, qui en eau sursalée donne les formes sténohyalines non réversibles aux formes chlorophylliennes, et colorées par un hématochrome."
[The organism which colors the salterns red and for which we can conserve the name Dunaliella salina is nothing but the final phase in the development of a very euryhyaline chlorophyll-containing flagellate related to the Volvocinae, which in hypersaline water produces stenohyaline foms that cannot revert to chlorophyll-containing forms, and are colored by a hematochrome.]
have not withstood the test of time. We now know that not all Dunaliella species produce massive amounts of carotene, and those that can, do so only under suitable conditions (exposure to high light intensities, nutrient limitation, etc., see also section 6 below). Lerche [30] thus saw that under suitable conditions all red clones became green, but after several weeks they turned olive to yellow-green and after several months they were red again.
Additional species were later added to the genus, especially thanks to the in-depth studies by Lerche [30] and Butcher [31] (Table 2). Lerche investigated material collected from salt lakes in Romania and in California, as well as the above-mentioned Cagliari salterns. She concluded that the former species D. viridis is heterogeneous and should be split into several new species. Thus the species D. media, D. euchlora, D. minuta, and D. parva were created. It must be stressed here that not all species mentioned tolerate the extremely high salt concentrations in which D. salina and D. viridis are found in nature. Some are typically marine organisms that were never reported to occur in hypersaline environments.
Table 2 Selected Dunaliella species
Name Author, year
D. salina Teodoresco, 1905, 1906 [2,20]
D. viridis „
D. peircei Nicolai and Baas Becking, 1935 [28]
D. parva Lerche, 1937 [30]
D. media „
D. euchlora „
D. minuta „
D. tertiolecta Butcher, 1959 [31]
D. primolecta „
D. quartolecta „
D. polymorpha „
An in-depth taxonomic treatment of the genus was given in Massyuk's 1973 monograph [32]. She divided the genus into two subgenera, Dunaliella (23 species) and Pascheria (5 species), the latter consisting of freshwater species only. Some of the species recognized by Massyuk may eventually be found to be polymorphic forms of a single taxon [33].
A species of considerable interest is Dunaliella acidophila, isolated from acidic waters and soils in the Czech Republic and in Italy [34,35]. This is not a true halophile but an acidophilic alga that grows optimally at pH values between 0.5 and 2. In recent years it has become a popular research object for the study of adaptation of life to low pH environments [36]. Its taxonomic/phylogenetic affiliation with the halophilic Dunaliella species has to my knowledge never been verified.
Molecular phylogeny techniques have been applied to the taxonomic study of Dunaliella from 1999 onwards. These studies have encompassed the 18S rRNA genes and the internal transcribed spacer regions, and have been based on gene sequence comparisons as well as on restriction fragment length polymorphism studies. Little correlation was found between the molecular data and the morphological-physiological attributes used in older studies to delineate species within the genus [37,38]. On the basis of 18S rRNA gene sequences, Olmos et al. [39] could differentiate between D. salina, D. parva and D. bardawil as species containing one, two and three introns, respectively, within the 18S rRNA gene. The molecular studies have made it clear that many culture collection strains are probably misnamed, and that some unnecessary species names may have been proposed in the past.
5. Life Stages and Sexual Reproduction in Dunaliella
Dunaliella salina and some of the other species undergo complex life cycles that encompass, in addition to division of motile vegetative cells, the possibility of sexual reproduction. Fusion of two equally sized gametes to form a zygote was documented in many of the early studies [7,20,29]. We thank a most detailed study of sexual reproduction in Dunaliella to Lerche [30], who reported sexual zygote formation in five of the six species studied (D. salina, D. parva, D. peircei, D. euchlora, and D. minuta). She reported zygote formation in D. salina to be induced by a reduction in salt concentration from 10 to 3%. In the process first the flagella touch, and then the gametes form a cytoplasmic bridge and fuse. The zygote has a thick outer layer. It can withstand exposure to freshwater and also survive prolonged periods of dryness. These zygotes germinate with the release of up to 32 haploid daughter cells through a tear in the cell envelope [30]. It is well possible that the cyst-like structures observed by Oren et al. [40] at the end of a bloom of green Dunaliella cells in the Dead Sea in 1992 were actually such zygotes. In this case, however, the formation of these rounded, thick-walled cells took place at a time of an increase in water salinity. Lerche [30] performed a series of elegant experiments in which carotenoid-rich red cells were crossed with green cells, enabling the investigator to follow the fusion of the two parent cells to form a zygote. A few of her drawings to illustrate the process are reproduced in Fig. 4. The possibility of formation of asexual resting cysts by D. salina was indicated by Hamburger [7], a finding that was disputed by Lerche. However, more recently, Loeblich [41] has reported formation of such cysts in media of reduced salinity (for a discussion see also [42]).
Figure 4 Aggregation of the red and the green form of Dunaliella salina (upper part) and zygote formation of D. salina (green and red form) (lower part). From [30].
Some Dunaliella species can also develop a vegetative palmelloid stage consisting of round non-motile cells. Lerche [30] has documented this phenomenon in D. salina cultures at lowered salinities, and Brock [43] observed such palmelloid forms of Dunaliella in benthic algal mats of Great Salt Lake, Utah.
6. Carotenoid Pigments of Dunaliella
The pigment responsible for the brightly red coloration displayed by D. salina, often designated in the older literature as "hematochrome", was recognized already very early as a carotenoid. As such it was identified by Blanchard [13], and Teodoresco [20], Lerche [30] and Ruinen [44] confirm this identification based on the solubility of the pigment in alcohol and in ether and on the blue color formed in the presence of concentrated sulfuric acid.
Before the modern electron microscope showed the β-carotene as granules between the thylakoids of the cell's single chloroplast, considerable differences of opinion existed regarding the intracellular location of this red carotenoid pigment. Thus, both Teodoresco [2,20] and Labbé [23] stated that the red pigment was distributed all over the cells' cytoplasm. Relating to a different claim by Hamburger [7], Teodoresco [20] wrote:
"je n'hésite pas à croire que ce pigment imprègne tout le corps des zoospores, excepté, bien entendu, l'extrémité antérieure, à l'endroit de l'insertion des flagellums."
[I don't hesitate to believe that the pigment impregnates the whole body of the zoospores, except, of course, the extreme anterior part, on the place where the flagella are inserted.]
Likewise, Hamel [29] claimed that at elevated salt concentrations, D. salina forms "hematochrome" that penetrates not only the "chromophore" (= chloroplast) but the entire cytoplasm as well. On the other hand, Hamburger believed the red pigment to be located as small droplets (which is true, see e.g. [45,46]), but she was mistaken about the location of the pigment:
"Er tritt in Form kleiner Tröpfchen auf, und ist, wie mir sicher scheint, nur der äußeren Alveolarschicht des Plasmas eingelagert, während das Chromatophor Träger des grünen Farbstoffes ist. Die Bemerkung Teodoresco's "hématochrome imprégnant non seulement le chromatophore, mais encore tout le corps des individus âgés", stimmt mit meinen Beobachtungen nicht überein."
[It occurs in the form of small droplets, and is, as seems sure to me, only deposited in the outer alveolar layer of the plasma, while the chromatophore is the bearer of the green pigment. The remark by Teodoresco that "the hematochrome that impregnates not only the chromatophore, but also the whole body of adult individuals" does not correspond with my observations.]
Baas Becking [27] correctly located the red-orange pigment in the chloroplast, and Leche [30] realized that the carotene masks the chlorophyll, so that the chloroplast can assume all shades from orange-red to yellow-green, olive and green:
"Der rote Farbstoff ist in Form öliger Tröpfchen zwischen den Wabe des Chloroplasten eingelagert und nicht wie Hamburger (1905) annimmt, in der äußeren Alveolarschicht des Protoplasmas."
[The red pigment is located in the form of oily droplets between the honeycomb structure of the chloroplast and not, as Hamburger (1905) assumes, in the outer cytoplasmic layer of the protoplast.]
β-Carotene, the major carotenoid accumulated by D. salina and D. bardawil, is a valuable chemical, in high demand as a natural food coloring agent, as pro-vitamin A (retinol), as additive to cosmetics, and as a health food [47]. Some Dunaliella strains can accumulate very large amounts of this carotenoid. Thus, as much as 13.8% of the total dry organic matter in the D. salina community in Pink Lake, Victoria, Australia, was estimated to be β-carotene [48]. Also in culture some strains may contain up to 10% and more of β-carotene in their dry weight, including a large percentage of the 9-cis isomer [46]. Therefore the biotechnological potential of Dunaliella as a source of β-carotene was investigated already relatively early. The first pilot plant for Dunaliella cultivation for β-carotene production was established in the USSR in 1966 [49,50]. The commercial cultivation of Dunaliella for the production of β-carotene throughout the world is now one of the success stories of halophile biotechnology [51-53]. Different technologies are used, from low-tech extensive cultivation in lagoons to intensive cultivation at high cell densities under carefully controlled conditions [54].
One of the methods used in such biotechnological operations to induce massive carotenoid accumulation is reduction of the growth rate by deprivation of nutrients. That a high carotenoid content of the cells may be caused by nutrient limitation as well as by high light intensities was already reported by Lerche [30]:
"Da die Rotfärbung besonders in alten Kulturen auftrat, lag die Annahme nahe, sie in Zusammenhang mit den Ernährungsbedingungen und speziell mit dem Fehlen eines oder mehrerer Stoffe zu bringen. Da Phosphor und Stickstoff bei der pflanzlichen Ernährung häufig die Stoffe sind, die im Minimum vorhanden sind, wurde das Augenmerk zunächst auf diese Stoffe gerichtet."
[As the red coloration occurred especially in old cultures, it was reasonable to assume a correlation with the nutritional conditions and in particular with the lack of one or more compounds. As phosphorus and nitrogen are in plant nutrition often the substances present in limiting amounts, we directed our attention first of all to these substances.]
7. Population Dynamics of Dunaliella in Salt Lakes and Salterns
Only few studies have been devoted to the quantitative evaluation of Dunaliella populations in salt lakes and salterns, the dynamics of their appearance and decline, and their contribution to the primary production in their habitats. Stephens and Gillespie (1976) reported measurements of the primary production in the south arm of Great Salt Lake, Utah, performed in 1973 (salinity around 135 g/l). Post [56] reported that in the cold season, round cyst-like cells of D. salina increased in numbers in the Great Salt Lake, especially on the lake's bottom. In the Dead Sea, green Dunaliella cells have been reported since the 1940s [57]. The first quantitative estimates of the Dunaliella population in the lake were made in 1964, and showed very high numbers: up to 4 × 104 cells per ml of surface water (sampling season not specified) [58]. Systematic monitoring of the population density at different seasons and depths in the Dead Sea from 1980 onwards have yielded a clear picture of the factors that determine development of the alga in this unusual environment. High concentrations of magnesium and calcium ions are known to be inhibitory to Dunaliella since Baas-Becking's earlier studies [27]. Dunaliella blooms therefore occur in the Dead Sea only when during unusually wet winters the upper water layers of the lake become sufficiently diluted to enable growth, and when phosphate, the limiting nutrient, is available. Such events have been observed in 1980 and again in 1992 [40,59].
Surprisingly, very little is known about the factors that determine the dynamics of Dunaliella in saltern pond systems. It is therefore interesting to note that some of the most in-depth studies on this topic were performed in the early 1920s in the salterns of Le Croisic on the Atlantic coast of France, where salt making is a seasonal operation. Labbé [23,25] showed changes in the algal community structure and related these to changes in salinity ("osmotic pressure; viscosity") of the brine, but he also recognized the role of the light intensity and the water temperature, as well as that of the pH. Based on the faulty assumption that the smaller green and the larger red Dunaliella cells are stages in the development of a single organism (see section 4 above), he described an annual cycle in which in the beginning of the winter few red motile cells ("érythrospores") and smaller green motile cells ("chlorospores") are present [24]. Dilution of the water by winter rains triggers the formation of red cysts ("érythrocystes"), but the "chlorospores" develop rapidly, conjugate, and form "chlorocystes". When the salt concentration increases in the summer season, red motile cells start to appear, always accompanied by green cells:
"Peu à peu, les érythrospores provenant de chlorospores prolifèrent, et leur dominance est fonction de la concentration saline."
[Gradually the "erythrospores" that are formed from "chlorospores" proliferate, and their dominance is a function of the salt concentration.]
8. Cultivation and Salt Tolerance of Dunaliella
The first controlled experiments to evaluate the effect of salinity on the growth rate of different Dunaliella isolates were reported in the 1930s. Baas-Becking [27] observed that D. viridis thrives equally well over the whole range of 1–4 M (6–23%) NaCl and over the pH range 6–9. He found calcium and magnesium ions in high concentrations to be inhibitory. More detailed and well-documented experiments, using a variety of species and isolates, were reported by Lerche [30]. She found most isolates to grow optimally between 2 and 8% salt, with very slow growth, if at all, at salt concentrations above 15% (Fig. 5). Between 0.47 and 1.22 divisions per day were recorded under optimal conditions.
Figure 5 Division rate ("Teilungsrate") (as number of divisions per day) of different Dunaliella isolates belonging to several species, as a function of the NaCl concentration of the medium. From [30].
The nutritional requirements of different Dunaliella strains were investigated in-depth by Gibor [60], Johnson et al. [61], Van Auken and McNulty [62], and others, enabling the optimization of media to grow the alga. Optimal salt concentrations for cultivation varies according to the strain, with values reported for D. viridis around 6%, for D. salina around 12% [42], while different Great Salt Lake isolates had optima of 10–15% or even 19% salt [43,62]. A general trend, observed in all these studies, is that the actual salinity of the environments from which the strains had been isolated was always much higher than the salt concentration found to be optimal in laboratory experiments. This may well reflect the fact that growth of an organism occurs in a certain environment not necessarily means that that environment is optimal for its development, but rather that the organism performs there better than all its competitors.
9. Osmotic Behavior of Dunaliella Cells
Dunaliella cells lack a rigid cell wall, and the cell is enclosed solely by a thin elastic plasma membrane. As a result, the cells' morphology is strongly influenced by osmotic changes. This was documented already in the early days. The descriptions by Teodoresco [2] are very exact here, and they deserve to be cited unabridged:
"Ces zoospores sont dépourvues de membrane cellulosique; celle-ci est représentée par une enveloppe qui possède une certaine souplesse et une certaine extensibilité, qui permet au corps de prendre les formes assez variées, suivant la concentration de l'eau. A ce point de vue, le genre Dunaliella diffère totalement de toutes les espèces de Chlamydomonas ..."
[These zoospores are devoid of a cellulose cell wall; instead there is a cell envelope that possesses a certain flexibility and a certain elasticity, which allows the body to take quite different forms, in accordance with the [salt] concentration of the water. In this respect the genus Dunaliella differs completely from all species of Chlamydomonas ...]
and:
"Ainsi, si nous plaçons une goutte d'eau salée, contenant des zoospores, sur le porte-objet, on constate, au microscope, qu'elles se présentent sous la forme mentionée plus haut. Mais si nous laissons la goutte s'évaporer un peu, on observe que le corps commence à s'allonger et à se difformer ... ; si alors nous ajoutons à la préparation une goutte d'eau douce, les zoospores s'arrondissent brusquement .... Cette expérience, que j'ai répétée un trés grand nombre de fois, m'a toujours donné les mêmes résultats."
[Thus, when we place a drop of salt water that contains zoospores [= motile vegetative cells] on a microscope slide, one detects in the microscope that these present themselves in the above-described form. However, when we let the drop evaporate a little, one observes that the body starts to elongate and to lose its shape. ... ; when we then add to the preparation a drop of fresh water, the zoospores suddenly round up. .... This experiment, which I have repeated a great number of times, has always given me the same results.]
The phenomena described above are illustrated in Fig. 2, drawings 9–29 and 30–31, respectively. Teodoresco further writes:
"Si à une goutte d'eau salée on ajoute une goutte plus grande d'eau douce, ce qui amène une abaissement brusque de la concentration, les zoospores non seulement s'arrondissent, mais encore cessent leurs mouvements; le volume du corps augmente et devient parfois deux fois plus grand et à la fin la zoospore éclate. La cause de cet éclatement n'est pas difficile à comprendre: c'est l'action méchanique de la pression osmotique trop élevée par rapport à la densité diminuée du milieu ambiant."
[If to a drop of salt water one adds a larger drop of fresh water, which leads to a sudden drop in concentration, the zoospores not only round up, but in addition cease their movements; the volume of the body increases and sometimes becomes twice as large, and finally the zoospore bursts. The cause of this burst is not difficult to understand: it is the mechanical action of the too high osmotic pressure in comparison to the decreased density of the ambient medium.]
Lerche [30] likewise observed the osmotic changes that occur when the salt concentration is changed. She noted that when a drop of D. salina cells suspended in 20% salt is flooded with distilled water, a large fraction of the cells burst, but some cells survived the treatment.
10. Intracellular Salt and Solute Concentrations of Dunaliella
Marrè and Servetta ([63], as cited in [61]) described measurements of the freezing point of the cytoplasmic fluid of D. salina to obtain information on the intracellular salt concentration. The results indicated an apparent "salt" concentration that exceeded the 3.9 M salt in which the cells were grown. At the time it was postulated that NaCl is taken up through the allegedly very permeable cell membrane during salt upshock, followed by free water flux to equalize intracellular and extracellular osmotic pressures [63-65].
That the salt concentrations within Dunaliella cells cannot be that high, was convincingly shown by the enzymological studies by Johnson et al. (1968), who demonstrated that some of the key enzymes of the algal metabolism such as pentose phosphate isomerase, ribulose bisphosphate carboxylase, glucose-6-phosphate dehydrogenase and phosphohexose isomerase, are strongly inhibited by NaCl. We now know that the intracellular ionic concentrations of Dunaliella are very low indeed. Using lithium ions as a marker for the extracellular water space to estimate the intracellular volume, the intracellular Na+ concentrations, both in cells grown in 0.5 M and in 4 M NaCl, was found not to exceed 100 mM [66]. Such low intracellular Na+ levels are achieved by the activity of a Na+/H+ antiporter in the cytoplasmic membrane [67], as well as by direct electron transport-coupled Na+ extrusion [68].
The enigma of the apparent incompatibility between the low intracellular ionic concentrations and the need for osmotic equilibrium of the cells' contents with the external medium was solved with the discovery that the cells accumulate photosynthetically produced glycerol as osmotic, "compatible" solute. It is interesting to note that the first experiments in which the effects of glycerol on Dunaliella were tested had already been performed by Teodoresco [20], almost hundred years ago. He examined the effect of glycerol and other non-ionic compounds that normally cause plasmolysis. He observed that D. salina cells temporarily lose their motility when suspended in 50% glycerol, but that motility is rapidly restored when the glycerol concentration is then slightly lowered in a humid environment. With 75% glycerol results were largely similar, except that a large fraction of cells died, and in 100% glycerol only few cells survived.
The first indications that glycerol is accumulated by Dunaliella to provide osmotic balance can be found in a short paper published in 1964 by Craigie and McLachlan [69]. They incubated D. tertiolecta with 14CO2, then extracted the cells with ethanol, separated the neutral fraction containing soluble carbohydrates and related compounds using ion exchange procedures, and characterized the compounds by two-dimensional paper chromatography and autoradiography. When the salinity of the medium was increased 100-fold from 0.025 to 2.5 M, 94-fold more radioactivity was found in the neutral fraction. Glycerol amounted to 56, 76, and 81% of the radioactivity of the neutral fraction extracted from cells incubated in 0.025, 0.5, and 2.5 M NaCl, respectively, most of the remainder probably consisting of soluble polysaccharides. In a subsequent study, Wegmann [70] confirmed that the proportion of the radiolabel from 14C-bicarbonate that ends up as glycerol increases with increasing salt concentration up to 2.8 M. He postulated that "The glycerol formation is considered to be a protective mechanism for the survival of Dunaliella in its natural habitat".
The role that glycerol plays in the salt adaptation of Dunaliella was firmly established by the studies of Ben-Amotz and Avron [71] and Borowitzka and Brown [72]. The concept of "compatible solutes", a term coined by Duncan Brown to indicate solutes that not only contribute to the osmotic status of the cell but also maintain enzyme activity under conditions of low water activity, was largely based on the study of the function of glycerol in Dunaliella.
Intracellular glycerol concentrations in Dunaliella can be very high indeed: cells grown in 4 M NaCl were reported to contain approximately 7.8 M glycerol inside, equivalent to a solution of 718 g l-1 glycerol in water [73]. Maintenance of such a high concentration requires special properties of the cell membrane in view of the fact that most biological membranes are relatively permeable to glycerol. It has been established that Dunaliella possesses a membrane with an unusually low permeability for glycerol [74,75], and this enables the cells to keep the glycerol inside the cell. The causes of the low glycerol permeability of the Dunaliella membrane are still not fully understood.
Attempts have been made to exploit the high concentrations of glycerol accumulated by Dunaliella as the basis for the commercial production of this compound. Although technically the production of glycerol from Dunaliella was shown to be possible [51,52,76] economic feasibility is low, and to my knowledge no biotechnological operation presently exists that exploits the alga for glycerol production.
11. Proteomics Approaches to the Understanding of Salt Tolerance in Dunaliella
A versatile organism such as Dunaliella that can adapt to a wide variety of salt concentrations can be used as a convenient model to study the formation of specific proteins as a function of changes in medium salinity. Such proteomic approaches have led to some interesting observations in recent years.
A number of such studies were directed to the detection of changes in the protein content of the cytoplasmic membrane, whose outer side is exposed to the medium salinity, when the cells are shifted from low to high salinity. Two membrane proteins were strongly induced by salt upshock, one with an apparent molecular mass of 60 kDa [77] and one of 150 kDa [78]. These proteins have been purified and characterized. The 60 kDa protein is a carbonic anhydrase that apparently helps the cell to take up carbon dioxide in concentrated brines in which the solubility of gases is decreased [79]. The 150 kDa protein is an unusual transferrin-like protein, involved in the transport of iron into the cell [80].
With a study published in 2004 by Liska et al. [81], Dunaliella research has entered the era of modern proteomics. Comparison of protein patterns of low-and of high-salt-grown cells were compared on two-dimensional gels led to the identification of 76 salt-induced proteins. Among the proteins up-regulated following salinity stress were key enzymes in the Calvin cycle, enzymes involved in starch mobilization and in redox energy production, regulatory factors in protein biosynthesis and degradation, and a homolog of bacterial Na+-redox transporters. The results indicate that Dunaliella responds to transfer to a high salinity by enhancement of photosynthetic CO2 assimilation and by diversion of carbon and energy resources for synthesis of glycerol. This beautiful study is a worthy conclusion of the first century of Dunaliella research, and provides us with a preview of the kind of information that may be expected to be obtained in the coming years, using approaches of genomics, proteomics and systems biology.
==== Refs
Dunal F Extrait d'un mémoire sur les algues qui colorent en rouge certains eaux des marais salants méditerranéens Ann Sc Nat Bot 2 Sér 1838 9 172
Teodoresco EC Organisation et développement du Dunaliella, nouveau genre de Volvocacée-Polyblepharidée Beih z Bot Centralbl 1905 Bd. XVIII 215 232
Ginzburg M Dunaliella : a green alga adapted to salt Adv Bot Res 1987 14 93 183
Avron M Ben-Amotz A Ed Dunaliella: Physiology, Biochemistry, and Biotechnology 1992 Boca Raton: CRC Press
Oren A Halophilic Microorganisms and their Environments 2002 Dordrecht: Kluwer Academic Publishers
Turpin PJF Quelques observations nouvelles sur les Protococcus, qui colorent en rouge les eaux des marais salants Comp Rend Acad Sci 1839
Hamburger C Zur Kenntnis der Dunaliella salina und einer Amöbe aus Salinenwasser von Cagliari Arch f Protistenkd 1905 6 111 131
Joly N Histoire d'un petit crustacé (Artemia salina) auquel on a faussement attribué la coloration en rouge des marais salants méditerranéens, suivie de recherches sur la cause réelle de cette coloration Ann Sc Nat Zool Ser 2 1840 13 225
Oren A Stambler N Dubinsky Z On the red coloration of saltern crystallizer ponds Int J Salt Lake Res 1992 1 2 77 89
Oren A Dubinsky Z On the red coloration of saltern crystallizer ponds. II. Additional evidence for the contribution of halobacterial pigments Int J Salt Lake Res 1994 3 9 13 10.1007/BF01990638
Geleznow N Über die Ursache der Färbung des Salzwassers im See Sak in der Krim Bull de l'Académie Imp des Sciences de St Petersburg 1872 17 557
Butschinsky P Die Protozoenfauna der Salzseelimane bei Odessa Zool Anz 1897 20
Blanchard R Résultats d'une excursion zoologique en Algérie Mém de la Soc Zool de France 1891 IV
Florentin R Faune des Mares Salées de Lorraine 1899 Nancy
Bujor P Contributions à la faune des lacs salées de Roumanie Ann Sci de l'Université de Jassy 1900 1
Blanchard R Note préliminaire sur Monas Dunalii, flagellé, qui cause la rubéfaction des marais salants Bull de la Soc Zool de France 1888 13 153
Dujardin F Histoire Naturelle des Zoophytes (Infusoires) 1841 Paris
Cohn F Chlamydomonas marina Cohn Hedwigia 1865 4 97
Hansgirg A Prodromus d. Algenflora von Böhmen 1866 1 106
Teodoresco EC Observations morphologiques et biologiques sur le genre Dunaliella Rev Gén Bot 1906 18 353 371 18: 409–427
Cavara F Alcune osservazione sulla Dunaliella salina (Dun.) Teod. delle saline di Cagliari Rend della R Acad delle Sc Fis e Matemat. di Napoli Ser. 3 1906 12
Peirce GJ The behavior of certain micro-organisms in brine. The Salton Sea Carnegie Inst Washington Publ 1914 193 49 70
Labbé A Sur les modifications adaptives de Dunaliella salina Dunal C R Acad Sci 1921 172 1074 1076
Labbé A Le cycle évolutif de Dunaliella salina C R Acad Sci 1921 172 1689 1690
Labbé A Les variations de la concentration en ions hydrogène dans les marais salants, comme facteur biologique C R Acad Sci 1922 175 843 845
Baas Becking LGM On organisms living in concentrated brine Tijdschr Ned Dierk Ver Ser III 1928 1 6 9
Baas-Becking LGM Salt effects on swarmers of Dunaliella viridis Teod J Gen Physiol 1931 14 765 779 10.1085/jgp.14.6.765 19872621
Nicolai FE Baas Becking LGM Einige Notizen über Salzflagellaten Arch f Protistenkd 1935 85 319 328
Hamel G Chlorophycées des côtes françaises Extr Rev Algol 1931 I-V 4 6
Lerche W Untersuchungen über Entwicklung und Fortpflanzung in der Gattung Dunaliella Arch f Protistenkd 1937 88 236 268
Butcher RW An Introductory Account of the Smaller Algae of British Coastal Waters. Part I: Introduction and Chlorophyceae 1959 London: Ministry for Agriculture, Fisheries and Food, Fishery Investigations, Series IV. Her Majesty's Stationery Office
Massyuk NP Morphology, Taxonomy, Ecology and Geographic Distribution of the Genus Dunaliella Teod. and Prospects for its Potential Utilization 1973 Kiev: Naukova Dumka (in Russian)
Preisig HR Ben-Amotz A, Avron M Morphology and taxonomy Dunaliella: Physiology, Biochemistry and Biotechnology 1992 Boca Raton: CRC Press 1 15
Kalina T Zur Morphologie und Taxonomie der Gattung Spermatozopsis Korschikow (Volvocales). Spermatozopsis acidophila sp. nov Preslia 1965 37 9 12
Albertano P Pinto G Santisi S Taddei R Spermatozopsis acidophila Kalina (Chlorophyta, Volvocales), a little known alga from highly acidic environments Giorn Bot Ital 1981 115 65 76
Gimmler H Weis U Ben-Amotz A, Avron M Dunaliella acidophila – life at pH 1.0 Dunaliella: Physiology, Biochemistry and Biotechnology 1992 Boca Raton: CRC Press 99 133
González MA Gómez PI Montoya R Comparison of PCR-RFLP analysis of the ITS region with morphological criteria of various strains of Dunaliella J Appl Phycol 1999 10 573 580 10.1023/A:1008035422784
González MA Coleman AW Gómez PI Montoya R Phylogenetic relationship among various strains of Dunaliella (Chlorophyceae) based on nuclear ITS rDNA sequences J Phycol 2001 37 604 611 10.1046/j.1529-8817.2001.037004604.x
Olmos J Paniagua J Contreras R Molecular identification of Dunaliella sp. utilizing the 18S rDNA gene Lett Appl Microbiol 2000 30 80 84 10.1046/j.1472-765x.2000.00672.x 10728567
Oren A Gurevich P Anati DA Barkan E Luz B A bloom of Dunaliella parva in the Dead Sea in 1992: biological and biogeochemical aspects Hydrobiologia 1995 297 173 185 10.1007/BF00019283
Loeblich LA Aplanospores of Dunaliella salina (Chlorophyta) J Protozool 1969 Suppl 16 22 23
Borowitzka LJ The microflora. Adaptations to life in extremely saline lakes Hydrobiologia 1981 81 33 46 10.1007/BF00048704
Brock TD Salinity and the ecology of Dunaliella from Great Salt Lake J Gen Microbiol 1975 89 285 292
Ruinen J Notizen über Salzflagellaten. II. Über die Verbreitung der Salzflagellaten Arch Protistenkd 1938 90 210 258
Ben-Amotz A Katz A Avron M Accumulation of β-carotene in halotolerant algae: purification and characterization of β-carotene-rich globules from Dunaliella bardawil (Chlorophyceae) J Phycol 1982 18 529 537 10.1111/j.0022-3646.1982.00529.x
Ben-Amotz A Lers A Avron M Stereoisomers of β-carotene and phytoene in the alga Dunaliella bardawil Plant Physiol 1988 86 1286 1291 16666068
Borowitzka MA Micro-algae as sources of fine chemicals Microbiol Sci 1986 3 372 375 3153571
Aasen AJ Eimhjellen KE Liaaen-Jensen S An extreme source of β-carotene Acta Chem Scand 1969 23 2544 2545 5368546
Massyuk NP Mass culture of the carotene bearing alga Dunaliella salina Ukr Bot Zh 1968 23 12 19
Drokova IH The alga Dunaliella salina as a source of β-carotene Ukr Bot Zh 1961 18 110
Ben-Amotz A Shelef G, Soeder CJ Glycerol, β-carotene and dry algal meal production by commercial cultivation of Dunaliella Algae Biomass 1980 Amsterdam: Elsevier 603 610
Ben-Amotz A Avron M Accumulation of metabolites by halotolerant algae and its industrial potential Ann Rev Microbiol 1983 37 95 119 10.1146/annurev.mi.37.100183.000523 6416147
Borowitzka LJ Borowitzka MA Moulton TP The mass culture of Dunaliella for fine chemicals: from laboratory to pilot plant Hydrobiologia 1984 116/117 115 121 10.1007/BF00027649
Ben-Amotz A Avron M Cresswell RC, Rees TAV, Shah, N The biotechnology of mass culturing Dunaliella for products of commercial interest Algal and Cyanobacterial Biotechnology 1989 Harlow: Longman Scientific and Technical Press 91 114
Stephens DW Gillespie DM Phytoplankton production in the Great Salt Lake, Utah, and a laboratory study of algal response to enrichment Limnol Oceanogr 1976 21 74 87
Post FJ The microbial ecology of the Great Salt Lake Microb Ecol 1977 3 143 165 24233467
Volcani BE The microorganisms of the Dead Sea Papers Collected to Commemorate the 70th Anniversary of Dr. Chaim Weizmann. Collective volume 1944 Rehovoth: Daniel Sieff Research Institute 71 85
Kaplan IR Friedmann A Biological productivity in the Dead Sea. Part I. Microorganisms in the water column Israel J Chem 1970 8 513 528
Oren A Shilo M Population dynamics of Dunaliella parva in the Dead Sea Limnol Oceanogr 1982 27 201 211
Gibor A The culture of brine algae Biol Bull 1956 111 223 229
Johnson MK Johnson EJ MacElroy RD Speer HL Bruff BS Effects of salts on the halophilic alga Dunaliella viridis J Bacteriol 1968 95 1461 1468 5646631
Van Auken OW McNulty IB The effect of environmental factors on the growth of a halophylic species of algae Biol Bull 1973 145 210 222 4732535
Marrè E Servettaz A Sul meccanismo di adattemento a condizioni osmotiche estreme in Dunaliella salina. II. Rapporto fra concentrazioni del mezzo esterno e composizione del succo cellulare Atti Accad Naz Lincei Cl Sci Fis Mat Nat Rend Ser 8 1959 26 272 278
Trezzi M Galli MG Bellini E L'osomo-resistenza di Dunaliella salina. Ricerche ultrastructurali Giorn Bot Ital 1965 72 255 263
Ginzburg M The unusual membrane permeability of two halophilic unicellular organisms Biochim Biophys Acta 1969 173 370 376 5769637
Katz A Avron M Determination of intracellular osmotic volume and sodium concentration in Dunaliella Plant Physiol 1985 78 817 820 16664332
Katz A Pick U Avron M Characterization and reconstitution of the Na+/H+ antiporter from the plasma membrane of the halophilic alga Dunaliella Biochim Biophys Acta 1989 983 1 4 2758043
Katz A Pick U Plasma membrane electron transport coupled to Na+ extrusion in the halotolerant alga Dunaliella Biochim Biophys Acta 2001 1504 423 431 11245805
Craigie JS McLachlan J Glycerol as a photosynthetic product in Dunaliella tertiolecta Butcher Can J Bot 1964 42 777 778
Wegmann K Osmotic regulation of photosynthetic glycerol production in Dunaliella Biochim Biophys Acta 1971 234 317 323 5117572
Ben-Amotz A Avron M The role of glycerol in the osmotic regulation of the halophilic alga Dunaliella parva Plant Physiol 1973 51 875 878 16658431
Borowitzka LJ Brown AD The salt relations of marine and halophilic species of the unicellular green alga, Dunaliella. The role of glycerol as a compatible solute Arch Microbiol 1974 96 37 52 10.1007/BF00590161
Brown AD Microbial Water Stress Physiology. Principles and Perspectives 1990 Chichester: John Wiley & Sons
Brown FF Sussman I Avron M Degani H NMR studies of glycerol permeability in lipid vesicles, erythrocytes and the alga Dunaliella Biochim Biophys Acta 1982 690 165 173 7126572
Gimmler H Hartung W Low permeability of the plasma membrane of Dunaliella parva for solutes J Plant Physiol 1988 133 165 172
Chen BJ Chi CH Process development and evaluation for algal glycerol production Biotechnol Bioengin 1981 23 1267 1287 10.1002/bit.260230608
Sadka A Himmelhoch S Zamir A A 150 kilodalton cell surface protein is induced by salt in the halotolerant green alga Dunaliella salina Plant Physiol 1991 95 822 831 16668060
Fisher M Pick U Zamir A A salt-induced 60-kilodalton plasma membrane protein plays a potential role in the extreme halotolerance of the alga Dunaliella Plant Physiol 1994 106 1359 1365 12232413
Fisher M Gokhman I Pick U Zamir A A salt-resistant plasma membrane carbonic anhydrase is induced by salt in Dunaliella salina J Biol Chem 1996 271 17718 17723 10.1074/jbc.271.30.17718 8663366
Fisher M Gokhman I Pick U Zamir A A structurally novel transferrin-like protein accumulates in the plasma membrane of the unicellular green alga Dunaliella salina grown in high salinities J Biol Chem 1997 272 1565 1570 10.1074/jbc.272.3.1565 8999829
Liska AJ Shevhenko A Pick U Katz A Enhanced photosynthesis and redox energy production contribute to salinity tolerance in Dunaliella as revealed by homology-based proteomics Plant Physiol 2004 136 2806 2817 10.1104/pp.104.039438 15333751
|
16176593
|
PMC1224875
|
CC BY
|
2022-09-08 23:19:43
|
no
|
Saline Syst. 2005 Jul 4; 1:2
|
utf-8
|
Saline Syst
| 2,005 |
10.1186/1746-1448-1-2
|
oa_comm
|
==== Front
Saline Syst
Saline Syst
Saline Systems
1746-1448
BioMed Central
1746-1448-1-3
16176594
10.1186/1746-1448-1-3
Research
UV irradiation induces homologous recombination genes in the model archaeon, Halobacterium sp. NRC-1
McCready Shirley [email protected]
Müller Jochen A [email protected]
Boubriak Ivan [email protected]
Berquist Brian R [email protected]
Ng Wooi Loon [email protected]
DasSarma Shiladitya [email protected]
1 School of Biological Molecular Sciences, Oxford Brookes University, Oxford OX3 0BP, UK
2 Center of Marine Biotechnology, University of Maryland Biotechnology Institute, 701 E. Pratt St., Suite 236, Baltimore, MD 21202 USA
3 Molecular and Structural Biology Program, Greenebaum Cancer Center, University of Maryland, Baltimore, MD 21201, USA
2005
4 7 2005
1 33
8 4 2005
4 7 2005
Copyright © 2005 McCready et al; licensee BioMed Central Ltd.
2005
McCready et al; licensee BioMed Central Ltd.
https://creativecommons.org/licenses/by/2.0/ This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
A variety of strategies for survival of UV irradiation are used by cells, ranging from repair of UV-damaged DNA, cell cycle arrest, tolerance of unrepaired UV photoproducts, and shielding from UV light. Some of these responses involve UV-inducible genes, including the SOS response in bacteria and an array of genes in eukaryotes. To address the mechanisms used in the third branch of life, we have studied the model archaeon, Halobacterium sp. strain NRC-1, which tolerates high levels of solar radiation in its natural hypersaline environment.
Results
Cells were irradiated with 30–70 J/m2 UV-C and an immunoassay showed that the resulting DNA damage was largely repaired within 3 hours in the dark. Under such conditions, transcriptional profiling showed the most strongly up-regulated gene was radA1, the archaeal homolog of rad51/recA, which was induced 7-fold. Additional genes involved in homologous recombination, such as arj1 (recJ-like exonuclease), dbp (eukaryote-like DNA binding protein of the superfamily I DNA and RNA helicases), and rfa3 (replication protein A complex), as well as nrdJ, encoding for cobalamin-dependent ribonucleotide reductase involved in DNA metabolism, were also significantly induced in one or more of our experimental conditions. Neither prokaryotic nor eukaryotic excision repair gene homologs were induced and there was no evidence of an SOS-like response.
Conclusion
These results show that homologous recombination plays an important role in the cellular response of Halobacterium sp. NRC-1 to UV damage. Homologous recombination may permit rescue of stalled replication forks, and/or facilitate recombinational repair. In either case, this provides a mechanism for the observed high-frequency recombination among natural populations of halophilic archaea.
==== Body
pmcBackground
In all organisms studied to date, UV irradiation causes inducible responses. In some bacteria the inducible SOS response involves about 40 genes that are up-regulated dependent on the pleiotropic regulator, LexA [1]. In eukaryotes a variety of genes are up- and down-regulated in response to UV-damage including several DNA repair genes, although no eukaryotic equivalent of the bacterial SOS response has been identified [2]. Among archaea, most studies have heretofore been limited to comparative genomic approaches where inducible mechanisms have not been discernable [3,4]. The hyperthermophilic archaea have been found to carry homologs of several eukaryotic nucleotide excision repair (NER) genes, including rad2 (FEN1/XPG), rad3 (XPD), eif4A (rad1/XPF), and rad25 (XPB) [5,6]. Remarkably, extremely halophilic archaea, such as the model organism Halobacterium sp. strain NRC-1, and some non-thermophilic methanogenic archaea, were found to harbor homologs of both eukaryotic NER genes and bacterial NER genes, uvrA, uvrB, uvrC and uvrD [6-8]. The classic SOS response system regulated by LexA is lacking in archaea, however, and the relationship between the bacterial and eukaryotic repair systems in archaea is not currently known [9].
Halophilic archaea, such as Halobacterium spp., are excellent experimental systems for studies of DNA repair because they are amongst the few archaeal microorganisms to encounter high levels of sunlight in their natural environment. They occupy an extreme environmental niche, where exposure to intense solar radiation leads to evaporation and concentration of NaCl to near- or even super-saturation. These microorganisms, including the sequenced wild-type model, Halobacterium sp. NRC-1, are highly resistant to the damaging effects of UV radiation in sunlight, principally due to extremely efficient photoreactivation of DNA damage [10,11]. In the presence of visible light they can survive UV doses many times higher than they would ever be exposed to naturally. Cell survival approaches 100 % after doses up to 100 J/m2 while 1 hour exposure to sunlight inflicts damage equivalent to, at most, only a few J/m2 [[12]; SM, unpublished]. UV tolerance of Halobacterium sp. NRC-1 is compared to other key organisms in Figure 1. Halobacterium is significantly more UV-tolerant, even without photoreactivating light, than Escherichia coli or Saccharomyces cerevisiae, though not as resistant as the extremely radiation-resistant Deinococcus radiodurans. Halophilic archaea also have excision repair mechanisms that can operate in the absence of photoreactivating light [12,13]. In addition, like bacteria and eukaryotes, they are likely to possess mechanisms that enable them to tolerate the presence of some unrepaired UV lesions, including lesion bypass by DNA polymerases that can circumvent photoproducts [14]) and recombination to facilitate recovery of stalled replication forks [15,16].
Figure 1 Survival of model organisms exposed to UV-C radiation. The percent survival (y axis logarithmic scale) is plotted versus dose of UV radiation (x axis linear scale) for human fibroblasts [41], Escherichia coli [42], Saccharomyces cerevisiae [43], Halobacterium sp. NRC-1 (in the dark or in presence of visible light) [11], and Deinococcus radiodurans [44].
We have taken a multifaceted approach to the study of UV responses in Halobacterium sp. NRC-1. Previously, the critical role for phr2 in light repair was demonstrated through a combination of genetics and biochemistry [11]. Here, we use DNA microarrays to show the importance of homologous recombination genes in the response of cells to UV damage. Interestingly, our results are distinct from those obtained in a previous study on Halobacterium sp. NRC-1 using significantly higher doses of UV irradiation [17].
Results and Discussion
In order to understand gene expression responses to UV at the whole genome level, we studied the model archaeon, Halobacterium sp. strain NRC-1, employing DNA microarrays. We used three different doses of UV-C, 30 J/m2, 50 J/m2, and 70 J/m2, with post-irradiation incubation times in the dark of 1 hour and 3 hours (Materials and Methods). At these UV doses, survival of cells is close to 100 % following DNA damage either in the light or in the dark (Figure 1) [11,18]. Cells were irradiated in growth medium with post-irradiation incubation in the same medium so as to introduce minimal additional stresses. To give an indication of repair rates at the doses used, we determined the relative occurrences of cyclobutane pyrimidine dimers (CPDs) and 6-4 photoproducts over time after irradiation with an intermediate dose of 50 J/m2 (Figure 2). The majority of photoproducts were repaired during the 3-hour period, though some cyclobutane dimers still remained at 3 hours.
Figure 2 Repair of the two principle UV-induced photoproducts in Halobacterium sp. Repair of cyclobutane pyrimidine dimers (CPDs) and 6-4 photoproducts (6-4 pp) was measured after a UV-C dose of 50 J/m2. About 55 % of CPDs and 25 % of 6-4 pp remain unrepaired after 1 hour; after 3 hours, the percentages are 28% and 2% respectively.
Custom DNA microarrays were fabricated using inkjet technology (Agilent Technologies, Palo Alto, CA) with in-house oligonucleotide design performed with the program OligoPicker [19]. The arrays contained 8,455 60-mer nucleotide features representing 2474 open reading frames (ORFs). The high specificity of 60-mer oligonucoletide arrays have been demonstrated previously [20]. Up to three probes were designed per ORF with a mean Tm of 81°C and a Tm range of 3°C. The microarray slides harbor both gene-probes (~8,000 features per array) as well as ~400 negative and positive control spots to test hybridization conditions and allow for error modeling. These microarrays were thoroughly tested for linearity of response and statistical significance in a related study of anaerobic respiration in Halobacterium sp. NRC-1 [21]. Signal intensities with a dynamic range in excess of three orders of magnitude were found allowing simultaneous analysis of low- and high intensity features.
For microarray analysis, two doses of UV-C, 30 J/m2 and 70 J/m2 were used. Microarray data quality was considered high. After removal of outliers, features replicated within a single array showed low differences in absolute processed signal intensities (7 % on average); and spot-to-spot variation for replicate experiments was 9 %. Of the 2474 open reading frames (ORFs) of Halobacterium sp. NRC-1 represented, 100 were significantly up-regulated and 150 were significantly down-regulated (1.5-fold above or below, respectively, p-value < 0.05) in at least one experimental setting. In the current study, we focused on genes involved in homologous recombination and DNA metabolism, which are the most significantly induced (Table 1). Expression levels for all genes represented on the array for the 30 J/m2 dose are also provided (see Additional file 1). Data obtained from UV irradiation with 70 J/m2 were essentially the same (not shown). At either dose, the pattern of inducible transcripts was very different from what would be seen in E. coli, not surprisingly, since there is no LexA homolog in Halobacterium sp. NRC-1. Moreover, there was no evidence of a classic coordinated SOS response and neither the prokaryotic DNA repair genes, including uvrA, uvrB, uvrC and uvrD nor any of the eukaryotic repair gene homologs were up-regulated.
Table 1 UV irradiation (30 J/m2)-inducible recombination- and DNA metabolism genes in Halobacterium sp. NRC-1.
Functional category Gene number Gene name Fold Induction Predicted function
1 hr 3 hr
recombination VNG2473 radA1 7.4 6.4 Rad51/RecA recombinase
VNG2160 rfa3 1.5a 1.5a RPA41 homolog, contains Zn- finger motif
VNG2162 rfa8 1.5a 1.5a RPA32 homolog
VNG2163 ral 1.5a 1.5a RPA linked ORF
VNG2167 dbp 1.9 1.4 Superfamily I helicase, DNA binding protein eukaryotic-like
VNG0779 arj1 1.4 1.5 RecJ like exonuclease
DNA VNG1644 nrdJ 2.8 3.1 class II ribonucleoside reductase
metabolism VNG1642 Vng1642 3.5 4.3 Unknown, contains Zn- finger motif
a Fold induction after irradiation with 70 J/m2.
Strikingly, however, the radA1 gene showed the most inducible transcript (~7 fold) at 1 hr and 3 hrs after both UV doses (Table 1; Figure 3). RadA1 is the Halobacterium sp. NRC-1 homolog of RecA/Rad51, which catalyses strand invasion and exchange during homologous recombination. RadA1 is more similar to Rad51 in eukaryotes than to bacterial RecA. Deletion of radA has been shown to cause severe UV sensitivity in the related halophililc archaeon, Haloferax volcanii [22]. The expression change of radA1 was considered statistically highly significant based on the following: The radA1 gene is represented by 3 different probes per array with one of those probes being duplicated, combining to a total of 24 data points for that gene in the present report. In addition, cDNA was prepared from three independently treated cultures per condition and array. The average standard deviation of the fold changes for all radA1-probes within an array was 0.85, while the array-to-array difference for identical probes in replicated experiments was 18 %. The average p-value of log ratios was < 10-22. Results from previous transcriptome profiling after environmental perturbations indicate that a 7-fold expression change is high in comparison to any gene in Halobacterium sp. NRC-1 [[21]; JM and SD, unpublished data]. In those experiments (5 different conditions not involving UV irradiation) the radA1 gene was not differentially expressed demonstrating that the induction presented here was caused by UV irradiation.
Figure 3 Whole genome microarray hybridization results comparing UV-irradiated cells to control cells. Irradiated cells received a UV-C dose of 30 J/m2 and incubated in the dark for 1 h (upper panel) and 3 h (lower panel). Control cells were treated exactly the same except for UV-irradiation. For each ORF represented on the array, the logarithm of the hybridization ratio of UV-irradiated cells (Cy5-labeled cDNA) to control cells (Cy3-labeled cDNA) is displayed in black marks on the y axis. The location of ORFs within the entire 2.6-megabase genome maps on the x axis. Expression ratios of selected genes are indicated.
The radA2 (radB) gene, a second homolog of recA in archaea, encoding a protein with unknown role in homologous recombination, was not up-regulated, in agreement with observations in other archaea [23]. Accompanying radA1 induction, several other genes involved in homologous recombination were significantly induced after UV irradiation. The dbp gene, encoding a eukaryote-like DNA binding protein of the superfamily I DNA and RNA helicases, was upregulated 1.95 (p-value 0.04) at 1 hr post UV-irradiation with 30 J/m2. The arj1 gene, encoding a recJ-like exonuclease was up-regulated 1.5 fold (p-value 0.008) at 3 hrs (30 J/m2). Additionally, an apparent operon encoding RPA ssDNA binding protein complex, rfa3, (VNG2160, RPA41 homolog, Zn-finger containing), rfa8 (VNG2162, RPA32 homolog), and an uncharacterized linked ORF (VNG2163) was induced 1.53 ± 0.02 fold (p-values < 0.001) at both time points after irradiation with 70 J/m2. Similar fold changes were measured after irradiation with 30 J/m2, however, p-values were around 0.1. In eukaryotes, RPA-ssDNA complexes are formed during almost all DNA-damage repair pathways. For example, RPA-proteins are recruited to Rad51 foci, protein complexes that accumulate at sites of DNA damage and stalled replication forks [24,25]. None of the other four RPA homologs of Halobacterium sp. NRC-1 (VNG0133, 1253, 1255, 6403) was significantly differentially expressed. As for radA1, neither of the above genes were found to be differentially expressed in other environmental perturbation experiments [[21]; JM and SD, unpublished data].
Our microarray data, together with these findings, strongly suggest a key role for homologous recombination in survival of UV damage in this class of archaea. By analogy with other organisms, there are at least two ways that recombination might contribute to survival of UV damage in Halobacterium sp. NRC-1. (i) Recombinational rescue of stalled replication forks may take place [15,26]. Rescue of stalled forks is likely to be error-free and this would fit with the low mutation rate in Halobacterium spp. [[1]27; SM, unpublished data]. (ii) Recombinational repair may occur by using duplicate copies of the genome, a mechanism that has been suggested to facilitate post-irradiation survival of D. radiodurans [28]. Halobacterium sp. NRC-1 is thought to contain multiple copies of its genome (J. Soppa pers. com.), which may greatly enhance recombinational repair.
In addition to genes involved in homologous recombination, several other genes were also highly induced. Most interestingly, the gene encoding a cobalamin-dependent class II RNR (nrdJ, formerly nrdB2) was strongly up-regulated (Figure 3, Table 1). A small open reading frame (VNG1642) located immediately downstream to nrdJ encoding a small uncharacterized zinc-finger containing protein (COG1645) was also strongly induced. RNR is the rate-limiting enzyme in the de novo synthesis of deoxyribonucleotide triphosphates which are utilized in both DNA synthesis and DNA repair; the up-regulation of the nrd genes may reflect the need for increased deoxynucleotide concentrations to allow rapid and accurate excision repair. In yeast, DNA damage elicits an increase in dNTP levels; and increased dNTP pools improve cell survival post DNA damage [29]. In many other organisms, RNR genes are up-regulated by UV and were among the very first UV-inducible genes identified in an early study of UV-inducible promoters in budding yeast [30]. RNR is regulated during the cell cycle in eukaryotes and the UV response depends on binding of transcription factor E2F to sites in the promoter [31]. It is noteworthy that, in E. coli, nrd genes are UV-inducible even in the absence of lexA [1].
In addition to nrdJ, arcBCA, required for fermentation of arginine [32] were also particularly strongly induced, though only at the early time, and with some variation in magnitude of induction between repeat experiments. Whether this induction reflects a demand for rapid supply of ATP during periods of DNA-damage repair is not known. Another possibility is that the expression of these genes is exquisitely responsive to small stresses and we view these results with some caution at this stage.
A previous transcriptome analysis of UV-irradiated Halobacterium sp. strain NRC-1 cultures [17] showed, as we do, the lack of an SOS-like response and the lack of up-regulation of any of the NER genes by UV. However, these authors did not report strong induction of recombination genes or of RNR genes. The UV dose used in the latter study was substantially higher (200 J/m2). This dose, which induces, approximately one photoproduct per 600 bp of DNA, causes about one hundred times as much DNA damage as induced by natural sunlight and resulted in compromised cell survival. It may be that the high UV dose caused interference with gene expression. This is in contrast to approximately one photoproduct per 4 kb at 30 J/m2 and one per 1.7 kb at 70 J/m2 [33-35], doses tolerated by Halobacterium sp. NRC-1 and used in the current study.
There have been questions raised as to the significance of gene expression responses to UV and other DNA damaging agents. In the yeast, S. cerevisiae, an analysis of deletion mutants lacking UV-inducible genes suggested that most do not contribute to survival of UV irradiation [36]. In addition, most yeast genes identified as having a role in surviving UV damage, by isolation of UV-sensitive mutants, including most NER genes, are not UV-inducible. In human cells too, most repair genes are not UV-inducible [37]. Thus, transcriptome analysis must be interpreted with caution and inferences about the extent of gene involvement should be supported by physiological and functional studies. In the present report, the absolute contribution of the different DNA-damage repair systems has not been investigated. However, the high level of up-regulation of radA1 and induction of other recombination genes while comparable few other gene expression changes occur, clearly suggests a significant role of homologous recombination in DNA-damage repair in Halobacterium sp. NRC-1.
Conclusion
Our transcriptome profiling work, together with our studies of the physiological response of a model archaeon, has shown that genes involved in homologous recombination are induced by UV irradiation in relatively low doses. Our results are consistent with homologous recombination playing an important role in the cellular response to UV damage in Halobacterium sp. NRC-1, either to permit rescue of stalled replication forks or to facilitate recombinational repair. In either case, we find that induction of recombination genes is prominent in the response of Halobacterium sp. NRC-1 to UV irradiation, which is particularly significant as it has been shown recently that recombination is important in facilitating genetic exchange in wild populations of halophilic archaea [38]. Our results suggest that homologous recombination is stimulated by sunlight in this model archaeon.
Methods
Culture conditions and UV-irradiation
Halobacterium sp. strain NRC- 1 was grown at 37°C under aerobic conditions to early exponential growth phase (OD600 0.19 – 0.23) in complete medium [39] and irradiated in the dark, in medium, using a UV-C source with dose rate of 1 J/sec/m2. Post-UV incubation was continued at the same temperature, without changing the medium. Typically, UV-irradiation is not carried out in growth media because of possible absorption of UV wavelengths. For the present microarray experiments, irradiation in growth medium was of particular importance to avoid additional stress caused to the cells by harvesting and changing media. Moreover, we ascertained that the absorption of UV-C by the growth medium used was minimal by measuring transmission of 260 nm light in a spectrophotometer. The effective UV dose was not significantly affected by irradiating in medium.
Measurement of photoproducts
A dot-blot immunoassay for differential quantitation of CPDs and 6-4 photoproducts was carried out as described in detail previously [40]. Briefly, exponential phase cells were harvested and irradiated in sterile salts solution with UV-C at a dose rate of 1 J/sec m2. Aliquots of yeast extract and casein hydrolysate solutions were added to restore nutrients and the irradiated cells were incubated aerobically at 37°C to allow repair to proceed. All irradiation and post-UV incubation was carried out either under yellow light illumination or in the dark. Samples were taken at timed intervals and DNA extracted for measurement of photoproducts. DNA concentrations in different samples were carefully equalised. Subsequently, each DNA sample was divided into two and one half was treated with hot NaOH to destroy 6-4 photoproducts. Two identical dot blots were prepared on nitrocellulose filters, each containing a set of dilutions of each DNA sample. One filter was exposed to a CPD photolyase to destroy cyclobutane dimers in all the DNA samples on that blot to allow measurement of 6-4 photoproducts alone. The dot blots were then exposed to rabbit polyclonal antiserum containing antibodies to 6-4 photoproducts and CPDs, then to biotinylated anti-rabbit antibody followed colorimetric quantitation using alkaline photsphatase-conjugated Extravidin (Sigma), Nitro Blue tetrazolium (NBT) and 5-bromo-4-chloro-indolyl phosphate (BCIP) substrate. The amount of color was measured using a scanning densitometer (BioRad GS-670) and compared to a set of standards included on each blot.
Microarray procedures
Relative mRNA levels were determined by parallel two-color hybridization to oligonucleotide (60-mer) microarrays representing 2474 ORFs representing 92 % of Halobacterium sp. NRC-1 ORFs according to Müller and DasSarma [21]. Transcriptome profiling of cells irradiated with 30 J/m2 was carried out in duplicate. Transcriptome analysis of cells irradiated with 70 J/m2 was carried out only once since results were essentially the same as for 30 J/m2. Total RNA was isolated from 50 ml cultures immediately after harvesting at 2°C using Agilent Total RNA isolation kit (Agilent) and DNA was hydrolysed using amplification grade DNase (Sigma, UK). In order to minimize biological noise, RNA preparations from three cultures grown under identical conditions were pooled to equal parts for cDNA synthesis. cDNA was prepared from 7 μg total RNA with Super Script III reverse transcriptase (Invitrogen, UK) and Cy3- or Cy5-dCTP (Amersham Biosciences, UK). Performance of duplicate experiments in which dyes were swapped during synthesis to account for labelling differences was not required. Previous results showed that differences in the relative intensity of the channels could be adjusted for by intensity-dependent LOWESS normalization [[21]; JM and SD, unpublished data]. cDNA preparations were purified after alkaline hydrolysis of RNA on Qiagen mini-elute columns (Qiagen, UK). The labeled cDNA targets were mixed with hybridization buffer (Agilent) and control targets (Agilent), and hybridized to microarray slides, assembled into a hybridization chamber (Agilent), for 17 h at 60°C in the dark. Post hybridization, the slides were washed as described [21] and scanned for the Cy3 and Cy5 fluorescent signals with an Agilent DNA-microarray scanner (Model no. G2565BA). Image processing and statistical analysis were carried out using Agilent Feature Extraction Software Version 7.1 as described previously [21]. Log ratios for each feature were calculated and the significance of the log ratio was assessed by calculating the most conservative log ratio error and significance value (p-value) using a standard error propagation algorithm (Agilent) and a universal error model (Rosetta Biosoftware).
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
SM designed the experiments, analyzed the data, and drafted the manuscript. JM assisted with experimental design, conducted the DNA microarray hybridization, and carried out bioinformatics and data analysis. IB assisted with experimental design and conducted the UV irradiation, RNA preparation, and cDNA labelling. BB assisted with bioinformatics analysis. WN assisted with UV irradiation, RNA preparation, and cDNA labelling. SD designed the experiments, assisted with data interpretation, and prepared the manuscript.
Supplementary Material
Additional File 1
Expression changes of Halobacterium NRC-1 open reading frames after UV irradiation with 30 J/m2 is provided as an additional XLS file.
Click here for file
Acknowledgements
We thank James P. Carney for critical reading of the manuscript. This work was supported by NSF grants MCB-0450695 and MCB-0296017 to SD and by BBSRC project grant P18099 to SM.
==== Refs
Courcelle J Khodursky A Peter B Brown PO Hanawalt PC Comparative gene expression profiles following UV exposure in wild-type and SOS-deficient Escherichia coli Genetics 2001 158 1 41 64 11333217
Cline SD Hanawalt PC Who's on first in the cellular response to DNA damage? Nat Rev Mol Cell Biol 2003 4 5 361 372 10.1038/nrm1101 12728270
Eisen JA Hanawalt PC A phylogenomic study of DNA repair genes, proteins, and processes Mutat Res 1999 435 3 171 213 10606811
DasSarma S Kennedy SP Berquist B Ng WV Baliga NS Spudich JL Krebs MP Eisen JA Johnson CH Hood L Genomic perspective on the photobiology of Halobacterium species NRC-1, a phototrophic, phototactic, and UV-tolerant haloarchaeon Photosyn Res 2001 70 3 17 10.1023/A:1013879706863
Grogan DW The question of DNA repair in hyperthermophilic archaea Trends Microbiol 2000 8 4 180 185 10.1016/S0966-842X(00)01729-7 10754577
Grogan DW Stability and repair of DNA in hyperthermophilic archaea Curr Issues Mol Biol 2004 6 2 137 144 15119824
DasSarma S Genome sequence of an extremely halophilic archaeon Microbial Genomes 2004 383 399
Ng V Genome sequence of Halobacterium species NRC-1 Proc Natl Acad Sci U S A 2000 97 22 12176 12181 10.1073/pnas.190337797 11016950
Walker GC Understanding the complexity of an organism's responses to DNA damage Cold Spring Harb Symp Quant Biol 2000 65 1 10 10.1101/sqb.2000.65.1 12760015
Hescox MA Carlberg DM Photoreactivation in Halobacterium cutirubrum Can J Microbiol 1972 18 7 981 985 5070716
McCready SJ Marcello L Repair of UV damage in Halobacterium salinarum Bioch Soc Trans 2003 31 3 694 698 10.1042/BST0310694
Quaite FE Takayanagi S Ruffini J Sutherland JC Sutherland BM DNA damage levels determine cyclobutyl pyrimidine dimer repair mechanisms in Alfalfa seedlings Plant Cell 1994 6 11 1635 1641 10.1105/tpc.6.11.1635 12244228
McCready SJ The repair of ultraviolet light-induced DNA damage in the halophilic archaebacteria, Halobacterium cutirubrum, Halobacterium halobium and Haloferax volcanii Mutat Res 1996 364 1 25 32 8814335
Boudsocq F Iwai S Hanaoka F Woodgate R Sulfolobus solfataricus P2 DNA polymerase IV (Dpo4): an archaeal DinB-like DNA polymerase with lesion-bypass properties akin to eukaryotic polη Nucleic Acids Res 2001 29 22 4607 4616 10.1093/nar/29.22.4607 11713310
Kuzminov A DNA replication meets genetic exchange: chromosomal damage and its repair by homologous recombination Proc Natl Acad Sci U S A 2001 98 15 8461 8468 10.1073/pnas.151260698 11459990
Courcelle J Donaldson JR Chow KH Courcelle CT DNA damage-induced replication fork regression and processing in Escherichia coli Science 2003 299 5609 1064 1067 10.1126/science.1081328 12543983
Baliga NS Bjork SJ Bonneau R Pan M Iloanusi C Kottemann MC Hood L DiRuggiero J Systems level insights into the stress response to UV radiation in the halophilic archaeon Halobacterium NRC-1 Genome Res 2004 14 6 1025 1035 10.1101/gr.1993504 15140832
Fitt PS Sharma N Castellanos G A comparison of liquid-holding recovery and photoreactivation in halophilic and non-halophilic bacteria Biochim Biophys Acta 1983 739 1 73 78 6338926
Wang X Seed B Selection of oligonucleotide probes for protein coding sequences Bioinformatics 2003 19 7 796 802 10.1093/bioinformatics/btg086 12724288
Hughes TR Expression profiling using microarrays fabricated by an ink-jet oligonucleotide synthesizer Nature Biotechnol 2001 19 4 342 347 10.1038/86730 11283592
Müller JA DasSarma S Genomic analysis of anaerobic respiration in the archaeon Halobacterium species NRC-1: dimethyl sulfoxide and trimethylamine N-oxide as terminal electron acceptors J Bacteriol 2005 187 5 1659 1667 10.1128/JB.187.5.1659-1667.2005 15716436
Woods WG Dyall-Smith ML Construction and analysis of a recombination-deficient (radA) mutant of Haloferax volcanii Mol Microbiol 1997 23 4 791 797 10.1046/j.1365-2958.1997.2651626.x 9157249
Reich CI McNeil LK Brace JL Brucker JK Olsen GJ Archaeal RecA homologues: different response to DNA-damaging agents in mesophilic and thermophilic archaea Extremophiles 2001 5 4 265 275 10.1007/s007920100197 11523896
Golub EI Gupta RC Haaf T Wold MS Radding CM Interaction of human rad51 recombination protein with single-stranded DNA binding protein, RPA Nucleic Acids Res 1998 26 23 5388 5393 10.1093/nar/26.23.5388 9826763
Tashiro S Walter J Shinohara A Kamada N Cremer T Rad51 accumulation at sites of DNA damage and in postreplicative chromatin J Cell Biol 2000 150 2 283 291 10.1083/jcb.150.2.283 10908572
Cox MM Historical overview: searching for replication help in all of the rec places Proc Natl Acad Sci U S A 2001 98 15 8173 8180 10.1073/pnas.131004998 11459950
Martin EL Reinhardt RL Baum LL Becker MR Shaffer JJ Kokjohn TA The effects of ultraviolet radiation on the moderate halophile Halomonas elongata and the extreme halophile Halobacterium salinarum Can J Microbiol 2000 46 2 180 187 10.1139/cjm-46-2-180 10721487
Makarova KS Aravind L Wolf YI Tatusov RL Minton KW Koonin EV Daly MJ Genome of the extremely radiation-resistant bacterium Deinococcus radiodurans viewed from the perspective of comparative genomics Microbiol Mol Biol Rev 2001 65 1 44 79 10.1128/MMBR.65.1.44-79.2001 11238985
Chabes A Georgieva B Domkin V Zhao X Rothstein R Thelander L Survival of DNA damage in yeast directly depends on increased dNTP levels allowed by relaxed feedback inhibition of ribonucleotide reductase Cell 2003 112 3 391 401 10.1016/S0092-8674(03)00075-8 12581528
Ruby SW Szostak JW Specific Saccharomyces cerevisiae genes are expressed in response to DNA-damaging agents Mol Cell Biol 1985 5 1 75 84 3920512
Lincker F Philipps G Chaboute ME UV-C response of the ribonucleotide reductase large subunit involves both E2F-mediated gene transcriptional regulation and protein subcellular relocalization in tobacco cells Nucleic Acid Res 2004 32 4 1430 1438 10.1093/nar/gkh310 14990748
Ruepp A Soppa J Fermentative arginine degradation in Halobacterium salinarium (formerly Halobacterium halobium): genes, gene products, and transcripts of the arcRACB gene cluster J Bacteriol 1996 178 16 4942 4947 8759859
Unrau P Wheatcroft R Cox B Olive T The formation of pyrimidine dimers in the DNA of fungi and bacteria Biochim Biophys Acta 1973 312 4 626 632 4200353
Williams JI Cleaver JE Excision repair of ultraviolet damage in mammalian cells. Evidence for two steps in the excision of pyrimidine dimers Biophys J 1978 22 2 265 279 656544
McCready SJ Cox BS Repair of 2 μm plasmid DNA in Saccharomyces cerevisiae Curr Genet 1980 2 3 207 210 10.1007/BF00435687 24189911
Birrell GW Brown JA Wu HI Giaever G Chu AM Davis RW Brown JM Transcriptional response of Saccharomyces cerevisiae to DNA-damaging agents does not identify the genes that protect against these agents Proc Natl Acad Sci U S A 2002 99 13 8778 8783 10.1073/pnas.132275199 12077312
Rieger KE Chu G Portrait of transcriptional responses to ultraviolet and ionizing radiation in human cells Nucleic Acids Res 2004 32 16 4786 4803 10.1093/nar/gkh783 15356296
Papke RT Koenig JE Rodriguez-Valera F Doolittle WF Frequent recombination in a saltern population of Halorubrum Science 2004 306 5703 1928 1930 10.1126/science.1103289 15591201
DasSarma S Fleischmann EM Halophiles 1995 Cold Spring Harbor Laboratory Press, Plainview, NY
McCready SJ Henderson DS A dot blot immunoassay for UV photoproducts Methods in Molecular Biology: DNA Repair Protocols: Eukaryotic Systems 1999 Humana Press
Bohr VA Okumoto DS Hanawalt PC Survival of UV-irradiated mammalian cells correlates with efficient DNA repair in an essential gene Proc Natl Acad Sci U S A 1986 83 11 3830 3833 3459159
Schenley RL Fisher WD Swenson PA Pyrimidine dimer excision in surviving and nonsurviving cells of ultraviolet-irradiated cultures of Escherichia coli J Bacteriol 1976 126 2 985 989 770461
Tuite MF Cox BS Ultraviolet mutagenesis studies of [psi], a cytoplasmic determinant of Saccharomyces cerevisiae Genetics 1980 95 3 611 630 7002721
Earl AM Rankin SK Kim K-P Lamendola ON Battista JR Genetic evidence that the uvsE gene product of Deinococcus radiodurans R1 is a UV damage endonuclease J Bacteriol 2002 184 4 1003 1009 10.1128/jb.184.4.1003-1009.2002 11807060
|
16176594
|
PMC1224876
|
CC BY
|
2022-09-08 23:19:43
|
no
|
Saline Syst. 2005 Jul 4; 1:3
|
utf-8
|
Saline Syst
| 2,005 |
10.1186/1746-1448-1-3
|
oa_comm
|
==== Front
Saline Syst
Saline Syst
Saline Systems
1746-1448
BioMed Central
1746-1448-1-5
16176595
10.1186/1746-1448-1-5
Review
Organic compatible solutes of halotolerant and halophilic microorganisms
Roberts Mary F [email protected]
1 Merkert Chemistry Center, Boston College, 2609 Beacon Street, Chestnut Hill, MA 02465, USA
2005
4 8 2005
1 55
3 6 2005
4 8 2005
Copyright © 2005 Roberts; licensee BioMed Central Ltd.
2005
Roberts; licensee BioMed Central Ltd.
https://creativecommons.org/licenses/by/2.0/ This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Microorganisms that adapt to moderate and high salt environments use a variety of solutes, organic and inorganic, to counter external osmotic pressure. The organic solutes can be zwitterionic, noncharged, or anionic (along with an inorganic cation such as K+). The range of solutes, their diverse biosynthetic pathways, and physical properties of the solutes that effect molecular stability are reviewed.
==== Body
pmc
A dynamic and important property of cells is their ability to rapidly adapt to changes in external media, for example, increasing NaCl. To adjust to increased external NaCl, cells in all three kingdoms accumulate a variety of small molecules in the cytoplasm to counteract the external osmotic pressure. Inorganic cations (K+ and in some cells Na+) are often key players in osmotic balance and the osmotic response. However, the diverse collection of organic solutes that organisms accumulate in response to salt stress (also termed osmolytes) is particularly intriguing. Cataloging the occurrence of different molecules and understanding their biosynthesis and regulation by external osmotic pressure have been active areas of research. Studies show that the accumulation of solutes has another role along with osmotic balance. Many osmolytes have been shown to increase the stability of proteins. They appear to act as chemical chaperones in cells, and the mechanism of this stabilization can provide insights into protein folding. The thermostabilizing role of osmolytes has also been exploited for various biotechnology purposes. The aim of this review is an examination of the chemical and biological scope of osmolytes in a wide range of halotolerant and halophilic organisms with an overview of experiments that address why these types of solutes have been naturally selected for osmotic balance. Also included is a brief summary of the biotechnological uses of these organic osmolytes.
The types of organic molecules used for osmotic balance include polyols and derivatives, sugars and derivatives, amino acids and derivatives, betaines, and ectoines and occasionally peptides suitably altered to remove charges [1]. As a general rule of thumb, bacteria and eukaryotes usually accumulate neutral compatible solutes whereas archaea prefer negatively charged solutes [2,3]. Interestingly, archaea tend to modify many of the same neutral or zwitterionic solutes accumulated by eukaryotes or bacteria to make them negatively charged. Osmolytes can either be synthesized by the cell or transported into the cell from the medium. A key feature of these molecules is that they do not inhibit overall cellular functions, although they may modulate individual enzyme activities. This behavior led to labeling them as 'compatible solutes' [4]. Their accumulation helps to maintain turgor pressure, cell volume, and concentration of electrolytes – all important elements for cell proliferation. It is thought that initial events that trigger osmolyte accumulation could include ion channels or other transmembrane proteins sensing differences in external and internal salt concentration, cell volume changes, and/or turgor pressure changes. However, except for transporters, how these physical changes are translated to increased osmolyte synthesis is not known.
Identification of Osmolytes
Although osmolytes tend to occur at high intracellular concentrations, they do not have unique chromophores and were not considered in much detail and in most cases even identified until high resolution NMR spectroscopy became a routine analytical method. From the 1970s onward, a variety of NMR approaches have been used to identify the organic solutes accumulated by halotolerant and halophilic organisms. Early natural abundance 13C NMR studies of cell extracts identified novel solutes, such as ectoine [5], several β-amino acids [6-8], and di-myo-inositol-1,1'-phosphate (DIP), the last associated with hyperthermophiles [9,10]. More recent methods using 1H NMR and two-dimensional experiments have significantly increased the sensitivity of solute detection [11]. 1H NMR methodology can also be used to detect and quantify osmolytes in cell cultures without extraction [12]. Other analytical methods such as HPLC have been used, often to quantify specific solutes as long as an appropriate detection method is available. Refractive index detection is the most general [13], but specific classes of molecules can be derivatized for rapid and sensitive detection (e.g., chromophores added to solutes containing free amino groups [14]). More recent advances have improved on the sensitivity of these other assays. For example, the combination of anion-exchange chromatography and pulse amperometric detection is a very sensitive method that can detect osmolytes such as ectoine after hydrolytic cleavage of the pyrimidine ring [15]. The methodology is sufficiently sensitive that it can be used to screen colonies on agar for solutes.
Organic osmolytes fall into three general chemical categories: (i) zwitterionic solutes, (ii) noncharged solutes, and (iii) anionic solutes. Structures of these molecules and their occurrence in halotolerant and halophilic microorganisms are presented in Figures 1,2,3,4. Intertwined with these organic solutes are K+ and Na+ which also contribute to osmotic balance in cells.
Figure 1 Zwitterionic organic osmolytes detected in bacteria and archaea.
Figure 2 Uncharged organic osmolytes detected in bacteria and archaea.
Figure 3 Anionic organic osmolytes containing carboxylates that have been detected in bacteria and archaea.
Figure 4 Anionic organic osmolytes containing phosphate or sulfate moieties that have been detected in bacteria and archaea.
A. Zwitterionic Solutes
Free polar amino acids in cells might be expected to play a role in osmotic balance. However, neutral amino acids are not accumulated to high concentrations, presumably because they are intermediates in protein biosynthesis. High and varying concentrations of these compounds could affect diverse cell pathways. Instead, many bacterial and archaeal cells synthesize and accumulate a few zwitterionic molecules derived from amino acids as compatible solutes. Structures of these solutes and where they are found are presented in Figure 1.
1. Betaine
This ubiquitous solute, glycine with the primary amine methylated to form a quaternary amine, is found in halophilic bacteria of diverse phylogenetic affiliation [16]. In most cells where it is accumulated as an osmolyte, the betaine is actively transported from complex medium. Betaine concentrations vary with external NaCl. For example, Imhoff and Rodriguez-Valera [16] showed that for the eight halophiles examined, the average betaine concentrations in 3, 10 and 20% NaCl (0.51, 1.7 and 3.4 M) were 0.21 ± 0.2, 0.65 ± 0.06, and 0.97 ± 0.09 M. A number of methanogens have also been observed to accumulate betaine when grown in rich medium [17]. In contrast to the large number of bacteria that transport betaine into the cell for use as an osmolyte, there are only a few bacteria (e.g., Actinopolyspora halophila and Halomonas elongata) and one methanogen (Methanohalophilus portucalensis FDF1) that are able to synthesize betaine either by oxidation of choline or methylate of glycine [18-20].
2. Ectoine and hydroxyectoine
Ectoine, a cyclic tetrahydropyrimidine (1,4,5,6-tetrahydro-2-methyl-4-pyrimidinecarboxylic acid) can almost be considered a marker for halophilic bacteria. As shown in Figure 1, it is synthesized by a wide range of bacteria, both halotolerant and halophilic varieties. This solute was first detected in the halophilic, phototrophic Halorhodospora halochloris [5]. The intracellular ectoine concentration was shown to increase with increased extracellular NaCl. Screens of a number of microorganisms have shown that ectoine is the major osmolyte in aerobic chemoheterotrophic bacteria [1]. It is also the major solute in bacterial strains isolated from alkaline, hypersaline Mono Lake [21]. More recently it has also been observed in the moderately halophilic methylotrophic bacteria Methylarcula marina, M. terricola, and Methylophaga sp. [22,23]. A variant of this solute, hydroxyectoine, has been detected in halotolerant Sporosarcina pasteurii grown in high osmolarity medium [24].
Growth conditions have been shown to affect the intracellular ectoine pool. For example, in halotolerant Brevibacterium sp., the size of the intracellular ectoine pool depends not only on the external salt concentration but on the type of carbon source and aeration level [25]. Ectoine accumulation can also depend on growth stage. In Chromohalobacter israelensis (formerly Bacterium Ba1), ectoine only accumulated when the cells were grown in greater than 0.6 M NaCl, and only in exponentially growing cells [26]. Some microorganisms, e.g., Brevibacterium epidermidis, can also metabolize ectoine [27], perhaps a useful trait if external salt concentrations decrease. In that halotolerant organism, ectoine accumulation occurs only with salt stress, not sugar stress.
The ability to accumulate ectoine can give an organism an ecological advantage. High-osmolarity-adapted Vibrio cholerae cells accumulate ectoine and betaine and out grow non-adapted cells [28]. This has implications for V. cholerae population dynamics when seawater and freshwater and their attendant microbes mix.
3. Nε-acetyl-β-lysine and β-glutamine
Methanogens have a notably different strategy than many other organisms in that they accumulate several β-amino acids for osmotic balance. These solutes provide an excellent strategy for producing a compatible solute since β-amino acids are not incorporated into proteins or other macromolecules. At high external NaCl (>1 M), two zwitterionic β-amino acids have been shown to accumulate in response to external NaCl. Nε-acetyl-β-lysine has been detected in a wide range of mesophilic and a few thermophilic methanogens [7,29-31]. β-Glutamine has been detected in Methanohalophilus species where it is synthesized and accumulated along with Nε-acetyl-β-lysine and betaine [8]. 13C-pulse/12C-chase and 15N-pulse/14N-dilution NMR experiments can be carried out where cells are grown in the presence of an NMR-active isotope (typically 13CO2 or 15NH4Cl) for some time. The labeled compound is then removed or significantly diluted with unlabeled material (12CO2 or 14NH4Cl). Loss of the NMR-active isotope then monitors the turnover of these β-amino acid solute pools in the cells. Both Nε-acetyl-β-lysine and β-glutamine exhibit little if any turnover in cells as expected if these are used only for osmotic balance [30,31].
B. Noncharged solutes
Few molecules that are polar but lack any formal charges have been identified as osmolytes in halophilic bacteria, although they are well represented in eukaryotes. For example, glycerol is prevalent as an osmolyte in marine and halophilic Dunaliella [32,33]. Glycerol accumulation is also a characteristic of halotolerant yeast Debaryomyces hansenii as well as the black yeast Hortea werneckii, and adaptation of this eukaryotic organism to high NaCl requires glycerol accumulation [33]. Myo-inositol, another polyol, is used as an osmolyte in several eukaryotes. Neither of these polar noncharged solutes has been identified as an osmolyte in bacteria or archaea (or associated with halophiles). However, negatively charged derivatives of both glycerol and inositol are accumulated by archaea. The few uncharged solutes that are used by halotolerant bacteria and archaea include several carbohydrates and an amino acid/dipeptide modified to neutralize all charged groups (Figure 2).
1. Carbohydrates
Few carbohydrates are used for osmotic balance, perhaps because those with a reducing end are chemically reactive, and in a sea of proteins these noncharged solutes would be likely to react with surface amino groups. To avoid this, the reactive end of the sugar forms a glycosidic bond with a small neutral molecule, either glycerol or glyceramide. The neutral derivatized sugars glucosylglycerol and α-mannosylglyceramide [34] have been detected in a few bacteria (Figure 2). α-Glucosylglycerol is accumulated by a member of the Proteobacteria, Stenotrophomonas [35]. This organism has a large number of potential biotechnology uses (many based on its ability to use uncommon carbon sources), one of which is the production of glucosylglycerol. α-Mannosylglyceramide is accumulated in Rhodothermus marinus [34].
The non-reducing glucose disaccharide trehalose is used by organisms to counteract drying, but it also serves as an osmolyte. In Actinopolyspora halophila trehalose represents 15% w/v in cells grown in 24% w/v NaCl [36]. However, in some cells, its accumulation is preferred at lower NaCl. For examples, in Chromohalobacter israelensis, trehalose is only an important solute when the cells are grown with <0.6 M external NaCl [26]. In the sulfate-reducing bacterium Desulfovibrio halophilus, trehalose is the major osmolyte. When grown in 15% (2.5 M) NaCl in the absence of a source of betaine, the cells accumulated 8 μmol trehalose/mg protein and ~2.5 μmol K+/mg protein [37].
Disaccharides without modifications, notably sucrose, can be transported by some halotolerant and halophilic organisms, and this can enhance growth in higher NaCl. Sucrose is synthesized in cyanobacteria and proteobacteria [38,39] where it is usually associated with lower salt tolerance strains. Synechocystis sp. strain PCC 6803 tolerates up to 1.2 M NaCl. In those cells, the sucrose is a minor solute (glucosylglycerol is the major osmolyte). However, the sucrose is critical for stationary phase survival under salt stress conditions [40]. This observation was rationalized by proposing that the sucrose could regulate metabolic pathways that are active under the nutritional stress conditions of stationary phase [40].
While it is rarely synthesized in bacteria, sucrose is a major osmoprotectant in plants, and synthesis of sucrose is similar in both cyanobacteria and plants. There are two distinct pathways. In freshwater and marine cyanobacteria, sucrose is synthesized synthesized from fructose-6-phosphate and a sugar nucleotide (UDP-glucose) in two steps using sucrose phosphate synthase and sucrose-phosphate phosphatase [41]. Filamentous cyanobacteria (e.g., Anabaena sp.) use a different pathway, sucrose synthase, which reversibly converts fructose and ADP-glucose (or UDP-glucose) to sucrose [42]. Since sucrose-synthesizing enzymes cannot be identified in other bacteria or archaea, it is thought that sucrose synthesis in eukaryotes was acquired by endosymbiotic cyanobacteria that were the ancestors of chloroplasts [43].
2. Uncharged Amino Acids and Peptides
Two solutes in this class have been identified as osmolytes: (i) a carboxamine, and (ii) an acetylated neutral glutamine dipeptide. In both solutes, modifications mask the charged α-amino and α-carboxyl groups. N-α-Carbamoyl-L-glutamine 1-amide, an unusual amino acid derivative, is accumulated by halophilic phototrophic bacterium Ectothiorhodospira marismortui (also known as Ectothiorhodospira mobilis) [44]. The dipeptide N-acetylglutaminylglutamine amide is synthesized by several halophilic purple sulfur bacteria [45,46].
C. Organic anions
Cells have a negative potential inside and often quite high intracellular K+. Negatively charged solutes could serve to balance high intracellular K+ as well as counteract osmotic pressure. Indeed, at lower external NaCl, many bacteria (including H. elongata which also synthesizes ectoine) and archaea use L-α-glutamate as an osmolyte. In methanogens, high NaCl often causes the cells to switch from anionic glutamate isomers to the zwitterionic solute Nε-acetyl-β-lysine for osmotic balance [29,31]. Anionic solutes used by bacteria and archaea for osmotic balance can have a carboxylate supply the negative charge (Table 3) or contain phosphate or sufate groups (Figure 4).
1. β-Glutamate
Methanogens tend to accumulate β-glutamate as well as α-glutamate for osmotic balance. 13C-pulse/12C-chase NMR experiments that monitor solute turnover for α- and β-glutamate have shown that the α-glutamate pool is metabolized and replenished while the β-amino acid pool is relatively static, hence it is an ideal compatible solute [30,31,47,48]. In Methanothermococcus thermolithotrophicus, both the α- and β-glutamate levels increase with increasing external NaCl [49]. However, there appears to be a threshold for the glutamates in this organism. The negatively charged glutamates are accumulated when the external NaCl is less than 1 M. In that regime, the total intracellular glutamates occur at concentrations comparable to the intracellular K+. Above 1 M NaCl, zwitterionic Nε-acetyl-β-lysine becomes the major solute [31,50]. The accumulation of the zwitterions at high NaCl could indicate that it is now energetically too costly to increase K+ and hence the anionic glutamates that aid in neutralizing much of the K+ are not needed. In support of this are observations in M. thermolithotrophicus that within 30 min of switching cells from 0.67 to 1.4 M NaCl, both K+ and glutamate concentrations increase transiently then later decrease as the zwitterion is eventually synthesized [48].
While most studies identifying β-glutamate have concentrated on methanogens, this solute has been detected in a few bacteria as well. For example, it has been detected in the Gram-positive organism Nocardiopsis halophila, which also accumulates the zwitterionic hydroxyectoine [51].
2. β-Hydroxybutyrate and derivatives
Soluble poly-β-hydroxybutyrates, normally used as carbon reservoirs in cells, have been detected in moderate concentrations in a number of organisms, including Methylarcula marina and Methylarcula terricola [22] and in the deep sea organism Photobacterium profundum SS9 [52]. The role of polyhydroxybutyrates in the deep sea bacterium is particularly intriguing. In P. profundum, betaine and glutamate represent the major solutes when the cells are grown at 1 atm. However, when grown at 280 atm, β-hydroxybutyrate and polymers of this solute accumulate and become the major solutes. At a fixed hydrostatic pressure, β-hydroxybutyrates also increase with increasing external NaCl (particular at high pressures), indicating that the monomer and possibly the polymer (although the enhanced intensity in the NMR resonances for this compound could also indicate increased chain length) function as conventional osmolytes. Because their intracellular levels respond to hydrostatic as well as osmotic pressure, these β-hydroxybutyrate solutes in this organism have been termed 'piezolytes' [52].
3. Anionic polyols and carbohydrates
In bacteria, high intracellular concentrations of negatively charged carbohydrates are not very common. Two such solutes that have been detected include α-glucosylglycerate and α-mannosylglycerate. These solutes tie up the reactive end of the sugar in a glycosidic bond with a hydroxyl group of glyceric acid. α-Mannosylglycerate, accumulated by several Rhodothermus spp., is higher in exponential phase cells and decreases abruptly as cells enter stationary phase [34]. These cells accumulate both the anion mannosylglycerate and the neutral α-mannosylglyceramide. Which of these two solutes dominates depends on stress conditions. Under temperature stress of the cells, R. marinus is biased towards accumulating mannosylglycerate; increased NaCl favored accumulation of the neutral α-mannosylglyceramide rather than the organic anion [34]. Glucosylglycerate has also been observed in Methanohalophilus portucalensis when those cells are grown with methanol rather than methylamine as the substrate for methanogenesis [30]. It is a relatively minor contributor to osmotic balance under those conditions, but the cells do synthesize it. Its turnover, measured by NMR, is roughly twice as slow as α-glutamate and 2–4 times faster than turnover of the zwitterions betaine and Nε-acetyl-β-lysine [30].
Halotolerant archaea (excluding most methanogens) tend to accumulate organic anions where the negative charge is often provided by a phosphate moiety and in some cases by sulfate added to a noncharged solute (Figure 4). Representatives of this class of compounds include the glycerol derivative α-diglycerol phosphate [53] and a series of myo-inositol phosphodiesters based on di-myo-inositol-1,1'-phosphate (DIP) [9,10]. Phosphodiesters are better choices than phosphomonoesters for accumulation at high concentrations since they will have weaker interactions with cations (particularly divalent cations). DIP and its relatives (e.g., mannosyl-DIP [54]) are associated with halotolerant hyperthermophiles. The intracellular concentration of these solutes increases with external NaCl, but the increase is usually more striking with growth temperatures above 80°C [10,55-57]. In Archaeoglobus fulgidus grown at 76°C, α-diglycerol phosphate is the major osmolyte, varying with external NaCl; little if any DIP is detected. However, at 87°C DIP concentrations are comparable to the α-diglycerol phosphate [58]. The association of DIP with very high temperatures suggests that this and related solutes have a role in stabilizing macromolecules to high temperature, although why this odd sugar is used is not clear. Synthesis of DIP from glucose-6-phosphate requires significant energy, so that there must be a reason for its accumulation.
A few archaea have been seen to accumulate cyclic-2,3-diphosphoglycerate, an unusual cyclic pyrophosphate with a net -3 charge. This solute was first detected in strains of Methanothermobacter thermoautotrophicus [59,60], a thermophilic methanogen that is usually grown in medium containing low NaCl. In the Marburg strain of that organism 1,3,4,6-tetracarboxyhexane, a component of methanofuran, is also a major solute [61]. Cyclic-2,3-diphosphoglycerate is also accumulated in Methanobrevibacter smithii, Methanopyrus kandleri and Methanothermus fervidus [58,62]. However, at least in M. thermoautotrophicus it has unusual behavior compared to most osmolytes. It exhibits very rapid turnover in the cells (compared to the cell doubling time) and appears to be fixed into a polymer pool from which it can be retrieved in times of stress [63,64]. Furthermore, its intracellular concentration does not vary much, even when the cells are grown in 0.4 NaCl [61]. This behavior could suggest it has a primary role as a carbon and phosphate storage compound in these methanogens.
Another unusual anionic solute found in haloalkaliphilic archaea is sulfotrehalose [65]. This derivative of trehalose with a sulfate at C2 of one of the glucose rings was the major solute in several Natronococcus and Natronobacterium species grown in defined media. The intracellular sulfotrehalose increases with increased external NaCl and is accumulated in amounts comparable to the intracellular K+ (P. Jablonski, unpublished results). However, the sulfotrehalose can be replaced by sucrose, in which case the cells have roughly double the amount of organic solute. Interestingly, sulfatides with sulfotrehalose (modified tehalose 2'-sulfate with acyl chains on the other glucose moiety attached to C2 and C3) are synthesized in Mycobacterium tuberculosis [66]. Whether or not such sulfatides exist in the archaea has not been determined.
D. K+ and other inorganic ions
The high concentration of organic anions in many halophiles requires counterions such as K+ and/or Na+. However, there are halophiles that exclusively use inorganic ions for osmotic balance. Halophilic aerobic archaea have been shown to have high K+ and Cl- [67], although absolute amounts appear to depend dramatically on the method of analysis and/or growth conditions. There are also bacteria with exceedingly high intracellular K+. In Salinibacter ruber, an obligatory aerobic chemoorganitrophic and very halophilic bacterium, K+ is the major intracellular component of osmotic balance with 11–15 μmol K+/mg protein [68]. Organic solutes are relatively low in this organism; these studies used elemental analysis with EM and an X-ray spectrum to describe constituents and flame-photometric determination of K+. S. ruber occupies a relatively unique position in the bacterial kingdom. Its closest relative, Rhodothermus marinus, uses α-mannosylglycerate and α-mannosylglyceramide as osmolytes for osmotic balance [34].
Other microorganisms that do not appear to use organic osmolytes for balance include anaerobic fementative members of the families Halobacteroidaceae and Halanaerobiaceae [67-70]. Halobacterium salinarum, an archaeon, has been reported to accumulate 12 μmol K+/mg protein. While some organic solutes were observed (≤50 mM), at those low concentrations they are unlikely to play a major role in osmotic balance, although they may aid in charge balance within the cells. Using a cell volume of 2.75 μl/mg proteins for H. salinarum, the estimated intracellular K+ is 4.4–4.8 M comparable to the Na+ concentration in the medium. Extracting cell pellets prior to K+ analyses led to considerably lower values (~1 M for S. ruber and ~3 M for H. salinarum) for K+, suggesting ion leakage. Intracellular Na+ was also high in the pellet extracts, which could suggest problems with this type of analysis. Alternatively, the difference might reflect complexed versus uncomplexed ions. Halophilic sulfate-reducing bacteria, e.g., Desulfohalobium retbaense and Desulfovibrio halophilus, like the haloarchea, appear to use inorganic salts for osmotic balance.
In contrast to the archaeal halophiles, many of the halophilic bacteria do not have exceptional high K+. For example, Halomonas elongata accumulates 1.1 μmol K+/mg protein when grow with yeast extract and 2.2 μmol/mg protein when grown in defined medium with glucose as the sole carbon source. In Halanaerobium acetethylicum grown in medium with 1.7 M NaCl, the internal cytoplasmic Na+ and Cl- are 0.92 and 1.2 M, respectively, while K+ and Mg2+ concentrations in cells are 0.24 and 0.02 M, respectively [70].
Although K+ (and occasionally Na+) appears to be the major intracellular cation, there are reports that, in some cells, Mg2+ can reach moderately high concentrations. Heldal and coworkers [71] have found high Mg2+ (close to 0.9 M) in native marine bacteria under conditions of low dissolved organic carbon. The intracellular Mg2+ was dramatically reduced (<0.2 M) when nutrient levels increased. They suggested that high intracellular Mg2+ is a marker of carbon limitation.
Chloride is the most prevalent inorganic anion in halophiles that do not accumulate organic anions. Molar concentrations of chloride have been detected in several halophilic archaea. This anion is pumped into cells by halorhodopsin or cotransported with Na+. While this anion certainly can contribute to osmotic balance, it appears to have more critical roles in haloadaptation [72]. For example, chloride has been shown to regulate betaine transport. Aside from chloride, little is known about the inorganic anion composition of halophiles. However, a recent FT-IR study of intact bacteria during growth indicates that in H. salinarum and Halococcus morrhuae, large changes occur in the concentration of sulfate ion in the cells [73]. Maximum sulfate occurs during the mid-part of the exponential phase.
E. Cocktails of organic solutes
One of the things arising from the studies of different halotolerant and halophilic organisms is that most cells use an array of solutes, not a single one, for osmotic balance. When a single solute is detected it is often supplied by the medium and efficiently transported into the cell. However, left to its own device, the typical bacterial or archaeal cell synthesizes several molecules that together contribute to osmotic balance. Sometimes this is a combination of anions and zwitterions, but often several solutes with the same net charge. Archaea provide particularly intriguing examples of this strategy, although an explanation for the diversity of solutes in a given organism is lacking.
Methanothermococcus thermolithotrophicus accumulates the anionic α- and β-glutamate when grown in medium with less than 1 M NaCl [49]. Cells adapted to higher external NaCl concentrations switch to accumulating a zwitterions, Nε-acetyl-β-lysine [31,48,50]. Since the glutamate concentration is roughly the same as intracellular K+, the switch to accumulating the zwitterions could be the result of an impaired K+ pump.
Methanohalophilus portucalensis, a halophilic methanogen, accumulates three zwitterions over its growth range: betaine, Nε-acetyl-β-lysine, and β-glutamine [8]. α-Glutamate is detected, but its intracellular concentration is relatively low and does not increase with increased external NaCl. Of the three zwitterions, β-glutamine is only accumulated to large amounts at the high NaCl end of the growth range. Several conditions can affect the balance among these three zwitterions [30]. The cells can grow on trimethylamine or methanol as the substrate for methanogensis, and the substrate with nitrogen promoted accumulation of the two solutes containing two nitrogen atoms (Nε-acetyl-β-lysine, and β-glutamine). Supplying precursors of these solutes (glycine, lysine, glutamate) has little effect on the distribution of the three zwitterions. Betaine is the only solute that could suppress synthesis of both Nε-acetyl-β-lysine and β-glutamine when it is present in the medium [74].
Balancing several different anionic osmolytes has also been observed in the halotolerant, hyperthermophilic Methanotorris igneus, which accumulates L-α-glutamate, β-glutamate, and DIP [10]. Increased external NaCl leads to preferential increases in the intracellular β-amino acid; thermal stress causes increases in DIP levels. Multiple anionic solutes are also accumulated in other hyperthermophilic archaea. In Archaeoglobus fulgidus, glutamate, DIP and α-diglycerol phosphate are used for osmotic balance [53]. In these cells, it is the α-diglycerol phosphate that was most sensitive to external NaCl while heat enhanced DIP synthesis and accumulation.
Biosynthesis of Osmolytes – Novel Pathways and Regulation
A. Betaine
Microorganisms have two different general pathways for synthesizing betaine (Figure 5). The oxidative pathway can occur with a single soluble enzyme (choline oxidase in Gram-positive soil bacteria [75]) or require two distinct soluble enzymes (choline monooxygenase and betaine-aldehyde dehydrogenase in higher plants [76]), or it can occur with a membrane-associated system coded by four genes in the bet operon (in marine invertebrates and bacteria including Escherichia coli). The last system has been studied genetically [77], with genes identified for choline dehydrogenase (betA), betaine-aldehyde dehydrogenase (betB), a choline transporter (betT) and a putative regulator (betI). The choline dehydrogenase catalyzes oxidation of choline to betaine aldehyde, which is then oxidized to betaine by the betB gene product. In Pseudomonas, an electron acceptor other than O2 is used for choline oxidation with suggestions that PQQ is the acceptor [78]. In Actinopolyspora halophila, choline is oxidized to betaine aldehyde then to betaine [36]. The aldehyde is produced with O2 consumption and H2O2 generation. The final oxidation to betaine uses reduction of NADP+.
Figure 5 Pathways for synthesizing betaine in bacteria and archaea.
Even organisms that do not accumulate betaine in response to osmotic stress may have homologues of the genes for synthesizing this solute. Halomonas elongata does not appear to accumulate betaine. However, the organism does have a gene that codes for the oxidation of choline to betaine [79]. Recently the choline dehydrogenase from that organism was expressed in E. coli and characterized. This enzyme can use O2 if no other electron acceptors are available, although Vmax decreases four-fold compared to kinetics with an acceptor such as phenazine methosulfate. Both choline and the betaine-aldehyde are converted to betaine. Although a glycine box suggestive of FAD+ as a cofactor was seen in the sequence, there is no experimental evidence for FAD+ as a cofactor. These observations prompt two questions: (1) what cofactor is used by this choline dehydrogenase, and (ii) under what conditions is this gene expressed?
Several microorganisms can also generate betaine by successively methylating glycine. GSMT (glycine sarcosine methyltransferase) and SDMT (sarcosine dimethylglycine methyltransferase) in Halorhodospira halochloris and Actinopolyspora halophila transfer the methyl group of S-adenosylmethionine to two different types of amines [18,36,80]. Betaine synthesis from glycine in a halotolerant photosynthetic organism Aphanothece halophytica was also carried out by GSMT and DNMT activities [81]. Only one methanogen, Methanohalophilus portucalensis, has been shown to synthesize betaine de novo [8]. In these anaerobic cells, 13C NMR labeling experiments suggest that betaine is generated by successive methylation of glycine [20]. If cells are grown with 10 mM betaine in the medium, accumulation of other osmolytes (Nε-acetyl-β-lysine, β-glutamine and glucosylglycerate) is suppressed [74]. However, supplying cells with betaine precursors does not necessarily alter the distribution of osmolytes. Exogenous glycine or sarcosine had no effect on betaine accumulation. Even though the glycine (13C-labeled) was shown to be internalized by the cells [74], betaine synthesis and suppression of other osmolytes only occurs with N, N-dimethylglycine added. This suggests that it is the N, N-dimethylglycine intracellular concentration that regulates betaine synthesis and accumulation. Recently a 240 kDa N-methyl transferase has been isolated and partially characterized (M.-C. Lai, C.-C. Wang, M.-J. Chuang, Y.-C. Wu, and Y.-C. Lee, personal communication). The source of the methyl groups transferred to glycine is (perhaps not surprisingly) S-adenosylmethionine. While mammalian activites usually show very specific methyltransferase activity (e.g., glycine N-methyltransferase), an aggregate of similar mass proteins (but with different pI values) carries out all three methylation activities in the methanogen. Different subunits in the aggregate are optimized for different methyl transfers; the K+ concentration also differentially modulates the methyltransferase activities [[82], M.-C. Lai and coworkers, personal communication]. The last result suggests that betaine accumulation is likely regulated by the internal K+ concentration in these cells.
B. Ectoine
The biosynthesis and regulation of ectoine in cells have been studied in several different bacteria, both Gram-negative and Gram-positive. Biosynthesis of ectoine in H. elongata has been studied in the greatest detail. The entry molecule into ectoine biosynthesis is aspartate semialdehyde, which is an intermediate in amino acid metabolism [83]. A shown in Figure 6, the aldehyde is converted to L-2,4-diaminobutyric acid, which is then acetylated to from Nγ-acetyldiaminobutyric acid (NADA). The final step is the cyclization of this solute to form ectoine. The genes for biosynthesis of this solute were identified after the isolation of salt-sensitive mutants led to cloning of genes [84]. Ectoine synthesis is carried out by the products of three genes: ectABC. The ectA gene codes for diaminobutyric acid acetyltransferase; ectB codes for the diaminobutyric acid aminotransferase; ectC codes for ectoine synthase [19,85]. The three recombinant H. elongata Ect enzymes have been characterized. The first enzyme (a 260 kDa complex of 44 kDa subunits) generates the diaminobutyrate by transaminating the aspartate semialdehyde with glutamate. Both pyridoxal 5'-phosphate and K+ are necessary for the diaminobutyrate aminotransferase activity [86]. The aminotransferase of step two is activated by 0.5 M NaCl (and similarly by KCl). The last enzyme involved, ectoine synthase, is also activated by NaCl. This suggests that ectoine accumulation is partially regulated by intracellular cations.
Figure 6 Biosynthetic pathway for ectoine.
In Chromohalobacter salexigens, the ectABC genes are regulated at the transcriptional level [87]. Osmoregulated promoters with sequence homology to general stress σ factor have been identified. Since ectoine levels are modulated with betaine present, there must be additional post-transcriptional control. The effect of betaine on ectABC expression and ectoine accumulation was also shown in Marinococcus halophilus [88]. In defined medium, the intracellular level of ectoine increases with NaCl and suppresses accumulation of trehalose. However, in complex medium, betaine is accumulated and ectoine synthesis is suppressed.
In most organisms, it is thought that hydroxyectoine is synthesized directly from ectoine. However, H. elongata has an alternate pathway that was observed in strains defective in EctC. These mutants that can not synthesize ectoine can still convert NADA directly to hydroxyectoine [89]. Canovas et al. [89] proposed that NADA is hydroxylated to 3-hydroxyl-Nγ-acetyldiaminobutyrate, which is then cyclized to hydroxyectoine by 'hydroxyectoine synthase.' The zwitterionic precursor of ectoine can also be accumulated for osmotic balance. NADA has been detected in a salt-sensitive strain of H. elongata [89] that grows optimally with 0.75 to 1.0 M NaCl. It accounts for 80% of the organic solute pool for cells grown in 1.5 M NaCl) with ectoine (6%) and hydroxyectoine (12%) also present. NADA confers higher osmotic stability to the cells than in a H. elongata mutant where diaminobutyrate accumulates [84]. Thus, this solute, but not its diaminobutyrate precursor (which would have a net positive charge) can act as a compatible solute if ectoine synthesis is blocked.
C. β-amino acids
Over the past few years, genes have been identified or proteins isolated or cloned that confirm pathways initially proposed based on 13C isotopic labeling of these solutes. The pathway originally proposed for biosynthesis of Nε-acteyl-β-lysine has two key enzymes: (i) isomerization of α-lysine to β-lysine catalyzed by a lysine aminomutase, then (ii) acetylation of the ε-amino group [20,31]. Recently the genes coding for these two enzymes were identified in Methanosarcina mazei Gö1 [90]: ablA codes for the aminomutase while ablB codes for the β-lysine acetyltransferase. Expression of the two genes, which are organized in an operon, is salt dependent in M. mazei. Several other methanogens, including Methanococcus maripaludis, have homologous genes [90]. Deletion of the abl operon in M. maripaludis generates cells incapable of growth in high salt medium. It will be interesting to characterize the methanogen lysine aminomutase and to compare it to the catabolic enzyme from bacteria that carries out the same chemistry.
Early NMR evidence ruled out a glutamate aminomutase activity as a means of generating β-glutamate [31] but did not identify precursors. As shown in Figure 7, the most likely pathway (proposed based on enzyme activities found in Methanocaldococcus jannaschii by M. Graupner, H. Xu, and R. H. White, personal communication) starts with the reduction of α-ketoglutarate to α-hydroxyglutarate, which is converted to its coenzyme A ester. Elimination of water from the α-hydroxyglutaryl-CoA generates glutaconyl-CoA, which forms β-glutamyl-CoA when ammonia is added (although the direct source of ammonia is not clear). Hydrolysis of the CoA ester generates β-glutamate. The products of the MJ0800 and MJ0400 genes have been identified as the enzymes responsible for water elimination in this pathway by R. H. White and coworkers (personal communication).
Figure 7 Proposed biosynthetic pathway for β-glutamate.
While not all the enzymes for synthesizing β-glutamate have been identified, conversion of β-glutamate to β-glutamine is done in Methanohalophilus portucalensis FDF1 by glutamine synthetase [20,91]. That GS has unusual properties compared to other studied GS enzymes. In particular, its activity with β-glutamate as substrate is much higher than that of other organisms [91]. Regulation of the enzyme must occur in the cell, because β-glutamine is only accumulated to NMR-detectable levels in M. portucalensis when the cells are grown at higher NaCl [8]. Although the in vitro Km values for both α- and β-glutamate in this organism appear quite high, there is likely to be another mechanism responsible for regulation of the synthesis of this solute by glutamine synthetase and accumulation for osmotic balance.
D. DIP
Data from NMR experiments using 13C-labeled precursors to label DIP in Methanotorris igneus coupled with in vitro assays with postulated intermediates [92] led to a pathway for the biosynthesis of DIP that includes four steps (Figure 8): (i) conversion of D-glucose-6-phosphate to L-inositol-1-phosphate (L-I-1-P) via inositol-1-phosphate synthase (IPS); (ii) hydrolysis of the L-I-1-P by inositol monophoshatase; (iii) coupling of the L-I-1-P with CTP to form CDP-inositol; and (iv) generation of the phosphodiester linkage by condensing CDP-inositol with L-I-1-P (via a 'DIP synthase' activity for whom there is yet no candidate in genomes of organisms that accumulate DIP). In P. woesei, but not in M. igneus, DIP could also be generated from incubations of crude cell extracts with GTP and I-1-P [93]. This finding can be explained by the same condensation mechanism, but assuming a multifunctional 'DIP synthase' that catalyzes not only the condensation of CDP-I and myo-inositol but the dephosphorylation of I-1-P as well (presumably without releasing the dephosphorylated product, myo-inositol).
Figure 8 Proposed pathway for DIP biosynthesis in hyperthermophilic organisms.
The IPS reaction has been examined in several hyperthermophiles (Archaeoglobus fulgidus, Methanotorris igneus, Pyrococcus furiosus, P. woesei, and Thermotoga maritima) known to accumulate this solute (L. Chen and M.F. Roberts, unpublished results). IPS activities in crude extracts are ubiquitous in these organisms and fall into two classes: (i) IPS dependent on divalent cations (Mn2+ or Zn2+) is detected in A. fulgidus, while (ii) the IPS activities from the other organisms are not activated by metal ions or NH4 + (the cofactor for all other known IPS). Although it is the first step in DIP synthesis, the IPS reaction is unlikely to be the point where DIP synthesis and accumulation are regulated since many archaea incorporate inositol into their lipids. If they incorporate L-I-1-P into lipids, then the second step, the generation of myo-inositol could be a way to regulate flow of resources into DIP.
The archaeal IMPase enzymes, easily identified by sequence homology to mammalian IMPases, have unusual properties. They exhibit similar substrate specificity to eukaryotic IMPases with one curious exception – they very specifically can dephosphorylate the phosphate on C1 of fructose bisphosphate [94]. The FBPase activity identifies this class of enzymes as dual phosphatases that can process substrates in completely different pathways. FBPase activity gives it a potential role in gluconeogenesis. However, a 'true' FBPase with a homologue in all archaeal genomes was recently purified and cloned [95,96], so that the IMPase/FBPase in archaea may normally function as an IMPase, (although, Methanocaldococcus jannaschii does not accumulate DIP, has no IPS sequence homologue, yet still has a gene for an IMPase/FBPase homologue which has been expressed and characterized [97]). Nonetheless, it is intriguing that an enzyme that could act as either an IMPase or FBPase under the right circumstances (salt or temperature stress?) could link carbohydrate synthesis with responses to stress. At least for the IMPase from A. fulgidus, there are hints as to what could regulate this enzyme. This IMPase has two spatially close cysteine residues that can be oxidized to form a disulfide (either by vigorus bubbling with O2 at 85°C or by adding oxidized E. coli thioredoxin [98]). Formation of the intramolecular disulfide inactivates the enzyme; treatment with either a reducing agent or a reduced thioredoxin can regenerate active enzyme. Unfortunately, the lack of genetics with A. fulgidus makes it difficult to see what role this protein does play in hyperthermophiles. Although enzymes for the third and fourth steps in DIP production have not been identified, the last step has been demonstrated with cell extracts and added CDP-inositol and myo-inositol (L. Chen and M.F. Roberts, unpublished results). DIP synthesis required the presence of Mg2+.
E. α-Mannosylglycerate (α MG)
The synthesis of this osmolyte has been examined in several hyperthermophiles. There appear to be two distinct pathways (Figure 9). In R. marinus, there is a direct condensation of GDP-mannose and D-glycerate to form α MG catalyzed by mannosylglycerate synthase [99]. A second pathway, used by hyperthermophilic archaea, converts GDP-mannose and D-3-phosphoglycerate to mannosyl-3-phosphoglycerate via mannosyl-3-phosphoglycerate synthase, followed by phosphoglycerate phosphatase activity to remove the phosphate group [100,101].
Figure 9 Two pathways exist for α-mannosylglycerate biosynthesis. In (I) GDP-mannose is directly converted to mannosylglycerate. In (II), the GDP-mannose condenses with 3-phosphoglycerate to for mannosyl-3-phosphoglycerate, which is subsequently dephosphorylated to form mannosylglycerate.
F. Cyclic-2,3-diphosphoglycerate (cDPG)
The biosynthesis of the solute cDPG diverts resources from gluconeogenesis by phosphorylation of 2-phosphoglycerate (2-PG) with ATP to form 2,3-diphosphoglycerate (DPG) via 2-phosphoglycerate kinase [102]. Different transformations involving cDPG production and hydrolysis are shown in Figure 10. A novel enzyme, cyclic 2,3-diphosphoglycerate synthetase (cDPGS) [103,104] then converts DPG to the cyclic form with ATP hydrolysis driving the reaction. Hydrolysis of cDPG to 3-PG would shuttle carbon and phosphate back into gluconeogenesis. In Methanothermus fervidus cDPGS is reversible, although it prefers the direction of cDPG synthesis [104]. The ability to generate ATP from ADP, inorganic phosphate Pi) and cDPG under these conditions could argue for a role in energy storage in that organism. However, in Methanobacter thermoautotrophicus, the K+-activated cDGPS appears irreversible (and membrane-bound), suggesting it may have a different role [105]. In soluble cell extracts, hydrolysis of cDPG in the presence of ADP and Pi could not generate ATP [106]. However, a membrane bound hydrolase that is inhibited by K+ and Pi has also been identified [105]. The regulation of cDPG degradation in M. thermautotrophicus is consistent with it playing a role as a carbon and phosphate storage compound, although its high concentration in cells clearly indicates it contributes to osmotic balance.
Figure 10 Proposed biosynthesis of cDPG as a pathway linked to gluconeogenesis through 2-PG and 3-PG. The dashed lines indicate the reversible cDPGS reaction of Methanothermus fervidus. The solid lines show cDPG and 2,3-DPG interconversions and illustrate the irreversible nature of the cDPGS in Methanobacter thermoautotrophicus.
Transport of Osmolytes
Osmolyte transporters also play important roles in the osmotic response. Some of these transporters are very specific and serve to retrieve any solute released by cells. Others have evolved to scavenge solute or osmolyte precursors so that the more wasteful biosynthetic resources of the cell are not used. Recent years have witnessed progress in identifying and characterizing the proteins responsible for uptake of the osmolytes betaine and ectoine from the medium. In other cases putative transporter genes have been identified but no accumulation of the solute is observed. A summary of the different types of betaine (and one ectoine) transporters is presented in Table 1. Another category of membrane proteins involved in osmolyte movement are the mechanosensitive channels. These are the major players in responding to hypoosmotic stress in that they serve as the conduits for solute removal from the cytoplasm. A recent review on many of the membrane proteins acting as osmosensors can be found in [107].
Table 1 Halotolerant or halophilic microorganisms that can transport betaine or ectoinse from the medium.
Solute & Organism Comments Reference
Betaine:
Acinetobacter sp. F-2-12 in 20% NaCl, cells accumulate 1.26 M betaine and 0.36 M glutamate [16]
Actinopolyspora halophila cells can synthesize it de novo (oxidation of choline) as well as transport it from the medium [36]
Alcaligenes sp. F-5-7 ~1 M betaine when cells grown in complex medium in 20% NaCl [16]
Alteromonas sp. A-387 [16]
Chromohalobacter israelensis betaine in the medium suppresses ectoine biosynthesis [26]
Chromobacterium marismortui A-65 in 20% NaCl, cells accumulate 0.5 M betaine and 0.10 M glutamate [16]
Corynebacterium glutamicum has genes for four uptake systems including high affinity BetP and a low capacity osmoregulated permease [111]
Desulfovibrio halophilus 1 mM external betaine suppresses sucrose synthesis [37]
Listeria monocytogenes halotolerant organism also accumulates acetylcarnitine, carnitine, γ-butyrobetaine and 3-dimethylsulfoniopropionate [110]
Marinococcus halophilus BCCT family transporter BetM [109]
Methanohalophilus portucalensis FDF1 accumulation of external betaine suppresses synthesis of osmolytes; bta gene responsible for ABC transporter activated by heat and salt stress [8]
Methanosarcina mazei Gö1 ota gene responds to salt shock [113]
Methanosarcina thermophila TM-1 High affinity ABC transporter [112]
Nesterenkonia halobia CCM 2591 in 20% NaCl, betaine is 1.10 M while glutamate is 0.05 M [16]
Pseudomonas sp. F-12-1 [16]
Tetragenococcus halophilus single component transporter (ButA) that is a member of BCCT family; specific for betaine [108]
Salinivibrio costicola A-514 [16]
Ectoine:
Halomonas elongata transporter similar to tripartite ATP-independent periplasmic transporter family (TRAP-T) [116]
Marinococcus halophilus role likely to be recovery of leaked ectoine [109]
Sinorhizobium meliloti ABC ectoine transporter identified [117]
A. Betaine
Betaine transport is common to a wide variety of halotolerant and halophilic organisms, both bacteria and archaea. There are basically two superfamilies of betaine transporters: (i) secondary transporters that use either the proton motive force or sodium motive force to drive betaine accumulation, and (ii) ATP binding cassette (ABC) transporters that couple ATP hydrolysis to uptake.
Most organisms that internalize this solute do so via a member of betaine choline carnitine transporter (BCCT) family of secondary transporters [108]. The transporter can be quite specific for betaine as in the moderate halophilic lactic acid bacterium Tetragenococcus halophilus [108] or Gram-positive Marinococcus halophilus [109]. Alternatively, the transporters available can internalize a wider range of solutes. For example, Listeria monocytogenes internalizes acetylcarnitine, carnitine, γ-butyrobetaine and 3-dimethylsulfoniopropionate as well as betaine, and the uptake increases the growth rate [110].
Other secondary transport systems have also been described. Corynebacterium glutamicum has the usual high affinity BetP uptake system. However, if this and the genes for three other compatible solute uptake systems are deleted, betaine can still be internalized, although the uptake is significantly reduced. The gene identified for the low capacity osmoregulated permease (lcoP) codes for a protein (LcoP) resembling a member of the BCCT-family [111]. External osmolarity regulates expression and activity of LcoP.
Methanogens that can transport betaine into the cell tend to use a high affinity transporter [112] that is an ABC transporter. ABC betaine transporters have a nucleotide binding domain that hydrolyzes ATP, a membrane spanning domain, and a substrate binding domain (and/or a periplasmic or extracellular binding protein with a high affinity for betaine). The ota (osmoprotectant transporter A) gene of Methanosarcina mazei responds to salt shock [113]. Methanohalophilus portucalensis FDF1 can transport betaine into the cell as well as synthesize it de novo [8,82]. The bta gene responsible for betaine transport in this organism is also an ABC-transporter and is activated by heat as well as salt stress [114]. It is highly specific for betaine – choline, proline and dimethylglycine, and carnitine could not compete with betaine uptake. Interestingly, addition of exogenous betaine or its biosynthetic intermediates induced bta expression immediately. The energy required for synthesis of betaine is 36 ATP whereas only two ATP are required for betaine transport by bta (S.-C. Chen and M.-C. Lai, unpublished results).
How does betaine finds the transporter? In hyperthermophiles, the high affinity ligand-binding protein ProX serves to bring the betaine to the transporter. Crystal structures of the A. fulgidus ProX in the absence and presence of betaine have identified cation-π interactions and non-classical hydrogen bonds between protein and ligand [114]. Similar ligand binding domains have been identified in ORFs in the genomes of other archaea.
B. Ectoine
Ectoine that is provided in the medium can be internalized by some microorganisms. Growth of halotolerant Brevibacterium sp. JCM 6894 is stimulated by exogenous ectoine or hydroxyectoine [115]. In H. elongata the transporter for ectoine and hydroxyectoine (TeaA, TeaB, TeaC) is similar to members of the tripartite ATP-independent periplasmic transporter family (TRAP-T) [116]. The Ks(ect) is 21.7 μM, indicating a high affinity for external ectoine. The role of this transporter appears to be recovery of ectoine leaked from the cell. Marinococcus halophilus also can transport external ectoine. In this cell, the EctM gene product is a BCCT family member [108].
In the same vein, a proteomic analysis of Sinorhizobium meliloti in media that was supplemented with ectoine detected increased synthesis of ten proteins, eight of which were identified by MALDI-TOF analysis of peptides from the two-dimensional gels [117]. Five of these belong to the same gene cluster (localized on the pSymB megaplasmid), whose components code for the ATP-binding cassette transporter ehu (ectoine/hydroxyectoine uptake). Another cluster of genes (eutABCDE) would produce proteins capable of ectoine catabolism. The net result of exposing S. meliloti to ectoine is to enhance the production of proteins to internalize and use any of these molecules that escape the cell.
C. Homologues of Transporter Genes
Halobacterium salinarum has two ORFs upstream of transducer genes with significant homology to binding proteins for amino acids and compatible solutes [118]. Deletion mutants indicate that the CoSB/CosT binding/transducer pair, in which the CosB is a membrane-anchored receptor, is critical for chemotaxis towards compatible solutes (in this case betaine). Whether or not the organism accumulates large amounts of betaine (which seems not to occur in Halobacterium NRC-1 [119]), this protein pair could function as a chemotaxis signaling pathway for organic osmolytes.
D. K+
A variety of K+ channels have been identified in microorganisms. Structures of various K+ channels, initially a closed, small bacterial channel [120] and more recently a gated K+-channel (MthK) from Methanobacter thermoautotrophicus [121], have contributed to understanding how these proteins are arranged in membranes. However, these K+-channels do not respond to altered osmotic pressure. Rather, different protein complexes appear to regulate intracellular K+ in response to osmotic stress. H. elongata uses K+-glutamate as an osmolyte. Recent work has identified three genes required for K+ uptake: trkA, trkH, and trkI [122]. The protein expressed by trkA would be analogous to the cytoplasmic NAD+/NADH binding protein TrkA in E. coli that is required for K+ uptake by the Trk system, while the H. elongata TrkH and TrkI are likely to be transmembrane proteins. Experiments with H. elongata indicate the TrkI is the main K+-transporter in this organism. Similar uptake systems may exist in other halophiles as well.
E. Membrane Osmosensors
Cells will swell upon hypoosmotic shock as water rushes into the cell. To return to the original cell volume, cells need a rapid means of cytoplasmic solute efflux. All microorganisms have families of gated transmembrane channels that open for solute release when the lateral pressure of the membrane drops below a critical value (for review see [123]). Mechanosensitive channels (Msc) are gated by membrane tension and thought to be primary biosensors for osmoregulation in bacteria [124,125]. Their major role appears to be the rapid and non-discriminating release of solutes upon hypoosmotic shock [125]. Msc fall into three classes: MscL, a pentamer with no solute preference that has large conducting activity; MscS, a heptamer that has smaller conductance, is sensitive to membrane tension, and can exhibit selectivity for anions or cations; MscK, likely a heptamer like MscS that is activated by cytoplasmic K+ [123]. These channels are usually closed but upon changes in membrane tension can open to allow solute efflux. Other osmosensors include the Volume-activated channels (VAC). These have been suggested to respond to hypoosmotic response as anion channels [126]. VAC sensors appear to be responsible for the expulsion of a variety of osmolytes, notably amino acids and polyols.
Macromolecule Stabilization By Osmolytes – Theories
Along with balancing external osmotic pressure, compatible solutes have also been shown to stabilize macromolecules. There are many theories regarding protein-solute interactions. These can be classified into two types: (i) those that postulate direct solute-macromolecule interactions and (ii) those that hypothesize that macromolecular stability is mediated by solute-induced changes in water structure.
A. Solute-Macromolecule Interactions: Preferential Solute Exclusion andHydration
Osmolytes in high concentrations compete with water molecules for interactions with protein surfaces. However, it has been proposed that these organic solutes are preferentially excluded from the surface of proteins [127-130]. This in turn leads to preferential hydration of the protein. The increased osmotic pressure generated by the solutes should favor compact folded proteins, which expose less surface area than denatured protein (Figure 11A). The size of internal cavities and internal water should be reduced as well [131]. Differential interactions of organic solutes with folded and denatured proteins also contribute to their stabilization effects. Bolen and coworkers [132] have proposed that, compared to water, solutes have more unfavorable interactions with the peptide backbone and since unfolded protein has more available backbone, this biases the equilibrium to a folded protein (Figure 11B). This would suggest that osmolytes that impart stability actually interact with the unfolded state of the protein, shifting the equilibrium to promote the folded configuration (the 'osmophobic effect' [132]). The free energy of the denatured state is higher than that of the native state, making population of this state energetically unfavorable (Figure 11B). Any interactions of osmolytes with hydrophobic residues of the unfolded protein do not overcome the osmophobic effect, nor do they interfere with the hydrophobic effect.
Figure 11 (A) Exclusion of solutes from the surface of a protein increases the concentration of solute in the bulk solution, which in turn increases water surface tension generating osmotic pressure that drives a protein to retain a more compact structure. (B) Osmolytes stabilize proteins to thermal denaturation by differentially raising the energy level of the unfolded state: U, unfolded state; F, folded state; aq, aqueous solution; S, solution containing osmolytes.
B. Solute-Induced Changes in Water Structure
Ionic solutes will have a pronounced effect on water structure and these interactions will affect macromolecule stability. Neutral salts do not have the same effect on structure and solubilities of proteins. The Hofmeister series of ions reflects the ability of different ions to bind water [133,134]. 'Kosmotropes' (order makers) exhibit a strong interaction with water, while 'chaotropes' (disorder makers) exhibit weaker interactions with water than water has with itself. It is thought that kosmotropes bind water strongly and aid in preserving the hydration layer around the macromolecules. These solutes prefer interactions with water rather than the protein surface, hence preserve preferential hydration of the protein. Chaotropes can displace water from the protein surface and contribute to destabilization of structure by dehydrating the macromolecule. Many osmolytes are strikingly similar to the ions of the Hofmeister series: amino acids resemble ammonium acetate and the methylamines are functionally similar to quarternary ammonium ions [135]. Their effects on proteins should then be similar to those in the Hofmeister series.
Collins and coworkers have shown that the Hofmeister series is actually a function of the apparent dynamic hydration number of the ion [136,137], with the more hydrated an ion, the greater the stabilization of macromolecules. The calculated hydrated radius for each ion nicely coincides with the ionic strength. Collins postulated that the effect of an osmolyte on another solute (in this case the macromolecule) depends on the extent it perturbs the solvation layer of the other solute. If the osmolyte is tightly hydrated, it cannot as easily interact with the macromolecule solvation layer, an event that would destabilize the macromolecule.
Other considerations of water structure and the influence of solutes on charged regions of macromolecules have suggested that the water layer at the protein surface is more dense and reactive than bulk water [138]. Patches of dense water, along with counterions, would cover charged surface regions of the protein, while inert zones of low-density water would be found around hydrophobic groups. Macromolecular crowding (see below) would also influence this. In dilute solutions, changing the density of the bulk water would be energetically unfavorable because of the large relative volume. When the solution is concentrated (a cell typically has 2 to 4 g water/g dry weight [138]), the volume of surface water becomes comparable to the volume of bulk water, allowing density changes to occur.
C. Osmolytes, Excluded Volume, and Pressure Effects
The crowded and inhomogeneous environment of the cell also contributes to stabilization of folded proteins [139-141]. The presence of solutes such as osmolytes, 'macrosolutes' such as cofactors, and other macromolecules aid in stabilizing proteins by decreasing the accessible volume, shifting the equilibrium between the folded and unfolded state of proteins to favor the more compact folded state.
This has relevance to pressure stress as well. Increasing hydrostatic pressure should promote water penetration from protein surface to the core. With osmolytes that are bigger than water, this penetration is less likely if critical water is around the protein and bulk water is diluted with the osmolyte. Studies of Photobacterium profundum solutes at 1 and 280 atm are perhaps the best evidence that small solutes can repel / inhibit water penetration at high hydrostatic pressures [52]. Elevated hydrostatic pressure tends to denature proteins [142], presumably by enhancing water penetration into the protein core [143]. In this deep sea organism, β-hydroxybutyrates accumulate to high intracellular levels at 280 atm. Perhaps these negatively charged solutes aid in preventing water penetration, both with an excluded volume effect but also by altering water structure in the vicinity of the proteins.
The observation of osmolyte cocktails in different types of cells likely reflects selection of solutes that cover different aspects of these effects – destabilization of the denatured state, retardation of water penetration in protein cores, and optimal modulation of water density for a particular cytoplasm. For example, one might expect thermophiles to have different solutes than mesophiles if solute exclusion is less important than other aspects of osmolyte effects (perhaps at high temperatures altering water structure is more important). The accumulation of multiple solutes may be explained by slightly different (and potentially overlapping) effects for each specific solute. For example, in a cell with multiple zwitterions, perhaps solute size or charge distribution are important in stabilizing water at some protein surfaces, while for other proteins preferential exclusion is the dominant effect. In cells with mixed zwitterions and anions, perhaps the anions are better excluded by proteins with a net negative charge, but if the organic anions are balanced by intracellular K+ there may be a limit on their intracellular concentration and so zwitterions are also accumulated.
Experimental Effects Of Osmolytes On Macromolecules
A. Thermoprotection of Proteins
Heat stress often provokes similar responses to salt stress. Organisms adapt to high external salinity by accumulating osmolytes, and the same solutes accumulated in vivo can also affect stability of microorganisms to thermal stress. E. coli cells adapted to grow in high salt contain increased betaine. Diamant et al. [144] showed that heat shock of these salt-adapted cells dramatically reduces the protein aggregation seen in non-adapted cells under the same stress. This behavior was suggested to result from osmolyte (specifically betaine, glycerol, proline or trehalose) activation of chaperones GroEL, DnaK, and ClpB. While such interactions could be shown at low osmolyte levels, at high osmolyte levels, refolding of proteins was reduced, possibly because of specific deleterious interactions of the osmolytes with chaperones.
In many cases, these small molecules assist in protein stabilization and/or refolding in vitro. The in vitro studies with purified enzymes allow one to explore any protective effects or unusual behavior of novel compatible solutes. These studies are consistent with the hypothesis that osmolytes are selected for their unfavorable interactions with peptide backbones [132].
Rabbit muscle lactate dehydrogenase has been used to test the effect of ectoine, hydroxyectroine, and their biosynthetic precursors DA and NADA on thermostability of this enzyme. At 55°C NADA enhances thermostability as measured by protection of the enzyme from thermal inactivation [89]. Hydroxyectoine is more effective than ectoine and NADA at stabilizing proteins to heat stress. The one real difference with hydroxyectoine is a hydroxyl group on the ring. Perhaps this further functionalization of the (now trihydro)pyrimidine ring aids in organizing water and maintaining high surface tension at high temperatures.
Ribonuclease has been a popular target for osmolyte stabilization studies. This disulfide crosslinked enzyme can be reversibly unfolded in the absence of reducing agents. The effects of some of the more exotic osmolytes have been examined with this enzyme. 2-O-α-Mannosylglycerate, 0.5 M, increases the mid-point of the thermal denaturation curve, Tm, by 7°C as well as increases the heat capacity for the protein [145], an effect consistent with the solute destabilizing the denatured state with respect to folded protein. Other studies with ribonuclease D show that a variety of zwitterionic osmolytes dramatically increase Tm for the protein. As an example, 6 M sarcosine increases Tm by 24.6° at pH 5 [146]. Crystal structures of the ribonuclease fail to detect any bound osmolyte or alterations in water bound to the protein. The data support the hypothesis that osmolytes stabilize proteins by perturbing unfolded states, which biases the equilibrium to a compact, folded state.
The charge distribution of an osmolyte can be important to its biological activity as well. Solutes used by bacteria and archaea have not been examined since the pKa of any functional groups are well outside accessible pH ranges for maintaining native proteins. Trimethylamine N-oxide (TMAO), a common solute in eukaryotes, is zwitterionic above pH 6 (pKa of 4.66), and it is this state of the molecule that is critical for its stabilization of proteins [147]. At low pH it no longer acts as a good thermoprotectant. While this observation may not have physiological relevance, it aids in our understanding of osmolyte properties important for their biological effects. TMAO has been shown to decrease the entropy of the unfolded state of onconase through a solvophobic effect [148]. This solute clearly diminishes the unfolding rate while having little effect on the stability of the native protein. For onconase, TMAO appears to induce a local structural change that retards unfolding.
The solute exclusion theory would argue for little specificity in osmolyte effects on macromolecules. However, there are many studies that clearly show preferential stabilization by solutes. Osmolytes can counteract denaturants such as urea. Potassium D- or L-glutamate (0.25 M) counteracts the effect of urea on glutaminyl-tRNA synthetase from Escherichia coli by shifting the equilibrium between the native and molten globule and molten globule to unfolded protein to a higher urea concentration [149]. However, for this protein other osmolytes (sorbitol, TMAO, inositol) cannot induce the same shift. A major conclusion of these studies is that the ability of an osmolyte to counteract urea denaturation depends on specific osmolyte-protein interactions.
As another example, two archaeal rubredoxins have been shown to be stabilized to quite different extents by α-diglycerol phosphate [150]. Their structures are similar, except that one is missing a hairpin loop. There are small conformational changes induced by α-DGP (or mannosylglycerate) and evidence for solute inducing a more compact state of the protein, and the occurrence of weak, specific interactions between osmolyte and protein surface.
Osmolytes can affect protein conformation and motions of native structures as well. TMAO induces α-helix formation of alanine-base peptides [151]. Compatible solutes also attenuate structural fluctuations as measured by amide hydrogen-deuterium exchange rates [152-154]. Osmolytes certainly inhibit slow, large unfolding transitions, but they can also modulate fast exchange rates as well [155]. Tryptophan phosphorescence has been used to probe the flexibility of the native structure of azurin and a number of mutants [131]. The sugar dampens fluctuations only for loose internally hydrated macromolecules and those with thermally expanded conformations. The sucrose (and presumably other polyols) will shift the equilibrium of protein conformations to a more compact rigid form.
One of the interesting questions is whether or not solutes in halophiles stabilize proteins in the same manner as for nonhalophiles. An interesting case in point is the effect of KCl on the dihydrofolate rductase (DHFR) from Haloferax volcanii compared to that from E. coli [156]. The protein from the extreme halophile is much more acidic and one might think the stabilization effects by K+ would differ compared to the mesophilic, non-halophilic homologue. The H. volcanii DHFR requires at least 0.5 M KCl to stay folded, while the E. coli protein is inactive above 1 M KCl. Yet the effect of salts on the stability of the proteins to urea is similar, if one compares stability at the appropriate physiological ionic strength. This work shows that salts stabilize the DHFRs by a common mechanism – preferential hydration and the Hofmeister effect of salt on the activity and entropy of the aqueous solvent. Although one could imagine hydrated salt networks occurring in the halophilic protein leading to halophile-specific stabilization, that is not the case.
B. Interaction with Nucleic Acids
Although most research into how osmolytes affect macromolecular stability has concentrated on proteins/enzymes, these solutes also affect nucleic acid stability. The addition of high concentrations of zwitterionic solutes increases the dielectric constant of the solution that, in turn, decreases ionic interactions and affects the DNA duplex. Isolated studies have explored the effect of zwitterionic solutes on nucleic acid stability. For example, betaine has been shown to eliminate the dependence of dsDNA melting on the base pair composition [157] and to enhance amplification of GC-rich templates [158] by lowering the Tm for the template. High concentrations of compatible solutes also alter accessibility of regions of the DNA to nucleases. Malin et al. [159] showed that ectoine and hydroxyectoine alter the DNA conformation such that endonucleases can no longer cleave it.
Biotechnological Applications Of Osmolytes
The properties of osmolytes make them suitable for a variety of uses in biotechnology as long as one can generate reasonable quantities either in vivo or in vitro. Induction of osmolytes in cells can increase protein folding, so that engineering osmolyte biosynthesis genes in an organism should improve its salt tolerance. The trick is to couple osmolyte production to salt stress. For in vitro uses, large amounts of pure solutes are needed. In many cases, the solutes can be supplied by 'bacterial milking.' Both ectoine and hydroxyectoine have been produced in large quantities using Halomonas elongata [160]. Bacteria in high NaCl are transferred to low osmolarity medium where they excrete the now excess solutes. Re-exposure of the bacteria to high salt induces them to re-synthesize the osmolytes. Repeated transfers between low and high osmolarity media should dramatically enrich the media in the osmolytes. Purification of the solutes then relies on chromatographic steps. This process is the basis of the German biotechnology company Bitop http://www.bitop.de/sources/html/e/index.htm that has developed preparative methods for many of the unique osmolytes produced by microorganisms.
A. Chemical Chaperones for Protein Folding
Insoluble or misfolded overexpressed proteins can often be partially denatured and refolded in the presence of osmolytes. A specific example is the use of osmolytes to enhance the yield of folded, functional cytotoxic proteins directed to the periplasm of E. coli [161]. Cells grown in 4% NaCl with 0.5 M sorbitol and supplemented with 10 mM betaine can accumulate large amounts of the target protein in the periplasm (this was tried with immunotoxins). Protein is released by freeze-thaw cycles. Both high osmotic strength and added compatible solutes (in this case betaine and sorbitol) are necessary for high yields of protein.
In the same vein, ectoine, betaine, trehalose, and citrulline have been shown to inhibit insulin amyloid formation in vitro [162]. This observation may provide directions for designing small molecules to inhibit myelin formation associated with neurodegenerative disorders.
B. Enhancing PCR
Several osmolytes (notably betaine, ectoine) have been shown to be useful in PCR amplification of GC-rich (72.6% GC) DNA templates with a high Tm. In particular, ectoine was shown to outperform regular PCR enhancers; it works by reducing the DNA Tm [163]. Interestingly, hydroxyectoine increases the Tm of duplex DNA. However, the optimal solute for these experiments is homoectoine (4,5,6,7-tetrahydro-2-methyl-1H-[1,3]-diazepine-4-carnoic acid), a synthetic derivative of ectoine with the ring expanded by one carbon. For betaine the effective range of solute is 0.5 to 2.0 M; for ectoine much less (0.25 to 0.5 M) is needed for the same effect. It would be intriguing to see what effect DIP type solutes have on PCR since they are synthesized by hyperthermophiles above 80°C.
C. Cryo-protection of microorganisms
Organic osmolytes have also been used as cryo-protectants. In a recent study, the ability of betaine to act as a cryo-protectant during freezing of diverse bacteria was examined. Betaine is often much better than two common cryo-protectant mixtures, serum albumin and trehalose/dextran, particularly under conditions simulating long-term storage [164]. It is better than the other treatments at preserving long term viability for microorganisms like Neisseria gonorrhoeae and Streptococcus pneumoniae. Betaine is as effective as glycerol for liquid nitrogen freezing of halophilic archaea, and neutrophilic Fe-oxidizing bacteria.
D. Use in cosmeceuticals and pharmaceuticals
The ability of osmolytes to aid in protecting cells from diverse stresses has led to the use of at least one of them, ectoine, in the cosmeceutical industry. Ectoine has been shown to protect skin from UVA-induced cell damage [65]. Based on this, RonaCare™ Ectoin, produced by Merck KgaA, Darmstadt, is presently in use as a moisturizer in cosmetics and skin care products.
Osmolytes have not been developed as reagents in the pharmaceutical industry, in part because as 'compatible solutes' they interact minimally with cellular machinery. However, their ability to stabilize biomolecules may have some very specific uses. As an example, the German company Bitop in collaboration with researchers at the Cologne University Clinic is exploring the use of these solutes in certain cancer therapies where they may protect tissues against vascular leak syndrome, a severe side effect of anti-caner agents.
E. Generation of Stress-Resistant Transgenic Organisms
Insertion of genes for osmolytes into non-halotolerant organisms should increase their ability to withstand salt stress. Plants are a good target for these types of experiments since they are often exposed to drought conditions that would concentrate salt. A few reports of transgenic plants suggest that eventually this strategy might be useful. Arabidopsis thaliana transformed with a choline oxidase gene (which is needed to synthesize betaine) from Arthrobacter globiformis has a significantly improved tolerance of salt stress along with improved cold and heat tolerance [166]. Transgenic tobacco with E. coli betA and bet B genes has also been constructed. This modified plant exhibits better salt and cold tolerance [167]. Inserting the H. elongata ectABC genes also confers hyperosmotic tolerance on cultured tobacco cells [168]. This was shown to increase the hyperosmotic tolerance of cultured cells, although only a small amount of ectoine accumulated. Other recent work to introduce genes for synthesizing osmolytes in plants [169] as a way to improve stress tolerance has, so far, not led to high accumulation of the osmolytes. Further developments await a determination of what limits osmolyte levels in plant cells.
Conclusion
Clearly, there are many different organic solutes used for osmotic balance in halotolerant and halophilic microorganisms [170]. Except for α-glutamate, the solutes that are accumulated are not intermediates in biochemical pathways. They are appropriately modified so that they are not chemically reactive even if they occur at high concentrations (i.e., no reactive groups in carbohydrates) and would have little affinity for the macromolecules that interact with precursors. In most cases these organic solutes are not metabolized by the cells that accumulate them. With these properties, they nicely fulfill Brown's original definition of a compatible solute. However, some of the more uncommon solutes raise interesting questions. For example, what is it about DIP that makes it a solute of choice in hyperthermohiles growing above 80°C?
One of the recurrent themes is that most microorganisms use a cocktail of solutes unless an external solute such as betaine is provided. One solute may be the major species, but there are usually several solutes at moderate concentrations in a cell, and the balance among the solutes can be modulated by growth stage and carbon and nitrogen substrates. In several of the cases described, there is a switch from one type of solute to another with increasing external NaCl. The most common change is from anionic organic solutes to neutral or zwitterionic solutes (e.g. glutamates to Nε-acetyl-β-lysine in several methanogens, α-mannosylglycerate to α-mannosylglyceramide in Rhodothermus marinus). This suggests that internal cation concentrations are intimately linked with the osmolyte pools. There is also evidence that when challenged with increased external NaCl, many organisms exhibit an initial response, which in terms of solutes may include K+ internalization and α-glutamate synthesis. The solutes in this first response are then replaced by a steady-state of other solutes reflecting the adapted cell osmolyte composition. How the cell coordinates this is an area that certainly needs to be explored.
In the past decade there has been significant progress in defining biosynthetic pathways and in identifying the enzyme components for many of these solutes. Ectoine, betaine, and mannosylglycerate synthesis have been examined in detail in several cells. Recent work has at least identified the genes coding for the lysine aminomutase and acetyltransferase needed for Nε-acetyl-β-lysine biosynthesis, and several of the enzymes along the biosynthetic pathways for β-glutamate and DIP have been characterized. Yet the enzymes responsible for the synthesis of other solutes (notably the neutral dipeptide and derivatized amino acid, α-diglycerol phosphate, and sulfotrehalose) have not been explored. In order to understand how external NaCl concentrations are linked to osmolyte synthesis (or removal) such information is critical.
Significant work has also been carried out to understand the thermoprotective features of osmolytes. It has become clear that solutes can have specific effects on protein dynamics and appear to limit some types of motions with the net result of stabilizing folded rather than unfolded structures. As more examples appear we should have a better idea of some of the unusual properties of some osmolytes. For example, modified polyols and carbohydrates are often used by cells that grow at high temperatures as well as high salt. Do these solutes alter solvent structure in a uniform way? Are they better than other solutes in aiding as chemical chaperones for protein folding/refolding?
Regardless of the details of osmolyte biosynthesis and interactions, it is clear that there is a use for these solutes in the biotechnology arena. Stabilization of proteins or enhancing PCR is an obvious application of these solutes, and only the most common solutes have been tried thus far. One might even improve upon the natural solutes with synthetic molecules if one has a firm grasp of how they affect different systems. Some of the more unusual solutes may have particularly interesting properties that could be exploited either in vitro or in vivo. Engineering foreign osmolyte pathways into other cells has not been very successful but such transgenic organisms will certainly be optimized in the future. It will certainly be interesting to see what new information is provided in the next 5 years or so.
Competing interests
The author(s) declare that they have no competing interests.
Acknowledgements
Preparation if this review was supported in part by a grant from the Department of Energy Biosciences DE-FG02-91ER20025.
==== Refs
Galinski EA Osmoadaptation in bacteria Adv Microb Physiol 1995 37 272 328 8540423
Martin DD Ciulla RA Roberts MF Osmoadaptation in archaea Appl Environ Microbiol 1999 65 1815 1825 10223964
Roberts MF Osmoadaptation and osmoregulation in archaea: update 2004 Front Biosci 2004 9 1999 2019 15353266
Brown AD Microbial water stress Bacteriol Rev 1976 40 803 846 1008746
Galinski EA Pfeiffer HP Trüper HG 1,4,5,6-Tetrahydro-2-methyl-4-pyrimidinecarboxylic acid. A novel cyclic amino acid from halophilic phototrophic bacteria of the genus Ectothiorhodospira Eur J Biochem 1985 149 135 139 10.1111/j.1432-1033.1985.tb08903.x 3838936
Robertson DE Lesage S Roberts MF β-Aminoglutaric acid is a major soluble component of Methanococcus thermolithotrophicus Biochim Biophys Acta 1989 992 320 326 2775788
Sowers KR Robertson DE Noll D Gunsalus RP Roberts MF Nε-acetyl-β-lysine: an osmolyte synthesized by methanogenic bacteria Proc Natl Acad Sci USA 1990 87 9083 9087 2123548
Lai MC Sowers KR Robertson DE Roberts MF Gunsalus RP Distribution of compatible solutes in the halophilic methanogenic archaebacteria J Bacteriol 1991 173 5352 5358 1909318
Scholz S Sonnenbichler J Schäfer W Hensel R Di-myo-inositol-1,1'-phosphate: a new inositol phosphate isolated from Pyrococcus woesei FEBS Lett 1992 306 239 242 10.1016/0014-5793(92)81008-A 1633880
Ciulla R Burggraf S Stetter KO Roberts MF Occurrence and role of di-myo-inositol-1,1'-phosphate in Methanococcus igneus Appl Environ Microbiol 1994 60 3660 3664 16349412
Fan TW-M Metabolite profiling by one and two-dimensional NMR analysis of complex mixtures Progr Nucl Magn Res Spec 1997 28 161 219
Motta A Romano I Gambacorta A Rapid and sensitive NMR method for osmolyte determination J Microbiol Methods 2004 58 289 294 10.1016/j.mimet.2004.04.012 15234528
Kets EPW de Bont JAM Heipieper HJ Physiological response of Pseudomonas putida S12 subjected to reduced water activity FEMS Microbiol Lett 1996 139 133 137 10.1016/0378-1097(96)00131-0
Kunte HJ Galinski EA Trüper HG A modified FMOC-method for the detection of amino acid-type osmolytes and tetrahydropyrimidines (ectoines) J Microbiol Methods 1993 17 129 136 10.1016/0167-7012(93)90006-4
Riis V Maskow T Babel W Highly sensitive determination of ectoine and other compatible solutes by anion-exchange chromatography and pulsed amperometric detection Anal Bioanal Chem 2003 377 203 207 10.1007/s00216-003-2085-9 12851735
Imhoff JF Rodriguez-Valera F Betaine is the main compatible solute of halophilic eubacteria J Bacteriol 1984 160 478 479 6148337
Robertson DE Noll D Roberts MF Menaia JA Boone DR Detection of the osmoregulator betaine in methanogens Appl Environ Microbiol 1990 56 563 565 2306094
Nyyssölä A Kerovuo J Kaukinen P von Weymarn N Reinikaiuem T Extreme halophiles synthesize betaine from glycine by methylation J Biol Chem 2000 275 22196 22201 10.1074/jbc.M910111199 10896953
Canovas D Vargas C Csonka LN Ventosa A Nieto JJ Synthesis of glycine betaine from exogenous choline in the moderately halophilic bacterium Halomonas elongata Appl Environ Microbiol 1998 64 4095 4097 9758852
Roberts MF Lai MC Gunsalus RP Biosynthetic pathway of the osmolytes Nε-acetyl-β-lysine, β-glutamine, and betaine in Methanohalophilus strain FDF1 suggested by nuclear magnetic resonance analyses J Bacteriol 1992 174 6688 6693 1400220
Ciulla RA Diaz MR Taylor BF Roberts MF Organic osmolytes in aerobic bacteria from Mono Lake, an alkaline, moderately hypersaline environment Appl Environ Microbiol 1997 63 220 226 16535487
Doronina NV Trotsenko YA Tourova TP Methylarcula marina gen. nov., sp. nov. and Methylarcula terricola sp. nov.: novel aerobic, moderately halophilic, facultatively methylotrophic bacteria from coastal saline environments Int J Syst Evol Microbiol 2000 50 1849 1859 11034496
Doronina NV Darmaeva TD Trotsenko YA Methylophaga alcalica sp. nov., a novel alkaliphilic and moderately halophilic, obligately methylotrophic bacterium from an East Mongolian saline soda lake Int J Syst Evol Microbiol 2003 53 223 229 10.1099/ijs.0.02267-0 12656177
Kuhlmann AU Bremer E Osmotically regulated synthesis of the compatible solute ectoine in Bacillus pasteurii and related Bacillus spp Appl Environ Microbiol 2002 68 772 783 10.1128/AEM.68.2.772-783.2002 11823218
Onraedt A Walcarius B Soetaert W Vandamme EJ Dynamics and optimal conditions of intracellular ectoine accumulation in Brevibacterium sp Commun Agric Appl Biol Sci 2003 68 241 246 15296171
Regev R Peri I Gilboa H Avi-Dor Y 13C NMR study of the interrelation between synthesis and uptake of compatible solutes in two moderately halophilic eubacteria, Bacterium Ba1 and Vibro costicola Arch Biochem Biophys 1990 278 106 112 10.1016/0003-9861(90)90237-S 2321951
Onraedt A De Muynck C Walcarius B Soetaert W Vandamme E Ectoine accumulation in Brevibacterium epidermis Biotechnol Lett 2004 26 1481 1485 10.1023/B:BILE.0000044448.86907.e4 15604783
Pflughoeft KJ Kierek K Watnick PI Role of ectoine in Vibrio cholerae osmoadaptation Appl Environ Microbiol 2003 69 5919 5927 10.1128/AEM.69.10.5919-5927.2003 14532045
Sowers KR Gunsalus RP Halotolerance in Methanosarcina spp.: role of Nε-acetyl-β-lysine, β-glutamate, glycine betaine and K+ as compatible solutes for osmotic adaptation Appl Environ Microbiol 1995 61 4382 4388 16535193
Robertson DE Lai M-C Gunsalus RP Roberts MF Composition, variation, and dynamics of major compatible solutes in Methanohalophilus strain FDF1 Appl Environ Microbiol 1992 58 2438 2443 16348748
Robertson DE Noll D Roberts MF Free amino acid dynamics in marine methanogens. β-Amino acids as compatible solutes J Biol Chem 1992 267 14893 14901 1353078
Borowitzka LJ Brown AD The salt relations of marine and halophilic species of the unicellular green alga, Dunaliella. The role of glycerol as a compatible solute Arch Mikrobiol 1974 96 37 52 4152062
Petrovic U Gunde-Cimerman N Plemenitas A Cellular responses to environmental salinity in the halophilic black yeast Hortaea werneckii Mol Microbiol 2002 45 665 672 10.1046/j.1365-2958.2002.03021.x 12139614
Silva Z Borges N Martins LO Wait R da Costa MS Santos H Combined effect of the growth temperature and salinity of the medium on the accumulation of compatible solutes by Rhodothermus marinus and Rhodothermus obamensis Extremophiles 1999 3 163 72 10.1007/s007920050112 10357003
Roder A Hoffmann E Hagemann M Berg G Synthesis of the compatible solutes glucosylglycerol and trehalose by salt-stressed cells of Stenotrophomonas strains FEMS Microbiol Lett 2005 243 219 226 10.1016/j.femsle.2004.12.005 15668022
Nyyssölä A Leisola M Actinopolyspora halophila has two separate pathways for betaine synthesis Arch Microbiol 2001 176 294 300 10.1007/s002030100325 11685374
Welsh DT Lindsay YE Caumette P Herbert RA Hannan J Identification of trehalose and glycine betaine as compatible solutes in the moderately halophilic sulfate reducing bacterium, Desulfovibrio halophilus FEMS Microbiol Lett 1996 140 203 207 10.1016/0378-1097(96)00180-2
Reed RH Borowitzka LJ Mackay MA Chudek JA Foster R Warr SRC Moore DJ Steart WDP Organic solute accumulation in osmotically stressed cyanobacteria FEMS Microbiol Rev 1986 39 51 56 10.1016/0378-1097(86)90060-1
Lunn JE Price GD Furbank RT Cloning and expression of a prokaryotic sucrose-phosphate synthase gene from the cyanobacterium Synechocystis sp. PCC 6803 Plant Mol Biol 1999 40 297 305 10.1023/A:1006130802706 10412908
Deplats P Folco E Salerno GL Sucrose may play an additional role to that of an osmolyte in Synechocystis sp.PCC 6803 salt-shocked cells Plant Physiol Biochem 2005 43 133 138 10.1016/j.plaphy.2005.01.008 15820660
Lunn JE Evolution of sucrose synthesis Plant Physiol 2002 128 1490 1500 10.1104/pp.010898 11950997
Mikkat S Hagemann M Schnoor A Active transport of glucosylglycerol is involved in salt adaptation of he cyanobacterium Synechocystis sp. strain PCC 6803 Microbiology 1996 142 1725 1732 8757737
Curatti L Porchia AC Herrera-Estrella L Salerno GL A prokaryotic sucrose synthase gene (susA) isolated from a filamentous nitrogen-fixing cyanobacterium encodes a protein similar to those of plants Planta 2000 211 729 735 10.1007/s004250000343 11089687
Galinski EA Oren A Isolation and structure determination of a novel compatible solute from the moderately halophilic purple sulfur bacterium Ectothiorhodospira marismortui Eur J Biochem 1991 198 593 598 10.1111/j.1432-1033.1991.tb16055.x 2050139
Smith LT Smith GM An osmoregulated dipeptide in stressed Rhizobium meliloti J Bacteriol 1989 171 4714 4717 2768187
D'Souza-Ault MR Smith LT Smith GM Roles of N-acetylglutaminylglutamine amide and glycine betaine in adaptation of Pseudomonas aeruginosa to osmotic stress Appl Environ Microbiol 1993 59 473 478 8434912
Roberts MF Choi BS Robertson DE Lesage S Free amino acid turnover in methanogens measured by 15N NMR spectroscopy J Biol Chem 1990 265 18207 18212 2211697
Martin DD Ciulla RA Robinson PM Roberts MF Switching osmolyte strategies: response of Methanococcus thermolithotrophicus to changes in external NaCl Biochim Biophys Acta 2001 1524 1 10 11078952
Robertson DE Roberts MF Belay N Stetter KO Boone DR Occurrence of β-glutamate, a novel osmolyte, in marine methanogenic bacteria Appl Environ Microbiol 1990 56 1504 1508 2339900
Ciulla RA Roberts MF Effects of osmotic stress on Methanococcus thermolithotrophicus: 13C-edited 1H-NMR studies of osmolyte turnover Biochim Biophys Acta 1999 1427 193 204 10216236
DasSarma S Arora P Halophiles Encyclopedia of Life Sciences 2002 8 Nature Publishing Group 458 466
Martin DD Bartlett DH Roberts MF Solute accumulation in the deep-sea bacterium Photobacterium profundum Extremophiles 2002 6 507 514 10.1007/s00792-002-0288-1 12486460
Lamosa P Burke A Peist R Huber R Liu MY Silva G Rodrigues-Pousada C LeGall J Maycock C Santos H Thermostabilization of proteins by diglycerol phosphate, a new compatible solute from the hyperthermophile Archaeoglobus fulgidus Appl Environ Microbiol 2000 66 1974 1979 10.1128/AEM.66.5.1974-1979.2000 10788369
Lamosa P Martins LO da Costa MS Santos H Effects of temperature, salinity, and medium composition on compatible solute accumulation by Thermococcus spp Appl Environ Microbiol 1998 64 3591 3598 9758772
Martins LO Carreto LS da Costa MS Santos H New compatible solutes related to di-myo-inositol-phosphate in members of the order Thermotogales J Bacteriol 1996 178 5644 5651 8824608
Da Costa MS Santos H Galinski EA An overview of the role and diversity of compatible solutes in Bacteria and Archaea Adv Biochem Eng Biotechnol 1998 61 117 153 9670799
Santos H da Costa MS Compatible solutes of organisms that live in hot saline environments Environ Microbiol 2002 4 501 509 10.1046/j.1462-2920.2002.00335.x 12220406
Martins LO Huber R Huber H Stetter KO da Costa MS Santos H Organic solutes in hyperthermophilic Archaea Appl Environ Microbiol 1997 63 896 902 16535556
Kanodia S Roberts MF Methanophosphagen: unique cyclic pyrophosphate isolated from Methanobacterium thermoautotrophicum Proc Natl Acad Sci USA 1983 80 5217 5221 16593359
Seeley RJ Farney DE A novel diphospho-P, P'-diester from Methanobacterium thermoautrophicum J Biol Chem 1983 258 10835 10838 6885803
Ciulla R Clougherty C Belay N Krishnan S Zhou C Byrd D Roberts MF Halotolerance of Methanobacterium thermoautotrophicum ΔH and Marburg J Bacteriol 1994 176 3177 3187 8195071
Tolman CJ Kanodia S Daniels L Roberts MF 31 P NMR spectra of methanogens: 2,3-cyclophosphoglycerate is detectable only in methanobacteria strains Biochim Biophys Acta 886 345 352 3754772
Gorkovenko A Roberts MF Cyclic 2,3-diphosphoglycerate as a component of a new branch in gluconeogensis in Methanobacterium thermoautotrophicum ΔH J Bacteriol 1993 175 4087 4095 8320225
Evans JNS Tolman CJ Kanodia S Roberts MF 2,3-Cyclopyrophosphate in methanogens: evidence by 13C NMR spectroscopy for a role in carbohydrate metabolism Biochemistry 1985 24 5693 5698 10.1021/bi00342a001 3841287
Desmarais D Jablonski PE Fedarko NS Roberts MF 2-Sulfotrehalose, a novel osmolyte in haloalkaliphilic archaea J Bacteriol 1997 179 3146 3153 9150208
Domenech P Reed MB Dowd CS Manea C Kaplan G Barry CE The role of MmpL8 in sulfatide biogenesis and virulence of Mycobacterium tuberculosis J Biol Chem 2004 279 21257 21265 10.1074/jbc.M400324200 15001577
Oren A Diversity of halophilic microorganisms: environments, phylogeny, physiology, and applications J Ind Microbiol Biotechnol 2002 28 56 63 10.1038/sj.jim.7000176 11938472
Oren A Heldal M Norland S Galinski EA Intracellular ion and organic solute concentrations of the extremely halophilic bacterium Salinibacter ruber Extremophiles 2002 6 491 498 10.1007/s00792-002-0286-3 12486458
Ventosa A Nieto JJ Oren A Biology of moderately halophilic aerobic bacteria Microbiol Mol Biol Rev 1998 62 504 544 9618450
Rengpipat S Lowe SE Zeikus JG Effect of extreme salt concentrations on the physiology and biochemistry of Halobacteroides acetoethylicus J Bacteriol 1988 170 3065 3071 3290195
Heldal M Norland S Erichsen ES Bratbak G Sandaa RA Larsen A Thingstad F Carbon limited marine bacteria keep high internal Mg2+ concentrations 2005 submitted for publication
Müller V Oren A Metabolism of chloride in halophilic prokaryotes Extremophiles 2003 7 261 266 10.1007/s00792-003-0332-9 12728360
Ede SM Hafner LM Fredericks PM Structural changes in the cells of some bacteria during population growth: a Fourier transform infrared-attenuated total reflectance study Appl Spectrosc 2004 58 317 322 10.1366/000370204322886672 15035713
Robinson PM Roberts MF Effects of osmolyte precursors on the distribution of compatible solutes in Methanohalophilus portucalensis Appl Environ Microbiol 1997 63 4032 4038 16535715
Rozwadowski KL Khachatourians GG Selvaraj G Choline oxidase, a catabolic enzyme in Arthrobacter pascens, facilitates adaptation to osmotic stress in Escherichia coli J Bacteriol 1991 173 472 478 1987142
Rathinasabapthi B Burnet M Russell BL Gage DA Liao PC Nye GJ Scott P Golbeck JH Hanson AD Choline monooxygenase, an unusual iron-sulfur enzyme catalyzing the first step of glycine betaine synthesis in plants: prosthetic group characterization and cDNA cloning Proc Natl Acad Sci USA 1997 94 3454 3458 10.1073/pnas.94.7.3454 9096415
Andresen PA Kaasen I Styrvold OB Boulnois G Strom AR Molecular cloning, physical mapping and expression of the bet genes governing the osmoregulatory choline-glycine betaine pathway of Escherichia coli J Gen Microbiol 1988 134 1737 1746 3065456
Russell R Scopes RK Use of hydrophobic chromatography for purification of the membrane-located choline dehydrogenase from a Pseudomonas strain Bioseparation 1994 4 279 284 7765090
Gadda G McAllister-Wilkins EE Cloning, expression, and purification of choline dehydrogenase form the moderate halophile Halomonas elongata Appl Environ Microbiol 2003 69 2126 2132 10.1128/AEM.69.4.2126-2132.2003 12676692
Nyyssölä A Reinikaiuem T Leisola M Characterization of glycine sarcosine N-methyltransferase and sarcosine dimethylglycine N-methyltransferase Appl Environ Microbiol 2001 67 2044 2050 10.1128/AEM.67.5.2044-2050.2001 11319079
Waditee R Tanaka Y Aoki K Hibino T Jikuya H Takano J Takabe T Takabe T Isolation and functional characterization of N-methyltransferases that catalyze betaine synthesis from glycine in a halotolerant photosynthetic organism Aphanothece halophytica J Biol Chem 2003 278 4932 49432 10.1074/jbc.M210970200 12466265
Lai M-C Yang D-R Chuang M-J Regulatory factors associated with synthesis of the osmolyte glycine betaine in the halophilic methanoarchaeon Methanohalophilus portucalensis Appl Environ Microbiol 1999 65 828 833 9925623
Peters P Galinski EA Trüper HG The biosynthesis of ectoine FEMS Microbiol Lett 1990 71 157 162 10.1016/0378-1097(90)90049-V
Canovas D Vargas C Iglesias-Guerra F Csonka LN Rhodes D Ventosa A Nieto JJ Isolation and characterization of salt-sensitive mutants of the moderate halophile Halomonas elongata and cloning of the ectoine synthesis genes J Biol Chem 1997 272 25794 25801 10.1074/jbc.272.41.25794 9325308
Canovas D Vargas C Calderon MI Ventosa A Nieto JJ Characterization of the genes for the biosynthesis of the compatible solute ectoine in the moderately halophilic bacterium Halomonas elongata DSM 3043 Syst Appl Microbiol 1998 21 487 497 9924816
Ono H Sawada K Khunajakr N Tao T Yamamoto M Hiramoto M Shinmyo A Takano M Murooka Y Characterization of biosynthetic enzymes for ectoine as a compatible solute in a moderately halophilic eubacterium, Halomonas elongata J Bacteriol 1999 181 91 99 9864317
Calderon MI Vargas C Rojo F Iglesias-Guerra F Csonka LN Ventosa A Nieto JJ Complex regulation of the synthesis of the compatible solute ectoine in the halophilic bacterium Chromohalobacter salexigens DSM 3043T Microbiol 2004 150 3051 3063 10.1099/mic.0.27122-0
Louis P Galinski EA Characterization of genes for the biosynthesis of the compatible solute ectoine from Marinococcus halophilus and osmoregulated expression in Escherichia coli Microbiology 1997 143 1141 1149 9141677
Canovas D Borges N Vargas C Ventosa A Nieto JJ Santos H Role of Nγ-acetyldiaminobutyrate as an enzyme stabilizer and an intermediate in the biosynthesis of hydroxyectoine Appl Environ Microbiol 1999 65 3774 3779 10473374
Pflüger K Baumann S Gottschalk G Lin W Santos H Müller V Lysine-2,3-aminomutase and β-lysine acetyltransferase genes of methanogenic archaea are salt induced and are essential for the biosynthesis of Nε-acetyl-β-lysine and growth in high salinity Appl Environ Microbiol 2003 69 6047 6055 10.1128/AEM.69.10.6047-6055.2003 14532061
Robinson P Neelon K Schreier HJ Roberts MF β-Glutamate as a substrate for glutamine synthetase Appl Environ Microbiol 2001 67 4458 4463 10.1128/AEM.67.10.4458-4463.2001 11571143
Chen L Spiliotis ET Roberts MF Biosynthesis of di-myo-inositol-1,1'-phosphate, a novel osmolyte in hyperthermophilic archaea J Bacteriol 1998 180 3785 3792 9683472
Scholz S Wolff S Hensel R The biosynthesis pathway of di-myo-inositol-1, 1'-phosphate in Pyrococcus woesei FEMS Microbiol Lett 1998 168 37 42 10.1016/S0378-1097(98)00415-7
Stec B Yang H Johnson KA Chen L Roberts MF MJ0109 is an enzyme that is both an inositol monophosphatase and the 'missing' archaeal fructose-1,6-bisphosphatase Nat Struct Biol 2000 7 1046 1050 10.1038/80968 11062561
Rashid N Imanaka H Kanai T Fukui T Atomi H Imanaka T A novel candidate for the true fructose-1,6-bisphosphatase in archaea J Biol Chem 2002 277 30649 30655 10.1074/jbc.M202868200 12065581
Sato T Imanaka H Rashid N Fukui T Atomi H Imanaka T Genetic evidence identifying the true gluconeogenic fructose-1,6-bisphosphatase in Thermococcus kodakaraensis and other hyperthermophiles J Bacteriol 2004 186 5799 5807 10.1128/JB.186.17.5799-5807.2004 15317785
Chen L Roberts MF Cloning and expression of the inositol monophosphatase gene from Methanococcus jannaschii and characterization of the enzyme Appl Environ Microbiol 1998 64 2609 2615 9647837
Stieglitz KA Seaton BA Head JF Stec B Roberts MF Unexpected similarity in regulation between an archaeal inositol monophosphatase/fructose bisphosphatase and chloroplast fructose bisphosphatase Protein Sci 2003 12 760 767 10.1110/ps.0236003 12649434
Martins LO Empadinhas N Marugg JD Miguel C Ferreira C da Costa MS Santos H Biosynthesis of mannosylglycerate in the thermophilic bacterium Rhodothermus marinus. Biochemical and genetic characterization of a mannosylglycerate synthase J Biol Chem 1999 274 35407 35414 10.1074/jbc.274.50.35407 10585410
Empadinhas N Marugg JD Borges N Santos H da Costa MS Pathway for the synthesis of mannosylglycerate in the hyperthermophilic archaeon Pyrococcus horikoshii. Biochemical and genetic characterization of key enzymes J Biol Chem 2001 276 43580 43588 10.1074/jbc.M108054200 11562374
Borges N Marugg JD Empadinhas N da Costa MS Santos H Specialized roles of the two pathways for the synthesis of mannosylglycerate in osmoadaptation and thermoadaptation of Rhodothermus marinus J Biol Chem 2004 279 9892 9898 10.1074/jbc.M312186200 14701859
Lehmacher A Hensel R Cloning, sequencing and expression of the gene encoding 2-phosphoglycerate kinase from Methanothermus fervidus Mol Gen Genet 1994 242 163 168 10.1007/BF00391009 8159166
Van Alebeek GJWM Tafazzul G Kreuwels MJJ Keltjens JT Vogels GD Cyclic 2,3-diphosphoglycerate metabolism in Methanobacterium thermoautotrophoicum ΔH Arch Microbiol 1994 162 193 198 10.1007/s002030050124
Matussek K Moritz P Brunner N Eckerskorn C Hensel R Cloning, sequencing, and expression of the gene encoding cyclic 2, 3-diphosphoglycerate synthetase, the key enzyme of cyclic 2, 3-diphosphoglycerate metabolism in Methanothermus fervidus J Bacteriol 1998 180 5997 6004 9811660
Van Alebeek GJWM Kreuwels MJJ Keltjens JT Vogels GD Methanobacterium thermoautotrophicum ΔH contains a membrane-bound cyclic 2,3-diphosphoglycerate hydrolase Arch Microbiol 1994 161 514 520 10.1007/s002030050090
Sastry MVK Robertson DE Moynihan A Roberts MF Enzymatic degradation of cyclic 2,3-diphosphoglycerate to 2,3-diphosphoglycerate in Methanobacterium thermoautotrophicum Biochemistry 1992 31 2926 2935 10.1021/bi00126a012 1550819
Poolman B Spitzer JJ Wood JM Bacterial osmosensing: roles of membrane structure and electrostatics in lipid-protein and protein-protein interactions Biochim Biophys Acta 2004 1666 88 104 15519310
Baliarda A Robert H Jebbar M Blanco C Le Marrec C Isolation and characterization of ButA, a secondary glycine betaine transport system operating in Tetragenococcus halophila Curr Microbiol 2003 47 347 351 10.1007/s00284-002-4017-x 14629018
Vermeulen V Kunte HJ Marinococcus halophilus DSM 20408T encodes two transporters for compatible solutes belonging to the betaine-carnitine-choline transporter family: identification and characterization of ectoine transporter EctM and glycine betaine transporter BetM Extremophiles 2004 8 175 184 10.1007/s00792-004-0375-6 14872322
Bayles DO Wilkinson BJ Osmoprotectants and cryoprotectants for Listeria monocytogenes Lett Appl Microbiol 2000 30 23 27 10.1046/j.1472-765x.2000.00646.x 10728555
Steger R Weinand M Kramer R Morbach S LcoP, an osmoregulated betaine/ectoine uptake system from Corynebacterium glutamicum FEBS Lett 2004 573 155 160 10.1016/j.febslet.2004.07.067 15327991
Proctor LM Lai R Gunsalus RP The methanogenic archaeon Methanosarcina thermophila TM-1 possesses a high-affinity glycine betaine transporter involved in osmotic adaptation Appl Environ Microbiol 1997 63 2252 2257 9172344
Roessler M Pflüger K Flach H Lienard T Gottschalk G Müller V Identification of a salt-induced primary transporter for glycine betaine in the methanogen Methanosarcina mazei Gö1 Appl Environ Microbiol 2002 68 2133 2139 10.1128/AEM.68.5.2133-2139.2002 11976081
Schiefner A Holtmann G Diederichs K Welte W Bremer E Structural basis for the binding of compatible solutes by ProX from the hyperthermophilic archaeon Archaeoglobus fulgidus J Biol Chem 2004 279 48270 48281 10.1074/jbc.M403540200 15308642
Nagata S Wang YB Interrelation between synthesis and uptake of ectoine for the growth of the halotolerant Brevibacterium species JCM 6894 at high osmolarity Microbios 2001 104 7 15 11229659
Grammann K Volke A Kunte HJ New type of osmoregulated solute transporter identified in halophilic members of the bacteria domain: TRAP transporter TeaABC mediates uptake of ectoine and hydroxyectoine in Halomonas elongata DSM 2581(T) J Bacteriol 2002 184 3078 3085 10.1128/JB.184.11.3078-3085.2002 12003950
Jebbar M Sohn-Bosser L Bremer E Bernard T Blanco C Ectoine-induced proteins in Sinorhizobium meliloti include an ectoine ABC-type transporter involved in osmoprotection and ectoine catabolism J Bacteriol 2005 187 1293 1304 10.1128/JB.187.4.1293-1304.2005 15687193
Kokoeva MV Storch K-F Klein C Oesterhelt D A novel mode of sensory transduction in archaea: binding protein-mediated chemotaxis towards osmoprotectants and amino acids EMBO J 2002 21 2312 2322 10.1093/emboj/21.10.2312 12006484
Neelon K Roberts MF Organic Solute Pools of Halobacterium NRC-1: Dependence on Growth Conditions and Response to Stress Extremophiles 2005 in press
Zhou Y Morais-Cabral JH Kaufman A MacKinnon R Chemistry of ion coordination and hydration revealed by a K+ channel-Fab complex at 2.0Å resolution Nature 2001 414 43 48 10.1038/35102009 11689936
Jiang Y Lee A Chen J Ruta V Cadene M Chait BT MacKinnon R X-ray structure of a voltage-dependent K+ channel Nature 2003 423 33 41 10.1038/nature01580 12721618
Kraegeloh A Amendt B Kunte HJ Potassium transport in a halophilic member of the Bacteria domain: identification and characterization of the K+ uptake systems TrkH and TrkI from Halomonas elongata DSM 2581T J Bacteriol 2004 187 1036 1043 10.1128/JB.187.3.1036-1043.2005
Edwards MD Booth IR Miller S Gating the bacvterial mechanosensitive channels: MscS a new paradigm Curr Opin Microbiol 2004 7 163 167 10.1016/j.mib.2004.02.006 15063854
Blount P Moe PC Bacterial mechanosensitive channels: integrating physiology, structure and function Trends Microbiol 1999 7 420 424 10.1016/S0966-842X(99)01594-2 10498951
Poolman B Blount P Folgering JHA Friesen RHE Moe PC van der Heide T How do membrane proteins sense water stress? Molec Microbiol 2002 44 889 902 10.1046/j.1365-2958.2002.02894.x 12010487
Perlman DF Goldstein L Organic osmolyte channels in cell volume regulation in vertebrates Exp Zool 1999 283 725 733 10.1002/(SICI)1097-010X(19990601)283:7<725::AID-JEZ10>3.0.CO;2-#
Liu Y Bolen DW The peptide backbone plays a dominant role in protein stabilization by naturally occurring osmolytes Biochemistry 1995 34 12884 12891 10.1021/bi00039a051 7548045
Timasheff SN Protein-solvent preferential interactions, protein hydration, and the modulation of biochemical reactions by solvent components Proc Natl Acad Sci USA 2002 99 9721 9726 10.1073/pnas.122225399 12097640
Timasheff SN Protein hydration, thermodynamic binding, and preferential hydration Biochemistry 2002 41 13473 13482 10.1021/bi020316e 12427007
Plaza di Pino IM Sanchez-Ruiz JM An osmolyte effect on the heat capacity change for protein folding Biochemistry 1995 34 8621 8630 10.1021/bi00027a011 7542026
Bolen DW Baskakov IV The osmophobic effect: natural selection of a thermodynamic force in protein folding J Mol Biol 2001 310 955 963 10.1006/jmbi.2001.4819 11502004
Cioni P Bramanti E Strambini GB Effects of sucrose on the internal dynamics of azurin Biophys J 2005 in press
Kunz W Henle J Ninham BW About the science of the effect of salts: Franz Hofmeister's historical papers Curr Opin Coll Interface Sci 2004 9 19 37 10.1016/j.cocis.2004.05.005
Baldwin RL How Hofmeister interactions affect protein stability Biophys J 1996 71 2056 2063 8889180
Yancey PH Clark ME Hand SC Bowlus RD Somero GN Living with water stress: evolution of osmolyte systems Science 1982 217 1214 1222 7112124
Collins KD Washabaugh MW The Hofmeister effect and the behavior of water at interfaces J Biol Chem 1985 18 323 422
Collins KD Ions from the Hofmeister series and osmolytes: effects on proteins in solution and in the crystallization process Methods 2004 34 300 311 10.1016/j.ymeth.2004.03.021 15325648
Wiggins PM Role of water in some biological processes Microbiol Rev 1990 54 432 449 2087221
Minton AP Effects of excluded surface area and adsorbate clustering on surface adsorption of proteins I. Equilibrium models Biophys Chem 2000 86 239 247 10.1016/S0301-4622(00)00151-4 11026688
Minton AP Effects of excluded surface area and adsorbate clustering on surface adsorption of proteins. II. Kinetic models Biophys J 2001 80 1641 1648 11259279
Hall D Minton AP Macromolecular crowding: qualitative and semiquantitative successes, quantitative challenges Biochim Biophys Acta 2003 1649 127 139 12878031
Robinson CR Sligar SG Hydrostatic and osmotic pressure as tools to study macromolecular recognition Methods Enzymol 1995 259 395 427 8538464
Hummer G Garde S Garcia AE Paulaitis ME Pratt LR The pressure dependence of hydrophobic interactions is consistent with the observed pressure denaturation of proteins Proc Natl Acad Sci USA 1998 95 1552 1555 10.1073/pnas.95.4.1552 9465053
Diamant S Eliahu N Rosenthal D Goloubinoff P Chemical chaperones regulate molecular chaperones in vitro and in cells under combined salt and heat stresses J Biol Chem 2001 276 39586 39591 10.1074/jbc.M103081200 11517217
Faria TQ Lima JC Bastos M Macanita AL Santos H Protein stabilization by osmolytes from hyperthermophiles: effect of mannosylglycerate on the thermal unfolding of recombinant nuclease a from Staphylococcus aureus studied by picosecond time-resolved fluorescence and calorimetry J Biol Chem 2004 279 48680 48691 10.1074/jbc.M408806200 15347691
Ratnaparkhi GS Varadarajan R Osmolytes stabilize ribonuclease S by stabilizing its fragments S protein and S peptide to compact folding-competent states J Biol Chem 2001 276 28789 2898 10.1074/jbc.M101906200 11373282
Singh R Haque I Ahmad F Counteracting osmoyte trimethylamine N-oxide destabilizes proteins at pH below its pKa J Biol Chem 2005 280 11035 11042 10.1074/jbc.M410716200 15653673
Gahl RF Narayan M Xu G Scheraga HA Trimethylamine-N-oxide modulates the reductive unfolding of onconase Biochem Biophys Res Commun 2004 325 707 710 10.1016/j.bbrc.2004.10.088 15541346
Mandal AK Samaddar S Banerjee R Lahiri S Bahttacharyya A Roy S Glutamate counteracts the denaturing effect of urea through its effect on the denatured state J Biol Chem 2003 278 36077 36084 10.1074/jbc.M211207200 12844489
Pais TM Lamosa P dos Santos W Legall W Turner DL Santos H Structural determinants of protein stabilization by solutes. The important of the hairpin loop in rubredoxins FEBS Lett 2005 272 999 1011
Celinski SA Scholtz JM Osmolyte effects on helix formation in peptides and the stability of coiled-coils Protein Sci 2002 11 2048 2051 10.1110/ps.0211702 12142459
Butler SL Falke JJ Effects of protein stabilizing agents on thermal backbone motions: a disulfide trapping study Biochemistry 1996 35 40595 40600 10.1021/bi961107v
Foord RL Leatherbarrow RJ Effects of osmolytes on the exchange rates of backbone amide hydrogens in proteins Biochemistry 1998 37 2969 2978 10.1021/bi9712798 9485449
Qu Y Bolen W Hydrogen exchange kinetics of Rnase A and the urea:TMAO paradigm Biochemistry 2003 42 5837 5849 10.1021/bi0206457 12741842
Kim YS Jones LS Dong A Kendrick BS Chang BS Manning MC Randolph TW Carpenter JF Effects of sucrose on conformational equilibria and fluctuations within the native-site ensemble of proteins Protein Sci 2003 12 1252 1261 10.1110/ps.0242603 12761396
Wright DB Banks DD Lohman JR Hilsenbeck JL Gloss LM The effect of salts on the activity and stability of Escherichia coli and Haloferax volcanii dihydrofolate reductases J Mol Biol 2002 323 327 344 10.1016/S0022-2836(02)00916-6 12381324
Rees WA Yager TD Korte J von Hippel PH Betaine can eliminate the base pair composition dependence of DNA melting Biochemistry 1993 32 137 144 10.1021/bi00052a019 8418834
Henke W Herdel K Jung K Schnorr D Loening SA Betaine improves the PCR amplification of GC-rich DNA sequences Nucl Acids Res 1997 25 3957 3958 10.1093/nar/25.19.3957 9380524
Malin G Iakobashvili R Lapidot A Effect of tetrahydropyrimidine derivatives on protein-nucleic acids interaction. Type II restriction endonucleases as a model system J Biol Chem 1999 274 6920 6929 10.1074/jbc.274.11.6920 10066745
Sauer T Galinski EA Bacterial milking: A novel bioprocess for production of compatible solutes Biotechnol Bioeng 1998 57 306 313 10.1002/(SICI)1097-0290(19980205)57:3<306::AID-BIT7>3.0.CO;2-L 10099207
Barth S Huhn M Matthey B Klimka A Galinski EA Engert A Compatible-solute-supported periplasmic expression of functional recombinant proteins under stress conditions Appl Environ Microbiol 2000 66 1572 1579 10.1128/AEM.66.4.1572-1579.2000 10742244
Arora A Ha C Park CB Inhibition of insulin amyloid formation by small stress molecules FEBS Lett 2004 564 121 125 10.1016/S0014-5793(04)00326-6 15094052
Schnoor M Voss P Cullen P Boking T Galla HJ Galinski EA Lorkowski S Characterization of the synthetic compatible solute homoectoine as a potent PCR enhancer Biochem Biophys Res Commun 2004 322 867 872 10.1016/j.bbrc.2004.07.200 15336543
Cleland D Krader P McCree C Tang J Emerson D Glycine betaine as a cryoprotectant for prokaryotes J Microbiol Methods 2004 58 31 38 10.1016/j.mimet.2004.02.015 15177901
Bünger J Driller H Ectoin: An Effective Natural Substance to Prevent UVA-Induced Premature Photoaging Skin Pharmacol Physiol 2004 17 232 237 10.1159/000080216 15452409
Alia HH Sakamoto A Murata N Enhancement of the tolerance of Arabidopsis to high temperatures by genetic engineering of the synthesis of glycinebetaine Plant J 1998 16 155 161 10.1046/j.1365-313x.1998.00284.x 9839462
Holmstrom KO Somersalo S Mandal A Palva TE Welin B Improved tolerance to salinity and low temperature in transgenic tobacco producing glycine betaine J Exp Bot 2000 51 177 185 10.1093/jexbot/51.343.177 10938824
Nakayama H Yoshida K Ono H Murooka Y Shinmyo A Ectoine, the compatible solute of Halomonas elongata, confers hyperosmotic tolerance in cultured tobacco cells Plant Physiol 2000 122 1239 1247 10.1104/pp.122.4.1239 10759521
Rontein D Basset G Hanson AD Metabolic engineering of osmoprotectant accumulation in plants Metab Eng 2002 4 49 56 10.1006/mben.2001.0208 11800574
Brown AD Compatible solutes and extreme water stress in eukaryotic micro-organisms Adv Microb Physiol 1978 17 181 242 352106
|
16176595
|
PMC1224877
|
CC BY
|
2022-09-08 23:19:43
|
no
|
Saline Syst. 2005 Aug 4; 1:5
|
utf-8
|
Saline Syst
| 2,005 |
10.1186/1746-1448-1-5
|
oa_comm
|
==== Front
Saline Syst
Saline Syst
Saline Systems
1746-1448
BioMed Central
1746-1448-1-6
10.1186/1746-1448-1-6
Book Review
Review of "Adaptation to life at high salt concentrations in Archaea, Bacteria, and Eukarya" Edited by Nina Gunde-Cimerman, Aharon Oren, and Ana Plemenitaš
Rensing Christopher [email protected]
1 Department of Soil, Water, and Environmental Science, University of Arizona, Tucson, AZ 85721, USA
2005
17 8 2005
1 66
27 7 2005
17 8 2005
Copyright © 2005 Rensing; licensee BioMed Central Ltd.
2005
Rensing; licensee BioMed Central Ltd.
https://creativecommons.org/licenses/by/2.0/ This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
The diversity of hypersaline environments and the physiology of representative organisms are only beginning to be understood. Recent progress in this area is documented in "Adaptation to life at high salt concentrations in Archaea, Bacteria, and Eukarya" – eds. Nina Gunde-Cimerman, Aharon Oren and Ana Plemenitas. The 34 chapters successfully paint a fascinating emerging picture of these environments and the microorganisms inhabiting them.
==== Body
pmcBook details
Adaption to life at high salt concentrations in Archaea, Bacteria and Eukarya. Series: Cellular Origin, Life in Extreme Habitatsand Astrobiology, Vol 9 2005. Hardcover ISBN: 1-4020-3632-9
Starting a new on-line journal such as Saline Systems presents the opportunity for editors and potential readers to familiarize themselves with recent ongoing research in this rather broad field. What better way to begin than to summarize new results and developments presented at a well-attended and received scientific meeting? This is the idea behind "Adaptations to life at high salt concentrations in Archaea, Bacteria and Eukarya" which explores the many-fold aspects of life under these extreme conditions. The book was initiated at an international conference "Halophiles 2004" in Ljubljana, Slovenia, in September 2004 organized by Nina Gunde-Cimerman and Ana Plemenitas, a conference which has been held roughly every three years since 1978 (Fig. 1). The first conference was held in Rehovot/Israel and organized by Roy Caplan and Margaret Ginzburg. A comparison of topics covered by conferences on halophilic microorganisms twenty years ago and more recently reflects how rapidly this field has expanded. The 34 chapters capture the current breadth of the field and illustrate the versatility of these microorganisms. Halophilic environments extend from the Great Salt Lake to the Dead Sea but also to unexpected environments such as arctic regions, deep-sea hypersaline brines and possibly habitats in outer space. The study of extremophiles might also give clues for a better understanding of the origin of life here on earth but possibly also elsewhere hence the link to exobiology explored by Joseph Seckbach and Rocco Mancinelli in this volume.
Figure 1 Participants of the recent international conference "Halophiles 2004" in Ljubljana.
If one were to give out last years Oscars for salt-lovers, "Two square off" would certainly take home the honor. In my and certainly also the media's opinion, the cultivation and initial characterization of Walsby's square archaeon was a major accomplishment. Henk Bolhuis' fun to read article certainly succeeds in capturing the fascination of this seemingly strange but widespread microbe. How did he finally manage to get a hold on 'Haloquadratum walsbyi', nothing fancy, just using agarose instead of agar plates. Seemingly simple, it certainly hides many frustrating hours spent in the lab. Another successful strategy was employed by David Burns out of Mike Dyall-Smith's lab using dilution cultures in natural saltern water with just a few supplements. Interestingly, they both managed to observe large sheets of Walsby's square archaeon, measuring approximately 40 × 40 μm. If these structures constitute single cells, or readily dissolve into multiple square components under the conditions used by D. Burns, this will certainly be an area of future studies.
Studying microbial communities in the natural environment is inherently difficult due to technical and biological aspects. One advantage of microbial ecology studies in extreme environments is the relatively low diversity of the respective microbial communities. It is therefore not surprising that this invitation to study the diversity and composition in hypersaline environments was followed up by a number of scientists and the results from Great Salt Lake, the Dead Sea, salterns in California and Spain and ancient salt deposits summarized here. In addition, often neglected predators such as Haloviruses or Protozoa feeding on halophilic archaea and bacteria are not only described but also evaluated on their impact on microbial communities in three chapters.
The physiology of halophilic and halotolerant organisms is also covered in great detail and insight in this volume. How do they do this, how can any organism survive under these conditions? To survive and thrive in a hypersaline environment, microorganisms needed to develop mechanisms of osmoadaptation and salt tolerance. In nature, different solutions brought about by convergent evolution to a low water activity environment can be found in various organisms able to withstand high salt conditions. Not only haloarchaea, as initially thought, but also bacteria accumulate high internal concentrations of potassium and chloride. An important, environmentally relevant representative, Salinibacter ruber, is described in detail from a genomic and environmental perspective. Another surprising aspect of the physiology of some halophiles is their requirement for chloride. A more detailed account is given in an interesting chapter on chloride requirement in moderately halophilic bacteria by Volker Müller and Stefan Saum. Their initial hypothesis for the necessity of chloride is not as counterion as one might intuitively believe but as an inducer for a novel regulatory network. However, if that were true, wouldn't one expect mutants in the signaling pathway without a requirement for chloride?
Transcriptional regulation in haloarchaea was also described using gas vesicle synthesis as an example. We are clearly only at the beginning of understanding the largely unexplored interplay between bacterial-like repressors and an eukaryae-like transcription machinery. This, I believe, is one of the surprises from the recent or on-going genome sequencing efforts. Brian Berquist and coworkers provide a genome-wide COG-based inventory of regulators, basal transcriptional machinery, and DNA replication and repair systems in the two completely sequenced haloarchaea, Halobacterium sp. NRC-1 and Haloarcula marismortui. This showed nearly all the regulators to be bacterial, along with a few (laterally transferred?) components of the DNA replication and repair systems. These aspects of haloarchaeal genetics are likely to be the subject of significant interest in the near future.
Ana Plemenitas, Nina Gunde-Cimerman and their respective group members have to be commended for their Herculean effort setting the record straight on fungal response to low water activity and their contribution and impact in hypersaline environments in three extensive and well-researched chapters.
Is this a book worth having: by all means YES! Overall it contains a wealth of information that is really useful having as a resource. In addition, research strategies and thoughtful discussions are presented in chapters such as Diversity of Microbial Communities: The Case of Solar Salterns by Carlos Pedros-Alio. This methodology is often at least as important as mere facts and findings, especially for the hopefully intended target audience of graduate students and postdoctoral researchers.
|
0
|
PMC1224878
|
CC BY
|
2022-09-08 23:19:43
|
no
|
Saline Syst. 2005 Aug 17; 1:6
|
utf-8
|
Saline Syst
| 2,005 |
10.1186/1746-1448-1-6
|
oa_comm
|
==== Front
Saline Syst
Saline Syst
Saline Systems
1746-1448
BioMed Central
1746-1448-1-7
16176596
10.1186/1746-1448-1-7
Research
Metaproteomic analysis of Chesapeake Bay microbial communities
Kan Jinjun [email protected]
Hanson Thomas E [email protected]
Ginter Joy M [email protected]
Wang Kui [email protected]
Chen Feng [email protected]
1 Center of Marine Biotechnology, University of Maryland Biotechnology Institute, Baltimore, MD 21202, USA
2 Graduate College of Marine Studies and Delaware Biotechnology Institute, University of Delaware, Newark, DE 19711, USA
3 Department of Chemistry and Biochemistry, University of Delaware, Newark, DE 19716, USA
2005
19 8 2005
1 77
17 6 2005
19 8 2005
Copyright © 2005 Kan et al; licensee BioMed Central Ltd.
2005
Kan et al; licensee BioMed Central Ltd.
https://creativecommons.org/licenses/by/2.0/ This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Natural microbial communities are extremely complex and dynamic systems in terms of their population structure and functions. However, little is known about the in situ functions of the microbial communities.
Results
This study describes the application of proteomic approaches (metaproteomics) to observe expressed protein profiles of natural microbial communities (metaproteomes). The technique was validated using a constructed community and subsequently used to analyze Chesapeake Bay microbial community (0.2 to 3.0 μm) metaproteomes. Chesapeake Bay metaproteomes contained proteins from pI 4–8 with apparent molecular masses between 10–80 kDa. Replicated middle Bay metaproteomes shared ~92% of all detected spots, but only shared 30% and 70% of common protein spots with upper and lower Bay metaproteomes. MALDI-TOF analysis of highly expressed proteins produced no significant matches to known proteins. Three Chesapeake Bay proteins were tentatively identified by LC-MS/MS sequencing coupled with MS-BLAST searching. The proteins identified were of marine microbial origin and correlated with abundant Chesapeake Bay microbial lineages, Bacteroides and α-proteobacteria.
Conclusion
Our results represent the first metaproteomic study of aquatic microbial assemblages and demonstrate the potential of metaproteomic approaches to link metagenomic data, taxonomic diversity, functional diversity and biological processes in natural environments.
==== Body
pmcBackground
Bacterioplankton contribute significantly to both primary production and biomass in the ocean and coastal water [1,2]. With an average concentration of approximately 106 cells ml-1, bacterioplankton is an important catalyst of biogeochemical processes including oceanic carbon and nitrogen cycles [3,4]. Studying bacterioplankton is challenging because most groups either have never been cultivated [5,6] or grow to very low density in the laboratory [7]. Culture-independent molecular approaches have indicated that environmental bacterial communities are more complex and diverse than previously thought [5,6,8]. Metagenomics is the direct cloning, sequencing, assembly and annotation of DNA from microbial communities and has been applied to waters, soils and extreme environments [9-12]. A recent metagenomic study of the Sargasso Sea revealed that substantial complex microbial diversity exists in the ocean: 148 novel bacterial phylotypes and more than a million of previously unknown genes were discovered and annotated [12].
As genomic data accumulates from pure cultures and environmental communities, it becomes critical to understand gene expression and protein function. While metagenome sequences provide valuable information on potential functions, accurately predicting ecological function from sequence is nearly impossible without information on what proteins are synthesized under specific conditions [13-15]. To address this question, post-genomic molecular approaches such as microarrays to monitor mRNA abundance [16] have been developed. In addition, as proteins/proteomes are the ultimate functional products of genes/genomes, proteomic studies of microbial communities (metaproteomics) are an obvious approach to advance our understanding of microbial community function.
Metaproteomics can provide a direct measurement of functional gene expression in terms of the presence, relative abundance and modification state of proteins [17,18]. Proteomics and metaproteomics rely on two-dimensional gel electrophoresis (2D-PAGE) coupled with mass spectrometry (MS) based protein identification relying on mass based (MALDI-TOF MS) or sequence based (LC-ESI-MS/MS) methods. These techniques have only been applied in limited scope to environmental microbial communities. One-dimensional gel electrophoresis (1D-PAGE) coupled with radioactive labelling or enzymatic activity assay has been used to study proteins induced in response to environmental stresses [19,20]. However, little concrete information on the sequences or identities of induced proteins emerged from these studies. A metaproteomic approach was applied to a laboratory-scale activated sludge bioreactor resulting in the identification of three highly expressed proteins presumably originating from an uncultured Rhodocyclus-type polyphosphate – accumulating organism [18]. More recently, using genomic and mass spectrometry-based proteomic methods, metaproteomes from an acid mine drainage (AMD) microbial biofilm community have been identified and linked their in situ functions to the challenging environments [21]. However, all these studies are dealing with low-complexity microbial communities. So far, no studies have yet applied proteomic approaches to natural aquatic microbial communities.
Estuaries represent one of the most complex and productive ecosystems. The Chesapeake Bay is the largest estuary in United States (Fig. 1). It has received a great deal of attention because of its large geographic span and economic significance. With strong environmental gradients, it provides an ideal model system for integrated investigations on composition and function of microbial communities. In this study, we developed a metaproteomic approach to document microbial community protein profiles along a transect of the Chesapeake Bay. Significant differences were noted between proteomes collected at different sites and metaproteome patterns accurately predicted the relationship of sites as determined by 16S rRNA gene PCR-DGGE (denaturing gradient gel electrophoresis). Furthermore, proteins identified from Chesapeake Bay samples appeared to originate from marine bacterioplankton. This study demonstrates that metaproteomic approaches can be successfully applied to naturally occurring and complex microbial communities in their native habitats.
Figure 1 Metaproteome sampling stations at the Chesapeake Bay.
Results
Microbial community collection
Epifluorescence microscopic counts showed that concentrated microbial communities mainly contained free-living bacteria (~ 95%). The recovery efficiency of bacterial cells using the tangential flow ultrafiltration system was 75 ± 5% (data not shown). With the average concentration of 2.5 × 106 cells ml-1 in the starting water samples, the density of microbial cells in the ultrafiltration retentate was about 2.5 × 108 cells ml-1. Thus, about 3.75 × 1010 cells were analyzed in each sample. Extracts typically contained between 140 and 192 μg of protein giving a value range of 3.7 × 10-15 to 5.1 × 10-15 g protein cell-1. This value is significantly lower than that determined for cultured strains in this study and in general for marine bacteria (60–330 × 10-15 g protein cell-1, [22]). It remains to be determined whether this discrepancy indicates that the extraction protocol needs further optimization or is a fundamental property of microbial cells in environmental samples.
1D-PAGE analysis of proteins from isolated bacterial strains and environmental samples
Individual proteins from cultivated marine bacteria were well resolved by 1D-PAGE and produced distinct patterns when 8 Chesapeake Bay bacterial isolates were compared (Fig. 2). The observed molecular masses ranged from ~10 to 250 kDa (Fig. 2, lanes 1–8) whereas proteins from microbial community samples were < 80 kDa (Fig. 2, lanes 9 and 10). Overall resolution was much poorer in community samples as evidenced by less sharply defined bands in these samples. This blurring effect was also noted in a very simple mixed microbial community described below and was not dependent on sampling manipulations (data not shown).
Figure 2 1D-PAGE patterns of total proteins obtained from 8 different bacteria isolated from Baltimore Inner Harbor. M, Marker; Mr, molecular weight; Lanes 1 – 8 correspond to Vibrio vulnificus, Marine Bacillus sp., Marinomonas sp., Psychrobacter pacificens, Pseudomonas sp., Pseudoalteromonas sp., Shewanella sp., and Hahella sp.. Lanes 9 and 10 are duplicated environmental microbial communities. For each lane, 20 μg of protein is loaded and the gel is stained by silver staining.
Analysis of isolated strains and artificial mixed communities
Artificial community consisting of Chlorobium tepidum strain WT2321, Escherichia coli strain JM109 and an uncharacterized strain of Pseudomonas fluorescens was analyzed by 2D-PAGE. Preliminary experiments indicated that a 300 ml sample containing 1 × 107 cells per ml of the community could be successfully analyzed by 2D-PAGE. Analysis by 1D-PAGE afforded greater sensitivity, ~1 × 104 cells per ml, but resolution of individual bands was poor as noted above. Protein assays on samples of the community before dilution and recovery and after indicated that the metaproteomic sample preparation recovered ~ 30% of the total microbial protein present in the original community sample.
Typical results from a 2D-PAGE experiment are shown in Fig. 3. The overlays indicate that 2D-PAGE patterns from single strains of community members only match a fraction of protein spots present in the mock metaproteome sample (Fig. 3a–d). This is qualitatively observed as a large number of green or pink protein spots in the overlay views showing unmatched protein spots. Each individual strain is expected to contribute only one third of the protein content of the community. In contrast, when a sample of the community prior to dilution and recovery is compared to a mock metaproteome that had been subjected to sample handling protocols, almost perfect matching of the samples is seen as evidenced by the large proportion of dark grey to black spots (Fig. 3d) when these images are overlain. Thus, no individual member of the community, which covers the range of cell sizes in the environmental samples, is selectively excluded by the sampling protocol.
Figure 3 The harvesting protocol for microbial communities does not bias against different types of bacteria. Proteomes of Chlorobium tepidum (a), Escherichia coli (b) and Pseudomonas fluorescens (c) and the metaproteomes of an artificially constructed community containing all three organisms (d) were overlain and compared to the metaproteomes of the artificial community after dilution and recovery using Compugen Z3 software. Green or pink colored protein spots are unmatched. Gray or black spots are matched. Total 100 μg proteins are loaded on each polyacrylamide gel and the gels are stained by SYPRO Ruby. pI, isoelectric point; Mr, molecular weight.
Extraction of metaproteomes from the Chesapeake Bay
In this study, in order to optimize the protein extraction of aquatic microbial communities, different protocols that varied all steps in protein extraction and purification were tested including (i) sample collection (filtration on membrane filter, tangential flow concentration with centrifugation); (ii) washing buffer to remove ambient salts and polysaccharides; (iii) extraction buffer (standard lysis buffer, SDS-PAGE buffer, urea-thiourea-CHAPS buffer); (iv) reducing agent (dithiothreitol (DTT) vs. tributyl phosphine (TBP));(v) cell lysis method (freeze-thaw, French pressure cell); (vi) protein precipitation (acetone vs. TCA); (vii) IPG strip range (pH 3–10 vs. pH 4–7); and (viii) staining method (Commassie blue, silver, SYPRO Ruby). From these trials, the following protocol emerged: (i) tangential flow concentration with centrifugation; (ii) TS washing buffer (Tris 10 mM, Sucrose 250 mM); (iii) urea-thiourea-CHAPS lysis buffer with TBP; (iv) lysis via French pressure cell; (v) TCA precipitation; (vi) First dimension pH 4–7 IPG strip; (vii) SYPRO Ruby staining. However, given the indigenous characteristics among diverse microbial communities, extraction of metaproteomes may vary by site, time and experiment as well.
Quantitative Comparison of Chesapeake Bay Metaproteome Samples
Metaproteome images from different Chesapeake Bay stations in the upper (station 858), middle (station 804, replicates a and b) and lower Bay (station 707) were compared (Fig. 4a–d). A number of protein spots were shared by all samples. Some of these are proteins present in RNase, Dnase and protease inhibitor cocktail in the extraction buffer (data not shown), but a number of proteins appear to be common in all samples examined. These are black to dark grey spots in the image overlays (Fig. 4a–d). A first level of quantitative comparison determined the specific numbers of protein spots shared between samples (Table 1). The total number of spots compared for each sample is relatively low as the analysis was restricted to spots with sufficient quality and intensify to permit subsequent attempts at protein identification. As expected, replicate metaproteome images from the middle Bay are more similar to one another than the metaproteomes of other stations, sharing ~92 % of all detected spots. Furthermore, the lower and middle Bay metaproteomes are significantly more similar to one another than either is to the upper Bay metaproteomes with ~70 % of all detected spots in common. The upper Bay metaproteomes only shared about ~30 % of detected spots with either the middle or lower Bay metaproteomes.
Figure 4 Comparisons of Chesapeake Bay metaproteomes. (a) Independent samples from Station 804, 804a and 804b; (b) Station 804a vs. Station 707; (c) Station 804a vs. Station 858; (d) Station 707 vs. 858. Image overlays were constructed with Compugen Z3 software. Spots circled in red are unmatched, those in yellow and blue are differentially expressed at a level of ≥ 3-fold between images. No unmatched or differential spots are shown in c and d because software based matching of these images failed. Red marks in panels c and d are alignment points used to produce the pictured overlay. Quantitative results of matching are reported in Table 1. Total 100 μg proteins are loaded on each polyacrylamide gel and the gels are stained by SYPRO Ruby. pI, isoelectric point; Mr, molecular weight.
Table 1 Quantitative comparison of Chesapeake Bay metaproteomes.
Samples compared spotsa unmatcheda differentiala,b
804a 207 7 3
vs. 804b 189 26 13
396 33 (8.3 %) 16(4.0%)
804a 207 37 23
vs. 707 198 86 6
405 123 (30.3 %) 29(7.1 %)
804a 207 156C --d
vs. 858 155 104c --
362 160 (71.8 %) --
707 198 142b --
vs. 858 155 99b --
353 241 (68.3 %) --
a-Spots from first gel, second gel and the sum are listed. Numbers in parentheses show the percentage of the total.
b-Matched spots that are ≥ 3-fold more intense than the comparative image.
c-Estimated by manual comparison of detected spots. Software was unable to match images.
d-No differential comparison possible as software based matching failed.
Relative spot intensity was extracted from comparisons of middle Bay to middle Bay and middle Bay with lower Bay metaproteome images. This was not possible with the upper Bay sample as manual matching was employed due to the low level of similarity between samples. Again, as expected, the number of differentially expressed proteins (≥ 3-fold change in matched spot intensity) was nearly twice as large when comparing middle Bay to a lower Bay metaproteomes as when comparing the replicated middle Bay samples (Table 1). These results indicate that both qualitative and highly quantitative comparisons between sites and between time series samples at the same site will be possible using the approaches developed in this study.
Identification of Proteins in Chesapeake Bay Metaproteomes
A total of 41 protein spots were excised from a number of 2-D gels reflecting various molecular weights, charges and relative abundance. Following MALDI-TOF MS, seven spots failed to yield interpretable MS profiles, while the remaining 34 proteins exhibited clear and distinct MS peaks. Database searches using the MASCOT search engine with varying parameter settings (peptide mass tolerance from 0.5 to 3 Da, missed cleavages from 1 up to 5) produced no significant matches for these 34 proteins. Subsequent publications from other laboratories and our own simulations using known protein sequences [[23,24], Hanson, unpublished data] suggest that greater than 97 % amino acid sequence identity is required to provide a positive match when searching with MALDI-TOF MS data.
Seven individual proteins (Fig. 5) isolated from middle Chesapeake Bay (station 804) metaproteome samples were further analyzed by both MALDI-TOF MS and LC-MS/MS sequencing coupled to MS-BLAST searching (Table 2). MALDI-TOF MS failed to provide identification for any of these samples, similar to the samples described above. LC-MS/MS based searches provided tentative identities for three Chesapeake Bay metaproteome samples. These were identified as homologues of hypothetical proteins annotated in the recently reported Sargasso Sea metagenome [12]. Information on potential functions of these proteins was obtained by downloading the full length proteins from the Sargasso Sea database and searching them against known databases by BLASTP (Table 3). The Sargasso Sea metagenome hypothetical protein corresponding to sample CB1 is not significantly similar to any known proteins in sequence databases. Sample CB3 may correspond to subunit 7 of the NADH:ubiquinone oxidoreductase (complex I) while sample CB6 is similar to a family of predicted aminopeptidases with unspecified functional significance. The tandem mass spectra of samples CB2, CB3 and CB5 had no match with any known proteins or hit keratin and bovine serum albumin that possibly came from background.
Figure 5 Proteins selected for identification from middle Chesapeake Bay (station 804). Total 100 μg protein are loaded on polyacrylamide gel and the gel is stained by SYPRO Ruby. CBl-CB6 samples are common to Chesapeake Bay stations while NC1 is found on negative control gels containing DNase, RNase and protease inhibitors. Results of protein identification are reported in Table 2 and 3. pI, isoelectric point; Mr, molecular weight.
Table 2 Identification of proteins from Chesapeake Bay station 804 metaproteomes (Fig. 5).
Sample pI MW MALDIID?a MS/MS ID?b Peptides Matched Scorec Accession
NC1 5.1 29 kDa No No - - -
CB1 5.3 60 kDa No Sargasso sea metagenome 2 110 EAH98995.1
CB2 4.9 40 kDa No Bovine serum albumin 2 138 P02769
CB3 5.7 42 kDa No Sargasso sea metagenome 3 116 EAH45127.1
CB4 4.4 35 kDa No Keratin 2 117 Q9DCV7
CB5 4.2 33 kDa No No - - - -
CB6 5.0 20 kDa No Sargasso sea metagenome 2 88 EAC65279.1
a-MASCOT search as described in Materials and Methods.
b-MS-BLAST search as described in Materials and Methods.
c-For a description of scoring, see reference 27.
Table 3 BLASTP analysis of Sargasso sea metagenome hits.
Sample Accession Best hit E-value Organism Accession
CB1 EAH98995.1 Hypothetical protein 0.47 Plasmodium berghei CAI00437
CB3 EAH45127.1 NADH:UQ oxidoreductase (49 kDa, subunit 7) 1 × 10-63 Cytophaga hutchinsonii ZP_00309190
CB6 EAC65279.1 Predicted aminopeptidase 2 × 10-16 Novosphingobium aromaticivorans ZP_00305215
Discussion
In this study, we deliberately focused on exploring the proteome profiles from bacterioplankton communities between 0.2 and 3.0 microns in size by the choice of prefiltration and ultrafiltration cut-off sizes. Although the epifluorescence microscopy observation confirmed that the major components are bacterioplankton (~95%), small numbers of eukaryotic microbes were possibly included. These likely did not affect the overall protein profiles observed as analyses were restricted to abundant proteins, which would give the best chance for positive identification.
Metaproteomic approaches have thus far only been applied to laboratory scale bioreactors with a specialized community selected for phosphate removal [18] and a low-complexity natural microbial biofilm [21]. Extending this approach to complex environmental samples was not trivial. Initial studies comparing isolated strains, artificial communities and natural community samples by 1D-PAGE indicated that more resolving power was needed to deal with even simplified communities (data not shown). Thus, a metaproteomic approach utilizing 2D-PAGE and MS based protein identification was adopted. The experimental protocol outlined in this study was designed to avoid metaproteome changes arising from bias in the sample collection or handling. This was tested using artificial constructed bacterial assemblage containing 3 different species with varied cell sizes and we found no significant biases.
The protocol was also field tested by comparing replicated samples from the middle Chesapeake Bay to each other and comparing a range of samples from upper, middle and lower Chesapeake Bay stations. The replicated samples shared more than ~92 % of proteins indicating that the metaproteomic approach applied in this study was robust. Furthermore, significant differences were noted when the middle Bay metaproteomes was compared with lower Bay and upper Bay metaproteomes with only 70 % and 30 % of protein spots in common. This pattern can be likely and partially explained by the difference among the population structures of these samples. Genetic fingerprints indicated that upper Bay bacterioplankton community was different from the middle and lower Bay (Fig. 6). Clustering analysis based on presence/absence of DGGE bands showed that the similarity between middle Bay to lower Bay was 64% while the upper Bay only shared 46% similarity to both of middle Bay and lower Bay. Finally, relative spot abundance was also much more tightly correlated when the replicated middle Bay samples were compared to each other than when they were compared to the lower Bay sample. These results demonstrate the approach outlined here is sufficiently sensitive to detect both coarse (shared spots) and fine (relative spot abundance) quantitative differences between samples, even when relatively low numbers of spots are included in the analysis. This is critical for any comparative approach.
Figure 6 DGGE fingerprints of bacterioplankton communities in Chesapeake Bay. 858, 804 and 707 are sampling stations. M: marker.
This study, in addition to others, indicates that protein identification is the major challenge for metaproteomics [18,21,23,24]. Although distinct mass spectra from 34 protein spots were obtained by MALDI-TOF MS, no significant matches were found in sequence databases. MALDI-TOF generally requires at least 97 % amino acid sequence identity between query and target to find a significant match [[25], Hanson unpublished]. It seems unlikely that many proteins in environmental samples will share this level of identity with proteins in sequence databases derived from cultured organisms. Post-translational modifications of proteins also account for the difficulty in the identifications. Thus, MALDI-TOF MS is unlikely to be useful for metaproteomic approaches.
In contrast, LC-MS/MS or N-terminal sequencing coupled to MS-BLAST searching is able to provide tentative identification for metaproteomes. However, the abundance of most proteins is too low to be identified through the venue of N-terminal sequencing. In the community proteomic analysis of a natural acid mine drainage microbial biofilm, the proteins could be identified by MS and assigned to five most abundant microbes because of the availability of metagenomic data. But the relative high likelihood of false-positive protein identification requires matching of two or more peptides per protein for confident detection [21]. Therefore, caution is required for interpretation of the data. In this study, three Chesapeake Bay metaproteome samples matched different hypothetical proteins annotated in the Sargasso Sea metagenome [12]. This result strongly supports a marine origin for these sequences as would be expected for a large number of proteins in the Chesapeake Bay, particularly in lower and middle Bay samples where there is significant salinity. Even with tentative identities, extending that identity to function must be done with some care. The Sargasso Sea metagenome hypothetical protein corresponding to sample CB1 is not significantly similar to any known proteins in sequence databases giving no clues to its function. Sample CB3 may correspond to subunit 7 of the NADH:ubiquinone oxidoreductase or complex I (Table 3). Complex I is a key component of most membrane bound electron transport chains that is responsible for the transfer of electrons from cytoplasmic NADH pools to the membrane bound quinone pool coupled to proton motive force generation. Subunit 7 is a peripheral membrane protein of the quinone reduction core of complex I [26]. The organism containing the closest match is Cytophaga hutchinsonii, a member of the Bacteroidetes assemblage of organisms, which is a substantial fraction of many marine communities [27]. A current study on population structure of Chesapeake Bay bacterioplankton showed that Bacteroidetes group accounts for ~10% of total community in summer time [e. g. Kan unpublished].
Sample CB6 is similar to a family of predicted aminopeptidases with unspecified functional significance. The closest matching protein is from Novosphingobium aromaticivorans. While N. aromaticivorans is normally considered terrestrial, other Novosphingobium and related Sphingobium and Sphingopyxis strains are widely distributed. As an important component of the α-proteobacteria, these groups can be detected in and isolated from marine and estuarine environments [[28,29], Kan unpublished]. This identification along with that of CB3 support an aquatic bacterial origin for these proteins that is consistent with their presence in the Chesapeake Bay.
Unanswered questions remain regarding the applicability of metaproteomics to natural communities. These include the following: Does a focused protein spot on a 2D SDS-PAGE gel from an environmental sample contain one protein or multiple proteins? What type of information is required to infer identity of spots between different samples? What is the sensitivity of metaproteomics to changes in community composition and the physiological status of community members? How can functional inferences provided by metaproteomics be further tested? Will the approach outlined here be applicable to other systems such as soils, sediments, and extreme environments? Clearly, much more work and complementary approaches need to be applied to these problems.
Conclusion
To our knowledge, this study represents the first application of a metaproteomic approach to a high-complexity aquatic microbial community. The main goals of this study were to develop a method capable of collecting planktonic microbial proteins in quantities suitable for analysis by 2D-PAGE. This was accomplished and attempts were made to identify a subset of these proteins. These attempts reinforced the notion that sequence based methods (LC-MS/MS) will be required to make any headway in protein identification in natural systems. Future studies will identify a much larger number of proteins from Chesapeake Bay microbial communities to address the questions raised above and provide insights into microbial community dynamics and function.
Methods
Bacterial cultures
Eight bacterial strains isolated from upper Chesapeake Bay (Baltimore Inner Harbor) were used in this study. Based on 16S rRNA gene sequences, these bacteria have been identified as Vibrio vulnificus, Marine Bacillus sp., Marinomonas sp., Psychrobacter pacificens, Pseudomonas sp., Pseudoalteromonas sp., Shewanella sp., and Hahella sp. respectively [Kan unpublished]. These bacteria were grown in 1/2 YTSS broth (4 g yeast extract, 2.5 g tryptone per liter dissolved in in situ water) and harvested at the exponential growth stage using centrifugation (10,000 × g, 5 min, 4°C).
Artificial Community Construction and Recovery
To determine if microbial community analysis by 2D SDS-PAGE is feasible and representative, a simple artificial mixed community was constructed using three bacterial strains of differing size: Chlorobium tepidum strain WT2321 (~0.5–0.8 μm cell length), Escherichia coli strain JM109 (~1.2–1.6 μm cell length), and an uncharacterized strain of Pseudomonas fluorescens (~8–10 μm cell length) (kindly provided by G. A. O'Toole, Dartmouth University). Protein content per cell for each strain was determined by measuring protein via a modified Bradford assay (Bio-Rad) and direct cell counting on replicate samples for each organism. Communities containing the same amount of protein for each strain were constructed by mixing appropriate volumes of pure cultures. The mock community was then diluted into 5 1 of 10 mM potassium phosphate buffer (pH = 7.2) to specific cell densities and the cells recovered. Total protein extracts of the mock community and each member strain were made by pelleting cell samples in a microfuge and extracting proteins by resuspending in 5 M urea + 2 M thiourea + 2 % (w/v) CHAPS + 2 % (w/v) SB 3–10 + 40 mM Tris + 0.2 % (w/v) BioLyte 3–10 (sequential extraction reagent 3, Bio-Rad) at room temperature and vortexing for 2 minutes.
Microbial community sampling
Picoplankton communities were collected at three stations along the middle axis of the Chesapeake Bay on 7 June 2003 aboard the R/V Cape Henlopen (Fig. 1). The stations 858, 804 and 707 represent the upper, middle and lower Bay, respectively. At each station, 0.2 g of chloramphenicol (Fisher Scientific, NJ) and 2 ml Protease inhibitor cocktail II (CalBiochem, CA) were added to 20 l of surface water (1 m below) to stop protein synthesis and inhibit activities of proteases. Samples were pre-filtered through 3-μm-pore-size polycarbonate filters (142-mm diameter; Millipore, Bedford, MA) to remove large particles and eukaryotes. The filter was replaced every 5 liters. Microbial cells in the filtrate were concentrated to a final volume of 150 ml using a tangential-flow ultrafiltration (30,000 MW cutoff) as described elsewhere [30]. Duplicate water samples were collected at station 804. Microbial cells in the retentate were pelleted using GS-15R centrifuge (Beckman, Fullerton, CA) at 13,000 × g, 4°C for 10 minutes. The collected cells were rinsed with TS washing buffer (Tris-HCl 10 mM, Sucrose 250 mM, pH 7.6) and resuspended with 0.5 ml of extraction buffer. The extraction buffer consisted of 0.01 M Tris-HCl, pH 7.4, 1 mM EDTA, 7 M urea and 2 M thiourea, 10% (v/v) glycerol, 2 % CHAPS, 0.2 % amphylotes, 0.002 M Tributyl phosphine (TBP), DNase (0.1 mg/ml), RNase (0.025 mg/ml) and proteinase inhibitor cocktail (CalBiochem, CA). TBP, DNase, RNase and proteinase inhibitor cocktail were freshly added to the buffer prior to applying to samples. Cells were stored frozen until further processing.
To estimate the recovery efficiency of ultrafiltration, bacterial cells were counted before and after ultrafiltration. Bacterial cells were stained with SYBR Gold (Molecular Probes, Inc., Eugene, Oreg.) following the protocol described previously [31]. Bacterial cells were enumerated under blue excitation (485 nm) on a Zeiss Axioplan epifluorescence microscope (Zeiss) using 63 × Antiflex Neoflua oil objective lens. At least 200 bacterial cells per sample were counted.
Protein extraction and purification
For 1D-PAGE, proteins from natural microbial communities and cultured bacteria were extracted using lysis buffer (50 mM Tris-HCl, 2% SDS, 10% v/v glycerol, 0.1 M DTT, 0.01% Bromophenol Blue, pH 6.8). Cells suspended in buffer were heated in a boiling water bath for 2 minutes followed by centrifugation (10,000 × g, 4°C for 3 min). The supernatant was collected and 20 μg protein for each was loaded onto polyacrylamide gels. Silver staining was applied to 1D-PAGE gels.
For 2D-PAGE samples, cell suspensions were passed through a French Pressure cell (SLM Aminco) at 20,000 lb/in2 twice and then incubated on ice for 20 minutes. During the ice incubation, samples were vortexed for 15 sec every 5 minutes. Large cellular debris was removed by centrifugation (10,000 × g, 4°C for 5 min). Proteins in the supernatant were precipitated with trichloracetic acid and resuspended in extraction buffer. Protein concentration of the sample was determined using the RC DC protein assay kit (Bio-Rad, Hercules, CA). Extracted proteins were stored at -80°C.
Isoelectric Focusing (IEF) and SDS-PAGE
The first dimension separation of proteins was carried out in the immobilized pH gradient (IPG) strips (11 cm, pH 3–10 or 4–7) on a Bio-Rad Protean IEF Cell system (Bio-Rad, Hercules, CA). Each 2D-PAGE was conducted using 100 μg of total protein. The IEF program was: 250V for 20 min followed with a linear ramp to 8000V for 2.5 hr, and 8000V for a total 40,000 V-hr with a rapid ramp. After the first dimension, the IEF strips were equilibrated in freshly made Buffer 1 (6 M urea, 2% SDS, 0.05 M Tris/HCl pH 8.8, 50% Glycerol) and Buffer 2 (6 M urea, 2% SDS, 0.375 M Tris/HCl pH 8.8, 20% Glycerol and 0.5 g iodoacetamide) (Bio-Rad, Hercules, Calif), respectively.
The second dimension of 2D-PAGE were performed using 8–16% gradient precast polyacrylamide gels (Bio-Rad, Hercules, CA) following the manufacturer's instructions. The gels were stained with SYPRO Ruby (Bio-Rad, Hercules, CA) after electrophoresis and scanned using a Typhoon 9410 fluorescent Imager (Amersham, NJ) with 488nm excitation and emission filter 610 BP30.
Metaproteome Image Analysis
Images were analyzed and quantitatively compared using the Z3 proteomics software package (Compugen, Israel). Gel images were compared in multiple gel mode using the total density in gel method for spot quantification. All gels were subjected to the same spot detection parameters followed by automated matching. Pairwise comparisons of gels were inspected and matches edited manually to eliminate poor quality or low intensity matches. When automatic matching failed, the number of matched and unmatched spots was estimated by manual examination of overlaid 2D SDS-PAGE images.
Protein Identification by Mass Spectrometry
Protein spots were manually excised from gels using Pasteur pipettes and digested as described by Mann et al. [32]. Tryptic peptides were analyzed both via MALDI-TOF and LC-MS/MS. MALDI spectra were acquired on a Bruker (Billerica, MA) Biflex III MALDI mass spectrometer operating in reflectron mode with delayed extraction. External calibration was performed using Calibration Mixture 2 from the Sequazyme Peptide Mass Standards Kit (Applied Biosystems, Foster City, CA). LC-MS/MS was performed on a Micromass (Beverly, MA) Q-TOF Ultima API-US coupled to a Micromass capLC. Tryptic digests were separated using both a C18 trapping column for washing and concentrating (LC Packings (Sunnyvale, CA) 300 μm × 5 mm C18) and a C18 analytical column for enhanced separation (LC Packings 180 μm × 15 cm C18). The solvent system consisted of 95% 0.1% formic Acid, 5% acetonitrile for the aqueous phase and 95% acetonitrile, 5% 0.1% formic Acid for the organic phase. A 60/60 gradient (to 60% organic in 60 mins) running at l μl/min was employed with most peptides eluting by ~30% organic. The LC eluent was electrosprayed directly into the Q-TOF using the nanosprayer source. Data dependent scanning was used with both MS and MS/MS spectra being acquired during an LC run. Spectra were processed and deconvoluted using programs found with the Micromass operating system, MassLynx v. 3.5.
MALDI-TOF peak lists were searched against protein sequence databases using the Matrix Science Mascot web interface http://www.matrixscience.com/search_form_select.html. Deconvoluted MS/MS spectra were analyzed using a demonstration version of PeaksStudio 3.0 software (Bioinformatics Solutions Inc., Canada) for de novo sequence prediction. All sequences for each protein spot were used as queries in MS-BLAST searches as described by Shevchenko et al. [33] via the MS-BLAST web interface http://dove.embl-heidelberg.de/Blast2/msblast.html.
Competing interests
The author(s) declare that they have no competing interests.
Acknowledgements
We thank Dr. Brian Bradley from The University of Maryland Baltimore County, Dr. Dele Ogunseitan from The University of California, Irvine, and Mike Schwalbach from University of Southern California for their comments on this manuscript. The authors thank Murray V. Johnston from the University of Delaware for providing the services of the UD Mass Spectrometry facility. This work was supported by grants from the National Science Foundation, Microbial Observatories Program (MCB-0132070, MCB-0238515, MCB-0537041, MCB-0536982), INBRE program of the National Center for Research Resources (NIH grant P20 RR16472-04), and support for JMG from NSF DBI-0096578.
==== Refs
Campbell L Nolla HA Vaulot D The importance of Prochlorococcus to community structure in the central North Pacific Ocean Limnol Oceanogr 1994 39 954 961
Li WKW Primary productivity of prochlorophytes, cyanobacteria, and eucaryotic ultraphytoplankton: measurements from flow cytometric sorting Limnol Oceanogr 1994 39 169 175
Hobbie JE Daley RJ Jasper J Use of Nuclepore filters for counting bacteria by fluorescence microscopy Appl Environ Microbiol 1977 33 1225 1228 327932
Azam F Microbial control of oceanic carbon flux: The plot thickens Science 1998 280 694 696 10.1126/science.280.5364.694
Giovannoni SJ Britschgi TB Moyer CL Field KG Genetic diversity in Sargasso Sea bacterioplankton Nature 1990 345 60 63 10.1038/345060a0 2330053
Amann RI Ludwig W Schleifer K Phylogenetic identification and in situ detection of individual microbial cells without cultivation Microbiol Rev 1995 59 143 169 7535888
Rappe MS Cannon SA Vergin KL Giovannoni SJ Cultivation of the ubiquitous SAR11 marine bacterioplankton clade Nature 2002 418 630 633 10.1038/nature00917 12167859
Ward DM Weller R Bateson MM 16S rRNA sequences reveal numerous uncultured microorganisms in a natural community Nature 1990 345 63 65 10.1038/345063a0 1691827
Beja O Suzuki MT Koonin EV Aravind L Hadd A Nguyen LP Villacorta R Amjadi M Garrigues C Jovanovich SB Feldman RA DeLong EF Construction and analysis of bacterial artificial chromosome libraries froma marine microbial assemblage Environ Microbiol 2000 2 516 529 10.1046/j.1462-2920.2000.00133.x 11233160
Rondon MR August PR Bettermann AD Brady SF Grossman TH Liles MR Loiacono KA Lynch BA MacNeil IA Minor C Tiong CL Gulman M Osburne MS Clardy J Handelsman J Goodman RM Cloning the soil metagenome: a strategy for accessing the genetic and functional diversity of uncultured microorganisms Appl Environ Microbiol 2000 66 2541 2547 10.1128/AEM.66.6.2541-2547.2000 10831436
Tyson GW Chapman J Hugenholtz P Allen EE Ram RJ Richardson PM Solovyev VV Rubin EM Rokhsar DS Banfield JF Community structure and metabolism through reconstruction of microbial genomes from the environment Nature 2004 428 37 43 10.1038/nature02340 14961025
Venter JC Remington K Heidelberg JF Halpern AL Rusch D Eisen JA Wu D Paulsen I Nelson KE Nelson W Fouts DE Levy S Knap AH Lomas MW Nealson K White O Peterson J Hoffman J Parsons R Baden-Tillson H Pfannkoch C Rogers Y Smith HO Environmental genome shotgun sequencing of the Sargasso Sea Science 2004 304 66 74 10.1126/science.1093857 15001713
Lopez MF Proteome analysis I. Gene products are where the biological action is J Chromatogr 1999 722 191 202
Petersohn A Brigulla M Haas S Hoheisel JD Volker U Hecker M Global analysis of the general stress response of Bacillus subtilis J Bacteriol 2001 183 5617 5631 10.1128/JB.183.19.5617-5631.2001 11544224
Eymann C Homuth G Scharf C Hecker M Bacillus subtilis functional genomics: global characterization of the stringent response by proteome and transcriptome analysis J Bacteriol 2002 184 2500 2520 10.1128/JB.184.9.2500-2520.2002 11948165
Conway T Schoolnik GK Microarray expression profiling: capturing a genome-wide protrait of the transcriptome Mol Microbiol 2003 47 879 889 10.1046/j.1365-2958.2003.03338.x 12581346
Blackstock WP Weir MP Proteomics: quantitative and physical mapping of cellular proteins Trends Biotechnol 1999 17 121 127 10.1016/S0167-7799(98)01245-1 10189717
Wilmes P Bond PL The application of two-dimensional polyacrylamide gel electrophoresis and downstream analyses to a mixed community of prokaryotic microorganisms Enivron Microbiol 2004 6 911 920 10.1111/j.1462-2920.2004.00687.x
Ogunseitan OA Direct extraction of catalytic proteins from natural microbial communities J Microbiol Methods 1997 28 55 63 10.1016/S0167-7012(96)00962-1
Ogunseitan OA Extraction of proteins from aquatic environments Mol Microbiol Ecol Mann 1998 4.1.6 1 12
Ram RJ VerBerkmoes NC Thelen MP Tyson GW Baker BJ Blake RC II Shah M Hettich RL Banfield JF Community proteomics of a natural microbial biofilm Science 2005 308 1915 1920 10.1126/science.1109070 15879173
Zubkov MV Fuchs BM Eilers H Burkill PH Amann R Determination of total protein content of bacterial cells by SYPRO staining and flow cytometry Appl Environ Microbiol 1999 65 3251 3257 10388732
Liska AJ Shevchenko AJ Combining mass spectrometry with database interrogation strategies in proteomics Trends Anal Chem 2003 22 291 299 10.1016/S0165-9936(03)00507-7
Habermann B Oegeman J Sunyaev S Shevchenko AJ The power and the limitations of cross-species protein identification by mass spectrometry-driven sequence similarity searches Mol Cell Proteomics 2004 3 238 249 10.1074/mcp.M300073-MCP200 14695901
Shevchenko A Sunyaev S Liska A Bork P Nanoelectrospray tandem mass spectrometry and sequence similarity searching for identification of proteins from organisms with unknown genomes Methods Mol Biol 2003 211 221 234 12489434
Zickermann V Bostina M Hunte C Ruiz T Radermacher M Brandt U Functional implications from an unexpected position of the 49-kDa subunit of NADH:ubiquinone oxidoreductase J Biol Chem 2003 278 29072 29078 10.1074/jbc.M302713200 12754256
Cottrell MT Kirchman DL Community composition of marine bacterioplankton determined by 16S rRNA gene clone libraries and fluorescence in situ hybridization Appl Environ Microbiol 2000 66 5116 5122 10.1128/AEM.66.12.5116-5122.2000 11097877
Ostrowski M Fegatella F Wasinger V Guilhaus M Corthals GL Cavicchioli R Cross-species identification of proteins from proteome profiles of the marine oligotrophic ultramicrobacterium, Sphigopyxis alaskensis Proteomics 2004 4 1779 1788 10.1002/pmic.200300695 15174144
Sohn JH Kwon KK Kang JH Jung HB Kim SJ Novosphingobium pentaromativoram sp. nov., a high-molecular-mass polycyclic aromatic hydrocarbon-degrading bacterium isolated from estuarine sediments Int J Syst Evol Microbiol 2004 54 1483 1487 10.1099/ijs.0.02945-0 15388699
Chen F Suttle CA Short SM Genetic diversity in marine algal virus communities as revealed by sequence analysis of DNA polymerase genes Appl Environ Microbiol 1996 62 2869 2874 8702280
Chen F Lu J Binder BJ Liu Y Hodson RE Application of digital image analysis and flow cytometry to enumerate marine viruses stained with SYBR Gold Appl Environ Microbiol 2001 67 539 545 10.1128/AEM.67.2.539-545.2001 11157214
Shevchenko A Wilm M Vorm O Mann M Mass spectrometric sequencing of proteins from silver-stained polyacrylamide gels Anal Chem 1996 68 850 858 10.1021/ac950914h 8779443
Shevchenko A Sunyaev S Loboda A Bork P Ens W Standing KG Charting the proteomes of organisms with unsequenced genomes by MALDI-quadrupole time-of-flight mass spectrometry and BLAST homology searching Anal Chem 2001 73 1917 1926 10.1021/ac0013709 11354471
|
16176596
|
PMC1224879
|
CC BY
|
2022-09-08 23:19:43
|
no
|
Saline Syst. 2005 Aug 19; 1:7
|
utf-8
|
Saline Syst
| 2,005 |
10.1186/1746-1448-1-7
|
oa_comm
|
==== Front
Thromb JThrombosis Journal1477-9560BioMed Central London 1477-9560-3-131614455610.1186/1477-9560-3-13Original Clinical InvestigationHomocysteine, MTHFR C677T gene polymorphism, folic acid and vitamin B 12 in patients with retinal vein occlusion Ferrazzi Paola [email protected] Micco Pierpaolo [email protected] Ilaria [email protected] Lisa Simona [email protected] Alessandro Giacco [email protected] Giorgio [email protected] Lidia Luciana [email protected] Corrado [email protected] Thrombosis Center, Istituto Clinico Humanitas "IRCCS", Milan, Italy2 Ophtalmology Unit, Istituto Clinico Humanitas "IRCCS", Milan, Italy2005 7 9 2005 3 13 13 19 6 2005 7 9 2005 Copyright © 2005 Terrazzi et al; licensee BioMed Central Ltd.2005Terrazzi et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Many available data have suggested that hyperhomocysteinaemia, an established independent risk factor for thrombosis (arterial and venous), may be associated with an increased risk of retinal vein occlusion (RVO).
Aim of the study
To evaluate homocysteine metabolism in consecutive caucasian patients affected by RVO from Northern Italy.
Patients and Methods
69 consecutive patients from Northern Italy (mean age 64.1 ± 14.6 yy) with recent RVO, were tested for plasma levels of homocysteine (tHcy: fasting and after loading with methionine), cyanocobalamine and folic acid levels (CMIA-Abbot) and looking for MTHFR C677T mutation (Light Cycler-Roche) and compared to 50 volunteers, enrolled as a control group.
Results
Fasting levels of tHcy were significantly higher in patients than in controls: mean value 14.7 ± 7.7 vs 10.2 ± 8 nmol/ml. Post load levels were also significantly higher: mean value 42.7 ± 23.7 vs 30.4 ± 13.3 nmol/ml; Total homocysteine increase was also evaluated (i.e. Δ-tHcy) after methionine load and was also significantly higher in patients compared to control subjects: mean Δ-tHcy 27.8 ± 21.5 vs 21.0 ± 16 nmol/ml (normal value < 25 nmol/ml). Furthermore, patients affected by RVO show low folic acid and/or vitamin B12 levels, although differences with control group did not reach statistical significance. Heterozygous and homozygous MTHFR mutation were respectively in study group 46% and 29% vs control group 56% and 4%.
Conclusion
our data confirm that hyperhomocysteinaemia is a risk factor for RVO, and also that TT genotype of MTHFR C677T is more frequently associated with RVO: if the mutation per se is a risk factor for RVO remains an open question to be confirmed because another study from US did not reveal this aspect.
Hyperomocysteinemia is modifiable risk factor for thrombotic diseases. Therefore, a screening for tHcy plasma levels in patients with recent retinal vein occlusion could allow to identify patients who might benefit from supplementation with vitamins and normalization of homocysteine levels, in fasting and after methionine load.
retinal vein occlusionthrombophiliahomocysteineMTHFR C677Tfolic acid
==== Body
Background
Retinal vein occlusion (RVO) is a multifactorial disease which may affect small, medium and large ocular vessels; central vein occlusion represents the most dangerous clinical entity [1]. However, RVO is considered an unusual site of thrombosis [2]. Pathogenesis of RVO may recognise a local disease such as glaucoma, ocular hypertension, optical neuropathy [3] and/or an underlying systemic disease such as hypertension, diabetes, atherosclerosis, dyslipidemia, hyperviscosity syndrome, local or systemic vasculytis [4-8]. Yet, also impairment of normal haemostasis with a trend toward hypercoagulable state has been described. Blandello et al. described increased levels of prothrombin fragment 1+2 and d-dimer [9] and Lijima et al. reported increased levels of thrombin-antithrombin complexes in subjects affected by RVO [10]. Inherited thrombophilia related to clotting inhibitors deficiency (i.e. protein C, Protein S and Antithrombin III deficiencies) [11,12] has been rarely reported such as clotting XII deficiency [13], while data on the role of factor V Leiden are still matter of discussion [14-18] and few data are available on the role of prothrombin A20210G gene polymorphism in RVO pathogenesis [17,19]. Acquired thrombophilia due to the presence of antiphospholipid syndrome (primary or secondary to immunopathological disease) is an established risk factor for RVO [20,21]. Also hypofibrinolysis is a thrombotic risk factor for RVO. Plasminogen deficiency and 4G/5G gene polymorphism of plasminogen activator inhibitor type 1, in fact, have been recently described as risk factors for RVO [22,23].
Homocysteine is a sulphur-containing amino acid, which results from the hydrolysis of S-adenosyl-homocysteine in the methionine metabolic cycle. Several condition may determine an increase of blood homocysteine such as an inadequate folate intake with diet, smoking, drugs (i.e. methotrexate, hormones, antiepileptic), renal failure and inherited gene polymorphism of methylene-tetra-hydro-folate-reductase (MTHFR). Increase in circulating homocysteine may trigger endothelial dysfunction through oxidative damage therefore inducing increased oxidation of low density lipoprotein, stimulation of smooth muscle cell proliferation and hypercoagulable state. Hyperhomocysteinemia has been reported as risk factor for arterial and/or venous thrombosis [24,25]. Yet, hyperhomocysteinaemia has also been described as risk factor for RVO [26], but data on large based population are still lacking.
The aim of this study is to evaluate the role of homocysteine metabolism in consecutive patients with RVO from Northern Italy.
Patients and methods
Patients
We selected 69 consecutive caucasian patients from Northern Italy (40 males and 29 females, mean age 64.1 ± 14.6 years) affected by retinal vein occlusion (RVO). RVO diagnosis was performed with fundus oculi examination and fluorangiography.
Thirtyeight/69 (55%) patients were affected by hypertension, 12/69 (17%) by diabetes, 18/69 (26%) by dyslipidemia, 4/69 (5%) by glaucoma, while 16/69 (23%) were smokers; none of them had been diagnosed for hyperviscosity syndrome or vasculitis.
Risk factors for RVO in all patients are summarised in table 1.
Table 1 Risk factor for RVO of study group and control subjects.
RVO subjects (69) CG subjects (50) P
Hypertension 38 (55%) 15 (30%) 0.014
Diabetes 12 (17%) 9 (18%) Ns
Glaucoma 6(8%) 0 (0%) 0.039
Hyperviscosity syndrome 0 (0%) 0 (0%) Ns
Vasculitis 0 (0%) 0 (0%) Ns
Smoking 16 (23%) 12 (24%) Ns
Dyslipidemia 18 (26%) 12 (24%) Ns
RVO: retinal vein occlusion
CG: control group
HYP: hypertension
DIAB: diabetes
SMOG: smoking
DYS: dyslipidemia
GLAU: glaucoma
HS: hyperviscosity syndrome
VASC: vasculytis
Ns: not significant
Anticardiolipin antibodies (IgM and IgG) anti-β2-glycoprotein I antibodies and lupus anticoagulant were assayed in all patients in order to detect antiphospholipid syndrome; so patients with antiphospholipid syndrome were excluded from the study. Furthermore, none patient showed kidney failure.
All patients agreed to enter in the study after an informed consent was obtained.
Control group
As control group we selected 50 age-matched volunteers of the same ethnic background (38 males and 12 females, mean age 58.4 ± 12.2 years) without personal and familial history of thrombotic disorders (i.e. previous venous and/or arterial thrombosis). All subjects agreed to enter the study after an informed consent was obtained. Fifteen/50 (30%) subjects were affected by hypertension, 9/50 (18%) by diabetes, 12/50 (24%) by dyslipidemia, while 12/50 (24%) were smokers; none of them had been diagnosed for glaucoma, hyperviscosity syndrome or vasculitis. Risk factors for RVO of control subjects are also summarised in table 1.
Common risk factors for RVO
Common risk factors for RVO (i.e. hypertension, diabetes, smog, dyslipidemia, vasculitys, hyperviscosity syndrome, glaucoma) were evaluated by a thorough anamnesis, including questions concerning familial anamnesis, personal anamnesis and pharmacological anamnesis.
Methods
Whole blood samples were collected from all subjects selected in the study by venipuncture from antecubital vein in order to screen possible involvement of homocysteine metabolism. Samples were collected nearly 5–7 days after diagnosis of RVO after an adequate treatment had been started.
All patients were assayed for plasma homocysteine (tHcy), fasting and after load with methionine, vitamin B 12 and folic acid levels and MTHFR C677T gene polymorphism.
First blood sample
The first blood sample was collected in EDTA to screen fasting homocysteine (FPIA-Abbot).
Second blood sample
Post-load homocysteine value was measured after methionine administration (3.8 g/m2) per os. Homocysteinaemia was tested on a blood sample collected after four hours, then evaluated by collection of a new blood sample in EDTA (FPIA-Abbot). No food containing methionine was allowed in the interval.
We evaluated total homocysteine increase after oral load of methyonine (i.e. Δ-tHcy) in the study group and in control subjects [27].
Third blood sample
The third blood sample was collected in SST II advanced tube in order to detect serum folic acid levels and serum vitamin B 12 levels (CMIA-Abbot).
Fourth blood sample
DNA was extracted using an automated procedure (MagNA PURE, Roche, Italy). Patients were screened for the C677T gene polymorphism of MTHFR using PCR amplification with specific primers and Light Cycler apparatus (Roche, Milan, Italy).
Statistical analysis
Data are expressed as mean ± standard deviation (SD) or as number and percentage, as appropriate. Statistical analysis was performed with STATA 6 . Significance of differences was assessed by Student's t test for unpaired data, χ2 test or Fisher exact test as appropriate; differences were considered to be significant if p < 0.05.
Results
Common risk factors for RVO were evaluated and are summarised in table 1. Hypertension was present in 55% of patients compared to 30% of control group (p: 0.014); diabetes was present in 17% of patients compared to 18% of control group (p: ns); smoking was present in 23% of patients compared to 24% of control group (p: ns); dyslipidemia was present in 26% of patients compared to 24% of control group (p: ns); glaucoma was present in 8% of patients was absent in control subjects (p: 0.039); hyperviscosity syndrome and/or vasculytis were not found in study group nor in control subjects (p: ns).
Fasting levels of homocysteinaemia were significantly higher in study group than in control subjects (14.7 ± 9.9 vs 10.2 ± 8 nmol/ml, p: 0.02) (table 2).
Table 2 Data of homocysteine metabolism parameters of study group and control subjects.
Test (unit of measurement) RVO subjects (69) GC subjects (50) P
Hcy (nmol/ml) 14.7 ± 9.9 10.2 ± 8 0.02
PL-Hcy (nmol/ml) 42.7 ± 23.7 30.4 ± 13.3 <0.01
Δ-tHcy (nmol/ml) 27.8 ± 21.5 21.0 ± 16.0 <0.01
Serum folic acid (ng/ml) 6.27 ± 3.8 7.22 ± 3.5 Ns
Serum B12 vit. (pg/ml) 403.3 ± 202.0 601.6 ± 145.4 Ns
MTHFR C677T heterozigosity (%) 46,7 56,25 Ns
MTHFR C677T homozigosity (%) 29,3 4,17 <0.01
Hcy: homocysteine
PL-Hcy, post load homocysteine
Δ-tHcy: total homocysteine increase
MTHFR: methylene tetra hydrofolate reductase
RVO: retinal vein occlusion
CG: control group
Ns: not significant
Post-load levels of homocysteine (PL-Hcy) were also significantly higher in study group than in control group (42.7 ± 23.7 vs 30.4 ± 13.3 nmol/ml, p: < 0.01; normal value < 38 nmol/ml), as far as total homocysteine increase (Δ-tHcy) (27.8 ± 21.5 vs 21.0 ± 16 nmol/ml, p: < 0.01; normal value < 25 nmol/ml) (table 2).
Δ-tHcy increase after methionine load was significantly higher in study group than in control subjects (mean Δ-tHcy 27.8 ± 21.5 vs 19.8 ± 16 nmol/ml, p: < 0.01; normal value < 25 nmol/ml) (table 2).
Serum folic acids levels were 6.27 ± 3.8 ng/ml in patients vs 7.22 ± 3.5 ng/ml in control group (p: ns) (table 2), while serum vitamin B 12 levels were 403.3 ± 202.1 pg/ml in patients vs 512.6 ± 144.2 pg/ml in control subjects (p: ns) (table 2). Three/69 (4%) patients of study group showed low vitamin B 12 levels (i.e. < 150 pg/ml) vs none in the control group, while 11/69 (15%) of patients showed low folic acids levels (i.e. < 3 ng/ml) vs 2/50 (2.5%) in the control group. Among these, six patients with low folic acid and/or vitamin B 12 serum levels showed high fasting Hcy levels.
MTHFR C677T gene polymorphism was searched in 63/69 patients (91%) and was found in 75% of patients. 29/63 (46%) showed heterozigosity for MTHFR C677T gene polymorphism, while 18/63 (29%) had homozigosity for MTHFR C677T gene polymorphism. 48/50 subjects in the control group underwent genetic test to detect MTHFR C677T gene polymorphism: 27/48 (56%) showed heterozygous mutation (p: ns), while 2/48 (4%) showed homozygous mutation (p: 0.01) (table 2).
Discussion
RVO is a multifactorial disease which includes also retinal vein thrombosis and its pathophysiology may be due to local and/or systemic risk factor [1]. We may recognise several types of RVO depending on the site of occlusion (branch RVO, central RVO, hemicentral RVO) and for localisation in large, medium or small-calibre veins; from a clinical point of view RVO is usually associated with visual loss of variable degree. An association of RVO with impairment of haemostasis with a trend toward hypercoagulable state has frequently been ruled out in case of retinal vein thrombosis by several reports, but a clear relationship seems to have been established only for antiphospholipid syndrome [1,6,19]; other conditions, such as inherited thrombophilia, are less commonly described and the real incidence in these cases seems to be different in several studies [13-18]. A possible explanation could also be related to an ethnic background and to the inclusion criteria of selected patients in the related studies. However, according to available data thrombophilia seems to be more frequent in young patients affected by RVO [28].
In this field, hyperhomocysteinemia has recently been identified as an emerging risk factor for RVO. Several reports in last few years, in fact, described frequently hyperhomocysteinemia in patients affected by RVO [29-34]. In the same reports the real incidence seems to differ not only for ethnical reasons but also for the difference in the criteria used for the inclusion. A relationship between hyperhomocysteinemia and RVO is well established for young subjects, but a lot of available studies focused on patients younger than 55 years [31,32]. Moreover, only a few studies on this topic enrolled more than 50 patients, such as the Blue Mountain Eye Study [28] and a meta-analysis by Cahill et al from US [30], this being a possible cause of underestimation of this issue. On the other hand, only few studies evaluated homocysteine metabolism in a more systematic way [31-35]. Although most studies focused, in fact, on MTHFR C677T gene polymorphism, not all researchers considered folate and vitamin B 12 serum levels.
In the present study we investigated homocysteinemia and MTHFR C677T gene polymorphism and other common variables associated to homocysteine metabolism, such as folate and vitamin B 12 serum levels.
We found data that patients affected by RVO had hyperhomocysteinemia, detected as fasting homocysteinemia and as post-load homocysteinemia, and the differences reached statistical significance (table 2). These data seem to be in agreement with those previously reported by other studies. However, this is the first report also showing total increase of homocysteinemia (i.e. Δ-Hcy) evaluated in patients affected by RVO and this difference also reached statistical significance (table 2).
Serum folate and vitamin B12 levels were also tested because strongly associated to homocysteine metabolism, in terms of a possible therapeutical role. Although serum folate and vitamin B12 levels were lower in patients compared also to control subjects, these differences did not reach statistical significance; however, also these data are in agreement with data reported by Yildirim et al on small population [33].
Yet, our data clearly show that hyperhomocysteinemia is more frequent than other common known risk factors for RVO, apart from hypertension and glaucoma (i.e. diabetes, dyslipidemia, smoking, hyperviscosity syndrome, vasculytis) (table 1). Only glaucoma and hypertension reached, in fact, statistical significance in patients affected by RVO versus control subjects in our study (table 1), confirming a frequent and an interesting possible pathophysiological role of homocysteine metabolism in RVO.
Moreover, the role of hyperhomocysteinemia in patients affected by RVO seems to be confirmed also by the high frequency of MTHFR C677T homozigosity in patients affected by RVO compared to control group (table 2). The TT genotype, in fact, has been associated more frequently to hyperhomocysteinemia, although this high incidence of TT genotype could be associated also to an ethnic background. This could be relevant because an epidemiological study written by Cappuccio et al underlined that TT genotype is more frequent in Caucasian subjects compared to Asian subjects [36] and a previous study by Cahill et al. [30] on subjects from US revealed that hyperhomocysteinemia but not TT genotype of MTHFR was a risk factor for RVO. So, although further studies on large based population should be performed also from a genetic point of view; our data suggest a prompt screening for homocysteine metabolism and MTHFR C677T gene polymorphism (in particular for Caucasian subjects) in patients affected by RVO. Further studied could also be addressed to understand also the role of another emerging MTHFR gene polymorphism (i.e. A1298C) sometime associated to hyperhomocysteinemia [37], alone and/or in association with MTHF C677T gene polymorphism, in large based population affected by RVO.
In conclusion, based on this study, we suggest a careful investigation on several metabolic parameters related to homocysteine metabolism in patients affected by RVO. A screening for hyperhomocysteinemia should be promptly performed particularly in patients without other common risk factors for RVO. These studies could have in fact a deep impact also on therapeutical aspects in order to better understand a possible role of folic acid and vitamin B12 fortification in the management of RVO.
Acknowledgements
Authors thank dr. Manuela Morenghi, Epidemiology Unit, Istituto Clinico Humanitas, for her helpful work on statistical analysis.
==== Refs
Prisco D Bertini L Marcucci R Poli D Le occlusioni venose retiniche: malattie internistiche? Ann Ital Med Int 2000 15 75 84 [Italian] 10842895
Martinelli I Unusual forms of venous thrombosis and thrombophilia Pathophysiol Haemost Thromb 2002 32 343 345 13679671 10.1159/000073595
The Eye Disease Case-Control Study Group Risk factors for central retinal vein occlusion Arch Ophtalmol 1996 114 545 554
Dodson PM Kubicki AJ Taylor KG Kritzinger EE Medical condition underlying recurrence or retinal vein occlusion Br J Ophtalm 1985 69 493 496
Dodson PM Galton DJ Winder AF Retinal vascular abnormalities in hyperlipidemias Trans Ophtalmol Soc U K 1981 101 17 21
Laroche L Saraux H Unilateral central vein occlusion in systemic lupus erythematosus Ophtalmology 1984 189 128 129
Orellana J Friedmann AH Ocular manifestation of multiple myeloma, Waldenstrom's macroglobulinemia and benign monoclonal gammopathy Surv Ophtalmol 1981 26 157 169 10.1016/0039-6257(81)90065-5
Hayasaka S Ugomori S Kodama T Noda S Setogawa T Central vein occlusion in two patients with immunoglobulin G multiple myeloma associated with blood hyperviscosity Ann Ophtalmol 1993 25 191 194
Blandello F D'Angelo SV Parlavecchia M Tavola A Della Valle P Brancato R D'Angelo A Hypercoagulability and high lipoprotein (a) levels in patients with central retinal vein occlusion Thromb Haemost 1994 72 39 43 7974373
Lijima H Godho T Imai M Tsukahara S Thrombin-antithrombin III complex in acute retinal vein occlusion Am J Ophtalmol 1998 126 677 682 10.1016/S0002-9394(98)00226-8
Guareschi G Viganò D'Angelo S Foresti V Blandello F D'Angelo A Normal activities of protein C, protein S and antithrombin III in young patients with retinal vein occlusion Haematologica 1990 75 483 484 2151449
Kruger K Anger V Ischemic occlusion of the central retinal vein and protein C deficiency J Fr Ophtalmol 1990 13 369 371 2290011
Khuli C Sharrer I Koch F Ohrloff C Hattenbach LO Factor XII deficiency: a thrombophilic risk factor for retinal vein occlusion Am J Ophtalmol 2004 137 459 464 10.1016/j.ajo.2003.10.024
Dhote R Bachmeyer C Horrellou MM Toulon P Christoforov B Central retinal vein thrombosis associated with resistance to activated protein C Am J Ophtalmol 1995 120 388 389
Graham SL Goldberg I Murray B Beaumont P Chong BH Activated protein C resistance-low incidence in glaucomatous optic disc haemorrage and central retinal vein occlusion Aust N Z J Ophtalmol 1996 24 199 205
Gottlieb JL Blice JP Mestichelli B Konle BA Benson WE Activated protein C resistance, factor V Leiden and central vein occlusion in young adults Arch Ophtalmol 1998 116 577 579
Di Minno G Albisinni R Coppola A Loffredo M Cerbone AM Greco GM Retinal vein occlusion and inherited conditions predisposing to thrombophilia Thromb Haemost 1998 80 702 703 9798994
Delahousse B Arsène S Piquemal R Iochmann S Le Lez ML Pouplard C Gruel Y The 20210 allele of the prothrombin gene is not a risk for retinal vein occlusion Blood Coagul Fibrinolysis 1998 9 447 448 9712295
Kleiner RC Najarian LV Schatten S Jabs DA Patz A Kaplan HJ Vaso-occlusive retinopathy associated with antiphospholipid antibodies (lupus anticoagulant retinopathy) Ophtalmology 1989 96 896 904
Lureau MA Glacet-Bernard A Coscas G Bilateral central retinal vein occlusion and lupus anticoagulant antibody J Fr Ophtalmol 1995 18 462 472
Tavola A Vigano D'Angelo S Bandello F Brancato R Parlavecchia M Safa O D'Angelo A Central retinal vein and branch artery occlusion associated with inherited plasminogen deficiency and high lipoprotein (a) levels: a case report Thromb Res 1995 80 327 331 8585045 10.1016/0049-3848(95)00183-R
Glueck CJ Bell H Vadlamani L Gupta A Fontaine RN Wang P Stroop D Gruppo R Heritable thrombophilia and hypofibirinolysis. Possible causes of retinal vein occlusion Arch Ophtalmol 1999 117 43 49
Wenzler EM Rademarkers AJJ Boers GH Cruysberg JRM Webers CAB Deurman AF Hyperhomocysteinemia in retinal artery and retinal vein occlusion Am J Ophtalmol 1993 115 162 167
Welch GN Loscalzo J Homocysteine and atherothrombosis N Engl J Med 1998 338 1042 1050 9535670 10.1056/NEJM199804093381507
Martinelli I Risk factors in venous thromboembolism Thromb Haemost 2001 86 395 403 11487029
Blousse V Newman NJ Sternberg P Retinal vein occlusion and transient visual loss associated with hyperhomocysteinemia Am J Ophtalmol 1997 124 257 260
Bolander-Gouaille C Bolander-Gouaille C How to interpret the test results and how to handle hyperhomocysteinemia Determination of homocysteine, why, when and how 1999 Tom Sandberg 54 57
Lahey JM Tunc M Kearney J Modlinski B Koo H Jhonson RN Tanaka S Laboratory evaluation of hypercoagulable states in patients with central vein occlusion who are less than 56 years of age Ophtalmology 2002 109 126 131 10.1016/S0161-6420(01)00842-9
Chua B Kifley A Wong TY Mitchell P Homocysteine and retinal vein occlusion: a population-based study Am J Ophtalmol 2005 139 181 182 10.1016/j.ajo.2004.06.084
Cahill MT Stinnett SS Fekrat S Meta-analysis of plasma homocysteine, serum folate, serum vitamin B (12), and thermolabil MTHFR genotype as risk fators for retinal vascular occlusive disease Am J Ophtalmol 2003 136 1136 1150 10.1016/S0002-9394(03)00571-3
Parodi MB Di Crecchio L Hyperhomocysteinemia in central retinal vein occlusion in young adults Semin Ophtalmol 2003 18 154 159
Di Crecchio L Parodi MB Sanguinetti G Iacono P Ravalico G Hyperhomocysteinemia and methylenetetrahydrofolate reductase 677C-T mutation in patients under 50 years of age affected by central vein occlusion Ophtalmology 2004 111 940 945 10.1016/j.ophtha.2003.08.028
Yildirim C Yaylali V Tatlipinar S Kaptanoglu B Akpinar S Hyperhomocysteinemia: a risk factor for retinal vein occlusion Ophtalmologica 2004 218 102 106 10.1159/000076144
Brown BA Marx JL Ward TP Hollifield RD Dick JS Brozettin JJ Howard RS Homocysteine: a risk factor for retinal vein occlusive disease Ophtalmology 2002 109 287 290 10.1016/S0161-6420(01)00923-X
Prisco D Marcucci R Bertini L Gori AM Cardiovascular and thrombophilic risk factors for central retinal vein occlusion Eur J Intern Med 2002 13 163 169 12020623 10.1016/S0953-6205(02)00025-0
Cappuccio FP Bell R Perry IJ Gilg J Ueland PM Refsum H Sagnella GA Jeffery S Cook DG Homocysteine levels in men and women of different ethnic and cultural background living in England Atherosclerosis 2002 164 95 102 12119198 10.1016/S0021-9150(02)00024-2
Abu El-Asrar AM Abdel Gader AG Al-Amro SA Al-Attas OS Hyperhomocysteinemia and retinal vascular occlusive disease Eur J Ophthalmol 2002 12 495 500 12510718
|
16144556
|
PMC1224880
|
CC BY
|
2021-01-04 16:36:24
|
no
|
Thromb J. 2005 Sep 7; 3:13
|
utf-8
|
Thromb J
| 2,005 |
10.1186/1477-9560-3-13
|
oa_comm
|
==== Front
Virol JVirology Journal1743-422XBioMed Central London 1743-422X-2-751612487710.1186/1743-422X-2-75Short ReportConstitutive expression of Atlantic salmon Mx1 protein in CHSE-214 cells confers resistance to Infectious Salmon Anaemia virus Kibenge Molly JT [email protected] Khalid [email protected] Frederick SB [email protected] Department of Pathology and Microbiology, Atlantic Veterinary College, University of Prince Edward Island, 550 University Avenue, Charlottetown, PE. C1A 4P3. Canada2005 26 8 2005 2 75 75 8 7 2005 26 8 2005 Copyright © 2005 Kibenge et al; licensee BioMed Central Ltd.2005Kibenge et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Infectious salmon anaemia (ISA) is a highly fatal viral disease affecting marine-farmed Atlantic salmon which is caused by ISA virus (ISAV), a fish orthomyxovirus that has recently been assigned to the new genus Isavirus within the family Orthomyxoviridae. Mx proteins are among the interferon (IFN)-induced proteins responsible for the development of an antiviral state in vertebrate cells. We used real-time reverse transcription-polymerase chain reaction (RT-PCR) and Chinook salmon embryo (CHSE-214) cells constitutively expressing Atlantic salmon Mx1 protein (ASMx1) to examine the antiviral properties of ASMx1 against two ISAV strains, NBISA01 and HKS-36, having phenotypically different growth properties (cytopathic vs non-cytopathic) in the CHSE-214 cell line. We present evidence that ISAV is sensitive to ASMx1. CHSE-214 cells constitutively expressing ASMx1 showed increased resistance to infection with the cytopathic ISAV strain NBISA01, manifested as delayed development of cytopathic effects (CPE) and significant reduction in the severity of CPE, as well as a 10-fold reduction in virus yield. However, by real-time RT-PCR we observed no significant difference in the mean threshold cycle (Ct) values of ISAV RNA levels, suggesting that the ASMx1 activity on ISAV occurs at the post-transcription steps of virus replication, possibly in the cytoplasm.
==== Body
Findings
Infectious salmon anaemia (ISA) virus (ISAV), the causative agent of a highly fatal disease of marine-farmed Atlantic salmon, is a fish orthomyxovirus that has recently been assigned to the new genus Isavirus within the family Orthomyxoviridae [1]. The disease has caused severe economic losses to the salmon-farming industry in several countries in the northern hemisphere, and confirmed positive diagnostic results for ISA are reportable to the "Office International des Epizooties", OIE [2]. Viruses in the genus Isavirus are enveloped particles of 90–140 nm diameter with surface projections consisting of a combined haemagglutinin-esterase (HE) protein [3] and a separate putative fusion (F) protein [4]. The genome is composed of eight segments of linear, single-stranded negative sense RNA ranging in length from 1.0 to 2.4 kb with a total molecular size of approximately 14.3 kb [5]. Sequence analysis of several ISAV isolates on the eight segments consistently reveals two genotypes, one European and one North American. The gene-coding assignments of the ISAV genome differ from those of other orthomyxoviruses [2,4].
Some strains of ISAV can replicate and cause CPE in the CHSE-214 cell line while others do not replicate in this cell line at all [6,7]. The molecular basis for this phenotypic difference is not clearly known. Although there is a strong geographical correlation in that practically all ISAV isolates of the European genotype do not cause CPE in the CHSE cell line, this phenotype is not related to sequence variation in the HE protein [8].
In general, fish viral diseases are difficult to control due to the high susceptibility of fish at an early age, and insufficient knowledge of pathogenesis of virus infections. In this context, we are studying the mechanisms by which ISAV interacts with its host, including the identification of markers for ISAV virulence. Vertebrates [9], including fish [10], mount an early strong innate immune response against virus infections, characterized by the induction and secretion of cytokines, such as type I interferons (IFN-α/β) that mediate an antiviral state. Secreted IFNs signal through a common receptor activating a JAK/STAT signaling pathway which leads to the transcriptional upregulation of numerous IFN-stimulated genes (ISGs), a number of which encode antiviral proteins [10-12]. The induced antiviral proteins include dsRNA-dependent protein kinase R (PKR), 2',5'-oligoadenylate synthetase (OAS), and the Mx proteins [10,12]. Viruses have evolved mechanisms to subvert the host IFN response [12,13]. Previous in-vitro studies revealed that ASMx1 inhibited the replication of infectious pancreatic necrosis virus (IPNV), a dsRNA virus belonging to the Birnaviridae family, but it did not appear to inhibit replication of ISAV [14,15]. It was concluded from those results that ISAV and IPNV have developed different strategies to avoid the IFN-system of Atlantic salmon. Thus, we were interested in analyzing ISAV strains having phenotypically different growth properties in the CHSE-214 cell line to gain further insight in ISAV virulence. We present evidence that ISAV is sensitive to the antiviral activity of ASMx1.
The following two ISAV isolates were selected for use in this study: NBISA01 and HKS-36. The growth properties of the isolates in CHSE-214 cells have been previously described [6]. NBISA01 is CPE-positive whereas HKS-36 is CPE-negative in CHSE-214 cells. For use, the viruses were propagated and titrated in CHSE-214 cells and/or TO cells as previously described [6,8]. To demonstrate the effects of ASMx1 on the replication of the two ISAV strains, we compared the cytopathogenicity and virus yields of the viruses in normal CHSE-214 cells and in cells constitutively expressing ASMx1. The CHSE-214 cells constitutively expressing ASMx1 [15] were a kind gift from Dr. Børre Robertsen, Norwegian College of Fishery Science, University of Tromsø, Norway. These cells were grown in presence of 0.5 mg/ml zeocine (Invitrogen Life Technologies) to maintain expression of the transfected genes (Dr. Børre Robertsen, personal communication). Virus stocks were prepared from TO cell cultures infected with the selected ISAV isolates which were harvested when CPE was complete; usually 7–9 days post inoculation (dpi). They contained 108.16 TCID50/ml for NBISA01, and 107.16TCID50/ml for HKS-36. CHSE-214 cell monolayers in 25 cm2 tissue culture flasks, each inoculated with 1 ml of virus, were monitored daily for CPE and were harvested 14 dpi. Viral titers of the harvests were determined on normal CHSE-214 cells in 48-well tissue culture plates as previously described [6]. A significant delay in development of CPE and a reduction in the severity of CPE, as well as a 10-fold reduction in virus yield were observed for NBISA01 grown in CHSE-214 cells constitutively expressing ASMx1 compared to normal CHSE-214 cells (Table 1). No CPE was seen with HKS-36 in both types of CHSE-214 cells.
Table 1 Inhibition of ISAV NBISA01 by AsMx1 in CHSE-214 cells.
CHSE-214 cells1 CPE development2 CPE severity3 Virus titer4
Normal 5–14 days Complete (5+) 5.50
ASMx1 overexpression 6-* days Incomplete (3+) 4.50
1CHSE-214 cells: Normal CHSE-214 cells and CHSE-214 cells constitutively expressing ASMx1.
2CPE development denotes time from first appearance to complete CPE in days post inoculation; * indicates incomplete CPE development by 14 days post inoculation.
3CPE severity was scored from 1+ to 5+: Complete CPE was scored 5+; Incomplete CPE scored 1+ to 4+).
4Virus titers are expressed as TCID50/ml.
In an attempt to assess the effects of ASMx1 on the viral mRNA levels, we used real-time RT-PCR in the LightCycler with RNA Amplification Kit SYBR Green I (Roche Applied Science) and PCR primers FA-3/RA-3 targeting a 220-bp product on ISAV segment 8 [7]. This assay has been shown to be 100 times more sensitive than the conventional one-tube RT-PCR assay, and to differentiate ISAV isolates into three CHSE phenotypes (replicating cytopathic, replicating non-cytopathic, and non-replicating) based on their ability to replicate and cause CPE in CHSE-214 cells [7]. Total RNA was extracted from 300 μl of cell culture supernatants of the uninfected and ISAV-infected tissue cultures harvested at 14 dpi by using TRIZOL reagent (Invitrogen Life Technologies). The RNA pellet was dissolved in 10 μl of RNase free water and 1 μl was used in real-time RT-PCR. The assay was performed on three replicates of each virus sample. The thermal conditions were one cycle of reverse transcription at 55°C for 30 min, pre-denaturation at 95°C for 30 s followed by 50 cycles of 95°C for 5s, 59°C for 10s, 72°C for 10s, and data acquisition at 80°C for 2s. Melting curve analysis was performed from 70°C to 95°C in 0.1°C/s increments to assess the specificity of the RT-PCR products. The quantitative (Ct values) and melting curve data were analyzed using LightCycler software version 3.5 (Roche Applied Science). The real-time amplification reaction products were also resolved by 1% agarose gel electrophoresis in 0.5 × TBE buffer and visualized by staining with ethidium bromide and photographed under 304 nm UV light.
The real-time RT-PCR data are shown in Figures 1,2,3. The mean Ct value of ISAV RNA levels for NBISA01 was 17.39 ± 0.24 in normal CHSE-214 cells (Fig. 1A) and 17.62 ± 0.08 in CHSE-214 cells constitutively expressing ASMx1 (Fig. 2G). For HKS-36, the mean Ct value was 22.94 ± 0.08 in normal CHSE-214 cells (Fig. 1D) and 21.08 ± 0.003 in CHSE-214 cells constitutively expressing ASMx1 (Fig. 2J). For each virus strain, there was no significant difference in the Ct values of both types of CHSE-214 cells. Melting curve analysis showed that the Tm of the ISAV templates was 82.5 ± 0.03°C and occurred as a single amplicon peak for both NBISA01 and HKS-36 in normal CHSE-214 cells (Fig. 1B, E) and in CHSE-214 cells constitutively expressing ASMx1 (Fig. 2H, K). All these amplicons had one band of 220 bp on agarose gel (Fig. 1C, F and Fig. 2I, L), indicating virus-specific amplification and uniform virus populations in the individual virus samples. The melting curve analyses of amplification products of uninfected CHSE-214 cells and water control indicated the presence of primer dimers having values of 77.4 ± 0.01°C (Fig. 3N, Q) and 78.1°C (Fig. 3T), respectively. No viral amplicon-specific melting peaks were observed in these samples. The absence of the 220 bp band on agarose gel in these samples (Fig. 3O, R, U) further confirmed the validity of melting curve analyses of these samples. Thus, by real-time RT-PCR we were not able to detect any effect of the ASMx1 activity on ISAV replication for both NBISA01 and HKS-36. In case of NBISA01, this indicated to us that there was no correlation between the virus yields (Table 1) and the Ct values (Fig. 1A and Fig. 2G). This would be expected since the real-time RT-PCR detected both mRNA and viral genomic RNA and was not indicative of infectious virus. Moreover, RNA detected by real-time RT-PCR could be from infectious as well as non-infectious viral progeny. These observations allow us to speculate that the ASMx1 activity on ISAV occurs at the post-transcription steps of virus replication. The human and mouse Mx proteins are known to inhibit different steps of the influenza virus multiplication cycle [16] which is dictated by their intracellular locations. The mouse Mx1 protein accumulates in the nucleus and interferes with primary transcription of influenza virus in the nucleus, whereas the human MxA protein which is localized in the cytoplasm inhibits a subsequent step that presumably takes place in the cytoplasm of infected cells. ASMx1 has been reported to localize in the cytoplasm [15]. Thus in this context, the ASMx1 activity on ISAV may resemble that of the human MxA protein on influenza A virus.
Figure 1 Amplification, melting curve, and agarose gel electrophoresis of RT-PCR targeting a 220-bp product on ISAV segment 8 using total RNA from ISAV-infected normal CHSE-214 cells: NBISA01 (A-C) and HKS-36 (D-F). Total RNA was subjected to real-time RT-PCR with 50 cycle amplification. The arrows in C and F indicate the expected band of 220 bp on agarose gel.
Figure 2 Amplification, melting curve, and agarose gel electrophoresis of RT-PCR targeting a 220-bp product on ISAV segment 8 using total RNA from ISAV-infected normal CHSE-214 cells constitutively expressing ASMx1: NBISA01 (G-I) and HKS-36 (J-L). Total RNA was subjected to real-time RT-PCR with 50 cycle amplification. The arrows in I and L indicate the expected band of 220 bp on agarose gel.
Figure 3 Amplification, melting curve, and agarose gel electrophoresis of RT-PCR targeting a 220-bp product on ISAV segment 8 using total RNA from uninfected controls: normal CHSE-214 cells (M-O), CHSE-214 cells constitutively expressing ASMx1 (P-R) and the water only control (S-U). Samples were subjected to real-time RT-PCR with 50 cycle amplification. The arrows in O, R and U represent DNA markers in lane M; note the absence of the 220 bp-band on these agarose gels.
To our knowledge these are the first studies demonstrating that ASMx1 interferes with ISAV replication and suggest that this interference occurs after viral mRNA synthesis. Orthomyxoviruses are known to be sensitive to the antiviral action of type I IFN, and have evolved viral IFN antagonists to suppress IFN induction [17-19]. Influenza A virus NS1 protein binds dsRNA [20] and prevents activation of PKR [21] and OAS [22]. Additionally, the NS1 protein is able to prevent activation of NFκ-B [23] and IRF-3 [24] which are necessary for IFN-β synthesis in virus infected cells. Influenza B virus NS1 protein also inhibits activation of IRF-3 [19], and the activity of ISG15, one of the major type I IFN inducible proteins [25]. Thogoto virus protein ML suppresses IRF-3 function [26]. It is therefore expected that ISAV also encodes protein(s) with IFN antagonistic properties, giving the virus an advantage in its fight against the antiviral host response to infection. The complete ISAV protein profile has not yet been definitively established. ISAV segment 8, like its influenza A virus counterpart, has been shown to encode two proteins; one open reading frame (ORF) was shown to encode a major structural protein corresponding to the virus matrix protein [3], and the other ORF is considered to encode the non-structural protein [2] albeit with no sequence similarity to the influenza NS1 proteins [27]. It would be interesting to see if IFN and ISGs knockdowns would enhance the growth and/or cytopathogenicity of ISAV. This would provide further insight into the exact mechanisms by which ISAV interacts with the Atlantic salmon IFN system with implications for ISAV virulence.
List of abbreviations
Infectious salmon anaemia virus (ISAV), Atlantic salmon Mx1 protein (ASMx1), cytopathic effect (CPE), Tris borate EDTA (TBE).
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
MJTK conducted all the experiments and wrote the manuscript. KM assisted with running the real-time RT-PCR assays and with writing the manuscript. FSBK conceived the study, coordinated the research efforts and edited the paper. All three co-authors read and approved the final manuscript.
Acknowledgements
We kindly thank Dr. Børre Robertsen for the CHSE-214 cells constitutively expressing ASMx1. This work was supported by Natural Sciences and Engineering Research Council (NSERC) of Canada Strategic Grants and an NSERC Discovery Grant.
==== Refs
Kawaoka Y Cox NJ Haller O Hongo S Kaverin N Klenk H-D Lamb RA McCauley J Palese P Rimstad E Webster RG CM Fauquet, MA Mayo, J Maniloff, U Desselberger, LA Bull Infectious Salmon Anaemia Virus Virus Taxonomy - Eight Report of the International Comitee on Taxonomy Viruses 2005 Elsevier Academic Press: New York 681 693
Kibenge FSB Munir K Kibenge MJT Joseph T Moneke E Infectious salmon anaemia virus: causative agent, pathogenesis and immunity Ani Hlth Res Rev 2004 5 65 78 10.1079/AHR200461
Falk K Aspehaug V Vlasak R Endresen C Identification andcharacterization of viral structural proteins of infectious salmon anaemia virus J Virol 2004 78 3063 3071 14990725 10.1128/JVI.78.6.3063-3071.2004
Aspehaug VT Characterization of major structural proteins of the infectious salmon anaemia virus (ISAV) Doctoral Thesis 2005 University of Bergen, Norway Paper III
Clouthier SC Rector T Brown NEC Anderson ED Genomic organization of infectious salmon anaemia virus J Gen Virol 2002 83 421 428 11807235
Kibenge FSB Lyaku JR Rainnie D Hammell KL Growth of infectious salmon anaemia virus in CHSE-214 cells and evidence for phenotypic differences between virus strains J Gen Virol 2000 81 143 150 10640552
Munir K Kibenge FSB Detection of infectious salmon anaemia virus by real-time RT-PCR J Virol Meth 2004 117 37 47 10.1016/j.jviromet.2003.11.020
Kibenge FSB Kibenge MJT McKenna PK Stothard P Marshall R Cusack RR McGeachy S Antigenic variation among isolates of infectious salmon anaemia virus correlates with genetic variation of the viral haemagglutinin gene J Gen Virol 2001 82 2869 2879 11714961
Samuel CE Antiviral actions of interferons Clin Microbiol Rev 2001 14 778 809 11585785 10.1128/CMR.14.4.778-809.2001
Robertsen B The interferon system of teleost fish Fish & Shellfish Immunol 2006 20 172 191 15939626 10.1016/j.fsi.2005.01.010
Goodbourn S Didcock L Randall RE Interferons: cell signalling, immune modulation, antiviral responses and virus countermeasures J Gen Virol 2000 81 2341 2364 10993923
Stark GR Kerr IM Williams BR Silverman RH Schreiber RD How cells respond to interferons Annu Rev Biochem 1998 67 227 264 9759489 10.1146/annurev.biochem.67.1.227
Garcia-Sastre A Inhibition of interferon-mediated antiviral responses by influenza A viruses and other negative-strand RNA viruses Virology 2001 279 375 384 11162793 10.1006/viro.2000.0756
Jensen I Robertsen B Effect of double-stranded RNA and interferon on the antiviral activity of Atlantic salmon cells against infectious salmon anaemia virus and infectious pancreatic necrosis virus Fish & Shellfish Immunol 2002 13 221 241 12365733 10.1006/fsim.2001.0397
Larsen R Røkenes TP Robertsen B Inhibition of infectious pancreatic necrosis virus replication by Atlantic salmon Mx1 protein J Virol 2004 78 7938 7944 15254166 10.1128/JVI.78.15.7938-7944.2004
Pavlovic J Haller O Staeheli P Human and mouse Mx proteins inhibit different steps of the influenza virus multiplication cycle J Virol 1992 2564 2569 1548781
Garcia-Sastre A Egorov A Matassov D Brandt S Levy DE Influenza A virus lacking the NS1 gene replicates in interferon-deficient systems Virology 1998 252 324 330 9878611 10.1006/viro.1998.9508
Hagmaier K Jennings S Buse J Weber F Kochs G Novel gene product of Thogoto virus segment 6 codes for an interferon antagonist J Virol 2003 77 2747 2752 12552016 10.1128/JVI.77.4.2747-2752.2003
Donelan NR Dauber B Wang X Basler CF Wolff T Garcia-Sastre A The N- and C-terminal domains of the NS1 protein of influenza B virus can independently inhibit IRF-3 and beta interferon promoter activation J Virol 2004 78 11574 11582 15479798 10.1128/JVI.78.21.11574-11582.2004
Hatada E Fukuda R Binding of influenza A virus NS1 protein to dsRNA J Gen Virol 1992 73 3325 3329 1469370
Bergmann M Garcia-Sastre A Carnero E Pehamberger H Wolff K Influenza virus NS1 protein counteracts PKR-mediated inhibition of replication J Virol 2000 74 6203 6206 10846107 10.1128/JVI.74.13.6203-6206.2000
Lu Y Wambach M Katze MG Krug RM Binding of the influenza virus NS1 protein to double-stranded RNA inhibits the activation of the protein kinase that phosphorylates the Eif-2 translation initiation factor Virology 1995 214 222 228 8525619 10.1006/viro.1995.9937
Wang X Li M Zheng H Muster T Palese P Influenza A virus NS1 protein prevents activation of NF-kappaB and induction of Alpha/Beta interferon J Virol 2000 74 11566 11573 11090154 10.1128/JVI.74.24.11566-11573.2000
Talon J Horvath CM Polley R Basler CF Muster T Activation of interferon regulatory factor 3 is inhibited by the influenza A virus NS1 protein J Virol 2000 74 7989 7996 10933707 10.1128/JVI.74.17.7989-7996.2000
Yuan W Krug RM Influenza B virus NS1 protein inhibits conjugation of the interferon (IFN)-induced ubiquitin-like ISG15 protein EMBO J 2001 20 362 371 11157743 10.1093/emboj/20.3.362
Jennings S Martinez-Sobrido L Garcia-Sastre A Weber F Kochs G Thogoto virus ML protein suppresses IRF3 function Virology 2005 331 63 72 15582653 10.1016/j.virol.2004.10.015
Mjaaland S Rimstad E Falk K Dannevig BH Genomic characterization of the virus causing infectious salmon anaemia in Atlantic salmon (Salmo salar L): an orthomyxo-like virus in a teleost J Virol 1997 71 7681 7686 9311851
|
16124877
|
PMC1224881
|
CC BY
|
2021-01-04 16:39:00
|
no
|
Virol J. 2005 Aug 26; 2:75
|
utf-8
|
Virol J
| 2,005 |
10.1186/1743-422X-2-75
|
oa_comm
|
==== Front
World J Surg OncolWorld Journal of Surgical Oncology1477-7819BioMed Central London 1477-7819-3-611616474210.1186/1477-7819-3-61ResearchRole of p-glycoprotein expression in predicting response to neoadjuvant chemotherapy in breast cancer-a prospective clinical study Chintamani [email protected] Jai Parakash [email protected] Mahesh K [email protected] Sunita [email protected] Anju [email protected] Ashima [email protected] Pranjal [email protected] Department of Surgery, Vardhman Mahavir Medical College Safdarjang Hospital New Delhi-110023-India2 Department of Radiology, Vardhman Mahavir Medical College Safdarjang Hospital New Delhi-110023-India3 Tumor Biology Laboratory, Indian Council Of Medical Research, Vardhman Mahavir Medical College Safdarjang Hospital New Delhi-110023-India2005 14 9 2005 3 61 61 15 1 2005 14 9 2005 Copyright © 2005 Chintamani et al; licensee BioMed Central Ltd.2005Chintamani et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Neoadjuvant chemotherapy (NACT) is an integral part of multi-modality approach in the management of locally advanced breast cancer. It is vital to predict response to chemotherapy in order to tailor the regime for a particular patient. The prediction would help in avoiding the toxicity induced by an ineffective chemotherapeutic regime in a non-responder and would also help in the planning of an alternate regime. Development of resistance to chemotherapeutic agents is a major problem and one of the mechanisms considered responsible is the expression of 170-k Da membrane glycoprotein (usually referred to as p-170 or p-glycoprotein), which is encoded by multidrug resistance (MDR1) gene. This glycoprotein acts as an energy dependent pump, which actively extrudes certain families of chemotherapeutic agents from the cells. The expression of p-glycoprotein at initial presentation has been found to be associated with refractoriness to chemotherapy and a poor outcome. Against this background a prospective study was conducted using C219 mouse monoclonal antibody specific for p-glycoprotein to ascertain whether pretreatment detection of p-glycoprotein expression could be utilized as a reliable predictor of response to neoadjuvant chemotherapy in patients with breast cancer.
Patients and methods
Fifty cases of locally advanced breast cancer were subjected to trucut® biopsy and the tissue samples were evaluated immunohistochemically for p-glycoprotein expression and ER, PR status. The response to neoadjuvant chemotherapy was assessed clinically and by using ultrasound after three cycles of FAC regime (cyclophosphamide 600 mg/m2, Adriamycin 50 mg/m2, 5-fluorourail 600 mg/m2 at an interval of three weeks). The clinical response was correlated with both the pre and post chemotherapy p-glycoprotein expression. Descriptive studies were performed with SPSS version 10. The significance of correlation between tumor response and p-glycoprotein expression was determined with chi square test.
Results
A significant relationship was found between the pretreatment p-glycoprotein expression and clinical response. The positive p-glycoprotein expression was associated with poor clinical response rates. When the clinical response was correlated with p-glycoprotein expression, a statistically significant negative correlation was observed between the clinical response and p- glycoprotein expression (p < 0.05). There was another significant observation in terms of development of post NACT p-glycoprotein positivity. Before initiation of NACT, 26 patients (52%) were p-glycoprotein positive and after three cycles of NACT, the positivity increased to 73.5% patients.
Conclusion
The study concluded that pretreatment p-glycoprotein expression predicts and indicates a poor clinical response to NACT. Patients with positive p-glycoprotein expression before initiation of NACT were found to be poor responders. Thus pretreatment detection of p-glycoprotein expression may be utilized, as a reliable predictor of response to NACT in patients with breast cancer The chemotherapy induced p-glycoprotein positivity observed in the study could possibly explain the phenomenon of acquired chemoresistance and may also serve as an intermediate end point in evaluating drug response particularly if the adjuvant therapy is planned with the same regime.
==== Body
Background
Neoadjuvant chemotherapy (NACT) is an integral part of multi-modality approach for the local and systemic management of locally advanced breast cancer (LABC) [1-5]. However chemoresistance is a major problem and one of the proposed mechanisms for its development is the expression of 170-kDa-membrane glycoprotein (usually referred to as P-170 or P-glycoprotein) encoded by multi drug resistance-1 (MDR-1) gene [6-9]. This membrane glycoprotein acts as an energy dependent pump, which actively inhibits accumulation of certain families of chemotherapeutic agents by extruding them from the cells thus leading to a poor response [8,9].
Immunohistochemical detection using monoclonal antibody for p-glycoprotein has been considered to be a more sensitive method than southern, northern or western blot analysis because even very low number of positive cells or lower grade of expression could be readily recognized [10,11].
In some tumors like leukemia's, multiple myelomas, Hodgkin's lymphomas, soft tissue sarcomas the expression of p-glycoprotein at initial presentation has been found to correlate significantly with refractoriness to chemotherapy thus highlighting its importance as a contributor to chemoresistance [12,13].
Pretreatment detection of p-glycoprotein has not been found to be a useful tool for predicting response before the initiation of chemotherapy in breast cancers as its expression has not been commonly observed in the untreated breast cancer cell lines [9,14,15]. There are only scattered reports in the literature of extraordinarily high incidence of p-glycoprotein expression in untreated breast cancer specimen and its role in the assessment of pretreatment p-glycoprotein expression for predicting treatment failure [15-19].
Against this background a prospective study was conducted using C-219 mouse monoclonal antibody specific for p-glycoprotein to ascertain whether pretreatment detection of p-glycoprotein in patients of LABC could be utilized as a reliable predictor of response to neoadjuvant chemotherapy.
Patients and methods
After approval by the Institution Review Board and the ethical committee of the hospital, 50 fine needle aspiration cytology (FNAC) proven cases of LABC according to AJCC (American Joint Committee on Cancer) classification were included in the study. The tumor size and the axillary lymph node status were measured clinically and by using ultrasonography. Core biopsy was performed for immunohistochemical estimations of p-glycoprotein and ER, PR status in the biopsy specimen before initiating the chemotherapy. Routine and metastatic work up (total blood count, platelet count), chest radiograph, electrocardiography (ECG) (echocardiography when ECG had a positive finding), liver function tests, bone scan, ultrasonography (USG) of the abdomen, renal function tests were routinely done in all the cases.
Three cycles of FAC regime (cyclophosphamide 600 mg/m2, adriamycin 50 mg/m2, 5-fluorourail 600 mg/m2) were given at three weekly intervals and the patients were assessed both clinically and by USG for response in the form of reduction in breast tumor size and axillary lymph node status. Patey's modified radical mastectomy was performed three weeks after the last cycle and the mastectomy specimen was examined for pathological response, resected margins, axillary lymph nodes, ER, PR status and p-glycoprotein expression (post NACT).
The pathological tumor response was evaluated by size measurement at the time of tumor resection macroscopically and by detecting tumor cell existence (or not) microscopically.
Clinical responders were defined as patients with a complete (CR) or partial response (PR) [CR: complete resolution of tumor, PR>50% regression in maximum diameter of initial tumor] after 3 cycles of NACT. Non-responders were patients with a minimal response (MR<50% regression in maximum diameter of initial tumor), no change (NC) or local progression. Pathological complete response (pCR) was defined as absence of any gross or microscopic evidence of residual tumor in the mastectomy specimen i.e. absence of residual invasive or in situ disease following NACT. Its assessment was done irrespective of the clinical response status. Clinical response was taken in to consideration for statistical analysis as the pCR was observed in only seven patients (n = 50).
Immunohistochemical methods
Biopsy specimen was preserved in buffered formalin solution and five-micron sections were prepared on poly-l-lysine coated glass slides. Sections were deparaffinized in xylene and hydrated in alcohol for 15 minutes. Further, incubation was done in 0.3% hydrogen peroxide in methanol solution for 45 minutes; the slides were washed with citrate buffer and kept in koplin jar with citrate buffer (pH-6) at 200 power in microwave (5–6 pulses). Sections were washed with Tris Buffer Saline (TBS) solution and incubated with blocking antibody (C-219 mouse monoclonal antibody) at 37°C overnight. Dilutions used were 1:20. Sections were washed with TBS solution, incubation done with avidin biotin complex (ABC) at 37°C for one hour and 3,3 Diaminobezidine tetra hydrochloride solution was applied for 3–5 minutes. Counter staining with hematoxylin solution was done for 3–5 minutes. The sections were washed with distilled water, air-dried and mounted using DPX mount.
For p-glycoprotein positive, controls were taken as positive breast cancerous cells and negative controls were taken as test slides without primary antibody. The pattern of positive staining was cytoplasmic. The monoclonal antibody used was C219 antibody (DAKO M3521) [It recognizes an epitope lying in cytoplasmic domain, 200 amino acids long, of the terminal regions of the p-glycoprotein polypeptide].
The p-glycoprotein expression was interpreted on the basis of percentage of p-glycoprotein positive cells against total population of cells [1+<25%, 2+ = 25–50 % and 3+ >50% positive cells] (figure 1, figure 2, figure 3, figure 4, figure 5, figure 6, figure 7).
Figure 1 Positive immunoreactivity for p-glycoprotein in breast carcinoma (200×).
Figure 2 Positive immunoreactivity for p-glycoprotein in breast carcinoma (400×).
Figure 3 p glycoprotein expression 1+.
Figure 4 p glycoprotein expression 2+.
Figure 5 p glycoprotein expression 3+.
Figure 6 p glycoprotein expression 3+1.
Figure 7 p glycoprotein expression 3+2.
The intensity of staining was also assessed and it was found that staining intensity correlated closely with percentage of positive cells and a single index i.e. percentage of positive cells was used for analysis.
Statistics
Descriptive studies were performed with SPSS version 10. The significance of correlation between tumor response and p-glycoprotein expression was determined with chi square test.
Results
Fifty cases of LABC were included in the study with the mean age being 43 years (range 25–60 years) and 26 patients (53.3 %) were premenopausal. The mean tumor size before NACT was 8 cm. Twenty-four (48%) had N1 disease while 26 patients (52%) presented with N2 disease in the axilla (table 1).
Table 1 Tumor size and axillary lymph node status before NACT (n = 50)
Tumor size (cm) Frequency Percent
<5 cm 8 16.7
5–8 cm 18[A1] 36.7
8–10 cm 13[A2] 26.7
10 cm 10 20
Lymph node status (N = 50) 100
N1 24 48
N2 26 52
In the biopsy specimen before initiation of NACT, 26 patients (52%) were p-glycoprotein positive and 24 were negative (n = 50). The distribution of p-glycoprotein positive patients based on grades was 1+ and 2+ in 8 patients each (16%), 3+ in 6(12%) and 4+ in 4(8%) (Additional file 1). Among the premenopausal patients 16 (59.2%) stained positive for p-glycoprotein while only 10 (43.4%) of postmenopausal patients stained positive. The difference was statistically not significant.
The clinical response was assessed using stringent World Health Organization (WHO) criteria and reduction in mean tumor size after three cycles of NACT was found to be statistically significant (p < 0.05)(table 2). Thirty patients (n = 50) were responders [complete response in 7 and partial response in 23 patients (n = 30)] while the rest 20 (40%) were non-responders (table 2).
Table 2 Mean tumor size before and after NACT (Paired sample statistics)
Status Mean (n) Std.Deviation Std. Error mean
Pre NACT tumor size(cm) 8 50 2.7 .50
Post NACT tumor size(cm) 4.3 50 2.1 .38
Paired differences t df Significance (2-tailed)
Paired samples test Mean Std. deviation Std. error mean 95% confidence interval of the difference
Pair 1: Pre NACT tumor size-Post NACT tumor size 3.677 1.6152 .2949 lower 3.074 upper 4.280 12.468 29 0.000
Significant clinical response was observed in the axillary lymph node status after NACT. There was complete response in N1 patients (n = 24) i.e. they were all down staged to N0. Amongst the patients with N2 disease (n = 26), 18(69.2%) were downstaged to N1, the rest 8(31.3%) patients showed no response (table 3). This downstaging in the axillary lymph node status was found to be statistically significant (p < 0.05).
Table 3 Axillary lymph node status before and after NACT(n = 50)
N = 50 N0 N1 N2
Before NACT nil 24(48%) 26(52%)
After NACT 24(48%) 18(36%) 8(16%)
The clinical response however did not show have any significant correlation with the pre NACT tumor size, age and menopausal status of the patients. Only seven patients (14%) showed pCR (pathological response) and significantly all these were p-glycoprotein negative. There was no statistically significant correlation observed between ER status and clinical response in the present study.
When pre NACT p-glycoprotein expression was correlated with the clinical response, it was observed that out of 30 clinical responders 21 patients (70%) were p-Glycoprotein -ve and 9(30%) were p-glycoprotein +ve while out of 20 clinical non-responders 17 patients were p-glycoprotein positive (85%). This was found to be statistically significant and as it was observed that most of the non-responders were p-glycoprotein positive.
Significantly all the seven complete responders were p-glycoprotein negative and out of nine p-glycoprotein positive patients that were partial responders, six showed very low levels of p-glycoprotein expression (1+). Thus, there was a statistically significant correlation observed between the degree of positivity of p-glycoprotein expression and poor clinical response (p < 0.05).
When p-glycoprotein positive patients were analyzed for the grades of positivity and clinical response, it was observed that increase in grade was associated with decreased response rates. There was thus an inverse relation ship observed between p-glycoprotein expression and the clinical response to NACT.
With an increase in the level of p-glycoprotein expression, the response rates dropped significantly. When the clinical response was correlated with pre-NACT p-glycoprotein expression, (With the confidence limit of 99%-p = 0.01, chi square test was applied) a statistically significant negative correlation was observed between p-glycoprotein expression and clinical response
The change in the p-glycoprotein expression before and after NACT was also found to be statistically significant in the present study. Before initiation of NACT 26 patients (52%)were p-glycoprotein positive and after three cycles of NACT, it increased to 73.5%. This chemotherapy induced p-glycoprotein positivity could possibly explain the phenomenon of acquired chemoresistance after NACT and may also serve as an intermediate end point in evaluating drug response particularly if the adjuvant therapy is planned with the same regime.
Discussion
Carcinoma of breast is a leading cause of cancer mortality in women all over the world and the second most common malignancy in India after carcinoma of the uterine cervix [1,4]. In India like in other developing countries 25–30% cases are locally advanced at the time of diagnosis [1,4]. The recommended approach for the management of LABC is a multimodality approach intended to provide both local and systemic control and studies have confirmed that surgery alone is an inadequate treatment [3]. The realization that patients with LABC are likely to have undetectable micro metastases at diagnosis has lead to systemic treatment assuming an important role, as even aggressive surgical techniques do not reduce the higher incidence of local recurrence. Most importantly surgery does not change the pattern of distant failure in these patients as they often have micrometastatic disease at the time of diagnosis [18-21].
Neoadjuvant chemotherapy was first introduced with a 70% objective response rate in 1970s and was initially utilized to convert unresectable tumors to smaller tumors making them more amenable to local control with either surgery or radiotherapy. Although the correlation between the tumor response and prognosis is still uncertain, it is generally believed that such a relationship may exist [2,16,17]. The other important advantage of NACT is that it provides an in vivo chemosensitivity test for assessment of tumor response from which prognostic information could be obtained.
Development of resistance to chemotherapeutic agents is a major and evolving problem and the search for an ideal predictor of response is still on [2]. One of the proposed contributory mechanisms is the expression of p-glycoprotein, which is encoded by a family of three genes (MDR1, MDR2 and MDR3) in rodents and two genes MDR1 (also known as PGY 1) and MDR 3 (also known on PGY3) in humans. The MDR 1 and MDR 2 gene products (class 1 and class II isoforms) are involved in drug resistance, whereas the biological properties of MDR3 gene product with class II, III isoform are not multidrug resistance [17-22]. The p-glycoprotein is a 17 Kda membrane glycoprotein which functions as an energy dependent drug efflux pump leading to poor response due to decreased accumulation of drugs inside the cells.
The electrophoretic methods such as Northern and Western blotting in detecting p-glycoprotein expression with tiny tissue samples containing very small number of p-glycoprotein expressing tumor cells have not proved satisfactory. Therefore immunohistochemical detection of p-glycoprotein using monoclonal antibodies is now widely accepted for detecting even a single p-glycoprotein expressing cell or low levels of expression, which could be missed by electrophoretic analysis [11].
The p-glycoprotein has been reported to be expressed in several normal human tissues also, notably epithelial cells with excretory/secretary functions (kidney, liver, colon), in endothelial cells at several blood tissue barrier sites (brain, testis), in secretary and gestational endometrium, in placental trophoblasts, and in adrenal glands (predominantly in the cortex). p-glycoprotein is also expressed in natural killer cells, lymphocytes, granulocytes, monocytes and in a minority of CD 34+ hematopoetic stem cells. The pattern of distribution of p-glycoprotein in normal humans suggests that its physiological role is to protect cells against xenobiotics and endogenous toxins [23-25].
Tumors arising in organs that normally express high levels of p-glycoprotein, such as kidneys, adrenals or colon are known to be intrinsically resistant to chemotherapy [9]. p-glycoprotein over expression has also been observed in leukemia's, cervix cancer as well as in soft tissue sarcomas [24]. The role of p-glycoprotein in human breast cancer is however unclear. Most of the published data suggests that p-glycoprotein expression in primary breast tumor is not a common phenomenon [23-25].
Sugawara et al detected one positive sample out of nine tumor samples using MRK-16 monoclonal antibody (Mab) in a classical immunoperoxidase staining study. In this study only a few tumor cells were stained within the positive specimen [14]. Using C219 MAb in an avidin biotin immunoperoxidase system, Schneider et al, tested 23 breast cancer specimens. None, or only minimal reactivity was found in specimens coming from untreated patients (12 cases) [9]. Dixon et al, reported no clear positivity for p-glycoprotein out of 26 primary breast tumors using 219 Mab [26] and Hyun C et al, used JSB-1 MAb and detected only 6 p-glycoprotein positive tumors out of 23(26%) primary breast tumors [26]. Verrelle et al,. used C494 MAb and detected 17 p-glycoprotein positive tumors out of 20 primary breast tumors. There was a significant negative correlation observed between p-glycoprotein expression and clinical response [18]. They had however used fresh frozen tissue unlike in the present study where paraffin embedded tissue was used. Schneider et al, [9] used C 494 and reached exactly opposite results to those observed in the study by Verrelle et al, [16-18] where the same antibody was used, this could also be because of the fact that they used fresh frozen tissue [9,18]. In the present study c219 CoAb antibody was used and the results obtained were opposite to those of Schneider et al. probably because of the same reasons.
The p-glycoprotein expression in primary breast cancer is therefore not a commonly observed phenomenon and only two reports of extraordinarily high incidence of p-glycoprotein involved in untreated breast cancer specimen, have appeared in the recent past [17,18]. Ro et al, had used 219 Mab and reported that intrinsic drug resistance (pretreatment p-glycoprotein positivity) may play a role in the failure of induction chemotherapy in locally advanced breast carcinoma.
In the present study, 26 out of 50 patients were p-glycoprotein positive (52%) and 30 patients (60%) showed clinical response to NACT however out of 30 clinical responders, 21 patients (70%) were p-glycoprotein negative. It was observed that of the 9 p-glycoprotein positive patients that were responders, 6 patients showed very low levels of p-glycoprotein expression (1+). With an increase in p-glycoprotein expression, the response rate was therefore found to drop significantly. The p-glycoprotein positivity correlated inversely with clinical response to NACT in the present study.
There are some published reports of a correlation between p-glycoprotein expression and menopausal status. The premenopausal patients have been found to have a higher p-glycoprotein expression than the postmenopausal patients [26-32]. In the present study however, there was no statistically significant correlation observed between the menopausal status and the p-glycoprotein expression. This discrepancy could have been due to a smaller sample size.
The presence of estrogen receptors provides a molecular basis for the distinction between human breast carcinoma that are responsive to hormone therapy and those that are not [31]. In various studies correlation between p-glycoprotein expression and estrogen receptor status has been found to be unclear [32]. In some studies it was found that patients whose tumors lacked ER had a higher response rate to chemotherapy [26-32]. In the present study no significant correlation between the ER status and the p-glyoprotein expression could be found although the sample size was small and most of the patients were loco regionally advanced.
The reported response rates following NACT vary between 49 to 94 % in various studies and this has been due to use of different chemotherapy combinations given at variable intervals and doses. The reported pCR rate in most studies has been 4–34(%)[2]. The largest trial (National Surgical Adjuvant Breast and Bowel Project NSABP trial B-18) reported a pCR rate of 13 % [2,33-35]. More recently pCR rate of 26–34% have been reported with the use of texanes and the 5-year overall and disease free survival rates were found to be significantly higher in the group whose primary tumor had a pCR than in the remaining responders [2,33-35].
The response rate in our study was 60% while the pCR rate was 14%(n = 7). The overall clinical response rate observed was lower than that reported in some published series (table 4). This could have been due to a smaller sample size or the fact that majority of our patients were locoregionally advanced or because of both these factors.
Table 4 Response to NACT observed in various studies (22,27–31)
Year Institute Neoadjuvant chemotherapy used Number of patients Response %
De Lena et al.[29] 1979 Southeastern cancer study group FAC 14 80
Morrow et al.[23] 1980 Guy's hospital AV 12 83
Aisner et al.[27] 1982 University of Maryland FAC 27 74
Sataloff et al.[30] 1994 Thomas Jefferson University Hospital CMF 189 85
Swain et al.[28] 1995 M.D. Anderson cancer center FAC 174 88
Singh et al.[22] 1996 PGI, Chandigarh India CMF 38 75.7
Present Study 2003–2004 VMMC, Safdarjang Hospital New Delhi FAC 50 60
AV: [Adriamycin, Vincristine]
CMF: [Cyclophosphamide. Methotrexate, 5-fluouracil]
FAC [5-fluouracil, Adriamycin, Cyclophosphamide]
There was a significant clinical response observed in the axillary lymph node status after NACT in the present study (table 3). It has also been observed in a recent NSABP B-18 trial where a significant down staging of the axillary lymph node status following NACTwas observed [35-37].
In the study of Kuerer et al, conducted at M D Anderson Cancer Center, on 372 cases of LABC the difference in the axillary lymph node status before and after NACT was not found to be statistically significant but a significant correlation has been reported between disease free and overall survival and the pathological response in the axillary lymph node status in various studies [2,38].
There are only few studies reporting an increase in the frequency of p-glycoprotein expression during or after NACT [15,16]. In the present study it was observed that before initiation of chemotherapy 26 patients (52%) were p-glycoprotein positive and after three cycles of NACT, 73.5% patients showed positivity. This increase in frequency of expression after three cycles of NACT could explain the phenomenon of acquired resistance to chemotherapy. The immunohistochemical detection of p-glycoprotein in surgical specimen after NACT would be an available tool to predict the acquired chemo resistance in breast carcinoma. It could also be used as an intermediate end point in determining drug sensitivity for adjuvant treatment, especially when adjuvant chemotherapy is planned with the same regimen as NACT.
Conclusion
This study highlights the importance of p-glycoprotein expression in predicting response to NACT in breast cancer patients. Patients with positive p-glycoprotein expression before initiation of NACT were found to be poor clinical responders. This pretreatment detection of p-glycoprotein expression may thus be utilized as a predictor of response to NACT. The increased expression of p-glycoprotein induced by the NACT probably explains the phenomenon of acquired resistance to chemotherapy and the detection of post NACT p-glycoprotein can be used as an intermediate end point in determining drug sensitivity for adjuvant treatment, especially when adjuvant chemotherapy is planned with the same regimen as induction chemotherapy. The toxic and ineffective chemotherapy may thus be avoided in non-responders.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
C was the surgeon in charge who designed the study and standardized the treatment.
JP, PK were the postgraduates in charge of the cases and contributed in the study design's
AB were responsible for the immunohistochemical analysis and molecular biology of the tissue,
AsB: contributed in the dose regulation and statistical analysis of the results.
MM was responsible for the ultrasound assessment of the breast to assess the response.
Supplementary Material
Additional File 1
The tumor response and staining pattern of tumor samples
Click here for file
Acknowledgements
The study was conducted after the approval of IRB and the Ethical committee of the hospital
==== Refs
Deo SV Bhutani M Shukla NK Raina V Rath GK Purkayasth J Randomized trial comparing neo-adjuvant versus adjuvant chemotherapy in operable locally advanced breast cancer (T4b N0-2 M0) J Surg Oncol 2003 84 192 197 14756429 10.1002/jso.10323
Charfare H Limongelli S Purushotham AD Neoadjuvant chemotherapy in breast cancer Br J Surg 2005 92 14 23 15635596 10.1002/bjs.4840
Heys SD Chaturvedi S Primary chemotherapy in breast cancer:The beginning of the end or end of the beginning for the surgical oncologist? World J Surg Oncol 2003 1 14 12932299 10.1186/1477-7819-1-14
Chintamani Singhal V Singh JP Lyall A Saxena S Bansal A Is drug-induced toxicity a good predictor of response to neoadjuvant chemotherapy in patients with breast cancer? – a prospective clinical study BMC Cancer 2004 4 48 15310398 10.1186/1471-2407-4-48
Smith IE Detre S Burton SA Salter J A'Hern R Walsh G Jounston SR Dowsett M Reduced apoptosis and proliferation and increased Bcl-2 in residual breast cancer following preoperative chemotherapy Breast Cancer Res Treat 1998 48 107 116 9596482 10.1023/A:1005933815809
Kerr JF Winterford CM Hormon BV Apoptosis: its significance in cancer and cancer therapy Cancer 1994 73 2013 2026 8156506
Hortobagyi GN Ames FC Buzdar AU Kau SW Mc Neese MD Paulus D Hug V Holmes FA Romsdahl MM Fraschini G Management of stage lIl primary breast cancer with primary chemotherapy, surgery and radiotherapy Cancer 1988 62 2507 2516 3056604
McCready Dr Hortobagyi GN Kau SW Smith TL Buzdar AU Balch CM The prognostic significance of lymph node metastases after preoperative chemotherapy for locally advanced breast cancer Arch Surg 1989 124 21 25 2910244
Gerlach JH Kartner N Bell DR Ling V Multidrug resistance Cancer Surv 1986 5 25 46 2885085
Schneider J Bak M Efferth T Kaufmann M Mattern J Volm M P-glycoprotein expression in treated and untreated human breast cancer Br J Cancer 1989 60 815 818 2574987
Chan HS Bradley G Thorner P Haddad G Gallie BL Ling V A sensitive method for immunocytochemical detection of P-glycoprotein in multidrug-resistant human ovarian carcinoma cell lines Lab Invest 1988 59 870 875 2904510
Merkel DE Fuqua SA Tandon AK Hill SM Buzdar AU McGuire WL Electrophoretic analysis of 248 clinical breast cancer specimens for p-glycoprotein overexpression or gene amplification J Clin Oncol 1989 7 1129 1136 2569033
Dalton WS Grogan TM Meltzer PS Scheper RJ Durie BG Taylor CW Miller TP Salmon SE Drug resistance in multiple myeloma and non-Hodgkin's lymphoma; Detection of p-glycoprotein circumvention by addition of verapamil to chemotherapy J Clin Oncol 1989 7 415 424 2564428
Chan HS Thorner PS Haddad G Ling V Immunohistochemical detection of p-glycoprotein: prognostic correlation in soft tissue sarcoma of childhood J Clin Oncol 1990 8 689 704 1968964
Sugawara I Kataoka I Morishita Y Hamada H Tsuruo T Itoyama S Mori S Tissue distribution of p-glycoprotein encoded by a multidrug resistant gene as revealed by a monoclonal antibody, MRK16 Cancer Res 1988 48 1926 1929 2894894
Koh EH Chung HC Lee KB Lim HY Kim JH Roh JK Min JS Lee KS Kim BS The value of immunohistochemical detection of p-glycoprotein in breast cancer before and after induction chemotherapy Yonsei Med J 1992 33 137 142 1357834
Chung HC Rha SY Kim JH Roh JK Min JS Lee KS Kim BS Lee KB p-Glycoprotein: the immediate end point of drug resistance to induction chemotherapy in locally advanced breast cancer Breast Cancer Res Treat 1997 42 65 72 9116319 10.1023/A:1005739525196
Ro J Shin A Ro JY Fritsche H Hortobagyi G Blick M Immunohistochemical analysis of P-glycoprotein expression correlated with chemotherapy resistance in locally advanced breast cancer Hum Pathol 1990 21 787 791 1974874 10.1016/0046-8177(90)90046-8
Verrelle P Meissonnier F Fonck Y Feillel V Dionet C Kwiatkowski F Plagne R Chassagne J Clinical relevance of immunohistochemical detection of multidrug resistance P-glycoprotein in breast carcinoma J Natl Cancer Inst 1991 83 111 116 1671103
Eltahir A Heys SD Hutcheon AW Sarkar TK Smith I Walker LG Ah-See AK Eremin O Treatment of large and locally advanced breast cancer using neoadjuvant chemotherapy Am J Surg 1998 175 127 132 9515529 10.1016/S0002-9610(97)00279-1
Bonadonna G Valagussa P Moliterni Zambetti M Brambilla C Adjuvant cyclophosphamide, methotrexate, and fluorouracil in node-positive breast cancer: the results of 20 years of follow-up N Engl J Med 1995 332 901 906 7877646 10.1056/NEJM199504063321401
Buzdar AU Singletary SE Booser DJ Frye DK Wasaff B Hortobagyi GN Combined modality treatment for stage III and inflammatory carcinoma of the breast. M.D.Anderson Cancer Center experience Surg Oncol Clin North Am 1995 4 715 734
Singh G Singh DP Gupta D Muralikrishna BV Neoadjuvant chemotherapy in locally advanced breast cancer J Surg Oncol 1996 61 38 41 8544458 10.1002/(SICI)1096-9098(199601)61:1<38::AID-JSO9>3.0.CO;2-T
Morrow M Braverman A Thelmo W Sohn CK Sand J Mora M Forlenza T Marti J Multimodal therapy for locally advanced breast cancer Arch Surg 1986 121 1291 1296 3778202 10.1001/archsurg.121.11.1291
Linn SC Giaccone G Van Diest PJ Blokhuis WM van der Valk P van Kalken CK Kuiper CM Pinedo HM Baak JP Prognostic relevance of p-glycoprotein expression in breast cancer Ann Oncol 1995 6 679 685 8664189
Wallner J Depisch D Hopfner M Haider K Spona J Ludwig H Pirker R MDR1 gene expression and prognostic factors in primary breast carcinomas Eur J Cancer 1991 27 1352 1355 1835847
Dixon AR Bell J Ellis IO Elston CW Blamey RW P-glycoprotein expression is locally advanced breast cancer treated by neoadjuvant chemotherapy Br J Cancer 1992 66 537 541 1355661
Aisner J Morris D Elias EG Wiernik PH Mastectomy as an adjuvant to chemotherapy for locally advanced or metastatic breast cancer Arch Surg 1982 117 882 887 6284089
Swain SM Sorace RA Bagley CS Danforth DN JrBader J Wesley MN Steinberg SM Lippman ME Neoadjuvant chemotherapy in combined modality approach of locally advanced non-metastatic breast cancer Cancer Res 1987 47 3889 3894 3036348
De Lena M Zucali R Viganotti G Valagussa P Bonadonna G Combination chemotherapy, radiotherapy approach in locally advanced breast cancer Cancer Chemother Pharmacol 1978 1 53 59 373908
Sataloff DM Mason BA Prestipino AJ Seinige UL Lieber CP Baloch Z Pathological response to induction chemotherapy in locally advanced carcinoma of breast, a determinant of out come J Am Coll Surg 1995 180 297 306 7874340
Seymour L Bezwoda WR Dansey RD p-Glycoprotein immunostaining correlates with ER and with high Ki67 expression but fails to predict anthracycline resistance in patients with advanced breast cancer Breast Cancer Res Treat 1995 36 61 69 7579508 10.1007/BF00690186
Winstanley J Cooke T George WD Murray G Holt S Croton R Griffiths K Nicholson R The long term prognostic significance of estrogen receptor analysis in early carcinoma of breast Br J Cancer 1991 64 99 101 1854633
Heys SD Hutcheon AW Sarkar TK Ogston KN Miller ID Payne S Smith I Walker LG Eremin O Aberdeen Breast Group Neoadjuvant docetaxel in breast cancer: 3-year survival results from Aberdeen trial Clin Breast Cancer 2002 3 S69 S74 12435290
Bear HD Anderson S Brown A Smith R Mamounas EP Fisher B Margolese R Theoret H Soran A Wickerham DL Wolmark N National Surgical Adjuvant Breast Bowel Project Protocol B-27 The effect on tumor response of adding sequential preoperative docetaxel to preoperative doxorubicin and cyclophosphamide:preliminary results from National Surgical Adjuvant Breast and Bowel Project Protocol B-27 J Clin Oncol 2003 21 4165 4174 14559892 10.1200/JCO.2003.12.005
Wolmark N Wang J Mamounas E Bryant J Fisher B Preoperative chemotherapy in patients with operable breast cancer: nine year results from National Surgical Adjuvant Breast and Bowel Project B-18 J Natl Cancer Inst Monogr 2001 30 96 102 11773300
Fisher B Brown A Mamounas E Wieand S Robinox A Margolese RG Cruz AB JrFisher ER Wickerham DL Wolmark N Decillis A Hoehn JL Lees AW Dimitrov NV Effect of pre-operative chemotherapy on loco-regional disease in women with operable breast cancer: findings from National Surgical Adjuvant Breast and Bowel Project B-18 J Clin Oncol 1997 15 2483 2493 9215816
Kuerer HM Newman LA Smith TL Ames FC Hunt KK Dhingra K Theriault RL Singh G Binkley SM Sneige N Buchholz TA Ross MI McNeese MD Buzdar AU Hortobagyi GN Singletary SE Clinical course of breast cancer patients with complete pathological tumor and axillary lymph node response to doxorubicin based neoadjuvant chemotherapy J Clin Oncol 1999 17 460 469 10080586
|
16164742
|
PMC1224882
|
CC BY
|
2021-01-04 16:39:04
|
no
|
World J Surg Oncol. 2005 Sep 14; 3:61
|
utf-8
|
World J Surg Oncol
| 2,005 |
10.1186/1477-7819-3-61
|
oa_comm
|
==== Front
Environ HealthEnvironmental Health1476-069XBioMed Central London 1476-069X-4-121601417710.1186/1476-069X-4-12ResearchA retrospective study of PBDEs and PCBs in human milk from the Faroe Islands Fängström Britta [email protected] Anna [email protected] Philippe [email protected] Pál [email protected] Åke [email protected] Department of Environmental Chemistry, Stockholm University, SE-106 91 Stockholm, Sweden2 Institute of Public Health, University of Southern Denmark, DK-5000 Odense, Denmark3 Department of Environmental Health, Harvard School of Public Health, Boston, MA 02215, USA4 Faroese Hospital System, FR-100 Tórshavn, Faroe Islands2005 14 7 2005 4 12 12 12 4 2005 14 7 2005 Copyright © 2005 Fängström et al; licensee BioMed Central Ltd.2005Fängström et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Persistent organic pollutants (POPs) in wildlife and humans remain a cause of global concern, both in regard to traditional POPs, such as the polychlorinated biphenyls (PCBs), and emerging POPs, such as the polybrominated diphenyl ethers (PBDEs). To determine the time related concentrations, we analyzed human milk for these substances at three time points between 1987 and 1999. Polychlorobiphenylols (OH-PCBs), the dominating class of PCB metabolites, some of which are known to be strongly retained in human blood, were also included in the assessment.
Methods
We obtained milk from the Faroe Islands, where the population is exposed to POPs from their traditional diet (which may include pilot whale blubber). In addition to three pools, nine individual samples from the last time point were also analyzed. After cleanup, partitioning of neutral and acidic compounds, and separation of chemical classes, the analyses were carried out by gas chromatography and/or gas chromatography/mass spectrometry.
Results
Compared to other European populations, the human milk had high PCB concentrations, with pool concentrations of 2300 ng/g fat 1987, 1600 ng/g fat in 1994, and 1800 ng/g fat in 1999 (based on the sum of eleven major PCB congeners). The nine individual samples showed great variation in PCB concentrations. The OH-PCBs were present in trace amounts only, at levels of approximately 1% of the PCB concentrations. The PBDE concentrations showed a clear increase over time, and their concentrations in human milk from 1999 are among the highest reported so far from Europe, with results of individual samples ranging from 4.7 to 13 ng/g fat
Conclusion
Although remote from pollution sources, the Faroe Islands show high concentrations of POPs in human milk, particularly PCBs, but also PBDEs. The PBDEs show increasing concentrations over time. The OH-PCB metabolites are poorly transferred to human milk, which likely is related to their acidic character.
==== Body
Background
Even though persistent organic pollutants (POPs) have been regulated in most countries in the world, in agreement with the Stockholm convention in 2001 [1], the POPs are still posing a global environmental threat on the environment, both to wildlife and humans [2,3]. This is particularly true at hot spot locations and in remote areas where dietary habits may influence uptake and accumulation of POPs. This can be exemplified with human exposures to polychlorinated biphenyls (PCBs) at two former PCB production sites in Michalovche, Slovakia, and in Anniston, Alabama, U.S.A [4,5]. Populations with elevated levels of PCB due to food intake include populations frequently consuming fish from the Baltic Sea [6,7] and Lake Michigan, [8,9], Inuits from Northern Quebec [10,11] and Greenland [12], and the residents of the Faroe Islands [13,14]. The aim of the present study was to indicate any concentraion changes of two major classes of environmental contaminants in humans, as determined in human milk from the Faroe Islands, sampled at three different time points between 1987 and 1999. The chemicals chosen for analysis were PCBs and polybrominated diphenyl ethers (PBDEs), as well as polychlorobiphenylols (OH-PCBs) as the major class of PCB metabolites.
The Faroe Islands is a small community in the North Atlantic, located north of Scotland, far from industrial sources of any POPs, including both the PCBs and PBDEs. The Faroese population is exposed to these substances through their food, and possibly via goods and products in their homes and work environment. Their lifestyle is entirely Western and their diet is generally based on seafood. Part of the population has a considerable intake of traditional local food, such as pilot whale meat and blubber, and seabirds (e.g., fulmar meat and eggs) [13,15,16]. These high-trophic-level marine species are known to contain high concentrations of POPs (e.g. PCBs and PBDEs) [15-17]. The PBDE concentrations in pilot whale are in a concentration range similar to those reported in marine mammals from other parts of the world and considerably higher than those found in most marine mammals from the arctic [18,19]. Because of dietary differences among the Faroese, a wide variability in PCB concentrations is present. In 1987 milk pools contained 1.8–3.5 μg PCB per gram fat [20], serum from pregnant Faroese women in 1994 confirmed the overall average, but individual PCB concentrations ranged from 0.15 up to as much as 22 μg/g fat [13]. In subjects from the latter study, OH-PCB concentrations of 0.019 to 1.8 μg/g fat were reported, thereby showing the importance of this class of PCB metabolites retained in human blood [13].
PBDEs were first reported in fish from Sweden [21] and subsequently in Japan [22] and in human milk from Germany [23]. These initial results have been followed by a large number of studies world-wide; in humans, fish, birds and mammals [19,24]. In a time-trend study for PBDEs in human milk from Sweden from the early 1970s to 1997, the concentrations showed a significant increase [25], while samples from 1997 to 2000 indicate a decrease, mainly due to reduced concentrations of BDE-47 [26]. A decrease of PBDEs in Swedish milk was also indicated by Lind and co-workers 2003 [27]. A good overview of data on human PBDE exposures were given in recent reviews, including those by Hites [19], Sjödin et al. [28] and Ryan [29]. In Norway, PBDE in human milk increased from 1986 to 2001, with similar concentration levels to those reported in Sweden and Japan [25,26,30,31]. In the Canadian Artic, the PBDE concentrations have increased with a factor of 3 during the 1990s (i.e., from 2.2 to 6.2 ng/g fat) [32]. A similar concentration range has been reported from the United Kingdom. In southern Canada, the concentrations were about 3 times higher [29,32], and the PBDE levels reported in human milk from the United States are about 2–3 times higher again than those from Southern Canada and ~20 times higher than those seen in Europe and Japan [19,25,26,29,31,33,34].
PCBs still constitute some of the dominating environmental contaminants among all POPs [3]. The human concentrations of PCBs have decreased over time, even though the changes seem to be less than for 2,2-bis(4-chlorophenyl)-1,1-dichloroethene (DDE). These reductions in human PCB concentrations are related to the legislative measures taken in the 1970s. Temporal studies on the decrease of PCB levels, as studied in Swedish mothers milk from the early 1970s through 1997, indicate a 30% decrease [35]. In Germany, the PCB concentrations in mother's milk have decreased by 60% from 1986 to 1997 [36]. Also in Germany, Fürst and co-workers showed PCB decreases of up to 85% from 1984 to 2003 [37] (pers. comm. P. Fürst). Declining PCB concentrations in human milk have also been observed in Norway [38] and Canada [39], with different rates of the decrease in each country.
Materials and methods
Samples
In connection with formation of mother-child pair cohorts in 1987, 1994/95, and 1999 from consecutive births at the National Hospital in Tórshavn, Faroe Islands, milk was collected on days 3–5 post partum and then deep frozen. For the present study, mothers were selected from the three cohorts using the following inclusion criteria: age between 20 and 29 years, parity no more than one, singleton birth, and delivery at term. Similar procedures were used in 1987, 1994, and 1999 [20,40]. Three pooled milk samples were generated from the three cohorts, each containing the same amount of milk from ten mothers (2 ml from each mother). Because dietary habits are changing, women were further selected for inclusion in the study from information on their diet during pregnancy. The pooled samples contained milk from four women who did not eat whale meat at all during pregnancy, and two each who ate whale meat once, 2–3 times, and at least 4 times per month. However, due to lack of sufficient sample volume, the pool from 1994 contained milk from only three women who had not eaten whale meat at all. To determine the full range and variability of POP exposure, nine individual samples taken from the same women selected in the pool from the most recent cohort (1999) were also analysed. Due to limited volumes of milk available, specimens of at least 10 ml were taken from the milk samples collected 2–3 weeks after delivery. Additionally, due to changes in dietary habits in the Faroese [14], the samples selected from 1999 may not be fully representative of Faroese pregnant women, but they are comparable to the previously collected samples, in regard to traditional dietary habits. All samples were analyzed for major PBDE, PCB and OH-PCB congeners.
Chemicals
The individual PBDE congeners (numbered according to Ballschmiter et al. [41]): BDE-47, BDE-77, BDE-99, BDE-100, BDE-153, BDE-154, and BDE-209 were synthesized in-house [42]. The individual PCB congeners: CB-101, CB-105, CB-118, CB-128, CB-138, CB-146, CB-153, CB-156, CB-170, CB-180, CB-183 and CB-200 [41] were purchased from Larodan Fine chemicals AB in Malmö, Sweden. The hydroxylated PCB standards 4-OH-CB146, 4-OH-CB187 and 4-OH-CB193 were synthesized (as described elsewhere [43]) and abbreviated according to Letcher et al. [44]. All solvents were of pro analysis quality. 2-Propanol from AnalaR (BDH laboratory supplies pool, England) and methyl tert-butyl ether (HPLC-grade; Rathburn, Walkerburn, Scotland) were glass-distilled prior to use. Silica gel (<0.063 mm) was purchased from Merck (Darmstadt, Germany) and activated (300°C, 12 h) before use.
Instruments
The PBDE analysis was performed by gas chromatography/mass spectrometry (GC/MS) using a Finnigan TSQ 700 instrument (ThermoFinnigan, Bremen, Germany) connected to a Varian 3400 gas chromatograph equipped with an AS200S CTC autosampler. The transfer line temperature was set to 290°C and the ion source temperature maintained at 200°C. On-column injections were made using a septum-equipped temperature programmable injector (SPI) fitted with a high performance insert directly connected to a DB-5 HT capillary column (15 m × 0.25 mm i.d., 0.1 μm film thickness; J&W Scientific) with helium as carrier gas at a head pressure of 3 psi. The injector was temperature programmed from 60°C to 320°C at 150°C/min and the oven from 80°C (1 min), 15°C/min to 300°C (16 min). The PBDE congeners were analysed with selected ion monitoring (SIM) by scanning for the negative bromide ion (isotopes m/z 79 and 81) formed by electron capture reactions at chemical ionization (ECNI) with methane (5.0, AGA, Stockholm, Sweden) as the electron thermalization buffer gas at 5.6 torr and a primary electron energy of 70 eV. All chromatographic data were collected, analysed and quantified using the proprietary ICIS2 software from Thermofinnigan.
The PCB and OH-PCB analyses were performed on a Varian 3400 gas chromatograph, equipped with a Varian 8200 autosampler, an electron capture detector (ECD), and a split-splitless injector operated in the splitless mode. Hydrogen was used as a carrier gas and nitrogen as a make-up gas. A CP-Sil-8-column (25 m × 0.15 mm internal diameter and 0.12 μm film thickness Chrompack, EA Middleburg, The Netherlands) was used. For the PCB analysis, the column temperature was 80°C (1 min), 20°C/min, to 300°C (5 min) and for the OH-PCB analysis the column temperature was 80°C (1 min), 50°C/min, to 200°C (1 min), 1°C/min, to 230°C, 50°C/min, to 330°C (2 min). The injector temperature was 280°C and the detector temperature 360°C. The data were collected using a PC-based ELDS Pro v2.0 system (Chromatograhic Data System AB, Stockholm, Sweden). The linear relationship of the GC/ECD and GC/MS system was determined and the quantifications were performed using a single point external standard within the concentration range of the linear relationship.
Extraction and cleanup procedure
The extraction and cleanup procedure for the milk samples is a modified version of a method for serum samples first described and validated by Hovander and co-workers [45]. In the modified version of the method, formic acid and diethyl ether are used instead of hydrochloric acid and methyl tert-butyl ether. Surrogate standards, BDE-77, CB-200 and 4-OH-CB193 were added to the samples prior to extraction. In short, formic acid (1 ml) and 2-propanol (6 ml) were added to a milk sample (5 g), subsequently extracted with a mixture of n-hexane/diethyl ether (1:1, 6 ml), and re-extracted once (3 ml). The lipid content was determined gravimetrically after gentle evaporation of the solvent. The neutral and phenol-type compounds were separated by partitioning with a potassium hydroxide solution. The bulk of lipids in the neutral fraction was removed with concentrated sulfuric acid and additional cleanup was performed on two subsequently applied sulfuric acid/silica gel columns, according to Hovander et al. [45]. Potential OH-PCB congeners were derivatized with diazomethane and any remaining lipids in the methylated phenol fraction were removed as described elsewhere [45]. All samples were protected from daylight during handling and storage to prevent any photochemical degradation of the brominated compounds to be analyzed.
Solvent blank samples representing every fifth sample were cleaned up and analyzed in the same way as the other samples. The PBDEs in the sample had to be three times greater than the PBDE amount in the blank to be considered as a quantifiable amount. The average blank sample amount has been subtracted from the results. The overall recoveries and standard deviation (SD) of the surrogate standards were 83% ± 6.1 for BDE-77, 83% ± 4.2 for CB-200 and 104% ± 6.7 for 4-OH-CB193.
Recovery experiment
The recovery study was performed on 5 g of cow milk with a fat content at 3%. Before extraction a selected number of PBDE congeners were added, BDE-47, BDE-99, BDE-100, BDE-153, BDE-154 and BDE-209 at two different spike levels (0.1 ng/sample and 1 ng/sample) both in triplicates. In parallel, six samples were extracted without any PBDE congeners added. Prior to analysis by GC/MS, the same PBDE congeners (BDE-47, BDE-99, BDE-100, BDE-153, BDE-154 and BDE-209) were added to the samples at the two levels (0.1 ng/sample and 1 ng/sample). These samples therefore represent 100% of the analytes, and from these, the absolute recoveries were calculated. In addition, three more samples were run in parallel as blank samples for control of any background contamination.
Results
The duplicate results (A/B) for PBDE congener concentrations in pooled milk samples from 1987, 1994/95 and 1999 are presented in Table 1, as are the mean/medians and ranges for the nine individual samples from 1999. All PBDE results except for one were above the limit of quantification (LOQ), a single individual milk sample having a non-detectable BDE-209 concentration. The PBDE results for all individual samples from 1999 exceed the pool value from 1987, except for BDE-209 and BDE-154/BB-153. Likewise, all total PBDE concentrations of the individual samples exceed the totals from both 1987 and 1994/95. The time related levels of PBDEs (Fig. 1) therefore seems to reflect a real and substantial increase. While this is obvious for BDE-47, an even more dramatic increase has occurred for BDE-153, especially when compared to relevant data from Sweden (Fig. 2).
Table 1 Concentrations of the PBDE congeners identified in human milk from the Faroe Islands. The samples were pooled with ten mothers in each pool, two samples (A/B) of the same pool were analyzed in parallel. ΣPBDE consist of BDE-47, BDE-99, BDE-100, BDE-153 and BDE-209.
Concentration (ng/g fat) Concentration (pmol/g fat)
Year 1987 1994/95 1999 1999 1987 1994/95 1999 1999
Samples (A/B)a (A/B)a (A/B)a n = 9b (A/B)a (A/B)a (A/B)a n = 9b
mean/median (range) mean/median (range)
BDE-47 0.37/0.40 1.1/1.1 1.7/1.6 1.9/1.3 (0.90–4.5) 0.77/0.83 2.2/2.2 3.5/3.4 4.0/2.6 (1.8–9.2)
BDE-99 0.15/0.17 0.50/0.51 0.91/0.94 0.84/0.73 (0.33–1.8) 0.26/0.30 0.88/0.90 1.6/1.7 1.5/1.3 (0.59–3.3)
BDE-100 0.23/0.26 0.58/0.59 0.92/0.96 1.0/0.58 (0.30–2.8) 0.41/0.45 1.0/1.1 1.6/1.7 1.8/1.0 (0.53–4.9)
BDE-153 0.57/0.65 1.4/1.4 3.3/3.6 2.4/2.1 (1.5–3.8) 0.89/1.0 2.1/2.2 5.1/5.5 3.8/3.2 (2.3–6.0)
BDE-209 0.59/0.59 0.47/0.55 1.1/1.3 1.0/0.60 (<0.14c–3.2) 0.62/0.62 0.49/0.58 1.2/1.3 1.1/0.63 (<0.15c–3.4)
ΣPBDE 1.9/2.1 4.0/4.2 8.0/8.4 7.2/5.8 (4.7–13) 2.9/3.2 6.8/6.9 13/14 12/8.9 (7.9–23)
BDE-154/BB-153 1.8/2.1 1.8/1.8 2.5/2.6 1.9/1.3 (0.66–4.0) 2.9/3.3 2.7/2.8 3.9/4.0 3.0/2.1 (1.0–6.3)
aThe samples were pooled with 10 mothers in each pool, two samples of the same pool were analyzed in parallel.
bThe nine individual samples are the same samples as those included in the pool, except one missing.
cLOQ for BDE-209
Figure 1 PBDE concentrations in pooled milk samples from the Faroese generated in 1987, 1994/95 and 1999. PBDE congener (BDE-47, BDE-99, BDE-100, BDE-153 and BDE-209) concentrations (pmol/g fat) in three pooled milk samples from mother-child pair cohorts generated in 1987, 1994/95 and 1999. Each pool consisted of 10 mothers and two samples from the same pool were analyzed in parallel.
Figure 2 A time related comparison of BDE-47 and BDE-153 in Swedish mother's milk and the present study. Three pooled milk samples from mother-child cohorts generated in 1987, 1994/95 and 1999 from the Faroe Islands are compared to the Swedish pooled milk samples [25,26].
The concentrations of eleven major PCB congeners in pooled milk samples and the median and ranges for the nine individual samples from 1999 are given in Table 2. All concentration data are presented both on a weight and molar basis to promote correct comparisons of concentrations of these compounds with their highly different molar masses. In contrast to the tendency seen in PBDE concentrations, the PCB results are rather stable (Fig. 3), with the three pools showing quite similar results, and all pool data well within the variability seen in the individual samples from 1999.
Table 2 Concentrations of the PCB and OH-PCB congeners identified in human milk from the Faroe Islands. The samples were pooled with 10 mothers in each pool, duplicate samples (A/B) were analyzed in parallel. ΣPCB consists of CB-105, CB-118, CB-128, CB-138, CB-146, CB-153, CB-156, CB-170, CB- 180, CB-183 and CB-187 and ∑OH-PCB is the sum of 4-OH-CB146 and 4-OH-CB187.
Concentration (ng/g fat) Concentration (pmol/g fat)
Year 1987 1994/95 1999 1999 1987 1994/95 1999 1999
Samples (A/B)a (A/B)a (A/B)a n = 9b (A/B)a (A/B)a (A/B)a n = 9b
mean/median (range) mean/median (range)
CB-105 32/39 28/27 27/32 31/19 (9.6–89) 98/120 85/83 83/98 95/57 (29–270)
CB-118 170/180 110/110 140/140 150/96 (42–430) 520/560 350/330 430/430 460/300 (130–1300)
CB-128/CB-167 30/31 21/21 22/23 22/17 (8.0–43) 82/85 58/59 62/63 60/48 (22–120)
CB-138 490/510 380/370 420/420 430/350 (160–1100) 1400/1400 1000/1000 1200/1200 1200/970 (450–3100)
CB-146 110/100 83/84 100/100 150/130 (55–450) 290/290 230/230 280/280 420/350 (150–1300)
CB-153 570/610 420/410 470/470 460/380 (180–1100) 1600/1700 1200/1100 1300/1300 1300/1100 (500–3000)
CB-156 85/88 53/50 56/55 52/42 (21–120) 230/240 150/140 150/150 140/120 (59–340)
CB-170 170/170 100/100 110/110 100/80 (41–230) 430/440 270/250 290/290 260/200 (100–570)
CB-180 380/400 250/240 270/270 250/210 (99–550) 970/1000 620/610 690/680 640/530 (250–1400)
CB-183 50/53 38/37 41/41 37/31 (14–88) 130/130 95/94 100/100 95/78 (36–220)
CB-187 160/170 130/130 140/140 130/110 (56–340) 410/420 330/330 360/360 330/270 (140–860)
ΣPCB 2200/2400 1600/1600 1800/1800 1800/1500(690–4600) 6100/6400 4400/4300 4900/4900 5000/4000 (1900–12000)
4-OH-CB146 0.80/0.87 0.33/0.26 0.35/0.40 0.66/0.36 (0.18–1.7) 2.1/2.3 0.88/0.69 0.93/1.1 1.7/0.96 (0.50–4.4)
4-OH-CB187 1.2/1.3 0.48/0.44 0.50/0.45 0.80/0.42 (0.26–1.8) 3.0/3.1 1.2/1.1 1.2/1.1 2.0/1.0 (0.64–4.5)
ΣOH-PCB 2.0/2.1 0.81/0.70 0.85/0.85 1.4/0.75 (0.45–3.5) 5.1/5.4 2.1/1.8 2.1/2.2 3.7/2.0 (1.1–8.9)
aThe samples were pooled with 10 mothers in each pool, two samples of the same pool were analyzed in parallel
bThe nine individual samples are the same samples as those included in the pool, except one missing.
Figure 3 PCB concentrations in pooled milk samples from the Faroese generated in 1987, 1994/95 and 1999. PCB congener concentrations in pmol/g fat (CB-105, CB-118, CB-128, CB-138, CB-146, CB-153, CB-156, CB-170, CB-180, CB-183 and CB-187) in three pooled milk samples from a mother-child pair cohorts generated in 1987, 1994/95 and 1999. Each pool consisted of 10 mothers and two samples from the same pool were analyzed in parallel.
Concentrations of two individual OH-PCB congeners (4-OH-CB146 and 4-OH-CB187) detected for all samples analyzed are presented in Table 2 (bottom). The ratio between the dominating OH-PCB metabolite and the PCB congener present in the highest concentration, 4-OH-CB187/CB-153 in the milk was below 0.002 in all cases.
The recoveries of BDE-47, BDE-99, BDE-100, BDE-153, BDE-154 and BDE-209 (Table 3) were about 90%, within a range of 79–107%, except for BDE-209 for which the recovery range was somewhat wider (86–160%). As an additional reflection of the analytical quality, the mean concentrations of the nine individual samples are very close to the value obtained from the pool, despite the fact that one of the single samples was lost. Although the individual samples were selected to represent differences in recent diets, there was no clear association between recent diet and the POP results obtained.
Table 3 Results from the PBDE recovery study performed in milk. PBDEs were added at two concentration levels.
Low level a High level b
0.1 ng/sample 1 ng/sample
BDE-47 79 – 82 84 – 86
BDE-99 98 – 100 88 – 88
BDE-100 88 – 90 87 – 88
BDE-153 104 – 107 86 – 94
BDE-154 103 – 107 87 – 92
BDE-209 116 – 160c 86 – 105
a Triplicates samples b Double samples c Only double samples
Discussion
The concentrations of each of the PBDE congeners appear to increase substantially in human milk from the Faroese women studied between 1987 and 1999 (Table 1 and Fig. 1), most likely independent of a skewed sampling strategy. The variability of the ΣPBDE concentrations is apparent from analyses of the nine individual samples from the most recent group of samples (1999). The rather wide range between the highest and the lowest ΣPBDE concentration could conceivably be due to differences in life style and dietary habits, as has previously been demonstrated for PCB concentrations. Other studies have also revealed high variations in PBDE concentrations between individuals [27,34,46,47]. Previous studies have also shown increasing PBDE concentrations with time. For example, Merionyté and co-workers [25,26] showed a similar trend for BDE-47, but a very different trend for BDE-153 (Fig. 2). The apparent high concentrations and/or increase over time of BDE-153 in the Faroese is more pronounced than reported in most previous studies [28,31,34] even though this congener has occasionally been shown to dominate among the PBDE congeners in blood from humans in the Netherlands [47] and the human milk from the U.S. [48]. Very recently BDE-153 was pointed out as the most common PBDE congener in a few human adipose tissue samples [49]. The pattern with a high concentration of BDE-153 was seen in seven out of nine of the individual milk samples from 1999. The other PBDE congeners were present at concentration levels as expected from previously published data.
BDE-209 is a PBDE congener with an apparent short half-life (15 days) [50], and its presence in concentrations up to 3.2 ng/g fat, with only one sample below LOQ (Table 1) is an indication of continuous exposure to this PBDE congener. The median concentration of BDE-209 in 1999 is approximately half the concentration in serum from Swedish abattoir workers (2.5 ng/g fat) from 2000 [51]. BDE-209 was also recently detected at very low concentrations in some human milk samples from Germany collected in 2001–2003, with a median at 0.1 ng/g fat [52]. A mean of 0.9 ng/g fat was found in U.S. mothers [34], but only seven samples out of 47 were above the LOQ. In the Faroese milk samples, BDE-209 was above LOQ in all but one sample and at a mean concentration of approx. 1 ng/g fat at the last time point 1999. This is twice the concentrations of the milk pools from 1987 and 1994/95 (Table 1).
The ΣPBDE median concentrations of the nine individual samples and the level of the pool sample of the Faroe Islands mother's milk from 1999 are in the same range as milk samples from the United Kingdom mothers sampled in 2002 [33] and German samples from 2002 [53], but higher than those reported by Fürst and co-workers from southern Germany [37]. The Faroese human milk PBDE levels are intermediate between European and North American concentrations, with the exceptions mentioned [34]. Even though it may be convenient to compare overall ΣPBDE concentrations, individual PBDE congener concentrations (Fig. 1) are much more important, because the compounds differ in their toxicities. Depending on the large molecular weight span of the PBDE congeners, and given the high atomic weight of bromine, the PBDE congener concentrations are most correctly reported on a molar basis and accordingly comparisons are most correctly done this way.
The PCB concentrations (Table 2) in the Faroese human milk are considerably higher than those reported from other studies of human milk (e.g., in Sweden and Belgium), which are exceeded by a factor of about five in the milk from the Faroe Islands [35,54]. The PCB time trend observed in Swedish human milk showed a steady decrease, with the PCB concentration in 1997 reaching about 30% of that in 1972 [35]. In Germany, a decline to 10–25% of the PCB concentration occurred from 1984 and 2003 (P. Fürst, pers. comm.). In an Inuit population (Nunavik, Québec, Canada), the PCB concentration in cord blood were shown to decrease by about 7% per year from 1994 to 2001 [11]. All of these studies indicate decreases of PCBs in contrast to the relatively constant concentrations found in the present study among women in the Faroe Islands. However, a wide variation in the ΣPCB concentration occurs between the individual samples from the most recent group of the Faroese samples, and a small average decrease cannot be excluded. The relatively high concentration of CB-118 is notable, although the human milk PCB profile is dominated by CB-138, -153 and -180 (Table 2 and Fig. 3). Additional analyses would be necessary to describe the temporal trend of PCBs in human samples in the Faroe Islands.
The OH-PCB concentrations are low in mother's milk (Table 2). This finding is in accordance with results previously reported from Sweden and in Canada [55,56]. Hence 4-OH-CB187 was the dominating congener, followed by 4-OH-CB146 in the Faroese milk. The low concentrations are also to be anticipated, because these metabolites are mainly protein bound in the blood [44] and not dissolved or partitioned to lipids as are their parent compounds. Accordingly, the OH-PCB exposure of a nursing child is very low, despite the toxicant being easily transferred to the fetus [55,57]. These characteristics make the OH-PCB different from the neutral PCBs and PBDEs.
Despite substantial decreases elsewhere, Faroese human milk still contains high concentrations of PCBs, with levels in the low ppm range. The major PCB congener in human milk (CB-153) exceeds the most abundant PBDE congener (BDE-153) by a factor of 150 in the most recent samples (Tables 1 and 2). These high PCB concentrations are likely explained by dietary habits, given that the Faroe Island population has a high seafood intake, which includes PCB and PBDE contaminated pilot whale blubber, fulmars and fulmar eggs [15-17].
The high concentrations and unique pattern of PBDEs, with BDE-153 being the dominant congener, could possibly be explained by a higher persistence of BDE-153 than of BDE-47. In contrast, pilot whale blubber has been shown to contain mainly BDE-47, followed by BDE-99, then BDE-153, thus indicating that this food item may not be the primary source of human PBDE exposure. Since fulmars contain only low PBDE concentrations, these birds also constitute only a minor source for human PBDE intake. The high PBDE concentrations in 1999 are unlikely to be explained by exposures to electronic and electric equipments, since such exposures would not exceed those present in the rest of Europe. It is possible that at least some PBDEs are passed through marine food chains, and that the Faroese are highly exposed from their traditional seafood diet.
A study from a Northern Canadian population showed that the traditional food sources in this area did not necessarily result in higher concentrations of PBDEs, as observed for other persistent organic pollutants (POPs), such as the PCBs [58]. Arctic mammals, including ringed seals, have shown relatively low levels of PBDEs, ranging from 0.40 to 4.3 ng/g fat, as compared to pilot whale blubber from the Faroe Islands, where PBDE concentrations were 0.84 to 3.2 μg/g fat [17,18]. Still, the fact remains that residents at the Faroe Islands have high concentrations of both PCB and PBDE, and the difficulty in explaining the sources of the PBDEs suggest that this issue needs further attention.
The method used for the PBDE analysis in human milk has previously been used for analysis of PCBs in human milk and was formerly used also for both organochlorine and organobrominated compounds in human serum. The human serum method was adopted for human milk analysis with a few modifications [45]. The recovery study was performed with cow's milk [59], because the PBDE concentrations in human milk are much higher and would not be suitable for blank samples. A small pilot study was performed to confirm that the cow's milk was virtually free from any PBDEs, which was the case. Milk with 3% fat was used, because human milk normally has a lipid concentration of about that level. The recovery study was performed only for the PBDEs, because the analytical method for PCBs has been previously documented [60]. The average recoveries of BDE-47, BDE-99, BDE-100, BDE-153 and BDE-154 added to the cow milk before extraction ranged from 79 – 107% (Table 3).
Conclusion
This study reports PBDE concentrations in human milk from the Faroe Islands, for the first time. A steep increase of PBDE concentrations is shown from 1987 to 1999, and the concentrations in milk from the late 1990s are among the highest in Europe. The PBDE pattern is different from the one reported elsewhere, with BDE-153 as the dominant congener, rather than BDE-47 [25,30]. The PBDE sources for the Faroese population still need to be identified, but could include food items affected by passage through marine food chains. Furthermore, the PCB concentrations remain high in Faroese mother's milk with no clear decrease over time up to the late 1990s. The hydroxylated metabolites of PCBs are poorly transferred to mother's milk leading to very low exposures to these metabolites via the milk.
List of abbreviations
GC/ECD gas chromatography / electron capture detector
GC/MS gas chromatography / mass spectrometry
OH-PCBs polychlorobiphenylols
PBDEs polybrominated diphenyl ethers
PCBs polychlorinated biphenyls
POPs persistent organic pollutants
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
BF led the writing of the manuscript, assisted in the clean-up procedure, the GC/ECD and GC/MS analysis of the samples, and designed the project. AS did the cleanup work for the samples and made the GC/ECD analysis. PG participated in the writing and design of the project. PW collected and selected all the samples. ÅB participated in planning, writing, and interpretation. All authors read and approved the final manuscript.
Acknowledgements
We are grateful to Ioannis Athanassiadis for doing the GC/MS analysis. We are also grateful to the laboratory personnel at University of Southern Denmark for preparing the pooled samples. Financial support for this study has been given by grants from the EU R&D programme, Anemone (QLK4-CT-2001-00186).
==== Refs
UNEP Final act of the conference of plenipotentaries on the Stockholm convention on persistent organic pollutants http://www chem unep ch/pops/ 2001 Stockholm, Sweden, United Nations Environment Programme
AMAP Jensson H Arctic Pollution 2002 2002 Oslo, AMAP
WHO Damstra T, Barlow S, Bergman , Kavlock R and van der Kraak G Global assessment of the state-of-the-science of endocrine disruptors 2002 International Programme on Chemical Safety, WHO
Hansen LG DeCaprio AP Nisbet ICT PCB congener comparisons reveal exposure histories for residents of Anniston, Alabama, USA Fresenius Envir Bull 2003 12 181 190
Pavuk M Cerhan JR Lynch CF Schecter A Petrik J Chovancova J Kocan A Environmental exposure to PCBs and cancer incidence in eastern Slovakia Chemosphere 2004 54 1509 1520 14659953 10.1016/j.chemosphere.2003.10.038
Grimvall E Rylander L Nilsson-Ehle P Nilsson U Strömberg U Hagmar L Östman C Monitoring of polychlorinated biphenyls in human blood plasma: Metodological developments and influence of age, lactation, and fish consumption Arch Environ Contam Toxicol 1997 32 329 336 9096084 10.1007/s002449900193
Sjödin A Hagmar L Klasson Wehler E Björk J Bergman Influence of the consumption of fatty Baltic Sea fish on plasma levels of halogenated environmental contaminants in Latvian and Swedish men Environ Health Perspect 2000 108 1035 1041 11102293
Humphrey HE Gardiner JC Pandya JR Sweeney AM Gasior DM McCaffrey RJ Schantz SL PCB Congener Profile in the Serum of Humans Consuming Great Lakes Fish Environ Health Perspect 2000 108 167 172 10656858
He JP Stein AD Humphrey HE Paneth N Courval J Time Trends in Sport-Caught Great Lakes Fish Consumption and Serum Polychlorinated Biphenyl Levels among Michigan Anglers, 1973-1993 Environ Sci Technol 2001 35 435 440 11351711 10.1021/es001067p
Dewailly E Ayotte P Bruneau S Laliberté C Muir D Norstrom R Inuit Exposure to Organochlorines through the Aquatic Food Chain in Arctic Québec Environ Health Perspect 1993 101 618 620 8143594
Dallaire F Dewailly E Muckle G Ayotte P Time Trends of Persistent Organic Pollutants and Heavy Metals in Umbilical Cord Blood of Inuit Infants Born in Nunavik (Québec, Canada) between 1994 and 2001 Environ Health Perspect 2003 111 1660 1664 14527847
Bjerregaard P Dewailly E Ayotte P Pars T Ferron L Mulvad G Exposure of Inuit in Greenland to organochlorines through the marine diet Journal of Toxicology and Environmental Health, Part A 2001 62 69 81 11209822 10.1080/009841001455490
Fängström B Athanasiadou M Grandjean P Weihe P Bergman Hydroxylated PCB metabolites and PCBs in serum from pregnant Faroese women Environ Health Perspect 2002 110 895 899 12204824
Weihe P Hoppe HW Grandjean P Sustained high concentrations of PCBs in Faroese pregnant women despite dietary intervention. Organohalogen Compounds 2003 63 Boston, MA, USA 389 392
Fängström B Athanasiadou M Athanassiadis I Bignert A Grandjean P Weihe P Bergman Polybrominated diphenyl ethers and traditional organochlorine pollutants in fulmars (Fulmarus glacialis) from the Faroe Islands Chemosphere 2005 60 836 843 15992590 10.1016/j.chemosphere.2005.01.065
Dam M Bloch D Screening of Mercury and Persistent Organochlorine Pollutants in Long-Finned Pilot Whale (Globicephala melas) in the Faroe Islands Mar Pollut Bull 2000 40 1090 1099 10.1016/S0025-326X(00)00060-6
Lindström G Wingfors H Dam M van Bavel B Identification of 19 polybrominated diphenyl ethers (PBDEs) in long-finned pilot whale (Globicephala melas) from the Atlantic Arch Environ Contam Toxicol 1999 36 355 363 10047605 10.1007/s002449900482
Ikonomou MG Rayne S Addison RF Exponential increases of the brominated flame retardants, polybrominated diphenyl ethers, in the Canadian arctic from 1981 to 2000 Environ Sci Technol 2002 36 1886 1892 12026966 10.1021/es011401x
Hites RA Polybrominated diphenyl ethers in the environment and in people: a meta-analysis of concentrations Environ Sci Technol 2004 38 945 956 14998004 10.1021/es049548m
Grandjean P Weihe P Needham LL Burse VW Patterson DGJ Sampson EJ Jorgensen PJ Vahter M Relation of a seafood diet to Mercury, Selenium, Arsenic and Polychlorinated Biphenyl and other organochlorine concentrations in human milk Environ Res 1995 71 29 38 8757236 10.1006/enrs.1995.1064
Andersson Blomkvist G Polybrominated aromatic pollutants found in fish in Sweden Chemosphere 1981 10 1051 1060 10.1016/0045-6535(81)90216-2
Watanabe I Kashimoto T Tatsukawa R Polybrominated biphenyl ethers in marine fish, shellfish and river and marine sediments in Japan Chemosphere 1987 16 2389 2396 10.1016/0045-6535(87)90297-9
Kruger C Polybromierte Biphenyle und Polybromierte Biphenylether-Nachweis und Bestimmung in ausgewählten Lebensmittel 1988 Westfälischen Wilhelms Universität zu Munster Germany
Law RJ Alaee M Allchin CR Boon JP Lebeuf M Lepom P Stern GA Levels and trends of polybrominated diphenylethers and other brominated flame retardants in wildlife Environ Int 2003 29 757 770 12850094 10.1016/S0160-4120(03)00110-7
Meironyté D Norén K Bergman Analysis of polybrominated diphenyl ethers in Swedish human milk. A time-related trend study, 1972-1997 Journal of Toxicology and Environmental Health, Part A 1999 58 329 341 10580757 10.1080/009841099157197
Meironyté Guvenius D Norén K Polybrominared diphenyl ethers in Swedish human milk. The follow-up study: 20010/5/14. The Second International Workshop on Brominated Flame Retardants 2001 Stockholm, Sweden 303 305
Lind Y Darnerud PO Atuma S Aune M Becker W Bjerselius R Cnattingius S Glynn A Polybrominated diphenyl ethers in breast milk from Uppsala county, Sweden Environ Res 2003 93 186 194 12963403 10.1016/S0013-9351(03)00049-5
Sjödin A Patterson DGJ Bergman A review on human exposure to brominated flame retardants (BFRs) - particularly polybrominated diphenyl ethers (PBDEs) Environ Int 2003 29 829 839 12850099 10.1016/S0160-4120(03)00108-9
Ryan JJ Polybrominated diphenyl ethers (PBDEs) in human milk; occurrence worldwide The Third International Workshop on Brominated Flame Retardants Toronto, Canada 2004 17 21
Thomsen C Froshaug M Leksnes H Becher G Brominated flame retardants in breast milk from Norway Organohalogen Compounds 2003 60-65 Boston, USA 33 36
Akutsu K Kitagawa M Nakazawa H Makino T Iwazaki K Oda H Hori S Time-trend (1973-2000) of polybrominated diphenyl ethers in Japanese mother's milk Chemosphere 2003 53 645 654 12962714 10.1016/S0045-6535(03)00764-1
Pereg D Ryan JJ Ayotte P Muckle G Patry B Dewailly E Temporal and spatial changes of brominated diphenyl ethers (BDEs) and other POPs in human milk from Nunavik (Arctic) and Southern Quebec Oraganohalogen Compounds 2003 60-65 Boston, USA 127 130
Kalantzi OI Martin FL Thomas GO Alcock RE Tang HR Drury SC Carmichael PL Nicholson JK Jones KC Different levels of polybrominated diphenyl ethers (PBDEs) and chlorinated compounds in breast milk from two U.K. Regions Environ Health Perspect 2004 112 1085 1091 15238282
Schecter A Pavuk M Päpke O Ryan JJ Birnbaum L Rosen R Polybrominated diphenyl ethers (PBDEs) in U.S. mothers' milk Environ Health Perspect 2003 111 1723 1729 14594622
Norén K Meironyté D Certain organochlorine and organobromine contaminants in Swedish human milk in perspective of past 20-30 years Chemosphere 2000 40 1111 1123 10739053 10.1016/S0045-6535(99)00360-4
Schade G Heinzow B Organochlorine pesticides and polychlorinated biphenyls in human milk of mothers living in northern Germany: Current extent of contamination, time trend from 1986 to 1997 and factors that influence the levels of contamination Sci Total Environ 1998 215 31 39 9599454 10.1016/S0048-9697(98)00008-4
Fürst P Organochlorine pesticides, dioxins, PCB and polybrominated biphenylethers in human milk from Germany in the course of time Organohalogen Compounds 2001 52 Gyeongju, Korea 185 188
Johansen HR Becher G Polder A Skaare J Congener-specific determination of polychlorinated biphenyls and organochlorine pesticides in human milk from Norwegian mothers living in Oslo J Toxicol Environ Health 1994 42 157 171 8207752
Newsome WH Davies D Doucet J PCB and organochlorine pesticides in Canadian human milk - 1992 Chemosphere 1995 30 2143 2153 7620848 10.1016/0045-6535(95)00086-N
Grandjean P Bjerve KS Weihe P Steuerwald U Birthweight in a fishing community: significance of essential fatty acids and marine food contaminants International Journal of Epidemiology 2001 30 1272 1278 11821327 10.1093/ije/30.6.1272
Ballschmiter K Mennel A Buyten J Long chain alkyl-polysiloxanes as non-polar stationary phases in capillary gas chromatography Fresenius J Anal Chem 1993 346 396 402 10.1007/BF00325850
Marsh G Hu J Jakobsson E Rahm S Bergman Synthesis and characterization of 32 polybrominated diphenyl ethers Environ Sci Technol 1999 33 3033 3037 10.1021/es9902266
Bergman Klasson Wehler E Kuroki H Nilsson A Synthesis and mass spectrometry of some methoxylated PCB Chemosphere 1995 30 1921 1938 10.1016/0045-6535(95)00073-H
Letcher RJ Klasson Wehler E Bergman Paasivirta J Methyl sulfone and hydroxylated metabolites of polychlorinated biphenyls New types of persistent halogenated compounds The Handbook of Environmental Chemistry 2000 11 1 Berlin, Springer-Verlag 315 359
Hovander L Athanasiadou M Asplund L Jensen S Klasson Wehler E Extraction and cleanup methods for analysis of phenolic and neutral organohalogens in plasma J Anal Toxicol 2000 24 696 703 11110024
Lunder S Sharp R Mothers' milk. Record levels of toxic fire retardants found in American mothers' breast milk 2004
Weiss J Meijer L Sauer P Linderholm L Athanassiadis I Bergman PBDE and HBCDD levels in blood from Dutch mothers and infants - Analysis project within the Dutch Groningen Infant Cohort The Third International Workshop on Brominated Flame Retardants, BFR 2004 Toronto, Canada 2004 71 74
She J Holden A Sharp M Tanner M Williams-Derry C Hooper K Unusual Pattern of Polybrominated Diphenyl Ethers (PBDEs) in US Breast Milk Organohalogen Compounds 2004 66 Berlin, Germany 3945 3950
Johnson-Restrepo B Kannan K Rapaport D Rodan B Polybrominated diphenyl ethers and polychlorinated biphenyls in human adipose tissue from New York Environ Sci Technol 2005 39 5177 5182 16082945 10.1021/es050399x
Thuresson K Occupational exposure to brominated flame retardants - with emphasis on polybrominated diphenyl ethers 2004 Stockholm University, Environmental Chemistry Department
Thuresson K Bergman Jakobsson K Occupational exposure to commercial decabromodiphenyl ether in workers manufacturing or handling flame retarded rubber Environ Sci Technol 2005 39 1980 1986 15871227 10.1021/es048511z
Vieth B Herrmann T Mielke H Ostermann B Päpke O Rudiger T PBDE levels in human milk: The situation in Germany and potential influencing factors - A controlled study Organohalogen Compounds 2004 66 Berlin, Germany 2643 2648
Weber H Heseker H Analysis of polybrominated diphenyl ethers in breast milk of German mothers - results of a pilot study Fresenius Envir Bull 2004 13 356 360
Pirard C De Pauw E Focant JF Levels of selected PBDEs and PCBs in Belgian human milk Organohalogen Compounds Boston, USA 2003 64 158 161
Meironyté Guvenius D Aronsson A Ekman-Ordeberg G Bergman Norén K Human prenatal and postnatal exposure to polybrominated diphenyl ethers, polychlorinated biphenyls, polychlorobiphenylols and pentachlorophenol Environ Health Perspect 2003 111 1235 1241 12842779
Newsome WH Davies D Determination of PCB Metabolites in Canadian Human Milk Chemosphere 1996 33 559 565 8680832 10.1016/0045-6535(96)00199-3
Soechitram SD Athanasiadou M Hovander L Bergman Sauer PJJ Fetal exposure to PCBs and their hydroxylated metabolites in a dutch cohort Environ Health Perspect 2004 112 1208 1212 15289169
Ryan JJ van Oostdam J Polybrominated diphenyl ethers (PBDEs) in maternal and cord blood plasma of several Northern Canadian Populations Organohalogen Compounds 2004 66 Berlin, Germany 2579 2585
Needham L Ryan JJ Fürst P Guidelines for analysis of human milk from environmental chemicals Journal of Toxicology and Environmental Health, Part A 2002 65 1893 1908 12470493 10.1080/00984100290071784
Weiss J Päpke O Bignert A Jensen S Greyerz E Agostoni C Besana R Riva E Giovannini M Zetterström R Concentrations of dioxins and other organochlorines (PCBs, DDTs, HCHs) in human milk from Seveso, Milan and a Lombardian rural area in Italy: a study performed 25 years after the heavy dioxin exposure in Seveso Acta Paediatr 2003 92 467 472 12801115
|
16014177
|
PMC1226148
|
CC BY
|
2021-01-04 16:36:32
|
no
|
Environ Health. 2005 Jul 14; 4:12
|
utf-8
|
Environ Health
| 2,005 |
10.1186/1476-069X-4-12
|
oa_comm
|
==== Front
Theor Biol Med ModelTheoretical Biology & Medical Modelling1742-4682BioMed Central London 1742-4682-2-301609553610.1186/1742-4682-2-30ResearchThe fractal geometry of nutrient exchange surfaces does not provide an explanation for 3/4-power metabolic scaling Painter Page R [email protected] Office of Environmental Health Hazard Assessment, California Environmental Protection Agency, P. O. Box 4010, Sacramento, California 95812, USA2005 11 8 2005 2 30 30 30 4 2005 11 8 2005 Copyright © 2005 Painter; licensee BioMed Central Ltd.2005Painter; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
A prominent theoretical explanation for 3/4-power allometric scaling of metabolism proposes that the nutrient exchange surface of capillaries has properties of a space-filling fractal. The theory assumes that nutrient exchange surface area has a fractal dimension equal to or greater than 2 and less than or equal to 3 and that the volume filled by the exchange surface area has a fractal dimension equal to or greater than 3 and less than or equal to 4.
Results
It is shown that contradicting predictions can be derived from the assumptions of the model. When errors in the model are corrected, it is shown to predict that metabolic rate is proportional to body mass (proportional scaling).
Conclusion
The presence of space-filling fractal nutrient exchange surfaces does not provide a satisfactory explanation for 3/4-power metabolic rate scaling.
==== Body
Background
Physiological variables (e.g., cardiac output) or structural variables (e.g., pulmonary alveolar surface area) in mammals of mass M in many cases are closely approximated by an exponential function, R = R1Mb, which is termed an allometric relationship [1,2]. A prominent example is Kleiber's law for scaling the basal metabolic rate (BMR) in mammals [3,4], B = B1M3/4, which is equivalent to scaling the specific BMR, B/M, proportionally to M-1/4.
In the report, "The Fourth Dimension of Life: Fractal Geometry and Allometric Scaling of Organisms," West, Brown and Enquist (WBE) [5] derive the 3/4-power law in part from the claim that mammalian distribution networks are "fractal like" and in part from the conjecture that natural selection has tended to maximize metabolic capacity "by maximizing the scaling of exchange surface areas" for the delivery of oxygen and nutrients to body tissues.
WBE derive an expression describing scaling of surface area for nutrient exchange by considering a scale transformation that increases the linear dimensions of arteries and other internal structures (with the exception of capillaries) by the factor λ. The dimensions of individual capillaries are assumed to be invariant. WBE express scaling of the total internal exchange area as
(1)
where a is the area following the transformation and ar is the area before. The authors describe the exponent 2+εa as the "fractal dimension of a" to justify the restriction 0 ≤ εa ≤ 1 (Assumption 1). They justify the upper limit of εa by stating that εa = 1 "represents the maximum fracticality of a volume-filling structure in which the effective surface area scales like a conventional volume." This makes it clear that the structures in their model exist in 3-dimensional Euclidean space, E3.
Similarly, they express scaling of v, the internal volume associated with a (and assumed to be proportional to body mass), as
(2)
where 3+εv is termed the "fractal dimension of v" and εv satisfies 0 ≤ εv ≤ I (Assumption 2). They then write v = al, defining a new function l, which is assumed to be an internal linear dimension other than that of capillaries. The scaling of l is described by the equation
(3)
where εl is again a parameter that satisfies 0 ≤ εl ≤ 1 (Assumption 3). From the above equation for v, it follows that
(4)
The last assumptions of the theory are that natural selection has tended to maximize metabolic capacity "by maximizing the scaling of exchange surface areas" (Assumption 4) and that BMR is proportional to a (Assumption 5). Maximization of a/v requires εa = 1, and εl = 0 (Result 1). Consequently, the fractal dimension of a is 3 (Result 2), and the fractal dimension of v is 4 (Result 3). Substitution of these values into Equations (2) and (4) followed by elimination of λ leads to
a/v ∝ v-1/4 (5)
which is a form of Kleiber's law if Assumption 5 is true.
In a critical review of the WBE model, Dodds et al. [6] claim that "the bounds 0 ≤ εa, εv, εl ≤ 1 are overly restrictive." They analyze the example where 0 ≤ εa, εv ≤ 1, as in the WBE model, but where -1 ≤ εl ≤ 1. Optimization leads to the conclusion that the fractal dimension of l is 0, that the fractal dimension of both a and v is 3 and that exchange surface area a scales with volume. Consequently, a/v is constant in this example.
Agutter and Wheatley [7] also critically reviewed the WBE model, pointing out that the maximal metabolic rate (MMR) is plausibly limited by nutrient supply while the BMR is not limited by nutrient supply. Therefore, the model of WBE should predict the scaling of MMR. However, the scaling exponent for MMR appears to be different from 3/4. Weibel et al. [8] estimate this exponent to be 0.872 with a 95% confidence interval of (0.812 – 0.931).
While the issue raised by Agutter and Wheatley may not be resolvable using mathematical analysis, the issue raised by Dodds et al. is readily addressed by mathematical analysis. In the following, the theory of WBE is evaluated by first using a model of a 3-dimensional fractal-like network. Then the rigor of the arguments used in deriving the results of the theory is evaluated using properties of Hausdorff n-dimensional measure, the concept that is the basis for the general definition of fractal dimension.
Results
If the argument used by WBE to "prove" 3/4-power scaling is valid, it should require 3/4-power scaling for a specific example of a fractal distribution network. Examples of the "fractal-like" arterial networks previously described by WBE [9] are shown in Figures 1 and 2. The supply network for a square starts with an H-shaped network that is connected to the nutrient source (Figure 1a). The network is extended by iteratively connecting each terminal site to an H-shaped structure that is one-half the size (in terms of linear dimension) of the structures added in the previous step (Figure 1b). For a network that supplies a cube, we start with two parallel H-shaped structures that are connected by a conduit. This structure, termed an H-H structure, is illustrated in Figure 2a. This network is extended by iterative additions of H-H structure of one-half the dimension of the previously added H-H structure. Each added structure is connected at its midpoint. Iterative addition of smaller and smaller H-shaped structures in Figure 1 gives the fractal lung model of Mandelbrot [10], and iterative addition of H-H structures gives a 3-dimensional fractal model. An infinite sequence of additions gives an area-filling network of fractal dimension 2 for the 2-dimensional network and a space-filling network of fractal dimension 3 for the 3-dimensional network. The 2-dimensional network in Figure 1 is equivalent to the fractal-like network illustrated in Figure 4 of Turcotte et al. [11], and the 3-dimensional network in Figure 2 is equivalent to the fractal-like network in Figure 7 of Turcotte et al.
Figure 1 A 2-dimensional fractal-like, branching network model for an arterial tree. Blood enters the network through the structure represented as a thick horizontal line. Terminal arteries are represented by thin horizontal lines. a. A network that uniformly supplies a 2 × 2 area where the unit distance is the spacing between adjacent termini of small arteries. b. A network that uniformly supplies a 4 × 4 area.
Figure 2 A 3-dimensional fractal-like, branching network model for an arterial tree. Blood enters the network through the structure represented as a thick horizontal line. Terminal arteries are represented by thin horizontal lines. a. A network that uniformly supplies a 2 × 2 × 2 volume where the unit distance is the spacing between adjacent termini of small arteries. b. A network that uniformly supplies a 4 × 4 × 4 volume.
We now compare the maximum nutrient exchange surface area for the network shown in Figure 2a with that of the network shown in Figure 2b. We assume that, for both networks, each terminus is connected to capillaries that have an associated fractal surface. Their "maximum fracticality" is the dimension 3, which corresponds to a space-filling surface. The measure of the exchange surface area is the volume of the space within V that is filled by the surface. This volume is assumed by WBE to be proportional to total body volume and to body mass. Therefore, we can write A = cV, where c ≤ 1. Consequently, exchange surface area and metabolic rate scale proportionally to volume for the networks in this example. This in turn implies that λ3 is proportional to V. However, WBE conclude that V is proportional to λ4. Consequently, there must be an error in their argument.
The source of the contradiction is Equation (4), which WBE justify by claiming "v can always be expressed as v = al, where l is some length characteristic of the internal structure of the organism." In conventional geometry, this assertion is correct for certain types of figures when area is defined as the cross-sectional area. For example, the volume of a right cylinder is equal to the area of its circular cross section multiplied by the length of the cylinder. The volume is not equal to the exterior surface area of the cylinder multiplied by its length. Unfortunately, WBE assume that in fractal geometry, unlike the arithmetic of conventional geometry, volume is exterior surface area multiplied by length. The assumption that fractal volume is equal to fractal cross-sectional area multiplied by length leads to the conclusion that volume scales as λ3. This is because a cross-section of a fractal in E3 is the intersection of a plane and the fractal, i.e., it is a set of points in 2-dimensional space, just as is the case for conventional geometric objects. With this correct calculation of the (maximum) dimension of a fractal object with surface area a and length l, it follows that the metabolic rate is proportional to volume and body mass, as illustrated by the example in Figure 2.
Discussion
At the heart of the argument of WBE is the hypothesis that a 3-dimensional object in E3 with volume V can be filled with a fractal surface to produce an object with fractal dimension 4. This hypothesis is false because the volume of an object has the same, finite value whether an object contains a space-filling fractal surface or not. Because the volume is finite (and not equal to 0), the Hausdorff dimension (which is a general definition of fractal dimension) must be equal to 3 [12]. If the presence of a space-filling surface within the object changes its fractal dimension to a value greater than 3, then the Hausdorff dimension is greater than 3 and the conventional volume is infinite. This contradiction in the WBE model is resolved by rejecting all assumptions that allow for the Hausdorff dimension of a mammalian body to be greater than 3, the maximum dimension of an object in E3. These are Equations (2), (3) and (4) and Assumptions (2) and (3). Equation (3) and Assumption (3) must be rejected because l is not a set of points in space and therefore has no fractal dimension. When these assumptions are removed, the maximum nutrient exchange surface area principle of WBE leads to the prediction that metabolic rate is directly proportional to body mass.
The assumption that relates exchange surface area to metabolic rate, Assumption 5, is a common assumption used to explain diffusion-limited nutrient uptake. It seems plausible for non-fractal exchange surfaces such as the walls of capillaries, which are approximately cylindrical. For such surfaces, area is proportional to Hausdorff 2-dimensional measure. Hausdorff 3-dimensional measure of such surfaces is 0. Therefore, metabolic rate must be proportional to 2-dimensional measure of the surface if the WBE formulation is to be biologically meaningful. However, when the capillary surface is extended to a space-filling fractal that scales as volume, Hausdorff 2-dimensional measure is infinite, but Hausdorff 3-dimensional measure is proportional to the volume filled by the fractal. Therefore, the Hausdorff 3-dimensional measure must be used to scale metabolic rate for space-filling surfaces if the formulation is to be biologically meaningful. While this dichotomy in the computation of rate-determining area is not a mathematical contradiction, it does result in losing the standard justification for Assumption 5, because nutrient diffusion rate and metabolic rate cannot be proportional to exchange surface area when the fractal dimension of exchange surface area is 3. Furthermore, Assumption 5 leads to the conclusion that all space-filling exchange surfaces filling the same volume V of a mammalian body must confer exactly the same metabolic rate on the organism. This seems peculiar because the connection of function with biological form appears to have been lost as a result of the application of WBE's maximization principle, Assumption 4.
As discussed in the background section, Dodds et al. [6] claim that the bounds on fractal dimensions in the WBE model are "too restrictive" and replace Assumption 3 by extending the allowable values of the fractal dimension of l to the interval [0, 2]. However, their criticism is not valid because the bounds on fractal dimensions in the WBE model are not too restrictive. They are not restrictive enough.
Competing interests
The author(s) declare that they have no competing interests.
Acknowledgements
I thank Dr. John Hoggard and Dr. Charles Salocks for their helpful comments on drafts of this article.
==== Refs
Schmidt-Nielsen K Scaling: Why is Animal Size so Important? 1984 New York: Cambridge University Press
Calder WA III Size, Function and Life History 1984 Cambridge. MA: Harvard University Press
Kleiber M Body size and metabolism Hilgardia 1932 6 315 353
Kleiber M Body size and metabolic rate Physiol Rev 1947 27 511 541
West GB Brown JH Enquist BJ The fourth dimension of life: Fractal geometry and allometric scaling of organisms Science 1999 284 1677 1679 10356399 10.1126/science.284.5420.1677
Dodds PS Rothman DH Weitz JS Re-examination of the "3/4-law" of metabolism J Theoret Biol 2001 209 9 27 11237567 10.1006/jtbi.2000.2238
Agutter PS Wheatley DN Metabolic scaling: consensus or controversy? Theoret Biol Med Modelling 2004 1 13 10.1186/1742-4682-1-13
Weibel ER Bacigalupe LD Scnmitt B Hoppeler H Allometric scaling of maximal metabolic rate in mammals: muscle aerobic capacity as determinant factor Respir Physiol Neurobiol 2004 140 115 132 15134660 10.1016/j.resp.2004.01.006
West GB Brown JH Enquist BJ A general model for the origin of allometric scaling laws in biology Science 1997 276 122 126 9082983 10.1126/science.276.5309.122
Mandelbrot BB The Fractal Geometry of Nature 1983 New York: Freeman
Turcotte DL Pelletier JD Newman WI Networks with side branching in biology J Theoret Biol 1998 193 577 592 9745754 10.1006/jtbi.1998.0723
Falconer K Fractal Geometry: Mathematical Foundations and Applications 2005 Second New York: John Wiley & Sons
|
16095536
|
PMC1226149
|
CC BY
|
2021-01-04 16:39:25
|
no
|
Theor Biol Med Model. 2005 Aug 11; 2:30
|
utf-8
|
Theor Biol Med Model
| 2,005 |
10.1186/1742-4682-2-30
|
oa_comm
|
==== Front
World J Surg OncolWorld Journal of Surgical Oncology1477-7819BioMed Central London 1477-7819-3-511604579310.1186/1477-7819-3-51Case ReportA novel combination of multiple primary carcinomas: Urinary bladder transitional cell carcinoma, prostate adenocarcinoma and small cell lung carcinoma- report of a case and review of the literature Koutsopoulos Anastassios V [email protected] Konstantina I [email protected] George [email protected] Elpida [email protected] Marios [email protected] Efstathios [email protected] Department of Pathology, Heraklion University Hospital, Greece2 Department of Pneumonology, Heraklion University Hospital, Greece2005 26 7 2005 3 51 51 19 3 2005 26 7 2005 Copyright © 2005 Koutsopoulos et al; licensee BioMed Central Ltd.2005Koutsopoulos et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
The incidence of multiple primary malignant neoplasms increases with age and they are encountered more frequently nowadays than before, the phenomenon is still considered to be rare.
Case presentation
We report a case of a man in whom urinary bladder transitional cell carcinoma, metachronous prostate adenocarcinoma and small cell lung carcinoma were diagnosed within an eighteen-month period. The only known predisposing factor was that he was heavy smoker (90–100 packets per year). The literature on the phenomenon of multiple primary malignancies in a single patient is reviewed and the data is summarized.
Conclusion
It is important for the clinicians to keep in mind the possibility of a metachronous (successive) or a synchronous (simultaneous) malignancy in a cancer patient. It is worthy mentioning this case because clustering of three primary malignancies (synchronous and metachronous) is of rare occurrence in a single patient, and, to our knowledge, this is the first report this combination of three carcinomas appearing in the same patient.
==== Body
Background
The phenomenon of multiple primary malignant neoplasms in the same individual was described firstly by Billroth at the end of the 19th century [1]. Since then, several cases of double or even triple primary malignant neoplasms have been reported. It is believed that multiple primary malignant neoplasms now occur more frequently than before. Although, not uncommon, they occur more often in elderly patients, as the incidence of malignancies increases with age. The diagnosis of second primary neoplasms is rising as a result of prolonged survival of patients treated for previous malignancy with alkylating agents, topoisomerase II inhibitors, and/or radiotherapy[2]. A review of the recent literature indicates clearly that they appear more frequently in the upper digestive tract, respiratory system, head and neck region, or urogenital system; the reported incidence ranges from 2% to 10% [3].
In this report we present a patient who developed primary bladder carcinoma and metachronous prostate and small cell lung carcinoma (SCLC) within an eighteen-month period. This combination of multiple primary carcinomas, to our knowledge, has never been reported in the literature.
Case presentation
A 75-year old ex-smoker (90–100 packet per year) underwent a transurethral resection of urinary bladder papilloma in February 2002. The histology of resected specimen was papillary transitional cell carcinoma grade II (Figure 1A). The tumor cells were positive for cytokeratin 7 (Figure 1B) and negative for cytokeratin 20. There were no muscle fibers in the examined tissue. The ultrasound examination of the urogenital system revealed nodular hyperplasia of the prostate. The tumor clinical stage according to the American Cancer Committee U.I.C.C. (1992) was Ta. Patient's cancer relapsed at the end of the same year and he underwent a programmed transurethral resection of the tumor, which proved to be papillary transitional cell carcinoma grade I-II. No lamina propria or muscle invasion was detected. The patient was also treated with intracystic infusion of bacille Calmette-Guerin (BCG). Ten days later, because of urine retention, he underwent transurethral resection of the prostate. Multiple tissue fragments of total dimensions 4.5 × 3.5 × 2.2 cm were examined histologically. Seven out of the 10 examined slides revealed foci of partially mucinous (Figure 2A) adenocarcinoma of the prostate (the greatest measured focus was 8.5 mm in maximum diameter), Gleason grade II-III and Gleason score 5 (Figure 2A, B). Immunohistochemical study was performed and showed strong positivity for Prostate Specific antigen (PSA) (Figure 2C) whereas; no expression of carcinoembryonic antigen (CEA) was detected in tumor cells. These findings confirmed the diagnosis of primary prostate adenocarcinoma. The tumor's stage according to the 1997 TNM staging system of prostatic adenocarcinoma was T1b. Serum prostate specific antigen (PSA) levels were elevated (9 ng/mL) before surgery. No additional surgical treatment was given and at follow-up visits prostate specific antigen (PSA) levels measurement and intracystic injection of BCG was performed. In September of the same year, due to progressively worsening dyspnea a computed tomography was performed that revealed a mediastinal mass in conjunction to the right lung hilum and to the right main bronchus with maximum diameter of 9 cm. Bronchoscopy showed a large mass which invaded the right main bronchus mucosa and extended to the carina. Histology of the bronchial mucosal sample showed infiltration of lamina propria by malignant cells (Figure 3A). Their immunophenotype was: CD56 (+) (Figure 3B), Pan-Cytokeratin (paranuclear dot stain positivity) (Figure 3C) and Leukocyte Common Antigen negative. Combining the morphological and the immunohistochemical results, we concluded that the patient was suffering from small cell lung carcinoma (SCLC). The patient's stage was IIIB. Ten days after the diagnosis was confirmed, the patient underwent the first cycle of chemotherapy (Cisplatin and Vepesid), during which he died from cardiac arrest due to chemotherapy toxicity.
Figure 1 Microscopically, the extracted urinary bladder tissue particles proved to be pieces of papillary transitional cell carcinoma grade II [A) hematoxylin and eosin × 40] and immunohistochemically they expressed cytokeratin 7 [B) cytokeratin 7 × 100].
Figure 2 Histologically, in most of the prostate tissue fragments were recognized areas of, partially mucinous, adenocarcinoma of the prostate, grade II-III (A. hematoxylin and eosin × 40, B: hematoxylin and eosin × 100). The tumor cells were strongly positive for PSA (C: PSA × 40).
Figure 3 The bronchial mucosa showed extensive invasion from small blue round cells (A: hematoxylin and eosin × 100) that were positive for the neuroendocrine marker CD56 (B: × 200) and pan-cytokeratin (C: × 200).
Discussion
We report a patient who developed three histologically distinct malignancies, i.e. primary bladder carcinoma and metachronous prostate and SCLC within an eighteen-month period. There are several predisposing or causal factors for each malignancy. For our patient there was only one common causal factor, the fact that he was a heavy smoker (90–100 packets per year). No other predisposing factor or a family history was found that might have contributed to the development of these three malignancies. The presence of bladder and prostate carcinomas in the same patient is not a rare event. Chun [3] reported that the rate of bladder carcinoma in patients with prostate carcinoma is eighteen times higher (p < 0,01) and the rate of prostate carcinoma in those with bladder carcinoma is nineteen times higher (p < 0,01) than expected. Although bladder and prostate carcinoma can coexist in the same individual frequently enough, the rare event is the appearance of a third malignancy. There is a case report by Rovinescu et al [4] referring to a patient with three primary malignancies. The first tumor was a clear cell carcinoma of the kidney, which was followed by a transitional cell carcinoma of the bladder and then by a distinct adenocarcinoma of the prostate. More recently, in 2003, Satoh et al [5] also reported the same combination of multiple primary malignancies in a patient. Our case is the first one of an individual having these two primary malignancies of the urogenital system and another tumor of the lower respiratory tract.
Table 1 summarizes the cases with three or more primary malignancies. As can be easily seen, although the appearance of three primary malignancies in one patient is not very common, should not be considered such a rare event.
Table 1 There are summarized the cases of triple or more malignancies, the first author, journal, year of publication and combination of neoplasms.
Year Author 1st Malignancy 2nd Malignancy 3rd Malignancy 4th Malignancy 5th Malignancy
1 1949 Crail H.W [6] Thyroid Carcinoma Rectal Carcinoma Medulloblastoma
2 1974 Hamoudi A.B.[7] Colon Carcinoma Thymus Carcinoma Skin Carcinoma Astrocytoma G3
3 1975 Ohsato K. [8] Colon Carcinoma Astrocytoma G3 Duodenal Carcinoma
4 1976 Kawanami K. [9] Ileum Carcinoma Glioblastoma Colon Carcinoma
5 1976 Rovinescu I.[4] Clear Cell Carcinoma Of Kidney Transitional Cell Carcinoma Of Bladder Prostate Carcinoma
6 1979 Itoh H.[10] Colon Carcinoma Stomach Carcinoma Astrocytoma G3
7 1979 Mullen J.L.[11] Hodgkin' Disease Squamous Cell Carcinoma Of Larynx Squamous Cell Carcinoma In Esophagus
8 1979 Pinel J.[12] 7 Squamous Cell Carcinomas In 9 Years
9 1980 Cohen C.[13] Multiple Cutaneous Squamous Cell Carcinomas Multiple Cutaneous Basal Cell Carcinomas Diffuse Poorly Differentiated Lymphocytic Lymphoma
10 1982 Friedman C.D. [14]. Breast Carcinoma Colon Carcinoma Glioblastoma In Brain
11 1983 Li F.P.[15] Colon Carcinoma Astrocytoma G3 Leukemia
12 1984 Haibach H.[16] Thyroid Carcinoma Renal Carcinoma Duodenal Carcinoma
13 1985 Alessi E.[17] Multiple Sebaceous Tumors Keratoacanthoma 3 Primary Adenocarcinomas Of Colon
14 1985 Kobayashi T. [18] Uterus Carcinoma Stomach Carcinoma Breast Carcinoma Glioblastoma In Brain
15 1985 Megighian D.[19] Squamous Cell Carcinoma Of Parotid Squamous Cell Carcinoma Of Tongue Squamous Cell Carcinoma Of Soft Palate Squamous Cell Carcinoma Of Larynx Squamous Cell Carcinoma Of Hypopharynx
16 1985 Staren E.D.[20] Carcinoma Of Larynx Carcinoma Of Floor Of Mouth Dual Primary Bronchogenic Carcinomas
17 1986 Craig D.M.[21] Squamous Cell Carcinoma Of The Floor Of The Mouth Adenocarcinoma Of Lung Squamous Cell Carcinoma Of Larynx Squamous Cell Carcinoma Of The Tongue
18 1986 Ogasawara K.[22] Breast Carcinoma Breast Carcinoma Lung Carcinoma Glioblastoma In Brain Thyroid Carcinoma
19 1987 Hayashi K.[23] Colon Carcinoma Rectal Carcinoma Glioblastoma In Brain
20 1987 Kobayashi T.[24] Uterus Carcinoma Stomach Carcinoma Glioblastoma In Brain
21 1988 Ohi H. [25] Skin Carcinoma Medulloblastoma Thyroid Carcinoma
22 1991 Baigrie R.J.[26] 7 Primary Carcinomas
23 1991 Solan M.J.[27] Two Breast Carcinomas Thyroid Carcinoma Multiple Skin Carcinomas
24 1992 Melkert P.W. [28] Squamous Cell Carcinoma Of Skin Squamous Cell Carcinoma Of Vulva Squamous Cell Carcinoma Of Vagina Squamous Cell Carcinoma Of Anus Squamous Cell Carcinoma Of Cervix Uteri
25 1992 Marcos Sanchez F. [29] Colon Carcinoma Renal Carcinoma Breast Carcinoma
26 1993 Brugieres L. [30] Soft Tissue Tumor Brain Tumor Thyroid Carcinoma Breast Carcinoma
27 1993 Kikuchi T. [31] Glioblastoma Colon Carcinoma Colon Carcinoma
28 1993 Shiseki M.[32] Skin Carcinoma Colon Carcinoma Glioblastoma In Brain
29 1994 Bumpers H.L.[33] Squamous Cell Carcinoma Of Larynx Squamous Carcinoma Of Lung Adenocarcinoma Of Breast Adenocarcinoma Of Colon
30 1994 Nishihara K. [34] Papillary Adenocarcinoma Of Papilla Of Vater Papillary Adenocarcinoma Of Common Bile Duct Papillary Adenocarcinoma Of Pancreas
31 1995 Angeli-Besson C. [35] Chronic Myeloid Leukemia, Multiple Squamous Cell Carcinomas
32 1996 Hayashi T.[36] Squamous Cell Carcinoma In Soft Palate Squamous Cell Carcinoma In Larynx Squamous Cell Carcinoma In Esophagus
33 1996 Nagane M.[37] Tubular Adenocarcinoma Of Stomach Transitional Cell Carcinoma Of Bladder Glioblastoma In Brain
34 1996 Nagane M. [37] Papillary Adenocarcinoma Of Lung Adenocarcinoma Of Rectum Glioblastoma In Brain
35 1997 Potzsch C.[38] Breast Carcinoma Small Cell Lung Carcinoma Renal Cell Carcinoma Acute Myelomonocytic Leukemia
36 1997 Shan L.[39] 14 Foci Of Primary SCC, Esophagus, Oral Floor, Soft Palate, Uvula, Lingual Radix, Piriform Recess, Hypopharynx, Trachea, Lingual Body
37 1999 Cribier B. [40] Eccrine Porocarcinoma Tricholemmal Carcinoma Multiple Squamous Cell Carcinomas
38 1999 Ramsay H.M.[41] Acute Myeloid Leukemia Chronic Lymphocytic Leukemia Basal Cell Carcinomas
39 1999 Schon M.P.[42] Basal Cell Carcinomas Hairy Cell Leukemia Basal Cell Carcinomas
40 2000 Beswick S.J.[43] Basal Cell Carcinomas Malignant Melanoma In Situ Basal Cell Carcinomas
41 2001 Mukai [44] Stomach Carcinoma Duodenal Carcinoma Esophageal Cancer Renal Cancer Colon Carcinoma In Situ
42 2003 Satoh H.[5] Carcinoma Of Kidney Transitional Cell Carcinoma Of Bladder Prostate Carcinoma
Additionally, studying the existing bibliography, we noticed that there is a little confusion regarding the terms used, such as synchronous, simultaneous and metachronous or successive neoplasms. All of these words have to do with the time that the neoplasms are discovered and have nothing to do with the time of their genesis. The word synchronous is a Greek one that should refer to neoplasms appearing in the same time. It is synonymous to the word simultaneous and they are interchangeable. Metachronous (meta- means after and -chronous is the time) is also a Greek word referring to a neoplasm that is discovered while there is already a known neoplasm in the same patient. The word successive could be used equally to metachronous.
Conclusion
Summarizing, it is important for the clinicians to keep in mind that the appearance of another tumor in a patient suffering from cancer could be either a metastasis or another malignancy and should always investigate the possibility of a metachronous (successive) or a synchronous (simultaneous) malignancy. Moreover, the combination of the three different neoplasms (bladder, prostate and SCLC) in one patient, to the best of our knowledge, has never been reported before.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
KAV wrote the original manuscript and performed histopathological evaluation of the lung lesion.
DKI participated in the writing of the original manuscript and prepared photomicrographs.
DG performed histopathological evaluation of the urinary bladder lesion.
GE performed histopathological evaluation of the prostate lesion.
FM performed bronchoscopy and patient's management.
SE prepared requested revisions of the manuscript.
Acknowledgements
The permission was obtained from the next of kin of patient for publication of this case report.
==== Refs
Billroth T [General surgical pathology and therapy. Guidance for students and physicians. Lecture] Khirurgiia (Mosk) 1991 136 143 1803082
Munker R Hiller E Melnyk A Gutjahr P Second malignancies: clinical relevance and basic research. Int J Oncol 1996 763 776
Chun TY Coincidence of bladder and prostate cancer J Urol 1997 157 65 67 8976217 10.1097/00005392-199701000-00018
Rovinescu I Rousseau E [Considerations relating to multiple primary carcinomas of the urinary tract (author's transl)] J Urol Nephrol (Paris) 1976 82 621 626 1034027
Satoh H Momma T Saito S Hirose S [A case of synchronous triple primary carcinomas of the kidney, bladder and prostate] Hinyokika Kiyo 2003 49 261 264 12822453
Crail HW Multiple primary malignancies arising in the rectum, brain, and thyroid. US Naval Med Bull 1949 49 123 128
Hamoudi AB Ertel I Newton WAJ Reiner CB Clatworthy HWJ Multiple neoplasms in an adolescent child associated with IGA deficiency Cancer 1974 33 1134 1144 4406171
Ohsato K Hashimoto H Itoh H et Familial adenomatosis of the colon associated with brain tumor (Turcot's syndrome): report of a case and review of the literature (Japanese) STOMACH INTEST (Tokyo) 1975 10 1511 1517
Kawanami K Ohno M Matsuura K et Turcot's syndrome: report of an autopsy case STOMACH INTEST (Tokyo) 1976 11 1075 1082
Itoh H Ohsato K Yao T Iida M Watanabe H Turcot's syndrome and its mode of inheritance Gut 1979 20 414 419 223949
Mullen JL Reichman J Rosato EF Multiple carcinomas following therapy for Hodgkin disease J Surg Oncol 1979 11 75 78 431085
Pinel J de Romanet JB Fontanel JP [Multiple carcinomas of the upper digestive tract. 7 tumour sites in 9 years (author's transl)] Ann Otolaryngol Chir Cervicofac 1979 96 619 621 525951
Cohen C Multiple cutaneous carcinomas and lymphomas of the skin Arch Dermatol 1980 116 687 689 6892866 10.1001/archderm.116.6.687
Friedman CD McCarthy JR Multiple primary malignant neoplasms IMJ Illinois Medical Journal 1982 161 115 116 6122671
Li FP Little JB Bech-Hansen NT Paterson MC Arlett C Garnick MB Mayer RJ Acute leukemia after radiotherapy in a patient with Turcot's syndrome. Impaired colony formation in skin fibroblast cultures after irradiation Am J Med 1983 74 343 348 6572034 10.1016/0002-9343(83)90643-5
Haibach H Rosenholtz MJ Synchronous thyroid, renal, and duodenal carcinomas Arch Pathol Lab Med 1984 108 272 274 6546663
Alessi E Brambilla L Luporini G Mosca L Bevilacqua G Multiple sebaceous tumors and carcinomas of the colon. Torre syndrome Cancer 1985 55 2566 2574 3995473
Kobayashi T Diagnosis and treatment of multiple primary neoplasm—brain tumors and cancer of other organs Saishin Igaku 1985 40 1613 1620
Megighian D Zorat PL A rarely observed case of primary multiple carcinomas of the head and neck Arch Otorhinolaryngol 1985 242 7 12 4038153 10.1007/BF00464399
Staren ED Roberts J Multiple primary cancers of the respiratory tract: a case of synchronous carcinoma of the larynx, carcinoma of the floor of the mouth, and dual primary bronchogenic carcinomas J Surg Oncol 1985 29 261 263 4079408
Craig DM Triedman LJ Four primary malignant neoplasms in a single patient J Surg Oncol 1986 32 8 10 3724193
Ogasawara K Ogawa A Shingai J Kayama T Wada T Namiki T Suzuki J [Synchronous multiple primary malignant tumors accompanied by glioblastoma. Case report] Neurol Med Chir (Tokyo) 1986 26 908 912 2437485
Hayashi K Ohtsuki Y Sonobe H Takahashi K Wada S Yoshida K [An autopsy case of triple primary cancers consisting of glioblastoma multiforme of the pons, colon cancer and rectal carcinoid--a statistical analysis of cases of brain tumor combined with other primary cancers in Japan autopsy annuals] Gan No Rinsho 1987 33 1846 1853 2826842
Kobayashi T Takahashi T Tanaka T Tateishi T Miura S [Multiple primary neoplasm--glioblastoma combined with cancer of other organs] No Shinkei Geka 1987 15 1011 1017 2827047
Ohi H Kikuchi K Futawatari K Kowada M A histologically-verified triple cancer--report of a rare case involving a primary brain tumor Gan No Rinsho Japan Journal Of Cancer Clinics 1988 34 1001 1005 3043037
Baigrie RJ Seven different primary cancers in a single patient. A case report and review of multiple primary malignant neoplasia Eur J Surg Oncol 1991 17 81 83 1995363
Solan MJ Multiple primary carcinomas as sequelae of treatment of pulmonary tuberculosis with repeated induced pneumothoraces. Case report and review of the literature Am J Clin Oncol 1991 14 49 51 1987738
Melkert PW Walboomers JM Jiwa NM Cuesta MA Kenemans P Meijer CJ Multiple HPV 16-related squamous cell carcinomas of the vulva, vagina, anus, skin and cervix in a 31-year-old woman Eur J Obstet Gynecol Reprod Biol 1992 46 53 56 1330760 10.1016/0028-2243(92)90280-C
Marcos Sanchez F Salvador Fernandez M Juarez Ucelay F Druet Ampuero J [Carcinomas of the colon, kidney, and breast in the same patient. A new case of multiple primary malignant neoplasms] An Med Interna 1992 9 257 257 1324018
Brugieres L Gardes M Moutou C Chompret A Meresse V Martin A Poisson N Flamant F Bonaiti-Pellie C Lemerle J Feunteun J Screening for germ line p53 mutations in children with malignant tumors and a family history of cancer Cancer Research 1993 53 452 455 8425176
Kikuchi T Rempel SA Rutz HP de TN Mulligan L Cavenee WK Jothy S Leduy L Van Meir EG Turcot's syndrome of glioma and polyposis occurs in the absence of germ line mutations of exons 5 to 9 of the p53 gene Cancer Res 1993 53 957 961 8439970
Shiseki M Nishikawa R Yamamoto H Ochiai A Sugimura H Shitara N Sameshima Y Mizoguchi H Sugimura T Yokota J Germ-line p53 mutation is uncommon in patients with triple primary cancers Cancer Lett 1993 73 51 57 8402598 10.1016/0304-3835(93)90187-E
Bumpers HL Natesha RK Barnwell SP Hoover EL Multiple and distinct primary cancers: a case report J Natl Med Assoc 1994 86 387 388 8046769
Nishihara K Tsuneyoshi M Shimura H Yasunami Y Three synchronous carcinomas of the papilla of Vater, common bile duct and pancreas Pathol Int 1994 44 325 332 8044300
Angeli-Besson C Koeppel MC Jacquet P Andrac L Sayag J Multiple squamous-cell carcinomas of the scalp and chronic myeloid leukemia Dermatology 1995 191 321 322 8573931
Hayashi T Sagawa H Kobuke K Fujii K Yokozaki H Tahara E Molecular-pathological analysis of a patient with three synchronous squamous cell carcinomas in the aerodigestive tract Jpn J Clin Oncol 1996 26 368 373 8895679
Nagane M Shibui S Nishikawa R Oyama H Nakanishi Y Nomura K Triple primary malignant neoplasms including a malignant brain tumor: report of two cases and review of the literature Surg Neurol 1996 45 219 229 8638217 10.1016/0090-3019(95)00305-3
Potzsch C Fetscher S Mertelsmann R Lubbert M Acute myelomonocytic leukemia secondary to synchronous carcinomas of the breast and lung, and to metachronous renal cell carcinoma J Cancer Res Clin Oncol 1997 123 678 680 9620229 10.1007/s004320050124
Shan L Nakamura Y Nakamura M Zhang Z Jing X Hara T Yokoi T Kakudo K Synchronous and metachronous multicentric squamous cell carcinomas in the upper aerodigestive tract Pathol Int 1997 47 68 72 9051695
Cribier B Lipsker D Grosshans E Eccrine porocarcinoma, tricholemmal carcinoma and multiple squamous cell carcinomas in a single patient Eur J Dermatol 1999 9 483 486 10491509
Ramsay HM Fryer A Strange RC Smith AG Multiple basal cell carcinomas in a patient with acute myeloid leukaemia and chronic lymphocytic leukaemia Clin Exp Dermatol 1999 24 281 282 10457131 10.1046/j.1365-2230.1999.00480.x
Schon MP Reifenberger J Von SS Megahed M Lang K Gattermann N Meckenstock G Goerz G Ruzicka T Multiple basal cell carcinomas associated with hairy cell leukaemia Br J Dermatol 1999 140 150 153 10215787 10.1046/j.1365-2133.1999.02626.x
Beswick SJ Garrido MC Fryer AA Strange RC Smith AG Multiple basal cell carcinomas and malignant melanoma following radiotherapy for ankylosing spondylitis Clin Exp Dermatol 2000 25 381 383 11012589 10.1046/j.1365-2230.2000.00668.x
Mukai M Macuuchi H Mukohyama S Oida Y Himeno S Nishi T Nakazaki H Satoh S Quintiple carcinomas with metachronous triple cancer of the esophagus, kidney, and colonic conduit following synchronous double cancer of the stomach and duodenum Oncol Rep 2001 8 111 114 11115580
|
16045793
|
PMC1226150
|
CC BY
|
2021-01-04 16:39:04
|
no
|
World J Surg Oncol. 2005 Jul 26; 3:51
|
utf-8
|
World J Surg Oncol
| 2,005 |
10.1186/1477-7819-3-51
|
oa_comm
|
==== Front
J Transl MedJournal of Translational Medicine1479-5876BioMed Central London 1479-5876-3-271598752210.1186/1479-5876-3-27ResearchThe salivary microbiota as a diagnostic indicator of oral cancer: A descriptive, non-randomized study of cancer-free and oral squamous cell carcinoma subjects Mager DL [email protected] AD [email protected] PM [email protected] CM [email protected] MR [email protected] JM [email protected] The Forsyth Institute, 140 The Fenway, Boston, MA, USA2 Brigham and Women's Hospital, 27 Francis Street, Boston, MA, USA3 Dana Farber Cancer Institute, 44 Binney Street, Boston, MA, USA2005 7 7 2005 3 27 27 22 2 2005 7 7 2005 Copyright © 2005 Mager et al; licensee BioMed Central Ltd.2005Mager et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
The purpose of the present investigation was to determine if the salivary counts of 40 common oral bacteria in subjects with an oral squamous cell carcinoma (OSCC) lesion would differ from those found in cancer-free (OSCC-free) controls.
Methods
Unstimulated saliva samples were collected from 229 OSCC-free and 45 OSCC subjects and evaluated for their content of 40 common oral bacteria using checkerboard DNA-DNA hybridization. DNA counts per ml saliva were determined for each species, averaged across subjects in the 2 subject groups, and significance of differences between groups determined using the Mann-Whitney test and adjusted for multiple comparisons. Diagnostic sensitivity and specificity in detection of OSCC by levels of salivary organisms were computed and comparisons made separately between a non-matched group of 45 OSCC subjects and 229 controls and a group of 45 OSCC subjects and 45 controls matched by age, gender and smoking history.
Results
Counts of 3 of the 40 species tested, Capnocytophaga gingivalis, Prevotella melaninogenica and Streptococcus mitis, were elevated in the saliva of individuals with OSCC (p < 0.001). When tested as diagnostic markers the 3 species were found to predict 80% of cancer cases (sensitivity) while excluding 83% of controls (specificity) in the non-matched group. Diagnostic sensitivity and specificity in the matched group were 80% and 82% respectively.
Conclusion
High salivary counts of C. gingivalis, P. melaninogenica and S. mitis may be diagnostic indicators of OSCC.
Oral Squamous Cell CarcinomaOral mucosabacterial markersbacteriaearly detection
==== Body
Background
Each year nearly 30,000 Americans are diagnosed with oral cancer. 90% of these lesions are oral squamous cell carcinomas [1]. Despite advances in surgery, radiation and chemotherapy, the five-year survival rate is 54%, one of the lowest of the major cancer sites, and this rate has not improved significantly in recent decades [2-4]. Worldwide, the problem is much greater, with over 350,000 to 400,000 new cases being found each year [5]. The disease kills one person every hour – more people than cancers of the cervix, brain, ovary, testes, liver, kidney, malignant melanoma or Hodgkin's lymphoma [5,6]. In the United States, African American males suffer the highest incidence and lowest survival rates of any group. From 1985 to 1996, the five-year survival rate for tongue carcinoma in African-American men was 27%, compared with a 47% five-year survival rate among white men [7]. In 2001, similar five-year survival rates were found in a study of oral and pharyngeal cancer among African-American and White men [8]. Notably, incidence in young adults (<40 years) is increasing in the U.S. [9,10] and worldwide [11,12].
Early detection followed by appropriate treatment, can increase cure rates to 80 or 90%, and greatly improve the quality of life by minimizing extensive, debilitating treatments [5,13]. Despite the accessibility of the oral cavity to direct examination, these malignancies are often not detected until a late stage [5,14,15]. Oral cancer is unusual in that it carries a high risk of second primary tumors. Patients who survive a first cancer of the oral cavity have up to a 20-fold increased risk of developing a second primary oral cancer and that risk lasts 5–10 years and sometimes longer [16].
Major risk factors for oral cancers in the United States are use of tobacco and alcohol, which account for 75 to 80% of all oral cancers [5,17]. Although tobacco is a well-recognized risk factor for OSCC, the public is generally unaware that alcohol synergizes with tobacco. Those who both smoke and drink have 15 times the risk of developing oral cancer [5]. Notably, some oral cancer patients have no known risk factors, and the disease in this population may pursue a particularly aggressive course [18].
The American Cancer Society recommends that doctors and dentists examine the mouth and throat during routine examinations [2] as early cancer lesions are often asymptomatic and may mimic benign lesions [19,20]. General population screening, however, has not been shown to reduce the incidence of and mortality from oral cancer. The reasons include the low prevalence and incidence of OSCC, the potential for false-positive diagnoses and poor compliance with screening and referral [6,21]. Thus the National Institute of Dental and Craniofacial Research and The Oral Cancer Foundation have recommended that research efforts focus on developing novel detection techniques [5,16].
Studies have reported that certain common oral bacteria are elevated on or in oral and esophageal cancer lesions and their associated lymph nodes [22-28]. Although increased colonization of facultative oral streptococci have been reported most often [24-27], anaerobic Prevotella, Veillonella, Porphyromonas and Capnocytophaga species were also elevated [25,26,28]. Currently, studies are examining whether bacteria may be incidentally or causally associated with oral cancer. Additional research is determining whether various salivary markers may be used as early diagnostic indicators for oral cancer.
The reason for these shifts in bacterial colonization of cancer lesions is unclear. Mechanistic studies of bacterial attachment provide some insights, however. Research has repeatedly shown that oral bacteria demonstrate specific tropisms toward different biological surfaces in the oral cavity such as the teeth, mucosa, and other bacteria [29-35]. The non-shedding surfaces of the teeth offer a far different habitat than the continually shedding surfaces of the oral mucosa. Due to the repeated shedding of epithelial cells, there is less time for a complex biofilm to develop on soft tissue surfaces; thus, a premium is placed on potent mechanisms of adhesion. The differences in bacterial tropisms for specific oral sites suggest that different intra-oral surfaces and bacterial species have different receptors and adhesion molecules that dictate the colonization of different oral surfaces.
It is now recognized that bacteria bind to and colonize mucosal surfaces in a highly selective manner via a "lock- and key" mechanism. Adhesins on bacteria bind specifically to complementary receptors on the mucosal surfaces of the host. These adhesins differ from species to species leading to specificity in attachment to different surfaces. Studies have shown that even within genera, colonization patterns of individual species may differ markedly [29-32]. Streptococcus salivarius, for example, preferentially colonized the oral soft tissues and saliva compared to the teeth, while the reverse was true of Streptococcus sanguis.
Cancer has been referred to as a molecular disease of cell membrane glycoconjugates, [36-38]. Certain glycoconjugates serve as receptors for specific bacteria and recent reports support the notion that shifts in the colonization of different cancer cells are associated with observed changes in cell surface receptors [36,40,41]. An in vitro study of S. sanguis, a common oral inhabitant, demonstrated that its binding capacity to normal exfoliated human buccal epithelial cells (HBEC) depended upon the availability of surface sialic acid residues [36]. Desialylation of HBEC invariably abolished adhesion of S. sanguis to these epithelial cells. In similar experiments carried out with a buccal carcinoma cell line, S. sanguis did not reliably attach. It was determined that the tumor cells did not express the sialylated membrane glycoprotein of normal cells suggesting that changes in the surface receptors had occurred in the buccal carcinoma cell line.
In a previous study of 225 OSCC-free subjects we found a high degree of specificity in the "preferred" intra-oral localization of species, even within a single genus such as Streptococcus [42]. This specificity in localization of individual species agreed with that described in previous studies. Our investigation extended earlier findings by describing the distribution of multiple species within the same genus on a wider range of intra-oral surfaces. For example, S. oralis, S. constellatus, S. mitis, S. intermedius and S. anginosus colonized the soft tissues in higher proportions than the teeth; however, their "preferred" soft tissue habitats differed. S. sanguis colonized different soft tissue locations in similar proportions, but was found in higher mean proportions on the teeth, particularly in the supragingival plaque.
The availability of a large amount of data from the OSCC-free subjects permitted this group to be subset according to periodontal and smoking status and the colonization patterns on the soft tissues compared among groups [43]. The clinical parameters among the populations were in accord with those found in previous studies and results were similar to previous investigations [44-47]. Few differences were found in the salivary or soft tissue microbiota among the subset populations. It was concluded that the presence or absence of periodontal infections or a smoking habit had minimal effects on salivary and soft tissue colonization. These findings were in accord with studies by Danser et al. 1996 and Lie et al. 1998 [48,49] but contrasted with earlier reports by Colman et al. 1976 and van Winkelhoff et al. 1986 [50,51]. Importantly, we found that when the microbiota of teeth, soft tissues and saliva were compared, the microbial profile of saliva was similar to that of the soft tissues, but saliva and soft tissue colonization differed markedly from that of dental plaque. These findings were similar to those of other investigations [46,51,52].
As previously mentioned, studies have reported that the microbiota of OSCC lesions differs from that found on the soft tissues of OSCC-free individuals. Little was known, however, about the salivary microbiota of oral cancer subjects. Thus, the purpose of the present investigation was to determine whether the salivary microbiota in subjects with an oral squamous cell carcinoma (OSCC) lesion would differ from that found in OSCC-free controls.
Materials and methods
OSCC-free Population
A total of 229 OSCC-free subjects were recruited from the patient pool at The Forsyth Institute. All subjects were 18 years or older, and immunocompetent. Exclusion criteria included: antibiotic therapy within the previous 3 months, pregnancy or lactation, systemic conditions associated with immune dysfunction (e.g., diabetes), previous chemotherapy or radiation and the presence of any oral mucosal lesions.
Oral Cancer Population
A total of 45 subjects diagnosed with OSCC via biopsy were recruited from the Partners' Hospitals (The Dana Farber Cancer Institute, Brigham and Women's Hospital and Massachusetts General Hospital). Inclusion criteria required that subjects be 18 years or older and immunocompetent, with a primary untreated OSCC. Exclusion criteria included systemic conditions associated with immune dysfunction (e.g., diabetes), previous chemotherapy or radiation, an inability to properly consent, and/or lesions that could not be sampled due to discomfort, anatomic location or that did not affect the surface oral epithelium. In the unmatched comparisons OSCC subjects were older and included a higher percentage of male subjects and smokers than OSCC-free subjects (Table 1). Thus a subset of 45 controls was matched by computer for age, gender and smoking with the 45 OSCC subjects (Table 2).
Table 1 Age, gender and smoking status of 229 OSCC-free and 45 OSCC subjects.
OSCC-free OSCC
N 229 45
Mean Age (±SEM) 42.06 (±1.04) 57.6 (±2.34)
Minimum Age 18 18
Maximum Age 81 92
% Males (N) 47% (107) 71% (32)
% Smokers (N) 20% (46) 40% (18)
Table 2 Age, gender and smoking status of 45 OSCC-free and 45 OSCC subjects matched by age, gender and smoking history
OSCC-free OSCC
N 45 45
Age Mean (±SEM) 53.67 (±2.06) 54.46 (±2.27)
Age Min 19 18
Age Max 81 85
Males 32 32
Smokers 18 18
Collection of samples and preparation of test membranes
Whole unstimulated saliva samples were collected by expectoration from 229 OSCC-free and 45 OSCC subjects. Samples were evaluated for their content of 40 common oral bacteria using checkerboard DNA-DNA hybridization as described by Socransky et al. [53] and whole genomic probes were prepared by the method described by Smith et al. (1989) [54]. The concentration of the purified DNA was determined by spectrophotometric measurement of the absorbance at 260 nm and purity of the preparations was assessed by the ratio of the absorbances at 260 and 280 nm. Whole genomic DNA probes were prepared from each of the 40 test strains by labeling 1–3 μg DNA with digoxigenin (Boehringer Mannheim, Indianapolis IN.) using a random primer technique [55]. The membranes were prehybridized to block nonspecific binding. The 40 species examined are commonly found in the oral cavity and are listed in Table 3 with their corresponding American Type Culture Collection (ATCC) numbers. Two lanes in each run had standards at 105 and 106 cells of each species and signals were converted to absolute counts by comparison with standards on the membrane. Signals were detected using a Storm Fluorimager (Molecular Dynamics, Sunnyvale CA). The sensitivity of this assay also detected 104 cells of a given species by adjusting the concentration of each DNA probe. Failure to detect a signal was recorded as zero, although counts in the 1 to 1000 range could have been present. Data available for all subjects were compared using the Mann-Whitney test and Bonferroni adjustment performed for multiple comparisons.
Table 3 The 40 test strains employed for the development of DNA probes.
Microorganism ATCC number
Actinobacillus actinomycetemcomitans 43718 and 29523
Actinomyces gerencseriae 23860
Actinomyces israelii 12102
Actinomyces naeslundii genospecies 1 12104
Actinomyces naeslundii genospecies 2 43146
Actinomyces odontolyticus 17929
Campylobacter gracilis 33236
Campylobacter rectus 33238
Campylobacter showae 51146
Capnocytophaga gingivalis 33624
Capnocytophaga ochracea 33596
Capnocytophaga sputigena 33612
Eikenella corrodens 23834
Eubacterium nodatum 33099
Eubacterium saburreum 33271
Fusobacterium nucleatum ss nucleatum 25586
Fusobacterium nucleatum ss polymorphum 10953
Fusobacterium nucleatum ss vincentii 49256
Fusobacterium periodonticum 33693
Gemella morbillorum 27824
Leptotrichia buccalis 14201
Neisseria mucosa 19696
Peptostreptococcus micros 33270
Porphyromonas gingivalis 33277
Prevotella intermedia 25611
Prevotella melaninogenica 25845
Prevotella nigrescens 33563
Propionibacterium acnes 11827 and 11828
Selenomonas noxia 43541
Streptococcus anginosus 33397
Streptococcus constellatus 27823
Streptococcus gordonii 10558
Streptococcus intermedius 27335
Streptococcus mitis 49456
Streptococcus oralis 35037
Streptococcus sanguis 10556
Tannerella forsythensis 43037
Treponema denticola B1
Treponema socranskii S1
Veillonella parvula 10790
Results
Unmatched subjects
The levels of salivary bacteria in subjects with and without OSCC are illustrated in Fig 1. Comparisons between the 229 OSCC-free and 45 OSCC subjects indicated that 6 common oral bacteria (P. melaninogenica, C. gingivalis, Capnocytophaga ochracea, Eubacterium saburreum, Leptotrichia buccalis and S. mitis) differed (p < 0.001). Elevated salivary counts of ≥0.4 × 105/ml of 3 bacteria, C. gingivalis, P. melaninogenica and S mitis, were found to have diagnostic sensitivity and specificity ≥80%. The remaining species, including the three bacteria that were recovered in lower levels in salivary samples of OSCC patients, E. saburreum, L. buccalis, and C. ochracea, did not contribute significantly to the diagnostic test as their diagnostic sensitivity and specificity values were ≤60%.
Figure 1 Salivary counts for the 40 test species in both populations. Mean counts (±SD) of 40 test species in saliva samples of 229 OSCC-free & 45 OSCC subjects (*** = p < 0.001).
Diagnostic sensitivity and specificity for different counts of C. gingivalis, P. melaninogenica and S. mitis/ml saliva for the unmatched populations are illustrated in Figure 2. C. gingivalis was the species most closely associated with oral cancer lesions but diagnostic sensitivity and specificity was highest when levels of the 3 bacteria were each ≥0.4 × 105. Median DNA probe counts of C. gingivalis P. melaninogenica and S. mitis in OSCC-free subjects were 0.25, 0.63, and 0.31 × 105 per ml of saliva. In contrast, the median DNA probe counts of these 3 species in the 45 OSCC subjects were 3.24, 5.62 and 1.62 × 105 per ml of saliva, respectively.
Figure 2 Diagnostic sensitivity and specificity of 3 bacterial species in the unmatched populations. a. Diagnostic sensitivity and specificity when C. gingivalis is at different salivary counts × 105/ml and both P. melaninogenica and S. mitis = 0 b. Diagnostic sensitivity and specificity when P. melaninogenica is at different salivary counts × 105/ml, C. gingivalis ≥ 0.4 × 105/ml and S. mitis = 0 c. Diagnostic sensitivity and specificity when S. mitis is at different salivary counts × 105/ml and both C. gingivalis and P. melaninogenica ≥0.4 × 105/ml
Matched Subjects
45 OSCC-free controls were matched by computer for age, gender and smoking history with the 45 OSCC subjects. Diagnostic sensitivity and specificity for detection of oral cancer using levels of salivary bacteria were computed as described above. The results for the matched population were similar to those for the unmatched comparisons namely that the increased counts of C. gingivalis, P. melaninogenica and S. mitis were 80% diagnostically sensitive and 82% diagnostically specific for the presence of OSCC (Figure 3). As before, the remaining 37 species did not improve sensitivity or specificity.
Figure 3 Diagnostic sensitivity and specificity of 3 bacterial species in the matched populations. a. Diagnostic sensitivity and specificity when C. gingivalis is at different salivary counts × 105/ml and both P. melaninogenica and S. mitis = 0 b. Diagnostic sensitivity and specificity when P. melaninogenica is at different salivary counts × 105/ml, C. gingivalis ≥0.4 × 105/ml and S. mitis = 0 c. Diagnostic sensitivity and specificity when S. mitis is at different salivary counts × 105/ml and both C. gingivalis and P. melaninogenica ≥0.4 × 105/ml
Discussion
Results from this investigation demonstrated that oral cancer subjects had elevated counts (p < 0.001) of C. gingivalis, P. melaninogenica and S. mitis in saliva compared to OSCC-free subjects. These results are borderline in significance after adjusting for multiple comparisons. However, when each species was ≥0.4 × 105 they indicated the presence of an OSCC lesion with 80% diagnostic sensitivity and ≥82% specificity in both matched and unmatched populations.
The reason for this finding is unclear. One explanation may relate to the altered cell surface receptors observed in cancer cells [36,39,41]. It seems reasonable that alterations in tumor cell receptors could change the adhesion of certain species of bacteria. This was shown, as previously discussed, in an in vitro study of HBEC and buccal cell carcinoma cell lines using the common oral bacterium S. sanguis by Neeser [36]. One might expect that as Neeser found decreased colonization of S. sanguis, a similar study using S. mitis would result in a reduced colonization of oral cancer cells. Interestingly, our previous investigation of 225 OSCC-free subjects provided evidence to the contrary. Colonization of different oral sites differed among the 40 test species, even among those of the same genera, such as streptococci. For example, S. sanguis and S. mitis both colonized the oral soft and hard tissues; however, marked differences in their proportions at these sites were noted. Highly species-specific oral colonization by streptococci has been reported by other investigators [29-32,57,58].
Saliva was found to be similar in microbial profile to the soft tissues. This was a significant finding from the study of the OSCC-free population. In contrast, the microbiota of the teeth and saliva differed markedly. These results agreed with previous studies [30,57,58]. Thus, if alterations in bacterial adhesion to OSCC cells observed in vitro exist in vivo, colonization of OSCC lesions would be affected. Shifts in the soft tissue microbiota of the oral cavity appear likely to affect salivary levels as well.
A screening test for oral cancer based on salivary counts of bacterial species is appealing. Saliva is now meeting the demand for inexpensive, noninvasive, and easy-to use diagnostic aids for oral and systemic diseases, and for assessing risk behaviors such as tobacco and alcohol use. Detection of HIV by the presence of virus-specific antibodies in saliva, for example, has led to the development of commercially available test kits [16]. If increased numbers of certain salivary species are shown to be a signature of oral cancer, an early diagnostic test for OSCC may be developed, reducing the morbidity and mortality of this devastating cancer.
Studies to examine the validity of these findings are planned. If the results of this study are validated it will be important to address whether oral bacteria can be used as indicators of oral cancer and whether certain oral species contribute to carcinogenesis.
Conclusion
Results of the present study suggest that high salivary counts of C. gingivalis, P. melaninogenica and S. mitis may be diagnostic indicators of OSCC. These findings taken with those of an earlier study indicate that the presence of an OSCC has a more powerful effect on the salivary microbiota than either smoking or periodontal infections.
Authors' contributions
DLM conceived of this investigation and the preliminary studies, constructed the study design, developed clinical sampling techniques, wrote hospital protocols and K-23 grant application that funded the project, collected and processed the majority of OSCC subject samples, modified laboratory protocols and drafted the manuscript. ADH made substantial contributions to the conception and study design. She coordinated the study of OSCC-free subjects, conducted the statistical analysis of the OSCC-free data and made substantial contributions to the interpretation of the data for the OSCC-free and OSCC populations. PMD, CMN and MRP made substantial contributions to the conception, design and coordination of the study during the preliminary studies and initiation of this investigation and were instrumental in coordinating clinical evaluations with the recruitment and sampling of OSCC subjects. JMG was instrumental in the conception and design of the study and performed the majority of statistical analyses and interpretations of data for the OSCC subjects. JMG made critical revisions for important intellectual content of the manuscript. All authors have given final approval of the manuscript.
Acknowledgements
My sincere gratitude goes to Rosemary Costello RN, MS, OCN, and Julia Kazakin MD for their unwavering support of this study. Many thanks to James Rocco MD, PhD, Sovanda Som, BS, Tina Yaskell, BS, the staff of the Head and Neck Clinics at the Dana Farber Cancer Institute and Massachusetts General Hospital for their invaluable assistance in conducting this investigation. Finally, special thanks go to the patients who enrolled in this study; it would not have been possible without them.
==== Refs
SEER Cancer Statistics Review 2002 US Department of Health and Human Services, Public Health Service, National Institutes of Health, Bethesda, MD
The American Cancer Society
Silverman S Oral cancer Semin Dermatol 2001 13 132 137 8060825
Canto MT Devesa SS Oral cavity and pharynx cancer incidence rates in the United States 1975–1998 Oral Oncol 2002 38 610 617 12167440 10.1016/S1368-8375(01)00109-9
The Oral Cancer Foundation
Mignogna MD Fedele S Russo LL The World Cancer Report and the burden of oral cancer Eur J Cancer Prev 2004 13 139 142 15100581 10.1097/00008469-200404000-00008
Goldberg HI Lockwood SA Wyatt SW Crossett LS Trends and differentials in mortality from cancers of the oral cavity and pharynx in the United States, 1973–1987 Cancer 1994 74 565 572 8033034
Moore RJ Doherty DA Do KA Chamberlain RM Khuri FR Racial disparity in survival of patients with squamous cell carcinoma of the oral cavity and pharynx Ethn Health 2001 6 165 77 11696928 10.1080/13557850120078099
Myers JN Elkins T Roberts D Byers RM Squamous cell carcinoma of the tongue in young adults: increasing incidence and factors that predict treatment outcomes Otolaryngol Head Neck Surg 2000 122 44 51 10629481
Pitman KT Johnson JT Wagner RL Myers EN Cancer of the tongue in patients less than forty Head Neck 2000 22 297 302 10748455 10.1002/(SICI)1097-0347(200005)22:3<297::AID-HED14>3.0.CO;2-3
SEER Cancer Statistics Review 1998 US Department of Health and Human services, Public Health Service, National Institutes of Health, Bethesda, MD
Neville BW Day TA Oral cancer and precancerous lesions CA: Cancer J Clin 2002 52 195 215 12139232
Silverman S Jr Demographics and occurrence of oral and pharyngeal cancers. The outcomes, the trends, the challenge J Am Dent Assoc 2001 7 11
Holmes JD Dierks EJ Homer LD Potter BE Is detection of oral and oropharyngeal squamous cancer by a dental health care provider associated with a lower stage at diagnosis? J Oral Maxillofac Surg 2003 61 285 291 12618965 10.1053/joms.2003.50056
Alfano MC Horowitz AM Professional and community efforts to prevent morbidity and mortality from oral cancer J Am Dent Assoc 2001 24S 29S 11803649
National Institute of Dental and Craniofacial Research
Blot WJ McLaughlin JK Winn DM Austin DF Greenberg RS Preston-Martin S Bernstein L Schoenberg JB Stemhagen A Fraumeni JF Jr Smoking and drinking in relation to oral and pharyngeal cancer Cancer Res 1988 48 3282 3287 3365707
Llewellyn CD Johnson NW Warnaulasuriya KA Risk factors for squamous cell carcinoma of the oral cavity in young people – a comprehensive literature review Oral Oncol 2001 37 401 418 11377229 10.1016/S1368-8375(00)00135-4
Shugars DC Patton LL Detecting, diagnosing, and preventing oral cancer Nurse Pract 1997 22 105, 109 10 113-5 passim 9211456
Thomas JE Faecher RS A physician's guide to early detection of oral cancer Geriatrics 1992 47 58 63 1729147
Hawkins RJ Wang EE Leake JL Preventive health care, 1999 update: prevention of oral cancer mortality. The Canadian Task Force on Preventive Health Care J Can Dent Assoc 1999 65 617 10859725
Rice DH Weimert TA Altered bacterial flora and clinical course with intraoral cancer Laryngoscope 1978 88 1861 1863 713677
Nagy KN Sonkodi I Szoke I Nagy E Newman HN The microflora associated with human oral carcinomas Oral Oncology 1998 34 304 308 9813727 10.1016/S1368-8375(98)00008-6
Sasaki H Ishizuka T Muto M Nezu M Nakanishi Y Inagaki Y Watanabe H Watanabe H Terada M Presence of Streptococcus anginosus DNA in esophageal cancer, dysplasia of esophagus, and gastric cancer Cancer Research 1998 58 2991 2995 9679961
Sakamoto H Naito H Ohta Y Tanakna R Maeda N Sasaki J Nord CE Isolation of bacteria from cervical lymph nodes in patients with oral cancer Archives of Oral Biology 1999 44 789 793 10530911 10.1016/S0003-9969(99)00079-5
Tateda M Shiga K Saijo S Sone M Hori T Yokoyama J Matsuura K Takasaka T Miyagi T Streptococcus anginosus in head and neck squamous cell carcinoma: implication in carcinogenesis Int J Mol Med 2000 6 699 703 11078831
Shiga K Tateda M Saijo S Hori T Sato I Tateno H Matsuura K Takasaka T Miyagi T Presence of Streptococcus infection in extra-oropharyngeal head and neck squamous cell carcinoma and its implication in carcinogenesis Oncol Rep 2001 8 245 248 11182034
Crean S-J Nair SP Fardy M Wilson M Spratt D Identification of bacterial DNA using PCR cloning within oral squamous cell carcinomas [abstract] J Dent Res 2002 81 A364
Krasse B The proportional distribution of Streptococcus salivarius and other streptococci in various parts of the mouth Odont Revy 1954 5 203 211
Liljemark WF Gibbons RJ Proportional distribution and relative adherence of Streptococcus mitis on various surfaces in the human oral cavity Infect Immun 1972 6 852 859 4637299
Van Houte J Gibbons RJ Banghart SB Adherence as a determinant of the presence of Streptococcus salivarius and Streptococcus sanguis on the human tooth surface Arch Oral Biol 1970 15 1025 1034 5275860 10.1016/0003-9969(70)90115-9
Van Houte J Gibbons RJ Pulkkinen AJ Adherence as an ecological determinant for streptococci in the human mouth Arch Oral Biol 1971 16 1131 1141 5289677 10.1016/0003-9969(71)90042-2
Gibbons RJ van Houte J Dental caries Annu Rev Med 1975 26 121 136 1096752 10.1146/annurev.me.26.020175.001005
Gibbons RJ Role of adhesion in microbial colonization of host tissues: a contribution of oral microbiology J Dent Res 1996 75 866 870 8675796
Gibbons RJ Bacterial adhesion to oral tissues: A model for infectious diseases J Dent Res 1989 68 750 760 2654229
Neeser JR Grafstrom RC Woltz A Brassart D Fryder V Guggenheim B A 23 kDa membrane glycoprotein bearing NeuNAc alpha 2-3Gal beta 1-3GalNAc O-linked carbohydrate chains acts as a receptor for Streptococcus sanguis OMZ 9 on human buccal epithelial cells Glycobiology 1995 5 97 104 7772872
Bhavanandan VP Cancer-associated mucins and mucin-type glycoproteins Glycobiology 1991 1 493 503 1822231
Hakomori S Tumor malignancy defined by aberrant glycosylation and sphingo(glyco)lipid metabolism Cancer Res 1996 56 5309 5318 8968075
Hoang-Vu C Bull K Schwarz I Krause G Schmutzler C Aust G Kohrle J Dralle H Regulation of CD97 protein in thyroid carcinoma J Clin Endocrinol Metab 1999 84 1104 1109 10084602 10.1210/jc.84.3.1104
Tieng V Le Bouguenec C du Merle L Bertheau P Desreumaux P Janin A Charron D Toubert A Binding of Escherichia coli adhesin AfaE to CD55 triggers cell-surface expression of the MHC class I-related molecule MICA Proc Natl Acad Sci 2002 99 2977 2982 11830641 10.1073/pnas.032668099
Williams P Chaudhry Y Goodfellow IG Billington J Powell R Spiller OB Evans DJ Lea S Mapping CD55 function. The structure of two pathogen-binding domains at 1.7 A J Biol Chem 2003 278 10691 10696 12499389 10.1074/jbc.M212561200
Mager DL Ximenez-Fyvie LA Haffajee AD Socransky SS Distribution of selected bacterial species on intra-oral surfaces J Clin Periodontol 2003 30 644 654 12834503 10.1034/j.1600-051X.2003.00376.x
Mager DL Haffajee AD Socransky SS The effects of periodontitis and smoking on the microbiota of oral mucous membranes and saliva in systemically healthy subjects J Clin Periodontol 2003 30 1031 1037 15002888 10.1046/j.0303-6979.2003.00418.x
van der Velden U Abbas F Hart AA Experimental gingivitis in relation to susceptibility to periodontal disease. (I.) Clinical observations J Clin Periodontol 1985 12 61 68 3855872
Lie MA Danser MM van der Weijden GA Timmerman MF de Graaff J van der Velden U Oral microbiota in subjects with a weak or strong response in experimental gingivitis J Clin Periodontol 1995 22 642 647 8583022
Muller H-P Zoller L Eger T Hoffmann S Lobinsky D Natural distribution of oral Actinobacillus actinomycetemcomitans in young men with minimal periodontal disease J Periodont Res 1996 31 373 380 8884631
Asikainen S Chen C Oral ecology and person-to-person transmission of Actinobacillus actinomycetemcomitans and Porphyromonas gingivalis Periodontology 2000 1999 20 65 81 10522223
Danser MM Timmerman MF van Winkelhoff AJ van der Velden U The effect of periodontal treatment on periodontal bacteria on the oral mucous membranes J Periodontol 1996 67 478 485 8724705
Lie MA van der Weijden GA Timmerman MF Loos BG van Steenbergen TJ van der Velden U Oral microbiota in smokers and non-smokers in natural and experimentally-induced gingivitis J Clin Periodontol 1998 25 677 686 9722273
Colman G Beighton D Chalk AJ Wake S Cigarette smoking and the microbial flora of the mouth Aust Dent J 1976 21 111 118 1068665
van Winkelhoff AJ van der Velden U Winkel EG De Graaff J Black-pigmented Bacteroides and motile organisms on oral mucosal surfaces in individuals with and without periodontal breakdown J Perio Res 1986 21 434 439
Eger T Zoller L Muller H-P Hoffmann S Lobinsky D Potential diagnostic value of sampling oral mucosal surfaces for Actinobacillus actinomycetemcomitans in young adults Eur J Oral Sci 1996 104 112 117 8804898
Socransky SS Smith C Martin L Paster BJ Dewhirst FE Levin AE "Checkerboard" DNA-DNA hybridization Biotechniques 1994 17 788 792 7833043
Smith GL Socransky SS Smith CM Rapid method for the purification of DNA from subgingival microorganisms Oral Microbiol Immunol 1989 4 47 51 2628868
Feinberg AP Vogelstein BA Technique for radiolabeling DNA restriction endonuclease fragments to high specific activity Anal Biochem 1983 132 6 13 6312838 10.1016/0003-2697(83)90418-9
Schramm W Angulo GB Torres PC Burgess-Cassler A A simple saliva-based test for detecting antibodies to human immunodeficiency virus Clin Diagn Lab Immunol 1999 6 577 80 10391866
Smith DJ Anderson JM King WF Van Houte J Taubman MA Oral streptococcal colonization of infants Oral Microbiol Immunol 1993 8 1 4 8510978
Krasse B The proportional distribution of different types of streptococci in saliva and plaque material Odont Revy 1953 4 304 312
|
15987522
|
PMC1226180
|
CC BY
|
2021-01-04 16:39:28
|
no
|
J Transl Med. 2005 Jul 7; 3:27
|
utf-8
|
J Transl Med
| 2,005 |
10.1186/1479-5876-3-27
|
oa_comm
|
==== Front
PLoS GenetPLoS GenetpgenplgeplosgenPLoS Genetics1553-73901553-7404Public Library of Science San Francisco, USA 1618419010.1371/journal.pgen.001003705-PLGE-RA-0082R2plge-01-03-09Research ArticleCell BiologyMolecular Biology - Structural BiologyGenetics/Gene FunctionGenetics/Gene ExpressionEukaryotesNematodesCaenorhabditisGenetic Interactions Due to Constitutive and Inducible Gene Regulation Mediated by the Unfolded Protein Response in C. elegans
Genetic and Microarray Analysis of
C. elegans UPR
Shen Xiaohua 1¤Ellis Ronald E 2Sakaki Kenjiro 1Kaufman Randal J 1*1 Howard Hughes Medical Institute, Department of Biological Chemistry, University of Michigan Medical Center, Ann Arbor, Michigan, United States of America
2 Department of Molecular Biology, The UMDNJ School of Osteopathic Medicine, Stratford, New Jersey, United States of America
Kim Stuart EditorStanford University School of Medicine, United States of America*To whom correspondence should be addressed. E-mail: [email protected]¤ Current address: Children's Hospital, Dana Farber Cancer Institute and Harvard Medical School, Boston, Massachusetts, United States of America
9 2005 23 9 2005 1 3 e3715 4 2005 8 8 2005 Copyright: © 2005 Shen et al.2005This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are properly credited.The unfolded protein response (UPR) is an adaptive signaling pathway utilized to sense and alleviate the stress of protein folding in the endoplasmic reticulum (ER). In mammals, the UPR is mediated through three proximal sensors PERK/PEK, IRE1, and ATF6. PERK/PEK is a protein kinase that phosphorylates the alpha subunit of eukaryotic translation initiation factor 2 to inhibit protein synthesis. Activation of IRE1 induces splicing of XBP1 mRNA to produce a potent transcription factor. ATF6 is a transmembrane transcription factor that is activated by cleavage upon ER stress. We show that in Caenorhabditis elegans, deletion of either ire-1 or xbp-1 is synthetically lethal with deletion of either atf-6 or pek-1, both producing a developmental arrest at larval stage 2. Therefore, in C. elegans, atf-6 acts synergistically with pek-1 to complement the developmental requirement for ire-1 and xbp-1. Microarray analysis identified inducible UPR (i-UPR) genes, as well as numerous constitutive UPR (c-UPR) genes that require the ER stress transducers during normal development. Although ire-1 and xbp-1 together regulate transcription of most i-UPR genes, they are each required for expression of nonoverlapping sets of c-UPR genes, suggesting that they have distinct functions. Intriguingly, C. elegans atf-6 regulates few i-UPR genes following ER stress, but is required for the expression of many c-UPR genes, indicating its importance during development and homeostasis. In contrast, pek-1 is required for induction of approximately 23% of i-UPR genes but is dispensable for the c-UPR. As pek-1 and atf-6 mainly act through sets of nonoverlapping targets that are different from ire-1 and xbp-1 targets, at least two coordinated responses are required to alleviate ER stress by distinct mechanisms. Finally, our array study identified the liver-specific transcription factor CREBh as a novel UPR gene conserved during metazoan evolution.
Synopsis
The endoplasmic reticulum (ER) is an intracellular organelle where proteins fold and assemble prior to transport to the cell surface. The ER contains a finely tuned quality control apparatus to ensure that improperly folded proteins are retained in the ER lumen. A variety of physiological demands, environmental perturbations, and pathological conditions compromise protein folding in the ER and lead to the accumulation of unfolded proteins. The unfolded protein response (UPR) is an evolutionarily conserved intracellular adaptive signaling pathway that alleviates protein-folding defects in the ER. The unfolded protein signal is transmitted from the ER to the nucleus by three pathways involving the proteins ATF-6, PEK-1, and IRE-1/XBP-1. However, it is not known how these three pathways coordinate downstream transcriptional activation to mediate either cell adaptation or cell death. The authors have studied the nematode Caenorhabditis elegans to present a comprehensive genetic and gene expression analysis of the three UPR pathways. The findings demonstrate that the UPR regulates the expression of hundreds of genes in the presence, as well as the absence, of ER stress in a manner that is more complex and diverse than previously known.
Citation:Shen X, Ellis RE, Sakaki K, Kaufman RJ (2005) Genetic interactions due to constitutive and inducible gene regulation mediated by the unfolded protein response in C. elegans. PLoS Genet 1(3): e37.
==== Body
Introduction
The endoplasmic reticulum (ER) is the primary site where all secretory and membrane proteins fold prior to transiting the secretory pathway. In addition, the ER is the major Ca++ storage organelle and the site of lipid and oligosaccharide synthesis [1]. Therefore, ER homeostasis is essential for cellular function and survival in all eukaryotes. The unfolded protein response (UPR) is a transcriptional and translational regulatory pathway that evolved to sense and alleviate protein-folding stress in the ER caused by physiological demands or environmental variation [2,3].
In yeast, the UPR is solely dependent on Ire1p [4,5]. Ire1p1 is a bifunctional protein kinase and endoribonuclease that cleaves an unconventional 252-base intron from HAC1 mRNA, which encodes a basic leucine zipper (bZIP)–containing transcription factor. Elegant studies in Saccharomyces cerevisiae identified 381 UPR-inducible genes that function primarily in ER protein folding and trafficking, ER-associated degradation (ERAD), and phospholipid metabolism [6]. In mammals, two homologs of IRE1, IRE1α and IRE1β, exist, and both are able to cleave a 26-base intron in XBP1 mRNA to create a translational frame-shift that alters the carboxyl terminus of the protein to produce a potent bZIP transcription factor [7–11]. The primary targets that require the IRE1/XBP1 pathway are genes encoding functions in ERAD, such as EDEM (ER degradation-enhancing α-mannosidase-like protein), which recognizes specific glycoforms on unfolded proteins and directs them to the 26S proteasome [12].
In mammals, two additional ER transmembrane proteins, PERK/PEK and ATF6 mediate the UPR [13,14]. PERK phosphorylates the α subunit of eukaryotic translation initiation factor 2 (eIF2α) to mediate translational attenuation [15,16]. In addition, eIF2α phosphorylation paradoxically increases translation of ATF4 mRNA, which encodes a transcriptional activator required for induction of an anti-oxidative response and amino acid biosynthesis and transport functions [16,17]. However, induction of the protein chaperone BiP, a primary marker of UPR activation, remained intact in ire1
−/− mouse embryonic fibroblasts (MEFs) [18] and was only partially reduced in perk−/− MEFs and in eIF2α-phosphorylation-resistant MEFs [16,19], suggesting the existence of another UPR signaling pathway in mammals. ATF6 p90 is a type II transmembrane protein that contains a bZIP transcription factor domain in its cytosolic amino terminus. When unfolded proteins accumulate in the ER, ATF6 transits to the Golgi compartment, where it is cleaved by site-1 protease (S1P) and site-2 protease (S2P) to produce a p50 cytoplasmic soluble bZIP-containing transcription factor [12,14,18,20]. In mammals, there are two homologs of ATF6, ATF6α and ATF6β [21,22]. BiP induction is completely abolished in cells that lack S2P and cannot process ATF6 [18], supporting the idea that the primary targets of ATF6 are protein chaperones that augment ER protein-folding capacity. In addition, forced expression of cleaved ATF6α (p50) induced genes encoding ER-resident protein chaperones [23]. However, reduction in ATF6α and/or ATFβ mRNA did not significantly affect UPR gene induction as analyzed by gene profiling [24]. Thus, the specific roles of ATF6 in activation of ER stress-induced gene expression are in question.
Although three UPR-signal-transducing pathways have been characterized in mammals, it is not known how they coordinate downstream transcriptional activation of different target genes to mediate responses that direct either adaptation or apoptosis when protein folding in the ER is compromised. Unfortunately, the absence of homologous ATF6 and PERK signaling pathways in yeast limits the applicability of studying the UPR in yeast to understanding this process in higher eukaryotes. On the other hand, analysis in mice is complicated by the presence of multiple homologs for IRE1 and ATF6, and the embryonic lethality of homozygous mutations in mammals. Furthermore, studies in MEFs cannot elucidate the physiological and developmental functions of these pathways, because different cell types have different requirements for the UPR sub-pathways [25]. Thus, we selected the nematode Caenorhabditis elegans as a model for studying the UPR in mammals. Although a previous cDNA microarray study in C. elegans identified only 26 genes that require xbp-1 for up-regulation in adult nematodes following ER stress [26], we suspected that many more target genes exist. In this study, we used genetic and microarray studies in C. elegans to identify the cellular functions of individual UPR signaling pathways and elucidate how these pathways are coordinated during ER stress.
Results
C. elegans atf-6 Complements the ire-1/xbp-1 Pathway
Based on sequence homology, we identified F45E6.2 as the only C. elegans homolog of ATF6α, and named it atf-6. C. elegans ATF-6 has 22% identity to human ATF6α (Figure 1) and is remotely homologous to human ATF6β (CREBL1, about 16% identity). Worm ATF-6 contains a serine-rich region at the N-terminus, a bZIP domain, and a hydrophobic stretch consisting of 22 residues that is likely to form a transmembrane domain (Figure 1). The C-terminus contains two regions with high homology to their mammalian counterparts that may be required for BiP association and translocation to the Golgi [27]. These similarities suggest that C. elegans ATF-6 is a type II ER transmembrane protein that may function like mammalian ATF6α.
Figure 1 The Sequence Alignment of C. elegans, Human, and Murine ATF6 Homologs
The sequence alignment of C. elegans (ce), human (hs), and mouse (ms) shows five conserved regions in worm ATF-6 including a serine-rich domain, a bZIP domain, a transmembrane domain, and two C-terminal homology regions (HR-I and HR-II). Blue indicates residues that are conservative across species, green indicates blocks of similar residues, yellow indicates identical residues, and grey indicates weak similarity.
To elucidate the function of atf-6, RNA interference (RNAi) was used to inactivate the gene. The atf-6(RNAi) animals had no obvious phenotype differences from wild-type. However, atf-6(RNAi); ire-1(v33) double mutants were sluggish and sick, arrested development at the L2 larval stage, and died soon thereafter. They showed intestinal degeneration similar to that found in ire-1; pek-1 double mutants (Figure 2), and developed many vacuoles in the intestinal cells. As expected, atf-6(RNAi); xbp-1(RNAi) animals also died early in larval development. Interestingly, RNAi-mediated silencing of C. elegans S2P together with ire-1 also caused L2 arrest and intestinal degeneration, although silencing S2P alone caused no phenotype abnormalities (data not shown). These results show that either signaling through ATF-6/S2P or through IRE-1/XBP-1 is sufficient for normal development. Furthermore, the findings imply that C. elegans ATF-6 may be regulated by S2P-mediated proteolytic cleavage as in mammals. Since C. elegans does not have an identifiable S1P homolog, S2P might be the only protease that cleaves ATF-6. Alternatively, it is possible that a protease that is not similar to S1P cleaves ATF-6 prior to cleavage by S2P. For example, in Escherichia coli the transmembrane protein RseA is cleaved sequentially by DegS and RseP, and although RseP is a homolog of S2P, DegS is not similar to S1P [28]. Finally, silencing both atf-6 and pek-1 caused no obvious phenotype abnormalities (Figure 2D), suggesting that atf-6 and pek-1 may function in the same pathway, and that this pathway is partially redundant to the ire-1/xbp-1 pathway.
Figure 2
C. elegans atf-6 and pek-1 Display Partially Redundant Roles in Complementing ire-1/xbp-1 for Larval Survival and Development
(A) Characterization of C. elegans atf-6 and its ok551 deletion allele. The atf-6 gene structure is depicted in boxes and lines, representing exons and introns, respectively. The atf-6(ok551) allele lacks 1,900 bp of genomic sequence, and has the potential to encode a protein without the leucine zipper portion of the bZIP domain, the transmembrane domain, and ER lumenal domain. The atf-6(ok551) deletion allele can be detected by PCR, using the primers indicated by arrows.
(B) Genetic interactions of atf-6, ire-1, and pek-1. Animals with the genotype ire-1(v33); atf-6(ok551) arrested as young larvae, showing that loss of both ire-1 and atf-6 is lethal. The ire-1(v33)/ mnC1; atf-6(ok551)/ pek-1(ok275) animals (P0) segregated healthy F1 progeny with the genotype ire-1(v33); atf-6(ok551)/ pek-1(ok275), which in turn produced dead F2 animals with exactly the same genotype, suggesting that ATF-6 and PEK-1 function synergistically to cope with endogenous ER stress during development.
(C) Nomarski micrograph of a 3-d-old atf-6(RNAi); ire-1(v33) animal. The germline of this animal did not develop past the L2 larval stage.
(D) Comparisons of intestinal degeneration in various double mutants: (i) ire-1(v33); pek-1(ok275), (ii) xbp-1(RNAi); pek-1(ok275), (iii) ire-1(v33); atf-6(RNAi), and (iv) atf-6(RNAi); pek-1(ok275). Normaski micrographs show a portion of the intestine. Mutants in (i)–(iii) arrested at the L2 larval stage and showed intestinal degeneration. The mutants in (iv) had an intestinal morphology similar to the wild-type. Yellow arrows indicate vacuoles in intestinal cells. Red arrowheads indicate light-reflective aggregates appearing in some mutants ([ii]and [iii]).
atf-6 and pek-1 Share a Common Regulatory Function
To confirm these RNAi results, we characterized the atf-6 allele ok551. DNA sequence analysis revealed a 1,900-bp deletion, extending from 1,276 bp to 3,175 bp downstream of the start codon (Figure 2A). The mutant transcript contains only the first four exons and part of exon 5, and encodes a short protein consisting of the N-terminal transcriptional activation domain and the basic region of the bZIP domain. The leucine zipper, the transmembrane domain, and the C-terminal ER luminal domain are deleted. Because the mutant protein can neither sense ER stress nor bind to its DNA targets, we suspect that it causes a loss of function. Despite this deficit, atf-6(ok551) animals appear wild-type, and respond normally to tunicamycin, an agent that induces ER stress by inhibition of asparagine-linked glycosylation (Figure S1).
To dissect genetic interactions among the three known ER stress transducers in C. elegans, we constructed the strain ire-1(v33) II/ mnC1; atf-6(ok551) +/ + pek-1(ok275) X. This trans-heterozygote for atf-6 and pek-1 is stable, since these genes are located near each other on the X chromosome, and the ire-1(v33) deletion is balanced by the marker chromosome mnC1 (Figure 2B). Next, we used PCR genotyping to isolate and study offspring of ire-1(v33)/ mnC1; atf-6(ok551) animals. From these heterozygous parents, we found that 0/50 adult offspring had the genotype ire-1(v33); atf-6(ok551), but 28/100 eggs were homozygous for both genes. We conclude that the ire-1; atf-6 double mutants die before adulthood, just as we observed using RNAi for atf-6.
Animals with the genotype ire-1(v33)/ mnC1; atf-6(ok551) +/ + pek-1(ok275) segregated viable ire-1(v33); atf-6(ok551) +/ + pek-1(ok275) F1 progeny, but these animals produced F2 progeny that all arrested early in larval development (Figure 2B). The lethality of the ire-1(v33); atf-6(ok551) genotype was expected based on our RNAi experiments and the lethality of ire-1(v33); pek-1(ok275) was reported previously [7]. However, the finding that ire-1(v33); atf-6(ok551) +/ + pek-1(ok275) heterozygotes died, whereas their parents (which had the same genotype) lived, shows that ire-1 has a maternal effect. This result suggests that ire-1 might act during embryogenesis or early development. More importantly, these results also show that the loss of a single copy each of the atf-6 and pek-1 genes was sufficient to kill ire-1-null animals. The observation that haplo-insufficiency for both atf-6 and pek-1 is equivalent to loss of both copies of atf-6, or both copies of pek-1, suggests that atf-6 and pek-1 share a common regulatory function.
IRE-1 Acts through XBP-1 to Induce Transcription of Many UPR Genes
Genetic interactions suggested that ire-1, pek-1, and atf-6 regulate the worm UPR and are required for growth and survival. To elucidate their functions, we performed microarray analysis. Defects in ire-1/xbp-1 signaling in the presence of a mutation in either pek-1 or atf-6 caused L2 larval arrest, implying that UPR signaling may be particularly important at this stage of development. Thus, we carried out a series of microarray analyses using synchronized L2 larvae.
Although the atf-6(ok551) allele is synthetic lethal with an ire-1 deletion, the mutant protein encoded by atf-6(ok551) could associate with other DNA-binding proteins to activate transcription. Because RNAi appeared effective at knocking down atf-6 function, we used atf-6(RNAi) animals for the microarray study. To confirm the efficacy of our RNAi treatment, we silenced atf-6 in ire-1(v33) control animals, and found that all ire-1(v33); atf-6(RNAi) mutants arrested and died at the L2 stage, which indicates that atf-6 was efficiently silenced. All synchronized L2 animals, including N2, ire-1(v33), xbp-1(zc12), pek-1(ok275), and atf-6(RNAi), were treated for 4 h in the absence or 4 h in the presence of 30 μg/ml of tunicamycin to induce ER stress [7,9]. To achieve statistical significance, we prepared RNA samples from independently treated animals for each chip analysis. We performed three biological repeats for each strain and treatment except for tunicamycin-treated xbp-1(zc12) animals, for which the analyses were only repeated twice.
To determine how tunicamycin-induced ER stress affects gene expression in each strain, we used analysis of variance (ANOVA) to study the interactions between each nematode strain and drug treatment [29,30], and identified 4,050 probes with an interaction p-value less than 0.01 (one gene could have several probes in the Affymetrix C. elegans genome array). This threshold was chosen arbitrarily to ensure that the ANOVA was stringent but would not miss significant genes. Among these 4,050 probes, 202 genes were up-regulated at least 2-fold in wild-type N2 animals by tunicamycin treatment, but not were properly induced in at least one of the mutant strains (i.e., the fold-induction in mutant strains was less than half that observed in the N2 strain). We call this set of 202 genes inducible UPR (i-UPR) genes (Figure 3). The list of 202 genes may underrepresent genes regulated by the i-UPR because of our stringent set of criteria. For example, cnx-1, which encodes calnexin, and crt-1, which encodes calreticulin, were up-regulated about 1.45-fold and 1.74-fold, respectively, by tunicamycin in an ire-1/xbp-1-dependent manner. However, they were not included in the list of i-UPR genes because they had less than 2-fold induction. To confirm the array results, we analyzed 30 genes by quantitative real-time RT-PCR (Figure S2). Expression patterns consistent with the array data were observed for all 30 genes, showing the sensitivity, quality, and validity of our array data.
Figure 3 Transcriptional Targets of ire-1, xbp-1, pek-1, and atf-6
(A) Venn diagram showing the sets of i-UPR genes regulated by ire-1, xbp-1, pek-1, and atf-6.
(B) Venn diagram showing the sets of c-UPR genes regulated by ire-1, xbp-1, pek-1, and atf-6.
Gene functions were obtained by comparing direct downloads from the Affymetrix Web site to annotations in Wormbase [31]. For uncharacterized worm genes, functions were inferred based on their mammalian or yeast homologs. About 84% of i-UPR genes (170 out of 202 genes) were regulated by both ire-1 and xbp-1 (Figure 3A; Table S1). The finding that most ire-1 and xbp-1 targets overlap supports the hypothesis that ire-1 and xbp-1 function in a single linear pathway. Of genes with known functions, about 40% were involved in the secretory pathway (Figure 4A). ire-1/xbp-1 was required for tunicamycin-mediated induction of previously characterized protein-folding catalysts including BiP (HSP-4, HSP-3), protein disulfide isomerase (PDI-1, PDI-2), DNJ-7 (also known as PERK inhibitor [p58IPK]), and ERO-1, which eliminates highly reactive oxygen species in the ER [6,7,12,17,24,26,32–34], as well as five signal peptidases that would facilitate protein processing under ER stress (Table S1). Previous studies have suggested that the UPR and ERAD are intimately coordinated and UPR induction increases ERAD capacity [6,12]. In addition to previously identified ERAD genes DER-1 (Derlin-1), HRD1, SEL-1 (HRD3), EDEM, and ERD-2 [6,12,18,24,26,35], Derlin-1-interacting AAA ATPase p97, vesicle-fusing ATPase NSF-1, and ER retention protein Rer1 were induced by the UPR in an ire-1- and xbp-1-dependent manner. Moreover, a set of genes encoding functions involved in various aspects of protein trafficking such as clathrin (mainly COP-II) vesicle formation, translocon assembly, vesicle docking, and signaling events regulating the above processes were part of the i-UPR. The gene apm-1 (associated protein complex medium chain-1) encodes the μ1 medium chain of the AP-1 clathrin-associated protein complex located at the trans-Golgi complex [36]. Loss of either ire-1 or xbp-1 reduced induction of apm-1, suggesting that the UPR can also regulate Golgi protein trafficking distal to the ER. Interestingly, apm-1 RNAi animals arrested at the L1 larval stage with an abnormal intestine [36], a phenotype similar to that observed in ire-1; pek-1 worms, indicating the importance of the UPR in maintaining organ integrity and function to ensure proper development.
Figure 4 Complex UPR Transcriptional Regulation of Genes with Known Functions in C. elegans
(A) The i-UPR pathway. Many conditions such as exogenous drug treatment, nutrient deprivation, viral infection, or protein overexpression block or overwhelm protein-folding reactions in the ER and result in ER stress. In this study, we used tunicamycin to block protein folding so as to activate the UPR. Following ER stress, ire-1 and xbp-1 act in a linear pathway that dominates the transcriptional response (total of 139 target genes with known functions), inducing genes that reshape the secretory pathway, adjust the metabolic profile, up-regulate functions involved in calcium homeostasis and anti-oxidative stress, and regulate other genes that might affect cell fate. Interestingly, ten genes require either ire-1 or xbp-1, but not both, for their induction upon ER stress. About 21 genes require only pek-1 for maximal induction, and eight genes regulated by ire-1/xbp-1 also share regulation by pek-1. Finally, atf-6 does not play a significant role in the i-UPR pathway (depicted by broken arrow). In addition to two genes that were also regulated by ire-1/xbp-1, the only gene that depends solely on atf-6 for its induction is cht-1, which encodes a chitinase orthologous to human chitotriosidase.
(B) The c-UPR pathway. During development, active protein synthesis and secretion require the UPR signaling molecules ire-1, xbp-1, and atf-6 to maintain the expression of c-UPR genes, defined by the fact that they are not up-regulated by tunicamycin but are dependent on ER stress transducers for expression. Among the genes with known functions, there are only 12 that overlap between the set of 45 genes regulated by ire-1 and the set of 160 genes regulated by xbp-1. In addition, atf-6 is required by nine genes that are regulated by ire-1 and 35 that are genes regulated by xbp-1. Moreover, the expression of 19 genes is solely dependent on atf-6, suggesting an important role of atf-6 in the c-UPR pathway. By contrast, pek-1 is largely dispensable for regulation of the c-UPR as only nine genes require pek-1 in addition to their requirements for ire-1 to maintain basal expression.
The second largest subset of i-UPR genes (~ 21%) were those involved in lipid, phospholipid, and sugar metabolism, suggesting that the ER couples its protein-folding status with membrane biogenesis and energy consumption and supply (Table S1; Figure 4A). In addition, the i-UPR up-regulated genes that encode functions directly involved in DNA binding and mRNA processing, transport, and translation. Finally, i-UPR genes regulated by ire-1/xbp-1 were also involved in other processes, such as cell proliferation, calcium homeostasis and intracellular signaling, ion transport, and mitochondrial function (Table S1; Figure 4A).
The Functions of UPR Genes Regulated by pek-1 and atf-6 Complement Those Regulated by ire-1/xbp-1
Although previous studies did not detect a set of genes in C. elegans regulated by pek-1, our array data show that approximately 23% (47 out of 202) of i-UPR genes require pek-1 for maximal induction (see Figure 3A; Table S2). This percentage is similar to that observed in mammals [16,17], suggesting that pek-1 plays a similar role in transcriptional regulation of the UPR. The expression of a gene (T04C10.4) homologous to mammalian ATF4 did not change upon tunicamycin treatment regardless of strain type, consistent with the observation in mammals that ATF4 expression is not regulated at the transcriptional level (Table S3–S6). i-UPR genes regulated by pek-1 were involved in various aspects of cell function including the secretory pathway, protein degradation, oxidative stress, metabolism, ion transport, gene expression, and cytoskeleton function (Figure 4A). However, the percentage of secretory pathway genes that required pek-1 was significantly lower than those that required ire-1/xbp-1. In addition, pek-1 appeared dispensable for the induction of genes directly involved in ER protein folding and ERAD. This result is consistent with our previous observation that pek-1 is not required for BiP (hsp-3 and hsp-4) induction [7]. Interestingly, 11 genes down-regulated in pek-1 mutants were up-regulated in ire-1 and xbp-1 mutants; five were up-regulated in all of the ire-1, xbp-1, and atf-6 mutants; two were up-regulated in ire-1 mutants; and one was up-regulated in both xbp-1 and atf-6 mutants (Table S2).
Intriguingly, atf-6 deficiency only affected about six i-UPR genes, none of which appeared to be directly involved in ER protein folding, secretion, or ERAD (Table S2). The negligible role of worm atf-6 in regulating i-UPR genes involved in the secretory pathway is unexpected based on the proposal that ATF6 helps to regulate protein folding in mammals [18,23]. However, it is consistent with a report showing that a reduction in ATF6α and/or ATF6β does not affect UPR induction in MEFs [24]. The only gene that required atf-6 was cht-1, which encodes a chitinase ortholog of human chitotriosidase, an enzyme belonging to the family of glycosylhydrolases that is massively expressed by lipid-laden macrophages in different lipid-storage diseases including atherosclerosis and Gaucher disease [37,38]. However, the molecular mechanism that underlies the tightly controlled expression of chitotriosidase and how chitotriosidase plays a role in accumulation of lipid material in the lysosomal apparatus is not yet known. One physiological role of human chitotriosidase is likely in innate immunity toward chitin-containing pathogens. In C. elegans, CHT-1 may play a role in embryogenesis, and may also be required for cuticle degradation during molting and degradation of chitin-containing pathogens as part of a host defense mechanism [39].
Constitutive UPR Genes Reveal Physiological Roles for ire-1, xbp-1, atf-6, and pek-1
During cell growth, differentiation, or physiological responses, there might be constant low-level stress in the ER that requires a basal UPR. Identification of genes that were differentially expressed in ire-1, xbp-1, pek-1, and atf-6 mutants independent of tunicamycin induction might identify normal physiological functions of the UPR. To detect genes that were differentially expressed in mutant worms, we analyzed 8,117 probes (genes) with an F-value-associated p-value less than 0.001 in the type analysis. Approximately 576 probes (genes) had an average expression that varied more than 2-fold in at least one of the knockout strains compared to wild-type animals (p ≤ 0.005). Approximately 228 probes (genes) were up-regulated in at least one of the mutant strains (Table S7), suggesting inhibitory roles of these UPR transducers in regulating these genes during development. In contrast, 324 genes were down-regulated in at least one of the mutant strains, suggesting a requirement for the UPR transducers to maintain their expression. Since expression of these 324 genes was tunicamycin-insensitive but was dependent on one or more of the UPR transducers, we called them constitutive UPR (c-UPR) genes (see Figure 3B). Only one gene (WB protein ID: CE20477) in this list was induced more than 2-fold by tunicamycin (Table S8). However, since all mutants displayed a similar 2-fold induction, CE20477 was classified as a c-UPR gene.
Out of 324 c-UPR genes, the expression of 72 required ire-1 and 239 required xbp-1 (Tables S8 and S9). Interestingly, there were only 13 overlapping genes that required both ire-1 and xbp-1, suggesting that ire-1 and xbp-1 have additional divergent functions that are separate from the classic i-UPR. In fact, about 816 genes were differentially expressed (>2-fold) in ire-1 and xbp-1 mutants (p ≤ 0.005; Table S10). Although C. elegans atf-6 appeared dispensable for the transcriptional regulation of i-UPR genes, it was required to maintain the expression of 26% of the c-UPR genes (84 out of 324 genes), suggesting the importance of atf-6 in normal cell processes and/or development. Finally, pek-1 regulated only nine c-UPR genes, and all of them were also dependent on ire-1 and/or xbp-1 (Table S9; Figure 4B). c-UPR genes encode proteins involved in a wide range of cellular functions including metabolism, gene expression, protein synthesis and degradation, intracellular trafficking, membrane transport, cytoskeleton function, cell cycle, apoptosis, and signal transduction.
CREBh is a Novel UPR Gene Dependent on ire-1, xbp-1, and atf-6
In the i-UPR gene list, we identified a gene—F57B10.1—encoding a bZIP transcription factor homologous to mammalian CREBh. C. elegans CREBh was up-regulated about 2.73-fold upon tunicamycin treatment in N2 worms. However, this up-regulation was abolished in ire-1 and xbp-1 mutant worms (Figure 5A). Although atf-6(RNAi) worms showed 2.6-fold induction of CREBh, both basal and stimulated levels were significantly reduced. In contrast, pek-1 was not required for CREBh expression. To confirm the requirement of ire-1/xbp-1 and atf-6 in CREBh expression, we performed real-time quantitative RT-PCR analysis (Figure 5A). Both the basal and stimulated expression levels of CREBh in ire-1, xbp-1, and atf-6 mutants were significantly reduced. While ire-1 mutants showed one-half the expression of wild-type animals, xbp-1 and atf-6 mutants showed more dramatic reductions in CREBh—about 4.2- and 10-fold, respectively. The small differences in CREBh expression between microarray and real-time PCR analysis were probably due to different normalization methods. In the microarray analysis, gene expression was normalized to the total hybridization intensity based on the assumption that the total amount of RNA per cell does not change with different conditions. In the RT-PCR analysis, gene expression was normalized to the expression of act-3, a house-keeping actin gene. Nevertheless, both sets of data show that C. elegans CREBh is a novel gene regulated by both ire-1/xbp-1 and atf-6. Both atf-6(RNAi) and atf-6(ok551) worms showed very similar CREBh expression patterns, supporting the hypothesis that ok551 is a loss-of-function allele and that atf-6 regulates CREBh expression.
Figure 5 CREBh Is a Novel UPR-Responsive Gene
(A) Microarray and quantitative RT-PCR analyses show that the expression of C. elegans (ce) CREBh requires ire-1, xbp-1, and atf-6. “Tuni” indicates tunicamycin treatment, as described in the Materials and Methods section.
(B) ER stress induced by dithiothreitol in HepG2 cells activates CREBh transcription. HepG2 cells were treated with dithiothreitol and harvested at various time points from 0 h to 8 h. The relative expression of CREBh and spliced xbp-1 transcripts (Xbp1s) was analyzed by quantitative RT-PCR and normalized to GADPH. The induction pattern of CREBh resembles that of spliced xbp-1 transcripts.
To determine whether mammalian CREBh also responds to ER stress, we analyzed CREBh expression in a human hepatoma cell line—HepG2, upon ER stress induced by dithiothreitol, which blocks protein folding by interfering with disulfide-bond formation. CREBh transcripts were induced and peaked at 6 h following dithiothreitol treatment with ~8.7-fold up-regulation, and then sharply declined at 8 h (Figure 5B). The transient induction pattern of CREBh mimicked that of spliced xbp-1, confirming that CREBh is a UPR-responsive gene in both C. elegans and mammals.
Discussion
The ire-1/xbp-1 Pathway Controls Most i-UPR Genes in Worms
In yeast, the UPR is solely dependent on Ire1p and its splicing target HAC1. By contrast, mammalian IRE1/XBP1 appears to be required primarily for the induction of genes involved in ERAD, based on the finding that the expression and induction of BiP but not EDEM is intact in ire1-deficient MEFs. We found that C. elegans ire-1/xbp-1 regulates the majority (~ 83%) of the classic i-UPR genes, including genes functioning in both protein folding and ERAD. This result suggests that worm ire-1 and xbp-1 have broader functions in the UPR than their mammalian homologs, and implies that the UPR of C. elegans strongly resembles that of yeast.
Homozygous mutations in either pek-1 or atf-6, or haplo-insufficiency for both pek-1 and atf-6 failed to complement ire-1-null mutants. By contrast, pek-1(RNAi); atf-6(RNAi) double mutants were normal. This result implies that the ire-1 pathway plays a more important role in the worm UPR and development than either pek-1 or atf-6. This notion is supported by our discovery that ire-1 and xbp-1 control the expression of the vast majority of both i-UPR and c-UPR genes. The broader role of ire-1/xbp-1 for the UPR in C. elegans compared to mammals is not surprising since worms are developmentally simpler than mammals. During evolution, mammalian genes gained more diverse, specific, as well as fine-tuned regulation. For example, both mammalian IRE1 and PERK are selectively critical for specific developmental programs and functions. Disruption of either pathway causes lethality and growth defects associated with hypoplasia or malfunctions of selective secretory organs, such as hepatocyte and B lymphocyte defects in ire1
−/− mice and pancreatic beta cell defects in perk
−/− mice [16,19,25,40–42].
The UPR Is Controlled by Complex Genetic Interactions and Is Essential for Development
The requirement for PERK and ATF6 homologs in both C. elegans and mammals indicates that they both differ from yeast, perhaps because metazoans normally experience ER stress during differentiation of highly specialized cells and tissues. Generating a complex set of differentiated tissues might require multiple ER sensors that detect and respond to a variety of demands placed on the ER during cell differentiation. This hypothesis is supported in C. elegans by the finding that double mutants defective in either ire-1/xbp-1 and pek-1 or ire-1/xbp-1 and atf-6 die specifically as L2 larvae with intestinal degeneration.
A few i-UPR and c-UPR genes were up-regulated in some UPR gene mutants but were down-regulated in others. These genes are highlighted in purple in tables. This dichotomous regulation is most apparent for i-UPR genes regulated by pek-1, as shown in Table S2. One possible explanation is that the loss of ire-1, xbp-1, or atf-6 induces ER stress, which up-regulates pek-1 signaling. Alternatively, ire-1/xbp-1 and atf-6 could inhibit the expression of pek-1-dependent genes. Currently, we cannot distinguish between these two possibilities. The synthetic lethality observed in either ire-1; pek-1 double mutants or ire-1; atf-6 mutants may result from both compensatory gene regulation among the ire-1/xbp-1, pek-1, and atf-6 pathways and/or functional complementation between the i-UPR and c-UPR pathways.
IRE-1 and XBP-1 Regulate Different Subsets of c-UPR Genes
In C. elegans, most i-UPR genes controlled by ire-1 also require xbp-1 (see Figure 4A; Table S1). This observation isn't surprising, since these genes are thought to act in a linear pathway, in which IRE-1 mediates the splicing of xbp-1 mRNA to create a potent bZIP transcription factor. There are only ten exceptions to this rule among i-UPR genes; examples include arf-1.1, which requires only xbp-1, and T12D8.5, which requires only ire-1. By contrast, almost all c-UPR genes require either ire-1 or xbp-1, but not both (Figure 4B; Tables S8 and S9). In addition, C. elegans xbp-1; pek-1 double mutants display many reflective crystal-like aggregates in their degenerating intestines, whereas these are absent in ire-1; pek-1 double mutants (see Figure 2D). This phenotypic difference supports the idea that ire-1 and xbp-1 functions don't completely overlap, especially since these differences are found in animals that were raised under non-UPR-inducing conditions.
Since xbp-1 mRNA does not appear to be spliced under physiological conditions [7], there is no reason why ire-1 and xbp-1 should act on the same target genes. What is surprising is that xbp-1 functions at all. IRE-1 could regulate other target genes as a kinase, or by controlling their splicing. However, it is unlikely that worm xbp-1 is cleaved by something other than ire-1, since spliced xbp-1 mRNA is not detected in ire-1 mutants [7]. Instead, unspliced xbp-1 mRNA must function independently of IRE-1. Although this doesn't occur in yeast, metazoan xbp-1 is not an exact homolog of yeast HAC1. Not only do they share minimal amino acid sequence homology, but unspliced HAC1 mRNA is translationally attenuated, whereas unspliced xbp-1 is not [9,43]. Furthermore, unspliced xbp-1 encodes a transcription factor that can act as a dominant-negative regulator of spliced XBP-1 [40,44].
Although ire-1 and xbp-1 regulate genes that function in similar processes, such as metabolism and gene expression, the depth and breadth of the regulation were shown to be different (see Figure 4B; Tables S8 and S9). For example, ire-1 regulated only eight genes involved in gene expression, while xbp-1 regulated 42 genes that function in many different aspects of gene regulation. Fifteen are transcription factors, and the others act in diverse processes ranging from RNA synthesis, processing, and export to RNA catabolism; from transcription to translation; and from DNA repair to gene silencing. In addition, many genes involved in protein turnover, intracellular trafficking, membrane transport, cell fate decisions, signal transduction, cytoskeletal structure, neuronal functions, etc., were dependent on xbp-1 but not on ire-1 (Table S9). In total, the expression of 816 genes depended on either ire-1 or xbp-1, but not both (Table S10).
Our findings are supported by recent evidence that murine IRE1α plays multiple roles in both the early and the late stages of B cell development, while XBP1 is only required for the terminal differentiation of B cells into plasma cells [25,40,41]. In the early stages, IRE1α regulates c-UPR genes (such as TDT [terminal deoxynucleotidyl transferase] and the recombination-activating genes RAG1 and RAG2), and is required for immunoglobulin gene rearrangement and B cell receptor formation. Interestingly, the IRE1α-dependent regulation of RAG1, RAG2, and TDT does not require either the IRE1α kinase or endoribonuclease activities [25]. During terminal B cell differentiation, the IRE1α endoribonuclease initiates splicing of XBP1 mRNA to produce the XBP1 transcription factor that is required for plasma cell differentiation and antibody production [40,41].
Furthermore, we found that ire-1, but not xbp-1, is required for basal expression of C49F5.1, which encodes S-adenosylmethionine synthetase, an enzyme that catalyzes the formation of S-adenosylmethionine, the principal biological methyl donor and precursor for the synthesis of adenosine and homocysteine [45]. In the mammalian liver, S-adenosylmethionine plays a pivotal role in the regulation of cellular proliferation, differentiation, and apoptosis, and its levels must be tightly controlled [46]. S-adenosylmethionine synthetase expression in fetal liver isolated from ire1α−/− mouse embryos was approximately 3.5-fold reduced compared to that observed in their wild-type littermates, whereas xbp1
−/− embryos showed proper expression of S-adenosylmethionine synthetase (K. Zhang and R. Kaufman, unpublished data). This result implies that S-adenosylmethionine synthetase is a novel IRE1-dependent c-UPR gene.
C. elegans ATF-6 Regulates a Set of Genes That Are Not Inducible by ER Stress
Blast search revealed that the C. elegans gene F45E6.2 is most closely related to mammalian ATF6α, with low homology to mammalian ATF6β, so we named it C. elegans atf-6. Our data show that C. elegans atf-6 does not regulate induction of i-UPR genes, but does control approximately one-quarter of c-UPR genes. Therefore, atf-6, ire-1, and xbp-1 are required for the expression of most c-UPR genes, and the pek-1 and the ire-1/xbp-1 pathways are required for the expression of most i-UPR genes (see Figure 3). An alternative model is that worm atf-6 might have evolved as a backup mechanism to the ire-1/xbp-1 pathway, since many genes regulated by atf-6 overlap with those regulated by ire-1 and xbp-1.
The worm microarray results showing that atf-6 does not regulate the i-UPR genes are consistent with the observation that atf-6(ok551) mutants and atf-6(RNAi) animals are as resistant to tunicamycin as wild-type worms (see Figure S1), whereas both ire-1 and pek-1 mutants are hypersensitive to tunicamycin [7]. However, it is possible that atf-6(RNAi) animals have residual activity because of incomplete silencing, and that this activity is sufficient to regulate gene expression, even though the atf-6 (RNAi); ire-1(v33) double mutant phenotype was lethal. In addition, we found that both the full-length and the N-terminal nuclear forms of worm atf-6 did not increase the activity of a mammalian ATF6α luciferase reporter in MEFs (data not shown). However, the DNA sequence recognition motif for the C. elegans and mammalian ATF6 homologs may have diverged. Alternatively, it is possible that during evolution mammalian ATF6 gained additional functions, such as up-regulating genes encoding proteins that facilitate protein folding in the ER.
The hypothesis that mammalian ATF6 regulates protein folding following ER stress was based on indirect evidence, including (1) the forced expression of ATF6α [23], and (2) analysis of a cell line that lacked S2P and thus failed to process ATF6 and activate BiP gene expression [18]. However, we do not believe that these data are conclusive. First, a reduction in ATF6α and/or ATFβ by RNAi produced minimal effects on UPR gene induction, as monitored by gene profiling [24]. Second, a number of transcription factors exist that are targets of S2P, such as Luman [47], Oasis [48], and CREBh [49,50]. One of these other S2P-dependent ER stress-induced transcription factors might control the induction of UPR target genes, like BiP. Our analysis of C. elegans atf-6 suggests that mammalian ATF6, identified as a serum response factor, may not be a typical UPR transducer like IRE1 and PERK. To firmly establish the function of mammalian ATF6, it will be necessary to analyze ATF6α and ATF6β knockout mice.
As the C. elegans intestine is the first organ to encounter many environmental toxins and infectious agents, it is likely that tunicamycin is more accessible to intestinal epithelial cells than to cells of other tissues, such as neurons. Expression analysis using promoter-driven green fluorescent protein (GFP) showed that both atf-6 and pek-1 are strongly expressed in the intestine, as well as neurons and muscles (Figure S3). In addition, an hsp-3 promoter–GFP fusion was expressed at a low level, but was highly inducible by tunicamycin treatment in the majority of the tissues in the worm (Figure S4). Finally, many of the i-UPR and c-UPR genes encode functions for general cell processes that are expressed ubiquitously, such as protein synthesis and degradation, intracellular trafficking, metabolism, and cell cycle. These observations suggest that the i-UPR does not selectively exist in the intestine, and likewise, the c-UPR is not selectively present in non-intestinal tissues.
Gene Targets of the UPR Regulate Diverse Functions
Our microarray studies identified most of the UPR genes reported previously, including 17 of the 26 xbp-1-dependent genes identified in a nematode cDNA array study [26] (Table S11). This cDNA study also identified a family of ten abu genes (activated in blocked UPR), that were induced to higher levels in ER-stressed xbp-1 mutant animals than in ER-stressed wild-type animals [26]. However, we found that none of the abu genes were up-regulated in either ire-1 or xbp-1 mutant worms (Table S11), perhaps because we studied L2 larvae and most of the abu genes appear to be expressed in older worms.
Our array data not only identified additional genes involved in key functions like protein folding, translocation, and ERAD, but also revealed additional cellular processes affected by the UPR. For example, a set of genes proposed to regulate calcium homeostasis is induced upon ER stress in an ire-1/xbp-1-dependent manner (Table S1; see Figure S2). This set includes SERCA (sca-1), ryanodine receptor (unc-68) [51], calumenin [52], nucleobindin [53,54], Herp [55], BAP31 [56], and Ca2+-independent phospholipase A2 (iPLA2) [57]. These findings suggest that the cell enhances its overall ability to handle calcium, a cation that is crucial to the delicate balance between cell survival and apoptosis upon ER stress. In addition, the i-UPR up-regulates genes that are directly involved in cell proliferation and differentiation such as CDK5 activator-binding protein C53 and the piwi/argonaute family protein eIF2C4.
Phosphatidylcholine (PC) is a major structural component of cell membranes and an important source for signaling molecules such as diacylglycerol [58]. The de novo synthesis of PC from choline consists of three steps: choline kinase converts choline into phosphocholine, which is then converted into CDP-choline. In the last step, choline/ethanolaninephosphotransferase catalyzes the transfer of phosphocholine from CDP-choline to diacylglycerol. We found two genes (F22F7.5 and B0285.9) encoding choline kinases that were up-regulated by tunicamycin in an ire-1/xbp-1-dependent manner (Table S1; Figure S2). In addition, ire-1, xbp-1, and atf-6 were all required to maintain expression of choline/ethanolaninephosphotransferase in either a physiological context or upon encountering acute environmental stress. Increased expression and activity of choline kinase were reported to associate with cancer, while decreased PC synthesis induced growth arrest and apoptosis [58,59]. Therefore, up-regulation or maintenance of PC synthesis by both the i-UPR and the c-UPR may provide a survival signal confronting ER stress. Recently, it was proposed that up-regulation of choline phosphotransferase may contribute to the membrane expansion that accompanies activation of the UPR during B lymphocyte differentiation into plasma cells [41,60].
CREBh Is a Novel UPR Gene
Finally, our microarray data identified CREBh as a novel UPR gene in both worms and mammals. C. elegans CREBh appears to be essential in worm development as suggested by embryonic lethality in CREBh-silenced worms [61]. CREBh belongs to the CREB/ATF family of transcription factors and its expression is restricted to the liver [49]. CREBh induction during fetal liver development is abolished in ire1α−/− mouse embryos (K. Zhang and R. Kaufman, unpublished data), suggesting an evolutionary conserved mechanism. However, it is not known whether mammalian XBP1 and/or ATF6 regulate CREBh expression. Interestingly, CREBh has a transmembrane domain that strongly resembles that of ATF6. Elucidating whether CREBh is cleaved and activated similar to ATF6 under ER stress and understanding its biological functions will broaden our view of the UPR in liver development and function.
Conclusion
In summary, transcription profiling in C. elegans revealed two aspects of the UPR: the i-UPR pathway directs cells to respond to acute environmental stress, and the c-UPR pathway is an essential component of the UPR during normal development (see Figure 4). In the i-UPR pathway, ire-1 and xbp-1 act in a linear process that dominates transcriptional regulation to reshape the secretory pathway and adjust cellular functions involved in calcium and phospholipid homeostasis, cell proliferation and death, anti-oxidative stress, metabolism, energy generation, cytoskeletal structure, and mitochondrial function. In addition, pek-1 is necessary for the maximal induction of one-quarter of i-UPR genes, but atf-6 plays little role in the classic i-UPR since few genes require atf-6 for their induction upon ER stress. By contrast, atf-6 plays a more important role than pek-1 does in the c-UPR pathway. Furthermore, ire-1 and xbp-1 regulate very different sets of c-UPR genes, so their normal physiological functions have diverged. This observation implies that alternative regulatory mechanisms can activate or transmit signals downstream of ire-1 and xbp-1. Finally, the UPR regulates genes functioning in a much broader range than those previously reported or expected from analysis of yeast and other model systems, and many of these regulatory interactions seem to be conserved in mammals. We believe that genetic and genomic profiling analysis in worms provides a missing link between the yeast and the mammalian UPRs, and sheds light on the role of the mammalian UPR in development and disease.
Materials and Methods
Strains and general methods.
C. elegans strains were cultivated at 20 °C [62]. The strain N2 (Bristol) was used as the wild-type. The atf-6(ok551) mutant was obtained from the C. elegans Gene Knockout Consortium (http://celeganskoconsortium.omrf.org/). The mutant strains ire-1(v33), pek-1(ok275), xbp-1(zc12), and mnC1 were described previously [7,9]. To construct the strain ire-1(v33)/ mnC1; atf-6(ok551)/ pek-1(ok275), we crossed ire-1(v33)/mnC1; pek-1(ok275) with atf-6(ok551). The ire-1(v33)/ +; atf-6(ok551)/ pek-1(ok275) genotypes were selected by PCR and crossed to ire-1(v33)/ mnC1; pek-1(ok275). Offspring with the genotype ire-1(v33)/ mnC1; atf-6(ok551)/ pek-1(ok275) were identified by PCR genotyping. Genotyping of ire-1(v33) and pek-1(ok275) was described previously [7].
Genotyping and characterization of the atf-6(ok551) deletion mutant.
The primers atf-6-If (5′-AATGACCAGGAAATGTGGGA-3′) and atf-6-Ir (5′-AAGTGTCAATTGGCCAGTCCCTGT-3′) were used to detect the atf-6(ok551) deletion allele. The wild-type atf-6 allele amplified a 2,980-bp fragment compared to a 1,100-bp fragment from the atf-6(ok551) allele. Homozygous atf-6(ok551) mutants were identified by PCR using primers atf-6-F4 (5′-CGGAAGAGTCATCACGTATGAAG-3′) and atf-6-R2 (5′-GGCAGAAGCACGTAGTCTTGAAG-3′) from inside the deletion region. These two reactions were performed in one PCR tube by mixing all four primers. Wild-type animals generated only one 780-bp fragment amplified by primers F4 and R2. Homozygous atf-6(ok551) mutants had only one 1,100-bp fragment amplified by If and Ir, whereas heterozygous mutants had both 1,100-bp and 780-bp fragments.
RNA interference by injection.
PCR was used to amplify cDNA fragments flanked by the T7 promoter at both the 5′ and 3′ ends. The primer pair T7-5ceATF6 (5′-GGATCCTAATACGACTCACTATAGGAGACGGCGGGAGTTTAGGAGATTC-3′) and T7-3ceATF6 (5′-GGATCCTAATACGACTCACTATAGGCTTGATTTGGCGTTGCGTAGC-3′) amplified an approximately 520-bp fragment encoding the N-terminus of ATF-6. Amplified templates were transcribed in vitro to yield dsRNA [63] for injection as described [64]. Only progeny hatched from eggs laid between 12 and 24 h postinjection were studied. Silencing of xbp-1 by RNAi was described previously [7].
RNAi-mediated silencing atf-6 genes by bacterial feeding.
The primers 5′-TGTTTTAGATCTGGCGGGAGTTTAGGAGATTC-3′ and 5′-AATATGGTACCTTGATTTGGCGTTGCG-3′ were used to amplify a 430-bp cDNA fragment of atf-6, which was cloned into the BglII and KpnI sites in the L4440 feeding vector (pPD129.36). The resulting atf-6 RNAi plasmids were transformed into the HT115 (DE3)—an RNase III–deficient E. coli strain [65]. Resistance to ampicillin and tetracycline was used to select transformed cells. A small LB culture containing ampicillin (100 μg/ml) and tetracycline (12.5 μg/ml) was inoculated overnight, and diluted 1:100 into a large LB culture that contained ampicillin (50 μg/ml). When the OD at 600 nm reached 0.6, IPTG was added to a final concentration of 1 mM to allow induction for 4 h. A batch of bacteria expressing atf-6 dsRNA was centrifuged and stored at 4 °C for further use. Another batch of bacteria was seeded directly onto NMG-Lite plates containing 1 mM IPTG and 50 μg/ml ampicillin. Seeded plates were dried at room temperature and induction was continued at room temperature overnight. To prepare atf-6(RNAi) worms for microarray analysis, synchronized L4 larvae (N2 or ire-1[v33]) were fed with bacteria expressing atf-6 dsRNA in liquid culture for 1.5 d and then bleached to produce many eggs, which were then transferred to seeded atf-6 RNAi plates. After 20 h, the resulting atf-6(RNAi) L2 worms were transferred and treated with or without 30 μg/ml tunicamycin for 4 h in liquid culture before harvesting RNA. As a positive control to ensure the efficiency of atf-6 RNAi, nearly all ire-1(v33); atf-6(RNAi) worms died in 2 d on atf-6(RNAi) plates and no survivors were observed at 5 d, suggesting that the atf-6 gene was efficiently silenced.
RNA isolation and cRNA synthesis for microarray analysis.
Synchronized adult worms including N2, ire-1(v33), pek-1(ok275), and xbp-1(zc12) animals were bleached to isolate eggs, which were transferred to 150-mm NGM-Lite plates to allow growth for 20 h. All the L2 larvae were then transferred to liquid culture in the presence or absence of 30 μg/ml tunicamycin for 4 h. After treatments, worms were centrifuged and pellets frozen in liquid nitrogen for RNA preparation. Total RNA was isolated by using TriReagent from MRC (Cincinnati, Ohio, United States) and purified using the Qiagen (Valencia, California, United States) RNeasy cleanup procedure. The quality of total RNAs was assessed by Agilent (Palo Alto, California, United States) Bioanalyzer. The cRNAs were synthesized from 10 μg of total RNA by CyScribe kit from Amersham Pharmacia Biotech (Piscataway, New Jersey, United States) and fragmented in the presence of heat and Mg2+ before hybridization to the Affymetrix C. elegans Genome Genechip, which represents 22,500 transcripts from C. elegans.
Microarray analysis.
Each RNA sample used for cRNA synthesis and chip hybridization was isolated from independently prepared worm samples so that the repeats for each strain and treatment were biologically significant. Although a total of 29 chips were processed in five different batches, RNA digestion plot and density curves ensured that they were all comparable. An ANOVA model (yij = αi + βj + [αβ]ij + ɛij) was fit to the expression values computed from all 29 chips. Here y is the expression value, α is the main effect for treatment, β is the main effect for strain type, and αβ is the type-plus-treatment interaction term. In this analysis we were only interested in the interaction and type terms. In the overall model-fit data, 13,706 probes had a q-value less than 0.001. Global significance was determined by using permutation tests. On the average of permutation tests, we identified only 27 probes with a q-value equal to 0.001, suggesting the significance of our data. Significant genes in overall model-fit data were then filtered into 4,050 probes by their interaction p-value being less than 0.01. About 265 genes represented by 307 probes were up-regulated at least 2-fold in the N2 strain upon tunicamycin treatment and 202 of them failed to be induced in at least one of the mutant strains as their folds of induction were less than a half of that observed in the N2 strain.
After filtering out probes significant in the interaction term, we analyzed the type term for the remaining significant genes in overall model-fit data. There were 8,117 probes with a p-value associated with F-value less than 0.001 when we did cross-comparison between any two strain types. The fold changes were calculated by comparing the average expression of both nontreated and treated mutants to that in N2 strain samples, and the corresponding p-value for each comparison was calculated. Among approximately 8,117 significant probes, the expression of about 576 varied at least 2-fold in at least one of the mutant strains compared to the N2 strain and had a corresponding p-value less than 0.005. About 324 genes were down-regulated in the mutants and further studied. About 228 probes were up-regulated in the mutants and not discussed in the paper (Table S7).
Quantitative real-time RT-PCR analysis.
The cDNA was synthesized from 400 ng of total RNA in a 20-μl reaction and then diluted into 800 μl. To detect targeted transcripts, 9 μl of the diluted cDNA was mixed with 1 μl of primer sets (1 μM for each primer) and 10 μl of sybr green mix from Applied Biosystems (Foster City, California, United States). The reactions were run at step 1, 95 °C for 10 min, and step 2, 95 °C for 10 s followed by 60°C for 50 s for 40 cycles on a Bio-Rad (Hercules, California, United States) iCycler. The PCR product was captured at real time during amplification at 60 °C, and Ct numbers were obtained. C. elegans gene expression was normalized to the expression of act-3, a house-keeping actin [7]. The primer pair 5′-TCAGAACTCAAATGGTCTTGTCAGA-3′ and 5′-ATGACGAAGATGGTGCATTGAG-3′ was used to detect C. elegans CREBh. The expression of human genes was normalized to GADPH (Applied Biosystems). The primers 5′-CATCATCCTCCCCTCCATCA-3′ and 5′-GAACACTCGTACAGGCGCAAA-3′ were used to detect human CREBh. The primers 5′-CCGCAGCAGGTGCAGG-3′ and 5′-GAGTCAATACCGCCAGAATCCA-3′ were used to analyze the spliced human XBP1 transcripts.
Supporting Information
Figure S1
atf-6 (ok551) Mutants Are Resistant to ER Stress Induced by Tunicamycin
Gravid adults were allowed to lay eggs for 4 h on plates in column a, and then were transferred to the corresponding plates in column b. After 4 h to permit egg laying, the animals were transferred to plates in column c for another 4 h of egg laying. Eggs from each strain were laid on plates containing either 5 μg/ml tunicamycin (A) or 7.5 μg/ml tunicamycin (B), and after 3 d the numbers of adult worms appearing on plates were counted. The percentages of worms maturing to adulthood were calculated.
(10 KB PDF)
Click here for additional data file.
Figure S2 RT-PCR Verification of Microarray Data
Thirty genes from the I-UPR gene list were picked randomly and analyzed their expression by quantitative RT-PCR. All gene expression levels were normalized to that of act-3.
(72 KB PDF)
Click here for additional data file.
Figure S3 Expression Patterns of C. elegans atf-6 and pek-1
(A) pek-1-promoted GFP.
(B) atf-6-promoted GFP. The upstream 3.5-kb promoter regions of the pek-1 gene and the atf-6 gene were fused to the coding region of GFP (Fire lab GFP expression vector ppD95.75). The expression patterns were confirmed in at least five independent stable lines carrying extrachromosomal arrays of the GFP fusion constructs.
(88 KB PDF)
Click here for additional data file.
Figure S4
hsp-3-Promoted GFP Worms
The 2-kb promoter region of the hsp-3 gene was inserted upstream of GFP and introduced as an extrachromosomal array into C. elegans. GFP was analyzed after 36 h following hatching in plates lacking (top) or containing (bottom) of 10 μg/ml tunicamycin.
(17 KB PDF)
Click here for additional data file.
Table S1 i-UPR Genes Regulated by ire-1/xbp-1
This table summarizes genes that are up-regulated at least 2-fold upon tunicamycin treatment in N2 animals but are not up-regulated in ire-1 and/or xbp-1 mutants. Fold changes were obtained by comparing tunicamycin-treated samples with untreated samples. The fold change and mean expression are represented in log(2) algorithm. Significant down-regulations in mutants are highlighted in yellow (p < 0.01). Some genes are down-regulated in some mutants but up-regulated (highlighted in purple) in others. Genes marked by an asterisk are essential genes based on knockout mutations or RNAi analysis as listed on the Wormbase Web site.
(76 KB XLS)
Click here for additional data file.
Table S2 i-UPR Genes Regulated by pek-1 and atf-6
(A) and (B) summarize genes that are up-regulated at least 2-fold upon tunicamycin treatment in N2 animals but are not up-regulated in pek-1 (A) or atf-6 (B) mutants. Fold changes were obtained by comparing tunicamycin-treated samples with untreated samples. The fold change and mean expression are represented in log(2) algorithm. Significant down-regulations in mutants are highlighted in yellow (p < 0.01). Some genes are down-regulated in some mutants but up-regulated (highlighted in purple) in others. Genes marked by an asterisk are essential genes based on knockout mutations or RNAi analysis as listed on the Wormbase Web site.
(128 KB XLS)
Click here for additional data file.
Table S3 Microarray Data Analysis: Expression Values
This table contains the expression values computed from all chips being analyzed. Every chip analysis is named as an alphabetic character followed by a number. X represents xbp-1(zc12) mutant, I is ire-1(vc33), P is pek-1(ok275), A is atf-6 (RNAi) and N means wild-type N2 animals. The odd numbers (1, 3, 5) represent samples without tunicamycin treatment, and the even numbers (2, 4, 6) are samples treated with tunicamycin. IA represents ire-1(v33); atf-6(RNAi) double mutant animals and IP represents ire-1(v33); pek-1(ok275). Because ire-1(v33); atf-6(RNAi) and ire-1(v33); pek-1(ok275) double mutants were very sick, they were not included in the ANOVA. We included the data from ire-1(v33); atf-6(RNAi) and ire-1(v33); pek-1(ok275) samples for reference. The expression values and fold changes are represented in log(2) algorithm.
(12.0 MB ZIP)
Click here for additional data file.
Table S4 Microarray Data Analysis: Overall Model-Fit Data
This table contains 13,706 genes (more exactly, probes) with a q-value less than 0.001.
(3.9 MB XLS)
Click here for additional data file.
Table S5 Microarray Data Analysis: Interaction Data
This table contains 4,050 genes (probes) with an interaction p-value less than 0.01. Fold changes were calculated as treated samples versus untreated samples.
(2.2 MB XLS)
Click here for additional data file.
Table S6 Microarray Data Analysis: Type Data
This table contains significant genes (probes) in the type term (an F-value-associated p-value less than 0.001). Mean gene expression levels for each strain type (regardless of treatments) were compared and represented as fold changes. Individual p-values for each comparison were calculated. The fold change values are red if the corresponding p-values are less than 0.005.
(4.6 MB XLS)
Click here for additional data file.
Table S7 Genes Up-Regulated in Mutants
Significant genes in the type data were filtered by at least 2-fold variance from N2 wild-type samples and significant p-values less than 0.005. This table contains genes (probes) that were up-regulated in at least one of the mutants. The significant fold changes are colored red (p < 0.005).
(146 KB XLS)
Click here for additional data file.
Table S8
ire-1-Dependent c-UPR Genes
This table summarizes c-UPR genes that are down-regulated in ire-1(v33) mutants compared to N2 animals. Fold changes were calculated by comparing the mean expression of mutants including both treated and untreated animals to that of N2 animals. IvN, XvN, PvN, and AvN indicate comparisons of ire-1, xbp-1, pek-1, and atf-6, respectively, to N2. Genes significantly down-regulated are highlighted in yellow and those up-regulated are in purple (p < 0.005). Genes marked by an asterisk are essential genes based on knockout mutations or RNAi analysis as listed on the Wormbase Web site.
(41 KB XLS)
Click here for additional data file.
Table S9
xbp-1- and/or atf-6-Dependent c-UPR Genes
This table summarizes c-UPR genes that are down-regulated in either xbp-1(zc12) and/or atf-6 (RNAi) mutants compared to N2 animals. Fold changes were calculated by comparing the mean expression of mutants including both treated and untreated animals to that of N2 animals. IvN, XvN, PvN, and AvN indicate comparisons of ire-1, xbp-1, pek-1, and atf-6, respectively, to N2. Genes significantly down-regulated are highlighted in yellow and those up-regulated are in purple (p < 0.005). Genes marked by an asterisk are essential genes based on knockout mutations or RNAi analysis as listed on the Wormbase Web site.
(112 KB XLS)
Click here for additional data file.
Table S10 Genes Differentially Regulated by ire-1 and xbp-1
Significant genes in the type term were further analyzed by comparing ire-1 versus xbp-1 (IvX). Genes (probes) in this spreadsheet had at least a 2-fold change between ire-1 and xbp-1 samples and had corresponding p-values less than 0.005. The significant fold changes are colored red (p < 0.005).
(350 KB XLS)
Click here for additional data file.
Table S11 Comparisons to Previous Reported Genes Involved in the ER Stress Response
We examined the 26 genes that were reported previously by Urano et al. to be dependent on xbp-1 signaling for induction upon ER stress [26]. Consistently, 17 genes were also ire-1/xbp-1-dependent UPR genes in our array data. However, the expression of four genes either didn't change or showed a down-regulation in an ire-1/xbp-1-independent manner upon tunicamycin treatment. LEC-8 showed a less than 2-fold induction in wild-type worms and thus was excluded from our UPR gene list. In addition, the abu genes were not up-regulated in either ire-1 or xbp-1 mutants. Finally, we looked at uda-1, which was reported as an ire-1-dependent UPR gene in C. elegans by Uccelletti et al. [66]. The expression of uda-1 was slightly induced in wild-type worms and showed moderate reduction in both ire-1 and xbp-1 worms. This result is consistent with the northern blot shown by Uccelletti et al.
(29 KB XLS)
Click here for additional data file.
Accession Numbers
The Wormbase (http://www.wormbase.org/) accession numbers for the C. elegans genes and gene products discussed in this paper are atf-6 (F45E6.2), BAP31 (Y54G2A.18), Ca2+-independent phospholipase A2 (F47A4.5), calumenin (M03F4.7), CDK5 activator-binding protein C53 (Y113G7B.16), choline/ethanolaninephosphotransferase (CE05695), CREBh (F57B10.1), Derlin-1-interacting AAA ATPase p97 (C41C4.8), EDEM (C47E12.3), eIF2C4 (C04F12.1), Herp (F25D7.2), HRD1 (F55A11.3), nucleobindin (F44A6.1), Rer1 (F46C5.8), and S2P (Y56A3A.2). The UniGene (http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=unigene) accession number for mammalian CREBh is Hs.247744.
We thank the Cancer Center Microarray Core at the University of Michigan Medical School for performing the cRNA synthesis and chip hybridization; particularly, James W. MacDonald for performing the biostatistic analysis on the array data; and Kezhong Zhang and Tom Rutkowski for critical comments on the manuscript. Supported in part by National Institutes of Health grant DK42394 (RJK) and National Science Foundation grant MCB 987598 (REE).
Competing interests. The authors have declared that no competing interests exist.
Author contributions. XS, REE, and RJK conceived and designed the experiments. XS performed the experiments. KS provided technical assistance on quantitative RT-PCR. XS and RJK analyzed the data. XS, REE, and RJK wrote the paper.
Abbreviations
ANOVAanalysis of variance
bZIPbasic leucine zipper
c-UPRconstitutive unfolded protein response
eIF2αthe alpha subunit of eukaryotic translation initiation factor 2
ERendoplasmic reticulum
ERADendoplasmic reticulum–associated degradation
i-UPRinducible unfolded protein response
GFPgreen fluorescent protein
MEFmouse embryonic fibroblast
PCphosphatidylcholine
RNAiRNA interference
S1Psite-1 protease
S2Psite-2 protease
UPRunfolded protein response
==== Refs
References
Kaufman RJ 1999 Stress signaling from the lumen of the endoplasmic reticulum: Coordination of gene transcriptional and translational controls Genes Dev 13 1211 1233 10346810
Schroder M Kaufman RJ 2005 The mammalian unfolded protein response Annu Rev Biochem 74 739 789 15952902
Shen X Zhang K Kaufman RJ 2004 The unfolded protein response—A stress signaling pathway of the endoplasmic reticulum J Chem Neuroanat 28 79 92 15363493
Cox JS Chapman RE Walter P 1997 The unfolded protein response coordinates the production of endoplasmic reticulum protein and endoplasmic reticulum membrane Mol Biol Cell 8 1805 1814 9307975
Kawahara T Yanagi H Yura T Mori K 1997 Endoplasmic reticulum stress-induced mRNA splicing permits synthesis of transcription factor Hac1p/Ern4p that activates the unfolded protein response Mol Biol Cell 8 1845 1862 9348528
Travers KJ Patil CK Wodicka L Lockhart DJ Weissman JS 2000 Functional and genomic analyses reveal an essential coordination between the unfolded protein response and ER-associated degradation Cell 101 249 258 10847680
Shen X Ellis RE Lee K Liu CY Yang K 2001 Complementary signaling pathways regulate the unfolded protein response and are required for C. elegans development Cell 107 893 903 11779465
Yoshida H Matsui T Yamamoto A Okada T Mori K 2001 XBP1 mRNA is induced by ATF6 and spliced by IRE1 in response to ER stress to produce a highly active transcription factor Cell 107 881 891 11779464
Calfon M Zeng H Urano F Till JH Hubbard SR 2002 IRE1 couples endoplasmic reticulum load to secretory capacity by processing the XBP-1 mRNA Nature 415 92 96 11780124
Tirasophon W Welihinda AA Kaufman RJ 1998 A stress response pathway from the endoplasmic reticulum to the nucleus requires a novel bifunctional protein kinase/endoribonuclease (Ire1p) in mammalian cells Genes Dev 12 1812 1824 9637683
Wang XZ Harding HP Zhang Y Jolicoeur EM Kuroda M 1998 Cloning of mammalian Ire1 reveals diversity in the ER stress responses EMBO J 17 5708 5717 9755171
Yoshida H Matsui T Hosokawa N Kaufman RJ Nagata K 2003 A time-dependent phase shift in the mammalian unfolded protein response Dev Cell 4 265 271 12586069
Harding HP Zhang Y Ron D 1999 Protein translation and folding are coupled by an endoplasmic-reticulum-resident kinase Nature 397 271 274 9930704
Haze K Yoshida H Yanagi H Yura T Mori K 1999 Mammalian transcription factor ATF6 is synthesized as a transmembrane protein and activated by proteolysis in response to endoplasmic reticulum stress Mol Biol Cell 10 3787 3799 10564271
Harding HP Zhang Y Bertolotti A Zeng H Ron D 2000 Perk is essential for translational regulation and cell survival during the unfolded protein response Mol Cell 5 897 904 10882126
Scheuner D Song B McEwen E Liu C Laybutt R 2001 Translational control is required for the unfolded protein response and in vivo glucose homeostasis Mol Cell 7 1165 1176 11430820
Harding HP Zhang Y Zeng H Novoa I Lu PD 2003 An integrated stress response regulates amino acid metabolism and resistance to oxidative stress Mol Cell 11 619 633 12667446
Lee K Tirasophon W Shen X Michalak M Prywes R 2002 IRE1-mediated unconventional mRNA splicing and S2P-mediated ATF6 cleavage merge to regulate XBP1 in signaling the unfolded protein response Genes Dev 16 452 466 11850408
Harding HP Zeng H Zhang Y Jungries R Chung P 2001 Diabetes mellitus and exocrine pancreatic dysfunction in perk−/− mice reveals a role for translational control in secretory cell survival Mol Cell 7 1153 1163 11430819
Ye J Rawson RB Komuro R Chen X Dave UP 2000 ER stress induces cleavage of membrane-bound ATF6 by the same proteases that process SREBPs Mol Cell 6 1355 1364 11163209
Yoshida H Okada T Haze K Yanagi H Yura T 2001 Endoplasmic reticulum stress-induced formation of transcription factor complex ERSF including NF-Y (CBF) and activating transcription factors 6alpha and 6beta that activates the mammalian unfolded protein response Mol Cell Biol 21 1239 1248 11158310
Thuerauf DJ Morrison L Glembotski CC 2004 Opposing roles for ATF6alpha and ATF6beta in endoplasmic reticulum stress response gene induction J Biol Chem 279 21078 21084 14973138
Okada T Yoshida H Akazawa R Negishi M Mori K 2002 Distinct roles of ATF6 and PERK in transcription during the mammalian unfolded protein response Biochem J 366 585 594 12014989
Lee AH Iwakoshi NN Glimcher LH 2003 XBP-1 regulates a subset of endoplasmic reticulum resident chaperone genes in the unfolded protein response Mol Cell Biol 23 7448 7459 14559994
Zhang K Wong HN Song B Miller CN Scheuner D 2005 The unfolded protein response sensor IRE1alpha is required at 2 distinct steps in B cell lymphopoiesis J Clin Invest 115 268 281 15690081
Urano F Calfon M Yoneda T Yun C Kiraly M 2002 A survival pathway for Caenorhabditis elegans with a blocked unfolded protein response J Cell Biol 158 639 646 12186849
Shen J Chen X Hendershot L Prywes R 2002 ER stress regulation of ATF6 localization by dissociation of BiP/GRP78 binding and unmasking of Golgi localization signals Dev Cell 3 99 111 12110171
Grigorova IL Chaba R Zhong HJ Alba BM Rhodius V 2004 Fine-tuning of the Escherichia coli sigmaE envelope stress response relies on multiple mechanisms to inhibit signal-independent proteolysis of the transmembrane anti-sigma factor, RseA Genes Dev 18 2686 2697 15520285
Kerr MK Martin M Churchill GA 2000 Analysis of variance for gene expression microarray data J Comput Biol 7 819 837 11382364
Churchill GA 2004 Using ANOVA to analyze microarray data Biotechniques 37 173 175,177 15335204
Stein L Sternberg P Durbin R Thierry-Mieg J Spieth J 2001 WormBase: Network access to the genome and biology of Caenorhabditis elegans
Nucleic Acids Res 29 82 86 11125056
van Huizen R Martindale JL Gorospe M Holbrook NJ 2003 P58IPK, a novel endoplasmic reticulum stress-inducible protein and potential negative regulator of eIF2alpha signaling J Biol Chem 278 15558 15564 12601012
Yan W Frank CL Korth MJ Sopher BL Novoa I 2002 Control of PERK eIF2alpha kinase activity by the endoplasmic reticulum stress-induced molecular chaperone P58IPK Proc Natl Acad Sci U S A 99 15920 15925 12446838
Mori K Ogawa N Kawahara T Yanagi H Yura T 1998 Palindrome with spacer of one nucleotide is characteristic of the cis- acting unfolded protein response element in Saccharomyces cerevisiae
J Biol Chem 273 9912 9920 9545334
Ye Y Shibata Y Yun C Ron D Rapoport TA 2004 A membrane protein complex mediates retro-translocation from the ER lumen into the cytosol Nature 429 841 847 15215856
Shim J Sternberg PW Lee J 2000 Distinct and redundant functions of mu1 medium chains of the AP-1 clathrin-associated protein complex in the nematode Caenorhabditis elegans
Mol Biol Cell 11 2743 2756 10930467
Boot RG Renkema GH Strijland A van Zonneveld AJ Aerts JM 1995 Cloning of a cDNA encoding chitotriosidase, a human chitinase produced by macrophages J Biol Chem 270 26252 26256 7592832
Aguilera B Ghauharali-van der Vlugt K Helmond MT Out JM Donker-Koopman WE 2003 Transglycosidase activity of chitotriosidase: Improved enzymatic assay for the human macrophage chitinase J Biol Chem 278 40911 40916 12890686
Maeda I Kohara Y Yamamoto M Sugimoto A 2001 Large-scale analysis of gene function in Caenorhabditis elegans by high-throughput RNAi Curr Biol 11 171 176 11231151
Reimold AM Iwakoshi NN Manis J Vallabhajosyula P Szomolanyi-Tsuda E 2001 Plasma cell differentiation requires the transcription factor XBP-1 Nature 412 300 307 11460154
Iwakoshi NN Lee AH Vallabhajosyula P Otipoby KL Rajewsky K 2003 Plasma cell differentiation and the unfolded protein response intersect at the transcription factor XBP-1 Nat Immunol 4 321 329 12612580
Reimold AM Etkin A Clauss I Perkins A Friend DS 2000 An essential role in liver development for transcription factor XBP-1 Genes Dev 14 152 157 10652269
Ruegsegger U Leber JH Walter P 2001 Block of HAC1 mRNA translation by long-range base pairing is released by cytoplasmic splicing upon induction of the unfolded protein response Cell 107 103 114 11595189
Lee AH Iwakoshi NN Anderson KC Glimcher LH 2003 Proteasome inhibitors disrupt the unfolded protein response in myeloma cells Proc Natl Acad Sci U S A 100 9946 9951 12902539
Mato JM Corrales FJ Lu SC Avila MA 2002 S-Adenosylmethionine: A control switch that regulates liver function FASEB J 16 15 26 11772932
Martinez-Chantar ML Latasa MU Varela-Rey M Lu SC Garcia-Trevijano ER 2003 L-methionine availability regulates expression of the methionine adenosyltransferase 2A gene in human hepatocarcinoma cells: Role of S-adenosylmethionine J Biol Chem 278 19885 19890 12660248
Raggo C Rapin N Stirling J Gobeil P Smith-Windsor E 2002 Luman, the cellular counterpart of herpes simplex virus VP16, is processed by regulated intramembrane proteolysis Mol Cell Biol 22 5639 5649 12138176
Kondo S Murakami T Tatsumi K Ogata M Kanemoto S 2005 OASIS, a CREB/ATF-family member, modulates UPR signalling in astrocytes Nat Cell Biol 7 186 194 15665855
Omori Y Imai J Watanabe M Komatsu T Suzuki Y 2001 CREB-H: A novel mammalian transcription factor belonging to the CREB/ATF family and functioning via the box-B element with a liver-specific expression Nucleic Acids Res 29 2154 2162 11353085
Chin KT Zhou HJ Wong CM Lee JM Chan CP 2005 The liver-enriched transcription factor CREB-H is a growth suppressor protein underexpressed in hepatocellular carcinoma Nucleic Acids Res 33 1859 1873 15800215
Wehrens XH Lehnart SE Marks AR 2005 Intracellular calcium release and cardiac disease Annu Rev Physiol 67 69 98 15709953
Yabe D Nakamura T Kanazawa N Tashiro K Honjo T 1997 Calumenin, a Ca2+-binding protein retained in the endoplasmic reticulum with a novel carboxyl-terminal sequence, HDEF J Biol Chem 272 18232 18239 9218460
Lin P Yao Y Hofmeister R Tsien RY Farquhar MG 1999 Overexpression of CALNUC (nucleobindin) increases agonist and thapsigargin releasable Ca2+ storage in the Golgi J Cell Biol 145 279 289 10209024
Lin P Le-Niculescu H Hofmeister R McCaffery JM Jin M 1998 The mammalian calcium-binding protein, nucleobindin (CALNUC), is a Golgi resident protein J Cell Biol 141 1515 1527 9647645
Chan SL Fu W Zhang P Cheng A Lee J 2004 Herp stabilizes neuronal Ca2+ homeostasis and mitochondrial function during endoplasmic reticulum stress J Biol Chem 279 28733 28743 15102845
Breckenridge DG Stojanovic M Marcellus RC Shore GC 2003 Caspase cleavage product of BAP31 induces mitochondrial fission through endoplasmic reticulum calcium signals, enhancing cytochrome c release to the cytosol J Cell Biol 160 1115 1127 12668660
Ramanadham S Hsu FF Zhang S Jin C Bohrer A 2004 Apoptosis of insulin-secreting cells induced by endoplasmic reticulum stress is amplified by overexpression of group VIA calcium-independent phospholipase A2 (iPLA2 beta) and suppressed by inhibition of iPLA2 beta Biochemistry 43 918 930 14744135
Cui Z Houweling M 2002 Phosphatidylcholine and cell death Biochim Biophys Acta 1585 87 96 12531541
Ramirez de Molina A Rodriguez-Gonzalez A Lacal JC 2004 From Ras signalling to ChoK inhibitors: A further advance in anticancer drug design Cancer Lett 206 137 148 15013519
Sriburi R Jackowski S Mori K Brewer JW 2004 XBP1: A link between the unfolded protein response, lipid biosynthesis, and biogenesis of the endoplasmic reticulum J Cell Biol 167 35 41 15466483
Simmer F Moorman C van der Linden AM Kuijk E van den Berghe PV 2003 Genome-wide RNAi of C. elegans using the hypersensitive rrf-3 strain reveals novel gene functions PLoS Biol 1 e12. DOI: 10.1371/journal.pbio.0000012 14551910
Brenner S 1974 The genetics of Caenorhabditis elegans
Genetics 77 71 94 4366476
Chen PJ Singal A Kimble J Ellis RE 2000 A novel member of the tob family of proteins controls sexual fate in Caenorhabditis elegans germ cells Dev Biol 217 77 90 10625537
Fire A Xu S Montgomery MK Kostas SA Driver SE 1998 Potent and specific genetic interference by double-stranded RNA in Caenorhabditis elegans
Nature 391 806 811 9486653
Kamath RS Martinez-Campos M Zipperlen P Fraser AG Ahringer J 2001 Effectiveness of specific RNA-mediated interference through ingested double-stranded RNA in Caenorhabditis elegans
Genome Biol 2 RESEARCH0002 11178279
Uccelletti D O'Callaghan C Berninsone P Zemtseva I Abeijon C 2004 ire-1-dependent transcriptional up-regulation of a lumenal uridine diphosphatase from Caenorhabditis elegans J Biol Chem 279 27390 27398 15102851
|
16184190
|
PMC1231716
|
CC BY
|
2021-01-05 07:59:36
|
no
|
PLoS Genet. 2005 Sep 23; 1(3):e37
|
utf-8
|
PLoS Genet
| 2,005 |
10.1371/journal.pgen.0010037
|
oa_comm
|
==== Front
PLoS GenetPLoS GenetpgenplgeplosgenPLoS Genetics1553-73901553-7404Public Library of Science San Francisco, USA 1618419110.1371/journal.pgen.001003805-PLGE-RA-0076R2plge-01-03-08Research ArticleDevelopmentGastroenterology - HepatologyGenetics/Gene FunctionGenetics/Disease ModelsEukaryotesAnimalsVertebratesMammalsMus (Mouse)Acinar Cell Apoptosis in Serpini2-Deficient Mice Models Pancreatic Insufficiency Pancreatic Insufficiency in
Serpini2 Null Mice
Loftus Stacie K 1*Cannons Jennifer L 1Incao Arturo 1Pak Evgenia 1Chen Amy 1Zerfas Patricia M 2Bryant Mark A 2Biesecker Leslie G 1Schwartzberg Pamela L 1Pavan William J 11 Genetic Disease Research Branch, National Human Genome Research Institute, National Institutes of Health, Bethesda, Maryland, United States of America
2 Division of Veterinary Resources, Office of Research Services, National Institutes of Health, Bethesda, Maryland, United States of America
Roopenian Derry EditorThe Jackson Laboratory, United States of America*To whom correspondence should be addressed. E-mail: [email protected] 2005 23 9 2005 18 8 2005 1 3 e3815 4 2005 18 8 2005 This is an open-access article distributed under the terms of the Creative Commons Public Domain declaration which stipulates that, once placed in the public domain, this work may be freely reproduced, distributed, transmitted, modified, built upon, or otherwise used by anyone for any lawful purpose.2005This is an open-access article distributed under the terms of the Creative Commons Public Domain declaration, which stipulates that, once placed in the public domain, this work may be freely reproduced, distributed, transmitted, modified, built upon, or otherwise used by anyone for any lawful purpose.Pancreatic insufficiency (PI) when left untreated results in a state of malnutrition due to an inability to absorb nutrients. Frequently, PI is diagnosed as part of a larger clinical presentation in cystic fibrosis or Shwachman–Diamond syndrome. In this study, a mouse model for isolated exocrine PI was identified in a mouse line generated by a transgene insertion. The trait is inherited in an autosomal recessive pattern, and homozygous animals are growth retarded, have abnormal immunity, and have reduced life span. Mice with the disease locus, named pequeño (pq), exhibit progressive apoptosis of pancreatic acinar cells with severe exocrine acinar cell loss by 8 wk of age, while the islets and ductal tissue persist. The mutation in pq/pq mice results from a random transgene insertion. Molecular characterization of the transgene insertion site by fluorescent in situ hybridization and genomic deletion mapping identified an approximately 210-kb deletion on Chromosome 3, deleting two genes. One of these genes, Serpini2, encodes a protein that is a member of the serpin family of protease inhibitors. Reintroduction of only the Serpini2 gene by bacterial artificial chromosome transgenic complementation corrected the acinar cell defect as well as body weight and immune phenotypes, showing that deletion of Serpini2 causes the pequeño phenotype. Dietary supplementation of pancreatic enzymes also corrected body size, body weight, and immunodeficiency, and increased the life span of Serpini2-deficient mice, despite continued acinar cell loss. To our knowledge, this study describes the first characterized genetic animal model for isolated PI. Genetic complementation of the transgene insertion mutant demonstrates that Serpini2 deficiency directly results in the acinar cell apoptosis, malabsorption, and malnutrition observed in pq/pq mice. The rescue of growth retardation, immunodeficiency, and mortality by either Serpini2 bacterial artificial chromosome transgenic expression or by pancreatic enzyme supplementation demonstrates that these phenotypes are secondary to malnutrition in pq/pq mice.
Synopsis
Pancreatic insufficiency is defined by the inability to digest and absorb nutrients due to the loss of pancreatic enzyme function or loss of the acinar cells that produce the enzymes. In this manuscript the authors have described a mouse model of pancreatic insufficiency characterized by the specific loss of pancreatic acinar cells. This specific acinar cell loss results in mice that are unable to digest and absorb nutrients from the diet, stunting the animal's growth and giving rise to immunological anomalies. The authors have identified a serendipitous transgene insertion/deletion encompassing the mouse Serpini2 gene locus as the source of the phenotypes observed. Reintroduction of the Serpini2 gene, a member of the serpin family of serine cysteine protease inhibitors, by bacterial artificial chromosome complementation corrects the pancreatic and immunological phenotypes of the disorder, confirming Serpini2 as the responsible gene. Reintroduction of pancreatic enzymes through diet supplementation is also capable of correcting the reduction in size and weight, reduction in viability, and immunological deficiencies, indicating that these phenotypes are secondary to malnutrition alone. This work provides a new mouse model for investigation of malnutrition/malabsorption due to pancreatic insufficiency and identifies a novel function for the serpin family member Serpini2.
Citation:Loftus SK, Cannons JL, Incao A, Pak E, Chen A, et al (2005) Acinar cell apoptosis in Serpini2-deficient mice models pancreatic insufficiency. PLoS Genet 1(3): e38.
==== Body
Introduction
The ability to ectopically express novel genes in the mouse genome via introduction of transgenic sequences has proved to be a valuable tool in the study of mammalian development and disease. However, in approximately 5% of the transgenic lines produced, an insertional mutation at the genomic integration locus also occurs [1]. Often the insertion of the transgene is accompanied by a loss of surrounding genomic sequences and, dependent on the location of nearby coding and regulatory sequences, will produce an observable phenotype. We have been able to take advantage of such a transgene insertion/deletion event to identify and characterize a mouse model of pancreatic insufficiency (PI).
PI or a lack of pancreatic enzymes can lead to malabsorption due to an inability to digest and absorb nutrients. The most common genetic disorder that exhibits PI is cystic fibrosis (CF), as 95% of patients with CF have pancreatic acinar cell loss, with fatty cell replacement and interstitial fibrosis [2]. A second example of a syndrome that involves PI is the Shwachman–Diamond syndrome (SDS). Patients present with PI due to pancreatic acinar cell loss, but also have clinical manifestations of short stature, neutropenia, pancytopenia and predisposition to acute myelogenous leukemia [3].
Animal models of acute PI have been created by inducing pancreatic acinar cell apoptosis, either by administration of the synthetic cholecystokinin analog caerulein [4] or by physical blockage of the pancreas by ductal ligation [5]. However, defined genetic animal models of PI have been lacking. Only recently has a genetic model for an inbred CF mouse been described that recapitulates pancreas involvement of the disorder; however, this animal also manifests other features of CF [6].
In this paper we describe a new mouse model for exocrine PI that arose serendipitously in a line of mice carrying a transgene insertion. This insertion event resulted in a genomic deletion of the gene encoding the protease inhibitor SERPINI2/ PANCPIN. The mutant allele, termed pequeño (which is the Spanish word for “small”), is characterized by severe exocrine acinar cell loss at 8 wk of age, while islets and ductal tissue are spared. The disorder is inherited in an autosomal recessive pattern, and untreated homozygotes are malnourished, with a body weight one-third smaller than control littermates. Secondary to the malnutrition, these animals also have compromised immunity and a reduced life span. Administration of pancreatic enzyme diet supplementation is sufficient to reverse the effects of malnutrition in pq/pq animals, counteracting the growth defects, decreased viability, and immunodeficiency.
Results
Identification and Pathology of the Pequeño Mouse Line
In offspring of a p3pTVA-B line of transgenic mice that were bred to homozygosity, we observed runted mice with a reduced viability phenotype that we called pequeño (Figure 1A). Line p3pTVA-B was one of five independent transgenic mouse lines generated using the Pax3 promoter to drive expression of the avian retroviral receptor tv-a [7]. The four other lines generated did not produce the pequeño phenotype, suggesting that it was the site of transgene insertion that caused the phenotype. Analyses of 34 progeny resulting from the intercross of two obligate heterozygous mice demonstrated that 24 transgenic offspring (71%) were produced, in line with the expected Mendelian ratio. Six of the transgene-positive mice (6/34, 17.6%) were significantly smaller at weaning, suggesting that they were homozygous for a recessive mutation. Fluorescent in situ hybridization (FISH) analyses of metaphase spreads from three pequeño animals using the p3pTVA construct as a probe were performed. All three mice were homozygous for the transgene insertion on mouse Chromosome 3 (data not shown). We concluded that the pequeño phenotype is due to a homozygous mutation (pq/pq) that is linked to the transgene insertion and that the phenotype is inherited in an autosomal recessive pattern, as the heterozygotes had no apparent manifestations.
Figure 1 Acinar Cell Loss in pequeño Mice
(A) Size of pq/pq mice in relation to wild-type and heterozygous littermates at 8 wk of age.
(B) Average weight of n = 6 males for each genotype over a 32-wk period.
(C–H) Pancreas hematoxylin and eosin histological analyses were performed on 8-wk-old (C, D, and E), 3-wk-old (F), and 1-wk-old (G and H) mice. Wild-type (+,+) acinar cells (C) and heterozygous (pq/+) acinar cells (D) at 8 wk of age were morphologically normal. This is in contrast to homozygous (pq/pq) animals at 8 wk of age (E) and 3 wk of age (F), where there is a severe loss of exocrine pancreatic acinar cells while duct cells and islet cells (arrow) are normal. Acinar cells are present in pancreas samples from 1-wk-old pq/pq mice (H), but the cells are smaller and show reduced cytosolic volume when compared to wild-type littermates (G). Bar represents 25 μm.
The pq/pq mice were on average one-third the weight of wild-type littermates (Figure 1B) and produced a gray-colored stool. Pathological assessment at 8 wk of age revealed that the spleen and thymus were dramatically reduced in size and the pancreas was abnormal by histological analyses of affected animals when compared to unaffected littermates. Gross skeletal abnormalities were not visible by X-ray or skeletal preparation (data not shown). The thymus of homozygous mutants (n = 6) were about 4-fold smaller than wild-type littermates (n = 6) (2.90% versus 11.44% thymus/brain weight; p < 2.5 × 10−5). Similar results were observed for the spleen (4.05% versus 16.7% spleen/brain weight; p < 9.7 × 10−7). Histological examination of the endocrine pancreas for pq/pq mice revealed that islets and ductal tissue appeared normal. Analysis of the exocrine pancreas showed a striking loss of acinar cells (Figure 1). Histological evaluation of the pancreas of younger animals at 1 wk of age showed abnormal acinar cells that contained fewer zymogen granules (ZGs) than those of littermate controls (Figure 1G and 1H). However, by 3 wk of age pq/pq animals had only about one-half the number of the pancreatic acinar cells when compared to wild-type, despite the normal appearance of ductal and endocrine islet cells (Figure 1F). This observation indicated that extensive acinar cell loss was occurring between 1 and 3 wk of age in pq/pq animals.
Further analyses of the pancreas from 1-wk-old pq/pq animals by electron microscopy revealed a reduction in size and number of dense ZGs in the cytoplasm of the pq/pq acinar cells and a lack of dense material within the ductal network (Figure 2). In contrast, the pancreas of wild-type littermates contained many ZGs, and electron-dense material representing large quantities of enzymes is found in the pancreatic ducts (Figure 2A). Additionally, many pq/pq acinar cells were found to exhibit hallmark features of cells undergoing apoptosis. These acinar cells were smaller than adjacent cells and contained compact cytoplasmic organelles, condensed nuclear chromatin, and concentrically whirled bodies of rough endoplasmic reticulum (Figure 2B–2D). In addition, these acinar cells were observed being phagocytosed by neighboring epithelial cells while some were sloughed directly into the ductal lumina (Figure 2B–2D) [8]. As a second independent method of assessing apoptosis, pancreatic tissue was analyzed for DNA fragmentation using terminal deoxynucleotidyl transferase biotin–dUTP nick end labeling (TUNEL) staining. Pancreas sections from two animals in each group (wild-type and pq/pq) were counted for TUNEL staining in 12 independent, randomly selected fields (Figure 2E and 2F). The number of TUNEL-positive cells for pq/pq animals (mean counts of 7.2 and 4.9 cells per field) was significantly increased when compared to that of the wild-type (0.33 and 0.25 cells per field). Analyses of the four pair-wise comparisons were statistically significant (p < 0.01). Taken together, these results are consistent with apoptosis as the major cause of acinar cell loss.
Figure 2 Electron Microscopy Shows pequeño Acinar Cells Are Present but Undergoing Apoptosis at 1 Wk of Age
(A) Wild-type acinar cells have normal cellular morphology and contained abundant numbers of ZGs (seen as black circular vesicles containing electron-dense material), and electron-dense material from ZGs is present in the ductal lumen (white arrow).
(B–D) pq/pq mice have fewer and smaller ZGs and no dense material in the ductal lumen (L) when compared to wild-type animals. A significant number of acinar cells are undergoing apoptosis, as indicated by the presence of cellular phagocytosis (C and D), rough endoplasmic reticulum whirling bodies (B and C) (black arrows), and nuclear condensation and fragmentation (D) (asterisk).
(E and F) TUNEL staining of 3-wk-old pancreas tissue from wild-type (E) and pq/pq (F) mice indicates an increase in TUNEL-positive, red-fushcin-staining cells in pq/pq mice as compared to wild-type.
L, lumen; N, nucleus; RER, rough endoplasmic reticulum. For (A–D) bar represents 5 μm. For (E and F) bar represents 50 μm.
To test the hypothesis that growth failure is caused by PI, pq/pq animals were treated with a dietary supplementation of pancreatic enzymes in addition to a normal diet and assessed for correction of the malnutrition (Figure 3). Animals supplemented with a pancreatic enzyme diet were monitored for size, weight, and viability. Treated pq/pq mice gained both weight and length when compared with untreated pq/pq animals (Figure 3A and 3B). As expected, the enzyme supplement did not correct the acinar cell deficiency (Figure 3G). We conclude that the PI alone was responsible for the growth retardation in pq/pq mice (Figure 3).
Figure 3
pequeño Malnutrition Is Rescued by BAC Transgenic Expressing Serpini2 Gene or Pancreatic Enzyme Diet Supplementation
(A–C) Treatment of pq/pq mice (blue) pharmacologically with daily pancreatic enzyme diet supplement (purple) or genetically with BAC 150E1 transgene (green), rescues (A) size, (B) weight, and (C) animal viability in comparison to wild-type mice (black). For (B) weight measurements are averaged from n = 6 males from each genotype.
(D–G) Hematoxylin and eosin staining of pancreas sections from 8-wk-old mice. Wild-type mice (D) have normal pancreatic acinar cells present and normal pancreatic morphology while in (E) pq/pq mice there is a loss of acinar cells. In the BAC+; pq/pq mice (F), pancreas acinar cells are present. The acinar cell loss in pq/pq mice treated with enzyme supplementation (G) is the same as untreated pq/pq mice (A–C), although diet supplementation is effective in combating the effects of malnutrition. Bar represents 25 μm.
Identification of the Gene Responsible for the Pequeño Phenotype
To refine the location of the transgene insertion site in pequeño mice, the p3pTVA transgene was hybridized to heterozygous pq/+ chromosome spreads in combination with bacterial artificial chromosomes (BACs) spanning Chromosome 3 using multicolor FISH. BACs spanning 52,290,000–85,000,000 bp of Chromosome 3 were identified with NCBI Map Viewer (http://www.ncbi.nlm.nih.gov/mapview/) mouse build 31, and obtained from Children's Hospital Oakland Research Institute (http://www.chori.org/bacpac). This analysis showed that the BAC clones RP23-388I15 and RP23-296I12 flanked the p3pTVA transgene insertion and mapped the insertion to a 2.3-MB region (Figure 4). This region contained four genes of known function and several putative genes represented by expressed sequence tags and/or gene prediction algorithms (Figure 4B). One candidate gene in the region, Serpini2, a serpin protein family member, is predicted to be a serine-cysteine protease inhibitor. In addition, Serpini2 has been shown to be expressed at high levels in pancreas and adipose tissue [9,10], making it an excellent candidate gene for the pequeño phenotype.
Figure 4 Pequeño Phenotype Is the Result of a 210-kb Genomic Deletion
(A) FISH to interphase pq/+ chromosome spreads with a p3pTVA transgene probe (red) locates the transgene insertion on mouse Chromosome 3 between BACs RP23-338I15 (yellow) and RP23-296I12 (green).
(B) The genomic region defined by BAC FISH localization contained six known or predicted genes, one of which, Serpini2, is highly expressed in pancreas and adipose tissue. PCR primer sets A–P (Table S1) were used to characterize and define the identified ~210-kb deletion, indicated by red dashed line. PCR primer sets that correspond to known and predicted exon sequences are indicated by black vertical dashed lines. The genomic deletion encompassed the entire Serpini2 gene locus and exons 4 to 17 of the predicted gene E-62199.1/E-38943.1.
The genomic draft sequence from the mouse February 2003 freeze build at the University of California at Santa Cruz (http://genome.ucsc.edu/cgi-bin/hgGateway) was used to design primers within the Serpini2 gene and to the flanking genomic DNA (Figure 4B; Table S1). PCR analysis established that the transgene insertion caused in an approximately 210-kb genomic deletion at the locus, which included the entire Serpini2 gene and continued distally to include several 3′ exons of Ensembl-predicted novel gene transcript sequences ENSMUST00000062199.1 (E-62199.1) and ENSMUST00000038943.1 (E-38943.1) (Figure 4B). The size and location of the deletion established the possibility that the pequeño phenotype could be a recessively inherited contiguous gene syndrome.
We reasoned that the restricted expression of the Serpini2 gene in pancreas and adipose tissue as well as its proposed function as a serine protease inhibitor suggested that deletion of the Serpini2 locus alone caused the pequeño phenotype. An alternative explanation was that deletion of the two Ensemble-predicted transcript sequences E-62199.1 and E-38943.1 contributed to the phenotype. Sequence analysis of the predicted exons for E-62199.1/E-38943.1 suggested they are a single gene and an ortholog of the human locus Unigene cluster Hs.376728. The human gene sequence and predicted mouse gene sequences are identical in exon–intron structure and are 75% identical at the amino acid level. The pequeño deletion includes the entire Serpini2 gene locus and exons 14 to 17 of E-62199.1/E-38943.1, leaving the first three exons intact.
BAC Transgene Complementation Cross Rescues Pequeño Acinar Cell Defect
In order to dissect the role of Serpini2 and rule out the predicted genes as candidates for pequeño, we produced a BAC transgenic line, 150E1, to perform Serpini2 complementation crosses with pequeño mice. BAC RP23-150E1 was selected using the University of California at Santa Cruz February 2003 browser because it contained the entire intact Serpini2 gene and 71 kb of the upstream promoter region, but not a functional E-62199.1/E-38943.1 gene. BAC-positive (BAC+) transgenic animals were crossed with pq/+ mice and the resulting double heterozygous BAC+; pq/+ animals were backcrossed to pq/+ animals. Animals were genotyped with primers to the p3pTVA transgene, a deletion-specific primer “77kb” (within the genomic deletion, but outside of BAC 150E1), and BAC vector end-specific primers. Transgenic BAC+; pq/pq animals were indistinguishable from BAC−; +/+ and BAC+; +/+ animals by size, weight and viability (see Figure 3A–3C). RT-PCR from pancreas RNA obtained from wild-type, BAC−; pq/pq null, and BAC+; pq/pq BAC-complemented animals was performed. The results demonstrated that Serpini2 is not expressed in BAC−; pq/pq null animals and that the BAC transgene in BAC+; pq/pq animals restored expression of Serpini2 in pancreas (data not shown). Histological analysis of BAC+; pq/pq animals revealed they had normal acinar cells, showing that Serpini2 expression alone corrected the acinar cell defect phenotype (Figure 3D–3F). We conclude that deletion of the Serpini2 gene, and not E-62199.1/E-38943.1, causes the pequeño phenotype of acinar cell loss leading to PI and subsequent growth retardation due to malabsorption and malnutrition.
Immune Cell Defects in pq/pq Mice
The pq/pq mice without pancreatic diet supplementation exhibited decreased life span (typically 2–7 mo compared to over 1 y for normal mice). Interestingly, when pq/pq mice without pancreatic diet supplementation were housed in sterile cages with acidified water, their condition improved and their life span increased compared to pq/pq animals housed in standard conditions. This suggested that the pq/pq mice had compromised immune function and infections were contributing to the morbidity (data not shown). We therefore sought to characterize the immune system of the pq/pq mice.
Analyses of lymphoid organs revealed that pq/pq mice had a reduction in thymic, splenic, and bone marrow cellularity by 21 and 28 d after birth (Figure 5). In particular, spleens from pq/pq mice at 3–4 wk of age showed a dramatic reduction in the number of lymphocytes present, with T cell numbers reduced approximately 5-fold and B cell numbers reduced more than 10-fold. Thymocyte numbers were also severely reduced (10- to 20-fold), suggesting that the decrease in mature T cells resulted from reduced thymic output. Despite the dramatic reduction in thymic size, pq/pq mice had a normal thymocyte profile, with cells progressing from the most immature double negative (CD4−8−) population to double positive CD4+8+ cells and both CD4 and CD8 single-positive mature cells, with no clear block at any stage (Figure 6). In particular, the increase in the percentage of single-positive CD4+ and CD8+ cells argues that the cells exhibit no defects in selection and can progress to the most mature stages of thymic development. These results suggested that thymic development was not blocked per se, but that there was reduced cell input or viability at multiple stages.
Figure 5 Decreased Lymphocyte Numbers and Impaired B cell Development in pequeño Mice
(A) pequeño mice exhibit a striking reduction in the cellularity of the thymus, spleen and bone marrow at days 21 and 28 of life (n = 3).
(B) pequeño mice show decreased early B cell progenitors in the bone marrow. Expression profiles of B220 and IgM on total lymphoid cells in bone marrow are shown for wild-type (top row, percentage of B220+IgM− pro-/pre-B cell fraction are indicated). The B220+IgM− pro-/pre-B cell fraction (filled arrow) was gated to analyze the developmental stages using CD43 (middle row). The B220+IgM−CD43+ (less mature) fraction (open arrow) was then gated on to analyze CD24 expression (bottom row). pq/pq mice exhibit an increase in the proportion of CD43+CD24− B cells, indicating a block in the pre-/pro-B to early pro-B transition.
(C) Flow cytometric analysis of surface IgM+/IgD− expression on B220+ cells in the spleen indicates a decreased percentage of immature (IgM+/IgD−) B cells in the pequeño mice.
Figure 6 Pancreatic Enzyme Supplementation and BAC Transgenesis Rescue Lymphocyte Cells Numbers and Developmental Defects in pequeño Mice
(A) pequeño mice with either enzyme supplementation (pq/pq; diet) or a BAC transgenic (pq/pq; BAC) display increased cellularity of the thymus, spleen, and bone marrow at day 28 of life (n = 4).
(B–D) Pancreatic enzyme supplementation and the BAC transgenic rescue the developmental defects in the pequeño mice. (B) CD4 and CD8 expression profiles of the thymus. (C) Expression profiles of B220 and IgM on total lymphoid cells in bone marrow (top row). The B220+IgM− pro-/pre-B cell fraction (filled arrow) was gated to analyze the developmental stages using CD43 (middle row). The B220+IgM−CD43+ (less mature) fraction (open arrow) was then gated on to analyze CD24 expression (bottom row). (D) Splenic expression profile analysis of surface IgM/IgD expression on B220+ cells .
We also observed a defect in B cell development. B lymphocytes develop in the bone marrow through distinct stages characterized by differential expression of cell surface markers and the ordered rearrangement of Ig heavy and light chain gene segments [11,12]. Wild-type mice have populations of B220+sIgM− pro- and pre-B cells in their bone marrow that upregulate and express increasing levels of sIgM on their cell surfaces as they develop into immature and mature B cells. In contrast, pq/pq mice had fewer pro- and pre-B cells (B220+IgM−), and expression of IgM on immature cells was lower, particularly at day 28 (see Figure 5B). Within the B220+sIgM− fractions, the most immature pro-B cell fractions, fractions A–C, are further defined by expression of CD43 (B220+IgM−CD43+) and differing expression of CD24 [13]. Fraction A is CD24− and contains cells that first appear to be committed to the B cell lineage (pre-pro-B cells). Fractions B (early pro-B cells) and C (late pro-B cells) now express CD24 and initiate rearrangement at the IgH locus (B220+IgM−CD43+CD24+). Compared to wild-type, pq/pq pre-B cells manifested a higher percentage of B220+IgM−CD43+ cells, indicative of a partial block in B cell development prior to the pre-B cell stage. This block was more apparent as the mice aged (compare day 28 to day 21, Figure 5B). Moreover, within the CD43+ fraction, pq/pq pro-B cells expressed lower levels of CD24, demonstrating that this block occurred between the pre-proB and early pro-B cell stage. Similarly, in the spleen, reduced numbers of the most immature population of B cells (IgM+IgD−) were observed (Figure 5C), arguing that there was reduced production of B cells in the bone marrow.
Pancreatic Diet Supplementation and BAC Transgene Rescues Immune System Defects
To determine whether the PI and subsequent malnourished state alone were contributing to the impaired immune system or whether there was an intrinsic defect in pq/pq lymphocytes due to the lack of SERPINI2, we sought to determine how the pq/pq mice would respond to dietetic supplementation with pancreatic enzymes in contrast to BAC+; pq/pq and pq/pq untreated animals. The pancreatic enzyme replacement therapy restored immune cell numbers in the pq/pq mice, suggesting immune system defects were secondary to malnutrition. Similarly, BAC+; pq/pq transgenic mice were also able to complement immune system defects. At 4 wk of age, enzyme supplementation resulted in a partial rescue, while animals with the BAC transgene were indistinguishable from wild-type animals with respect to cellularity of the spleen, thymus, and bone marrow and the numbers of mature T and B cells within the spleen (see Figure 6A). Moreover, both the diet supplementation and the BAC transgene insertion complemented bone marrow differentiation defects. pq/pq mice with either pancreatic enzyme diet supplementation or the BAC transgene had an increased B220+IgM− fraction and normalized development of IgM+ immature B cells and the percentage of early progenitor B cells, as defined by CD43 and CD24, when compared to pq/pq animals without either enzyme supplementation or BAC transgene introduction (Figure 6C). Splenic B cell profiles were also normalized, with increased percentages of early newly emigrated B cells (Figure 6D) in enzyme-treated and BAC transgenic pq/pq animals. Thus, the immune defects in pq/pq mice appear to be the in large part secondary to pancreatic exocrine deficiency, which leads to malnourishment caused by malabsorption.
Discussion
We have identified a mutant with PI caused by a deletion of the protease inhibitor Serpini2. This phenotype (pequeño) is inherited in an autosomal recessive pattern and is to our knowledge the first published mouse mutant to genetically model PI. The mutation, caused by a transgene insertion, led to an approximately 210-kb genomic deletion that completely removes the Serpini2 gene and a portion of the predicted gene E-62199.1/E-38943.1. Reintroduction of the Serpini2 gene alone by BAC complementation corrected the acinar cell defect and improved immune function, demonstrating that deletion of the Serpini2 gene is the primary genetic cause of the pequeño mouse phenotype.
Serpini2 was initially identified and implicated as a potential tumor suppressor gene because it was downregulated in pancreatic cancer cell lines when compared to normal pancreas [9]. Serpini2 is also known as MEPI, a gene expressed in normal breast myoepithelial cells, but not in malignant breast carcinoma cells [14]. We did not observe any overt phenotype in the pequeño mice with regard to mammary gland development/function. Female pq/pq mice that had been maintained on a pancreatic enzyme diet supplementation were able to rear litters of 8–10 pups with no difficulty. It is possible that loss of expression in mammary tissue does not affect mice mammary gland function. Alternatively, there is the potential that another serpin family member may be compensating for the loss of Serpini2 activity, as has previously been observed for loss of Serpinb6 function resulting in upregulation of Serpinb1 expression [15].
SERPINI2 is a member of the serpin superfamily of proteins that have been implicated in a variety of functions including blood coagulation, angiogenesis, inflammation, and programmed cell death [16]. In humans, 33 serpins have been identified [17]. These proteins are further subdivided into two classes, inhibitory and non-inhibitory. The former group, which includes SERPINI2, is characterized by conformational structure changes that allow for formation of covalent acyl-enzyme intermediates with the proteins' target protease, and thus have been termed “suicide substrates” [18]. The target substrates for SERPINI2 are unknown.
Further insight into the cellular location and function of SERPINI2 has come from work with the rat homolog, ZG-46p. SERPINI2/ZG-46p has been characterized in the pancreas, located predominately in the Golgi and in ZGs [19], and was isolated from ZG membranes [20]. ZGs in the acinar cells contain inactive proenzyme precursors for digestive enzymes including trypsin, amylase, chymotrypsin, and carboxypeptidase. Normal acinar cells produce large quantities of granules that are transported from the trans-Golgi network to the luminal side of the acinar cell and are then secreted to the ductal lumen upon hormonal and neural signaling mechanisms. The contents of the ZGs are transported via ducts to the duodenum (pp. 429–430 of [21]), where these enzymes are required for proper digestion and absorption of food.
Given the subcellular location of SERPINI2 in the Golgi and ZGs, in addition to the predicted gene function as a serine protease inhibitor, we hypothesize that SERPINI2 may play a role in regulating the level of inactive zymogen precursors. The importance of tight regulation of active protease levels by protease inhibitors in the exocrine pancreas is well documented. This regulation is demonstrated by considering mutations in two genes that are mutated in human hereditary pancreatitis: the cationic trypsinogen gene PRSS1 and serine protease inhibitor Kazal type 1 (SPINK1). For a small number of individuals with chronic PI, mutations in either gene have been shown to alter this balance and lead to autodigestion of tissue with inflammation [22,23]. However, genetic animal models for these two genes are lacking.
It is also important to note that differences in the severity of pancreatitis may be regulated by the type of cellular response to the inflammation, whether apoptosis or necrosis. In several models of PI, acinar cell apoptosis causes a less severe pancreatitis phenotype [24–26], in contrast to the more severe pancreatitis that has been documented with acinar cell necrosis [25,27]. It is interesting to speculate that other less severe alleles than the null allele present in the pequeño mouse model may not have severe apoptosis and may demonstrate a pancreatitis phenotype similar to that observed in patients with SPINK1 mutations.
Alternatively, SERPINI2 may function to allow proper sorting and/or transport of ZGs from the Golgi to the ductal lumen of the acinar cell. This hypothesis is consistent with the observations made of the acinar cells of 1-wk-old pq/pq mice that have yet to undergo apoptosis. In these acinar cells there is a dramatic loss of ZGs present in the cell and the ZGs appear smaller in size than those in wild-type and heterozygous littermates. Further analysis is required to differentiate these potential modes of action for SERPINI2.
It is clear that malnutrition is the primary cause of the observed pequeño phenotypes. The inability of pq/pq mice to adequately digest and absorb food from a young age adversely affects growth and immune system function. A large body of evidence argues that malnutrition, stress-induced hormones, and corticosteroids can adversely affect immune cell numbers and survival [28]. In addition, leptin, a pleiotropic molecule that regulates food intake and metabolic endocrine functions, also plays a role in immune and inflammatory responses [29]. Although the defects we observed here do not completely correlate with those associated with increased steroid exposure in mice, most studies of steroid exposure have examined short-term glucocorticoid exposure [30]. However, over-expression of a glucocorticoid receptor in the thymus produced results similar to ours, with reduced cellularity observed at multiple stages of T cell development [31]. Similar long-term studies have not examined B cell numbers or function. While we cannot rule out a direct effect of SERPINI2 on lymphocytes, the improvement of immune function by pancreatic diet supplementation strongly suggests that these defects are secondary to malnutrition, which could result in increased stress hormone and steroid production, decreased leptin levels, and/or a lack of nutrients.
It is interesting that growth retardation, PI, and immunodeficiency are present in both the pequeño mouse and the human disorder SDS. The human disease is a rare and clinically heterogeneous disorder that manifests PI, short stature, lipidosis of the pancreas, neutropenia, pancytopenia, and predisposition to acute myelogenous leukemia. Although we have not observed neutropenia in the blood of pq/pq mice (Figure S1; Table S2), it may be difficult to compare secondary immune defects in mouse and human, as effects of malnutrition and stress hormones may differ between the species. Nonetheless, the combination of PI and immune defects in this mouse mutant is intriguing, given that several recent studies indicate SDS may exhibit genetic heterogeneity [32–34].
The most common mutation for SDS has been identified and is a gene conversion caused by recombination of the ubiquitously expressed SBDS gene with the pseudogene SBDSP located 5.8 Mb away on Chromosome 7 [32]. The SBDS protein is predicted to be an RNA-processing enzyme based upon sequence homology to a yeast ortholog and cluster analysis data from microarray experiments [32]. Given the overlap in phenotypes of the pequeño mice and genetic heterogeneity predicted for patients with SDS, Serpini2 is a candidate gene for analysis in the subset of patient samples in which no mutation in the SBDS gene can be identified.
Materials and Methods
Pancreatic enzyme diet.
The pancrezyme diet was provided as a supplementation to normal mouse chow Rodent NIH-31 Auto18–4 (Ziegler, Gardeners, Pennsylvania, United States). The supplement (Pancrezyme, Daniels Pharmaceuticals, St. Petersburg, Florida, United States) was mixed with Transgenic Dough diet (Bio-Serv, Frenchtown, New Jersey, United States) at a ratio of 0.9 g supplement/50 g of transgenic dough. Five grams per animal of supplemented transgenic dough mixture was provided once every 24 h.
FISH.
Metaphase spreads were performed with standard air-drying technique from blood collected from retro-orbital bleeds and mouse spleens [35]. DNA was labeled by nick translation technique, essentially as described by Pinkel et al. [36] and Lichter et al. [37] with Spectrum Orange (Vysis, Downers Grove, Illinois, United States). On each slide 200 ng of labeled probe was applied. Repeat sequences were blocked with 10× excess Cot1 DNA, 10 μl of a hybridization mixture containing the labeled DNA, 50% formamide, 2× SSC, and 10% dextran sulfate (pH 7) and were denatured at 75 °C for 10 min, then incubated at 37 °C for 30 min. Slides were incubated for at least 1 h at 37 °C in 2× SSC, then dehydrated with 2-min ethanol washes (70%, 80%, and 90%). Slide denaturation was performed in 70% formamide/2× SSC for 2 min, then dehydrated at −20 °C in 2-min ethanol washes (70%, 80%, 90%, and 100%). Post-hybridization washes were performed at 45 °C as follows: (1) 50% formamide and 2× SSC for 15 min, (2) 0.1× SSC for 10 min. Slides were counterstained with DAPI-Antifade (250 ng/ul BM, Vector, Burlingame, California, United States).
Production of BAC transgenic mice.
P3pTVA-transgenic mice were generated as described [7]. For BAC complementation analysis, fertilized oocytes obtained from matings of C57BL/6J mice were used for pronuclear injection of RP23-150E1 BAC DNA. BAC DNA was purified by CsCl gradient isolation and not linearized prior to injection. The concentrations of BAC DNA used for microinjection were 0.5–0.75 μg/ul. A total of 69 pups were born, five of which were positive for the BAC by PCR.
Histology and electron microscopy.
Tissues evaluated by hematoxylin and eosin staining and light microscopy were fixed in 10% neutral buffered formalin routinely processed, embedded into paraffin, sectioned at 5 μm thick, and stained with hematoxylin and eosin, performed by Histoserve (Rockville, Maryland, United States). Pancreas tissue evaluated by electron microscopy was fixed in 2.5% glutaraldehyde in 0.1 M cacodylate buffer (pH 7.4) overnight. The tissue was washed with cacodylate buffer and postfixed with 1% OsO4 for 2 h. Tissue was washed again with 0.1 M cacodylate buffer, serially dehydrated in ethanol and propylene oxide and embedded in Eponate 12 resin (Ted Pella, Redding, California, United States). Thin sections, approximately 80 nm, were obtained by utilizing the Leica Ultracut-UCT ultramicrotome (Leica, Deerfield, Illinois, United States) and placed onto 400 mesh copper grids and stained with saturated uranyl acetate in 50% methanol and then with lead citrate. The grids were viewed in the Philips 410 electron microscope (FEI, Hillsboro, Oregon, United States) at 80 kV and images were recorded on Kodak (Rochester, New York, United States) SO-163 film.
TUNEL staining.
Pancreas tissue for TUNEL staining was fixed in 4% paraformaldehyde, sectioned at 5 μm thick, with TUNEL staining performed by Histoserve. Visualization of TUNEL-positive cells was done using New Fushcin dye. Each of the results for the two wild-type animals (mean counts of 0.33 and 0.25 per field) was compared to each of the two results for the pq/pq animals (mean counts of 7.2 and 4.9 per field) in a Bonferroni multiple-comparison test (pp. 255–262 of [38]). Because the distributions had significantly different standard deviations, the test was repeated with the non-parametric Kruskal-Wallis test followed by Dunn's post test [38]. All four pair-wise comparisons were statistically significant at p < 0.01.
Cell preparation and flow cytometry.
Bone marrow cells were obtained by injecting PBS containing 5% FCS through surgically extracted femurs. Splenic cells and thymocytes were obtained by dissociation of the tissue with a plastic mesh and rubber stopper from a 3-ml syringe. All cells were treated with red blood cell lysis solution. For flow cytometric analysis, cells were blocked with normal mouse serum and Fc block for 10 min, then stained with FITC-, PE-, CyChrome-, or biotin-conjugated antibodies for 30 min. Biotin-conjugated antibodies were detected with a secondary stain: streptavidin-APC. Analysis was performed on a FACScan (Becton-Dickinson, Palo Alto, California, United States) using Flowjo software. The following monoclonal antibodies were obtained from BD PharMingen (San Diego, California, United States): CD3, CD4, CD8, B220, CD19, CD43, CD24, CD11b, GR-1, Ter119, DX5, and IgM. Anti-IgD was purchased from Southern Biotechnology (Birmingham, Alabama, United States).
Genotyping.
PCR reactions used 1× Genamp PCR buffer, 2.5 mM MgCl2, 0.2 mM dNTPs, 0.2 nM each primer, and 1 unit Amplitaq (Applied Biosystems, Foster City, California, United States). PCR conditions were 5 min. denaturation at 93 °C followed by 35 cycles of 93 °C for 30 sec, 58 °C for 30 sec, and 72 °C for 30 sec. BAC primers designed to flanking arm sequences were 3199R (CGCTGCAAAACGCTGACGGAACAGTAG) and 3670R (CTCAGCGTATGGTTGTCGCCGGATGTAT), and 11016F (ATGGGCAAATATTATACGCAGGCGACAAG) and 11595R (CGTATTAGCGGCCGCAAATTTATTAGAGCA). Primers within the pequeño deletion, but outside of BAC sequences were 77kbF (CTTGCCTCATCAATGCTAGG) and 77kbR (GGATAAGTCTATCTGACTTC). Primers specific to transgene insertion were P3p39 (CTGGAGCCTGTGGACTTGGAT) and TVA Rout2 (CAGTGATCAGCATCCACATGC).
For primers used to characterize the pequeño genomic deletion, see Table S1.
Supporting Information
Figure S1 Detection of Neutrophils in the Bone Marrow of Wild-Type and pq/pq Mice
(A) Flow cytometric analysis was conducted by gating on granulocytes using forward versus side light scatter and GR-1and CD11b staining in the bone marrow. Increased percentage of neutrophils in the bone marrow of pq/pq mice is due to elevated percentage of GR-1lo CD11b+ immature neutrophils.
(696 KB PDF)
Click here for additional data file.
Table S1 Primers Sets Used to Define Genomic pequeño Deletion
(63 KB DOC)
Click here for additional data file.
Table S2
pequeño Mice Show Increased Percentage of Neutrophils in Blood
(54 KB PDF)
Click here for additional data file.
Accession Numbers
The OMIM (http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=OMIM) accession numbers for the genes and gene products discussed in this article are CF (602421) and SDS (260400).
The MGI (http://www.informatics.jax.org/) accession number for Serpini2 is 1915181.
The Ensembl (http://www.ensembl.org) accession numbers for the genes discussed in this article are PRSS1 (ENSG00000173636), SPINK1 (ENSG00000164266), and E-62199.1/E-38943.1 (ENSMUSG00000058653).
The UniGene (http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=unigene) accession number of the human ortholog of the mouse gene Serpini2 discussed in this article is Hs.376728.
We gratefully acknowledge Karen Stoos and Jennifer Puck for providing the BAC genotyping primer sequences, Ana Venegas and Li Jun Yu for providing technical assistance, and Mike Cichanowski and Julia Fekecs for graphics assistance. Animal procedures were performed in accordance with protocol G94–7 from the National Human Genome Research Institute,National Institutes of Health Institutional Review Board. This research was supported by the Intramural Research program of the National Human Genome Research Institute, National Institutes of Health.
Competing interests. The authors have declared that no competing interests exist.
Author contributions. SKL, LGB, PLS, and WJP conceived and designed the experiments. SKL, JLC, AI, EP, AC, PMZ, and MAB performed the experiments. SKL, JLC, AI, EP, AC, PMZ, MAB, LGB, PLS, and WJP analyzed the data. SKL, JLC, EP, AC, PMZ, MAB, and PLS contributed reagents/materials/analysis tools. SKL, JLC, PMZ, MAB, LGB, PLS, and WJP wrote the paper.
A previous version of this article appeared as an Early Online Release on August 18, 2005 (DOI: 10.1371/journal.pgen.0010038.eor).
Abbreviations
BACbacterial artificial chromosome
CFcystic fibrosis
FISHfluorescent in situ hybridization
PIpancreatic insufficiency
pq,
pequeño;
SDSShwachman-Diamond Syndrome
TUNELterminal deoxynucleotidyl transferase biotin–dUTP nick end labeling
ZGzymogen granule
==== Refs
References
Meisler MH 1992 Insertional mutation of ‘classical' and novel genes in transgenic mice Trends Genet 8 341 344 1475845
Rowntree RK Harris A 2003 The phenotypic consequences of CFTR mutations Ann Hum Genet 67 471 485 12940920
Dror Y Freedman MH 2002 Shwachman-diamond syndrome Br J Haematol 118 701 713 12181037
Reid LE Walker NI 1999 Acinar cell apoptosis and the origin of tubular complexes in caerulein-induced pancreatitis Int J Exp Pathol 80 205 215 10583630
Hamamoto N Ashizawa N Niigaki M Kaji T Katsube T 2002 Morphological changes in the rat exocrine pancreas after pancreatic duct ligation Histol Histopathol 17 1033 1041 12371130
Durie PR Kent G Phillips MJ Ackerley CA 2004 Characteristic multiorgan pathology of cystic fibrosis in a long-living cystic fibrosis transmembrane regulator knockout murine model Am J Pathol 164 1481 1493 15039235
Hou L Loftus SK Incao A Chen A Pavan WJ 2004 Complementation of melanocyte development in SOX10 mutant neural crest using lineage-directed gene transfer Dev Dyn 229 54 62 14699577
Cummings MC Winterford CM Walker NI 1997 Apoptosis Sternberg SS Histology for pathologists, 2nd ed Philadelphia Lippincott-Raven 3 21
Ozaki K Nagata M Suzuki M Fujiwara T Miyoshi Y 1998 Isolation and characterization of a novel human pancreas-specific gene, pancpin, that is down-regulated in pancreatic cancer cells Genes Chromosomes Cancer 22 179 185 9624529
Chang WS Chang NT Lin SC Wu CW Wu FY 2000 Tissue-specific cancer-related serpin gene cluster at human chromosome band 3q26 Genes Chromosomes Cancer 29 240 255 10992299
Meffre E Casellas R Nussenzweig MC 2000 Antibody regulation of B cell development Nat Immunol 1 379 385 11062496
Melchers F ten Boekel E Seidl T Kong XC Yamagami T 2000 Repertoire selection by pre-B-cell receptors and B-cell receptors, and genetic control of B-cell development from immature to mature B cells Immunol Rev 175 33 46 10933589
Hardy RR Carmack CE Shinton SA Kemp JD Hayakawa K 1991 Resolution and characterization of pro-B and pre-pro-B cell stages in normal mouse bone marrow J Exp Med 173 1213 1225 1827140
Xiao G Liu YE Gentz R Sang QA Ni J 1999 Suppression of breast cancer growth and metastasis by a serpin myoepithelium-derived serine proteinase inhibitor expressed in the mammary myoepithelial cells Proc Natl Acad Sci U S A 96 3700 3705 10097100
Scarff KL Ung KS Nandurkar H Crack PJ Bird CH 2004 Targeted disruption of SPI3/Serpinb6 does not result in developmental or growth defects, leukocyte dysfunction, or susceptibility to stroke Mol Cell Biol 24 4075 4082 15082799
Silverman GA Bird PI Carrell RW Church FC Coughlin PB 2001 The serpins are an expanding superfamily of structurally similar but functionally diverse proteins. Evolution, mechanism of inhibition, novel functions, and a revised nomenclature J Biol Chem 276 33293 33296 11435447
van Gent D Sharp P Morgan K Kalsheker N 2003 Serpins: Structure, function and molecular evolution Int J Biochem Cell Biol 35 1536 1547 12824063
Huntington JA Carrell RW 2001 The serpins: Nature's molecular mousetraps Sci Prog 84 125 136 11525014
Chen CY Cronshagen U Kern HF 1997 A novel pancreas-specific serpin (ZG-46p) localizes to the soluble and membrane fraction of the Golgi complex and the zymogen granules of acinar cells Eur J Cell Biol 73 205 214 9243181
Schmidt K Dartsch H Linder D Kern HF Kleene R 2000 A submembranous matrix of proteoglycans on zymogen granule membranes is involved in granule formation in rat pancreatic acinar cells J Cell Sci 113 2233 2242 10825295
Banks WJ 1986 Applied veterinary histology Baltimore Williams and Wilkins 527 p.
Whitcomb DC Gorry MC Preston RA Furey W Sossenheimer MJ 1996 Hereditary pancreatitis is caused by a mutation in the cationic trypsinogen gene Nat Genet 14 141 145 8841182
Witt H Luck W Hennies HC Classen M Kage A 2000 Mutations in the gene encoding the serine protease inhibitor, Kazal type 1 are associated with chronic pancreatitis Nat Genet 25 213 216 10835640
Bhatia M Wallig MA Hofbauer B Lee HS Frossard JL 1998 Induction of apoptosis in pancreatic acinar cells reduces the severity of acute pancreatitis Biochem Biophys Res Commun 246 476 483 9610387
Kaiser AM Saluja AK Sengupta A Saluja M Steer ML 1995 Relationship between severity, necrosis, and apoptosis in five models of experimental acute pancreatitis Am J Physiol 269 C1295 C1304 7491921
Gukovskaya AS Gukovsky I Jung Y Mouria M Pandol SJ 2002 Cholecystokinin induces caspase activation and mitochondrial dysfunction in pancreatic acinar cells. Roles in cell injury processes of pancreatitis J Biol Chem 277 22595 22604 11964411
Gukovskaya AS Perkins P Zaninovic V Sandoval D Rutherford R 1996 Mechanisms of cell death after pancreatic duct obstruction in the opossum and the rat Gastroenterology 110 875 884 8608898
Vacchio MS Ashwell JD 2000 Glucocorticoids and thymocyte development Semin Immunol 12 475 485 11085180
Fantuzzi G Faggioni R 2000 Leptin in the regulation of immunity, inflammation, and hematopoiesis J Leukoc Biol 68 437 446 11037963
Laakko T Schwartz RC Fraker PJ 2002 IL-7-mediated protection of pro and pre-B cells from the adverse effects of corticosterone Cell Immunol 220 39 50 12718938
Pazirandeh A Xue Y Prestegaard T Jondal M Okret S 2002 Effects of altered glucocorticoid sensitivity in the T cell lineage on thymocyte and T cell homeostasis FASEB J 16 727 729 11923224
Boocock GR Morrison JA Popovic M Richards N Ellis L 2003 Mutations in SBDS are associated with Shwachman-Diamond syndrome Nat Genet 33 97 101 12496757
Nakashima E Mabuchi A Makita Y Masuno M Ohashi H 2004 Novel SBDS mutations caused by gene conversion in Japanese patients with Shwachman-Diamond syndrome Hum Genet 114 345 348 14749921
Woloszynek JR Rothbaum RJ Rawls AS Minx PJ Wilson RK 2004 Mutations of the SBDS gene are present in most patients with Shwachman-Diamond syndrome Blood 104 3588 3590 15284109
Lundsteen C Lind AM 1985 A test of a climate room for preparation of chromosome slides Clin Genet 28 260 262 3905090
Pinkel D Straume T Gray W 1986 Cytogenetic analysis using quantitative, high-sensitivity fluorescence hybridization Proc Natl Acad Sci U S A 83 2934 2938 3458254
Lichter P Cremer T Borden J Manuelidis L Ward DC 1988 Delineation of individual human chromosomes in metaphase and interphase cells by in situ suppression hybridization using recombinant DNA libraries Hum Genet 80 224 234 3192212
Motulsky H 1995 Intuitive Biostatistics New York Oxford University Press 386 p.
|
16184191
|
PMC1231717
|
CC0
|
2021-01-05 07:59:37
|
no
|
PLoS Genet. 2005 Sep 23; 1(3):e38
|
utf-8
|
PLoS Genet
| 2,005 |
10.1371/journal.pgen.0010038
|
oa_comm
|
==== Front
PLoS GenetPLoS GenetpgenplgeplosgenPLoS Genetics1553-73901553-7404Public Library of Science San Francisco, USA 1618955110.1371/journal.pgen.001004005-PLGE-RA-0155R2plge-01-03-10Research ArticleMolecular Biology - Structural BiologyGenetics/Chromosome BiologyDrosophilaREC, Drosophila MCM8, Drives Formation of Meiotic Crossovers Drosophila MCM8 in Meiotic Recombination
Blanton Hunter L 1Radford Sarah J 1McMahan Susan 2Kearney Hutton M 3Ibrahim Joseph G 4Sekelsky Jeff 123*1 Curriculum in Genetics and Molecular Biology, University of North Carolina, Chapel Hill, North Carolina, United States of America
2 Program in Molecular Biology and Biotechnology, University of North Carolina, Chapel Hill, North Carolina, United States of America
3 Department of Biology, University of North Carolina, Chapel Hill, North Carolina, United States of America
4 Department of Biostatistics, University of North Carolina, Chapel Hill, North Carolina, United States of America
Hawley R. Scott EditorStowers Institute for Medical Research, United States of America* To whom correspondence should be addressed. E-mail: [email protected] 2005 23 9 2005 17 8 2005 1 3 e406 7 2005 17 8 2005 Copyright: © 2005 Blanton et al.2005This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are properly credited.Crossovers ensure the accurate segregation of homologous chromosomes from one another during meiosis. Here, we describe the identity and function of the Drosophila melanogaster gene recombination defective (rec), which is required for most meiotic crossing over. We show that rec encodes a member of the mini-chromosome maintenance (MCM) protein family. Six MCM proteins (MCM2–7) are essential for DNA replication and are found in all eukaryotes. REC is the Drosophila ortholog of the recently identified seventh member of this family, MCM8. Our phylogenetic analysis reveals the existence of yet another family member, MCM9, and shows that MCM8 and MCM9 arose early in eukaryotic evolution, though one or both have been lost in multiple eukaryotic lineages. Drosophila has lost MCM9 but retained MCM8, represented by REC. We used genetic and molecular methods to study the function of REC in meiotic recombination. Epistasis experiments suggest that REC acts after the Rad51 ortholog SPN-A but before the endonuclease MEI-9. Although crossovers are reduced by 95% in rec mutants, the frequency of noncrossover gene conversion is significantly increased. Interestingly, gene conversion tracts in rec mutants are about half the length of tracts in wild-type flies. To account for these phenotypes, we propose that REC facilitates repair synthesis during meiotic recombination. In the absence of REC, synthesis does not proceed far enough to allow formation of an intermediate that can give rise to crossovers, and recombination proceeds via synthesis-dependent strand annealing to generate only noncrossover products.
Synopsis
Most of our cells have two copies of each chromosome. For sexual reproduction, these must separate from one another to produce sperm or eggs with one copy of each chromosome. This occurs during meiosis, when chromosomes pair and exchange DNA segments. This exchange— meiotic recombination—creates physical linkages between chromosome pairs and is also a source of genetic diversity. To learn more about the process of meiotic recombination, the authors characterized the gene recombination defective (rec) from the fruit fly Drosophila melanogaster. Molecular analysis revealed that rec is related to a large family of genes found in all animals, plants, and protists. These genes are thought to be important in DNA replication, but rec appears to have a novel function. The authors found that mutants lacking rec are unable to copy enough DNA during meiotic recombination to form linkages between chromosomes. This results in chromosomes segregating randomly during meiosis, so that most eggs have an incorrect number or composition of chromosomes.
Citation:Blanton HL, Radford SJ, McMahan S, Kearney HM, Ibrahim JG, et al. (2005) REC, Drosophila MCM8, drives formation of meiotic crossovers. PLoS Genet 1(3): e40.
==== Body
Introduction
Faithful segregation of homologous chromosomes in meiosis requires crossovers, which, in concert with sister chromatid cohesion, form the chiasmata that hold and orient homologs on the meiotic spindle. Crossovers are distributed nonrandomly between chromosomes, along each chromosome arm, and relative to one another, indicating that meiotic recombination is tightly regulated. One aspect of this regulation is the process that determines whether a recombination event becomes a crossover or a noncrossover.
Models of meiotic recombination must account for the production of both crossovers and noncrossovers. Current models are derived from the double-strand break (DSB) repair model of Szostak et al. [1]. In this model, recombination is initiated with a DSB on one chromatid (Figure 1A, parts a–f). Resection of the 5′ ends leaves 3′ single-stranded overhangs. One of these overhanging ends invades a homologous, non-sister duplex and primes repair DNA synthesis. The strand displaced by the migrating synthesis bubble is captured by the other 3′ overhang, which primes synthesis using the displaced strand as a template. Ligation of the newly synthesized ends to the resected 5′ ends generates an intermediate with two Holliday junctions. This double Holliday junction (DHJ) intermediate is resolved by an unknown endonuclease to form either crossover or noncrossover products.
Figure 1 Meiotic Recombination Models
(A) Model for the formation of crossovers (CO) and noncrossovers (NCO). Recombination is initiated with a DSB (a), that is resected to generate 3′ single-stranded ends (b). A 3′ end invades the homologous chromosome (c) and primes repair DNA synthesis (d). If repair synthesis proceeds far enough, second-end capture can occur (e). After repair synthesis from the second end, ligation of newly synthesized 3′ ends to resected 5′ ends results in a DHJ intermediate (f). Resolution of the DHJ can generate a CO or a NCO. In SDSA, the nascent strand dissociates after repair synthesis and anneals to the second broken end (g). Gap filling and ligation generates an NCO. Note that NCO products can come from either a DHJ intermediate or SDSA, or both.
(B) Sources of gene conversion. hDNA is first generated during strand invasion (a, dotted box); mismatch repair may replace the invading sequence with the sequence of the template (b). After synthesis (c) and dissociation, annealing of the nascent strand to the other end of the break also generates hDNA (d), which can also be acted on by mismatch repair (e). The final NCO product in this example (f) has a region of gene conversion derived from two rounds of mismatch repair, but one round is sufficient to generate a gene conversion. Also, mismatch repair can act on the DHJ intermediate (A) or the products of DHJ resolution, and thus gene conversion can also be associated with crossing over.
Recent data from yeast has resulted in modification of this model. Allers and Lichten [2] physically monitored formation of recombination intermediates and products in Saccharomyces cerevisiae using an ectopic recombination system and found that noncrossover products appear before DHJ intermediates. They proposed that noncrossovers arise not through a DHJ intermediate, but through synthesis-dependent strand annealing (SDSA). In SDSA, the nascent strand dissociates from the template and anneals to the other resected end (Figure 1A, part g). Trimming of any overhangs and filling in of any gaps, followed by ligation, results in noncrossover products. Subsequent genetic tests of this model in S. cerevisiae are consistent with most noncrossovers coming from SDSA, while the remainder are derived from a DHJ intermediate [3].
These models also take into account the occurrence of gene conversion—nonreciprocal transfer of information from one duplex to another—that can be associated with both crossovers and noncrossovers. Figure 1B illustrates possible origins of gene conversion during SDSA. Heteroduplex DNA (hDNA), in which the two strands are derived from different parental molecules, is produced by both invasion of a single-stranded overhang into a homologous template and annealing of a newly synthesized strand to the other single-stranded overhang. Sequence differences between homologous chromosomes result in base/base mismatches and insertion/deletion heterologies in hDNA, and these can be recognized and repaired by the mismatch repair system. The product contains a region of sequence derived from the homologous chromosome, referred to as a gene conversion tract. If heterologies are not repaired, each strand will convey different genetic information to the haploid product of meiosis. Upon the first round of DNA replication and mitosis after fertilization or germination, these strands separate, resulting in the post-meiotic segregation (PMS) of parental alleles. PMS results in a mosaic individual, or, for unicellular eukaryotes, a sectored colony.
Though it is more difficult to physically observe intermediates formed during meiotic recombination in Drosophila, a wealth of evidence indicates that recombination is also initiated by DSBs in this organism. MEI-W68, the Drosophila ortholog of Spo11, which catalyzes meiotic DSB formation in S. cerevisiae, is required to generate both crossovers and noncrossovers, and in mei-W68 mutants recombination is restored by treatment with ionizing radiation [4,5]. Mutations in Drosophila genes required for strand invasion cause female sterility that is suppressed by mutation of mei-W68 [6–10]. Thus, the early steps in meiotic recombination appear to be similar between Drosophila and S. cerevisiae. In contrast, later stages of crossover production are different, since most crossovers in Drosophila require the XPF/Rad1 ortholog MEI-9 [10,11], its binding partner ERCC1 [12], and several novel proteins, including MUS-312 [13] and MEI-218 [14]. In addition, it is not known whether noncrossovers in Drosophila are derived from a DHJ intermediate or SDSA, although SDSA is the most common pathway for repair of mitotic DSBs generated by transposable element excision [15–17].
In Drosophila, mutagenesis screens have been used to identify many novel genes required for meiotic recombination. The gene recombination defective (rec) was identified more than 25 years ago by Rhoda Grell in an ethyl methanesulfonate (EMS) screen for temperature-sensitive recombination mutants [18]. Her preliminary characterization of two null alleles showed that rec mutants have high levels of chromosome nondisjunction and reduced fertility, both indicative of homologous chromosome segregation defects. Since these mutants are able to pair homologous chromosomes normally, but exhibit a severe reduction in crossing over, Grell concluded that rec is involved in generating meiotic crossovers.
To gain insight into the function of the REC protein in meiotic recombination, we identified rec molecularly and found that it encodes the Drosophila ortholog of MCM8. The eukaryotic mini-chromosome maintenance (MCM) family of proteins contains six members (MCM2–7) that form a heterohexameric helicase required for replication [19]. Though MCM2–7 are essential in all eukaryotes, rec mutants are viable, and we have been unable to find any function for REC outside of meiosis. To explore the defect in meiotic recombination further, we analyzed the distribution of crossing over in rec mutants and found that residual crossovers are distributed abnormally. This finding, coupled with epistasis analysis, suggests that REC might act at an intermediate step in recombination. Further insight into the function of REC comes from our finding that the frequency of noncrossovers is substantially increased in rec mutants, and that these noncrossover events have significantly shorter gene conversion tracts than those of wild-type females. Based on these phenotypes and the structural similarity between REC and MCM proteins, we propose that REC facilitates processive repair DNA synthesis, and is a prerequisite for formation of the DHJ intermediate during meiotic recombination. In the absence of REC, recombination proceeds through SDSA to generate noncrossovers.
Results
Molecular Identification of rec
Grell's early work on the null alleles of rec found that homozygous mutant females produce progeny with high levels of chromosome nondisjunction [18]. Elevated levels of nondisjunction are indicative of, among other things, defects in homolog synapsis or meiotic recombination. Grell found that although synaptonemal complex formation was normal in rec mutant females, the frequency of crossing over among progeny recovered was 3.6% of wild-type levels. This crossover reduction indicates that the product of rec is involved in the meiotic recombination pathway and is required to generate most crossovers. To understand the function of REC in generating crossovers, we first sought to identify the gene molecularly.
Matsubayashi and Yamamoto [20] used deletion and restriction mapping to place rec in a region on the right arm of Chromosome 3 between the genes c(3)G and spn-E. Subsequently, c(3)G and spn-E were molecularly identified and the region was sequenced, revealing nine predicted genes in the region to which rec was mapped. We identified CG31293 as a candidate for rec because it was the only gene in the region that encodes a protein whose proposed function is associated with DNA metabolism. CG31293 has two exons separated by a 6-kilobase intron. We sequenced CG31293 in rec1 and rec2 mutant flies and found mutations in both (Figure 2). Sequencing CG31293 in rec1 homozygotes revealed a C→T transition at position 889, which changes a glutamine codon to a stop codon. If translated, this mutation would result in a truncation after 295 of 885 amino acids. The rec2 allele has an A→T transversion that disrupts the CG31293 splice acceptor site.
Figure 2 Mutations in rec
The rec gene contains two exons and an intron. Untranslated regions are open and protein-coding regions are filled, with the MCM core indicated by gray fill. Amino acid numbers are shown below the model, and the positions of the four mutations sequenced are shown above. The locations of the Walker A and B boxes are shown as black bars and are labeled below the model. Only the position of the intron is indicated. The full intron is 5.5 kilobases and contains another predicted gene (CG4576), transcribed in the opposite direction.
To generate additional alleles of rec, we screened the progeny of females with an EMS-mutagenized Chromosome 3 in trans to rec2 for high levels of X chromosome nondisjunction (see Materials and Methods for details). Out of 1,238 lines screened, we obtained two alleles of rec, both of which have mutations in CG31293. The rec4 allele has a C→T transition at nucleotide 2251 that generated a nonsense codon near the end of the first exon; rec4 mutants could potentially produce a protein of the same approximate size as the rec2 mutant. The rec5 allele has a C→T transition at nucleotide 142 that generated a nonsense codon predicted to truncate the protein after only 47 residues. Based on the sequence of these four alleles, we conclude that rec corresponds to CG31293 and that all four alleles are likely to be null.
During the course of this study, Matsubayashi and Yamamoto, using a different strategy also concluded that rec corresponds to CG31293 [21]. Their sequence analysis of rec1 and rec2 corresponds with our findings. They also sequenced the temperature-sensitive allele rec3 and found a C→T transition at nucleotide 1379, which causes the substitution of a nonconserved serine for a phenylalanine at residue 455.
MCM8 and MCM9 Arose Early in Eukaryotic Evolution
CG31293 encodes an 885–amino acid protein that is homologous to the MCM family of proteins found throughout eukaryotes and archaea [19]. Archaeal species each have a single MCM that forms a homohexamer believed to be the replicative helicase. Eukaryotes have six related MCM proteins, MCM2–7, which constitute a hexameric helicase involved in replication, recombination, and transcription [22]. REC is the Drosophila ortholog of MCM8, a seventh eukaryotic member of this family that was identified recently [21,23–26]. MCM8 has been reported to be present in vertebrates and Drosophila, but not in fungi or nematodes, and thus the origin of this family member is unclear.
To learn more about the MCM family of proteins, we analyzed MCMs from a number of diverse eukaryotes. On the basis of molecular and ultrastructural analysis, eukaryotes can be divided into eight major phylogenetic groups [27]. Previously, MCMs have been analyzed from only two of these groups: plants, represented by Arabidopsis thaliana, and opisthokonts, which include fungi and animals. In addition, Gozuacik et al. [26] noted the existence of an MCM8 ortholog in the discicristate Leishmania major, suggesting a broad phylogenetic distribution of this family member among eukaryotes. To expand this analysis, we identified all of the MCMs from species for which complete or nearly complete genome sequence is available, including at least one from six of the eight eukaryotic phylogenetic groups (Table S1). We conducted alignments on the predicted protein sequence of each conserved MCM domain and used these alignments to construct phylogenetic trees (see Materials and Methods for details). Eukaryotic MCM domain proteins cluster into eight subfamilies comprising the six replicative MCMs, MCM8, and a novel group that we refer to as MCM9 (Figure 2). Classification in these eight subfamilies was the same with two different algorithms for tree construction (maximum likelihood and neighbor joining), although the relationship of each group to one another and group members to each other varies with different methods of tree construction.
As expected, the six replicative MCMs are present in all eukaryotes, indicating that these six were present in the ancestral eukaryote. MCM8 and MCM9 are also widely distributed among eukaryotes, found in five of the six phylogenetic groups we examined, suggesting that the MCM8 and MCM9 subfamilies also arose early in eukaryotic evolution. Because both are absent from the excavate Giardia lamblia, which is perhaps the most deeply branching species surveyed, we cannot determine whether these family members originated after a split between excavates and other eukaryotes, or whether they were lost in the lineage giving rise to Giardia. MCM8 and MCM9 are also absent from the nematode Caenorhabditis elegans and from all available fungal genomes, with the exception of the microsporidium Encephalitozoon cuniculi. Thus, MCM8 and MCM9 do appear to have been lost multiple times during the eukaryotic radiation. Interestingly, our survey suggests that in most eukaryotic genomes MCM8 and MCM9 are either both present or both lacking. This raises the possibility that these two family members function together.
Mammalian MCM9 is unique in that it lacks the carboxy-terminal half of the conserved MCM domain, including the Walker B box (Figure S1). This is unlikely to be an error in annotation, since all mammalian MCM9 sequences available have a similar structure, and several full-length cDNA sequences encoding the truncated human protein are present in the database. Although mammalian MCM9 would lack ATPase activity on its own, it is possible that this protein retains some other function.
Three features of the phylogenetic tree in Figure 3 indicate that Drosophila REC has diverged from other MCM8 proteins. First, REC never clusters adjacent to human MCM8, despite the fact that this is the most closely related of the species shown (compare clustering of replicative MCMs). Second, Drosophila melanogaster is the only species surveyed for this figure that has MCM8 but lacks MCM9. This is true for other Drosophila species, including D. pseudoobscura and D. virilis, which diverged from D. melanogaster approximately 25 and 40 million years ago, respectively (Figure S2). However, another arthropod, the mosquito Anopheles gambiae, has retained both MCM8 and MCM9.
Figure 3 Phylogenetic Analysis of Eukaryotic MCM Family Proteins
The tree shown was generated by the neighbor-joining method of ClustalW, using an alignment of core MCM domains, excluding positions with gaps and correcting for multiple substitutions. The numbers on each node are the percentage of trees with the given branch from 10,000 independent boot-strapped iterations. Human MCM9 was excluded from this analysis because it lacks the carboxy-terminal half of the MCM domain. The approximate position of this protein was inferred from other analyses and is indicated with a dotted line. This tree was rooted with a branch containing two archaeal MCMs. The scale represents the relationship of branch length to phylogenetic distance expressed as the number of substitutions per site. Organisms are abbreviated as follows: Pfu, Pyrococcus furiosum; Sso, Sulpholobus sulfataricus; Hsa, Homo sapiens; Dme, Drosophila melanogaster; Cel, Caenorhabditis elegans; Sce, Saccharomyces cerevisiae; Ecu, Encephalitozoon cuniculi; Ath, Arabidopsis thaliana; Ehi, Entamoeba histolytica; Pfa, Plasmodium falciparum; Lma, Leishmania major; Gla, Giardia lamblia.
The third notable feature of REC is that it has a longer branch length than other MCMs, indicating that it has accumulated more substitutions than its orthologs. Alignment of MCMs from several diverse eukaryotes and several Drosophila species (Figure S2) shows some significant differences between REC and other MCMs. Replicative MCMs have consensus sequences for the Walker A and B boxes, which are involved in nucleotide binding and hydrolysis [28], and differ from those of other ATPases. The A box consensus sequence of GxxGxGKT is GDP[GS]x[SA]KS (x represents any residue; residues in brackets are alternative possibilities at a given site) in replicative MCMs. Although other MCM8 and MCM9 sequences match the replicative MCM consensus, REC has a sequence closer to the canonical ATPase consensus: GDPGIGKT. Replicative MCMs also have a highly conserved Walker B signature of IDEFDKM. Even within the deeply branching protists, the only substitutions are replacement of the initial isoleucine with a different bulky aliphatic residue (L, V, or M), replacement of the central phenyalanine with a different hydrophobic residue (leucine in several G. lamblia MCMs), or replacement of the second aspartic acid residue with glutamic acid. The corresponding sequence of REC, however, is LDDVDKL, which has five substitutions in the seven-residue sequence. However, each substitution represents a conservative change, suggesting that ATPase function may have been conserved. We also note that the DK residues in this motif are not conserved in the MCM9 subfamily. A third highly conserved motif found in MCM proteins is the arginine finger found downstream of the Walker B box. The consensus for this motif is UUSRFDUU, where U is a bulky aliphatic residue (I, L, V, or M). This motif is found in all eukaryotic and archaeal MCMs surveyed except Arabidopsis MCM9 and Drosophila REC. In REC, this sequence is LLREFHLV, and thus the absolutely conserved SR and D residues are missing.
In spite of the important sequence differences between REC and other MCMs, it is clear that REC is a member of the MCM family. When REC is used as a query in BLAST searches, all MCM proteins are returned with a high score, while other sequences do not receive significant scores. In addition, searches for conserved domains identify the MCM domain with high confidence. The divergence of Drosophila REC from other MCMs may be related to the absence of MCM9 in Drosophila, perhaps by allowing REC to assume functions of both proteins. Alternatively, REC may have acquired a novel function that does not require MCM9.
Though REC has sequence homology to replicative MCMs, rec mutants do not have phenotypes consistent with an essential role in replication. The genes encoding Drosophila MCM2–7 are essential [29–33], but rec mutants are viable and do not exhibit any of the gross developmental defects observed in the replicative MCM mutants. However, rec is expressed in somatic tissues—cDNAs have been identified in libraries made from embryos, adult head, and cell lines [34]—suggesting a function outside of meiosis. Since many of the genes required for meiotic recombination are also required for mitotic repair of DNA damage [35], we hypothesized that REC may be involved in repair of DNA damage in somatic cells. To test this hypothesis, we measured the sensitivity of rec mutants to agents that cause DSBs (gamma irradiation) and block replication (methyl methanesulfonate [36] and nitrogen mustard), following previously published methods [37,38]. We found that rec mutants are not hypersensitive to either of these agents when compared to their heterozygous siblings (unpublished data). We conclude that REC is not essential for repair of DSBs or blocked replication forks in mitotically dividing cells. Matsubayashi and Yamamoto also found that rec mutants are not hypersensitive to X-rays or methyl methanesulfonate [21].
REC Functions in an Intermediate Step of Meiotic Recombination
The only known function for REC is in generating meiotic crossovers. As a putative replicative helicase, there are many steps in recombination where REC might act to promote crossing over, including pre-meiotic S phase, resection, strand invasion, repair synthesis, and resolution. Because rec mutants have normal synaptonemal complex formation [18], it is unlikely that REC is required to complete pre-meiotic S phase. To determine where REC acts in subsequent steps of meiotic recombination, we first placed rec genetically in the Drosophila recombination pathway relative to previously characterized genes. To accomplish this, we conducted epistasis studies and analyzed the distribution of residual crossovers in rec mutants. In Drosophila, null mutations in genes whose products are required to generate meiosis-specific DSBs (mei-W68, mei-P22) abolish essentially all crossovers [4]. Because null rec mutants have residual crossovers, we conclude that REC is not involved in DSB formation like MEI-W68 and MEI-P22. After DSB formation, broken ends are resected and Rad51 homologs catalyze strand invasion. Females mutant for spn-A, which encodes the Drosophila ortholog of Rad51, are sterile and lay eggs with developmental patterning defects [9]. These phenotypes are due to an activated DNA damage-dependent cell-cycle checkpoint caused by persistent unrepaired meiotic DSBs, which disrupts communication between the oocyte and the somatic follicle cells that pattern the eggshell [10]. The proteins required for resection are not known, but we expect Drosophila resection mutants to have a phenotype similar to spn-A mutants because they would also be unable to repair meiotic DSBs. Because rec mutants are not sterile and have normal eggshell patterning, it is unlikely that REC is required for resection or strand invasion. To determine whether REC functions before or after strand invasion, we generated rec spn-A double mutants. The phenotype of the double mutant females was similar to that of spn-A single mutants, including sterility and patterning defects (unpublished data). Though this does not exclude the possibility that REC could be acting in an accessory role at either resection or strand invasion, it suggests that REC acts after strand invasion.
Further insight into the function of REC can be gained by examining the distribution of residual crossovers in rec mutants. In wild-type females, most crossovers are located in the middle of a chromosome arm and very few occur near centromeres or telomeres. In most recombination mutants that do not completely abolish crossing over, the reduction in crossing over is polar, with a stronger reduction in medial intervals and less reduction in intervals proximal to the centromere [39,40]. Null mutations of mei-218 and hypomorphic alleles of genes required for DSB formation and strand invasion have this phenotype [41]. The only known recombination mutants that do not have a polar reduction are mei-9, which encodes the Drosophila ortholog of the endonuclease XPF, and mus312, which encodes a protein required for the meiotic function of MEI-9 [13,39]. MEI-9 and MUS312 are believed to act together in the final steps of the crossover pathway, perhaps in resolution of DHJ intermediates. Resolution of Holliday junctions in Escherichia coli requires the branch migration activity of the helicase RuvB (reviewed in [42]), and in mammalian cells Holliday junction branch-migration activity co-purifies with resolvase [43]. If REC has helicase activity that is required for resolution, we would expect rec mutants to exhibit a uniform reduction in crossing over, as in mei-9 mutants.
To examine crossover distribution in rec mutants, we measured the frequency of recombination on Chromosome 2 using intervals flanked by visible markers that span the entire left arm and centric heterochromatin. We calculated the ratio between the reduction in crossover frequency in the centromere-proximal interval (pr-cn) and the reduction across the entire arm (al-cn) (Table 1). In mei-9 mutants this ratio was 1.1, showing that the reduction in crossing over is the same in the centromere-proximal interval as in other regions of the chromosome. In mei-218 mutants, however, this ratio was 4.0, due to a more modest reduction in the interval proximal to the centromere. When we analyzed the frequency and distribution of crossovers in rec mutants, we found a more modest reduction in crossing over in the centromere-proximal interval compared to the total crossover reduction (pr-cn to al-cn ratio of 3.3). Based on this analysis, we conclude that REC does not act with MEI-9 at the resolution step of meiotic recombination.
Table 1 Crossing Over on the Second Chromosome in Recombination Mutants
Because mei-9 and rec have different distributions of residual crossovers, we were able to perform epistasis analysis between these two genes. As in the rec single mutant, the reduction in crossing over is polar in the double mutant, with a pr-cn to al-cn ratio of 6.2. Although this ratio is actually higher than that of the rec single mutant, it is unclear whether this difference is functionally significant or is merely an anomaly caused by genetic background effects or relatively low sample size. (Due to the difficulty of the genetic manipulations involved in generating and assaying the double mutant, we were not able to generate as large a sample size for this genotype as for the single mutants; for example, we recovered only a single crossover within the b-pr region, resulting in an apparently more severe decrease in this region than in other regions, though recovery of one additional crossover would have made this region appear more like the al-dp region.) Regardless of the reason for the differences between the double mutant and the rec single mutant, it is clear that the double mutant exhibits a polar phenotype, confirming that REC does not function at the same time as MEI-9 but most likely acts at an intermediate step in the recombination pathway.
rec Mutants Have Increased Rates of Noncrossover Recombination but Shorter Gene Conversion Tracts
Our data place REC in an intermediate step of the recombination pathway, sometime after strand invasion. These steps include repair DNA synthesis, capture of the second resected end, and ligation to form a DHJ intermediate. Because REC is homologous to MCM replication proteins, we hypothesized that REC acts during repair DNA synthesis. Repair synthesis is necessary for formation of both crossovers and noncrossover gene conversions. If REC is essential for all repair synthesis during meiosis, then rec mutants should exhibit a reduction in the frequency of noncrossovers similar to the reduction observed in crossovers. Alternatively, if REC merely facilitates repair synthesis, then there may be no reduction in the frequency of noncrossovers in rec mutants, but gene conversion tracts should be shorter than in wild-type.
To determine the frequency of noncrossovers and the length of conversion tracts, we recovered noncrossover gene conversions. Because there are no hotspots for recombination in Drosophila, we used a system originally developed by Chovnick and colleagues to select for recombination events that occur at the rosy (ry) locus [44,45]. The ry gene encodes xanthine dehydrogenase, an enzyme that metabolizes purine and is required for normal eye pigmentation; ry mutant larvae die when the food is supplemented with purine. Females trans-heterozygous for ry531 and ry606, missense mutations separated by 3.8 kilobases, are crossed to males that are homozygous for a deletion of ry (Figure 4). The progeny are treated with purine so that only rare ry+ recombinants and mosaics that have both ry+ and ry mutant tissue (products of PMS) survive. Visible markers flanking ry (kar and cv-c) were used to distinguish crossovers from noncrossover recombination events.
Figure 4 Intragenic Recombination at rosy
Cross scheme used to recover recombination events within the ry gene. Brackets indicate that females were either trans-heterozygous for different rec alleles or were completely wild-type at rec. The three types of ry+ progeny recovered after purine selection are indicated.
We screened more than 2.3 million progeny of wild-type females and more than 0.7 million progeny of rec mutant females (Table 2). As expected, the frequency of crossing over in rec mutants was less than 10% of the wild-type frequency. In contrast, noncrossover gene conversions were recovered almost twice as frequently in rec mutants as in wild-type. Increased recovery of noncrossovers in rec mutants could reflect an actual increase in the frequency of events, or it could be caused by an increase in gene conversion tract length. An increase in tract length increases the frequency of recovery of recombination events because selecting for ry+ recombinants enriches for longer conversion tracts. Since there are no recombination hotspots in Drosophila, recombination could be initiated at any site within or near ry [46]. Therefore, the longer a conversion tract is, the greater the probability that it crosses either the ry606 or ry531 mutant site. (There is also a selection against extremely long tracts, because these might span both the ry606 and ry531 sites and convert the mutant site to wild-type and the wild-type site to mutant, producing a mutant allele of ry. However, this effect is negligible due to the relatively large distance separating the two mutant sites [47].)
Table 2 Intragenic Recombination in Wild-Type and rec Mutants
Because we did not observe a reduction in the frequency of noncrossovers, we can conclude that REC is not essential for all repair synthesis during meiosis. To test our hypothesis that REC facilitates repair synthesis, and to determine why we recovered more noncrossover gene conversions in rec mutants, we measured gene conversion tract lengths in the events recovered. We were able to map gene conversion tracts because the ry chromosomes we used are polymorphic across their entire length. We mapped 33 single nucleotide polymorphisms, or small insertion/deletion heterologies, within a 7.3-kilobase region that includes the ry531 and ry606 sites (Figure 5A and Table S2). In contrast to the case in fungi, this level of polymorphism (~0.5%) has no effect on recombination frequency in Drosophila [47].
Figure 5 Gene Conversion Tracts from Wild-Type and rec Mutants
(A) Schematic of the rosy locus. Intron/exon structure is shown, with coding sequences filled. The positions of the selected sites corresponding to the ry606 and ry531 mutations are shown. Heterologies between the ry606 and ry531 chromosomes are indicated as lollipops on the scale bar. These are all single nucleotide polymorphisms, except for −1029 and −685, which are insertions of one- and four-bp, respectively, in ry531 relative to ry606. The scale is in bp, using the coordinate system of Bender et al. [56].
(B and C) Tract lengths observed in NCOs recovered from wild-type (B) and rec mutants (C). Each bar represents an independent event, with the open circle denoting the selected marker (ry606 or ry531 mutant sites). Black bars represent the minimum tract length for each event, with co-converted sites marked by white lines. Dotted lines represent the maximum tract length possible based on the next unconverted polymorphism. The dashed line in the second ry531 conversion on panel B indicates a possible discontinuous conversion tract.
Previous work to determine the mean length of conversion tracts at ry has utilized models that take into account the selection for longer tract lengths [48,49]; we employed a similar strategy. We sequenced the region flanking the selected site of conversion (ry531 or ry606) in 29 wild-type and 18 rec noncrossover gene conversion events and determined which nonselectable polymorphisms had also been converted (co-convertants: Figure 5B and C). To calculate the mean conversion tract length for each genotype, we derived a maximum likelihood model that incorporates standard errors, allowing us to compare the mean tract lengths of wild-type and rec mutants with 95% confidence intervals (see Materials and Methods). Conversion tract lengths generated by rec mutants are shorter than those from wild-type, with mean tract lengths of 250 basepairs (bp) and 441 bp, respectively. Although the 95% confidence intervals (160–340 bp and 323–558 bp, respectively) overlap slightly, the difference in tract lengths between rec and wild-type is statistically significant (p = 0.01). The increased recovery of noncrossover gene conversions from rec mutants therefore reflects an actual increase in the frequency of events and is not the result of longer conversion tracts. Indeed, because the conversion tracts are significantly shorter in rec mutants, the frequency of events recovered actually underestimates the true increase in noncrossover gene conversions.
Gene conversion tracts are the product of hDNA formation and repair. To determine whether the shorter tract length in rec mutants is due to a defect in the repair of hDNA, we used two different methods to determine whether PMS, a readout for hDNA repair, occurs in wild-type or rec mutants. Since the ry gene product is non–cell autonomous, any ry+ recombinant could be mosaic, containing genetic information for both ry+ and ry in different cells. To test for germ-line mosaicism, recombinant progeny were mated to ry flies; germ-line mosaics produce both ry and ry+ progeny. To test for somatic mosaicism, PCR was performed using primers specific for the ry606 and ry531 alleles. Mosaicism was not detected in any of the 29 noncrossovers recovered from wild-type females or the 18 noncrossovers from rec mutants. Since mosaicism has been detected using both methods in noncrossover events from mei-9 mutants (SJR and JS, unpublished data), we conclude that PMS does not occur at an appreciable frequency in the absence of REC.
Discussion
Understanding how crossovers form is crucial to understanding the mechanisms eukaryotes use to faithfully pass half of their genetic information to the next generation. In Drosophila, many components of the meiotic recombination pathway have been identified, but a complete picture of the process has yet to emerge. In this paper, we molecularly and genetically characterized an important participant in this pathway—REC, the Drosophila homolog of MCM8—giving us new insight into requirements for crossover formation.
Our data support a model in which REC acts at an intermediate step of meiotic recombination. REC is not required for pre-meiotic S phase because homologous chromosomes in rec mutant females form normal synaptonemal complex, indicative of complete replication of genomic DNA [18]. Our finding that rec mutant females have about twice the normal number of noncrossover gene conversions indicates that initiation of recombination is not impaired in rec mutants; rather, very few DSBs are repaired as crossovers. Our data suggest that REC functions after strand invasion, since females mutant for both rec and spn-A, which encodes the Rad51 ortholog, phenocopy spn-A single mutants. Based on the distribution of residual crossovers in rec mutants and in mei-9; rec double mutants, it is likely that REC does not function with MEI-9 at resolution but acts at some previous step.
Normally, some recombination events become crossovers and some become noncrossovers. An increase in noncrossovers would occur if the crossover pathway were blocked so that most or all events followed the noncrossover pathway. In the ry intragenic recombination assay, noncrossover gene conversions are recovered only if they span a mutant site and convert that site to the wild-type sequence. In contrast, a crossover can be recovered if it occurs anywhere between the two mutations, as long as it generates a wild-type chromosome. Based on conversion tract lengths and the distance between the two mutations, we expect that many of the crossovers we recover would not be detected if they instead became noncrossovers, because they would not contain a conversion tract long enough to span a mutant site. The increase in noncrossovers that we observed in rec mutants, therefore, appears to be more than expected from this simple interpretation. A possible explanation for the increased frequency of noncrossovers in rec mutants comes from a hypothesis proposed by Bhagat et al. [41], who suggested that crossover distribution is disrupted as the result of a feedback mechanism that ensures one crossover per chromosome. The proposed feedback mechanism senses some intermediate in the crossover pathway (e.g., the DHJ structure). In mutants in which this intermediate does not form, a signal causes the cell to initiate additional recombination events to ensure that a crossover is obtained. These initiations may occur outside the normal constraints, leading to a disruption of the normal distribution and an apparent polar reduction in crossing over. According to this model, rec mutants are impaired in formation of some crucial intermediate leading to crossovers. As a result, more recombination events are initiated, but most of these still become noncrossovers. Thus, the frequency of noncrossovers is elevated, and the crossovers that are produced do not follow the normal distribution.
The defect in rec mutants is not limited to an increased production of noncrossovers at the apparent expense of crossovers. We also found that noncrossover gene conversion tract length is significantly reduced in rec mutants. This could result from a defect in generating hDNA or a defect in repairing hDNA. Defects in repair of hDNA result in PMS of markers within the heteroduplex tract. We did not detect PMS in any of the events from wild-type or rec mutant females. Thus, rec mutants are not defective in repair of hDNA; rather, formation of hDNA may be compromised.
The length of hDNA can be affected by the extent of strand invasion and the amount of repair synthesis. In S. cerevisiae, the Mer3 helicase has been shown in vitro to stimulate Rad51-mediated strand invasion [50]. As in rec mutants, mutations in the gene that encodes Mer3 cause a reduction in the frequency of crossovers and an increase in the frequency of noncrossovers [51]. However, in physical assays mer3 mutants are defective in the transition from DSB to strand invasion intermediate. Our data suggest that REC acts after strand invasion, so we do not favor the notion that REC performs a function similar to that of Mer3. Furthermore, based on the similarity of REC to replicative MCMs, we think it plausible that rec mutants have shorter conversion tracts because repair synthesis is diminished.
What is the relationship between reduced repair synthesis and decreased crossing over in rec mutants? In S. cerevisiae, crossovers are believed to arise through resolution of the DHJ intermediate. Although this process can also give rise to noncrossovers, most noncrossovers are thought to arise through SDSA [2]. There is evidence in S. cerevisiae that formation of a DHJ intermediate requires more repair synthesis than SDSA [3]. If this is also the case in Drosophila, then decreased repair synthesis would increase the probability that a meiotic DSB will be repaired through SDSA instead of the DHJ pathway.
We propose a model in which REC drives crossover formation by acting at the repair synthesis step of meiotic recombination (Figure 6). In the absence of REC, synthesis does not proceed far enough to allow second-end capture and formation of the DHJ intermediate, resulting in a deficit of crossovers. Noncrossovers may still be formed through SDSA. There are two versions of this model. First, REC may facilitate repair synthesis at all sites of recombination (Figure 6A). In this version, noncrossovers may normally arise through the DHJ pathway or the SDSA pathway, but in rec mutants the SDSA pathway is favored; the decrease in gene conversion tract length in rec mutants reflects an overall decrease in repair synthesis. Alternatively, REC may facilitate synthesis only at those recombination sites designated to become DHJ intermediates (Figure 6B). In this version of the model, sites lacking REC in wild-type flies undergo SDSA. The decrease in mean tract length in rec mutants is due to loss of those noncrossovers that would have arisen via a DHJ intermediate.
Figure 6 Models for REC Function in Meiotic Recombination
(A) REC facilitates repair synthesis at all sites of meiotic recombination. Longer synthesis tracts allow second-end capture leading to DHJ intermediate formation, which can be resolved to generate COs and NCOs. If there is less repair synthesis, the newly synthesized strand can dissociate, anneal to the other broken end, and give rise to a NCO via SDSA.
(B) REC is present only at sites that will mature into DHJ intermediates. After initial repair DNA synthesis, some single-end invasions dissociate and anneal to the broken chromosome giving rise to NCOs via SDSA. Other single-end invasion, intermediates produce more repair synthesis in a REC-dependent manner to give rise to DHJs that can be resolved to generate either a CO or a NCO.
Our data do not indicate whether noncrossovers in wild-type flies arise through SDSA, DHJ, or a combination of the two. In Drosophila, SDSA is a primary pathway for DSB repair in nonmeiotic cells [16,17]. It may be that SDSA is the “default” pathway for recombinational repair of DSBs, and that meiosis-specific modifications promote formation of DHJs to allow crossing over. REC does not appear to play a role in SDSA in nonmeiotic cells (JS and M. Adams, personal communication), and therefore REC may be a component of the meiosis-specific modifications to DSB repair in Drosophila. To better understand the role of REC and the process of meiotic recombination, it will be important to determine the source of noncrossover recombinants in wild-type females.
Materials and Methods
Genetics.
Flies were propagated at 25 °C on standard medium containing agar, cornmeal, dextrose, and brewer's yeast. Information on loci mentioned, but not described here, is available on FlyBase (http://flybase.bio.indiana.edu/search/) [34].
Screen for new rec alleles.
Two- to three-day-old y/y+Y ; kar ry506 cv-c males were starved for 1 h in an empty vial and transferred to a bottle containing 25 mM EMS in 1% sucrose. After eating EMS overnight, the males were transferred to a clean bottle for 1 d. Treated males were crossed to y ; MKRS/TM6B Hu ry females (60 females and 30 males per bottle). Single males with a balanced mutagenized Chromosome 3 were crossed to y ; ru mus312 Z2035 st rec1 Ubxbx34e/TM6B Hu ry females. Nonvirgin females that were y ; ru mus312 Z2035 st rec1 Ubxbx34e/* kar ry cv-c were placed in a new vial with brothers. The progeny were screened for non-yellow females and yellow males which result from X chromosome nondisjunction. Stocks were made of all lines that exhibited nondisjunction, and new alleles of rec were confirmed by complementation analysis. Z2035 denotes an uncharacterized meiotic recombination mutation recovered in a screen of a collection of EMS-mutagenized chromosomes (SM and JS, unpublished data).
Phylogenetic analysis of MCM domain proteins.
MCM proteins were identified by BLASTP and TBlastN searches of the genome sequence databases at http://www.ncbi.nlm.nih.gov/BLAST/. Accession numbers of each protein are listed in Table S1. Initial alignments were done with ClustalX [52], using the default settings. Regions outside of the core MCM domain were removed, as were large insertions within the core. A de novo alignment was generated with these core sequences, using the BLOSUM scoring matrices. This alignment was edited, and used to generate phylogenetic trees using the neighbor-joining method implemented in ClustalX, with 10,000 boot-strap trials, and independently using the maximum likelihood method of Tree Puzzle [53].
Crossover frequency and distribution.
Crossing over on Chromosome 2 was assayed by crossing al dp b pr Bl cn/+ females in different genetic backgrounds to al dp b pr cn/CyO males. Recombination frequencies are expressed in map units; 1 map unit is equal to a recombination frequency of 1%.
Sequencing of rec mutants.
Single homozygous mutant flies were homogenized in Squishing Buffer (10 mM Tris-HCl [pH 8.2], 1 mM EDTA, 25 mM NaCl, 1 mg/ml Proteinase K) incubated at 37 °C for 30 min and then 95 °C for 5 min. PCR was performed using gene-specific primers. Products were isolated by agarose gel electrophoresis, purified from a gel slice, and sequenced at the University of North Carolina Genome Analysis facility. Mutations were verified by sequencing the opposite strand from a independent amplification.
Intragenic recombination at the rosy locus.
Females trans-heterozygous for ry606 and ry531 were crossed to y/y+Y ; kar ry506 cv-c males (see Figure 4 for all crosses used). Crosses were set up in bottles containing 25 ml of food medium using 90 females and 30 males (rec mutants) or 30 females and 10 males (wild-type). After 3 d, the adults were transferred to new bottles to generate the second brood, and the first brood bottles were treated with 0.75 ml of 0.2% purine. This was repeated to generate a third brood. One out of every 25 bottles was not treated with purine so that the total number of flies screened could be estimated.
For every ry+ progeny recovered, the type of recombinant was determined by examining the presence of flanking markers kar (0.3 map units to the right of ry) and cv-c (2.1 map units to the left of ry). Crossover progeny are wild-type for all three markers; a noncrossover gene conversion of ry531 produces a crossveinless (cv-c) fly, and a noncrossover gene conversion of ry606 results in a karmoysin (kar)-eyed fly. Because of the proximity of these markers to one another, it is unlikely that individual ry+ progeny represent more than one recombination event in the region. Recombinant progeny were crossed to kar ry506 cv-c flies. After mating the recombinant fly, it was removed and homogenized in Squishing Buffer. To determine whether PMS occurred, allele-specific PCR was performed on the noncrossover gene conversion progeny using primers specific to either the ry531 or ry606 allele.
In all experiments, the cross to generate females trans-heterozygous for ry606 and ry531 was done at 25 °C. When adults began to emerge, bottles were kept at 18 °C overnight for virgin collection, and 25 °C during the day. For wild-type, crosses of these females to ry506 males were incubated at room temperature, typically 20–22 °C. We analyzed 22 conversion tracts from these crosses. Crosses of rec mutant females to ry506 males were incubated at 25 °C. We expected that the difference in temperatures would not have a strong effect on meiotic recombination, since recombination occurs about 4 d prior to mature oocyte formation, which in most cases would be during virgin collection. Nonetheless, to ensure that the decreased tract length in rec mutants was not due to the different temperatures, we generated eight noncrossover gene conversions from wild-type females at 25 °C. These conversion tracts were in fact longer than those from the room temperature experiments (mean of 666 bp versus 374 bp). Because of this small sample size, however, we have based our analysis on all 29 wild-type tracts. We note that if tracts are truly longer at 25 °C, then the decreased tract length in rec mutants is even more severe than our analysis suggests.
Statistical analysis of gene conversion tract lengths.
We assume that the number of co-conversions follow a binomial distribution with parameter φk, where k is the number of bp between the selected and nonselected sites. In this analysis, we consider 44 distinct values of k. The values of k, co-conversions, and total conversions for all the wild-type data are given in Table S3. Let ci be the number of co-conversions at the ith site, and let ni be the total number of conversions. The likelihood of the entire dataset is given by a product of binomial probabilities:
The log of the likelihood in (1) is given by
To find the maximum likelihood estimate of φ, we take the derivative of (2) with respect to φ and set it equal to zero. We denote the maximum likelihood estimate of φ by
. Using standard large sample theory arguments [54], it can be shown that
is approximately normally distributed with large sample variance approximately equal to
where
The formulae in (3) and (4) facilitate the computation of large sample confidence intervals for φ, which take the form
, where
and
is the appropriate percentile of the standard normal distribution. For example, for a 95% confidence interval, α = 0.05 and z0.975 = 1.96.
We follow Hilliker et al. [49] and assume that the tract length distribution follows a geometric distribution with parameter φ, and hence the mean tract length is
and the expected selected tract length is
. Using standard large sample theory along with the delta method [54], it can be shown that
is approximately normally distributed with approximate large sample variance
and thus the (1 – α) × 100% confidence interval for m is given by
, where
.
Using the above results, we can formally test for differences between the mean tract lengths between any two datasets i and j using a Z-statistic, which takes the form
All computations of φ were done in XLIPSTAT [55] using double precision. The maximum likelihood estimate was computed in XLIPSTAT using the Nelder-Mead algorithm, which converged within 500 iterations for all datasets using a tolerance level of 10−10.
Supporting Information
Figure S1 Alignment of the Most Highly Conserved Region of MCM Proteins
This alignment shows the central region from all MCMs from Homo sapiens (Has), Drosophila melanogaster (Dma), Arabidopsis thaliana (Ath), Entamoeba histolytica (Ehi), and Giardia lamblia (Gla); the single MCM from the archaeal species Sulpholobus sulfataricus (Ssu); and MCM8 and MCM9 from Anopheles gambiae (Aga), Drosophila pseudoobscura (Dps), Drosophila virilis, and Encephalitozoon cuniculi (Ecu). A consensus is shown below the alignment, in which U, bulky aliphatic (I, L, M, V); @, aromatic (F, W, Y); &, bulky hydrophobic (I, L, M, V, F, W, Y); dot, any residue or no strong consensus. For this figure, a consensus residue is defined as one that is found in more than 80% of the sequences shown. Since 14 of the 44 sequences shown are from MCM8 and MCM9 orthologs, this figure emphasizes positions that are conserved between these more divergent subfamilies and canonical MCMs. Residues that match the consensus are shown in white text on a black background; conserved substitutions from the consensus are shown as white text on a gray background. The positions of the Walker A and B boxes and the arginine finger (RF) are indicated.
(43 KB PDF)
Click here for additional data file.
Figure S2 Phylogenetic Analysis of Eukaryotic MCM Family Proteins
The tree shown was generated by the neighbor-joining method of ClustalW, using the alignment of the most highly conserved region of the MCM core domain shown in Figure S1 (correcting for multiple substitutions but including positions with gaps; unrooted). Note that Giardia MCM2 clusters with MCM8 and MCM9 in this analysis. The numbers on each node are the percentage of trees with the given branch from 10,000 independent boot-strapped iterations. The scale represents the relationship of branch length to phylogenetic distance expressed as the number of substitutions per site. See Figure S1 legend for species names.
(13 KB PDF)
Click here for additional data file.
Table S1 Sequences Used for Phylogenetic Analysis
(40 KB DOC)
Click here for additional data file.
Table S2 Polymorphisms Used for Conversion Tract Length Determination
(80 KB DOC)
Click here for additional data file.
Table S3 Co-Conversion Data Used for Conversion Tract Length Determination
(67 KB DOC)
Click here for additional data file.
Accession Numbers
The FlyBase (http://flybase.bio.indiana.edu/search/) accession numbers for genes and gene products discussed in this paper are c(3)g (FBgn0000246), ERCC1 (FBgn0028434), MEI-9 (FBgn0002707), MEI-218 (FBgn0002709), MEI-P22 (FBgn0016036), MEI-W68 (FBgn0002716), MUS-312 (FBgn0002909), rosy(ry) (FBgn0003308), spn-A (FBgn0003479), and spn-E (FBgn0003483).
We thank Mitch McVey, Jan LaRocque, Lisa Antoszewski, and Corbin Jones for helpful discussions and comments on the manuscript, and members of the Sekelsky lab for assistance with virgin collection. SJR was supported by a Thomas S. and Caroline H. Royster, Jr. Graduate fellowship. HMK was supported by a Ruth Kirchstein National Research Service Award. This work was supported by a grant from the National Institutes of Health (RO1 GM61252 to JS).
Competing interests. The authors have declared that no competing interests exist.
Author contributions. HLB, SJR, and JS conceived and designed the experiments. HLB, SJR, SM, HMK, and JS performed the experiments. HLB, SJR, JGI, and JS analyzed the data. HLB, SJR, and JS wrote the paper.
A previous version of this article appeared as an Early Online Release on August 17, 2005 (DOI: 10.1371/journal.pgen.0010040.eor).
Abbreviations
bpbasepair
DHJdouble Holliday junction
DSBdouble-strand break
EMSethyl methanesulfonate
hDNAheteroduplex DNA
MCMmini-chromosome maintenance
PMSpost-meiotic segregation
SDSAsynthesis-dependent strand annealing
==== Refs
References
Szostak JW Orr-Weaver TL Rothstein RJ Stahl FW 1983 The double-strand break repair model for recombination Cell 33 25 35 6380756
Allers T Lichten M 2001 Differential timing and control of noncrossover and crossover recombination during meiosis Cell 106 47 57 11461701
Merker JD Dominska M Petes TD 2003 Patterns of heteroduplex formation associated with the initiation of meiotic recombination in the yeast Saccharomyces cerevisiae
Genetics 165 47 63 14504217
McKim KS Green-Marroquin BL Sekelsky JJ Chin G Steinberg C 1998 Meiotic synapsis in the absence of recombination Science 279 876 878 9452390
McKim KS Hayashi-Hagihara A 1998
mei-W68 in Drosophila melanogaster encodes a Spo11 homolog: Evidence that the mechanism for initiating meiotic recombination is conserved Genes Dev 12 2932 2942 9744869
Ghabrial A Ray RP Schupbach T 1998
okra and spindle-B encode components of the RAD52 DNA repair pathway and affect meiosis and patterning in Drosophila oogenesis Genes Dev 12 2711 2723 9732269
Yoo S McKee BD 2005 Functional analysis of the Drosophila Rad51 gene (spn-A) in repair of DNA damage and meiotic chromosome segregation DNA Repair (Amst) 4 231 242 15590331
Abdu U Gonzalez-Reyes A Ghabrial A Schupbach T 2003 The Drosophila spn-D gene encodes a RAD51C-like protein that is required exclusively during meiosis Genetics 165 197 204 14504227
Staeva-Vieira E Yoo S Lehmann R 2003 An essential role of DmRad51/SpnA in DNA repair and meiotic checkpoint control EMBO J 22 5863 5874 14592983
Ghabrial A Schupbach T 1999 Activation of a meiotic checkpoint regulates translation of Gurken during Drosophila oogenesis Nat Cell Biol 1 354 357 10559962
Sekelsky JJ McKim KS Chin GM Hawley RS 1995 The Drosophila meiotic recombination gene mei-9 encodes a homologue of the yeast excision repair protein Rad1 Genetics 141 619 627 8647398
Radford SJ Goley E Baxter K McMahan S Sekelsky J 2005
Drosophila ERCC1 is required for a subset of MEI-9-dependent meiotic crossovers Genetics In press.
Yildiz O Majumder S Kramer B Sekelsky JJ 2002
Drosophila MUS312 interacts with the nucleotide excision repair endonuclease MEI-9 to generate meiotic crossovers Mol Cell 10 1503 1509 12504024
McKim KS Dahmus JB Hawley RS 1996 Cloning of the Drosophila melanogaster meiotic recombination gene mei-218 : A genetic and molecular analysis of interval 15E Genetics 144 215 228 8878687
Kurkulos M Weinberg JM Roy D Mount SM 1994 P element-mediated in vivo deletion analysis of white-apricot: Deletions between direct repeats are strongly favored Genetics 136 1001 1011 8005410
Nassif N Penney J Pal S Engels WR Gloor GB 1994 Efficient copying of nonhomologous sequences from ectopic sites via P element-induced gap repair Mol Cell Biol 14 1613 1625 8114699
Adams MD McVey M Sekelsky JJ 2003
Drosophila BLM in double-strand break repair by synthesis-dependent strand annealing Science 299 265 267 12522255
Grell RF 1984 Time of recombination in the Drosophila melanogaster oocyte. III. Selection and characterization of temperature-sensitive and -insensitive recombination-deficient alleles in Drosophila
Genetics 108 435 443
Forsburg SL 2004 Eukaryotic MCM proteins: Beyond replication initiation Microbiol Mol Biol Rev 68 109 131 15007098
Matsubayashi H Yamamoto MT 1998 Dissection of chromosome region 89A of Drosophila melanogaster by local transposition of P elements Genes Genet Syst 73 95 103 9718674
Matsubayashi H Yamamoto MT 2003 REC, a new member of the MCM-related protein family, is required for meiotic recombination in Drosophila
Genes Genet Syst 78 363 371 14676427
Patel SS Picha KM 2000 Structure and function of hexameric helicases Annu Rev Biochem 69 651 697 10966472
Maiorano D Cuvier O Danis E Mechali M 2005 MCM8 is an MCM2–7 -related protein that functions as a DNA helicase during replication elongation and not initiation Cell 120 315 328 15707891
Volkening M Hoffmann I 2005 Involvement of human MCM8 in prereplication complex assembly by recruiting hcdc6 to chromatin Mol Cell Biol 25 1560 1568 15684404
Johnson EM Kinoshita Y Daniel DC 2003 A new member of the MCM protein family encoded by the human MCM8 gene, located contrapodal to GCD10 at chromosome band 20p12.3–13 Nucleic Acids Res 31 2915 2925 12771218
Gozuacik D Chami M Lagorce D Faivre J Murakami Y 2003 Identification and functional characterization of a new member of the human MCM protein family: hMCM8 Nucleic Acids Res 31 570 579 12527764
Baldauf SL 2003 The deep roots of eukaryotes Science 300 1703 1706 12805537
Walker JE Saraste M Runswick MJ Gay NJ 1982 Distantly related sequences in the alpha- and beta-subunits of ATP synthase, myosin, kinases, and other ATP-requiring enzymes and a common nucleotide binding fold EMBO J 1 945 951 6329717
Feger G Vaessin H Su TT Wolff E Jan LY 1995
dpa , a member of the MCM family, is required for mitotic DNA replication but not endoreplication in Drosophila
EMBO J 14 5387 5398 7489728
Treisman JE Follette PJ O'Farrell PH Rubin GM 1995 Cell proliferation and DNA replication defects in a Drosophila MCM2 mutant Genes Dev 9 1709 1715 7622035
Su TT Feger G O'Farrell PH 1996
Drosophila MCM protein complexes Mol Biol Cell 7 319 329 8688561
Feger G 1999 Identification and complete cDNA sequence of the missing Drosophila MCMs: DmMCM3, DmMCM6, and DmMCM7 Gene 227 149 155 10023044
Ohno K Hirose F Inoue YH Takisawa H Mimura S 1998 cDNA cloning and expression during development of Drosophila melanogaster MCM3, MCM6, and MCM7 Gene 217 177 185 9795205
Drysdale RA, Crosby MA, The Flybase Consortium 2005 Flybase: Genes and gene models Nucleic Acids Res 33 D390 D395 15608223
Baker BS Carpenter AT Ripoll P 1978 The utilization during mitotic cell division of loci controlling meiotic recombination in Drosophila melanogaster
Genetics 90 531 578 17248870
Boorstein RJ Pardee AB 1983 Coordinate inhibition of DNA synthesis and thymidylate synthase activity following DNA damage and repair Biochem Biophys Res Commun 117 30 36 6419736
Boyd JB Golino MD Setlow RB 1976 The mei-9a mutant of Drosophila melanogaster increases mutagen sensitivity and decreases excision repair Genetics 84 527 544 826452
McVey M Radut D Sekelsky JJ 2004 End-joining repair of double-strand breaks in Drosophila melanogaster is largely DNA ligase IV-independent Genetics 168 2067 2076 15611176
Baker BS Carpenter AT 1972 Genetic analysis of sex chromosomal meiotic mutants in Drosophila melanogaster
Genetics 71 255 286 4625747
Sandler L Lindsley DL Nicoletti B Trippa G 1968 Mutants affecting meiosis in natural populations of Drosophila melanogaster
Genetics 60 525 558 5728740
Bhagat R Manheim EA Sherizen DE McKim KS 2004 Studies on crossover-specific mutants and the distribution of crossing over in Drosophila females Cytogenet Genome Res 107 160 171 15467361
West SC 1997 Processing of recombination intermediates by the RuvABC proteins Annu Rev Genet 31 213 244 9442895
Constantinou A Chen XB McGowan CH West SC 2002 Holliday junction resolution in human cells: Two junction endonucleases with distinct substrate specificities EMBO J 21 5577 5585 12374758
Chovnick A Ballantyne GH Baillie DL Holm DG 1970 Gene conversion in higher organisms: Half-tetrad analysis of recombination within the rosy cistron of Drosophila melanogaster
Genetics 66 315 329 5512365
Chovnick A Ballantyne GH Holm DG 1971 Studies on gene conversion and its relationship to linked exchange in Drosophila melanogaster
Genetics 69 179 209 5002749
Clark SH Hilliker AJ Chovnick A 1988 Recombination can initiate and terminate at a large number of sites within the rosy locus of Drosophila melanogaster
Genetics 118 261 266 2834266
Hilliker AJ Clark SH Chovnick A 1991 The effect of DNA sequence polymorphisms on intragenic recombination in the rosy locus of Drosophila melanogaster
Genetics 129 779 781 1752421
Curtis D Bender W 1991 Gene conversion in Drosophila and the effects of the meiotic mutants mei-9 and mei-218
Genetics 127 739 746 2029970
Hilliker AJ Harauz G Reaume AG Gray M Clark SH 1994 Meiotic gene conversion tract length distribution within the rosy locus of Drosophila melanogaster
Genetics 137 1019 1026 7982556
Mazina OM Mazin AV Nakagawa T Kolodner RD Kowalczykowski SC 2004
Saccharomyces cerevisiae Mer3 helicase stimulates 3'-5′ heteroduplex extension by Rad51; implications for crossover control in meiotic recombination Cell 117 47 56 15066281
Nakagawa T Ogawa H 1999 The Saccharomyces cerevisiae MER3 gene, encoding a novel helicase-like protein, is required for crossover control in meiosis EMBO J 18 5714 5723 10523314
Thompson JD Gibson TJ Plewniak F Jeanmougin F Higgins DG 1997 The ClustalX windows interface: Flexible strageties for multiple sequence alignment aided by quality analysis tools Nucleic Acids Res 25 4876 4882 9396791
Schmidt HA Strimmer K Vingron M von Haeseler A 2002 TREE-PUZZLE: Maximum likelihood phylogenetic analysis using quartets and parallel computing Bioinformatics 18 502 504 11934758
Ferguson TS 1996 A course in large sample theory New York Chapman and Hall 256 p.
Tierney L 1990 Lisp-Stat: An object-oriented environment for statistical computing and dynamic graphics New York Wiley 397 p.
Bender W Spierer P Hogness DS 1983 Chromosomal walking and jumping to isolate DNA from the Ace and rosy loci and the bithorax complex in Drosophila melanogaster
J Mol Biol 168 17 33 6410077
|
16189551
|
PMC1231718
|
CC BY
|
2021-01-05 07:59:36
|
no
|
PLoS Genet. 2005 Sep 23; 1(3):e40
|
utf-8
|
PLoS Genet
| 2,005 |
10.1371/journal.pgen.0010040
|
oa_comm
|
==== Front
PLoS Comput BiolPLoS Comput. BiolpcbiplcbploscompPLoS Computational Biology1553-734X1553-7358Public Library of Science San Francisco, USA 1618419210.1371/journal.pcbi.001004305-PLCB-RA-0052R1plcb-01-04-04Research ArticleBioinformatics - Computational BiologyGenetics/Genome ProjectsGenetics/Chromosome BiologyPlant SciencePlantsOryzaReAS: Recovery of Ancestral Sequences for Transposable Elements from the Unassembled Reads of a Whole Genome Shotgun ReAS: Recovery of Ancestral SequencesLi Ruiqiang 12Ye Jia 12Li Songgang 23Wang Jing 2Han Yujun 2Ye Chen 2Wang Jian 12Yang Huanming 12Yu Jun 12Wong Gane Ka-Shu 124*Wang Jun 1256*1 James D. Watson Institute of Genome Sciences of Zhejiang University, Hangzhou, China
2 Beijing Institute of Genomics of Chinese Academy of Sciences, Beijing Genomics Institute, Beijing, China
3 College of Life Sciences, Peking University, Beijing, China
4 UW Genome Center, Department of Medicine, University of Washington, Seattle, Washington, United States of America
5 The Institute of Human Genetics, University of Aarhus, Aarhus, Denmark
6 Department of Biochemistry and Molecular Biology, University of Southern Denmark, Odense M, Denmark
Baxter Susan EditorThe National Center for Genome Resources, United States of America* To whom correspondence should be addressed. E-mail: [email protected] (GKW), [email protected] (JW)9 2005 23 9 2005 1 4 e4311 3 2005 23 8 2005 Copyright: © 2005 Li et al.2005This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are properly credited.We describe an algorithm, ReAS, to recover ancestral sequences for transposable elements (TEs) from the unassembled reads of a whole genome shotgun. The main assumptions are that these TEs must exist at high copy numbers across the genome and must not be so old that they are no longer recognizable in comparison to their ancestral sequences. Tested on the japonica rice genome, ReAS was able to reconstruct all of the high copy sequences in the Repbase repository of known TEs, and increase the effectiveness of RepeatMasker in identifying TEs from genome sequences.
Synopsis
Transposable elements (TEs) are a major component of the genomes of multicellular organisms. They are parasitic creatures that invade the genome, insert multiple copies of themselves, and then die. All we see now are the decayed remnants of their ancestral sequences. Reconstruction of these ancestral sequences can bring dead TEs back to life. Algorithms for detecting TEs compare present-day sequences to a library of ancestral sequences. Unknown to many, pervasive use of whole genome shotgun (WGS) methods in large-scale sequencing have made TE reconstructions increasingly problematic. To minimize assembly errors, WGS methods must reject the highly repetitive sequences that characterize most TEs, especially the most recent TEs, which are the least diverged from their ancestral sequences (and most informative for reconstruction). This is acceptable to many, because the most important parts of the genes are not repetitive, but for the TE aficionados, it is a problem. ReAS is a novel algorithm that does TE reconstruction using only the unassembled reads of a WGS. Tested against the WGS for japonica rice, it is shown to produce a library that is superior to the manually curated Repbase database of known ancestral TEs.
Citation:Li R, Ye J, Li S, Wang J, Han Y, et al (2005) ReAS: Recovery of ancestral sequences for transposable elements from the unassembled reads of a whole genome shotgun. PLoS Comput Biol 1(4): e43.
==== Body
Introduction
Transposable elements (TEs) make up a significant proportion of many eukaryotic genomes, totaling almost 45% for human [1], and 50% to 80% for rice, maize, and wheat [2–4]. They play important evolutionary roles [5,6], and can be essential tools for genome analyses [7]. RepeatMasker (A. Smit and P. Green, unpublished data) is one of the most commonly used algorithms for detecting TEs. It relies on a comparison to known TEs from libraries like Repbase [8], which represents many years of manual labor. Computational tools like REPuter [9], RECON [10], and RepeatGluer [11] automate this process. However, almost all of the genomes sequenced today employ a whole genome shotgun (WGS) method that is incapable of assembling the most recent TEs, and any efforts to force such an assembly together generally increase the probability of assembly errors. For example, in the mouse [12] and rice [13,14] genomes, 15% and 14% of the reads were left out of the assemblies, respectively. Some of the unassembled reads are due to centromeres or telomeres, but we know in rice that many are recent TEs. Unassembled reads are the most informative reads for TE recovery as they are the least diverged from their ancestral sequences. Despite this fact, all of the above tools were only tested on assembled genomes and it is not clear how effectively or efficiently they might incorporate the information in the unassembled reads. Hence, we developed a new algorithm, ReAS (from “recovery of ancestral sequences”), to produce the requisite TE library using only the unassembled reads of a WGS.
Ancestral sequence refers to the sequence of a TE when it was first inserted in the genome, and present-day sequence refers to the sequence of a TE as it exists today. With the passage of time, all TE sequences degenerate, and after a hundred million years or so, they become unrecognizable. It is the present-day sequence that cannot be assembled by a WGS (or by ReAS), but it is the ancestral sequence that is preferred by RepeatMasker, as divergence between an ancestral sequence and a present-day copy is half of that between two present-day copies. ReAS works on TEs that satisfy two assumptions. First, these TEs must exist at high copy numbers across the genome. Second, they must not be so old that they are no longer recognizable in comparison to their ancestral sequences. For such TEs, pieces of the ancestral sequence may still exist at high copy numbers, scattered across the genome, even if nowhere in the genome is there an intact version. Reconstruction of such ancestral sequences ought to be possible, as follows.
TEs are under no selective constraints once they insert into a genome. The process by which they subsequently decay is complex [15,16]. It includes mutational, insertional, and deletional events, plus transposition, amplification, and TE-mediated rearrangements. To the extent that this process is random, a consensus of present-day sequences should be a reasonable approximation of the ancestral sequence. Of course, for molecular evolution studies, a simple majority consensus is not good enough, and more detailed knowledge of the underlying biology for each TE is required to construct the correct ancestral sequence. The consensus is a good starting point, and for use as a RepeatMasker library in genome annotations, it suffices. Figure 1 depicts the ReAS process. We select a high-depth K-mer and retrieve all of the reads that contain this K-mer. Then we assemble these reads into an initial consensus sequence (ICS). Next, we look for new K-mers at the ends of the consensus and iteratively extend it until no further extensions are possible. Because most WGS projects generate reads from both ends of the clone inserts, we also take advantage of this linking information to resolve ambiguities and break misassemblies, but not to join fragmented assemblies. The end result is a ReAS TE.
Figure 1 The ReAS Algorithm
We start by computing K-mer depth, which is the number of times that a K-mer appears in the shotgun data. Copy number refers to how often a K-mer appears in the assembled genome. Depth divided by copy number is the coverage. We seed the process using a randomly chosen high-depth K-mer. All shotgun reads containing this K-mer are retrieved and trimmed into 100-bp segments centered at that K-mer. When the sequence identity between them exceeds a preset threshold, they are assembled into an ICS using ClustalW. We perform an iterative extension by selecting high-depth K-mers at both ends of the ICS and repeating the above procedure. After all such extensions are done, clone-end pairing information is used to resolve ambiguous joins and to break misassemblies, but not to join fragmented assemblies. The final consensus is our ReAS TE.
We will demonstrate how ReAS operates on japonica rice, as most of the rice TEs in Repbase are from that subspecies, and unassembled WGS reads [17] are available from the GenBank trace archives. Many of the software components for ReAS were developed for RePS, the WGS assembler first used in the indica rice genome [18,19]. Following the nomenclature in those papers, we define the depth of a K-mer as the number of times that it appears in the unassembled data. Copy number is how often it appears in the assembled genome. Shotgun coverage is the ratio of depth to copy number, which for japonica is 6×. The essential difference between RePS and ReAS is that the former avoids high-depth K-mers, and the latter seeks them out. These contrasting objectives are fundamentally at odds with each other. RePS must choose between leaving behind a lot of unassembled reads, versus having larger contigs with potentially more assembly errors. ReAS focuses on TE recovery alone, and is therefore more likely to get the right answer, in contrast to the other reconstruction algorithms like REPuter, RECON, or RepeatGluer, which must operate on a genome that is already misassembled by algorithms like RePS. All C subroutines and Perl scripts for ReAS are freely available from [email protected].
Results
One of the luxuries of doing this analysis on japonica is the fact that we have data from Syngenta [17], which uses a WGS method, and from the International Rice Genome Sequencing Project (IRGSP) [20–22], which uses a mapped-clone method. ReAS was run on the unassembled WGS reads from Syngenta, but when an assembled genome sequence was needed, we used the IRGSP results. The recovery process was seeded with K-mers of length K = 17, and a depth threshold of D = 14 was used. Mitochondria and chloroplast sequences were removed before analyzing the resultant ReAS TEs. The “gold standard” against which we benchmarked the recovered TEs was Repbase version 8.4.
Repbase Comparison
Figure 2 is an example of a perfectly recovered TE that exists in fragmented form in Repbase. This gypsy-like element is 10,841 bp. The region from 1 to 10,387 bp matches Repbase RIRE2_I (Internal) at 96% nucleotide identity, and the region from 10,401 to 10,841 bp matches Repbase RIRE2_LTR (long terminal repeat [LTR]) at 93% nucleotide identity. Notice that ReAS only recovered the LTR at one end of this TE, even though the correct ancestral sequence should have an LTR at both ends. This is a consequence of our restriction that reads already in a consensus be excluded from the next extension. Figure 3 depicts a recovered TE that had no counterpart within Repbase. We believe it is a valid TE because it has a BlastX alignment at 98% identity over 869 amino acids to a TE-related protein in GenBank (gi|34896386|ref|NP_909537.1| Putative mutator like transposase).
Figure 2 Complete Recovery of Known TE
RIRE2 is a gypsy-like TE that is found in two pieces in Repbase, as RIRE2_I (Internal) and RIRE2_LTR (LTR).
These have 96% and 93% nucleotide identity to our ReAS TE, respectively.
Figure 3 Recovery of Unknown TE
Although not found in Repbase, we believe that this ReAS TE is a valid reconstruction, because it has a BlastX match with identity 98% over 869 amino acids to a TE-related protein (gi|34896386|ref|NP_909537.1| Putative mutator like transposase) that is annotated in a GenBank clone.
More than 95% of the 17-mers in the recovered ReAS TEs had depths over 14, as we show in Figure S1. This is of course by design. In contrast, for the Repbase TEs, only half of the 17-mers had depths over 14. This is relevant for the comparisons, because not every one of these low-depth Repbase TEs is correct. It is clear, however, that ReAS cannot recover them, so for the comparisons, we considered only high-depth TEs in Repbase. WGS reads were aligned to Repbase TEs using BlastN. Good alignments were those of size greater than 100 bp and nucleotide identity of better than 85%. High-depth TEs were covered over 80% of their length with at least 14 aligned reads. A total of 54 Repbase TEs qualified, and these are described in Table S1. As we show in Table 1, 95.8% of 54 high-depth Repbase TEs were recovered, although sometimes in fragmentary form. If fragmentary recovery is not acceptable, and we credited only the best-matched ReAS TEs, 88.1% were recovered. The reduction in the recovery rate is mostly attributable to copia elements, which tend to have lower 17-mer depths and more divergent sequences.
Table 1 Nearly Complete Recovery of All 54 High-Depth TEs in Repbase
Table S2 compares each of these Repbase TEs to its best-matched ReAS TE. Over aligned regions, sequence identities averaged 96.8%. Where they failed to align, we computed error rates. False negative (FN) is the fraction of the Repbase TE that remains unaligned. It averages 11.9% in our dataset. False positive (FP) is the fraction of the ReAS TE that remains unaligned, but we must make some exceptions because there are many plausible explanations for these unaligned regions. We know that Repbase can be incomplete. ReAS TEs were larger than Repbase TEs in 42 of 54 cases. In seven instances, ReAS TEs were 2–18 times larger than Repbase TEs. This was due either to incomplete Repbase TEs, in the manner of Figure 2, and/or to concatemers of the form |- LTR -| |- Internal -| |- LTR -| |- Internal -|, which tend to occur when the LTRs are extremely diverged. Ignoring these seven instances, the average ratio of ReAS TE to Repbase TE size was actually 1.01. For our definition of FP, therefore, we ignored any unaligned regions at the ends and counted only unaligned regions in the middle, as these are the problems that are most likely to mislead a user. The average FP over the dataset was 1.6%.
Utility as TE Library
ReAS recovered 8,411 high-depth ancestral sequences with mean length of 640 bp and mean depth of 152, as we indicate in Table 2. Of these, 1,275 matched to known TEs in Repbase and 707 matched to TE-related proteins (keywords include retrotransposon, transposase, reverse transcriptase, gypsy, and copia); the remaining 6,429 were less easily classified. This last category is primarily composed of small low-depth repeats. Indeed, they drag down the overall mean length and depth. If we included only those repeats that matched to Repbase, the mean length was 1,634 bp and the mean depth was 464. Based on the arguments that we discuss below, we further subdivided this last category into 1,792 potentially interesting repeats (of length over 500 bp and depth over 35) and 4,637 probable artifacts.
Table 2 Ranking of All 8,411 ReAS TEs Based on Their Likelihood of Being TEs
Many of the ReAS TEs could be clustered together, on the criterion that a repeat was collapsed into a cluster when 80% of its length aligned with another member of that cluster at BlastN E-values of 10−5. This reduced the dataset to 7,015 clusters. The collapse was most pronounced among repeats with a match to Repbase, where 1,275 ReAS TEs were collapsed into 242 clusters. To explore the extent to which highly divergent TEs are assembled into different ReAS TEs, we performed a simulation. We started with an ancestral sequence of 500 bp, 2 kb, and 10 kb. From this, we simulated 100 present-day copies of the TE, with a range of divergences from the ancestral sequence. We simulated a 6× WGS and applied ReAS to that. Results were averaged over ten such simulations, as shown in Table S3. Some fragmentation was observed in the more divergent TEs, especially at larger sizes. However, even in the worst case of 0% to 40% divergence at 10 kb size, a single best-matched ReAS TE covered 86% of the ancestral TE. All of the other pieces were smaller than 500 bp in size and less than 35 in depth. Some collapsed into the best-matched ReAS TE. Almost all aligned to the ancestral TE, of which we invariably recovered more than 95%.
Table 3 classifies the 1,275 known TEs in the same manner as our previous indica genome analyses [13,14]. We recovered 691 (in 113 clusters) Class I TEs. There were 51 (in 25 clusters) copia and 381 (in 41 clusters) gypsy LTR retrotransposons, relatively few LINEs (long interspersed nuclear elements) and SINEs (short interspersed nuclear elements), plus 239 (in 39 clusters) unknown retrotransposons. This outcome is consistent with the finding that LTR retrotransposons are the single largest component of most plant genomes [23]. The ratio for copia to gypsy elements was 51/381 (or 25/41 by clusters). This is consistent with the finding that copia is less abundant than gypsy [24]. We recovered 217 (in 48 clusters) Class III TEs. These were typically quite small, with a mean size of 396 bp, and found in noncoding regions adjacent to genes [25].
Table 3 Classification of Recovered Sequences into Known TE Families
We subdivided the unclassified repeats by comparing their characteristics with those of our 1,275 known TEs. If we set a threshold at depth 35, then 9.3% of known TEs fell below this threshold, as opposed to 4,637 (84.7%) of 5,475 unclassified repeats of length under 500 bp. Conversely, we created chromosome-sized random sequences of length 10 Mb, simulated a 6× WGS on these, and ran ReAS. From ten such chromosomes, 3,419 ReAS TEs were recovered. These were obviously artifacts, which was easy to see because only 31 (0.9%) had length over 500 bp, while 38 (1.1%) had depth over 35. As a result, we used these cutoff thresholds to clean up our 6,429 unclassified repeats. Of the remaining 1,792 repeats, 89 (in 29 clusters) were of length over 1,000 bp and depth over 100, as shown in Tables S4–S9. One cluster of 45 repeats was centromeric in origin. Two clusters were attributable to ribosomal RNA and a seed prolamine gene. To check for pseudogenes, we searched for similarity to a nonredundant set of 19,079 full-length cDNAs [26] that are called nr-KOME. Seven clusters matched at BlastX E-values of 10−5, but five of these seven matched to cDNAs that show similarity to recently discovered TE-related proteins. If we eliminate these, we are left with 23 repeats (in 19 clusters) of mean length 1,795 bp and mean depth 739. This is comparable to known TEs. All but two of these clusters are 80% intact in the Syngenta or IRGSP assemblies, indicating they are not ReAS artifacts.
The ultimate test of utility is to use these ReAS TEs as a library for RepeatMasker and see how well that masks the rice genome. We did this analysis on the IRGSP genome, because this sequence was taken by a mapped-clone method, and is more representative of the true repeat content. Figure 4 and Table 4 show the comparison to and improvement over Repbase. Only 30.8% of the genome was masked using Repbase as the library, whereas 36.3% was masked when we used the ReAS TEs in the known and TE-related categories. This increased to 41.0% if we added the third category of “other repeats.” To consider whether gene duplications were a confounding factor, we ran RepeatMasker on the 19,079 gene regions defined by nr-KOME cDNAs. Only 1.7% of the nucleotides for the coding exons were masked, but 9.3% were masked if introns were included. In fact, even those few percent that were masked are not likely due to gene duplications, as there are TEs in cDNAs. Of the 246 genes where more than half of the coding exons were masked, four were ribosomal RNAs and the rest were BlastP homologous to TE-related proteins at E-values of 10−5.
Figure 4 Masking of Entire Rice Genome by RepeatMasker
We use different TE libraries and indicate the overlaps between these different results in a Venn diagram. ReAS (1to2) refers to the first two categories of Table 2 (Repbase and TE-related). ReAS (1to3) includes the third category (other repeats). Numbers are percentage of genome.
Table 4 Application of RepeatMasker to the Entire Rice Genome, Using Different Versions of the Repeat Library
Discussion
Given the inherent complexity of the ReAS algorithm, it is astonishing how well it does work, especially when benchmarked against the years of skilled labor that went into the production of Repbase. We would caution that our parameters were specifically tuned for rice, and they will likely need to be changed for non-grass genomes, to adjust for how TE history is different in other species. Even in rice, our current parameters worked much better for gypsy than for copia TEs. ReAS works best for high-copy TEs that have not had too much time to diverge from their ancestral sequences. The same constraints apply even in a manual reconstruction procedure, and a lot of hard work will be required to do much better than ReAS. Although it is true that Repbase has many low-depth TEs that are not in ReAS, there is often no evidence that these Repbase entries represent the correct ancestral sequence of any particular TE. Indeed, by using ReAS as the library in RepeatMasker, we can detect many more TEs than Repbase while missing little of what is in Repbase. Some manual curation is required for the post-analysis of recovered TEs. For example, |- LTR -| |- Internal -| |- LTR -| |- Internal -| concatemers must be resolved manually. ReAS removes a lot of the drudgery, but it is not the final answer. What it does is provide a starting point for expert annotation of the complete TE contents of a sequenced genome.
Materials and Methods
The ReAS algorithm.
We explain our choice of parameter settings in a later section. Here, we wish to discuss the principles behind the ReAS algorithm. In this paper, K = 17. Selection of the initial K-mer seed must satisfy three conditions. First, to avoid spurious matches, it cannot be a simple repeat like a poly-A tail. Second, the K-mer depth must exceed a predetermined threshold D. In this paper, D = 14. At a coverage of 6×, this corresponds to a copy number of 2.3. Finally, it should not be in a previously recovered ReAS TE.
Using the selected K-mer as a bait, we retrieve all sequence reads that contain this K-mer, and trim the reads down to 100-bp fragments centered at that K-mer. We align the fragments with each other and look for groups of D or more fragments with what we call 95% “mutual identity.” That is, we add fragments successively, and at each step, ask that the new fragments be 95% identical to at least 2/3 of the pre-existing fragments in that group. To avoid the combinatorial explosion, we employ a greedy algorithm. For the initial seeding, we consider only the biggest resultant group. For the extensions, we must consider all the resultant groups. This will create ambiguity problems. How these are resolved is discussed in a later section. Two commonly used alignment tools were tested: ClustalW [27] and Phrap (P. Green, unpublished data). ClustalW excelled at assembling a large number of fragments of somewhat differing sequence content, while Phrap excelled at assembling a small number of fragments of nearly identical sequence content. ClustalW tolerated the inevitable discrepancies, while Phrap often failed to assemble fragment groups as a result of these discrepancies. We therefore chose ClustalW.
To extend the consensus, we select high-depth K-mers from both ends of the ICS, and use these as secondary seeds to retrieve additional reads. Also, to reduce the chances that something might be missed, because of mutations or rearrangements, we consider all qualified K-mers (i.e., those satisfying the above seed conditions) within 50 bp of the ends. We require that the newly retrieved reads agree with the previous ICS in this 50-bp region, to within 95% identity. A constraint is imposed on the number of shared reads b. Suppose there are a + b reads in the previous ICS, and b + c reads in the new consensus. We ask that b/(a + b) > 20% or b/(b + c) > 20% or b > 200. Assuming that these conditions are satisfied, we extend the ICS by100 bp, starting from the end of the previous ICS. Since most reads are 500 bp long, no read is allowed to participate in more than five extensions. This process is repeated until no further extensions are possible, after which, a finishing step is used to get the last few bases. We walk out one base at a time, and stop whenever we see a 20-bp region where fewer than 60% of the reads agree to 95% mutual identity.
General difficulties.
The idealized algorithm described above is of course a simplification. In practice, there are three problems: ambiguity/misassembly, fragmentation, and segmental duplication. To resolve ambiguities and break misassemblies, we consider the long-range information that is available to us. The reads themselves at typical lengths of 500 bp are one source of such information. Use of read overlap data is implicitly incorporated, because we ask that the parent ICS and its extension share some portion of their reads. Another source of such information arises when reads are sampled from both ends of the clone inserts, which are typically 3 kb apart. Consider the following example.
Figure 5 shows a situation where two distinct TEs, a-e-c and b-e-d, share a similar fragment, e. Four reconstructions are possible: a-e-c, a-e-d, b-e-c, and b-e-d. If the shared region is not too long, for example, less than300 bp, the correct path may be identified by read overlaps. If not, the correct path may still be identified by clone-end pairing data. A few possibilities are listed here: (1) if a-e-c and b-e-d are supported, and nothing else, the other two paths can safely be eliminated; (2) if a-e-c is the only supported path, it is a less certain situation, but since the other path is most likely to be b-e-d, these are the paths we keep; (3) if a-e-c and a-e-d are both supported, it is difficult to know which path is correct, and so we keep all four; (4) if none of the four paths are supported, we keep them all. The operating philosophy is that we try to resolve all the ambiguities as safely as possible, but if it cannot be done, we keep all possible paths for future consideration.
Figure 5 Fork Problem during Consensus Extension
Suppose we have two TEs, each in three segments, a-e-c and b-e-d, where segment e is identically shared between the two TEs. Four results are possible (a-e-c, a-e-d, b-e-c, and b-e-d), and ReAS will compute all four paths. It then uses the overlapping reads or clone-end pairs for bridging information, and where possible, eliminates any incorrect paths.
Some TEs will not be recovered as a single consensus if, at some point along their length, the depth falls below D. The example in Figure 6 is for a gypsy-like TE called SZ-43LTR. Of the two recovered consensus fragments, one covers positions 1 to 927 bp and, compared to the Repbase TE, shows 98% nucleotide identity. The other covers positions 1,568 to 4,039 bp and shows 97% nucleotide identity. As expected, the break is in a region of low depth. Although in principle one can use clone-end pairing data to join fragmented assemblies, in practice we discovered that this procedure is too error-prone. For example, two distinct TEs may be adjacent to each other on the genome, and therefore they will be linked by a clone-end pair. Hence, we decided not to use the clone-end pairing data in this context. The information lost turns out to be minimal.
Figure 6 Fragmentation due to Low K-mer Depth
SZ-43LTR is the LTR region from a TE that is found as one piece in Repbase, but is recovered by ReAS as two nonoverlapping pieces, with 98% and 97% nucleotide identity to the Repbase entry.
Segmental duplication of large regions of a genome [28] causes another problem, because when the duplication is of sufficiently high copy number, it will be assembled as a “ReAS TE.” To the extent that there are TEs inside this duplication, we must determine their boundaries. We use an idea from RECON [10], taking advantage of the fact that TEs tend to occur at much higher copy numbers than duplications. Figure 7 explains the basic concept. TE boundaries are identified by sudden changes in depth, accompanied by many partially aligned reads. For any particular read, we define the endpoint as the boundary of its alignment with the ReAS TE. We search for any regions with a significant aggregation of endpoints and a significant depth discrepancy. On the low side, the depth is required to be less than 300. We ask that the ratio of high to low depth be greater than three. On the high side, we also require that there are endpoints for at least 50% of reads within 20 bp of the putative boundary. TEs so defined are then excised for further analysis.
Figure 7 TEs within Segmental Duplications
If the duplication is of sufficiently high copy number, it will be assembled as a “ReAS TE,” and what we need to do afterwards is find the boundaries of the TEs within this assembled duplication. On the assumption that TEs have much higher copy numbers, TE boundaries can be identified by sudden changes in depth, accompanied by many partially aligned reads.
Parameter settings.
In principle, ReAS is applicable to any genome (not just rice), with the appropriate changes in parameter settings. Table 5 lists the settings used for the Syngenta japonica 6× WGS, and explains how they might be adjusted for other genomes. For the specialists, we discuss the technical issues here. Consider the choice of K. It must be large enough for 4K to exceed the genome size. K ≥ 15 suffices for rice. Our algorithm needs a byte of memory for every possible K-mer, so there is a limit to how large K can be. Sixteen gigabytes are used at K = 17. Notice also that larger Ks might make it more difficult to recover the older TEs, where only the smallest fragments are still recognizable.
Table 5 Default Parameter Settings for Rice WGS
The threshold depth D is selected based on coverage and error rate considerations. If we assume that, in the unique portions of the genome, the read depths follow a Poisson distribution, then for a nominal coverage of 6× and a threshold depth of D = 14, one would expect a 0.1% probability for a unique sequence to be mistakenly called repetitive. This is of course only an approximate guideline, as the read depths do not really follow a Poisson distribution. As we show in Figure S2, 32.4% of the K-mers have a depth of one, because of sequencing errors. The number is large because a single error in one base pair will ruin every K-mer that overlaps with it. The data were filtered for base pairs of error probability worse than 10−2, but it is possible that our error probability is too optimistic, since we did not control the experimental conditions, and calibration was difficult.
We decided on the 95% mutual identity rule after a process of trial and error. For more divergent TEs in smaller genomes, a less stringent rule may be used. We would not recommend that it be less than 90%, because of the increased likelihood of misassemblies and the increased strain on computational resources. The 2/3 grouping factor is not a sensitive parameter. We simulated 1,000 present-day sequences by mutating an ancestral sequence to give divergences of 0% to 10%. The greedy algorithm treats each sequence as a node. Arcs are placed between every pair of nodes that pass the mutual identity threshold. We select the node with the most arcs, and consider all the other nodes in succession. The rule is that, when the number of arcs to the cluster exceeds the grouping factor, that node is added to the cluster. The number of nodes per cluster does vary with mutual identity, but at a 95% setting, it varies only 11% for grouping factors of 1/2 to 4/5.
For the extension process, we require that the number of shared reads exceed 20% or 200 reads. In theory, if all of the TE copies are full length, 80% of the reads should be shared because most reads are 500 bp and the ICS grows by 100 bp at a time. In practice, the 20% rule is easily satisfied by over 99% of valid (i.e., based on comparison to known TEs) reconstructions. However, without some sort of shared reads criterion, we would get too many misassemblies. The 200-reads threshold is not a sensitive parameter. Both parameters are robust, and we would not change either of them regardless of the genome.
We benchmarked parameters for the RECON-inspired splitting against segmental duplications from the assembled rice genome. We used known TEs to determine the read depth and breakpoint distributions. Reads that belong to the duplication are fully aligned, but those that belong to a TE from somewhere else are only partially aligned and break at the TE boundary. One such example is shown in Figure S3. In practice, we would expect the situation to be different in every genome, reflecting differences in duplication history, and these parameters must be adjusted accordingly.
Supporting Information
Figure S1 Comparison of Repbase and ReAS Showing Amount of Dataset at the Stated 17-mer Depth
Amount of data is defined by the total length of the TEs, not the number of TEs, so longer TEs contribute more to the ordinate. The vertical line marks our D = 14 threshold.
(593 KB TIF)
Click here for additional data file.
Figure S2 Observed Distribution of 17-mer Depths for Syngenta WGS versus Expected Poisson Distribution at Shotgun Coverage of 6×
Observed values are indicated by a solid line; expected values are indicated by a dashed line.
(560 KB TIF)
Click here for additional data file.
Figure S3 Example of a TE in a Segmental Duplication
This duplication has five copies in the entire genome, but living inside it is a TE with hundreds of copies. Reads are aligned in BlastN. Hits must exceed 100 bp and 85% nucleotide identity. Endpoints are declared when a read has over 50 unaligned bases at the end.
(578 KB TIF)
Click here for additional data file.
Table S1 Description of All 54 High-Depth TEs in Repbase
We compared shotgun reads to Repbase TEs. BlastN alignments that exceed 100 bp and 85% nucleotide identity are mapped back to the Repbase TEs. High-depth TEs are those that are over 80% covered by high-depth alignments (HD blocks) of depth over 14.
(20 KB XLS)
Click here for additional data file.
Table S2 Best-Matched ReAS TE for All 54 High-Depth TEs in Repbase
Asterisks have been appended to the names of those seven outliers that are clearly due to incomplete Repbase TEs and/or LTR concatemers. We define FN as the fraction of a Repbase TE that remains unaligned. Conversely, FP is the fraction of a ReAS TE that remains unaligned, where we ignore unaligned regions at the ends and count only those in the middle. Total and average are given in the final row. For the size ratio, we exclude the seven outliers. For FP and FN, we compute length-weighted averages.
(29 KB XLS)
Click here for additional data file.
Table S3 Effects of TE Sequence Divergence
We started with an ancestral sequence of 500 bp, 2 kb, and 10 kb. From this, we simulated 100 TE copies with the stated range of divergences from the ancestral TE. We simulated a 6× WGS and ran ReAS. Results are averaged over ten such simulations. This table indicates the number of recovered ReAS TEs, the number of clusters they collapse into, the number of likely artifacts of size less than 500 bp and 17-mer depth less than 35, the percentage of ancestral TE covered by all ReAS TEs, the percentage covered by only the best-matched ReAS TE, and the number of recovered pieces that cannot be aligned to the ancestral sequence.
(18 KB XLS)
Click here for additional data file.
Table S4 Description of Known ReAS TEs
We show length, mean 17-mer depth, cluster number, and classification information.
(228 KB XLS)
Click here for additional data file.
Table S5 Description of TE-Related ReAS TEs
We show length, mean 17-mer depth, cluster number, and classification information.
(134 KB XLS)
Click here for additional data file.
Table S6 Description of Other ReAS TEs
We show length, mean 17-mer depth, cluster number, and classification information. Here, we also consider non-TE interpretations by indicating (under “class”) if the entity is a simple or low-complexity repeat, and (under “notes”) if it is a ribosomal RNA or centromeric repeat. To check for likely pseudogenes, we searched for similarity to the nr-KOME cDNAs using BlastX at an E-values of 10−5. To verify that what we recovered is not an artifact of the ReAS process, we indicate the number of intact copies found in the Syngenta (WGS) and IRGSP (clone-by-clone) assemblies, where by “intact” we mean 80% of the ReAS TE is aligned at 85% nucleotide identity.
(238 KB XLS)
Click here for additional data file.
Table S7 FASTA-Formatted Sequence for Known ReAS TEs
(2.1 MB TXT)
Click here for additional data file.
Table S8 FASTA-Formatted Sequence for TE-Related ReAS TEs
(1.0 MB TXT)
Click here for additional data file.
Table S9 FASTA-Formatted Sequence for Other ReAS TEs
(1.3 MB TXT)
Click here for additional data file.
This work was sponsored by the Chinese Academy of Sciences (KSCX1-SW-03 and KSCX2-SW-223), Commission for Economy Planning, Ministry of Science and Technology (2001AA225041, 2002AA229021, 2002AA2Z1001, 2002AA104250, 2002AA234011, 2001AA231061, 2001AA231011, 2004AA231050, 2003AA207160, and 2002AA229061), the National Natural Science Foundation of China (30399120, 30200159, 30370330, 30370872, 30200163, and 90208019), Zhejiang University, and China National Grid. Some additional funding came from the National Human Genome Research Institute (1 P50 HG02351) and Danish National Research Foundation (Danish Platform for Integrative Biology). We thank Eric Lander for an important insight, and one of the anonymous reviewers for many constructive suggestions.
Competing interests. The authors have declared that no competing interests exist.
Author contributions. Gane Ka-Shu Wong and Jun Wang conceived and designed the experiments. Ruiqiang Li, Songgang Li, Jing Wang, and Jun Yu performed the experiments. Ruiqiang Li, Jia Ye, Songgang Li, and Yujun Han analyzed the data. Chen Ye, Jian Wang, and Huanming Yang contributed reagents/materials/analysis tools. Ruiqiang Li, Gane Ka-Shu Wong, and Jun Wang wrote the paper.
A previous version of this article appeared as an Early Online Release on August 23, 2005 (DOI: 10.1371/journal.pcbi.0010043.eor).
Abbreviations
FNfalse negative
FPfalse positive
ICSinitial consensus sequence
IRGSPInternational Rice Genome Sequencing Project
LTRlong terminal repeat
TEtransposable element
WGSwhole genome shotgun
==== Refs
References
Lander ES Linton LM Birren B Nusbaum C Zody MC 2001 Initial sequencing and analysis of the human genome Nature 409 860 921 11237011
SanMiguel P Bennetzen JL 1998 Evidence that a recent increase in maize genome size was caused by massive amplification of intergene retrotransposons Ann Bot 81 37 44
Vicient CM Suoniemi A Anamthawat-Jonsson K Tanskanen J Beharav A 1999 Retrotransposon BARE-1 and its role in genome evolution in the genus Hordeum Plant Cell 11 1769 1784 10488242
Meyers BC Tingey SV Morgante M 2001 Abundance, distribution, and transcriptional activity of repetitive elements in the maize genome Genome Res 11 1660 1676 11591643
Kidwell MG Lisch D 1997 Transposable elements as sources of variation in animal and plants Proc Natl Acad Sci U S A 94 7704 7711 9223252
Avise JC 2001 Evolving genomic metaphors: A new look at the language of DNA Science 294 86 87 11588247
Kumar A Hirochika H 2001 Applications of retrotransposons as genetic tools in plant biology Trends Plant Sci 6 127 134 11239612
Jurka J 1998 Repeats in genomic DNA: Mining and meaning Curr Opin Struct Biol 8 333 337 9666329
Kurtz S Choudhuri JV Ohlebusch E Scheleiermacher C Stoye J 2001 REPuter: The manifold applications of repeat analysis on a genomic scale Nucleic Acids Res 29 4633 4642 11713313
Bao Z Eddy SR 2002 Automated de novo identification of repeat sequence families in sequenced genomes Genome Res 12 1269 1276 12176934
Pevzner PA Tang H Tesler G 2004 De novo repeat classification and fragment assembly Genome Res 14 1786 1796 15342561
Waterston RH Lindblad-Toh K Birney E Rogers J Abril JF 2002 Initial sequencing and comparative analysis of the mouse genome Nature 420 520 562 12466850
Yu J Hu S Wang J Wong GK Li S 2002 A draft sequence of the rice genome (Oryza sativa L. ssp. indica ) Science 296 79 92 11935017
Yu J Wang J Lin W Li S Li H 2005 The genomes of Oryza sativa: A history of duplications PLoS Biol 3 e38. DOI: 10.1371/journal.pbio.0030038 15685292
Smit AF 1996 The origin of interspersed repeats in the human genome Curr Opin Genet Dev 6 743 748 8994846
Fedoroff N 2000 Transposons and genome evolution in plants Proc Natl Acad Sci U S A 97 7002 7007 10860963
Goff SA Ricke D Lan TH Presting G Wang R 2002 A draft sequence of the rice genome (Oryza sativa L. ssp. japonica ) Science 296 92 100 11935018
Wang J Wong GK Ni P Han Y Huang X 2002 RePS: A sequence assembler that masks exact repeats identified from the shotgun data Genome Res 12 824 831 11997349
Zhong L Zhang K Huang X Ni P Han Y 2003 A statistical approach designed for finding mathematically defined repeats in shotgun data and determining the length distribution of clone-inserts Genomics Proteomics Bioinformatics 1 43 51 15626332
Sasaki T Matsumoto T Yamamoto K Sakata K Baba T 2002 The genome sequence and structure of rice chromosome 1 Nature 420 312 316 12447438
Feng Q Zhang Y Hao P Wang S Fu G 2002 Sequence and analysis of rice chromosome 4 Nature 420 316 320 12447439
Rice Chromosome 10 Sequencing Consortium 2003 In-depth view of structure, activity, and evolution of rice chromosome 10 Science 300 1566 1569 12791992
Kumar A Bennetzen JL 1999 Plant retrotransposons Annu Rev Genet 33 479 532 10690416
McCarthy EM Liu J Lizhi G McDonald JF 2002 Long terminal repeat retrotransposons of Oryza sativa
Genome Biol 3 RESEARCH0053 12372141
Kikuchi K Terauchi K Wada M Hirano HY 2003 The plant MITE mPing is mobilized in anther culture Nature 421 167 170 12520303
Kikuchi S Satoh K Nagata T Kawagashira N Doi K 2003 Collection, mapping, and annotation of over 28,000 cDNA clones from japonica rice Science 301 376 379 12869764
Thompson JD Gibson TJ Plewniak F Jeanmougin F Higgins DG 1997 The CLUSTAL_X windows interface: Flexible strategies for multiple sequence alignment aided by quality analysis tools Nucleic Acids Res 25 4876 4882 9396791
Bailey JA Yavor AM Massa HF Trask BJ Eichler EE 2001 Segmental duplications: Organization and impact within the current human genome project assembly Genome Res 11 1005 1017 11381028
|
16184192
|
PMC1232128
|
CC BY
|
2021-01-05 09:18:22
|
no
|
PLoS Comput Biol. 2005 Sep 23; 1(4):e43
|
utf-8
|
PLoS Comput Biol
| 2,005 |
10.1371/journal.pcbi.0010043
|
oa_comm
|
==== Front
BMC BioinformaticsBMC Bioinformatics1471-2105BioMed Central London 1471-2105-6-2141612488310.1186/1471-2105-6-214Research ArticleSources of variation in Affymetrix microarray experiments Zakharkin Stanislav O [email protected] Kyoungmi [email protected] Tapan [email protected] Lang [email protected] Stephen [email protected] Katherine E [email protected] Rudolph S [email protected] David B [email protected] Grier P [email protected] Section on Statistical Genetics, University of Alabama at Birmingham, Birmingham, Alabama, USA2 Departments of Pharmacology and Toxicology, University of Alabama at Birmingham, Birmingham, Alabama, USA3 Heflin Center for Human Genetics, University of Alabama at Birmingham, Birmingham, Alabama, USA4 Department of Bioinformatics and Biostatistics, University of Louisville, Louisville, Kentucky, USA2005 29 8 2005 6 214 214 19 4 2005 29 8 2005 Copyright © 2005 Zakharkin et al; licensee BioMed Central Ltd.2005Zakharkin et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
A typical microarray experiment has many sources of variation which can be attributed to biological and technical causes. Identifying sources of variation and assessing their magnitude, among other factors, are important for optimal experimental design. The objectives of this study were: (1) to estimate relative magnitudes of different sources of variation and (2) to evaluate agreement between biological and technical replicates.
Results
We performed a microarray experiment using a total of 24 Affymetrix GeneChip® arrays. The study included 4th mammary gland samples from eight 21-day-old Sprague Dawley CD female rats exposed to genistein (soy isoflavone). RNA samples from each rat were split to assess variation arising at labeling and hybridization steps. A general linear model was used to estimate variance components. Pearson correlations were computed to evaluate agreement between technical and biological replicates.
Conclusion
The greatest source of variation was biological variation, followed by residual error, and finally variation due to labeling when *.cel files were processed with dChip and RMA image processing algorithms. When MAS 5.0 or GCRMA-EB were used, the greatest source of variation was residual error, followed by biology and labeling. Correlations between technical replicates were consistently higher than between biological replicates.
==== Body
Background
Microarray chips are a powerful technology capable of measuring expression levels of thousands of genes simultaneously. Expression profiling has led to dramatic advances in the understanding of cellular processes at the molecular level, which may lead to improvements in molecular diagnostics and personalized medicine [1]. The number of experiments involving microarrays grows nearly exponentially each year [2]. Several platforms are currently available, including the commonly used short oligonucleotide-based Affymetrix GeneChip® arrays, which utilize multiple probes for each gene and automated control of the experimental process from hybridization to quantification. Although microarrays have tremendous potential, great effort and care is required in planning and designing microarray experiments, analyzing gene expression data, and interpreting results [3-6].
A typical microarray experiment has many different sources of variation which can be attributed to biological and technical causes [4]. Biological variation results from tissue heterogeneity, genetic polymorphism, and changes in mRNA levels within cells and among individuals due to sex, age, race, genotype-environment interactions and other factors [7-10]. Biological variation reflects true variation among experimental units (i.e. individual mice, rats, tissue samples, etc.) and is of interest to investigators. However, preparation of samples, labeling, hybridization, and other steps of microarray experiment can contribute to technical variation, which can significantly impact the quality of array data [11-16]. To ensure highly reproducible microarray data, technical variation should be minimized by controlling the quality of the RNA samples, and by efficient labeling and hybridization [17].
Identifying sources of experimental variation and assessing their magnitude are important for optimal experimental design, as for example, in the planning of mRNA pooling in microarray experiments [18]. Similarly, this information is useful for estimating the optimal number of required technical replicates because measurement accuracy and reliability affect researchers' power to identify differentially expressed genes [19]. However, other considerations, such as the goals of the study, the features of a particular microarray platform, or the cost of arrays and samples may influence experimental design [4-6]. Several studies have been conducted to examine the relative contributions of various factors in different experimental settings [7-15]. Here, we estimated the relative magnitudes of sources of variation in experiments involving Affymetrix GeneChip® arrays and evaluated agreement between biological and technical replicates.
Results
Experimental design
The experiment was set up as described in Materials and Methods (see Figure 1). Source *.cel data files from 24 GeneChip® arrays were subjected to image processing by four popular methods for probe-level data implemented in BioConductor [20]: DNA Chip Analyzer (dChip) [21], MAS 5.0 [22], RMA [23], and GCRMA-EB [24].
Figure 1 Experimental design. The scheme of hierarchical unbalanced design used in our experiment is shown. A total of 8 rats and 24 chips were used.
Variance components estimation
For each probe set, expression levels were modeled as follows: yg = μg + Bg + L(B)g + εg, where Bg ~ N(0, ) is the effect of biological variation among experimental units; L(B)g ~ N(0, ) is the effect of labeling nested within biological replications; and, εg ~ N(0, ) is the residual error. It should be noted that in our case biological variation could be confounded by technical variation arising during tissue isolation and preparation of mRNA samples. Dobbin et al., 2005, found that variation at this stage of microarray processing was small compared to variation at the hybridization step [25]. The model was fit separately on the gene expression measurements of each of dChip, MAS 5.0, RMA and GCRMA-EB probe set summaries. Both the effects of biological replication and the labeling effect nested within biological cases were treated as random. We estimated variance components and applied shrinkage variance estimators to them. These shrunken estimators borrow information across genes and have been shown to improve statistical tests [26]. Figure 2 shows the density plots of the distributions of relative magnitudes of different sources of variation. The results indicated that for most of the genes, the biggest source of variation was biological when using dChip and RMA, whereas the biggest source of variation was residual error when using GCRMA-EB or MAS 5.0. For all algorithms, a significant number of probe sets had biological and labeling variance components estimates equal or very close to zero. The findings are summarized in Table 1.
Figure 2 Density plots of different sources of variation. Density plots of relative magnitudes of different sources of variation are shown for data analyzed with four image processing algorithms. The proportions of different variance components are shown on x-axis and frequencies of probe sets are shown on y-axis.
Table 1 Proportions of different sources of variation
Source dChip MAS 5.0 RMA GCRMA-EB
Mean SD Mean SD Mean SD Mean SD
Biological variation 0.431 0.304 0.292 0.300 0.393 0.306 0.310 0.292
Labeling variation 0.206 0.230 0.136 0.198 0.221 0.224 0.147 0.192
Residual error 0.363 0.274 0.572 0.311 0.386 0.272 0.543 0.298
Assessment of reproducibility
We investigated agreement between technical replicates and biological replicates using Pearson correlations between chips. The correlations for the following three groups were compared: (1) Correlations between two technical replicates at the hybridization stage within a biological replicate (i.e., chips i_2A vs. i_2B; total of 8 correlations); (2) Correlations between two technical replicates at the labeling stage within a biological replicate (i.e. chips i_1 vs. i_2A and i_1 vs. i_2B; total of 16 correlations); (3) Correlations between different biological replicates (all possible pairwise comparisons; total of 252 correlations). Results indicated that technical replicates at the hybridization step agree more closely (i.e. have consistently higher correlations) either than technical replicates at the labeling stage or than different biological replicates (Figure 3). This finding can be illustrated using scatter plots: regardless of the image processing method, technical replicates of the same biological replicate (Figure 4) show less dispersion than data from different animals (Figure 5).
Figure 3 Boxplots of pairwise correlations between chips. Box plots of Pearson correlations between technical replicates at the hybridization step (Hybr; i_2A vs. i_2B chips, where i is biological replicate), labeling step (Label; i_1 vs. i_2A and i_1 vs i_2B chips), and between different biological replicates (Bio; all pairwise combinations) are shown for four image processing algorithms (dChip, MAS 5.0, RMA, GCRMA-EB). Technical replicates have consistently higher correlations than different biological replicates.
Figure 4 Comparison of two technical replicates of the same biological replicate using different image processing techniques. Expression levels detected on the 1_2A chip (x-axis) are plotted against levels detected on the 1_2B chip (y-axis). Results obtained with different image processing algorithms are shown. dChip and MAS 5.0 are shown on the log scale for compatibility with RMA and GCRMA-EB. Good agreement between two chips will result in data grouped along the identity line, while lack of agreement will lead to dispersion.
Figure 5 Comparison of two different biological replicates using different image processing techniques. Expression levels detected on the 1_2A chip (x-axis) are plotted against levels detected on the 6_1 chip (B) (y-axis). Results obtained with different image processing algorithms are shown. dChip and MAS 5.0 are shown on the log scale for compatibility with RMA and GCRMA-EB. Good agreement between two chips will result in data grouped along the identity line, while lack of agreement will lead to dispersion.
The reproducibility at the hybridization stage was assessed by testing the significance of the differences between expression levels of technical replicates at the hybridization step using a paired t-test analysis as described in Material and Methods. Briefly, for each probe set we tested the hypothesis that a difference in expression levels between two technical replicates (i.e., between i_2A and i_2B chips) is equal to zero. A total of 15,923 paired t-tests were conducted and 15,923 p-values obtained for each image processing algorithm. The distribution of p-values was modeled using a mixture model approach [27]. Under a global null hypothesis, there are no differentially expressed genes and distribution of p-values is expected to be uniform on [0, 1]. If some genes are truly differentially expressed, we expect an increased number of small p-values (near 0). Distributions of p-values for the data obtained by four image processing methods are presented on Figure 5. By fitting the mixture of two beta distributions, one can estimate proportion of differentially expressed genes. We obtained the following estimates: dChip – 10.8%; MAS 5.0 – 4.8%; RMA – 2.3%, and GCRMA-EB – 13.6%. Thus, at the nominal α-level 0.05, the number of differentially expressed genes was smaller than expected by chance when data were processed with MAS 5.0 or RMA, but above the nominal α-level when data was processed with dChip or GCRMA-EB.
Discussion
Using Affymetrix GeneArray® chips, we examined the relative magnitudes of different sources of variation in microarray experiment. Analysis of variance using mixed-effects linear models is a common way to account for and test the significance of various factors contributing to overall variation [3]. Due to limitations of our hierarchical unbalanced experimental design and relatively small number of degrees of freedom, we did not include factors that can potentially contribute to variation such as day of processing, scanning order, mRNA preparation, etc. We assume that such factors were not significant. However, to formally test this assumption, another experiment is needed.
We used a general linear model to partition variance for each probe set into three components. The first source was biological (i.e. animal-to-animal) variation. The biological variation may be confounded by technical variation at the mRNA preparation step, but this variation is probably relatively small compared to variation at the hybridization step [25]. Thus, we assume that most of the variation for this effect was due to true biological differences among animals. The second source of variation was the effect of labeling. Although our experiments were carried out by the same person, using the same equipment, under the same experimental conditions as much as realistically possible, there is always some variation caused by minor environmental differences in temperature, duration, pipetting etc., which influences labeling efficiency. The third source of variation other than animal-to-animal variation and labeling-effect variation was residual error caused by differences in hybridization, scanning and other factors. To compare the relative magnitudes of different sources of variation, we estimated variance components and applied shrunken variance estimators that borrow information across genes. We constructed these shrunken variance estimators by shrinking a group of individual variance estimators toward their common corrected geometric mean [26]. The amount of shrinkage depends on the variation on the individual variance components estimators. These estimators were shown to be robust in respect to variance heterogeneity in gene expression data among groups [26].
We found that our results depend on the image processing algorithm used: biological variation was the largest source when dChip or RMA were used, but when *.cel files were processed with GCRMA-EB or MAS 5.0, the largest source was residual error. Bakay et al., 2002, found that biological variation presumably caused by tissue heterogeneity and genetic polymorphism was a major source of variation while technical variation was minor [12]. Han et al., 2004, found that biological variation was about of the same size as other sources combined [14]. Whitney et al., 2003, found that inter-individual variation in gene expression profiles was correlated with gender, age, and the time of day at which the sample was taken. These intrinsic differences in expression patterns were likely caused by differences in genotype, although they might also reflect epigenetic or environmental factors [9]. Oleksiak et al., 2002, in their studies of teleost fish have observed significant differences in gene expression levels between individuals from the same population and between different populations. These differences could be caused by genetic variation as well as other factors, including maternal effects and genotype-environment interactions [10]. On the contrary, Dumur et al., 2004, found that day-to day variation was the main source of variation [17]. Woo et al., 2004, in studies of inbred mice strains, detected that most of the genes had small biological variance, but about 10% of genes showed large variation between individuals [28].
We found that technical replicates within a biological replicate had higher and more consistent correlations with each other than with other biological replicates. Generally, our correlations were higher than those observed by Dobbin et al., 2005, for interlaboratory correlations between tumor samples [25] and were compatible with values for in-lab correlations obtained in another study [29].
The consistency of the hybridization step was evaluated using paired t-tests following by modeling of distribution of resulting p-values. The significance depends on the image processing algorithm used: the hybridization effect was not significant for MAS 5.0 (4.8% of genes were differentially expressed between two technical replicates) and RMA (2.3% of genes), but the proportion of differentially expressed genes was higher than expected by chance for dChip (10.8% of genes) and GCRMA-EB (13.6% of genes).
The low-level data were analyzed using four popular methods implemented in the BioConductor [20] package: dChip [21], MAS 5.0 [22], RMA [23], and GCRMA-EB [24]. We found that different low-level data processing algorithms produced different results. We provide comparisons mainly to illustrate the compatibility of several algorithms. Evaluation of the strengths and weaknesses of different image processing algorithms may require other experimental settings, such as spike-in data. Shedden et al., 2005, performed a comprehensive comparison of seven image processing methods for Affymetrix arrays and demonstrated that the choice of image processing algorithm has a major impact on the results of microarray data analysis [30]. The authors found that the dChip method operates consistently well, while MAS 5.0 and GCRMA-EB consistently performed poorly. GCRMA-EB had a particular disagreement with other methods when a t-test was used for group comparison, presumably because it might be more sensitive to the underlying statistical assumptions of a test (e.g. independence of genes). Similarly, we observed that estimates of the proportion of differentially expressed genes between two technical replicates at the hybridization stage were different than those for data processed with GCRMA-EB compared to other methods, which is consistent with finding of Shedden et al. [30].
The results presented here are specific for the systems being studied, and other experimental conditions may yield different estimates. For example, we used an outbred strain of rats, which had greater inherent biological variation than inbred strains. In cell cultures of inbred mice strains under otherwise equal conditions, the relative magnitude of biological variation presumably would be smaller. Different steps in microarray data analysis, such as normalization, transformation, and gene filtering, may affect results as well [31-35]. A microarray platform and microarray facility can also have a significant impact, as was demonstrated in several recent studies [25,36-38]. Testing the influence of these various factors could be an interesting topic of future research.
Conclusion
Identification of sources of variation and their relative magnitudes, among other factors, is important for optimal experimental design and the development of quality control procedures. In this study, we evaluated the relative magnitudes of different sources of variation in Affymetrix microarray experiments. Different image processing algorithms gave different variance components estimates: the greatest source was animal-to-animal (i.e. biological) variation when dChip and RMA were used, and residual error when MAS 5.0 or GCRMA-EB were used. We observed that correlations between technical replicates within one biological replicate were consistently higher than between different biological replicates. It should be noted that estimates obtained here were specific for our experimental system, and results would probably change if we used another organism or tissue, or another microarray platform.
Methods
Samples and microarrays
This study included samples taken from eight 21-day-old Sprague Dawley CD female rats exposed to genistein (a soy isoflavone) via their mother's milk. The mothers were fed AIN-76A diet supplemented with 200 mg genistein / kg chow. Young rats were sacrificed at day 21 and the 4th mammary glands extracted and flash-frozen in liquid nitrogen within 3 minutes of ex-sanguination. Samples were frozen at -70°C for approximately 90 days, at which point the extraneous fat was dissected off and samples processed in Trizol RNA extraction buffer. Total RNA was generated using Affymetrix RNA extraction and labeling kits according to manufacturer's protocols, and each of the RNA samples was split in half. The first half was labeled and run on a RAE 230A Affymetrix GeneChip®, and the other half was labeled, split, and run across two RAE 230A chips (see Figure 1). Affymetrix arrays were run in the Genomics Core facility of the Heflin Center for Human Genetics at the University of Alabama at Birmingham. Images were scanned on a HP 2500 scanner.
Image processing
Each of the low-level *.cel data files was processed using four popular image analysis algorithms: DNA Chip Analyzer (dChip) [21], MAS 5.0 [22], RMA [23], and GCRMA-EB [24]. The processing was done in R 1.8.1 / R 1.9.1 [39]. The default settings for all normalization procedures were used as implemented in the BioConductor [20]; in particular, the scale normalization for MAS 5.0; the quantile-quantile normalization for RMA; the invariant-set normalization for dChip; and the quantile-quantile normalization for GCRMA-EB (see [35] for the details of the different normalization methods). The default implementation of dChip, RMA, and GCRMA-EB used only the PM (perfect match) intensity matrix, while MAS 5.0 by default used both PM and MM (mismatch) matrices.
Evaluation of relative magnitudes of different sources of variation
The relative magnitudes of different sources of variation were estimated using a general linear model in PROC VARCOMP procedure of SAS 9.1 (SAS Institute Inc., Cary, NC) using REML option. The expression levels of each probe set, yg, were modeled as follows: yg = μg + Bg + L(B)g + εg, where Bg ~ N(0, ) is the effect of biological variation among experimental units; L(B)g ~ N(0, ) is the effect of labeling variation nested within biological replications; and εg ~ N(0, ) is the residual error, i.e. technical variation caused by factors other than labeling. Biological effect could be confounded by technical variation arising during mRNA sample preparation. For each probe set, variance components were estimated. We applied shrinkage variance estimators that borrow information across probe sets and improve individual variance estimators by shrinking them toward their corrected geometric mean [26]. The total variance was assumed to be the sum of three components: VARTot = VARBio + VARLabel + VARResidual, where VARBio is the shrunken estimate of biological variance; VARLabel is the shrunken estimate of variance due to labeling; and VARResidual is the shrunken variance estimate of residual error. The relative proportion of each source of variation was calculated as a ratio of the shrunken variance estimate to the sum of all three shrunken variance estimates:, i.e. calculates the proportion of biological variation, calculates the proportion of variation due to labeling within biological replicates, and calculates the proportion of variation due to unaccounted technical variation (residual error).
Assessment of reproducibility across different replicates
Pearson correlations between chips were calculated for the following three groups: (1) Correlations between two technical replicates at the hybridization step (i.e., chips i_2A vs. i_2B; total of 8 correlations); (2) Correlations between two technical replicates at the labeling step (i.e. chips i_1 vs. i_2A and i_1 vs. i_2B; total of 16 correlations); (3) Correlations between different biological replicates (all possible pairwise comparisons; total of 252 correlations).
To evaluate the significance of variation introduced at the hybridization step, paired t-tests were performed on 16 chips (i_2A and i_2B chips from each of 8 separate rats). For each probe set, the null hypothesis was that the difference between the expression levels of two replicates was equal to zero. A total of 15,923 t-tests were performed and 15,923 p-values were generated for each image processing algorithm. The distribution of resulting p-values was modeled using a mixture of two beta distributions [24]. If the global null hypothesis is true, there are no differentially expressed genes and the distribution of p-values is expected to be uniform [0, 1]. We expect an increased number of p-values close to 0 if some genes are truly differentially expressed. By fitting the mixture of two beta distributions, one can estimate a proportion of differentially expressed genes. At the nominal α-level 0.05, one expects 5% of genes to be differentially expressed just by chance. Thus, the differences between replicates were considered significant only if the proportion of differentially expressed genes was > 5%.
Authors' contributions
SOZ performed analysis of the data, drafted and finalized the manuscript. KK and RP helped with analysis and contributed to discussion. KES performed microarray experiment that generated the data. SB provided support for microarray experiment and contributed to discussion. TM and LC analyzed the data with four image-analysis algorithms. GPP planned and designed the experiment. GPP and DBA supervised and coordinated the project and assisted with the interpretation. All authors have read and approved the manuscript.
Figure 6 Distributions of p-values for the paired t-test for hybridization effect. Histograms of p-values for four image processing algorithms. If the global null hypothesis is true, the distribution of p-values would be uniform from 0 to 1 (dotted line). If differentially expressed genes are present, the number of small p-values will be increased.
Acknowledgements
This work was supported in part by NIH grants U54CA100949 and T32HL072757, and NSF grants 0217651 and 0090286. We thank Dr. Xiangqin Cui for critical reading of the manuscript and for making valuable suggestions.
==== Refs
Jain KK Applications of biochips: from diagnostics to personalized medicine Curr Opin Drug Discov Devel 2004 7 285 289 15216931
Gracey AY Cossins AR Application of microarray technology in environmental and comparative physiology Annu Rev Physiol 2003 65 231 259 12471169 10.1146/annurev.physiol.65.092101.142716
Kerr MK Martin M Churchill GA Analysis of variance for gene expression microarray data J Comput Biol 2000 7 819 837 11382364 10.1089/10665270050514954
Churchill GA Fundamentals of experimental design for cDNA microarrays Nat Genet 2002 32 Suppl 490 495 12454643 10.1038/ng1031
Yang YH Speed T Design issues for cDNA microarray experiments Nat Rev Genet 2002 3 579 88 12154381
Kerr MK Churchill GA Experimental design for gene expression microarrays Biostatistics 2001 2 183 201 12933549 10.1093/biostatistics/2.2.183
Molloy MP Brzezinski EE Hang J McDowell MT VanBogelen RA Overcoming technical variation and biological variation in quantitative proteomics Proteomics 2003 3 1912 1919 14625853 10.1002/pmic.200300534
Spruill SE Lu J Hardy S Weir B Assessing sources of variability in microarray gene expression data BioTechniques 2002 33 916 923 12398201
Whitney AR Diehn M Popper SJ Alizadeh AA Boldrick JC Relman DA Brown PO Individuality and variation in gene expression patterns in human blood Proc Natl Acad Sci 2003 100 1896 1901 12578971 10.1073/pnas.252784499
Oleksiak MF Churchill GA Crawford DL Variation in gene expression within and among natural populations Nat Genet 2002 32 261 266 12219088 10.1038/ng983
Brown JS Kuhn D Wisser R Power E Schnell R Quantification of sources of variation and accuracy of sequence discrimination in a replicated microarray experiment BioTechniques 2004 36 324 332 14989098
Bakay M Chen YW Borup R Zhao P Nagaraju K Hoffman EP Sources of variability and effect of experimental approach on expression profiling data interpretation BMC Bioinformatics 2002 3 4 11936955 10.1186/1471-2105-3-4
Breit S Nees M Schaefer U Pfoersich M Hagemeier C Muckenthaler M Kulozik AE Impact of pre-analytical handling on bone marrow mRNA gene expression Br J Haematol 2004 126 231 243 15238145 10.1111/j.1365-2141.2004.05017.x
Han ES Wu Y McCarter R Nelson JF Richardson A Hilsenbeck SG Reproducibility, sources of variability, pooling, and sample size: important considerations for the design of high-density oligonucleotide array experiments J Gerontol A Biol Sci Med Sci 2004 59 306 315 15071073
Liu J Walker N Waalkes MP Hybridization buffer systems impact the quality of filter array data J Pharmacol Toxicol Methods 2004 50 67 71 15233970 10.1016/j.vascn.2004.02.001
Page GP Edwards JW Barnes S Weindruch R Allison DB A design and statistical perspective on microarray gene expression studies in nutrition: the need for playful creativity and scientific hard-mindedness Nutrition 2003 19 997 1000 14624952 10.1016/j.nut.2003.08.001
Dumur CI Nasim S Best AM Archer KJ Ladd AC Mas VR Wilkinson DS Garrett CT Ferreira-Gonzalez A Evaluation of quality-control criteria for microarray gene expression analysis Clin Chem 2004 50 1994 2002 15364885 10.1373/clinchem.2004.033225
Kendziorski CM Zhang Y Lan H Attie AD The efficiency of pooling mRNA in microarray experiments Biostatistics 2003 4 465 477 12925512 10.1093/biostatistics/4.3.465
Allison DB Allison RL Faith MS Paultre F Pi-Sunyer FX Power and money: designing statistically powerful studies while minimizing financial costs Psychological Methods 1997 2 20 33 10.1037//1082-989X.2.1.20
Gentleman RC Carey VJ Bates DM Bolstad B Dettling M Dudoit S Ellis B Gautier L Ge Y Gentry J Hornik K Hothorn T Huber W Iacus S Irizarry R Leisch F Li C Maechler M Rossini AJ Sawitzki G Smith C Smyth G Tierney L Yang JY Zhang J BioConductor: open software development for computational biology and bioinformatics Genome Biol 2004 5 R80 15461798 10.1186/gb-2004-5-10-r80
Li C Wong WH Model-based analysis of oligonucleotide arrays: expression index computation and outlier detection Proc Natl Acad Sci 2001 98 31 36 11134512 10.1073/pnas.011404098
Hubbell E Liu WM Mei R Robust estimators for expression analysis Bioinformatics 2002 18 1585 92 12490442 10.1093/bioinformatics/18.12.1585
Irizarry RA Hobbs B Collin F Beazer-Barclay YD Antonellis KJ Scherf U Speed TP Exploration, normalization, and summaries of high density oligonucleotide array probe level data Biostatistics 4 249 264 12925520 10.1093/biostatistics/4.2.249
Wu Z Irizarry R Gentleman R Murillo F Spencer F A model based background adjustment for oligonucleotide expression arrays Technical Report 2004 John Hopkins University, Department of Biostatistics Working Papers, Baltimore, MD
Dobbin KK Beer DG Meyerson M Yeatman TJ Gerald WL Jacobson JW Conley B Buetow KH Heiskanen M Simon RM Minna JD Girard L Misek DE Taylor JM Hanash S Naoki K Hayes DN Ladd-Acosta C Enkemann SA Viale A Giordano TJ Interlaboratory comparability study of cancer gene expression analysis using oligonucleotide microarrays Clin Cancer Res 2005 11 565 572 15701842
Cui X Hwang JT Qiu J Blades NJ Churchill GA Improved statistical tests for differential gene expression by shrinking variance components estimates Biostatistics 2005 6 59 75 15618528 10.1093/biostatistics/kxh018
Allison DB Gadbury GL Heo M Fernandez JR Lee C Prolla TA Weindruch R A mixture model approach for the analysis of microarray gene expression data Comput Stati Data Anal 2002 39 1 20 10.1016/S0167-9473(01)00046-9
Woo Y Affourtit J Daigle S Viale A Johnson K Naggert J Churchill G A comparison of cDNA, oligonucleotide, and Affymetrix GeneChip gene expression microarray platforms J Biomol Tech 2004 15 276 284 15585824
Wang H He X Band M Wilson C Liu L A study of inter-lab and inter-platform agreement of DNA microarray data BMC Genomics 2005 6 71 15888200 10.1186/1471-2164-6-71
Shedden K Chen W Kuick R Ghosh D Macdonald J Cho KR Giordano TJ Gruber SB Fearon ER Taylor JM Hanash S Comparison of seven methods for producing Affymetrix expression scores based on False Discovery Rates in disease profiling data BMC Bioinformatics 2005 6 26 15705192 10.1186/1471-2105-6-26
Geller SC Gregg JP Hagerman P Rocke DM Transformation and normalization of oligonucleotide microarray data Bioinformatics 2003 19 1817 1823 14512353 10.1093/bioinformatics/btg245
Thygesen HH Zwinderman AH Comparing transformation methods for DNA microarray data BMC Bioinformatics 2004 5 77 15202953 10.1186/1471-2105-5-77
Qin LX Kerr KF Contributing Members of the Toxicogenomics Research Consortium Empirical evaluation of data transformations and ranking statistics for microarray analysis Nucleic Acids Res 2004 32 5471 5479 15479783 10.1093/nar/gkh866
Pounds S Cheng C Statistical development and evaluation of microarray gene expression data filters J Comput Biol 2005 12 482 495 15882143 10.1089/cmb.2005.12.482
Bolstad BM Irizarry RA Astrand M Speed TP A comparison of normalization methods for high density oligonucleotide array data based on variance and bias Bioinformatics 2003 19 2 12499286 10.1093/bioinformatics/19.2.185
Toxicogenomics Research Consortium Standardizing global gene expression analysis between laboratories and across platforms Nat Methods 2005 2 351 356 15846362 10.1038/nmeth754
Larkin JE Frank BC Gavras H Sultana R Quackenbush J Independence and reproducibility across microarray platforms Nat Methods 2005 2 337 344 15846360 10.1038/nmeth757
Irizarry RA Warren D Spencer F Kim IF Biswal S Frank BC Gabrielson E Garcia JG Geoghegan J Germino G Griffin C Hilmer SC Hoffman E Jedlicka AE Kawasaki E Martinez-Murillo F Morsberger L Lee H Petersen D Quackenbush J Scott A Wilson M Yang Y Ye SQ Yu W Multiple-laboratory comparison of microarray platforms Nat Methods 2005 2 345 350 15846361 10.1038/nmeth756
Ihaka R Gentleman R A language for data analysis and graphics J Comp Graph Stat 1996 5 299 314
|
16124883
|
PMC1232851
|
CC BY
|
2021-01-04 16:27:27
|
no
|
BMC Bioinformatics. 2005 Aug 29; 6:214
|
utf-8
|
BMC Bioinformatics
| 2,005 |
10.1186/1471-2105-6-214
|
oa_comm
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.