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
Genome BiolGenome Biology1465-69061465-6914BioMed Central London gb-2005-6-6-r541596080610.1186/gb-2005-6-6-r54MethodSingle-feature polymorphism discovery in the barley transcriptome Rostoks Nils 1Borevitz Justin O [email protected] Peter E 1Russell Joanne 1Mudie Sharon 1Morris Jenny 1Cardle Linda 1Marshall David F 1Waugh Robbie [email protected] Scottish Crop Research Institute, Genome Dynamics, Invergowrie, Dundee, DD2 5DA, Scotland, UK2 University of Chicago, Department of Ecology and Evolution, Chicago, IL 60637, USA2005 11 5 2005 6 6 R54 R54 8 2 2005 22 3 2005 14 4 2005 Copyright © 2005 Rostoks 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.
A probe level model for analysis of GeneChip gene expression data is presented which identified more than 10,000 single-feature polymorphisms between two barley genotypes, with a high sensitivity. This method is applicable to all oligonucleotide microarray data.
A probe-level model for analysis of GeneChip gene-expression data is presented which identified more than 10,000 single-feature polymorphisms (SFP) between two barley genotypes. The method has good sensitivity, as 67% of known single-nucleotide polymorphisms (SNP) were called as SFPs. This method is applicable to all oligonucleotide microarray data, accounts for SNP effects in gene-expression data and represents an efficient and versatile approach for highly parallel marker identification in large genomes.
==== Body
Background
Whole-genome sequences of Arabidopsis and rice have provided a fundamental platform for the discovery of gene content and function in dicot and monocot plants. Research on the model species has provided a wealth of knowledge on universal biochemical and genetic processes, as well as the development of analytical tools that are applicable to other plant species [1-3].
The availability of abundant, high-throughput sequence-based markers is the key for detailed genome-wide trait analysis. Single-nucleotide polymorphisms (SNP) are the most common sequence variation and a significant amount of effort has been invested in resequencing alleles to discovery SNPs. In fully sequenced small-genome model organisms SNP discovery is relatively straightforward, although high-throughput SNP discovery in natural populations remains both expensive and time-consuming [4].
A number of recent studies have reported the use of oligonucleotide arrays, including expression arrays, for SNP detection in a highly parallel manner [5]. In these studies, whole genomic DNA was demonstrated to work very well for simple organisms such as yeast [6,7], and even complex, albeit relatively small genomes, such as Arabidopsis [8]. However, the application of oligonucleotide arrays for SNP detection in large genomes, such as human, has relied on prior complexity reduction using PCR-based enrichment [9,10]. The use of oligonucleotide arrays for simultaneous genotyping and gene-expression analysis using RNA target has also been reported in yeast [11]. While there is arguably little need for enhanced SNP discovery in yeast, the real power of the approach came from coupling genotyping and gene expression analysis.
For large-genome species, including crops such as wheat and barley, full-genome sequences may not be available in the near future. This has been compensated to some extent by model species that have allowed conserved biological processes to be studied. However, while Arabidopsis and rice provide insights into universal genetic, structural and developmental processes, they fail to address many topics relevant to crop-plant species, such as yield, yield stability and quality. Rice has a long history as a genetic model that has been strengthened by release of draft genome sequences [12,13]. As a result of conservation of synteny at the genomic level it has been promoted as a model for the grasses [14]. However, unlike the temperate cereals such as wheat and barley, rice cultivation occurs under short days and rather specific environmental conditions, its end uses are distinct and numerous exceptions to conserved synteny have now emerged [15-17]. Together, these highlight the limitations of rice as a universal genetic model for the cereal grasses.
Wheat and barley together constitute one third of world cereal production [18]. Barley in particular is cultivated throughout the world, in environments as diverse as arctic regions of Northern Europe, subtropical regions of Africa and the highlands of the Andes and the Himalayas [19]. Barley breeding has created varieties tailored mainly for animal feed, malt production and human food [20]. Ultimately, environmental and agronomical variation is based on genetic (sequence) diversity of the barley genome, with expression of agronomic traits closely linked to environmental adaptability.
With genome sizes of around 5,200 megabase pairs (Mbp) for barley [21,22] and around 16,100 Mbp for bread wheat [21] and genomic structure consisting of gene islands interspersed with highly repetitive retrotransposon sequences [15,23], access to sequence-based markers is currently provided through highly developed expressed sequence tag (EST) resources [24].
The most important traits in crop species are generally polygenic. These have traditionally been studied using biparental mapping populations and a large pool of mapped restriction fragment length polymorphism (RFLP) and/or simple sequence repeat (SSR) markers [25]. However, with the strong trend towards genome-wide association analyses based on linkage disequilibrium (LD) [26,27] there is a clear need for robust high-density and high-throughput markers that can be effectively deployed, often in closely related elite germplasm. While the number and distribution of markers for LD studies in barley remains to be empirically determined, SNP markers offer both the sequence specificity and throughput necessary for the success of this approach. SNP discovery in large-genome species is currently limited to identifying SNPs in silico in EST assemblies and resequencing of EST-derived unigenes in relevant germplasm [27], and scaling-up such approaches requires significant investment of both time and funding [28-30]. An approach that would allow parallel screening of the whole 'gene space' for SNPs is therefore highly desirable.
An Affymetrix GeneChip that allows simultaneous expression analysis of 22,000 transcripts has recently become available for barley [31]. Transcription provides a native mechanism for the enrichment of gene sequences. Polymorphisms present in DNA are transcribed into the messenger RNA and can potentially affect the hybridization to the GeneChip probes, if present in a region complementary to the probe. Polymorphisms generated during mRNA processing, such as alternative splicing and polyadenylation, could also affect hybridization of the target RNA.
Here we report the use of the Affymetrix Barley1 GeneChip to identify single-feature polymorphisms (SFP), which include not only SNPs but also the processing polymorphisms mentioned above, in barley transcript profiling data from cultivars Morex and Golden Promise. The statistical algorithm presented here allowed us to distinguish genotype-dependent hybridization differences at the probe level once overall gene-expression level was accounted for, leading to the identification of 10,504 SFPs.
Results
Identification of SFP in Barley1 GeneChip transcription-profiling data
Gene-expression data for barley cultivars Morex and Golden Promise was generated within an international collaborative project of barley researchers (unpublished results, see Acknowledgements) and consisted of 36 GeneChip hybridizations (three replicates of six tissue types) for two genotypes. Raw microarray data are available from ArrayExpress [32,33], BarleyBase [34] and [35]. The analysis code, lists of RNA and genomic SFPs, primer sequences, and the SFP sequence confirmation table are available from our website as supplementary information [35]. The hybridization intensities for each of the perfect match (PM) probes were extracted from the .CEL files. Background correction and quantile normalization was performed using the Bioconductor package RMA [36,37]. The resulting data matrix of 22,801 probe sets with 11 PM probes each was analyzed using probe-level linear models that accounted for main fixed effects of genotype, tissue, and individual probe intensity, as well as tissue-specific differences across genotypes. One replicate from a single tissue sample of Golden Promise consistently clustered with the analogous Morex replicates and this sample was reclassified as Morex. The residuals from the linear model were saved into a matrix of 250,811 probes by 36 arrays and subsequently fitted for a genotype effect at the probe level to identify SFPs between the 17 Golden Promise and 19 Morex arrays. The Bioconductor package siggenes [37] was used to determine SFPs according to statistical analysis of microarrays (SAM) [38,39].
Figure 1 shows effects of the normalization steps on the expression profile of the probe set Contig10034_at and identification of a SFP in the probe 3 by removing probe and tissue effects. The large number of replicates for each genotype and the reduced genome complexity of the transcribed RNA allowed 10,504 SFPs to be identified at less than 0.1% false discovery rate (FDR) (Table 1). These SFPs resided in 3,734 Affymetrix probe sets, with one quarter of probe sets containing four or more SFPs. The magnitude of the d-statistic indicated the likelihood of a probe being called an SFP, while the sign indicated which genotype was polymorphic with regard to the reference 25mer probe on the array. Positive values predicted an SFP in Golden Promise, while negative values indicated an SFP in Morex (a complete list of SFP probes and corresponding d-statistics are available from [35]). Figure 2a shows the distribution of observed d-statistics (y-axis) of all probes on the array against the expected mean permutation null distribution (x-axis). Probes exceeding the threshold of less than 0.1% FDR, and thus containing SFP, are shown in green. Figure 2b is a histogram of the distribution of d-statistics truncated at ± 10 with thresholds shown. Figure 2b is a histogram of d-statistics truncated at ± 10 with Golden Promise SFPs in the right tail and Morex SFPs in the left tail.
Sequence confirmation of SFP
Confirmation of SFP was done by comparison with three barley sequence datasets. Barley EST [40] is EST unigene assembly 21 [40] and contained 234 contigs with 624 predicted SFP probes where both Morex and Golden Pomise sequence were available. These were examined manually to identify SNP that overlapped 25mers on the array (see SFP confirmation table in [35] (EST dataset)).
The second set is an experimental cDNA sequence set targeting regions with predicted SFPs. Comparative DNA sequence was generated from each genotype by targeted resequencing of reverse-transcription PCR (RT-PCR) products covering 262 probes. For each genotype we combined an equal amount of RNA from all six tissue types used for hybridization to the GeneChips and converted it to a single-stranded cDNA. PCR amplification and subsequent sequencing allowed us to obtain good-quality sequence from both genotypes (see SFP confirmation table in [35] (targeted dataset)).
The third set was an experimental random genomic DNA sequence set used as a tool for SNP discovery in barley [30]. This dataset (SFP confirmation table in [35] (random dataset)) consisted of barley unigenes that had been resequenced from genomic DNA from eight barley lines, including Morex and Golden Promise, within an ongoing SNP discovery project [30]. The selection of these genes was considered random with respect to the genes predicted to have SFP. The SNP discovery project targeted the 3' ends of unigenes, the region also selected for Affymetrix probe design. The random-sequencing dataset consisted of sequences for 300 unigene contigs and covered a total of 2,204 Affymetrix probes with high-quality sequences from both genotypes.
In total, 2,699 probes were analyzed in the three datasets, of which 2,667 were unique and 31 were present in multiple datasets. Sixty-six probes were polymorphic compared to both genotypes and, since they could not be detected by our algorithm, they were excluded from further analysis. 401 unique probes contained sequence polymorphisms - 223 features were polymorphic compared to Golden Promise and 178 to Morex. 2,200 probes did not have a sequence polymorphism (Table 2; SFP confirmation table in the supplementary information at [35]).
The sequence polymorphism information was compared with the expression SFP genotype calls. Of the 401 known sequence polymorphisms, 270 were correctly predicted by our analysis, indicating 67% sensitivity. Only 25 SFPs were called where sequence confirmation revealed the polymorphism in an opposite genotype, while 155 known SNPs escaped detection. How many of the 10,504 predicted SFPs were found actually to contain a sequence polymorphism? We have sequence information for 450 of these probes, of which 270 contained SNP in the predicted genotype. This suggested that up to 40% of the 10,504 predicted SFPs may be 'falsely discovered' sequence polymorphisms (Tables 2, 3). The large discrepancy between the permutation FDR threshold of 0.1% and that determined by sequencing is due to several factors. Expression polymorphisms, such as alternative splicing or polyadenylation, do not affect primary sequence, and are also detected in our statistical model. Genes with multiple adjacent SFPs may fall into this category. In addition, true SNPs near the 25mer may be identified as SFPs due to labeling polymorphisms.
The ability to detect sequence polymorphisms in the RNA-profiling data depends on several properties, including the expression status of the gene in a particular tissue type, the location of the SNP within the 25mer and the hybridization properties of the particular feature. We further investigated the effect of SNP position on the ability to identify a sequence polymorphism as an SFP in transcription data. SNP position was recorded as distance from the edge of the probe, position 1 being either end and 13 being the middle of the 25mer. Figure 3 shows that, as expected, when a SNP was located in the central region (positions 6-13) it was more often called as a SFP. SNP residing in the flanking three nucleotides were called at near the background rate. Probes containing multiple SNPs were also efficiently predicted (Figure 3). A similar pattern has been seen in genomic DNA hybridizations in Arabidopsis [8] and yeast [6], and in RNA hybridizations in yeast [11].
Comparison of SFP prediction in individual tissues against the full sample
We tested the sensitivity and false SNP discovery rates of our analysis with single tissue/genotype comparisons to observe how it would perform in smaller experiments. Datasets containing three replicates per genotype for each tissue type were analyzed at the threshold that again identified 10,504 SFP. In general there was a 4-16% decrease in sensitivity of the SFP prediction, which was the expected result of reducing power. On the other hand, SFP prediction in a single tissue type decreased the false SNP discovery rate by 4-5%. This was probably due to the reduction of probe-level variation in expression across tissues. In all, more than 10,000 SFP could be reliably identified even when expression profiles of single tissues were analyzed.
Genomic DNA hybridizations
To assess the feasibility of SFP identification from barley total genomic DNA (around 5200 Mbp) [21,22], we labeled and hybridized three replicates of three highly polymorphic genotypes, Oregon Wolfe Barley Dominant and Oregon Wolfe Barley Recessive [41], and wild barley species Hordeum vulgare ssp. spontaneum (accession Mehola), to the same Affymetrix Barley1GeneChip expression array. Raw microarray data are available from [35]. Raw .CEL files were background corrected and quantile normalized and the package siggenes [37,38] was subsequently used to identify probes showing significant hybridization differences between genotypes. To assess significance, 100 random permutations were performed, FDRs were evaluated at different thresholds (Table 1) and 1,090 SFPs were identified at a 22% FDR. Although there was less power to identify SFP with nine replicates in the genomic DNA dataset compared to 36 replicates in the RNA dataset, there was also much more noise relative to signal from barley genomic DNA. This was most probably due to the complexity of the large barley genome and a lower proportion of gene regions in the labeled DNA. However, if SFPs identified in genomic DNA were real, common polymorphisms in barley should be identified by both RNA and DNA approaches, even though different genotypes were used. As shown in Table 4, a significant overlap was identified between the two SFP sets, with 114 SFPs in common where only 46 are expected by chance (p < 3.863e-25). More replicates and alternative gene-specific labeling conditions should improve genomic DNA SFP identification from organisms with very large genomes [9].
Discussion
Affymetrix GeneChips designed for gene-expression analysis can be utilized for genome-wide identification of sequence polymorphisms [5]. Whole-genome DNA has been used as a hybridization target in yeast [6,7] and in Arabidopsis [8] to identify SFPs using expression arrays. While such an approach was valid in yeast and a small-genome model plant, the transfer of this approach to cereal crop plants with up to 100-fold larger genome sizes is problematic. The number of genes in barley is likely to be comparable to the estimated number of genes in Arabidopsis and rice [42,43]. However, the amount of repetitive DNA in barley will dilute the gene-specific signal in the target labelled DNA.
Until now, PCR-based artificial enrichment for a subset of sequences has been used to tackle the complexity of large genomes [10,9,44]. Using RNA as a hybridization target provides a natural way of enriching for gene sequences while maintaining all the sequence diversity present in transcribed sequences. However, sequence polymorphism effects on hybridization are concealed within the overall variation in gene-expression levels and tissue-dependent and genotype-dependent differential gene expression. Additional complexity comes from posttranscriptional sequence polymorphisms, such as alternative splicing and alternative polyadenylation. New array designs that tile probes across genes and intergenic regions will help unravel this complexity as nucleotide polymorphisms may affect single features while alternative transcripts may more often affect adjacent features.
We present here a statistical approach that allows us to reliably discern the probe-level differential hybridization between two genotypes that is often caused by sequence polymorphisms once variation in overall gene-expression level is normalized. Our approach allows the use of expression array data generated from different tissue types, and thus increases its versatility and applicability to the wide range of currently available oligonucleotide microarray data.
The analysis algorithm was applied to gene-expression microarray data generated from two barley genotypes with six tissue types each for a total of 36 array hybridizations. At a stringent 0.1% FDR, 10,504 SFPs were identified. Comparison to the available sequence-verified SNP data suggested that 67% of the known SNPs were predicted, confirming a good sensitivity. Approximately 40% of the SFP probes that were sequence-verified did not reveal any polymorphisms at the sequence level; thus, the FDR was up to 13-fold higher compared to the rate for Arabidopsis genomic DNA hybridizations [8]. The higher false-positive rates can be at least partly explained by variation in mRNA structure (for example, alternative splicing and polyadenylation) between tissues, and possibly between genotypes, which would lead to differential hybridization to probes but could not be detected by sequencing. A recent study using an EST collection concluded that at least 4% of barley genes may undergo alternative splicing [45]; however, more experimental data may be required to correctly model the rate of probe level variation in plant gene-expression data.
For practical application the balance between the cost of replicates and the number of replicates necessary to maintain sensitivity is important. We therefore analyzed the microarray data comparing just three replicates of each tissue type from the two genotypes (Table 3). Overall sensitivity decreased, but remained above 50%. Remarkably, the false SNP discovery rate was better for single tissue comparisons, probably because variation in mRNA transcript processing among tissues was eliminated.
Certain molecular marker applications require the precise nature of sequence changes to be known. The conventional approach to SNP discovery is based on resequencing alleles, which is particularly inefficient if the polymorphism levels are low. Prescreening for polymorphisms using, for example, single-strand conformation polymorphism (SSCP) [46] or Eco-TILLING [47], allows a reduction in sequencing costs, but these approaches are time-consuming, relatively expensive and rely on PCR. SFP detection in gene-expression microarray data allows parallel screening of a large proportion of all the organisms' gene space in one experiment. The stringency of SFP calls can also be adjusted for a particular application, that is, decreasing stringency will result in additional calls at the expense of higher false-positive rates.
Gene-expression levels are currently being treated as quantitative traits and transcript abundance variation is being mapped as quantitative trait loci (QTL) [48,49]. Incorporating SFP effects into calculations will improve accuracy of gene-expression studies and will facilitate correct assessment of allele-specific gene-expression differences. Furthermore, an SFP identified in a coding region of a gene that is differentially expressed in an allele-specific manner represents a marker linked to the regulatory regions of the gene, and as such may help distinguish between cis and trans effects in allele-specific gene expression [50-52].
Materials and methods
Affymetrix Barley1 GeneChip data
Affymetrix Barley1 GeneChip data was produced within an international collaborative project (A. Druka, G. Muehlbauer, I. Druka, R. Caldo, U. Baumann, N. Rostoks, A. Schreiber, R. Wise, T. Close, A. Kleinhofs, et al., unpublished work). Six tissue types were analyzed from two genotypes, Golden Promise (GP) and Morex (MX), with three type I replicates for a total of 36 arrays. We found that the GP genotype of one particular tissue replicate had a very high correlation with the three replicates from the comparable tissue from the MX genotype. We therefore re-assigned that replicate as genotype MX.
Genomic DNA from the wild barley Hordeum vulgare ssp. spontaneum (accession Mehola; arrays 1-3) and two morphologically diverse lines Oregon Wolfe Barley Recessive (arrays 4-6) and Oregon Wolfe Barley Dominant (arrays 7-9) [41] were prepared according to [53] and hybridized to the Affymetrix Barley1 GeneChip in triplicate according to standard methods for RNA.
SFP prediction in gene expression data
Raw .CEL files were background corrected and quantile normalized according to Bolstad et al. [36]. Subsequently, only the 11 Perfect Match (PM) features from each of 22,801 probe sets were fit with the following linear model
log(Ytgrp) = u + tissue + genotype + genotype × tissue + probe + error,
where Y is the background corrected normalized intensity of t (tissue), g (genotype), r (replicate), and p (probe) in a probe set. u is the mean probe intensity, while tissue has six states, and genotype has two states. The genotype by tissue effect accounted for tissue specific effects dependent on genotype. The residuals (22,801 probe sets × 11 probes = 250,811) from this model were fitted for a genotype effect at the probe level to reveal SFP using the Bioconductor package siggenes [37,36]. False discovery rates were estimated according to SAM [38,39] by performing 500 random permutations for RNA analysis or 100 permutations for genomic DNA analysis. The expected proportion of significantly different features (p0) was set to 0.95.
SFP confirmation by SNP analysis in silico
The EST unigene assembly 21 [40] that was used to produce the Affymetrix Barley1 GeneChip [31] contains 349,709 ESTs, of which 52,556 were derived from Morex (11 libraries) and 7,439 from Golden Promise (1 library). Library details are available from the HarvEST EST database [40]. HarvEST was used to identify a total of 1,758 unigene contigs containing both Morex and Golden Promise EST.
SFP confirmation by sequencing
192 primer pairs for 188 contigs were designed using Primer3 software [54] targeting 262 probes. Primers were supplied by Illumina. Single-stranded DNA template for PCR was synthesized from the same RNA samples that were used for hybridization to the Affymetrix GeneChips using SuperScript First-Strand Synthesis System for RT-PCR (Invitrogen). For each genotype, we combined 1 μg of RNA from each of the six tissue types and converted it to a single-stranded cDNA according to the manufacturer's recommendations using oligo(dT)12-18 as a primer. Single-stranded DNA was diluted fivefold and 2 μl was used for PCR amplification using gene-specific primers and HotStart Taq polymerase (Qiagen) with the following thermocycling parameters: 15 min 95°C, followed by 40 cycles of 30 sec 95°C, 45 s 60°C and 2 min 72°C, with a 10 min final extension at 72°C. PCR products were treated with ExoSAP-IT reagent (USB Corporation) and sequenced with the same primers using BigDye Terminator v3.1 cycle sequencing kit on an ABI PRISM 3700 sequencer (Applied Biosystems). Base-calling of ABI chromatograms and assembly of each unigene were done using Mutation Surveyor software (SoftGenetics, State College, PA). Synthetic chromatograms generated for all probe and EST unigene sequences were included in assemblies for comparison. Polymorphisms were called using Mutation Surveyor software and examined manually. SNP positions were recorded symmetrically, that is, a SNP in the central nucleotide of a 25-mer was in position 13, while SNPs in either first or twenty-fifth position was assigned position 1. Probes with multiple SNPs were allocated to a single group (Figure 3). Insertions and deletions were scored as polymorphisms, but the positions of polymorphisms were not scored.
SNP discovery in a random EST contig set
An SNP discovery project is currently underway in our laboratory which is based on resequencing alleles of barley genes in a set of eight barley lines, including Morex and Golden Promise [30]. The same EST unigene assembly that was used to design the Affymetrix Barley1 GeneChip was used in this SNP discovery study; PCR was carried out on genomic DNA templates, however. The Morex and Golden Promise sequences were reassembled separately as described for the SFP sequence set. Three hundred contigs representing essentially a random sample without any prior knowledge of polymorphisms were selected from this set on the basis that they included sequences from both genotypes; did not contain introns; sequences from both genotypes covered at least six Affymetrix Barley1 GeneChip probes for each probe set.
Acknowledgements
The gene-expression data for the barley cultivars Morex and Golden Promise was generated as part of an international collaborative project between barley researchers and is presented in a biological context in a separate manuscript (A. Druka, G. Muehlbauer, I. Druka, R. Caldo, U. Baumann, N. Rostoks, A. Schreiber, R. Wise, T. Close, A. Kleinhofs, A. Graner, A. Schulman, P. Langridge, K. Sato, P. Hayes, J. McNicol, D. Marshall, R. Waugh, personal communication). We thank those listed for pre-publication access to this dataset. Special thanks are due to Arnis Druka and Ilze Druka for assistance with microarray data and helpful discussions. We thank Yunda Huang for help and discussion with analysis. This project was funded by a BBSRC/SEERAD grant to R.W. and by start-up funds to J.O.B. from the University of Chicago.
Figures and Tables
Figure 1 Normalization of hybridization intensity profile of 25mer probes in a probe set. The y axis is background-corrected normalized log intensity and the x-axis shows the positions of the 11 features along the unigene. Black lines trace the Golden Promise arrays, while red trace the Morex arrays. Different line types differentiate tissues. Each panel illustrates normalization for one of the major sources of variation: probe effect; probe and tissue; probe and genotype; probe, genotype and tissue; probe, genotype, tissue and genotype by tissue. 100 such plots are available from [35].
Figure 2 Distribution of single-feature polymorphisms. (a) The observed d-statistics (y-axis) is plotted against the expected d-statistics (x-axis) as determined by permutations. 10,504 significant SFPs exceeding the threshold of 0.1% FDR are shown in green. (b) Histogram of d-statistics truncated at ± 10. Positive scores above the threshold 3.38 are Golden Promise SFPs, and negative scores below -3.37 are Morex SFPs.
Figure 3 Effect of SNP position on SFP identification. The positions of the SNP in 25mers are shown on the x-axis as distance from the edge in nucleotides (1 - 13 nucleotides). Multiple SNP category is provided separately by a single column. The y-axis indicates total number of probes identified for each SNP position. Each bar is divided into the SFP categories - mxSFP, nonSFP and gpSFP (see Table 2), and shows that more accurate SFP identification is made for SNPs that reside at internal sites. The number of 25mers in each category is shown within the bars.
Table 1 SFP false discovery rate (FDR) estimates in RNA and genomic DNA hybridization data
RNA hybridization: 17 Golden Promise 19 Morex, 6 tissues; SAM analysis for the two-class unpaired case assuming unequal variances; s0 = 0.0342 (the 5% quantile of the s values); number of permutations, 500. Mean number of falsely called genes is computed. Delta p0 Called False FDR
0.5 0.95 27,159 5,884 0.206
1.0 0.95 17,744 594 0.032
1.5 0.95 13,285 65 0.005
2.0 0.95 10,504 7 0.001
2.5 0.95 8,583 0 0.000
Genomic DNA hybridization three replicates three genotypes; SAM analysis for the multi-class case with three classes; s0 = 0.0123 (the 25 % quantile of the s values); number of permutations: 100; mean number of falsely called genes is computed. Delta p0 Called False FDR
1 0.95 4,017 2,073 0.47
2 0.95 1,728 583 0.31
3 0.95 1,090 258 0.22
4 0.95 789 139 0.16
5 0.95 631 86 0.13
The Bioconductor package siggenes [37,36] was used to derive SFP calls at various thresholds in the original data and randomly permuted data according to SAM [39]. Delta, the threshold; p0, the prior probability of the proportion of SFP in the null dataset; Called, the number of SFP at each threshold; False, the number of SFP in the mean permuted dataset.
Table 2 Single feature polymorphism (SFP) comparison with sequence-characterized SNPs
GeneChip
mxSFP nonSFP gpSFP
RNA sequence 5,301 240,307 5,203
MX 178 115 45 18
Non-polymorphic 2,200 27 2,045 128
GP 223 7 61 155
Chi-square = 2,049.2, df = 4, p-value = 0
The categories for SFP calls from RNA data are shown in columns: mxSFP, SFP in Morex; nonSFP, no SFP at the 0.1% FDR; gpSFP, SFP in Golden Promise. The categories of sequence-characterized probes are in rows: MX, polymorphism in Morex; non-polymorphic, no polymorphism between probe and any of the two genotypes; GP, polymorphism in Golden Promise. Intersections of the columns and rows indicate different combinations of sequence-verified polymorphisms and SFP.
Table 3 SFP discovery in individual tissue types
Tissue ALL COL CRO GEM LEA RAD ROO
Replicates (GP, MX) 18,18 3, 3 3, 3 3, 3 3, 3 3, 3 2, 4
Sensitivity 67% 52% 58% 63% 51% 62% 60%
False sequence polymorphism rate 40% 35% 34% 34% 34% 34% 35%
% variance explained 38% 30% 33% 37% 32% 31% 34%
Replicates indicate the number of arrays from each genotype analyzed for a given tissue type. Sensitivity is a percentage of correctly predicted SFP (270; Table 2) from the number of known sequence polymorphisms (401; Table 2). False sequence polymorphism rate is the percentage of predicted SFP that were found not to contain a DNA base-pair change. The % variance explained is that from a linear model fit of genotype (-1:MX; 0: no polymorphism; 1:GP) versus SFP d-statistic.
Table 4 Comparison of SFP prediction in RNA and genomic DNA hybridizations
GeneChip RNA
SFPs nonSFPs
GeneChip gDNA 10,504 240,307
SFPs 1,090 114 976
nonSFPs 24,9721 10,390 239,331
Chi-square = 107.28, df = 1, p-value = 3.863e-25
SFP and non-SFP probes in the gene-expression data are in columns, while the genomic data are in rows.
==== Refs
Borevitz JO Nordborg M The impact of genomics on the study of natural variation in Arabidopsis. Plant Physiol 2003 132 718 725 12805600 10.1104/pp.103.023549
Borevitz JO Chory J Genomics tools for QTL analysis and gene discovery. Curr Opin Plant Biol 2004 7 132 136 15003212 10.1016/j.pbi.2004.01.011
Rensink WA Buell CR Arabidopsis to rice. Applying knowledge from a weed to enhance our understanding of a crop species. Plant Physiol 2004 135 622 629 15208410 10.1104/pp.104.040170
Kwok PY Chen X Detection of single nucleotide polymorphisms. Curr Issues Mol Biol 2003 5 43 60 12793528
Hazen SP Kay SA Gene arrays are not just for measuring gene expression. Trends Plant Sci 2003 8 413 416 13678907 10.1016/S1360-1385(03)00186-9
Winzeler EA Castillo-Davis CI Oshiro G Liang D Richards DR Zhou Y Hartl DL Genetic diversity in yeast assessed with whole-genome oligonucleotide arrays. Genetics 2003 163 79 89 12586698
Winzeler E Richards D Conway A Goldstein A Kalman S McCullough M McCusker JH Stevens D Wodicka L Lockhart D Direct allelic variation scanning of the yeast genome. Science 1998 281 1194 1197 9712584 10.1126/science.281.5380.1194
Borevitz JO Liang D Plouffe D Chang HS Zhu T Weigel D Berry CC Winzeler E Chory J Large-scale identification of single-feature polymorphisms in complex genomes. Genome Res 2003 13 513 523 12618383 10.1101/gr.541303
Kennedy GC Matsuzaki H Dong S Liu WM Huang J Liu G Su X Cao M Chen W Zhang J Large-scale genotyping of complex DNA. Nat Biotechnol 2003 21 1233 1237 12960966 10.1038/nbt869
Dong S Wang E Hsie L Cao Y Chen X Gingeras TR Flexible use of high-density oligonucleotide arrays for single- nucleotide polymorphism discovery and validation. Genome Res 2001 11 1418 1424 11483583 10.1101/gr.171101
Ronald J Akey J Whittle J Smith E Yvert G Kruglyak L Simultaneous genotyping, gene expression measurement, and detection of allele-specific expression with oligonucleotide arrays. Genome Res 2005 15 284 291 15687292 10.1101/gr.2850605
Goff SA Ricke D Lan TH Presting G Wang R Dunn M Glazebrook J Sessions A Oeller P Varma H A draft sequence of the rice genome (Oryza sativa L. ssp. japonica). Science 2002 296 92 100 11935018 10.1126/science.1068275
Yu J Hu S Wang J Wong GK Li S Liu B Deng Y Dai L Zhou Y Zhang X A draft sequence of the rice genome (Oryza sativa L. ssp. indica). Science 2002 296 79 92 11935017 10.1126/science.1068037
Bennetzen J Freeling M The unified grass genome: synergy in synteny. Genome Res 1997 7 301 306 9110169
Caldwell KS Langridge P Powell W Comparative sequence analysis of the region harboring the hardness locus in barley and its colinear region in rice. Plant Physiol 2004 136 3177 3190 15466237 10.1104/pp.104.044081
Brunner S Keller B Feuillet C A large rearrangement involving genes and low-copy DNA interrupts the microcollinearity between rice and barley at the Rph7 locus. Genetics 2003 164 673 683 12807788
Bennetzen J Ramakrishna W Numerous small rearrangements of gene content, order and orientation differentiate grass genomes. Plant Mol Biol 2002 48 821 827 11999852 10.1023/A:1014841515249
FAO
Stanca M von Bothmer R, van Hintum T, Knuepffer H, Sato K Diversity in abiotic stress tolerance. Diversity in Barley 2003 Amsterdam: Elsevier Science 179 199
Fischbeck G von Bothmer R, van Hintum T, Knuepffer H, Sato K Diversification through breeding. Diversity in Barley 2003 Amsterdam: Elsevier Science 29 52
Bennett M Leitch I Nuclear DNA amounts in angiosperms. Annls Bot Lond 1995 76 113 176
Jakob SS Meister A Blattner FR The considerable genome size variation of Hordeum species (poaceae) is linked to phylogeny, life form, ecology, and speciation rates. Mol Biol Evol 2004 21 860 869 15014163 10.1093/molbev/msh092
Rostoks N Park YJ Ramakrishna W Ma J Druka A Shiloff BA SanMiguel PJ Jiang Z Brueggeman R Sandhu D Genomic sequencing reveals gene content, genomic organization, and recombination relationships in barley. Funct Integr Genomics 2002 2 51 59 12021850 10.1007/s10142-002-0055-5
NCBI dbEST summary
Graingenes
Flint-Garcia SA Thornsberry JM Buckler ES Structure of linkage disequilibrium in plants. Annu Rev Plant Biol 2003 54 357 374 14502995 10.1146/annurev.arplant.54.031902.134907
Rafalski A Applications of single nucleotide polymorphisms in crop genetics. Curr Opin Plant Biol 2002 5 94 100 11856602 10.1016/S1369-5266(02)00240-6
Kota R Rudd S Facius A Kolesov G Thiel T Zhang H Stein N Mayer K Graner A Snipping polymorphisms from large EST collections in barley (Hordeum vulgare L.). Mol Genet Genomics 2003 270 24 33 12938038 10.1007/s00438-003-0891-6
Kota R Varshney RK Thiel T Dehmer KJ Graner A Generation and comparison of EST-derived SSRs and SNPs in barley (Hordeum vulgare L.). Hereditas 2001 135 145 151 12152327 10.1111/j.1601-5223.2001.00145.x
Rostoks N Cardle L Svensson J Walia H Rodriguez E Wanamaker S Hedley P Liu H Ramsay L Russell J Single nucleotide polymorphism mapping of the barley genes involved in abiotic stresses. Czech J Genet Plant Breed 2004 40 52
Close TJ Wanamaker SI Caldo RA Turner SM Ashlock DA Dickerson JA Wing RA Muehlbauer GJ Kleinhofs A Wise RP A new resource for cereal genomics: 22K barley GeneChip comes of age. Plant Physiol 2004 134 960 968 15020760 10.1104/pp.103.034462
ArrayExpress
Brazma A Parkinson H Sarkans U Shojatalab M Vilo J Abeygunawardena N Holloway E Kapushesky M Kemmeren P Lara GG ArrayExpress - a public repository for microarray gene expression data at the EBI. Nucleic Acids Res 2003 31 68 71 12519949 10.1093/nar/gkg091
BarleyBase
Naturalvariation
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 19 185 193 12538238 10.1093/bioinformatics/19.2.185
Bioconductor
Schwender H Krause A Ickstadt K Comparison of the Empirical Bayes and the Significance Analysis of Microarrays 2003 Technical Report. SFB 475: Dortmund, Germany: University of Dortmund
Tusher VG Tibshirani R Chu G Significance analysis of microarrays applied to the ionizing radiation response. Proc Natl Acad Sci USA 2001 98 5116 5121 11309499 10.1073/pnas.091062498
HarvEST
Costa JM Corey A Hayes PM Jobet C Kleinhofs A Kopisch Obusch A Kramer SF Kudrna D Li M Riera Lizarazu O Molecular mapping of the Oregon Wolfe Barleys: A phenotypically polymorphic doubled-haploid population. Theor Appl Genet 2001 103 415 424 10.1007/s001220100622
Bancroft I Insights into cereal genomes from two draft genome sequences of rice. Genome Biol 2002 3 reviews1015.1 1015.3 12093379 10.1186/gb-2002-3-6-reviews1015
Bennetzen JL Coleman C Liu R Ma J Ramakrishna W Consistent over-estimation of gene number in complex plant genomes. Curr Opin Plant Biol 2004 7 732 736 15491923 10.1016/j.pbi.2004.09.003
Vos P Hogers R Bleeker M Reijans M van de LT Hornes M Frijters A Pot J Peleman J Kuiper M AFLP: a new technique for DNA fingerprinting. Nucleic Acids Res 1995 23 4407 4414 7501463
Zhang H Sreenivasulu N Weschke W Stein N Rudd S Radchuk V Potokina E Scholz U Schweizer P Zierold U Large-scale analysis of the barley transcriptome based on expressed sequence tags. Plant J 2004 40 276 290 15447653 10.1111/j.1365-313X.2004.02209.x
Andersen PS Jespersgaard C Vuust J Christiansen M Larsen LA Capillary electrophoresis-based single strand DNA conformation analysis in high-throughput mutation screening. Hum Mutat 2003 21 455 465 12673787 10.1002/humu.10169
Comai L Young K Till BJ Reynolds SH Greene EA Codomo CA Enns LC Johnson JE Burtner C Odden AR Efficient discovery of DNA polymorphisms in natural populations by Ecotilling. Plant J 2004 37 778 786 14871304 10.1111/j.0960-7412.2003.01999.x
Brem RB Yvert G Clinton R Kruglyak L Genetic dissection of transcriptional regulation in budding yeast. Science 2002 296 752 755 11923494 10.1126/science.1069516
Schadt EE Monks SA Drake TA Lusis AJ Che N Colinayo V Ruff TG Milligan SB Lamb JR Cavet G Genetics of gene expression surveyed in maize, mouse and man. Nature 2003 422 297 302 12646919 10.1038/nature01434
Cowles CR Hirschhorn JN Altshuler D Lander ES Detection of regulatory variation in mouse genes. Nat Genet 2002 32 432 437 12410233 10.1038/ng992
Morley M Molony CM Weber TM Devlin JL Ewens KG Spielman RS Cheung VG Genetic analysis of genome-wide variation in human gene expression. Nature 2004 430 743 747 15269782 10.1038/nature02797
Wittkopp PJ Haerum BK Clark AG Evolutionary changes in cis and trans gene regulation. Nature 2004 430 85 88 15229602 10.1038/nature02698
Arabidopsis methods
Rozen S Skaletsky H Primer3 on the WWW for general users and for biologist programmers. Methods Mol Biol 2000 132 365 386 10547847
| 15960806 | PMC1175974 | CC BY | 2021-01-04 16:05:39 | no | Genome Biol. 2005 May 11; 6(6):R54 | utf-8 | Genome Biol | 2,005 | 10.1186/gb-2005-6-6-r54 | oa_comm |
==== Front
Genome BiolGenome Biology1465-69061465-6914BioMed Central London gb-2005-6-6-r551596080710.1186/gb-2005-6-6-r55SoftwareRefinement and prediction of protein prenylation motifs Maurer-Stroh Sebastian [email protected] Frank [email protected] IMP - Research Institute of Molecular Pathology, Dr. Bohr-Gasse 7, A-1030 Vienna, Austria2005 27 5 2005 6 6 R55 R55 17 1 2005 22 3 2005 20 4 2005 Copyright © 2005 Maurer-Stroh and Eisenhaber; 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.
Three prenylation motif predictors are presented that allow discrimination between proteins that are unique substrates of farnesyltransferase (FT) and those that can be alternatively processed by geranylgeranyltransferase I (GGT1).
We refined the motifs for carboxy-terminal protein prenylation by analysis of known substrates for farnesyltransferase (FT), geranylgeranyltransferase I (GGT1) and geranylgeranyltransferase II (GGT2). In addition to the CaaX box for the first two enzymes, we identify a preceding linker region that appears constrained in physicochemical properties, requiring small or flexible, preferably hydrophilic, amino acids. Predictors were constructed on the basis of sequence and physical property profiles, including interpositional correlations, and are available as the Prenylation Prediction Suite (PrePS, ) which also allows evaluation of evolutionary motif conservation. PrePS can predict partially overlapping substrate specificities, which is of medical importance in the case of understanding cellular action of FT inhibitors as anticancer and anti-parasite agents.
==== Body
Rationale
Prenylation refers to the posttranslational modification of proteins with isoprenyl anchors [1-3]. These lipid moieties are typically involved in mediating not only protein-membrane but also protein-protein interactions. Three eukaryotic enzymes are known to catalyze the lipid transfer. The first two, farnesyltransferase (FT) and geranylgeranyltransferase 1 (GGT1), recognize the so-called CaaX box in the carboxy termini of substrate proteins and attach farnesyl (15-carbon polyisoprene) or geranylgeranyl (20-carbon polyisoprene), respectively, to a required and spatially fixed cysteine in that motif. The third enzyme, geranylgeranyltransferase 2 (GGT2 or RabGGT) recognizes the complex [4] of Rab GTPase substrate proteins with a specific Rab escort protein (REP) to attach one or two geranylgeranyl anchors to cysteines in a more flexible but also carboxy-terminal motif.
The CaaX box was initially understood to consist of a cysteine (C), followed by two aliphatic residues (aa) and a terminal residue (X) that would direct modification by either FT or GGT1, but newly found substrates and kinetic studies of mutated substrate peptides and enzyme inhibitors have shown that the motif recognized by the enzymes appears to be more flexible [2]. Furthermore, the determination of preference for FT or GGT1 is more complex and a function of the overall sequence context rather than specific amino acids at single positions. Whereas GGT2 appears to be specific to Rab GTPases as substrates, the recognition mechanism is not well understood. Overlapping substrate specificities between all three prenylating enzymes further complicate the understanding of the lipid modification process [5,6].
An unsolved problem so far is accounting for the complexity of the prenylation substrate recognition motifs in theoretical models in order to identify substrate proteins from their amino-acid sequence. No available method has been able to selectively assign the correct modifying enzyme, which determines the types and number of lipid anchors. The high probability of motifs similar to the small CaaX box occurring by chance is a general problem that has so far prohibited large-scale proteome analyses [7]. We describe here a method that aims to model the substrate-enzyme interactions on the basis of refinement of the recognition motifs for each of the prenyltransferases. The Prenylation Prediction Suite (PrePS) selectively assigns the modifying enzyme to predicted substrate proteins and sensitively filters out false-positive predictions based on the general methodology that has already been applied successfully for the prediction of glycosylphosphatidylinositol (GPI) anchors [8], myristoylation [9] and PTS1 peroxisomal targeting [10].
Known substrates and their motif-compliant homologs as learning sets
The first task consists of collecting sequences that are known substrates for the respective enzymes. Typically, a good starting point is the Swiss-Prot database [11]. However, according to earlier experience with annotation inaccuracies [12], any annotated experimental evidence has to be confirmed by following up all the related literature sources. As newly available data can be missing in the Swiss-Prot annotation, the searches have also to be extended to non-Swiss-Prot proteins. In most cases, the annotations for prenylation in Swiss-Prot are assigned by similarity to only a few entries with experimental validation. A major concern is the annotation of the correct anchor type attached to FT and GGT1 substrates, which could previously only tentatively be estimated without experimental data. This includes several entries with overall sequence similarity to a verified prenylated protein but totally different carboxy-terminal motifs. Given that single mutations can abolish recognition or switch enzyme specificities [13] and that not all homologs of lipid-modified proteins necessarily have to share the same modification type or membrane attachment factor (MAF) [14], entries with annotations only by similarity should not be included without critical consideration in a learning set.
Unfortunately, such justified concerns dramatically lower the amount of data in the learning set. However, because of earlier interest in developing peptide-based inhibitors of FT and GGT1 as anticancer treatments, the kinetics of the enzymes with various tetrapeptide substrates already modified with lipid anchors by the enzymes have been measured [15]. Hence, a protein homologous to a verified prenylated protein can be included in the learning set if its CaaX box has already been shown to interact productively with one of the prenyltransferases at least as a tetrapeptide.
However, possession of a valid CaaX box might not be a sufficient selection criterion. Typically, short terminal sequence motifs are connected to the rest of the protein by a linker region that experiences only limited constraints on specific amino acids per position but often has a compositional bias towards small or hydrophilic amino acids in connecting sequence stretches [16]. This property is found in a preliminary assembly of verified FT and GGT1 substrates and has been confirmed in the actual learning set for up to 11 residues upstream (amino-terminal) of the cysteine in the CaaX box (see below). Hence, learning-set sequences should also not violate the physicochemical properties constraining the sequence stretch amino-terminal to the CaaX box.
Taking account of the considerations above, the following procedure has been applied to obtain conservative and reliable learning sets of FT and GGT1 substrates. First, a literature search for known prenylated proteins and valid tetrapeptides (see [17]). Second, BLASTP [18] with an E-value threshold of 0.005 starting with known prenylated proteins against the National Center for Biotechnology Information (NCBI) nonredundant database to find homologs and cluster all collected sequences into groups of homologous proteins using the Markov-chain clustering algorithm (MCL) [19]. Third, check the validity of all CaaX boxes with experimental evidence for at least tetrapeptides. Fourth, check compliance with the physical properties of the full motif (including linker) by applying a preliminary predictor based on corrected Swiss-Prot entries in a similar style as described here (penalizing deviations from the physical property landscape of the motif).
This resulted in learning sets of 692 FT and 486 GGT1 substrates, respectively (see [17]). Among the FT substrates, 31 artificial constructs or mutations of naturally occurring sequences that have been shown to be processed by FT have also been included. Prenylation by GGT2 follows totally different mechanistic requirements than FT and GGT1 and will be treated separately after the sections about CaaX prenylation.
Refinement of the CaaX box motif descriptions
Compositional analysis of residue frequencies at single motif positions reveals that major restrictions to specific amino acid types exist only for positions within the CaaX box (see sequence logos in Figure 1). The previously reported preferences for aliphatic residues at positions +1 and +2 (the aa in CaaX) were recovered, but there is a clear tendency for other residue types to also be allowed, especially at position +1 (the first a in CaaX). Correlation analysis of residue frequencies at single motif positions with amino-acid property scales [20,21] can quantify the conservation of a physical property pattern (see Materials and methods). Although correlations higher than 0.6 can only be obtained for aliphatic property at position +2 (FT: 0.85, GGT1: 0.87), the average aliphatic property at position +1 within both FT and GGT1 learning sets still appears elevated when compared to an average calculated from the carboxy-termini of the nonredundant UniRef50 database [22] (see physical property profile in Figure 1). Similarly, there are correlations at position +2 and deviations from the UniRef50 average at position +1 for a property describing preference for extended conformations (see Tables 1 and 2). This appears to be best explained by the need to have the final peptide part in extended conformation rather than coiled or helical in order to fit into the binding pocket, as can be seen in the resolved structures of prenyltransferases with their substrate peptides [23].
The major difference between FT and GGT1 substrates remains at position +3 (the X in CaaX). Whereas a broad variety of residues are allowed in motifs recognized by FT (including several substrates with leucine at +3), mainly leucine and methionine appear to be preferred by GGT1 in agreement with experimental evidence [13]. Interestingly, position +3 correlates (FT, 0.7; GGT1, 0.8) with a physical property that measures membrane-buried preference parameters (see Tables 1 and 2). This feature does not seem to be important to support membrane interaction at a later stage for the protein, as the three carboxy-terminal residues (-aaX) are often cleaved off in a further processing step after attachment of the anchor [24]. However, hydrophobicity and volume of position +3 appear important for interaction with the binding pocket because of the rather lipophilic character of the latter (isoprenyl anchor on one side and hydrophobic residues on the others). The importance of position +3 for specificity between FT and GGT1 is further strengthened by differing conservation of residues in the binding pockets of the respective enzymes (Figure 2). Not surprisingly, the whole region of the binding pocket harboring the end of the prenylpyrophosphate (geranylgeranyl [C20] is one isoprene unit longer than farnesyl [C15]) and the X of the CaaX box (position +3) appear to comprise the major differences in residue conservation (Figure 2).
Using the Fisher criterion (see Materials and methods), interpositional correlations of residue sizes within positions +1, +2 and +3 (the carboxy-terminal three residues of the CaaX box that are buried in the binding pocket) from both FT and GGT1 substrates have been identified. Often, when a very large residue occurs at specific positions, neighboring residues compensate to obey the overall physicochemical constraints (for example, size limitation) in the binding pocket. Similarly, compensatory effects appear to exist regarding hydrophobicity between positions +1 and +3 in FT and between +1, +2 and +3 in GGT1 substrates (see Tables 1 and 2). Compensatory effects also seem responsible for the toleration of even large positively charged residues at positions +1 or +2, if the other residues are small enough to accommodate the whole peptide in the binding pocket. On the other hand, negative charges are apparently incompatible with the substrate recognition motif at these positions.
Extension of the CaaX prenylation motif by a flexible linker region
While the requirement for specific amino acids at single positions appears to be marginal outside of the CaaX box, physicochemical constraints that extend up to 11 residues amino-terminal from the modified cysteine can be found (Figure 1, Tables 1 and 2). At position -1 of the motif, there begins a pronounced tendency for residues with either small or flexible hydrophilic side chains. GGT1 especially appears to prefer amino acids like serine or lysine at this position. In general, GGT1 substrates have a higher number of lysines within positions -1 and -7 compared with the FT substrates.
The hydrophilic linker region with correlations over multiple positions to several hydrophobicity- and flexibility-related property scales might be required to allow accessibility of the carboxy terminus for the lipid-attaching enzymes. Indeed, in several resolved structures of in vivo prenylated GTPases, secondary structural elements such as helices that stabilize the fold of the protein are typically found only at the amino-terminal side of that linker region (beginning of helix at positions -12 (PDB identifier 1FTN), -13 (PDB 1MH1), -15 (PDB 1AM4), -12 (PDB 1A4R)). In the structure of a G protein gamma subunit, the linker region also appears to be extended and wrapped around the beta subunit in the heterotrimeric G protein signaling complex (PDB 1GG2). It needs to be emphasized that the linker region must not necessarily be in an unstructured conformation after the anchor has been attached (see also carboxy-terminal helix in structure PDB 1F5N of human 67 kDa guanylate binding protein 1 [25]), as folding back or lipid-mediated interaction with other proteins or membranes can also induce changes in the three-dimensional structure of the linker region. However, there appears to be a requirement for the ability to easily unfold/fold into flexible and more extended conformations that allow the carboxy terminus to be accessed and modified by the prenyltransferases. It is noteworthy that this length estimation of a flexible, hydrophilic linker is consistent with earlier findings in the GPI anchor [21], myristoylation [12] and PTS1 targeting [26] motifs. Hence, the actual motif length of substrates for CaaX prenylation appears longer than previously thought (total 15 residues = 4 CaaX + 11 linker).
Prediction function and validation
Following the approach already applied to the prediction of GPI and myristoyl anchors and PTS1-mediated targeting [8-10], a scoring function measuring compliance with the prenylation motif separately for the enzymes FT and GGT1, respectively, has been constructed (see Materials and methods). In brief, the composite prediction function S consists of a term Sprofile scoring a query sequence against the redundancy-corrected profile of the learning-set sequences and another term Sppt that penalizes deviation from the physicochemical motif requirements.
S = Sprofile + Sppt
The term Sprofile distinguishes the three positions +1, +2 and +3 of the CaaX box as well as the linker region (-1 to -11). Sppt comprises a sum of terms that are constructed from the physical property requirements for FT and GGT1 substrates that were outlined in the section describing the motif refinement (and listed in Tables 1 and 2 together with their rationale for inclusion in Sppt).
The threshold for a query protein to be a predicted farnesylation or geranylgeranylation target by FT or GGT1, respectively, is set to include all sequences in the learning set. Hence, the self-consistencies or upper bounds of sensitivities of the FT and GGT1 predictors are 100%. Additionally, the robustness of the method has been cross-validated in jackknife tests (see Materials and methods). In the cross-validation over the complete scoring function, the rates of finding known substrates after excluding them and their close homologs from the learning procedure (and, therefore, lower bounds for sensitivities) were 92.6% for FT and 98.6% for GGT1, respectively.
As required for a good predictor [16], the scores are translated into probabilities of false-positive prediction. For this purpose, a sigmoidal function (analytically based on the extreme-value distribution) is fitted to the distribution of score values calculated from non-prenylatable proteins (see Materials and methods). The general probabilities of false-positive prediction (that complement the specificities to 100%) are estimated to be 0.11% for the FT and 0.02% for the GGT1 predictor, respectively.
Capability to distinguish FT and GGT1 substrates
Previously, the assignment of CaaX box substrate proteins to either FT or GGT1 has been based mainly on the identity of the final residue in the motif (position +3) where FT allows several amino-acid types and GGT1 clearly prefers leucine [13,27]. This view has not changed but it has become clear that several substrates with leucine at position +3 can also be modified (if only to a lesser extent) by FT and not only GGT1. For example, in vitro studies have shown that motifs like CVIL, CVLL, CAIL and CCIL (single-letter amino-acid code) are valid for FT as well [28]. Mutation of the CVIA motif of yeast A-factor to CVIL results in geranylgeranylated as well as farnesylated proteins in vivo [29]. Also, RhoB (with a CKVL motif) is known to be both farnesylated and geranylgeranylated in vivo [30]. Similarily, substrate proteins ending with phenylalanine, such as the CVIF of R-Ras2/TC21, are not specific to either enzyme and can be substrates to FT and GGT1 [31].
In the same way that FT can accept CaaX box motifs ending in leucine and phenylalanine, GGT1 appears to tolerate methionine at this position, which was previously thought to direct farnesylation. This has important consequences in the case of the oncoprotein K-Ras (in variants with CVIM and CIIM motifs) which becomes geranylgeranylated in vivo when farnesyltransferase is inhibited [32].
As we have experienced with our earlier predictors for myristoylation and PTS1 targeting, we find even some correlations of the prediction scores with experimentally measured substrate-enzyme affinities. Interestingly, the scores of the GGT1 predictor give better agreement with the experimental data when divided by 3, in agreement with a threefold lower in vivo activity of GGT1 compared to FT [5]. To estimate the capability of the FT and GGT1 predictors to model the overlapping but distinct substrate specificities, we analyzed a set of heterogeneous substrate motifs that have been measured under the same experimental conditions for their affinities to either FT or GGT1 [5] and we tried to correlate these experimental data with our prediction scores. The set of motifs (CVLS, CIIS, CIIC, CVLF, CVIM, CAIM, CAIV, CAII, CAIL, CVVL, CIIL, and CTIL) contains a large fraction of examples that have been previously shown to be cross-reactive between FT and GGT1 or where the assignment based on simple heuristics depending on hydrophobicity of the final residue fails. In Figure 3, we have plotted the difference of predicted FT and GGT1 scores against the difference of experimentally measured logarithmic affinities for FT and GGT1. A correlation of 0.74 indicates that the theoretical interaction model implemented in the prediction function at least semi-quantitatively resembles the relative substrate specificities between FT and GGT1.
Prediction of prenylation by GGT2
Unlike FT and GGT1, substrate recognition by GGT2 is less dependent on strictly defined carboxy-terminal motifs, but on the complex formation of the substrate with an escort protein [4]. As illustrated in Figure 4, the substrate-escort protein complex then binds to GGT2 (consisting of the alpha and beta subunit typical of prenyltransferases) and, thereby, positioning the flexible substrate carboxy terminus towards the site of modification. Typically, the carboxy-terminal arrangement of cysteines is -XXXCC, -XXCXC, -XXCCX, -XCCXX or -CCXXX and, if available, both cysteines in such a motif will be geranylgeranylated. Currently, only the prenylation of Rab GTPases [33] with the help of Rab escort proteins (REP; two copies in higher organisms, otherwise only one copy) is known for the enzyme GGT2 which is, therefore, also called Rab geranylgeranyltransferase. Reports of lipid modification of fungal casein kinase I apparently represent carboxy-terminal palmitoylation [34] rather than the earlier postulated GGT2 prenylation [35].
Rab proteins are small GTPases (around 60 different have been identified in humans) [36] that share the general fold of the Ras superfamily as well as conserved residues in the nucleotide-binding site. Distinct motifs have been identified that are specific to the Ras, Rho, or Rab families [37]. By virtue of contributing to the binding site of Rabs with their REP, the Rab-specific F3F4 motif can be indirectly used to distinguish possible GGT2 substrates within the Ras superfamily (see sequence logos in Figure 4). However, the REP interaction motif (Rab F3F4) alone could be too short (13 residues) to allow highly sensitive large-scale database scans with thresholds that recognize the learning set (100% self-consistency requires a bit score greater than 5). Interestingly, a search with the final predictor against NCBI's nonredundant database finds only 34 hits with the F3F4 region alone that do not represent Rab proteins or their folds. To avoid these false positives, the hit to the overall alignment of Rab proteins with HMMer [38] (E-value < 0.1) is applied as additional prediction criterion to simulate recognition of the correct fold of related sequences.
Two alignments (F3F4 region and full length) were therefore constructed and after removal of entries with a maximal redundancy of 90% identity over the whole sequence length (117 of 179 entries annotated in Swiss-Prot remaining), hidden Markov models (HMMs) were created and calibrated. The choice of this methodology for the GGT2 prediction was strongly influenced by the fact that the HMMer [38] algorithm is well established in conservatively detecting fold homologies for globular domains at the sequence level. The final GGT2 prediction algorithm checks the carboxy termini for cysteines (at least one cysteine among the five last residues) and parses the HMMer outputs to combine the searches for final results. Estimates of false-positive prediction can be derived from the HMMer E-values.
PrePS: Webinterface and EvOluation
The three tools to predict lipid modification by FT, GGT1 and GGT2 are available as Prenylation Prediction Suite (PrePS), which is accessible online [39]. Users can submit their query sequences to all three or selections of the single predictors. Details of the profile and physical property terms of the scoring function are provided and can also be used to check and rationalize whether and why certain query sequences or artificial constructs intended for membrane targeting might be less suitable prenylation targets. Additionally, an option is provided that allows the user to retrieve homologs of the query protein from NCBI's nonredundant database using BLASTP and automatically annotates them with their respective PrePS results. From the scores for the different predictors (left screenshot in Figure 5) as well as the alignment of the carboxy termini of homologous sequences (right screenshot in Figure 5), the evolutionary motif conservation can be evaluated (evOluation) and used for further rationalization of the biological importance of the predicted motif.
Comparison with alternative methods
Until now, the only available tool to predict protein prenylation has been the Prosite [40] search with the pattern PS00294, which is also used in the PSORT II software [41]. However, this method can neither predict prenylation by GGT2 nor can it distinguish between modifications by FT or GGT1 and, hence, the attached anchor type. During preparation of this paper, an excellent study by Beese, Casey and colleagues [23] has been published that tries to define rules for substrate selectivity by crystallographic analysis of FT and GGT1 complexed with eight cross-reactive substrates. These detailed descriptions of the binding-pocket interactions of a few selected substrate peptides are in good agreement with the motif characteristics identified in this work. While the information gathered from the structural analysis exceeds the capability of any other purely theoretical method to judge interaction for the specific resolved enzyme-substrate pairs, it is difficult to generalize an interaction model from such a small dataset only on the basis of amino-acid constraints at single motif positions. Hence, applying these rules to a more restrictive Prosite-style pattern fails to identify around 30% of substrates experimentally verified in tetrapeptide interaction assays. When taking a closer look at known substrates that are not recognized by the rules of Beese, Casey and colleagues [23] it becomes apparent that this is mainly due to only a few factors. These are the exclusion of leucine at position +3 for alternative FT substrates (known example CKVL of RhoB), the exclusion of phenylalanine at position +3 for alternative FT substrates (known example CVIF of R-Ras2/TC21), the exclusion of glutamine at position +2 for FT substrates (known example serine/threonine kinase 11 or LKB1 with the motif CKQQ) and the exclusion of methionine at position +3 for alternative GGT1 substrates (known example CVIM of K-Ras). In addition, the rules of Beese, Casey and colleagues [23] assign isoleucine and valine at position +3 to GGT1 but not FT substrates. However, these two amino acids were shown to be valid for both FT and GGT1, with at least comparable affinities [13].
The inadequacy of the Beese, Casey and colleagues [23] motif in finding true-positive examples could be counteracted by loosening the motif description, as is already the case in the original Prosite entry PS00294, which nevertheless fails to predict known substrates with glutamine (LKB1) or proline (hepatitis delta antigen) at position +2. However, any reduction in motif stringency concomitantly results in a dramatic increase in the number of false-positive predictions. Table 3 compares typical prediction parameters for the different methods, if applicable. Neither the old nor an adjusted Prosite pattern can compete with the performance of PrePS in finding true substrates while, at the same time, only having a minimal number of false positives. The short Prosite patterns also do not take into account the linker region preceding the CaaX box, which is not defined by clear amino-acid type preferences but rather by general physicochemical property restrictions. The answers of Prosite-style predictions are only binary (yes/no), whereas PrePS gives continuous scores that can be split into interpretable motif-region contributions and that are shown to correlate with experimentally measured relative substrate affinities for FT or GGT1, respectively. Furthermore, only PrePS includes prediction of prenylation by GGT2 and provides an evaluation of evolutionary conservation of the prenylation motif among homologs of the query sequence.
Medical implications and prediction examples
Farnesyltransferase inhibitors (FTIs) have been developed to prevent prenylation of oncogenic Ras proteins and are currently undergoing phase II and III clinical trials [42]. While FTIs have been suggested also to target parasitic diseases [24,43], their efficacy as cancer treatments has been found to be ambivalent in respect of different cancer types. This could be due to the alternative prenylation of oncogenic proteins by GGT1 under FT inhibition, such as K-Ras, in contrast to the total inhibition of prenylation for unique FT substrates, such as H-Ras [2,44]. Identifying these two types of substrate behavior is critical for understanding FTI action as well as identifying their real cellular targets [45,46]. One of the applications of PrePS is in the distinction of substrates that are specific to FT (FTI target) or GGT1 or that are modified by both (less affected by FTIs).
We would like to mention here one example prediction of PrePS for a protein that would be a candidate for a previously unknown FTI target. The human nucleosome assembly protein I-like protein [47] (NAP1-like (GenBank:NP_004528)) has a CKQQ farnesylation motif that is further retained in mouse, rat, frog, fish, fungi and plants, as predicted by PrePS. This taxonomically widespread evolutionary conservation would rather indicate a relevance of the lipid anchor for the function of this protein, which is part of a family involved in transcriptional activation and chromatin formation, including histone binding [48] and nucleocytoplasmic shuttling [49]. The lack of the ability to be alternatively prenylated by GGT1 and, hence, being a unique FT substrate and putative FTI target, is also conserved in the other organisms, possibly pointing to the importance of the specific farnesyl anchor length. It should be noted that this protein is not predicted by the Prosite pattern PS00294 nor by the pattern derived from the rules of a few substrate-enzyme structures [23], but there exist other experimentally verified examples where the same CaaX box motif CKQQ has been shown to be farnesylated (yeast Pex19p [50] and human serine/threonine kinase 11 [51]).
While this paper was in preparation, farnesylation of the NAP1-like protein has been suggested experimentally through a special tagging and purification technique [52], giving support to the PrePS prediction. The same analysis, however, also suggests farnesylation of annexin A2 (GenBank accession number P07355) terminating in a CGGDD motif, which is not at all predicted by PrePS as it is mechanistically unlikely to be processed by farnesyltransferase. Another rather surprising prediction resulting from the tagging experiment is the farnesylation of Rab21 (Q9UL25), which has a double cysteine motif followed by three additional residues (CCSSG) which, at least formally, resembles a CaaX box. Rab proteins with CaaX boxes such as Rab5 (CCSN), Rab8 (CVLL/CSLL), Rab11 (CQNI) and Rab13 (CSLG) are usually modified by GGT2 in vivo [6,53,54] but Rab8 and Rab11 were shown also to be modified by GGT1 and FT in vitro [6,55]. PrePS predicts Rab21 to be geranylgeranylated by GGT2, but the prediction limit for farnesylation is not missed by far. The evOluation shows that the Rab21 orthologs in Xenopus (GenBank AAH60498.1) and Drosophila (AAH60498.1) share the double cysteines but their motif is different and shorter by one residue, pointing to a higher importance of the conservation of the cysteine doublet than the rest of the motif. The evOluation, furthermore, shows that Rab5 is the most closely related prenylated Rab-family member. Interestingly, both cysteines in the CCSN CaaX box motif of Rab5 were shown not only to be geranylgeranylated by GGT2 in vivo but are also required for proper localization and function of the GTPase [54]. Hence, a similar scenario for the two cysteines of the Rab21 prenylation motif cannot be excluded.
A complete analysis of large-scale predictions of prenylated proteins ranked by functional importance as estimated by evolutionary motif conservation and medical implications will be published in a follow-up work.
Materials and methods
Correlation of positional amino-acid frequencies with physical property scales
We identified physicochemical requirements for each motif position by correlating 20-dimensional vectors filled with the positional frequencies of occurrence of the 20 amino-acid types in the carboxy-terminally aligned learning set with a library of over 650 amino-acid physical properties [20,21]. This has been done over a largest subset of the learning set with removed redundancy of greater than 40% identity in the last 30 positions (nonredundant learning set = nrLS) and over positional vectors filled with frequencies derived from the profile (= prof) that has been corrected for redundancy with the position-specific independent counts (PSIC) method [56]. Such correlations have been estimated previously [12] to be significant for confidence levels α = 0.0025 and α = 0.001 if the values are greater than 0.62 and 0.7, respectively.
Fisher criterion to find interpositional correlations
The Fisher ratio F of the sum of variances of single positions with the variance over multiple positions for pairs and triplets of positions is calculated, allowing gaps of up to two residues between pairs.
The F values for probabilities p = 0.05 are taken as thresholds for evaluating significance of interdependence of physical properties between motif positions. These are 1.288 for the FT and 1.355 for the GGT1 nonredundant learning set, respectively.
Enzyme-specific binding pocket residue conservation
To identify regions in the FT and GGT1 binding pockets that are characteristic for the respective enzyme, we analyzed the pattern of residue conservation in vicinity to the CaaX substrate peptide and the prenylpyrophosphate (Figures 2 and 5). Alignments were created with FT and GGT1 beta subunits of a diverse subset of organisms (see legend of Figure 6) using ClustalX [57]. The conservation of alignment positions was measured using al2co [58] with the Henikoff-weighted variance-based options. The enzyme-specific conservation was calculated as conservation values in the FT or GGT1 alignments minus conservation values in a joined FT+GGT1 alignment (Figure 6) and then mapped to the binding pocket surface using a customized script in Swiss-Pdb Viewer [59]. Color transitions from white to yellow to red represent increasing conservation difference, where red signifies the highest enzyme-specific conservation. White parts are not characteristically different between FT and GGT1, but might also be well conserved between the two enzymes. As the observed differences in the binding pocket conservation relate to features of the enzyme and not the substrates, this information could not be taken directly into account in our scoring scheme. Indirectly, however, the resulting image mirrors the requirement for conservation also for the substrate peptides and, thereby, confirms the relative importance of motif positions as estimated by the Shannon entropy (see below).
Prediction score function
Construction and calculation of the scoring function essentially follows the methodologies [7,16,60] applied to the prediction of GPI anchors [8], myristoylation [9] and PTS1 peroxisomal targeting [10] and is summarized shortly below with emphasis on additions and problem-specific variations.
The composite score function consists of a profile and physical property term.
S = Sprofile + Sppt (2)
The profile term Sprofile
The profile matrix is calculated using the PSIC algorithm to account for redundancy in the learning set. The frequency of occurrence f(a,i) of an amino acid type a at a given alignment position i is down-weighted proportionally to the number of other positions that are identical in sequences sharing a at i, resulting in subscores SPSIC(a,i) representing the natural logarithms of these redundancy-corrected frequencies. These are summed over the respective regions (+3, +2, +1 and -1 to -11)
before they enter Sprofile:
αprofile is a normalization factor to compensate for differing lengths of the regions and is derived from:
Cregion is a relative weighting of motif regions. We propose a rational basis for its approximated computation. Cregion is derived from the average Shannon entropies of the regions in the alignment of the nonredundant learning set sequences. Shannon's information theoretic entropy has already previously been investigated as measurement of sequence variability in alignment positions and residue conservation [61]. We calculate the average of the exponential of the negative Shannon entropy as conservation measurement
where na,i is the number of occurrences of amino-acid type a at the investigated position i of the region and N the total number of sequences in the alignment. Only amino-acid types that occur at least once (na,i ≥ 1) are included in the sum. In order to avoid overly precise values, the final values are rounded up to the last decimal that showed variation in the conservation measure when comparing the small learning set based on Swiss-Prot annotation with the larger learning set described in this work. The position of the absolutely conserved cysteine in the CaaX box would have the maximum conservation value of 1. Table 4 lists Cregion for the regions +1, +2, +3 and -11 to -1.
The physical property term Sppt
The primary role of the physical property term is to exclude query sequences that do not fit into the determined physical property profile of the motif (see Figure 1). Hence, it does not add positive scores for compliance but only penalizes deviations from physical property requirements of the motif. These requirements have been identified as described above (listed in Tables 1 and 2; including rationale for inclusion into Sppt) and the biophysical meaning discussed earlier.
As an example, a penalty for a physical property P of the query being greater than the average physical property Pj over the nonredundant learning set (with σj being the corresponding standard deviation of a Gauss-like distribution) can be defined as follows:
The natural logarithms (to be comparable to the profile score) of these single terms enter Sppt as a sum with a term-specific weighting factor αj that emphasizes varying importance of the single terms.
As part of Sppt, the FT and GGT1 score functions also include a penalty for query proteins when the carboxy terminus appears inaccessible as a result of structural constraints. For example, helices with strongly hydrophobic sides are either buried in the structure or fold back to the protein core and reduce the flexibility and accessibility of the carboxy terminus. We recognize these if hydrophobic residues (LIVMFYW) appear in patterns like i,i+3,i+6 (hXXhXXh) or i,i+3,i+7 (hXXhXXXh) or i,i+4,i+7 (hXXXhXXh) within 20 residues preceding the cysteine of the CaaX box.
Cross-validation (jackknife) tests
We have performed three jackknife tests to validate the robustness of our predictors. First, cross-validation over the whole scoring function (Sprofile + Sppt recalculated after removal of the entry to be validated from the learning set) resulted in sensitivities of 99.4% for FT and 99.8% for GGT1, respectively. This generally indicates that the learning sets are large enough for the method not to suffer from removal of individual entries and, hence it should also be able to find valid motifs not yet included in the learning set. At the same time, we acknowledge that the similarity of sequences within homologous groups in the learning set influence the above estimated sensitivity. However, we emphasize that the values calculated for the profile matrix vary even when only one sequence from a group of homologs is left out, since the method used for profile extraction [56] does take into account such redundancy with both sequence- and position-specific weightings.
To address the role of homologous sequences remaining in the learning set for jackknife tests, we have extended the cross-validation to the more stringent case where not only the sequence to be predicted is excluded from the learning procedure but also its close homologs (estimated by a threshold of 40% sequence identity over the 30 carboxy-terminal residues). We find that in the case of the predictor for GGT1, the sensitivity of 98.6% is very close to the first jackknife test. For the FT predictor, a sensitivity of 92.6% is obtained in this test, which is only slightly lower than in the other jackknife procedures. The decrease can be almost exclusively attributed to the large group of highly similar sequences of hepatitis delta antigen that all share the uncommon motif CRPQ with both arginine at position +1 and proline at +2 being rather exceptional amino acids. Hence, the CRPQ sequences remain below the prediction threshold in the jackknife test if no example of this motif is present in the learning set. When leaving the CRPQ group out of the cross-validation, the FT predictor is calculated to have a sensitivity of 97.9%.
A third cross-validation test aims to elucidate whether the 39 parameters (13 terms with weightings, averages and variances, see Equations 7 and 8) introduced through the physical property terms in Sppt are overfitting the learning data. To exclude bias through similar sequences in homologous groups, the test was executed only over the learning set after removing redundancy (see section on correlation analysis above). Sppt alone is recalculated after removal of the entry to be validated from the parameter calculation procedure. The obtained sensitivities of 100% for FT and 99.8% for GGT1 indicate that the parameterizations of physical property terms in Sppt are not overfitting the learning data.
Probability of false-positive prediction
To estimate probabilities of false-positive prediction, a set of sequences had to be defined whose carboxy-terminal amino-acid pattern should not be subject to selection for a valid CaaX prenylation motif. This is fulfilled for sequences in the NCBI nonredundant database lacking a cysteine at the fourth position from the carboxy terminus. Hence, any compliance with motif restrictions of such sequences apart from the fourth last position could only be attributed to random or at least prenylation-independent appearance. Following earlier experience with GPI- and myristoyl-anchor prediction [8,9], a polynomial extreme-value distribution function has been fitted to the scores obtained for the described sequence set when ignoring the requirement of a CaaX box cysteine in the procedure. The probability of obtaining a score S greater than a threshold score Sth can be formulated as follows:
A polynom of the sixth degree was used to improve the residual fit. Polynoms with degrees higher than six would not result in an increase of the relative improvement of the residual. The probabilities of false-positive prediction for scores at the prediction threshold are extrapolated to 6.3% for FT and 1.2% for GGT1 for sequences that contain a cysteine at the fourth-last position (canonical CaaX box). Given that appearance of a cysteine at this position is rare in databases (1.7%), the independent general probabilities of false-positive prediction by FT and GGT1 for all protein sequences are as low as 0.11% and 0.02%, respectively. This corresponds to specificities of 99.89% and 99.98%.
Since the subcellular context of a protein can be relevant to judging the likelihood of in vivo prenylation when a corresponding motif has been predicted, we also tried to estimate probabilities of false-positive prediction for differently localized subsets of proteins. These subsets were retrieved from the Swiss-Prot database [22] and assigned according to the following keywords in the 'Subcellular Location' comment lines: 'cytoplasmic' (24,284), 'nuclear' (9800), 'membrane' (24,823) and 'extracellular' (509). Parentheses indicate the number of unambiguously annotated examples per subset. Again, only proteins lacking a -CXXX motif were taken into account, to ensure prenylation-independent selection of carboxy-terminal amino-acid residues. Although there appear to be some subcellular localization-specific fluctuations of the probabilities of false-positive prediction (partly due to limited or differing subset sizes), the relative advantages among the methods evaluated seem to remain stable (see Table 3).
Acknowledgements
We thank Boehringer Ingelheim for continuous support. This project has been partly funded by the Fonds zur Förderung der wissenschaftlichen Forschung Österreichs (FWF grant P15037), the Austrian National Bank (Österreichische Nationalbank) and by the bioinformatics contract study 2002-2004 for Bundesministerium fuer Wirtschaft und Arbeit.
Figures and Tables
Figure 1 Sequence logos [74] and physicochemical property profiles of FT and GGT1 substrates. Selected physical properties (hydrophilicity = KRIW790102; flexibility = KARP850103, size = CHOC760101; aliphatic = ZVEL_ALI_1; see Tables 1 and 2 for details) are calculated as average over the nonredundant learning sets of FT and GGT1. The plotted lines correspond to the relative deviation of the respective properties from an average calculated over carboxy termini from the UniRef50 database [22].
Figure 2 The two CaaX prenyltransferases. (a) Ribbon representations of FT (PDB 1D8D [75]) and GGT1 (PDB 1N4Q [76]); pink, alpha subunit; yellow, beta subunit. (b) The prenylpyrophosphates (green) and CaaX tetrapeptides (blue) inside the binding pockets with enzyme-specific conservation (conservation in FT or GGT1 minus conservation in joined FT+GGT1 alignment) mapped to binding-pocket surface. Increasing conservation difference is shaded from white to yellow to red. FPP, farnesyl-, GGPP, geranylgeranylpyrophosphate. The alignment of the sequences of these proteins is shown in Figure 6. Visualized with Swiss-Pdb Viewer [59].
Figure 3 Correlation between predicted and experimental FT/GGT1 substrate selectivity. The correlation of the difference between predicted FT and GGT1 scores with the difference of the experimentally measured logarithmic affinities for FT and GGT1 of the same substrates is plotted.
Figure 4 Determinants of GGT2 prenylation. (a) Sequence logos [74] of Ras superfamily members around part of the Rab-REP interaction site (colored red in the otherwise yellow GTPase structure). (b) Structural model of the Rab-REP-GGT2 prenylation complex based on PDB entries 1LTX [77] and 1VG0 [4]. REP1 (green) has a prenyl-binding pocket which is proposed to be involved in the dual geranylgeranylation mechanism (bound geranylgeranyl is shown in green). However, the catalytic attachment to the substrate cysteines takes place in the center of the GGT2 alpha-beta complex (light and dark blue) where the prenylpyrophosphate that will be transferred is also bound (blue space-filling representation, zinc in red). The structure was visualized using Swiss-Pdb Viewer [59].
Figure 5 Screenshot of the output provided by the PrePS server [39]. On the left is the prediction result for the query protein H-Ras (GenBank P01112) and the three prenylating enzymes. On the right, is shown the carboxy-terminal alignment and PrePS predictions of homologs of the query protein for evaluation of evolutionary motif conservation. Note that H-Ras is predicted to be prenylated only by FT, whereas the homologs K-Ras and N-Ras can also be prenylated by GGT1.
Figure 6 Alignment of FT and GGT1 beta subunits (FTb, GGT1b) in the regions of binding-pocket residues (marked with arrow) using ClustalX [57]. Residue ranges shown above and below correspond to the numbering in the PDB structures of rat FT beta (PDB 1D8D) and GGT1 beta (PDB 1N4Q), respectively. Accession numbers are as follows (GenBank unless indicated otherwise): Hs (Homo sapiens) FTb, NP_002019; GGT1b, NP_005014; Mm (Mus musculus) NP_666039; NP_766215; Rn (Rattus norvegicus) PDB 1D8D; 1N4Q; Tn (Tetraodon nigroviridis) CAG09215; CAF904630; Dm (Drosophila melanogaster) NP_650540; NP_525100; Ag (Anopheles gambiae) XP_321357; XP_317045; Ce (Caenorhabditis elegans) NP_506580; NP_496848; At (Arabidopsis thaliana) NP_198844; NP_181487; Sp (Schizosaccharomyces pombe) NP_594251; NP_594142; Sc (Saccharomyces cerevisiae) P22007; NP_011360. Standard ClustalX coloring (according to conserved amino acid type).
Table 1 Physical property terms in the FT scoring function
Property Position Rationale Explanation
ARGP820103 [62] +3 Corr = 0.7(nrLS) Membrane-buried preference, lipid contact when entering binding pocket
logPREN_CKQX_FT [15] +3 Corr = -0.72(nrLS) Kinetic measurement, relative unprocessed FPP amounts with tetrapeptide CKQX
CHOC760101 [63] +1 to +3 Fisher = 1.3 Side chain volume
ZVEL_CHARG [64] +1 to +3 LS composition General charge penalty
ZVEL_CHNEG [64] +1 to +3 LS composition Special negative charge penalty
WERD780102 [65] +1 and +3 Fisher = 1.51 Hydrophobicity compensation for inside preference
ZVEL_ALI_1 [64] +1 and +2 +2: Corr = 0.85(prof)
+1: continuing deviation from Uniref50 average Amino-acid property: aliphatic
LIFS790102 [66] +1 and +2 +2: Correlation = 0.76(prof)
+1: continuing deviation from Uniref50 average Preference for extended conformations
ZVEL_TINY_ [64] -1 Corr = 0.68(prof) Size, bulkiness
MOBILITY_2 [21] -1 Corr = 0.61(nrLS) Side chain mobility
VINM940101 [67] -11 to -1 -2: Corr = 0.72(prof)
-3: Corr = 0.75(prof)
-4: Corr = 0.78(nrLS)
-5: Corr = 0.82(nrLS)
-6: Corr = 0.84(nrLS)
-7: Corr = 0.79(nrLS)
-8: Corr = 0.74(prof)
-9: Corr = 0.82(nrLS)
-10: Corr = 0.84(nrLS)
-11: Corr = 0.79(nrLS)
Rest: continuing deviation from Uniref50 average Normalized flexibility average
KRIW790102 [68] -11 to -1 -2: Corr = 0.76(prof)
-6: Corr = 0.83(nrLS)
-7: Corr = 0.83(nrLS)
-8: Corr = 0.76(prof)
Rest: continuing deviation from Uniref50 average Fraction of site occupied with water
Buried helix (see Materials and methods) -20 to -1 Remove false positives Helix with strongly hydrophobic sides folds back to protein core and reduces flexibility and accessibility of C-terminus
Corr, correlation; LS, learning set; nrLS, nonredundant; prof, profile.
Table 2 Physical property terms in the GGT1 scoring function
Property Position Rationale Explanation
ARGP820103 [62] +3 Corr = 0.8(prof) Membrane-buried preference, lipid contact when entering binding pocket
LEVM760105 [69] +1 to +3 Fisher = 1.36 Size limitation (radius of gyration of side-chain)
YUTK870101 [70] +1 to +3 Fisher = 1.38 Hydrophobicity compensation (Unfolding Gibbs energy in water, pH7.0)
ZVEL_CHARG [64] +1 to +3 LS composition General charge penalty
ZVEL_CHNEG [64] +1 to +3 LS composition Special negative charge penalty
ZVEL_ALI_1 [64] +1 and +2 +2: Corr = 0.87(prof)
+1: continuing deviation from Uniref50 average Amino-acid property: aliphatic
LIFS790102 [66] +1 and +2 +2: Corr = 0.77(prof)
+1: continuing deviation from Uniref50 average Preference for extended conformations
FAUJ880101 [71] -1 and +2 Fisher = 1.52 Size, bulkiness (residues although 10 Å apart, face to same side of base pair)
FINA910103 [72] -1 Corr = 0.75(prof) Helix termination (for example, K, S favored, D,E,L,I,V disfavored)
KARP850103 [73] -7 to-1 -1: Corr = 0.69(prof)
-2: Corr = 0.70(prof)
-3: Corr = 0.71(prof)
-4: Corr = 0.74(nrLS)
-5: Corr = 0.75(prof)
-6: Corr = 0.70(nrLS)
-7: Corr = 0.78(nrLS) Flexibility (GGT1 lysine preference)
VINM940101 [67] -11 to -1 -4: Corr = 0.72(prof)
-5: Corr = 0.82(prof)
-6: Corr = 0.84(nrLS)
-7: Corr = 0.75(nrLS)
-8: Corr = 0.77(nrLS)
-9: Corr = 0.68(prof)
-10: Corr = 0.86(prof)
Rest: continuing deviation from Uniref50 average Normalized flexibility average
KRIW790102 [68] -11 to -1 -3: Corr = 0.70(prof)
-4: Corr = 0.73(prof)
-5: Corr = 0.84(prof)
-6: Corr = 0.81(prof)
-7: Corr = 0.83(nrLS)
-8: Corr = 0.85(nrLS)
-9: Corr = 0.76(prof)
-10: Corr = 0.86(prof)
Rest: continuing deviation from Uniref50 average Fraction of site occupied with water
Buried helix (see Materials and methods) -20 to -1 Remove false positives Helix with strongly hydrophobic sides folds back to protein core and reduces flexibility and accessibility of carboxy terminus
Table 3 Comparison of prediction performances
FT GGT1
Prosite PS00294 Beese, Casey and colleagues' rules PrePS FT Prosite PS00294 Beese, Casey and colleagues' rules PrePS GGT1
Sensitivity I 85%* 72% 100% 95%* 67% 100%
Sensitivity II NA NA 92.6/97.9%† NA NA 98.6%
Probability of false positive prediction (POFP) for -CXXX motifs (GenBank sequences) 17.1%* 9.9% 6.3% 17.1%* 10.0% 1.2%
POFP -CXXX 'cytoplasmic'‡ 18.2%* 8.9% 5.1% 18.2%* 8.6% 1.4%
POFP -CXXX 'nuclear'‡ 13.9%* 10.5% 5.5% 13.9%* 9.6% 1.1%
POFP -CXXX 'membrane'‡ 17.5%* 10.3% 3.8% 17.5%* 12.0% 0.8%
POFP --CXXX 'extracellular'$ 8.6%* 7.9% 3.3% 8.6%* 9.0% 0.2%
Overall probability of false positive prediction (GenBank sequences, assuming 1.7% with -CXXX) 0.29%* 0.16% 0.11% 0.29%* 0.17% 0.02%
*Prosite pattern PS00294 does not distinguish between prenylation by FT and GGT1.
†Sensitivity rises to 97.9% when the exceptional motif CRPQ of hepatitis delta antigen is removed. ‡For details see Materials and methods. Sensitivity I is the rate of finding known substrates from described learning set = self-consistency. Sensitivity II is the rate of finding known substrates after their exclusion (including homologs) from the learning set = cross-validation (see Materials and methods). Probabilities of false-positive predictions (POFP) complement the specificities to 100% (Specificity = 100 - POFP). The first listed POFP estimates the rates of false positives among query proteins that have a canonical -CXXX motif (which corresponds to 1.7% of all sequences). Below are estimations of POFPs for subsets of Swiss-Prot proteins that differ in their annotated subcellular localization (see Materials and methods). The final POFP is the estimate for false-positive predictions for all sequences (for example, when analyzing complete proteomes or large databases), independent of existence of a -CXXX motif. Formatting signifies: best (bold), intermediate (plain text), worst (italic) performance.
Table 4 Relative weightings of motif positions in profile term
Position(s) FT GGT1
-11 to -1 0.07 0.07
+1 0.15 0.14
+2 0.29 0.41
+3 0.17 0.42
==== Refs
Casey PJ Seabra MC Protein prenyltransferases. J Biol Chem 1996 271 5289 5292 8621375 10.1074/jbc.271.10.5289
Maurer-Stroh S Washietl S Eisenhaber F Protein prenyltransferases. Genome Biol 2003 4 212 12702202 10.1186/gb-2003-4-4-212
Roskoski R Jr Protein prenylation: a pivotal posttranslational process. Biochem Biophys Res Commun 2003 303 1 7 12646157 10.1016/S0006-291X(03)00323-1
Rak A Pylypenko O Niculae A Pyatkov K Goody RS Alexandrov K Structure of the Rab7:REP-1 complex: insights into the mechanism of Rab prenylation and choroideremia disease. Cell 2004 117 749 760 15186776 10.1016/j.cell.2004.05.017
Caplin BE Hettich LA Marshall MS Substrate characterization of the Saccharomyces cerevisiae protein farnesyltransferase and type-I protein geranylgeranyltransferase. Biochim Biophys Acta 1994 1205 39 48 8142482
Wilson AL Erdman RA Castellano F Maltese WA Prenylation of Rab8 GTPase by type I and type II geranylgeranyl transferases. Biochem J 1998 333 497 504 9677305
Eisenhaber F Eisenhaber B Maurer-Stroh S Andrade MM Prediction of post-translational modifications from amino acid sequence: problems, pitfalls, methodological hints. Bioinformatics and Genomes: Current Perspectives 51 2003 Wymondham, UK: Horizon Scientific Press 81 105
Eisenhaber B Bork P Eisenhaber F Prediction of potential GPI-modification sites in proprotein sequences. J Mol Biol 1999 292 741 758 10497036 10.1006/jmbi.1999.3069
Maurer-Stroh S Eisenhaber B Eisenhaber F N-terminal N-myristoylation of proteins: prediction of substrate proteins from amino acid sequence. J Mol Biol 2002 317 541 557 11955008 10.1006/jmbi.2002.5426
Neuberger G Maurer-Stroh S Eisenhaber B Hartig A Eisenhaber F Prediction of peroxisomal targeting signal 1 containing proteins from amino acid sequence. J Mol Biol 2003 328 581 592 12706718 10.1016/S0022-2836(03)00319-X
Boeckmann B Bairoch A Apweiler R Blatter MC Estreicher A Gasteiger E Martin MJ Michoud K O'Donovan C Phan I The SWISS-PROT protein knowledgebase and its supplement TrEMBL in 2003. Nucleic Acids Res 2003 31 365 370 12520024 10.1093/nar/gkg095
Maurer-Stroh S Eisenhaber B Eisenhaber F N-terminal N-myristoylation of proteins: refinement of the sequence motif and its taxon-specific differences. J Mol Biol 2002 317 523 540 11955007 10.1006/jmbi.2002.5425
Roskoski R JrRitchie P Role of the carboxyterminal residue in peptide binding to protein farnesyltransferase and protein geranylgeranyltransferase. Arch Biochem Biophys 1998 356 167 176 9705207 10.1006/abbi.1998.0768
Maurer-Stroh S Gouda M Novatchkova M Schleiffer A Schneider G Sirota FL Wildpaner M Hayashi N Eisenhaber F MYRbase: analysis of genome-wide glycine myristoylation enlarges the functional spectrum of eukaryotic myristoylated proteins. Genome Biol 2004 5 R21 15003124 10.1186/gb-2004-5-3-r21
Boutin JA Marande W Petit L Loynel A Desmet C Canet E Fauchere JL Investigation of S-farnesyl transferase substrate specificity with combinatorial tetrapeptide libraries. Cell Signal 1999 11 59 69 10206346 10.1016/S0898-6568(98)00032-1
Eisenhaber B Eisenhaber F Maurer-Stroh S Neuberger G Prediction of sequence signals for lipid post-translational modifications: insights from case studies. Proteomics 2004 4 1614 1625 15174131 10.1002/pmic.200300781
Valid CaaX tetrapeptides at the PrePS site
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. Nucleic Acids Res 1997 25 3389 3402 9254694 10.1093/nar/25.17.3389
Enright AJ Van Dongen S Ouzounis CA An efficient algorithm for large-scale detection of protein families. Nucleic Acids Res 2002 30 1575 1584 11917018 10.1093/nar/30.7.1575
Tomii K Kanehisa M Analysis of amino acid indices and mutation matrices for sequence comparison and structure prediction of proteins. Protein Eng 1996 9 27 36 9053899
Eisenhaber B Bork P Eisenhaber F Sequence properties of GPI-anchored proteins near the omega-site: constraints for the polypeptide binding site of the putative transamidase. Protein Eng 1998 11 1155 1161 9930665 10.1093/protein/11.12.1155
Apweiler R Bairoch A Wu CH Barker WC Boeckmann B Ferro S Gasteiger E Huang H Lopez R Magrane M UniProt: the Universal Protein knowledgebase. Nucleic Acids Res 2004 32 Database D115 D119 14681372 10.1093/nar/gkh131
Reid TS Terry KL Casey PJ Beese LS Crystallographic analysis of CaaX prenyltransferases complexed with substrates defines rules of protein substrate selectivity. J Mol Biol 2004 343 417 433 15451670 10.1016/j.jmb.2004.08.056
Maurer-Stroh S Washietl S Eisenhaber F Protein prenyltransferases: anchor size, pseudogenes and parasites. Biol Chem 2003 384 977 989 12956414 10.1515/BC.2003.110
Prakash B Praefcke GJ Renault L Wittinghofer A Herrmann C Structure of human guanylate-binding protein 1 representing a unique class of GTP-binding proteins. Nature 2000 403 567 571 10676968 10.1038/35000617
Neuberger G Maurer-Stroh S Eisenhaber B Hartig A Eisenhaber F Motif refinement of the peroxisomal targeting signal 1 and evaluation of taxon-specific differences. J Mol Biol 2003 328 567 579 12706717 10.1016/S0022-2836(03)00318-8
Moores SL Schaber MD Mosser SD Rands E O'Hara MB Garsky VM Marshall MS Pompliano DL Gibbs JB Sequence dependence of protein isoprenylation. J Biol Chem 1991 266 14603 14610 1860864
Boutin JA Marande W Goussard M Loynel A Canet E Fauchere JL Chromatographic assay and peptide substrate characterization of partially purified farnesyl- and geranylgeranyltransferases from rat brain cytosol. Arch Biochem Biophys 1998 354 83 94 9633601 10.1006/abbi.1998.0678
Caldwell GA Wang SH Naider F Becker JM Consequences of altered isoprenylation targets on a-factor export and bioactivity. Proc Natl Acad Sci USA 1994 91 1275 1279 8108401
Prendergast GC Actin' up: RhoB in cancer and apoptosis. Nat Rev Cancer 2001 1 162 168 11905808 10.1038/35101096
Carboni JM Yan N Cox AD Bustelo X Graham SM Lynch MJ Weinmann R Seizinger BR Der CJ Barbacid M Farnesyltransferase inhibitors are inhibitors of Ras but not R-Ras2/TC21 transformation. Oncogene 1995 10 1905 1913 7761092
Rowell CA Kowalczyk JJ Lewis MD Garcia AM Direct demonstration of geranylgeranylation and farnesylation of Ki-Ras in vivo. J Biol Chem 1997 272 14093 14097 9162034 10.1074/jbc.272.22.14093
Pereira-Leal JB Hume AN Seabra MC Prenylation of Rab GTPases: molecular mechanisms and involvement in genetic disease. FEBS Lett 2001 498 197 200 11412856 10.1016/S0014-5793(01)02483-8
Roth AF Feng Y Chen L Davis NG The yeast DHHC cysteine-rich domain protein Akr1p is a palmitoyl transferase. J Cell Biol 2002 159 23 28 12370247 10.1083/jcb.200206120
Vancura A Sessler A Leichus B Kuret J A prenylation motif is required for plasma membrane localization and biochemical function of casein kinase I in budding yeast. J Biol Chem 1994 269 19271 19278 8034689
Pereira-Leal JB Seabra MC Evolution of the Rab family of small GTP-binding proteins. J Mol Biol 2001 313 889 901 11697911 10.1006/jmbi.2001.5072
Pereira-Leal JB Seabra MC The mammalian Rab family of small GTPases: definition of family and subfamily sequence motifs suggests a mechanism for functional specificity in the Ras superfamily. J Mol Biol 2000 301 1077 1087 10966806 10.1006/jmbi.2000.4010
Eddy SR Profile hidden Markov models. Bioinformatics 1998 14 755 763 9918945 10.1093/bioinformatics/14.9.755
PrePS - Prenylation Prediction Suite
Falquet L Pagni M Bucher P Hulo N Sigrist CJ Hofmann K Bairoch A The PROSITE database, its status in 2002. Nucleic Acids Res 2002 30 235 238 11752303 10.1093/nar/30.1.235
Horton P Nakai K Better prediction of protein cellular localization sites with the k nearest neighbors classifier. Proc Int Conf Intell Syst Mol Biol 1997 5 147 152 9322029
Doll RJ Kirschmeier P Bishop WR Farnesyltransferase inhibitors as anticancer agents: critical crossroads. Curr Opin Drug Discov Devel 2004 7 478 486 15338957
Gelb MH Van Voorhis WC Buckner FS Yokoyama K Eastman R Carpenter EP Panethymitaki C Brown KA Smith DF Protein farnesyl and N-myristoyl transferases: piggy-back medicinal chemistry targets for the development of antitrypanosomatid and antimalarial therapeutics. Mol Biochem Parasitol 2003 126 155 163 12615314 10.1016/S0166-6851(02)00282-7
Sebti SM Hamilton AD Farnesyltransferase and geranylgeranyltransferase I inhibitors and cancer therapy: lessons from mechanism and bench-to-bedside translational studies. Oncogene 2000 19 6584 6593 11426643 10.1038/sj.onc.1204146
Tamanoi F Gau CL Jiang C Edamatsu H Kato-Stankiewicz J Protein farnesylation in mammalian cells: effects of farnesyltransferase inhibitors on cancer cells. Cell Mol Life Sci 2001 58 1636 1649 11706990
Sebti SM Der CJ Opinion: Searching for the elusive targets of farnesyltransferase inhibitors. Nat Rev Cancer 2003 3 945 951 14737124 10.1038/nrc1234
Simon HU Mills GB Kozlowski M Hogg D Branch D Ishimi Y Siminovitch KA Molecular characterization of hNRP, a cDNA encoding a human nucleosome-assembly-protein-I-related gene product involved in the induction of cell proliferation. Biochem J 1994 297 389 397 8297347
McBryant SJ Park YJ Abernathy SM Laybourn PJ Nyborg JK Luger K Preferential binding of the histone (H3-H4)2 tetramer by NAP1 is mediated by the amino-terminal histone tails. J Biol Chem 2003 278 44574 44583 12928440 10.1074/jbc.M305636200
Miyaji-Yamaguchi M Kato K Nakano R Akashi T Kikuchi A Nagata K Involvement of nucleocytoplasmic shuttling of yeast Nap1 in mitotic progression. Mol Cell Biol 2003 23 6672 6684 12944491 10.1128/MCB.23.18.6672-6684.2003
Gotte K Girzalsky W Linkert M Baumgart E Kammerer S Kunau WH Erdmann R Pex19p, a farnesylated protein essential for peroxisome biogenesis. Mol Cell Biol 1998 18 616 628 9418908
Sapkota GP Kieloch A Lizcano JM Lain S Arthur JS Williams MR Morrice N Deak M Alessi DR Phosphorylation of the protein kinase mutated in Peutz-Jeghers cancer syndrome, LKB1/STK11, at Ser431 by p90(RSK) and cAMP-dependent protein kinase, but not its farnesylation at Cys(433), is essential for LKB1 to suppress cell growth. J Biol Chem 2001 276 19469 19482 11297520 10.1074/jbc.M009953200
Kho Y Kim SC Jiang C Barma D Kwon SW Cheng J Jaunbergs J Weinbaum C Tamanoi F Falck J Zhao Y A tagging-via-substrate technology for detection and proteomics of farnesylated proteins. Proc Natl Acad Sci USA 2004 101 12479 12484 15308774 10.1073/pnas.0403413101
Gromov P Celis JE Rab11a is modified in vivo by isoprenoid geranylgeranyl. Electrophoresis 1998 19 1803 1807 9719562 10.1002/elps.1150191043
Gomes AQ Ali BR Ramalho JS Godfrey RF Barral DC Hume AN Seabra MC Membrane targeting of Rab GTPases is influenced by the prenylation motif. Mol Biol Cell 2003 14 1882 1899 12802062 10.1091/mbc.E02-10-0639
Joberty G Tavitian A Zahraoui A Isoprenylation of Rab proteins possessing a C-terminal CaaX motif. FEBS Lett 1993 330 323 328 8375503 10.1016/0014-5793(93)80897-4
Sunyaev SR Eisenhaber F Rodchenkov IV Eisenhaber B Tumanyan VG Kuznetsov EN PSIC: profile extraction from sequence alignments with position-specific counts of independent observations. Protein Eng 1999 12 387 394 10360979 10.1093/protein/12.5.387
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
Pei J Grishin NV AL2CO: calculation of positional conservation in a protein sequence alignment. Bioinformatics 2001 17 700 712 11524371 10.1093/bioinformatics/17.8.700
Guex N Peitsch MC SWISS-MODEL and the Swiss-PdbViewer: an environment for comparative protein modeling. Electrophoresis 1997 18 2714 2723 9504803 10.1002/elps.1150181505
Eisenhaber F Eisenhaber B Kubina W Maurer-Stroh S Neuberger G Schneider G Wildpaner M Prediction of lipid posttranslational modifications and localization signals from protein sequences: big-Pi, NMT and PTS1. Nucleic Acids Res 2003 31 3631 3634 12824382 10.1093/nar/gkg537
Valdar WS Scoring residue conservation. Proteins 2002 48 227 241 12112692 10.1002/prot.10146
Argos P Rao JK Hargrave PA Structural prediction of membrane-bound proteins. Eur J Biochem 1982 128 565 575 7151796
Chothia C The nature of the accessible and buried surfaces in proteins. J Mol Biol 1976 105 1 12 994183 10.1016/0022-2836(76)90191-1
Zvelebil MJ Barton GJ Taylor WR Sternberg MJ Prediction of protein secondary structure and active sites using the alignment of homologous sequences. J Mol Biol 1987 195 957 961 3656439 10.1016/0022-2836(87)90501-8
Wertz DH Scheraga HA Influence of water on protein structure. An analysis of the preferences of amino acid residues for the inside or outside and for specific conformations in a protein molecule. Macromolecules 1978 11 9 15 621952 10.1021/ma60061a002
Lifson S Sander C Antiparallel and parallel beta-strands differ in amino acid residue preferences. Nature 1979 282 109 111 503185
Vihinen M Torkkila E Riikonen P Accuracy of protein flexibility predictions. Proteins 1994 19 141 149 8090708 10.1002/prot.340190207
Krigbaum WR Komoriya A Local interactions as a structure determinant for protein molecules: II. Biochim Biophys Acta 1979 576 204 248 760806
Levitt M A simplified representation of protein conformations for rapid simulation of protein folding. J Mol Biol 1976 104 59 107 957439 10.1016/0022-2836(76)90004-8
Yutani K Ogasahara K Tsujita T Sugino Y Dependence of conformational stability on hydrophobicity of the amino acid residue in a series of variant proteins substituted at a unique position of tryptophan synthase alpha subunit. Proc Natl Acad Sci USA 1987 84 4441 4444 3299367
Fauchere JL Charton M Kier LB Verloop A Pliska V Amino acid side chain parameters for correlation studies in biology and pharmacology. Int J Pept Protein Res 1988 32 269 278 3209351
Finkelstein AV Badretdinov AY Ptitsyn OB Physical reasons for secondary structure stability: alpha-helices in short peptides. Proteins 1991 10 287 299 1946339 10.1007/BF01025627
Karplus PA Schulz GE Prediction of chain flexibility in proteins. A tool for the selection of peptide antigens Naturwiss 1985 72 212 213 10.1007/BF01195768
Crooks GE Hon G Chandonia JM Brenner SE WebLogo: a sequence logo generator. Genome Res 2004 14 1188 1190 15173120 10.1101/gr.849004
Long SB Casey PJ Beese LS The basis for K-Ras4B binding specificity to protein farnesyltransferase revealed by 2 A resolution ternary complex structures. Structure Fold Des 2000 8 209 222 10673434 10.1016/S0969-2126(00)00096-4
Taylor JS Reid TS Terry KL Casey PJ Beese LS Structure of mammalian protein geranylgeranyltransferase type-I. EMBO J 2003 22 5963 5974 14609943 10.1093/emboj/cdg571
Pylypenko O Rak A Reents R Niculae A Sidorovitch V Cioaca MD Bessolitsyna E Thoma NH Waldmann H Schlichting I Structure of rab escort protein-1 in complex with rab geranylgeranyltransferase. Mol Cell 2003 11 483 494 12620235 10.1016/S1097-2765(03)00044-3
| 15960807 | PMC1175975 | CC BY | 2021-01-04 16:05:39 | no | Genome Biol. 2005 May 27; 6(6):R55 | utf-8 | Genome Biol | 2,005 | 10.1186/gb-2005-6-6-r55 | oa_comm |
==== Front
Genome BiolGenome Biology1465-69061465-6914BioMed Central London gb-2005-6-6-r551596080710.1186/gb-2005-6-6-r55SoftwareRefinement and prediction of protein prenylation motifs Maurer-Stroh Sebastian [email protected] Frank [email protected] IMP - Research Institute of Molecular Pathology, Dr. Bohr-Gasse 7, A-1030 Vienna, Austria2005 27 5 2005 6 6 R55 R55 17 1 2005 22 3 2005 20 4 2005 Copyright © 2005 Maurer-Stroh and Eisenhaber; 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.
Three prenylation motif predictors are presented that allow discrimination between proteins that are unique substrates of farnesyltransferase (FT) and those that can be alternatively processed by geranylgeranyltransferase I (GGT1).
We refined the motifs for carboxy-terminal protein prenylation by analysis of known substrates for farnesyltransferase (FT), geranylgeranyltransferase I (GGT1) and geranylgeranyltransferase II (GGT2). In addition to the CaaX box for the first two enzymes, we identify a preceding linker region that appears constrained in physicochemical properties, requiring small or flexible, preferably hydrophilic, amino acids. Predictors were constructed on the basis of sequence and physical property profiles, including interpositional correlations, and are available as the Prenylation Prediction Suite (PrePS, ) which also allows evaluation of evolutionary motif conservation. PrePS can predict partially overlapping substrate specificities, which is of medical importance in the case of understanding cellular action of FT inhibitors as anticancer and anti-parasite agents.
==== Body
Rationale
Prenylation refers to the posttranslational modification of proteins with isoprenyl anchors [1-3]. These lipid moieties are typically involved in mediating not only protein-membrane but also protein-protein interactions. Three eukaryotic enzymes are known to catalyze the lipid transfer. The first two, farnesyltransferase (FT) and geranylgeranyltransferase 1 (GGT1), recognize the so-called CaaX box in the carboxy termini of substrate proteins and attach farnesyl (15-carbon polyisoprene) or geranylgeranyl (20-carbon polyisoprene), respectively, to a required and spatially fixed cysteine in that motif. The third enzyme, geranylgeranyltransferase 2 (GGT2 or RabGGT) recognizes the complex [4] of Rab GTPase substrate proteins with a specific Rab escort protein (REP) to attach one or two geranylgeranyl anchors to cysteines in a more flexible but also carboxy-terminal motif.
The CaaX box was initially understood to consist of a cysteine (C), followed by two aliphatic residues (aa) and a terminal residue (X) that would direct modification by either FT or GGT1, but newly found substrates and kinetic studies of mutated substrate peptides and enzyme inhibitors have shown that the motif recognized by the enzymes appears to be more flexible [2]. Furthermore, the determination of preference for FT or GGT1 is more complex and a function of the overall sequence context rather than specific amino acids at single positions. Whereas GGT2 appears to be specific to Rab GTPases as substrates, the recognition mechanism is not well understood. Overlapping substrate specificities between all three prenylating enzymes further complicate the understanding of the lipid modification process [5,6].
An unsolved problem so far is accounting for the complexity of the prenylation substrate recognition motifs in theoretical models in order to identify substrate proteins from their amino-acid sequence. No available method has been able to selectively assign the correct modifying enzyme, which determines the types and number of lipid anchors. The high probability of motifs similar to the small CaaX box occurring by chance is a general problem that has so far prohibited large-scale proteome analyses [7]. We describe here a method that aims to model the substrate-enzyme interactions on the basis of refinement of the recognition motifs for each of the prenyltransferases. The Prenylation Prediction Suite (PrePS) selectively assigns the modifying enzyme to predicted substrate proteins and sensitively filters out false-positive predictions based on the general methodology that has already been applied successfully for the prediction of glycosylphosphatidylinositol (GPI) anchors [8], myristoylation [9] and PTS1 peroxisomal targeting [10].
Known substrates and their motif-compliant homologs as learning sets
The first task consists of collecting sequences that are known substrates for the respective enzymes. Typically, a good starting point is the Swiss-Prot database [11]. However, according to earlier experience with annotation inaccuracies [12], any annotated experimental evidence has to be confirmed by following up all the related literature sources. As newly available data can be missing in the Swiss-Prot annotation, the searches have also to be extended to non-Swiss-Prot proteins. In most cases, the annotations for prenylation in Swiss-Prot are assigned by similarity to only a few entries with experimental validation. A major concern is the annotation of the correct anchor type attached to FT and GGT1 substrates, which could previously only tentatively be estimated without experimental data. This includes several entries with overall sequence similarity to a verified prenylated protein but totally different carboxy-terminal motifs. Given that single mutations can abolish recognition or switch enzyme specificities [13] and that not all homologs of lipid-modified proteins necessarily have to share the same modification type or membrane attachment factor (MAF) [14], entries with annotations only by similarity should not be included without critical consideration in a learning set.
Unfortunately, such justified concerns dramatically lower the amount of data in the learning set. However, because of earlier interest in developing peptide-based inhibitors of FT and GGT1 as anticancer treatments, the kinetics of the enzymes with various tetrapeptide substrates already modified with lipid anchors by the enzymes have been measured [15]. Hence, a protein homologous to a verified prenylated protein can be included in the learning set if its CaaX box has already been shown to interact productively with one of the prenyltransferases at least as a tetrapeptide.
However, possession of a valid CaaX box might not be a sufficient selection criterion. Typically, short terminal sequence motifs are connected to the rest of the protein by a linker region that experiences only limited constraints on specific amino acids per position but often has a compositional bias towards small or hydrophilic amino acids in connecting sequence stretches [16]. This property is found in a preliminary assembly of verified FT and GGT1 substrates and has been confirmed in the actual learning set for up to 11 residues upstream (amino-terminal) of the cysteine in the CaaX box (see below). Hence, learning-set sequences should also not violate the physicochemical properties constraining the sequence stretch amino-terminal to the CaaX box.
Taking account of the considerations above, the following procedure has been applied to obtain conservative and reliable learning sets of FT and GGT1 substrates. First, a literature search for known prenylated proteins and valid tetrapeptides (see [17]). Second, BLASTP [18] with an E-value threshold of 0.005 starting with known prenylated proteins against the National Center for Biotechnology Information (NCBI) nonredundant database to find homologs and cluster all collected sequences into groups of homologous proteins using the Markov-chain clustering algorithm (MCL) [19]. Third, check the validity of all CaaX boxes with experimental evidence for at least tetrapeptides. Fourth, check compliance with the physical properties of the full motif (including linker) by applying a preliminary predictor based on corrected Swiss-Prot entries in a similar style as described here (penalizing deviations from the physical property landscape of the motif).
This resulted in learning sets of 692 FT and 486 GGT1 substrates, respectively (see [17]). Among the FT substrates, 31 artificial constructs or mutations of naturally occurring sequences that have been shown to be processed by FT have also been included. Prenylation by GGT2 follows totally different mechanistic requirements than FT and GGT1 and will be treated separately after the sections about CaaX prenylation.
Refinement of the CaaX box motif descriptions
Compositional analysis of residue frequencies at single motif positions reveals that major restrictions to specific amino acid types exist only for positions within the CaaX box (see sequence logos in Figure 1). The previously reported preferences for aliphatic residues at positions +1 and +2 (the aa in CaaX) were recovered, but there is a clear tendency for other residue types to also be allowed, especially at position +1 (the first a in CaaX). Correlation analysis of residue frequencies at single motif positions with amino-acid property scales [20,21] can quantify the conservation of a physical property pattern (see Materials and methods). Although correlations higher than 0.6 can only be obtained for aliphatic property at position +2 (FT: 0.85, GGT1: 0.87), the average aliphatic property at position +1 within both FT and GGT1 learning sets still appears elevated when compared to an average calculated from the carboxy-termini of the nonredundant UniRef50 database [22] (see physical property profile in Figure 1). Similarly, there are correlations at position +2 and deviations from the UniRef50 average at position +1 for a property describing preference for extended conformations (see Tables 1 and 2). This appears to be best explained by the need to have the final peptide part in extended conformation rather than coiled or helical in order to fit into the binding pocket, as can be seen in the resolved structures of prenyltransferases with their substrate peptides [23].
The major difference between FT and GGT1 substrates remains at position +3 (the X in CaaX). Whereas a broad variety of residues are allowed in motifs recognized by FT (including several substrates with leucine at +3), mainly leucine and methionine appear to be preferred by GGT1 in agreement with experimental evidence [13]. Interestingly, position +3 correlates (FT, 0.7; GGT1, 0.8) with a physical property that measures membrane-buried preference parameters (see Tables 1 and 2). This feature does not seem to be important to support membrane interaction at a later stage for the protein, as the three carboxy-terminal residues (-aaX) are often cleaved off in a further processing step after attachment of the anchor [24]. However, hydrophobicity and volume of position +3 appear important for interaction with the binding pocket because of the rather lipophilic character of the latter (isoprenyl anchor on one side and hydrophobic residues on the others). The importance of position +3 for specificity between FT and GGT1 is further strengthened by differing conservation of residues in the binding pockets of the respective enzymes (Figure 2). Not surprisingly, the whole region of the binding pocket harboring the end of the prenylpyrophosphate (geranylgeranyl [C20] is one isoprene unit longer than farnesyl [C15]) and the X of the CaaX box (position +3) appear to comprise the major differences in residue conservation (Figure 2).
Using the Fisher criterion (see Materials and methods), interpositional correlations of residue sizes within positions +1, +2 and +3 (the carboxy-terminal three residues of the CaaX box that are buried in the binding pocket) from both FT and GGT1 substrates have been identified. Often, when a very large residue occurs at specific positions, neighboring residues compensate to obey the overall physicochemical constraints (for example, size limitation) in the binding pocket. Similarly, compensatory effects appear to exist regarding hydrophobicity between positions +1 and +3 in FT and between +1, +2 and +3 in GGT1 substrates (see Tables 1 and 2). Compensatory effects also seem responsible for the toleration of even large positively charged residues at positions +1 or +2, if the other residues are small enough to accommodate the whole peptide in the binding pocket. On the other hand, negative charges are apparently incompatible with the substrate recognition motif at these positions.
Extension of the CaaX prenylation motif by a flexible linker region
While the requirement for specific amino acids at single positions appears to be marginal outside of the CaaX box, physicochemical constraints that extend up to 11 residues amino-terminal from the modified cysteine can be found (Figure 1, Tables 1 and 2). At position -1 of the motif, there begins a pronounced tendency for residues with either small or flexible hydrophilic side chains. GGT1 especially appears to prefer amino acids like serine or lysine at this position. In general, GGT1 substrates have a higher number of lysines within positions -1 and -7 compared with the FT substrates.
The hydrophilic linker region with correlations over multiple positions to several hydrophobicity- and flexibility-related property scales might be required to allow accessibility of the carboxy terminus for the lipid-attaching enzymes. Indeed, in several resolved structures of in vivo prenylated GTPases, secondary structural elements such as helices that stabilize the fold of the protein are typically found only at the amino-terminal side of that linker region (beginning of helix at positions -12 (PDB identifier 1FTN), -13 (PDB 1MH1), -15 (PDB 1AM4), -12 (PDB 1A4R)). In the structure of a G protein gamma subunit, the linker region also appears to be extended and wrapped around the beta subunit in the heterotrimeric G protein signaling complex (PDB 1GG2). It needs to be emphasized that the linker region must not necessarily be in an unstructured conformation after the anchor has been attached (see also carboxy-terminal helix in structure PDB 1F5N of human 67 kDa guanylate binding protein 1 [25]), as folding back or lipid-mediated interaction with other proteins or membranes can also induce changes in the three-dimensional structure of the linker region. However, there appears to be a requirement for the ability to easily unfold/fold into flexible and more extended conformations that allow the carboxy terminus to be accessed and modified by the prenyltransferases. It is noteworthy that this length estimation of a flexible, hydrophilic linker is consistent with earlier findings in the GPI anchor [21], myristoylation [12] and PTS1 targeting [26] motifs. Hence, the actual motif length of substrates for CaaX prenylation appears longer than previously thought (total 15 residues = 4 CaaX + 11 linker).
Prediction function and validation
Following the approach already applied to the prediction of GPI and myristoyl anchors and PTS1-mediated targeting [8-10], a scoring function measuring compliance with the prenylation motif separately for the enzymes FT and GGT1, respectively, has been constructed (see Materials and methods). In brief, the composite prediction function S consists of a term Sprofile scoring a query sequence against the redundancy-corrected profile of the learning-set sequences and another term Sppt that penalizes deviation from the physicochemical motif requirements.
S = Sprofile + Sppt
The term Sprofile distinguishes the three positions +1, +2 and +3 of the CaaX box as well as the linker region (-1 to -11). Sppt comprises a sum of terms that are constructed from the physical property requirements for FT and GGT1 substrates that were outlined in the section describing the motif refinement (and listed in Tables 1 and 2 together with their rationale for inclusion in Sppt).
The threshold for a query protein to be a predicted farnesylation or geranylgeranylation target by FT or GGT1, respectively, is set to include all sequences in the learning set. Hence, the self-consistencies or upper bounds of sensitivities of the FT and GGT1 predictors are 100%. Additionally, the robustness of the method has been cross-validated in jackknife tests (see Materials and methods). In the cross-validation over the complete scoring function, the rates of finding known substrates after excluding them and their close homologs from the learning procedure (and, therefore, lower bounds for sensitivities) were 92.6% for FT and 98.6% for GGT1, respectively.
As required for a good predictor [16], the scores are translated into probabilities of false-positive prediction. For this purpose, a sigmoidal function (analytically based on the extreme-value distribution) is fitted to the distribution of score values calculated from non-prenylatable proteins (see Materials and methods). The general probabilities of false-positive prediction (that complement the specificities to 100%) are estimated to be 0.11% for the FT and 0.02% for the GGT1 predictor, respectively.
Capability to distinguish FT and GGT1 substrates
Previously, the assignment of CaaX box substrate proteins to either FT or GGT1 has been based mainly on the identity of the final residue in the motif (position +3) where FT allows several amino-acid types and GGT1 clearly prefers leucine [13,27]. This view has not changed but it has become clear that several substrates with leucine at position +3 can also be modified (if only to a lesser extent) by FT and not only GGT1. For example, in vitro studies have shown that motifs like CVIL, CVLL, CAIL and CCIL (single-letter amino-acid code) are valid for FT as well [28]. Mutation of the CVIA motif of yeast A-factor to CVIL results in geranylgeranylated as well as farnesylated proteins in vivo [29]. Also, RhoB (with a CKVL motif) is known to be both farnesylated and geranylgeranylated in vivo [30]. Similarily, substrate proteins ending with phenylalanine, such as the CVIF of R-Ras2/TC21, are not specific to either enzyme and can be substrates to FT and GGT1 [31].
In the same way that FT can accept CaaX box motifs ending in leucine and phenylalanine, GGT1 appears to tolerate methionine at this position, which was previously thought to direct farnesylation. This has important consequences in the case of the oncoprotein K-Ras (in variants with CVIM and CIIM motifs) which becomes geranylgeranylated in vivo when farnesyltransferase is inhibited [32].
As we have experienced with our earlier predictors for myristoylation and PTS1 targeting, we find even some correlations of the prediction scores with experimentally measured substrate-enzyme affinities. Interestingly, the scores of the GGT1 predictor give better agreement with the experimental data when divided by 3, in agreement with a threefold lower in vivo activity of GGT1 compared to FT [5]. To estimate the capability of the FT and GGT1 predictors to model the overlapping but distinct substrate specificities, we analyzed a set of heterogeneous substrate motifs that have been measured under the same experimental conditions for their affinities to either FT or GGT1 [5] and we tried to correlate these experimental data with our prediction scores. The set of motifs (CVLS, CIIS, CIIC, CVLF, CVIM, CAIM, CAIV, CAII, CAIL, CVVL, CIIL, and CTIL) contains a large fraction of examples that have been previously shown to be cross-reactive between FT and GGT1 or where the assignment based on simple heuristics depending on hydrophobicity of the final residue fails. In Figure 3, we have plotted the difference of predicted FT and GGT1 scores against the difference of experimentally measured logarithmic affinities for FT and GGT1. A correlation of 0.74 indicates that the theoretical interaction model implemented in the prediction function at least semi-quantitatively resembles the relative substrate specificities between FT and GGT1.
Prediction of prenylation by GGT2
Unlike FT and GGT1, substrate recognition by GGT2 is less dependent on strictly defined carboxy-terminal motifs, but on the complex formation of the substrate with an escort protein [4]. As illustrated in Figure 4, the substrate-escort protein complex then binds to GGT2 (consisting of the alpha and beta subunit typical of prenyltransferases) and, thereby, positioning the flexible substrate carboxy terminus towards the site of modification. Typically, the carboxy-terminal arrangement of cysteines is -XXXCC, -XXCXC, -XXCCX, -XCCXX or -CCXXX and, if available, both cysteines in such a motif will be geranylgeranylated. Currently, only the prenylation of Rab GTPases [33] with the help of Rab escort proteins (REP; two copies in higher organisms, otherwise only one copy) is known for the enzyme GGT2 which is, therefore, also called Rab geranylgeranyltransferase. Reports of lipid modification of fungal casein kinase I apparently represent carboxy-terminal palmitoylation [34] rather than the earlier postulated GGT2 prenylation [35].
Rab proteins are small GTPases (around 60 different have been identified in humans) [36] that share the general fold of the Ras superfamily as well as conserved residues in the nucleotide-binding site. Distinct motifs have been identified that are specific to the Ras, Rho, or Rab families [37]. By virtue of contributing to the binding site of Rabs with their REP, the Rab-specific F3F4 motif can be indirectly used to distinguish possible GGT2 substrates within the Ras superfamily (see sequence logos in Figure 4). However, the REP interaction motif (Rab F3F4) alone could be too short (13 residues) to allow highly sensitive large-scale database scans with thresholds that recognize the learning set (100% self-consistency requires a bit score greater than 5). Interestingly, a search with the final predictor against NCBI's nonredundant database finds only 34 hits with the F3F4 region alone that do not represent Rab proteins or their folds. To avoid these false positives, the hit to the overall alignment of Rab proteins with HMMer [38] (E-value < 0.1) is applied as additional prediction criterion to simulate recognition of the correct fold of related sequences.
Two alignments (F3F4 region and full length) were therefore constructed and after removal of entries with a maximal redundancy of 90% identity over the whole sequence length (117 of 179 entries annotated in Swiss-Prot remaining), hidden Markov models (HMMs) were created and calibrated. The choice of this methodology for the GGT2 prediction was strongly influenced by the fact that the HMMer [38] algorithm is well established in conservatively detecting fold homologies for globular domains at the sequence level. The final GGT2 prediction algorithm checks the carboxy termini for cysteines (at least one cysteine among the five last residues) and parses the HMMer outputs to combine the searches for final results. Estimates of false-positive prediction can be derived from the HMMer E-values.
PrePS: Webinterface and EvOluation
The three tools to predict lipid modification by FT, GGT1 and GGT2 are available as Prenylation Prediction Suite (PrePS), which is accessible online [39]. Users can submit their query sequences to all three or selections of the single predictors. Details of the profile and physical property terms of the scoring function are provided and can also be used to check and rationalize whether and why certain query sequences or artificial constructs intended for membrane targeting might be less suitable prenylation targets. Additionally, an option is provided that allows the user to retrieve homologs of the query protein from NCBI's nonredundant database using BLASTP and automatically annotates them with their respective PrePS results. From the scores for the different predictors (left screenshot in Figure 5) as well as the alignment of the carboxy termini of homologous sequences (right screenshot in Figure 5), the evolutionary motif conservation can be evaluated (evOluation) and used for further rationalization of the biological importance of the predicted motif.
Comparison with alternative methods
Until now, the only available tool to predict protein prenylation has been the Prosite [40] search with the pattern PS00294, which is also used in the PSORT II software [41]. However, this method can neither predict prenylation by GGT2 nor can it distinguish between modifications by FT or GGT1 and, hence, the attached anchor type. During preparation of this paper, an excellent study by Beese, Casey and colleagues [23] has been published that tries to define rules for substrate selectivity by crystallographic analysis of FT and GGT1 complexed with eight cross-reactive substrates. These detailed descriptions of the binding-pocket interactions of a few selected substrate peptides are in good agreement with the motif characteristics identified in this work. While the information gathered from the structural analysis exceeds the capability of any other purely theoretical method to judge interaction for the specific resolved enzyme-substrate pairs, it is difficult to generalize an interaction model from such a small dataset only on the basis of amino-acid constraints at single motif positions. Hence, applying these rules to a more restrictive Prosite-style pattern fails to identify around 30% of substrates experimentally verified in tetrapeptide interaction assays. When taking a closer look at known substrates that are not recognized by the rules of Beese, Casey and colleagues [23] it becomes apparent that this is mainly due to only a few factors. These are the exclusion of leucine at position +3 for alternative FT substrates (known example CKVL of RhoB), the exclusion of phenylalanine at position +3 for alternative FT substrates (known example CVIF of R-Ras2/TC21), the exclusion of glutamine at position +2 for FT substrates (known example serine/threonine kinase 11 or LKB1 with the motif CKQQ) and the exclusion of methionine at position +3 for alternative GGT1 substrates (known example CVIM of K-Ras). In addition, the rules of Beese, Casey and colleagues [23] assign isoleucine and valine at position +3 to GGT1 but not FT substrates. However, these two amino acids were shown to be valid for both FT and GGT1, with at least comparable affinities [13].
The inadequacy of the Beese, Casey and colleagues [23] motif in finding true-positive examples could be counteracted by loosening the motif description, as is already the case in the original Prosite entry PS00294, which nevertheless fails to predict known substrates with glutamine (LKB1) or proline (hepatitis delta antigen) at position +2. However, any reduction in motif stringency concomitantly results in a dramatic increase in the number of false-positive predictions. Table 3 compares typical prediction parameters for the different methods, if applicable. Neither the old nor an adjusted Prosite pattern can compete with the performance of PrePS in finding true substrates while, at the same time, only having a minimal number of false positives. The short Prosite patterns also do not take into account the linker region preceding the CaaX box, which is not defined by clear amino-acid type preferences but rather by general physicochemical property restrictions. The answers of Prosite-style predictions are only binary (yes/no), whereas PrePS gives continuous scores that can be split into interpretable motif-region contributions and that are shown to correlate with experimentally measured relative substrate affinities for FT or GGT1, respectively. Furthermore, only PrePS includes prediction of prenylation by GGT2 and provides an evaluation of evolutionary conservation of the prenylation motif among homologs of the query sequence.
Medical implications and prediction examples
Farnesyltransferase inhibitors (FTIs) have been developed to prevent prenylation of oncogenic Ras proteins and are currently undergoing phase II and III clinical trials [42]. While FTIs have been suggested also to target parasitic diseases [24,43], their efficacy as cancer treatments has been found to be ambivalent in respect of different cancer types. This could be due to the alternative prenylation of oncogenic proteins by GGT1 under FT inhibition, such as K-Ras, in contrast to the total inhibition of prenylation for unique FT substrates, such as H-Ras [2,44]. Identifying these two types of substrate behavior is critical for understanding FTI action as well as identifying their real cellular targets [45,46]. One of the applications of PrePS is in the distinction of substrates that are specific to FT (FTI target) or GGT1 or that are modified by both (less affected by FTIs).
We would like to mention here one example prediction of PrePS for a protein that would be a candidate for a previously unknown FTI target. The human nucleosome assembly protein I-like protein [47] (NAP1-like (GenBank:NP_004528)) has a CKQQ farnesylation motif that is further retained in mouse, rat, frog, fish, fungi and plants, as predicted by PrePS. This taxonomically widespread evolutionary conservation would rather indicate a relevance of the lipid anchor for the function of this protein, which is part of a family involved in transcriptional activation and chromatin formation, including histone binding [48] and nucleocytoplasmic shuttling [49]. The lack of the ability to be alternatively prenylated by GGT1 and, hence, being a unique FT substrate and putative FTI target, is also conserved in the other organisms, possibly pointing to the importance of the specific farnesyl anchor length. It should be noted that this protein is not predicted by the Prosite pattern PS00294 nor by the pattern derived from the rules of a few substrate-enzyme structures [23], but there exist other experimentally verified examples where the same CaaX box motif CKQQ has been shown to be farnesylated (yeast Pex19p [50] and human serine/threonine kinase 11 [51]).
While this paper was in preparation, farnesylation of the NAP1-like protein has been suggested experimentally through a special tagging and purification technique [52], giving support to the PrePS prediction. The same analysis, however, also suggests farnesylation of annexin A2 (GenBank accession number P07355) terminating in a CGGDD motif, which is not at all predicted by PrePS as it is mechanistically unlikely to be processed by farnesyltransferase. Another rather surprising prediction resulting from the tagging experiment is the farnesylation of Rab21 (Q9UL25), which has a double cysteine motif followed by three additional residues (CCSSG) which, at least formally, resembles a CaaX box. Rab proteins with CaaX boxes such as Rab5 (CCSN), Rab8 (CVLL/CSLL), Rab11 (CQNI) and Rab13 (CSLG) are usually modified by GGT2 in vivo [6,53,54] but Rab8 and Rab11 were shown also to be modified by GGT1 and FT in vitro [6,55]. PrePS predicts Rab21 to be geranylgeranylated by GGT2, but the prediction limit for farnesylation is not missed by far. The evOluation shows that the Rab21 orthologs in Xenopus (GenBank AAH60498.1) and Drosophila (AAH60498.1) share the double cysteines but their motif is different and shorter by one residue, pointing to a higher importance of the conservation of the cysteine doublet than the rest of the motif. The evOluation, furthermore, shows that Rab5 is the most closely related prenylated Rab-family member. Interestingly, both cysteines in the CCSN CaaX box motif of Rab5 were shown not only to be geranylgeranylated by GGT2 in vivo but are also required for proper localization and function of the GTPase [54]. Hence, a similar scenario for the two cysteines of the Rab21 prenylation motif cannot be excluded.
A complete analysis of large-scale predictions of prenylated proteins ranked by functional importance as estimated by evolutionary motif conservation and medical implications will be published in a follow-up work.
Materials and methods
Correlation of positional amino-acid frequencies with physical property scales
We identified physicochemical requirements for each motif position by correlating 20-dimensional vectors filled with the positional frequencies of occurrence of the 20 amino-acid types in the carboxy-terminally aligned learning set with a library of over 650 amino-acid physical properties [20,21]. This has been done over a largest subset of the learning set with removed redundancy of greater than 40% identity in the last 30 positions (nonredundant learning set = nrLS) and over positional vectors filled with frequencies derived from the profile (= prof) that has been corrected for redundancy with the position-specific independent counts (PSIC) method [56]. Such correlations have been estimated previously [12] to be significant for confidence levels α = 0.0025 and α = 0.001 if the values are greater than 0.62 and 0.7, respectively.
Fisher criterion to find interpositional correlations
The Fisher ratio F of the sum of variances of single positions with the variance over multiple positions for pairs and triplets of positions is calculated, allowing gaps of up to two residues between pairs.
The F values for probabilities p = 0.05 are taken as thresholds for evaluating significance of interdependence of physical properties between motif positions. These are 1.288 for the FT and 1.355 for the GGT1 nonredundant learning set, respectively.
Enzyme-specific binding pocket residue conservation
To identify regions in the FT and GGT1 binding pockets that are characteristic for the respective enzyme, we analyzed the pattern of residue conservation in vicinity to the CaaX substrate peptide and the prenylpyrophosphate (Figures 2 and 5). Alignments were created with FT and GGT1 beta subunits of a diverse subset of organisms (see legend of Figure 6) using ClustalX [57]. The conservation of alignment positions was measured using al2co [58] with the Henikoff-weighted variance-based options. The enzyme-specific conservation was calculated as conservation values in the FT or GGT1 alignments minus conservation values in a joined FT+GGT1 alignment (Figure 6) and then mapped to the binding pocket surface using a customized script in Swiss-Pdb Viewer [59]. Color transitions from white to yellow to red represent increasing conservation difference, where red signifies the highest enzyme-specific conservation. White parts are not characteristically different between FT and GGT1, but might also be well conserved between the two enzymes. As the observed differences in the binding pocket conservation relate to features of the enzyme and not the substrates, this information could not be taken directly into account in our scoring scheme. Indirectly, however, the resulting image mirrors the requirement for conservation also for the substrate peptides and, thereby, confirms the relative importance of motif positions as estimated by the Shannon entropy (see below).
Prediction score function
Construction and calculation of the scoring function essentially follows the methodologies [7,16,60] applied to the prediction of GPI anchors [8], myristoylation [9] and PTS1 peroxisomal targeting [10] and is summarized shortly below with emphasis on additions and problem-specific variations.
The composite score function consists of a profile and physical property term.
S = Sprofile + Sppt (2)
The profile term Sprofile
The profile matrix is calculated using the PSIC algorithm to account for redundancy in the learning set. The frequency of occurrence f(a,i) of an amino acid type a at a given alignment position i is down-weighted proportionally to the number of other positions that are identical in sequences sharing a at i, resulting in subscores SPSIC(a,i) representing the natural logarithms of these redundancy-corrected frequencies. These are summed over the respective regions (+3, +2, +1 and -1 to -11)
before they enter Sprofile:
αprofile is a normalization factor to compensate for differing lengths of the regions and is derived from:
Cregion is a relative weighting of motif regions. We propose a rational basis for its approximated computation. Cregion is derived from the average Shannon entropies of the regions in the alignment of the nonredundant learning set sequences. Shannon's information theoretic entropy has already previously been investigated as measurement of sequence variability in alignment positions and residue conservation [61]. We calculate the average of the exponential of the negative Shannon entropy as conservation measurement
where na,i is the number of occurrences of amino-acid type a at the investigated position i of the region and N the total number of sequences in the alignment. Only amino-acid types that occur at least once (na,i ≥ 1) are included in the sum. In order to avoid overly precise values, the final values are rounded up to the last decimal that showed variation in the conservation measure when comparing the small learning set based on Swiss-Prot annotation with the larger learning set described in this work. The position of the absolutely conserved cysteine in the CaaX box would have the maximum conservation value of 1. Table 4 lists Cregion for the regions +1, +2, +3 and -11 to -1.
The physical property term Sppt
The primary role of the physical property term is to exclude query sequences that do not fit into the determined physical property profile of the motif (see Figure 1). Hence, it does not add positive scores for compliance but only penalizes deviations from physical property requirements of the motif. These requirements have been identified as described above (listed in Tables 1 and 2; including rationale for inclusion into Sppt) and the biophysical meaning discussed earlier.
As an example, a penalty for a physical property P of the query being greater than the average physical property Pj over the nonredundant learning set (with σj being the corresponding standard deviation of a Gauss-like distribution) can be defined as follows:
The natural logarithms (to be comparable to the profile score) of these single terms enter Sppt as a sum with a term-specific weighting factor αj that emphasizes varying importance of the single terms.
As part of Sppt, the FT and GGT1 score functions also include a penalty for query proteins when the carboxy terminus appears inaccessible as a result of structural constraints. For example, helices with strongly hydrophobic sides are either buried in the structure or fold back to the protein core and reduce the flexibility and accessibility of the carboxy terminus. We recognize these if hydrophobic residues (LIVMFYW) appear in patterns like i,i+3,i+6 (hXXhXXh) or i,i+3,i+7 (hXXhXXXh) or i,i+4,i+7 (hXXXhXXh) within 20 residues preceding the cysteine of the CaaX box.
Cross-validation (jackknife) tests
We have performed three jackknife tests to validate the robustness of our predictors. First, cross-validation over the whole scoring function (Sprofile + Sppt recalculated after removal of the entry to be validated from the learning set) resulted in sensitivities of 99.4% for FT and 99.8% for GGT1, respectively. This generally indicates that the learning sets are large enough for the method not to suffer from removal of individual entries and, hence it should also be able to find valid motifs not yet included in the learning set. At the same time, we acknowledge that the similarity of sequences within homologous groups in the learning set influence the above estimated sensitivity. However, we emphasize that the values calculated for the profile matrix vary even when only one sequence from a group of homologs is left out, since the method used for profile extraction [56] does take into account such redundancy with both sequence- and position-specific weightings.
To address the role of homologous sequences remaining in the learning set for jackknife tests, we have extended the cross-validation to the more stringent case where not only the sequence to be predicted is excluded from the learning procedure but also its close homologs (estimated by a threshold of 40% sequence identity over the 30 carboxy-terminal residues). We find that in the case of the predictor for GGT1, the sensitivity of 98.6% is very close to the first jackknife test. For the FT predictor, a sensitivity of 92.6% is obtained in this test, which is only slightly lower than in the other jackknife procedures. The decrease can be almost exclusively attributed to the large group of highly similar sequences of hepatitis delta antigen that all share the uncommon motif CRPQ with both arginine at position +1 and proline at +2 being rather exceptional amino acids. Hence, the CRPQ sequences remain below the prediction threshold in the jackknife test if no example of this motif is present in the learning set. When leaving the CRPQ group out of the cross-validation, the FT predictor is calculated to have a sensitivity of 97.9%.
A third cross-validation test aims to elucidate whether the 39 parameters (13 terms with weightings, averages and variances, see Equations 7 and 8) introduced through the physical property terms in Sppt are overfitting the learning data. To exclude bias through similar sequences in homologous groups, the test was executed only over the learning set after removing redundancy (see section on correlation analysis above). Sppt alone is recalculated after removal of the entry to be validated from the parameter calculation procedure. The obtained sensitivities of 100% for FT and 99.8% for GGT1 indicate that the parameterizations of physical property terms in Sppt are not overfitting the learning data.
Probability of false-positive prediction
To estimate probabilities of false-positive prediction, a set of sequences had to be defined whose carboxy-terminal amino-acid pattern should not be subject to selection for a valid CaaX prenylation motif. This is fulfilled for sequences in the NCBI nonredundant database lacking a cysteine at the fourth position from the carboxy terminus. Hence, any compliance with motif restrictions of such sequences apart from the fourth last position could only be attributed to random or at least prenylation-independent appearance. Following earlier experience with GPI- and myristoyl-anchor prediction [8,9], a polynomial extreme-value distribution function has been fitted to the scores obtained for the described sequence set when ignoring the requirement of a CaaX box cysteine in the procedure. The probability of obtaining a score S greater than a threshold score Sth can be formulated as follows:
A polynom of the sixth degree was used to improve the residual fit. Polynoms with degrees higher than six would not result in an increase of the relative improvement of the residual. The probabilities of false-positive prediction for scores at the prediction threshold are extrapolated to 6.3% for FT and 1.2% for GGT1 for sequences that contain a cysteine at the fourth-last position (canonical CaaX box). Given that appearance of a cysteine at this position is rare in databases (1.7%), the independent general probabilities of false-positive prediction by FT and GGT1 for all protein sequences are as low as 0.11% and 0.02%, respectively. This corresponds to specificities of 99.89% and 99.98%.
Since the subcellular context of a protein can be relevant to judging the likelihood of in vivo prenylation when a corresponding motif has been predicted, we also tried to estimate probabilities of false-positive prediction for differently localized subsets of proteins. These subsets were retrieved from the Swiss-Prot database [22] and assigned according to the following keywords in the 'Subcellular Location' comment lines: 'cytoplasmic' (24,284), 'nuclear' (9800), 'membrane' (24,823) and 'extracellular' (509). Parentheses indicate the number of unambiguously annotated examples per subset. Again, only proteins lacking a -CXXX motif were taken into account, to ensure prenylation-independent selection of carboxy-terminal amino-acid residues. Although there appear to be some subcellular localization-specific fluctuations of the probabilities of false-positive prediction (partly due to limited or differing subset sizes), the relative advantages among the methods evaluated seem to remain stable (see Table 3).
Acknowledgements
We thank Boehringer Ingelheim for continuous support. This project has been partly funded by the Fonds zur Förderung der wissenschaftlichen Forschung Österreichs (FWF grant P15037), the Austrian National Bank (Österreichische Nationalbank) and by the bioinformatics contract study 2002-2004 for Bundesministerium fuer Wirtschaft und Arbeit.
Figures and Tables
Figure 1 Sequence logos [74] and physicochemical property profiles of FT and GGT1 substrates. Selected physical properties (hydrophilicity = KRIW790102; flexibility = KARP850103, size = CHOC760101; aliphatic = ZVEL_ALI_1; see Tables 1 and 2 for details) are calculated as average over the nonredundant learning sets of FT and GGT1. The plotted lines correspond to the relative deviation of the respective properties from an average calculated over carboxy termini from the UniRef50 database [22].
Figure 2 The two CaaX prenyltransferases. (a) Ribbon representations of FT (PDB 1D8D [75]) and GGT1 (PDB 1N4Q [76]); pink, alpha subunit; yellow, beta subunit. (b) The prenylpyrophosphates (green) and CaaX tetrapeptides (blue) inside the binding pockets with enzyme-specific conservation (conservation in FT or GGT1 minus conservation in joined FT+GGT1 alignment) mapped to binding-pocket surface. Increasing conservation difference is shaded from white to yellow to red. FPP, farnesyl-, GGPP, geranylgeranylpyrophosphate. The alignment of the sequences of these proteins is shown in Figure 6. Visualized with Swiss-Pdb Viewer [59].
Figure 3 Correlation between predicted and experimental FT/GGT1 substrate selectivity. The correlation of the difference between predicted FT and GGT1 scores with the difference of the experimentally measured logarithmic affinities for FT and GGT1 of the same substrates is plotted.
Figure 4 Determinants of GGT2 prenylation. (a) Sequence logos [74] of Ras superfamily members around part of the Rab-REP interaction site (colored red in the otherwise yellow GTPase structure). (b) Structural model of the Rab-REP-GGT2 prenylation complex based on PDB entries 1LTX [77] and 1VG0 [4]. REP1 (green) has a prenyl-binding pocket which is proposed to be involved in the dual geranylgeranylation mechanism (bound geranylgeranyl is shown in green). However, the catalytic attachment to the substrate cysteines takes place in the center of the GGT2 alpha-beta complex (light and dark blue) where the prenylpyrophosphate that will be transferred is also bound (blue space-filling representation, zinc in red). The structure was visualized using Swiss-Pdb Viewer [59].
Figure 5 Screenshot of the output provided by the PrePS server [39]. On the left is the prediction result for the query protein H-Ras (GenBank P01112) and the three prenylating enzymes. On the right, is shown the carboxy-terminal alignment and PrePS predictions of homologs of the query protein for evaluation of evolutionary motif conservation. Note that H-Ras is predicted to be prenylated only by FT, whereas the homologs K-Ras and N-Ras can also be prenylated by GGT1.
Figure 6 Alignment of FT and GGT1 beta subunits (FTb, GGT1b) in the regions of binding-pocket residues (marked with arrow) using ClustalX [57]. Residue ranges shown above and below correspond to the numbering in the PDB structures of rat FT beta (PDB 1D8D) and GGT1 beta (PDB 1N4Q), respectively. Accession numbers are as follows (GenBank unless indicated otherwise): Hs (Homo sapiens) FTb, NP_002019; GGT1b, NP_005014; Mm (Mus musculus) NP_666039; NP_766215; Rn (Rattus norvegicus) PDB 1D8D; 1N4Q; Tn (Tetraodon nigroviridis) CAG09215; CAF904630; Dm (Drosophila melanogaster) NP_650540; NP_525100; Ag (Anopheles gambiae) XP_321357; XP_317045; Ce (Caenorhabditis elegans) NP_506580; NP_496848; At (Arabidopsis thaliana) NP_198844; NP_181487; Sp (Schizosaccharomyces pombe) NP_594251; NP_594142; Sc (Saccharomyces cerevisiae) P22007; NP_011360. Standard ClustalX coloring (according to conserved amino acid type).
Table 1 Physical property terms in the FT scoring function
Property Position Rationale Explanation
ARGP820103 [62] +3 Corr = 0.7(nrLS) Membrane-buried preference, lipid contact when entering binding pocket
logPREN_CKQX_FT [15] +3 Corr = -0.72(nrLS) Kinetic measurement, relative unprocessed FPP amounts with tetrapeptide CKQX
CHOC760101 [63] +1 to +3 Fisher = 1.3 Side chain volume
ZVEL_CHARG [64] +1 to +3 LS composition General charge penalty
ZVEL_CHNEG [64] +1 to +3 LS composition Special negative charge penalty
WERD780102 [65] +1 and +3 Fisher = 1.51 Hydrophobicity compensation for inside preference
ZVEL_ALI_1 [64] +1 and +2 +2: Corr = 0.85(prof)
+1: continuing deviation from Uniref50 average Amino-acid property: aliphatic
LIFS790102 [66] +1 and +2 +2: Correlation = 0.76(prof)
+1: continuing deviation from Uniref50 average Preference for extended conformations
ZVEL_TINY_ [64] -1 Corr = 0.68(prof) Size, bulkiness
MOBILITY_2 [21] -1 Corr = 0.61(nrLS) Side chain mobility
VINM940101 [67] -11 to -1 -2: Corr = 0.72(prof)
-3: Corr = 0.75(prof)
-4: Corr = 0.78(nrLS)
-5: Corr = 0.82(nrLS)
-6: Corr = 0.84(nrLS)
-7: Corr = 0.79(nrLS)
-8: Corr = 0.74(prof)
-9: Corr = 0.82(nrLS)
-10: Corr = 0.84(nrLS)
-11: Corr = 0.79(nrLS)
Rest: continuing deviation from Uniref50 average Normalized flexibility average
KRIW790102 [68] -11 to -1 -2: Corr = 0.76(prof)
-6: Corr = 0.83(nrLS)
-7: Corr = 0.83(nrLS)
-8: Corr = 0.76(prof)
Rest: continuing deviation from Uniref50 average Fraction of site occupied with water
Buried helix (see Materials and methods) -20 to -1 Remove false positives Helix with strongly hydrophobic sides folds back to protein core and reduces flexibility and accessibility of C-terminus
Corr, correlation; LS, learning set; nrLS, nonredundant; prof, profile.
Table 2 Physical property terms in the GGT1 scoring function
Property Position Rationale Explanation
ARGP820103 [62] +3 Corr = 0.8(prof) Membrane-buried preference, lipid contact when entering binding pocket
LEVM760105 [69] +1 to +3 Fisher = 1.36 Size limitation (radius of gyration of side-chain)
YUTK870101 [70] +1 to +3 Fisher = 1.38 Hydrophobicity compensation (Unfolding Gibbs energy in water, pH7.0)
ZVEL_CHARG [64] +1 to +3 LS composition General charge penalty
ZVEL_CHNEG [64] +1 to +3 LS composition Special negative charge penalty
ZVEL_ALI_1 [64] +1 and +2 +2: Corr = 0.87(prof)
+1: continuing deviation from Uniref50 average Amino-acid property: aliphatic
LIFS790102 [66] +1 and +2 +2: Corr = 0.77(prof)
+1: continuing deviation from Uniref50 average Preference for extended conformations
FAUJ880101 [71] -1 and +2 Fisher = 1.52 Size, bulkiness (residues although 10 Å apart, face to same side of base pair)
FINA910103 [72] -1 Corr = 0.75(prof) Helix termination (for example, K, S favored, D,E,L,I,V disfavored)
KARP850103 [73] -7 to-1 -1: Corr = 0.69(prof)
-2: Corr = 0.70(prof)
-3: Corr = 0.71(prof)
-4: Corr = 0.74(nrLS)
-5: Corr = 0.75(prof)
-6: Corr = 0.70(nrLS)
-7: Corr = 0.78(nrLS) Flexibility (GGT1 lysine preference)
VINM940101 [67] -11 to -1 -4: Corr = 0.72(prof)
-5: Corr = 0.82(prof)
-6: Corr = 0.84(nrLS)
-7: Corr = 0.75(nrLS)
-8: Corr = 0.77(nrLS)
-9: Corr = 0.68(prof)
-10: Corr = 0.86(prof)
Rest: continuing deviation from Uniref50 average Normalized flexibility average
KRIW790102 [68] -11 to -1 -3: Corr = 0.70(prof)
-4: Corr = 0.73(prof)
-5: Corr = 0.84(prof)
-6: Corr = 0.81(prof)
-7: Corr = 0.83(nrLS)
-8: Corr = 0.85(nrLS)
-9: Corr = 0.76(prof)
-10: Corr = 0.86(prof)
Rest: continuing deviation from Uniref50 average Fraction of site occupied with water
Buried helix (see Materials and methods) -20 to -1 Remove false positives Helix with strongly hydrophobic sides folds back to protein core and reduces flexibility and accessibility of carboxy terminus
Table 3 Comparison of prediction performances
FT GGT1
Prosite PS00294 Beese, Casey and colleagues' rules PrePS FT Prosite PS00294 Beese, Casey and colleagues' rules PrePS GGT1
Sensitivity I 85%* 72% 100% 95%* 67% 100%
Sensitivity II NA NA 92.6/97.9%† NA NA 98.6%
Probability of false positive prediction (POFP) for -CXXX motifs (GenBank sequences) 17.1%* 9.9% 6.3% 17.1%* 10.0% 1.2%
POFP -CXXX 'cytoplasmic'‡ 18.2%* 8.9% 5.1% 18.2%* 8.6% 1.4%
POFP -CXXX 'nuclear'‡ 13.9%* 10.5% 5.5% 13.9%* 9.6% 1.1%
POFP -CXXX 'membrane'‡ 17.5%* 10.3% 3.8% 17.5%* 12.0% 0.8%
POFP --CXXX 'extracellular'$ 8.6%* 7.9% 3.3% 8.6%* 9.0% 0.2%
Overall probability of false positive prediction (GenBank sequences, assuming 1.7% with -CXXX) 0.29%* 0.16% 0.11% 0.29%* 0.17% 0.02%
*Prosite pattern PS00294 does not distinguish between prenylation by FT and GGT1.
†Sensitivity rises to 97.9% when the exceptional motif CRPQ of hepatitis delta antigen is removed. ‡For details see Materials and methods. Sensitivity I is the rate of finding known substrates from described learning set = self-consistency. Sensitivity II is the rate of finding known substrates after their exclusion (including homologs) from the learning set = cross-validation (see Materials and methods). Probabilities of false-positive predictions (POFP) complement the specificities to 100% (Specificity = 100 - POFP). The first listed POFP estimates the rates of false positives among query proteins that have a canonical -CXXX motif (which corresponds to 1.7% of all sequences). Below are estimations of POFPs for subsets of Swiss-Prot proteins that differ in their annotated subcellular localization (see Materials and methods). The final POFP is the estimate for false-positive predictions for all sequences (for example, when analyzing complete proteomes or large databases), independent of existence of a -CXXX motif. Formatting signifies: best (bold), intermediate (plain text), worst (italic) performance.
Table 4 Relative weightings of motif positions in profile term
Position(s) FT GGT1
-11 to -1 0.07 0.07
+1 0.15 0.14
+2 0.29 0.41
+3 0.17 0.42
==== Refs
Casey PJ Seabra MC Protein prenyltransferases. J Biol Chem 1996 271 5289 5292 8621375 10.1074/jbc.271.10.5289
Maurer-Stroh S Washietl S Eisenhaber F Protein prenyltransferases. Genome Biol 2003 4 212 12702202 10.1186/gb-2003-4-4-212
Roskoski R Jr Protein prenylation: a pivotal posttranslational process. Biochem Biophys Res Commun 2003 303 1 7 12646157 10.1016/S0006-291X(03)00323-1
Rak A Pylypenko O Niculae A Pyatkov K Goody RS Alexandrov K Structure of the Rab7:REP-1 complex: insights into the mechanism of Rab prenylation and choroideremia disease. Cell 2004 117 749 760 15186776 10.1016/j.cell.2004.05.017
Caplin BE Hettich LA Marshall MS Substrate characterization of the Saccharomyces cerevisiae protein farnesyltransferase and type-I protein geranylgeranyltransferase. Biochim Biophys Acta 1994 1205 39 48 8142482
Wilson AL Erdman RA Castellano F Maltese WA Prenylation of Rab8 GTPase by type I and type II geranylgeranyl transferases. Biochem J 1998 333 497 504 9677305
Eisenhaber F Eisenhaber B Maurer-Stroh S Andrade MM Prediction of post-translational modifications from amino acid sequence: problems, pitfalls, methodological hints. Bioinformatics and Genomes: Current Perspectives 51 2003 Wymondham, UK: Horizon Scientific Press 81 105
Eisenhaber B Bork P Eisenhaber F Prediction of potential GPI-modification sites in proprotein sequences. J Mol Biol 1999 292 741 758 10497036 10.1006/jmbi.1999.3069
Maurer-Stroh S Eisenhaber B Eisenhaber F N-terminal N-myristoylation of proteins: prediction of substrate proteins from amino acid sequence. J Mol Biol 2002 317 541 557 11955008 10.1006/jmbi.2002.5426
Neuberger G Maurer-Stroh S Eisenhaber B Hartig A Eisenhaber F Prediction of peroxisomal targeting signal 1 containing proteins from amino acid sequence. J Mol Biol 2003 328 581 592 12706718 10.1016/S0022-2836(03)00319-X
Boeckmann B Bairoch A Apweiler R Blatter MC Estreicher A Gasteiger E Martin MJ Michoud K O'Donovan C Phan I The SWISS-PROT protein knowledgebase and its supplement TrEMBL in 2003. Nucleic Acids Res 2003 31 365 370 12520024 10.1093/nar/gkg095
Maurer-Stroh S Eisenhaber B Eisenhaber F N-terminal N-myristoylation of proteins: refinement of the sequence motif and its taxon-specific differences. J Mol Biol 2002 317 523 540 11955007 10.1006/jmbi.2002.5425
Roskoski R JrRitchie P Role of the carboxyterminal residue in peptide binding to protein farnesyltransferase and protein geranylgeranyltransferase. Arch Biochem Biophys 1998 356 167 176 9705207 10.1006/abbi.1998.0768
Maurer-Stroh S Gouda M Novatchkova M Schleiffer A Schneider G Sirota FL Wildpaner M Hayashi N Eisenhaber F MYRbase: analysis of genome-wide glycine myristoylation enlarges the functional spectrum of eukaryotic myristoylated proteins. Genome Biol 2004 5 R21 15003124 10.1186/gb-2004-5-3-r21
Boutin JA Marande W Petit L Loynel A Desmet C Canet E Fauchere JL Investigation of S-farnesyl transferase substrate specificity with combinatorial tetrapeptide libraries. Cell Signal 1999 11 59 69 10206346 10.1016/S0898-6568(98)00032-1
Eisenhaber B Eisenhaber F Maurer-Stroh S Neuberger G Prediction of sequence signals for lipid post-translational modifications: insights from case studies. Proteomics 2004 4 1614 1625 15174131 10.1002/pmic.200300781
Valid CaaX tetrapeptides at the PrePS site
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. Nucleic Acids Res 1997 25 3389 3402 9254694 10.1093/nar/25.17.3389
Enright AJ Van Dongen S Ouzounis CA An efficient algorithm for large-scale detection of protein families. Nucleic Acids Res 2002 30 1575 1584 11917018 10.1093/nar/30.7.1575
Tomii K Kanehisa M Analysis of amino acid indices and mutation matrices for sequence comparison and structure prediction of proteins. Protein Eng 1996 9 27 36 9053899
Eisenhaber B Bork P Eisenhaber F Sequence properties of GPI-anchored proteins near the omega-site: constraints for the polypeptide binding site of the putative transamidase. Protein Eng 1998 11 1155 1161 9930665 10.1093/protein/11.12.1155
Apweiler R Bairoch A Wu CH Barker WC Boeckmann B Ferro S Gasteiger E Huang H Lopez R Magrane M UniProt: the Universal Protein knowledgebase. Nucleic Acids Res 2004 32 Database D115 D119 14681372 10.1093/nar/gkh131
Reid TS Terry KL Casey PJ Beese LS Crystallographic analysis of CaaX prenyltransferases complexed with substrates defines rules of protein substrate selectivity. J Mol Biol 2004 343 417 433 15451670 10.1016/j.jmb.2004.08.056
Maurer-Stroh S Washietl S Eisenhaber F Protein prenyltransferases: anchor size, pseudogenes and parasites. Biol Chem 2003 384 977 989 12956414 10.1515/BC.2003.110
Prakash B Praefcke GJ Renault L Wittinghofer A Herrmann C Structure of human guanylate-binding protein 1 representing a unique class of GTP-binding proteins. Nature 2000 403 567 571 10676968 10.1038/35000617
Neuberger G Maurer-Stroh S Eisenhaber B Hartig A Eisenhaber F Motif refinement of the peroxisomal targeting signal 1 and evaluation of taxon-specific differences. J Mol Biol 2003 328 567 579 12706717 10.1016/S0022-2836(03)00318-8
Moores SL Schaber MD Mosser SD Rands E O'Hara MB Garsky VM Marshall MS Pompliano DL Gibbs JB Sequence dependence of protein isoprenylation. J Biol Chem 1991 266 14603 14610 1860864
Boutin JA Marande W Goussard M Loynel A Canet E Fauchere JL Chromatographic assay and peptide substrate characterization of partially purified farnesyl- and geranylgeranyltransferases from rat brain cytosol. Arch Biochem Biophys 1998 354 83 94 9633601 10.1006/abbi.1998.0678
Caldwell GA Wang SH Naider F Becker JM Consequences of altered isoprenylation targets on a-factor export and bioactivity. Proc Natl Acad Sci USA 1994 91 1275 1279 8108401
Prendergast GC Actin' up: RhoB in cancer and apoptosis. Nat Rev Cancer 2001 1 162 168 11905808 10.1038/35101096
Carboni JM Yan N Cox AD Bustelo X Graham SM Lynch MJ Weinmann R Seizinger BR Der CJ Barbacid M Farnesyltransferase inhibitors are inhibitors of Ras but not R-Ras2/TC21 transformation. Oncogene 1995 10 1905 1913 7761092
Rowell CA Kowalczyk JJ Lewis MD Garcia AM Direct demonstration of geranylgeranylation and farnesylation of Ki-Ras in vivo. J Biol Chem 1997 272 14093 14097 9162034 10.1074/jbc.272.22.14093
Pereira-Leal JB Hume AN Seabra MC Prenylation of Rab GTPases: molecular mechanisms and involvement in genetic disease. FEBS Lett 2001 498 197 200 11412856 10.1016/S0014-5793(01)02483-8
Roth AF Feng Y Chen L Davis NG The yeast DHHC cysteine-rich domain protein Akr1p is a palmitoyl transferase. J Cell Biol 2002 159 23 28 12370247 10.1083/jcb.200206120
Vancura A Sessler A Leichus B Kuret J A prenylation motif is required for plasma membrane localization and biochemical function of casein kinase I in budding yeast. J Biol Chem 1994 269 19271 19278 8034689
Pereira-Leal JB Seabra MC Evolution of the Rab family of small GTP-binding proteins. J Mol Biol 2001 313 889 901 11697911 10.1006/jmbi.2001.5072
Pereira-Leal JB Seabra MC The mammalian Rab family of small GTPases: definition of family and subfamily sequence motifs suggests a mechanism for functional specificity in the Ras superfamily. J Mol Biol 2000 301 1077 1087 10966806 10.1006/jmbi.2000.4010
Eddy SR Profile hidden Markov models. Bioinformatics 1998 14 755 763 9918945 10.1093/bioinformatics/14.9.755
PrePS - Prenylation Prediction Suite
Falquet L Pagni M Bucher P Hulo N Sigrist CJ Hofmann K Bairoch A The PROSITE database, its status in 2002. Nucleic Acids Res 2002 30 235 238 11752303 10.1093/nar/30.1.235
Horton P Nakai K Better prediction of protein cellular localization sites with the k nearest neighbors classifier. Proc Int Conf Intell Syst Mol Biol 1997 5 147 152 9322029
Doll RJ Kirschmeier P Bishop WR Farnesyltransferase inhibitors as anticancer agents: critical crossroads. Curr Opin Drug Discov Devel 2004 7 478 486 15338957
Gelb MH Van Voorhis WC Buckner FS Yokoyama K Eastman R Carpenter EP Panethymitaki C Brown KA Smith DF Protein farnesyl and N-myristoyl transferases: piggy-back medicinal chemistry targets for the development of antitrypanosomatid and antimalarial therapeutics. Mol Biochem Parasitol 2003 126 155 163 12615314 10.1016/S0166-6851(02)00282-7
Sebti SM Hamilton AD Farnesyltransferase and geranylgeranyltransferase I inhibitors and cancer therapy: lessons from mechanism and bench-to-bedside translational studies. Oncogene 2000 19 6584 6593 11426643 10.1038/sj.onc.1204146
Tamanoi F Gau CL Jiang C Edamatsu H Kato-Stankiewicz J Protein farnesylation in mammalian cells: effects of farnesyltransferase inhibitors on cancer cells. Cell Mol Life Sci 2001 58 1636 1649 11706990
Sebti SM Der CJ Opinion: Searching for the elusive targets of farnesyltransferase inhibitors. Nat Rev Cancer 2003 3 945 951 14737124 10.1038/nrc1234
Simon HU Mills GB Kozlowski M Hogg D Branch D Ishimi Y Siminovitch KA Molecular characterization of hNRP, a cDNA encoding a human nucleosome-assembly-protein-I-related gene product involved in the induction of cell proliferation. Biochem J 1994 297 389 397 8297347
McBryant SJ Park YJ Abernathy SM Laybourn PJ Nyborg JK Luger K Preferential binding of the histone (H3-H4)2 tetramer by NAP1 is mediated by the amino-terminal histone tails. J Biol Chem 2003 278 44574 44583 12928440 10.1074/jbc.M305636200
Miyaji-Yamaguchi M Kato K Nakano R Akashi T Kikuchi A Nagata K Involvement of nucleocytoplasmic shuttling of yeast Nap1 in mitotic progression. Mol Cell Biol 2003 23 6672 6684 12944491 10.1128/MCB.23.18.6672-6684.2003
Gotte K Girzalsky W Linkert M Baumgart E Kammerer S Kunau WH Erdmann R Pex19p, a farnesylated protein essential for peroxisome biogenesis. Mol Cell Biol 1998 18 616 628 9418908
Sapkota GP Kieloch A Lizcano JM Lain S Arthur JS Williams MR Morrice N Deak M Alessi DR Phosphorylation of the protein kinase mutated in Peutz-Jeghers cancer syndrome, LKB1/STK11, at Ser431 by p90(RSK) and cAMP-dependent protein kinase, but not its farnesylation at Cys(433), is essential for LKB1 to suppress cell growth. J Biol Chem 2001 276 19469 19482 11297520 10.1074/jbc.M009953200
Kho Y Kim SC Jiang C Barma D Kwon SW Cheng J Jaunbergs J Weinbaum C Tamanoi F Falck J Zhao Y A tagging-via-substrate technology for detection and proteomics of farnesylated proteins. Proc Natl Acad Sci USA 2004 101 12479 12484 15308774 10.1073/pnas.0403413101
Gromov P Celis JE Rab11a is modified in vivo by isoprenoid geranylgeranyl. Electrophoresis 1998 19 1803 1807 9719562 10.1002/elps.1150191043
Gomes AQ Ali BR Ramalho JS Godfrey RF Barral DC Hume AN Seabra MC Membrane targeting of Rab GTPases is influenced by the prenylation motif. Mol Biol Cell 2003 14 1882 1899 12802062 10.1091/mbc.E02-10-0639
Joberty G Tavitian A Zahraoui A Isoprenylation of Rab proteins possessing a C-terminal CaaX motif. FEBS Lett 1993 330 323 328 8375503 10.1016/0014-5793(93)80897-4
Sunyaev SR Eisenhaber F Rodchenkov IV Eisenhaber B Tumanyan VG Kuznetsov EN PSIC: profile extraction from sequence alignments with position-specific counts of independent observations. Protein Eng 1999 12 387 394 10360979 10.1093/protein/12.5.387
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
Pei J Grishin NV AL2CO: calculation of positional conservation in a protein sequence alignment. Bioinformatics 2001 17 700 712 11524371 10.1093/bioinformatics/17.8.700
Guex N Peitsch MC SWISS-MODEL and the Swiss-PdbViewer: an environment for comparative protein modeling. Electrophoresis 1997 18 2714 2723 9504803 10.1002/elps.1150181505
Eisenhaber F Eisenhaber B Kubina W Maurer-Stroh S Neuberger G Schneider G Wildpaner M Prediction of lipid posttranslational modifications and localization signals from protein sequences: big-Pi, NMT and PTS1. Nucleic Acids Res 2003 31 3631 3634 12824382 10.1093/nar/gkg537
Valdar WS Scoring residue conservation. Proteins 2002 48 227 241 12112692 10.1002/prot.10146
Argos P Rao JK Hargrave PA Structural prediction of membrane-bound proteins. Eur J Biochem 1982 128 565 575 7151796
Chothia C The nature of the accessible and buried surfaces in proteins. J Mol Biol 1976 105 1 12 994183 10.1016/0022-2836(76)90191-1
Zvelebil MJ Barton GJ Taylor WR Sternberg MJ Prediction of protein secondary structure and active sites using the alignment of homologous sequences. J Mol Biol 1987 195 957 961 3656439 10.1016/0022-2836(87)90501-8
Wertz DH Scheraga HA Influence of water on protein structure. An analysis of the preferences of amino acid residues for the inside or outside and for specific conformations in a protein molecule. Macromolecules 1978 11 9 15 621952 10.1021/ma60061a002
Lifson S Sander C Antiparallel and parallel beta-strands differ in amino acid residue preferences. Nature 1979 282 109 111 503185
Vihinen M Torkkila E Riikonen P Accuracy of protein flexibility predictions. Proteins 1994 19 141 149 8090708 10.1002/prot.340190207
Krigbaum WR Komoriya A Local interactions as a structure determinant for protein molecules: II. Biochim Biophys Acta 1979 576 204 248 760806
Levitt M A simplified representation of protein conformations for rapid simulation of protein folding. J Mol Biol 1976 104 59 107 957439 10.1016/0022-2836(76)90004-8
Yutani K Ogasahara K Tsujita T Sugino Y Dependence of conformational stability on hydrophobicity of the amino acid residue in a series of variant proteins substituted at a unique position of tryptophan synthase alpha subunit. Proc Natl Acad Sci USA 1987 84 4441 4444 3299367
Fauchere JL Charton M Kier LB Verloop A Pliska V Amino acid side chain parameters for correlation studies in biology and pharmacology. Int J Pept Protein Res 1988 32 269 278 3209351
Finkelstein AV Badretdinov AY Ptitsyn OB Physical reasons for secondary structure stability: alpha-helices in short peptides. Proteins 1991 10 287 299 1946339 10.1007/BF01025627
Karplus PA Schulz GE Prediction of chain flexibility in proteins. A tool for the selection of peptide antigens Naturwiss 1985 72 212 213 10.1007/BF01195768
Crooks GE Hon G Chandonia JM Brenner SE WebLogo: a sequence logo generator. Genome Res 2004 14 1188 1190 15173120 10.1101/gr.849004
Long SB Casey PJ Beese LS The basis for K-Ras4B binding specificity to protein farnesyltransferase revealed by 2 A resolution ternary complex structures. Structure Fold Des 2000 8 209 222 10673434 10.1016/S0969-2126(00)00096-4
Taylor JS Reid TS Terry KL Casey PJ Beese LS Structure of mammalian protein geranylgeranyltransferase type-I. EMBO J 2003 22 5963 5974 14609943 10.1093/emboj/cdg571
Pylypenko O Rak A Reents R Niculae A Sidorovitch V Cioaca MD Bessolitsyna E Thoma NH Waldmann H Schlichting I Structure of rab escort protein-1 in complex with rab geranylgeranyltransferase. Mol Cell 2003 11 483 494 12620235 10.1016/S1097-2765(03)00044-3
| 15998445 | PMC1175987 | CC BY | 2021-01-04 16:05:40 | no | Genome Biol. 2005 Jun 29; 6(7):R56 | latin-1 | Genome Biol | 2,005 | 10.1186/gb-2005-6-7-r56 | oa_comm |
==== Front
Genome BiolGenome Biology1465-69061465-6914BioMed Central London gb-2005-6-7-r571599844610.1186/gb-2005-6-7-r57ResearchFunctional analysis of human and chimpanzee promoters Heissig Florian [email protected] Johannes [email protected] Jaroslaw [email protected] Philipp [email protected] Wolfgang [email protected]ääbo Svante [email protected] Max Planck Institute for Evolutionary Anthropology, Deutscher Platz 6, D-04103 Leipzig, Germany2005 1 7 2005 6 7 R57 R57 10 3 2005 13 6 2005 8 6 2005 Copyright © 2005 Heissig 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.
Twelve promoters of genes differentially expressed between humans and chimpanzees were tested for expression activity in culture cells. Seven promoters showed a significant difference in expression level between the human and chimpanzee promoter, but only three were in the same direction as the tissues, indicating that relevant expression differences between humans and chimpanzees will be difficult to predict from cell culture experiments or DNA sequences
Background
It has long been argued that changes in gene expression may provide an additional and crucial perspective on the evolutionary differences between humans and chimpanzees. To investigate how often expression differences seen in tissues are caused by sequence differences in the proximal promoters, we tested the expression activity in cultured cells of human and chimpanzee promoters from genes that differ in mRNA expression between human and chimpanzee tissues.
Results
Twelve promoters for which the corresponding gene had been shown to be differentially expressed between humans and chimpanzees in liver or brain were tested. Seven showed a significant difference in activity between the human promoter and the orthologous chimpanzee promoter in at least one of the two cell lines used. However, only three of them showed a difference in the same direction as in the tissues.
Conclusion
Differences in proximal promoter activity are likely to be common between humans and chimpanzees, but are not linked in a simple fashion to gene-expression levels in tissues. This suggests that several genetic differences between humans and chimpanzees might be responsible for a single expression difference and thus that relevant expression differences between humans and chimpanzees will be difficult to predict from cell culture experiments or DNA sequences.
==== Body
Background
Thirty years ago, King and Wilson [1] proposed that phenotypic differences between humans and chimpanzees are mainly caused by quantitative changes in gene expression rather than by structural changes in gene products. This idea is promoted also in some reviews [2] and seems to be supported by recent studies [3-8], which show that as many as 10% of all genes expressed in the brain differ in their expression levels between humans and chimpanzees. However, a causative connection between phenotypic differences and gene-expression differences in the two species remains to be established [9]. Similarly, the molecular basis of gene-expression differences between the two species is largely unknown.
The regulation of gene expression is a complex process involving chromatin structure, DNA methylation, transcription initiation, alternative splicing, RNA degradation, translational control, and posttranslational modifications [10,11]. However, initiation of transcription is thought to be a major factor determining the level of gene expression in most systems [12,13]. Studies in maize, yeast, mice, rats and humans indicate that both cis- and trans-acting factors are involved in transcriptional regulation [14-21]. Although trans-acting factors are clearly important, allelic DNA sequence variation in several human promoters has been shown to profoundly influence transcriptional activity [22-26]. Furthermore, the only functional comparison of a human and a chimpanzee promoter published to date shows that three nucleotide differences can lead to large differences in promoter activity [27].
To estimate what fraction of mRNAs differently expressed between human and chimpanzee tissues may be caused by DNA sequence differences in core promoters, we analyzed the activity of human and chimpanzee promoters from 12 genes that differ in their mRNA expression between the two species in brain and liver as measured by microarrays [6]. In each case, 2 kilobases (kb) of the putative human and chimpanzee promoter regions were cloned and tested for their ability to drive the transcription of a reporter gene during transient expression in human cervical carcinoma and neuroblastoma cell lines. The results show that no simple relationship exists between in vitro promoter activity and mRNA levels in tissues of the organisms.
Results
Gene-expression data measured with Affymetrix U95A arrays from livers and the prefrontal cortex of the brains of three humans and three chimpanzees [6] were used to identify genes that differ significantly in expression between the species. To avoid the influence of sequence differences on the hybridization of chimpanzee transcripts to microarray probes designed for human transcripts, we excluded all probes showing inconsistent hybridization patterns in the two species as described elsewhere [8]. The genes were required to be differentially expressed in at least one of the two tissues with a magnitude of at least 1.4-fold (false-discovery rate < 1%). The 71 genes that satisfied these criteria were further selected on the basis of the availability of an annotated transcription start site [28], the availability of human and chimpanzee DNA sequence of high quality, the possibility to place primers for amplification of the promoters as well as the successful amplification and cloning of the promoter fragments (see Materials and methods).
This left us with human and chimpanzee promoters from 12 genes. From each promoter, a fragment covering approximately 1,500 base pairs (bp) upstream and ≤ 500 bp downstream of the transcription start site without including the start codon was cloned in a plasmid in front of a firefly luciferase reporter gene. For each species, three independent clones were isolated and the insert of each clone was sequenced in its entirety. Each plasmid was mixed with a plasmid containing a sea-pansy luciferase reporter gene under the control of a constitutive promoter and transfected into a human neuroblastoma cell line and a human cervical carcinoma cell line, respectively. Experiments were performed in triplicate and the activities of the two luciferases were measured. All constructs showed an activity at least ninefold higher than the promoter-less control vector in at least one of the two cell lines. To control for transfection efficiency, the measurement of the firefly luciferase was normalized to the measurement of the sea-pansy luciferase and the difference in activity between the three human clones and the three chimpanzee clones analyzed.
The results are summarized in Figure 1 and Table 1. Out of the 12 promoters tested, two (ACADSB, C10orf10) show a significant difference (ANOVA p-value < 0.05) in both cell lines whereas five (IMPA1, CGI-51, SH3BGR, UNG, TERF) show a significant difference in only one of the two cell lines. Five promoters show no significant difference in either cell line. The average sequence divergence (Table 1) for these five promoters (1.2%) and the remaining seven (1.3%) is not significantly different from each other, neither for the complete promoter fragment (two-tailed t-test, p = 0.65) nor for 220 bp around the transcription start site (p = 0.43) in which most of the conserved regulatory motifs are found [29]. One promoter (THEM2) contains a chimpanzee-specific Alu insertion, but does not show a significant difference in its activity in either of the two cell lines.
Three promoters (ACADSB, C10orf10, IMPA1) show activity differences in the promoter assays that go in the same direction as the expression differences of the corresponding genes in the tissues. Interestingly, the two promoters (ACADSB, C10orf10) that show qualitatively similar differences in the two cell lines are both in concordance with the tissue expression differences. For four promoters (CGI-51, SH3BGR, UNG, TERF) that show differences in only one of the cell lines, the difference goes in the opposite direction to the expression differences in the tissues.
Discussion
We have compared the transcriptional activity of human and chimpanzee promoters from 12 genes that differ in their expression levels between humans and chimpanzees in tissues. We find that seven of the 12 promoter pairs differ significantly in their transcriptional activity in at least one of the two cell lines used (Figure 1, Table 1). This is in agreement with the finding that many proximal promoters that show sequence differences among humans differ in their activity in promoter assays [22,25,26,30]. Furthermore, we find that in five cases promoter activity differences are restricted to one of the two cell lines, showing that interspecies differences in promoter activity are often specific to cell line or tissue, an observation that is compatible with previous work on allelic promoter differences among different human cell lines [22,26] and tissue-specific cis-acting variation in mice [31] and humans [32]. Clearly, in order to predict such differences in promoter activity from their DNA sequences, much more knowledge on the occurrence of transcription factors in cells and their binding sites in vivo is needed.
What might seem more unexpected is that the differences in promoter activity observed in the cell lines seem to be independent of the differences in expression seen in the tissues. The transcript levels of only three genes (ACADSB, C10orf10, IMPA1) go in the same direction in at least one of the cell lines as they do in the tissues, whereas they go in the opposite direction for four genes (CGI-51, SH3BGR, UNG, TERF) in one of the two cell lines. One possible explanation is that the expression differences observed in tissues are due to differences in environmental factors between the species. This, however, seems unlikely, given the high reproducibility of expression differences between studies [5,33] that use different individuals. Assuming that the observed expression differences in tissues are indeed genetic in nature, one possibility to account for the discrepancy with the promoter activities observed in vitro is that the same sequence differences in the proximal promoters have opposite effects in different tissues or cell lines. However, none of the 12 promoters tested showed a significant opposite effect in the two cell lines (Figure 1), nor did any of 43 allelic variants of human promoters tested in a previous study show an opposite effect in different cell lines [22,26]. Furthermore, over 99% of the genes that show a significant expression difference between humans and chimpanzees in two tissues show the same direction of change in the two tissues [6,34]. Thus, it seems unlikely that a tissue- or cell-type effect is responsible for the opposite expression patterns seen here between the cell lines and the tissues.
A remaining possibility is that additional genetic differences between humans and chimpanzees outside the proximal promoters are numerous and/or strong enough to lead to gene-expression levels in tissues that are often qualitatively opposite to what would be inferred from the activity of the proximal promoter in vitro. This agrees with the observation that many gene-expression differences are inherited as quantitative traits, that is, several genetic loci are responsible for allelic gene expression differences observed among humans and among mice [14,21,35]. In fact, when gene-expression differences between humans and chimpanzees are compared, the number of loci affecting the expression of single genes is likely to be even higher than for allelic differences, as these two species are more diverged than individual mice or humans. Of possible relevance in this respect is the recent finding that promoters are much less conserved, relative to intronic regions, between human and chimpanzee than between mouse and rat [36]. A probable cause of this is the smaller effective population size of primates than of rodents, which would have allowed slightly deleterious regulatory variants to become fixed in periods when the population size was small [36]. When the population size, and hence the effectiveness of selection, subsequently increased, compensatory mutations outside the proximal promoters might have become fixed. An example of such compensatory mutations has been described for the even-skipped stripe 2 enhancer in two Drosophila species [37].
It is worth noting that the genes studied here are all differentially expressed in human and chimpanzee tissues. It is still unclear to what extent the discordance between behavior of promoters in vitro and the corresponding tissue mRNA levels also holds true for genes that do not show an expression difference between human and chimpanzee tissues. If many genetic differences do indeed influence the expression of a single gene, the proximal promoters of these non-differentially expressed genes would be expected to differ in their activity almost as frequently as the promoters of differentially expressed genes.
Our results imply that although many promoters may differ in activity between humans and chimpanzees, it will be difficult to predict physiologically relevant gene-expression differences from promoter activities observed in cell lines, even between two closely related species such as humans and chimpanzees. Further work is necessary to elucidate to what extent this applies also to allelic DNA sequence differences in promoters observed within a species. Further work is also needed to elucidate whether a general paradigm for how genome structure translates to gene expression activity can be derived.
Materials and methods
Selection of promoters for study
Genes for promoter analysis were selected on the basis of a large-scale transcriptome comparison between three humans and three chimpanzees in brain and liver using Affymetrix HG U95A and HG U95Av2 arrays [6]. All microarray analysis was performed using the MAS 5.0 software package from Affymetrix. Selected genes were required to be differentially expressed in at least one of the two tissues with average change p-value < 0.05 or > 0.95 (two-sided test) and with a fold-change magnitude of at least 1.4-fold. The false-discovery rate was determined by applying these selection criteria to 10,000 permutations of the original dataset with randomly assigned sample labels. Expression differences were confirmed by masking all probes showing inconsistent hybridization patterns in the two species using a custom mask file as described elsewhere [8].
We used the DBTSS databases of transcriptional start sites [38] (August 2003) to identify the transcription start sites and to collect promoter sequences 2,000 bp upstream and 1,000 bp downstream of the start sites. Out of a total of 71 genes satisfying the gene-expression-based criteria (see Additional data file 1), 35 had annotated transcription start sites and chimpanzee sequence was available (July 2003). For 24 of them, primers were designed using Primer 3 software [39] to amplify approximately 1,500 bp upstream and 500 bp downstream of the transcription start site, if possible without including any coding sequence or the start codon. Restriction sites for BglII or XhoI (depending on the presence of restriction sites in the promoter sequence) were added to the 5'-end of the primer for cloning. Using these primers (see Additional data file 2), we were able to amplify and isolate three independent clones from both species for 12 genes.
Cloning and reporter gene assay
DNA from one human and one chimpanzee was amplified using Expand 20 kb Plus PCR system (Roche) according to the manufacturer's protocols with an extension time of 3 min, or Pfu DNA polymerase (Stratagene), using the following conditions: 1 min at 96°C, 45 sec at 96°C, 45 sec at 61°C, 5 min at 72°C for 38 cycles, followed by a final extension at 72°C for 10 min, on Tetracyclers (MJ Research).
PCR products were purified up by QIAquick eight-well cleanup kit (Qiagen), digested with either BglII (NEB) or XhoI (NEB), purified on 1% low-melting agarose gels (Promega) and isolated using QIAquick gel purification kit (Qiagen). These fragments were cloned upstream of the firefly luciferase gene into the BglII or XhoI site of the pGl3 vector (Promega), using T4 Quick Ligase (NEB) and One Shot Top 10 F cells (Invitrogen). Colonies were picked and heated in 10 μl water for 5 min at 96°C, and 2 μl was used as template in a 25 μl PCR reaction, using one primer in the vector (GL2) and one primer in the promoter.
Positive clones were grown in 7 ml LB medium (Invitrogen) containing 100 μg/ml ampicillin (Sigma) at 37°C overnight. Vector DNA was isolated using a Miniprep kit (Qiagen), and DNA concentration was measured on a Nanodrop UV spectrophotometer (NanoDrop Technologies). All inserts were sequenced (Additional data file 4) using Big Dye Terminator chemistry (Applied Biosystems).
The human neuroblastoma cell line (SHEP [27]) was obtained from Martin Reick, University of Texas Southwestern Medical Center, and the human cervical carcinoma cell line (c33a) (ATCC Number HTB-31) was obtained from Kurt Engeland, University of Leipzig. SHEP cells were grown in DMEM (Gibco) medium supplemented with 15% fetal bovine serum (Sigma) and c33a cells in DMEM/MIX F12 (Gibco) supplemented with 10% fetal bovine serum (Sigma), and plated at ~85% confluence a day before transfection. One microgram of the promoter constructs was mixed with 67.4 ng of the pRL-SV40 vector (Promega) containing the sea-pansy luciferase gene in 96.8 μl serum-free medium (Optimem1, Gibco) and 2.5 μl Lipofectamine 2000 (Gibco). Cells were transfected in triplicate for 4 h at 37°C, 5% CO2 and 100% humidity, grown for 20 h and then lysed in 100 μl lysis buffer (Promega).
A 5 μl sample of lysate was used in a Dual-Luciferaser Reporter Assay System (Promega) in a Wallac Victor 2 Luminometer (PerkinElmer). Promoter activity was measured by normalizing the luciferase activity of the promoter constructs to the sea-pansy luciferase activity of the control plasmid from the same well (see Additional data file 3). We assessed the significance (p < 0.05) of different activity in human and chimpanzee promoters using a multi-way ANOVA including species, clones, and replicates.
Additional data files
Additional data is available with the online version of this paper. Additional data file 1 is a table listing the 71 genes differentially expressed between humans and chimpanzees in liver and/or brain. For the probe sets the change p-values (MAS 5.0) averaged over the 36 pairwise comparisons between human and chimpanzee samples [6] and the signal log2 ratio (slr; MAS 5.0) for the probe sets are given. p-values close to 1 suggest a higher expression in chimpanzees, as does a negative slr. Additional data file 2 is a table listing the primers used for promoter amplification. Primer sequences are written 5' to 3' and lower-case letters indicate added restriction sites. Additional data file 3 is a table listing measured promoter activities (the measured luciferase activity of the promoter constructs divided by the sea-pansy luciferase activity, for each of the three replicates that were done for each of the three clones for humans and chimpanzees in the neuroblastoma cell line and the cervix carcinoma cell line, respectively, for the 12 genes analyzed). Additional data file 4 contains the nucleotide sequences of the insert of the used vectors.
Supplementary Material
Additional File 1
Supplementary Table 1: Listed are all 71 genes that were identified as being at least 1.4-fold differently expressed in liver and/or brain. For the probe sets the change p-values (MAS 5.0) averaged over the 36 pairwise comparisons between human and chimpanzee samples [6] and the signal log2 ratio (slr; MAS 5.0) for the probe sets are given. p-values close to 1 suggest a higher expression in chimpanzees, as does a negative slr.
Click here for file
Additional File 2
Supplementary Table 2: The primer sequences used to amplify the promoter fragments are given here. Primer sequences are written 5' to 3' and lower-case letters indicate added restriction sites.
Click here for file
Additional File 3
Supplementary Table 3: The normalized promoter activities are listed; that is, the measured luciferase activity of the promoter constructs divided by the sea-pansy luciferase activity, for each of the three replicates that were done for each of the three clones for humans and chimpanzees in the neuroblastoma cell line and the cervix carcinoma cell line, respectively, for the 12 genes analyzed.
Click here for file
Additional File 4
The nucleotide sequences of the cloned promoters.
Click here for file
Acknowledgements
We are grateful to Ines Hellmann, Naim Matasci and Katja Nowick for comments on the manuscript and to the Max Planck Society for financial support.
Figures and Tables
Figure 1 Human and chimpanzee promoter activity and mRNA expression in tissues. (a) Promoter activity in cell cultures. Normalized activities of promoters of the indicated genes are compared to the average of the three human and the three chimpanzee clones for each cell line and promoter indicated. Blue indicates lower activity than average, whereas yellow indicates higher activity. The color scale for the fold-change is below. Significant differences in activity between the two species are indicated by red frames. (b) Expression of the genes in tissues as assayed by mRNA levels measured by oligonucleotide arrays for brain and liver [6]. The values are averages of the three human and three chimpanzee individuals for which expression levels were determined [6] and are compared to the average level of the corresponding gene.
Table 1 DNA sequence and expression divergence of human and chimpanzee promoters
Promoter Divergence† Divergence TSS‡ Brain§ Liver§ Neuroblastoma¶ Cervical carcinoma¶
ACADSB 1.4 3.3 -1.6 -1.3 -0.5* -0.3*
C10orf10 1.3 1.8 ND 4.0 1.4** 0.7***
CGI-51 1.6 1.9 -0.4 -1.3 0.3* 0.1
DDX17 1.8 2.0 -4.2 -4.4 0 -0.1
DHX16 1.5 0.9 -0.8 -1.7 0 0.1
HLA-DPB1 0.8 0.9 2.5 4.1 0.2 0.2
IMPA1 1.5 2.6 2.3 2.6 0.6* 0.2
PCOLCE 0.9 0.6 ND 2.1 0 0.2
SH3BGR 1.6 1.3 1.9 2.4 -0.8 -0.7***
TERF1 1.5 1.4 2.3 1.6 -0.7* 0
THEM2 1.2 2.6 -0.5 -0.8 0.1 0.1
UNG 0.6 0.5 2.5 1.8 -0.3* 0
†Point substitutions per compared site in the whole promoter (~ 2 kbp); ‡point substitutions per compared site around the transcription start site (-200 to +20); §signal log2 ratio as determined from arrays, where positive values indicate higher expression in humans (that is, a value of 1 means twofold higher expression in humans); ¶average signal log2 ratio for the promoter activity in cell lines; *, ** and *** indicate p-values (ANOVA) < 0.05, 0.01, and 0.001, respectively.
==== Refs
King MC Wilson AC Evolution at two levels in humans and chimpanzees Science 1975 188 107 116 1090005
Wray GA Hahn MW Abouheif E Balhoff JP Pizer M Rockman MV Romano LA The evolution of transcriptional regulation in eukaryotes Mol Biol Evol 2003 20 1377 1419 12777501 10.1093/molbev/msg140
Uddin M Wildman DE Liu G Xu W Johnson RM Hof PR Kapatos G Grossman LI Goodman M Sister grouping of chimpanzees and humans as revealed by genome-wide phylogenetic analysis of brain gene expression profiles Proc Natl Acad Sci USA 2004 101 2957 2962 14976249 10.1073/pnas.0308725100
Karaman MW Houck ML Chemnick LG Nagpal S Chawannakul D Sudano D Pike BL Ho VV Ryder OA Hacia JG Comparative analysis of gene-expression patterns in human and African great ape cultured fibroblasts Genome Res 2003 13 1619 1630 12840040 10.1101/gr.1289803
Caceres M Lachuer J Zapala MA Redmond JC Kudo L Geschwind DH Lockhart DJ Preuss TM Barlow C Elevated gene expression levels distinguish human from non-human primate brains Proc Natl Acad Sci USA 2003 100 13030 13035 14557539 10.1073/pnas.2135499100
Enard W Khaitovich P Klose J Zollner S Heissig F Giavalisco P Nieselt-Struwe K Muchmore E Varki A Ravid R Intra- and interspecific variation in primate gene expression patterns Science 2002 296 340 343 11951044 10.1126/science.1068996
Gu J Gu X Induced gene expression in human brain after the split from chimpanzee Trends Genet 2003 19 63 65 12547510 10.1016/S0168-9525(02)00040-9
Khaitovich P Muetzel B She X Lachmann M Hellmann I Dietzsch J Steigele S Do HH Weiss G Enard W Regional patterns of gene expression in human and chimpanzee brains Genome Res 2004 14 1462 1473 15289471 10.1101/gr.2538704
Khaitovich P Weiss G Lachmann M Hellmann I Enard W Muetzel B Wirkner U Ansorge W Paabo S A neutral model of transcriptome evolution PLoS Biol 2004 2 E132 15138501 10.1371/journal.pbio.0020132
Strachan T Read AP Human molecular genetics 1999 2nd New York: Wiley
Carey M Smale ST Transcriptional Regulation in Eukaryotes: Concepts, Strategies, and Techniques 2000 Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press
Roeder RG The role of general initiation factors in transcription by RNA polymerase II Trends Biochem Sci 1996 21 327 335 8870495 10.1016/0968-0004(96)10050-5
Mitchell PJ Tjian R Transcriptional regulation in mammalian cells by sequence-specific DNA binding proteins Science 1989 245 371 378 2667136
Schadt EE Monks SA Drake TA Lusis AJ Che N Colinayo V Ruff TG Milligan SB Lamb JR Cavet G Genetics of gene expression surveyed in maize, mouse and man Nature 2003 422 297 302 12646919 10.1038/nature01434
Raser JM O'Shea EK Control of stochasticity in eukaryotic gene expression Science 2004 304 1811 1814 15166317 10.1126/science.1098641
Morley M Molony CM Weber TM Devlin JL Ewens KG Spielman RS Cheung VG Genetic analysis of genome-wide variation in human gene expression Nature 2004 430 743 747 15269782 10.1038/nature02797
Brem RB Yvert G Clinton R Kruglyak L Genetic dissection of transcriptional regulation in budding yeast Science 2002 296 752 755 11923494 10.1126/science.1069516
Doss S Schadt EE Drake TA Lusis AJ Cis-acting expression quantitative trait loci in mice Genome Res 2005 15 681 691 15837804 10.1101/gr.3216905
Monks SA Leonardson A Zhu H Cundiff P Pietrusiak P Edwards S Phillips JW Sachs A Schadt EE Genetic inheritance of gene expression in human cell lines Am J Hum Genet 2004 75 1094 1105 15514893 10.1086/426461
Hubner N Wallace CA Zimdahl H Petretto E Schulz H Maciver F Mueller M Hummel O Monti J Zidek V Integrated transcriptional profiling and linkage analysis for identification of genes underlying disease Nat Genet 2005 37 243 253 15711544 10.1038/ng1522
Bystrykh L Weersing E Dontje B Sutton S Pletcher MT Wiltshire T Su AI Vellenga E Wang J Manly KF Uncovering regulatory pathways that affect hematopoietic stem cell function using 'genetical genomics' Nat Genet 2005 37 225 232 15711547 10.1038/ng1497
Hoogendoorn B Coleman SL Guy CA Smith K Bowen T Buckland PR O'Donovan MC Functional analysis of human promoter polymorphisms Hum Mol Genet 2003 12 2249 2254 12915441 10.1093/hmg/ddg246
Bray NJ Buckland PR Owen MJ O'Donovan MC Cis-acting variation in the expression of a high proportion of genes in human brain Hum Genet 2003 113 149 153 12728311
Trinklein ND Aldred SJ Saldanha AJ Myers RM Identification and functional analysis of human transcriptional promoters Genome Res 2003 13 308 312 12566409 10.1101/gr.794803
Buckland PR Allele-specific gene expression differences in humans Hum Mol Genet 2004 13 Spec No 2 R255 R260 15358732 10.1093/hmg/ddh227
Hoogendoorn B Coleman SL Guy CA Smith SK O'Donovan MC Buckland PR Functional analysis of polymorphisms in the promoter regions of genes on 22q11 Hum Mutat 2004 24 35 42 15221787 10.1002/humu.20061
Huby T Dachet C Lawn RM Wickings J Chapman MJ Thillet J Functional analysis of the chimpanzee and human apo(a) promoter sequences: identification of sequence variations responsible for elevated transcriptional activity in chimpanzee J Biol Chem 2001 276 22209 22214 11301336 10.1074/jbc.M102204200
Suzuki Y Yamashita R Sugano S Nakai K DBTSS, DataBase of Transcriptional Start Sites: progress report 2004 Nucleic Acids Res 2004 32 Database D78 D81 14681363 10.1093/nar/gkh076
Xie X Lu J Kulbokas EJ Golub TR Mootha V Lindblad-Toh K Lander ES Kellis M Systematic discovery of regulatory motifs in human promoters and 3' UTRs by comparison of several mammals Nature 2005 434 338 345 15735639 10.1038/nature03441
Rockman MV Wray GA Abundant raw material for cis-regulatory evolution in humans Mol Biol Evol 2002 19 1991 2004 12411608
Cowles CR Hirschhorn JN Altshuler D Lander ES Detection of regulatory variation in mouse genes Nat Genet 2002 32 432 437 12410233 10.1038/ng992
Lo HS Wang Z Hu Y Yang HH Gere S Buetow KH Lee MP Allelic variation in gene expression is common in the human genome Genome Res 2003 13 1855 1862 12902379 10.1101/gr.885403
Preuss TM Caceres M Oldham MC Geschwind DH Human brain evolution: insights from microarrays Nat Rev Genet 2004 5 850 860 15520794 10.1038/nrg1469
Khaitovich P Hellmann I Enard W Nowick K Leinweber M Franz H Weiss G Lachmann M Pääbo S Parallel patterns of evolution in the genomes and transcriptomes of humans and chimpanzees Science 2005
Chesler EJ Lu L Shou S Qu Y Gu J Wang J Hsu HC Mountz JD Baldwin NE Langston MA Complex trait analysis of gene expression uncovers polygenic and pleiotropic networks that modulate nervous system function Nat Genet 2005 37 233 242 15711545 10.1038/ng1518
Keightley PD Lercher MJ Eyre-Walker A Evidence for widespread degradation of gene control regions in hominid genomes PLoS Biol 2005 3 E42 15678168 10.1371/journal.pbio.0030042
Ludwig MZ Bergman C Patel NH Kreitman M Evidence for stabilizing selection in a eukaryotic enhancer element Nature 2000 403 564 567 10676967 10.1038/35000615
Suzuki Y Yamashita R Nakai K Sugano S DBTSS: DataBase of human Transcriptional Start Sites and full-length cDNAs Nucleic Acids Res 2002 30 328 331 11752328 10.1093/nar/30.1.328
Rozen S Skaletsky H Primer3 on the WWW for general users and for biologist programmers Methods Mol Biol 2000 132 365 386 10547847
| 15998446 | PMC1175988 | CC BY | 2021-01-04 16:35:47 | no | Genome Biol. 2005 Jul 1; 6(7):R57 | utf-8 | Genome Biol | 2,005 | 10.1186/gb-2005-6-7-r57 | oa_comm |
==== Front
Genome BiolGenome Biology1465-69061465-6914BioMed Central London gb-2005-6-7-r581599844710.1186/gb-2005-6-7-r58ResearchCreation and disruption of protein features by alternative splicing - a novel mechanism to modulate function Hiller Michael [email protected] Klaus 2Platzer Matthias 2Backofen Rolf [email protected] Institute of Computer Science, Friedrich-Schiller-University Jena, Chair for Bioinformatics, Ernst-Abbe-Platz 2, 07743 Jena, Germany2 Genome Analysis, Institute of Molecular Biotechnology, Beutenbergstrasse 11, 07745 Jena, Germany2005 22 6 2005 6 7 R58 R58 25 2 2005 19 4 2005 9 5 2005 Copyright © 2005 Hiller et al.; licensee BioMed Central LtdThis 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.
A new mechanism of alternative splicing is proposed which creates a protein feature by putting together two non-consecutive exons and destroys a feature by inserting an exon in its body. Evidence for this rare mechanism is provided by a genome-wide search with four specific protein features.
Background
Alternative splicing often occurs in the coding sequence and alters protein structure and function. It is mainly carried out in two ways: by skipping exons that encode a certain protein feature and by introducing a frameshift that changes the downstream protein sequence. These mechanisms are widespread and well investigated.
Results
Here, we propose an additional mechanism of alternative splicing to modulate protein function. This mechanism creates a protein feature by putting together two non-consecutive exons or destroys a feature by inserting an exon in its body. In contrast to other mechanisms, the individual parts of the feature are present in both splice variants but the feature is only functional in the splice form where both parts are merged. We provide evidence for this mechanism by performing a genome-wide search with four protein features: transmembrane helices, phosphorylation and glycosylation sites, and Pfam domains.
Conclusion
We describe a novel type of event that creates or removes a protein feature by alternative splicing. Current data suggest that these events are rare. Besides the four features investigated here, this mechanism is conceivable for many other protein features, especially for small linear protein motifs. It is important for the characterization of functional differences of two splice forms and should be considered in genome-wide annotation efforts. Furthermore, it offers a novel strategy for ab initio prediction of alternative splice events.
==== Body
Background
Alternative splicing is an important post-transcriptional process and mainly contributes to the complexity of a transcriptome and proteome [1-3]. Alternative splicing often produces two or more proteins with functional differences from one gene [4] but can also downregulate the overall protein level by producing targets for nonsense-mediated mRNA decay [5], which is used, for example, in the autoregulation of splicing factors [6]. Furthermore, defects in splicing are the basis for a number of diseases [7].
One major mechanism of alternative splicing to alter protein function is the insertion/deletion of functional units such as protein domains, transmembrane (TM) helices, signal peptides, or coiled-coil regions. Alternative splicing tends to insert/delete complete functional units instead of affecting parts of a unit [8]. Moreover, several protein domains have a tendency to be spliced out in some transcripts [9,10]. Many proteins occur in a soluble as well as a membrane-bound form. When encoded by a single gene, the soluble form can be produced by post-translational ectodomain shedding [11] or alternative splicing of exons that encode the TM helices. Indeed, 40-50% of the alternatively spliced, single-pass TM proteins have a splice form that specifically removes the TM domain [12,13]. Furthermore, protein forms can differ in their affinity to bind ligands [14,15] or in their subcellular location [16].
In this paper, we present a novel mechanism to modulate function and/or subcellular localization of a protein by alternative splicing. Assuming a protein feature is encoded in two parts by two non-consecutive exons, for example, exon 2 and 4, inclusion of exon 3 results in a protein lacking this feature since it is disconnected at the sequence level. In contrast, the skipping of exon 3 leads to a protein with this feature. We provide evidence for this mechanism by considering four protein features: TM helices, phosphorylation and glycosylation sites, and Pfam domains. In general, this mechanism is conceivable for many other protein features and provides a novel strategy for ab initio prediction of alternative splice events.
Results and discussion
In order to find genes that encode a protein feature by two non-consecutive exons, we searched all human RefSeq transcripts for annotated features that span an exon boundary. For these exon pairs, we searched dbEST to find alternative splice events that insert a sequence between them. Thus, we only selected pairs of exons if they had expressed sequence tag (EST)-confirmed, alternative exons between them that are skipped in the given RefSeq. Apart from alternative exons, intron retention or an alternative donor/acceptor site located in the intron can lead to such an insert. We only selected inserts that preserved the open reading frame. Then we evaluated whether the longer transcript (with the insert) still encodes the feature or not. We only considered two exons for small features like TM helices and post-translational modification contexts since it is unlikely that more than two exons encode the feature. For more complex features like Pfam domains, we allowed for the domain to be encoded by more than two exons.
The first protein feature we considered was TM domains. We annotated TM helices in all RefSeq transcripts with the TMHMM program [17]. We found 1,807 TM domains (14% of all TM domains) that are encoded by two exons (Additional data file 1). For ten cases, we found EST evidence for an insert due to alternative splicing. As TM domains are short stretches of hydrophobic amino acids, an insert with polar residues will result in the destruction of the TM helix. Indeed, the evaluation of these ten longer transcripts with TMHMM showed that six clearly lacked the TM domain which, in three cases, leads to a soluble protein (Table 1). An example of the disruption of the single TM domain is depicted for DIABLO in Figure 1a. A more complex example is at the Rhesus blood group antigen gene (RHCE) where the inclusion of two exons resulted in a loss of one TM domain as well as the gain of three others (Figure 1b). The massive reconstruction of TM domains in the respective protein isoforms can have considerable consequences for the orientation of the proteins within the cellular membrane and for their interaction with other membrane components.
To find further cases of feature disruption by sequence insertion, we applied the procedure to experimentally verified post-translational modification sites. Post-translational modification of proteins plays a role in various important processes. For example, phosphorylation of splicing factors can influence splicing decisions [18] and glycosylation is associated with a modulation of proteolytic resistance and ligand binding [19]. The residue to be modified must be located in a favorable sequence context to be recognized by the enzyme. If this residue is close to an exon boundary, an alternative splice event can change the context to an unfavorable one with the consequence that the modification cannot take place anymore. We inspected the O-GlycBase [19] and Phospho.ELM [20], and found 435 modified residues that are close to 213 different exon-exon junctions. Among them, four exon junctions showed an insert due to alternative splicing. CCL14 has a glycosylated serine at position 26, which is the last residue encoded by exon 1. We found two ESTs (AA612866, Z70293) with an included 48-nucleotide exon between exon 1 and 2. The NetOGlyc [21] score for the serine in the new sequence context dropped from 0.97 to 0.35 (threshold 0.5). Thus, the new context might prevent glycosylation of this residue. For CDK5, an alternative acceptor (BU529114) that inserts nine amino acids upstream of exon 8 alters the context of the phosphorylated serine at position 159 of the protein. The NetPhos [22] scores of both contexts differ (0.93 vs 0.43, threshold 0.5), which indicates that only one context allows recognition by the kinase and, thus, the phosphorylation of the serine. Additionally, we found two examples (MGP and CDK2) where an included exon alters the context of a phosphorylated residue, however, the scores for the new contexts dropped only marginally.
For the fourth feature, we considered functional protein domains using the Pfam database [23]. We found 473 inserts into a Pfam domain and nine of those resulted in a disruption of the Pfam (Table 2). Additionally, using the algorithm described in [24], we found three cases where the skipping of a RefSeq exon creates a new Pfam (Table 2). For example, skipping exon 4 of NM_024565 created the cyclin N-terminal domain (Figure 2a). Since exons 5 to 7 of this transcript encode the cyclin C-terminal domain (PF02984), only the exon skipping variant might perform the function of a cyclin. Moreover, skipping exon 2 of NM_139174 resulted in a new double-stranded RNA binding domain (Figure 2b). Downstream of this domain, the transcript encodes an adenosine-deaminase (editase) domain. Thus, the loss of the RNA binding property might act as a negative regulation of the editase activity. Most Pfam domains fold into three-dimensional structures and we cannot rule out that these 12 domains also adopt the correct folding with the insert. However, using standard cut-off scores, these Pfam domains cannot be found in the longer transcripts since the scores for both individual parts are always below the threshold.
In general, any EST-based approach is hampered by the bias of publicly available EST databases towards cancer-related tissues or cell lines that may exhibit aberrant splicing [25,26]. Furthermore, a splice form that is only represented by a single EST may be a rare error by the spliceosome. Therefore, we determined the number and tissue source of the ESTs that match both splice variants for the described examples (Additional data file 2). For seven of the 20 examples, only one splice form is represented by a single EST or by cancer-related ESTs. However, the remaining examples are supported by several ESTs as well as ESTs from normal tissue, and in four cases both splice variants are contained in the RefSeq database. Thus, we conclude that the majority of the described examples are real splice variants and not artifacts or aberrant splice events.
Besides the four features investigated here, there are many others that can only function if they are connected on the sequence level. Such functional sites or motifs often have a linear structure and comprise, for example, signal peptides, post-translational cleavage sites and subcellular localization signals as well as sites for protein-protein interaction. Many of these motifs are collected in the Eukaryotic Linear Motif (ELM) database [27]. Such features can lose their function if an insert separates them on the sequence level. For example, splicing at an alternative donor site of the protein kinase C delta leads to an insert of 26 amino acids into a caspase-3 cleavage site and to an isoform that is caspase-insensitive [28]. We have not investigated such features here since only a fraction of them have been experimentally verified and a prediction results in a high number of false positives. With further efforts in verifying and characterizing these features, we expect an increasing number of examples for the proposed mechanism of modulating protein function by alternative splicing. Interestingly, the same principle was recently used to experimentally characterize exon splicing silencers (ESS) [29]. In this study, ESS candidates were inserted in the middle exon of a three-exon minigene. If a candidate ESS acts as a silencer, the middle exon is skipped and only in this case a functional green fluorescent protein is encoded. Furthermore, this mechanism is not restricted to protein features but it is also conceivable for sequence and structural features at the mRNA level. For example, some of the variable first exons of NOS1 together with exon 2, form a hairpin structure that is involved in translational regulation, whereas other alternative first exons do not allow hairpin formation [30].
From an evolutionary viewpoint, this mechanism can be explained in two ways depending on whether the protein feature is ancestral or not. If the feature is ancestral, it means it is initially encoded by two neighboring exons and the inserted part must have appeared in the intronic sequences [31-33]. In this case, the insert simply has the function of a spacer. If the feature is not ancestral, it means the longer splice form is evolutionarily older and, therefore, the alternative exon or splice site must have been converted from a constitutive to an alternative one. This can happen, for example, by the weakening of splice sites or the creation of ESS [34]. Complex features with a high sequence specificity such as Pfam domains are likely to be ancestral. In contrast, small features with a loose sequence motif such as the context of a post-translational modification site can arise just by chance and can therefore be evolutionarily younger.
Not all alternative splice events are represented in EST databases and, thus, the development of non-EST-based methods for ab initio prediction of splice events is a necessary but challenging task. Currently, there is only one method that mainly uses genomic conservation of exons and flanking introns to discriminate between alternative and constitutive exons [35]. Although alternative splicing often deletes functional units, it is very hard to predict such events on the protein level without ESTs. However, a search for protein features that are put together by exon skipping would provide a new way to predict alternative splice events. For that purpose, it has to be assumed that the split feature is unlikely to be encoded by two non-consecutive exons just by chance. Since Pfam domains usually have a high sequence specificity, we tested this assumption for Pfams by skipping 10,962 constitutive exons. We found only four cases (0.036%) where skipping of a constitutive exon results in an additional Pfam domain (Additional data file 3). In contrast, nine of the 473 (1.9%) alternatively spliced inserts into Pfam domains resulted in a loss of the Pfam. The odds ratio of 53 indicates that Pfam domains are unlikely to be encoded by non-consecutive exons just by chance.
Conclusion
Alternative splicing frequently modulates protein function by insertion or deletion of functional units. In this case, the functional difference is directly associated with the sequence of the inserted or deleted part. Here, we provide evidence for an additional mechanism that acts by putting together a feature from two parts encoded by non-consecutive exons. Thus, the functional difference is not related to a specific insert and the two parts of the feature are present on both the long and the short splice form. The general idea is shown in Figure 3.
Recent alternative splicing databases include the annotation of the functional differences between two protein forms [36]. For this purpose, the novel mechanism described here has to be taken into account since it is obviously not sufficient to inspect the alternative exons in the context of the splice form that includes these exons. The functional difference of the examples shown here can only be found if the complete shorter splice form is investigated simultaneously.
Materials and methods
General procedure
All transcripts were taken from the RefSeq annotations in the UCSC Genome Browser (assembly hg16 with annotation March 2004) [37]. For exon pairs that together encode a protein feature, we extracted a 40-nucleotide context (20 nucleotides from the upstream and 20 nucleotides from the downstream exon) and searched, with BLAST, the human fraction of dbEST (August 2004) [38]. We only kept EST hits with two separate HSPs (high-scoring segment pairs).We discarded splice events that resulted in a frameshift and/or introduced a premature termination codon (PTC) since a frameshift leads to a new protein sequence downstream of the alternative splice site and transcripts with PTCs are frequently degraded by nonsense-mediated mRNA decay. Intron retention events were only included if the EST had a spliced intron up- or downstream. For the insertions, we checked presence of AG-GT splice sites. All splice forms were translated with the insertion and a check was made to see if the insert destroyed the feature.
TM domains
We predicted TM helices with TMHMM for all translated transcripts since, currently, TMHMM was found to be the best-performing TM prediction program [39]. The TM domain location was mapped to the exon structure and we considered a TM helix as encoded by two exons if each exon encoded at least 25% of the domain.
Glycosylation and phosphorylation contexts
We used Phospho.ELM version 2.0 and O-GlycBase v6.00. The SwissProt IDs were converted to RefSeq IDs with the table from the HUGO gene nomenclature committee website [40]. The location of the modified residues was mapped to the exon structure and we retained those close to an exon boundary (<10 amino acid distance for glycosylated and <5 amino acid distance for phosphorylated residues). To compute the scores for the glycosylated serine, we used NetOGlyc 2.0 because the latest version (3.1) is not able to recognize the serine in the annotated context.
Pfam domains
Pfam domains were found with hmmpfam using the 'gathering cutoff' scores as given in the Pfam database (version 14). We considered domains with less than 200 residues that are encoded by two or more exons (each exon encodes at least two residues of the Pfam). Additionally, we used the algorithm described in [24] to find cases where the RefSeq transcript is the longer splice form and a shorter exon skipping variant exists that encodes a new Pfam domain. To confirm such candidate splice forms, we searched dbEST with BLAST and the 40-nucleotide context from the up- and downstream exon.
Test of Pfam domain creation by chance
We compiled a set of 10,962 internal coding exons with a size divisible by three that had at least six ESTs showing their inclusion but no EST indicating their skipping. Those exons were considered to be constitutive. We produced the full-length protein and the shorter protein that corresponds to the hypothetical splice form without such an exon. Then, we used hmmpfam with the gathering cut-offs to search the Pfam database and compared the Pfam family hits for the full-length and the shorter protein.
Additional data files
The following additional data are available with the online version of this paper. Additional data file 1 is a table listing the TM domains that are encoded by two exons. Additional data file 2 contains the number of ESTs/RefSeqs and information about the tissues or libraries for both splice variants of the examples. Additional data file 3 contains the four cases where skipping of a constitutive exon results in a new Pfam domain.
Supplementary Material
Additional File 1
TM domains that are encoded by two exons
Click here for file
Additional File 2
ESTs/RefSeqs and if available their tissue/library source for the described examples
Click here for file
Additional File 3
Pfam creation events by skipping of a constitutive exon
Click here for file
Acknowledgements
We thank Anke Busch for helpful comments on the manuscript.
Figures and Tables
Figure 1 TM domain destruction by exon insertion. (a) Exons 2 and 3 of NM_138929 of DIABLO encode a TM domain (shown as blue boxes). This TM domain is destroyed in another transcript (NM_019887) that includes an additional exon. The inserted exon (shown in red) encodes many polar amino acids. (b) Exons 3 and 4 of NM_138617 of RHCE encode a TM domain that is destroyed in NM_138618 by the inclusion of two exons. Interestingly, the two included exons encode three new TM domains. Thus, the skipping of exon 4 and 5 of NM_138618 results in a protein that has only two instead of three TM domains fewer. Exon numbers refer to the respective transcript. TM, transmembrane.
Figure 2 Pfam creation by exon skipping. The alternative exon is shown in red. The two partial Pfam alignments for the RefSeq transcript and the complete alignment for the exon-skipping variant are shown above and below the partial gene structure, respectively. Dashed lines indicate parts of the exon for which a Pfam alignment has been found. (a) NM_024585 has a splice form that skips exon 4 (shown in red), which results in the creation of a new domain. The Pfam scores for the separated parts are far below the threshold score of 17 and, thus, the Pfam is not found for the longer transcript. (b) Skipping exon 2 of NM_139174 results in a new double-stranded RNA binding Pfam.
Figure 3 General mechanisms to alter linear protein features by alternative splicing. (a) A widespread mechanism is to skip or include an alternative exon (red box) that encodes a functional unit (indicated by the light bulb). The longer splice form with the alternative exon encodes a protein with this feature, the shorter splice form encodes a protein without this feature. (b) The novel mechanism involves a functional unit that is encoded by two non-consecutive exons (the two parts of the light bulb). In contrast to the mechanism mentioned above, the longer splice form encodes a protein without the functional unit although both parts are present on the protein sequence. The disruption of the unit results in a loss of function. The shorter splice form encodes a protein that puts together both parts of the unit which results in a gain of function (complete light bulb).
Table 1 RefSeq transcripts where, due to alternative splicing, sequence insertion destroys a TM helix
Gene symbol Gene name RefSeq with TM* RefSeq/EST without TM† Alternative splice event‡ Impact
DIABLO Diablo homolog (Drosophila) NM_138929 NM_019887 Exon between exon 2 and 3 Disruption of the single TM domain, soluble protein
DPP8 Dipeptidylpeptidase 8 NM_017743 NM_197961 Exon between exon 15 and 16 Disruption of the single TM domain, soluble protein
COX7A2 Cytochrome c oxidase subunit VIIa polypeptide 2 (liver) NM_001865 BU570379 Donor downstream of exon 3 Disruption of the single TM domain, soluble protein
RHCE Rhesus blood group, CcEe antigens NM_138617 NM_138618 Two exons between exon 3 and 4 Disruption of the fifth TM domain, insert contains three new TM domains
na na NM_014738 BM693684 Intron between exon 30 and 31 Disruption of the eighth TM domain
na na NM_152672 CF147426 Acceptor upstream of exon 4 Disruption of the second TM domain
*RefSeq transcript without the insert (shorter variant) that encodes a TM domain. †Transcript with the insert (longer variant) that destroys a TM helix. ‡Exon numbers refer to the RefSeq transcript with the TM helix. na, not approved; TM, transmembrane.
Table 2 RefSeq transcripts with an exon skipping splice form that puts together a new Pfam domain
Gene symbol Gene name RefSeq/EST with Pfam* RefSeq/EST without Pfam† Pfam ID Pfam description Alternative splice event‡ Pfam cutoff score§ Score upstream¶ Score downstream¥ Score combined#
na na NM_144604 AK056632 PF00642 Zinc finger C-x8-C-x5-C-x3-H type (and similar) Exon between exon 3 and 4 17.5 -1.2 9.4 23.6
PRSS25 protease, serine, 25 NM_145074 AF141306 PF00089 Trypsin Acceptor upstream of exon 4 23.4 3.0 1.1 30.8
FOSL2 FOS-like antigen 2 NM_005253 BX647822 PF00170 bZIP transcription factor Acceptor upstream of exon 4 23.2 16.1 -4.6 31.3
na na NM_003622 AB033056 PF02920 Integrase_DNA Exon between exon 8 and 9 18.0 13.4 -5.0 21.9
na na NM_006832 AK091532 PF00373 FERM domain (Band 4.1 family) Exon between exon 12 and 13 14.0 -15.9 10.3 15.6
PQBP1 Polyglutamine binding protein 1 NM_144494 BM692479 PF00397 WW domain Acceptor upstream of exon 3 17.0 5.0 9.7 32.5
MRPL27 Mitochondrial ribosomal protein L27 NM_148570 BQ028639 PF01016 Ribosomal L27 protein Acceptor upstream of exon 4 25.0 2.1 8.2 34.0
PLEKHB1 Pleckstrin homology domain containing, family B (evectins) member 1 NM_021200 BE703269 PF00169 PH domain Acceptor upstream of exon 3 22.8 -3.3 11.4 29.7
na na NM_020679 BP265352 PF02854 MIF4G domain Donor downstream of exon 6 14.0 1.1 0.2 17.2
TRUB2 TruB pseudouridine (psi) synthase homolog 2 (E. coli) BE793897 NM_015679 PF00849 RNA pseudouridylate synthase Skip exon 2 14.0 -2.1 -1.3 14.7
na na BM903757 NM_024565 PF00134 Cyclin, N-terminal domain Skip exon 4 17.0 0.3 9.6 52.9
na na BC033491 NM_139174 PF00035 Double-stranded RNA binding motif Skip exon 2 17.0 -5.2 13.5 21.7
*Transcript without the insert (shorter variant) that encodes a Pfam domain. †Transcript with the insert (longer variant) that does not encode a Pfam domain. ‡Exon numbers refer to the RefSeq transcript. §Per-domain 'gathering cut-offs' as given in the Pfam database. ¶,¥Pfam score for the partial domain encoded by the upstream and downstream exon, respectively. #Pfam score for the domain that is encoded by the splice form without the insert. na, not approved.
==== Refs
Graveley BR Alternative splicing: increasing diversity in the proteomic world. Trends Genet 2001 17 100 107 11173120 10.1016/S0168-9525(00)02176-4
Roberts GC Smith CWJ Alternative splicing: combinatorial output from the genome. Curr Opin Chem Biol 2002 6 375 383 12023119 10.1016/S1367-5931(02)00320-4
Hiller M Huse K Szafranski K Jahn N Hampe J Schreiber S Backofen R Platzer M Widespread occurrence of alternative splicing at NAGNAG acceptors contributes to proteome plasticity. Nat Genet 2004 36 1255 1257 15516930 10.1038/ng1469
Stamm S Ben-Ari S Rafalska I Tang Y Zhang Z Toiber D Thanaraj TA Soreq H Function of alternative splicing. Gene 2005 344 1 20 15656968 10.1016/j.gene.2004.10.022
Lewis BP Green RE Brenner SE Evidence for the widespread coupling of alternative splicing and nonsense-mediated mRNA decay in humans. Proc Natl Acad Sci USA 2003 100 189 192 12502788 10.1073/pnas.0136770100
Wollerton MC Gooding C Wagner EJ Garcia-Blanco MA Smith CWJ Autoregulation of polypyrimidine tract binding protein by alternative splicing leading to nonsense-mediated decay. Mol Cell 2004 13 91 100 14731397 10.1016/S1097-2765(03)00502-1
Garcia-Blanco MA Baraniak AP Lasda EL Alternative splicing in disease and therapy. Nat Biotechnol 2004 22 535 546 15122293 10.1038/nbt964
Kriventseva EV Koch I Apweiler R Vingron M Bork P Gelfand MS Sunyaev S Increase of functional diversity by alternative splicing. Trends Genet 2003 19 124 128 12615003 10.1016/S0168-9525(03)00023-4
Liu S Altman RB Large scale study of protein domain distribution in the context of alternative splicing. Nucleic Acids Res 2003 31 4828 4835 12907725 10.1093/nar/gkg668
Resch A Xing Y Modrek B Gorlick M Riley R Lee C Assessing the impact of alternative splicing on domain interactions in the human proteome. J Proteome Res 2004 3 76 83 14998166 10.1021/pr034064v
Hooper NM Karran EH Turner AJ Membrane protein secretases. Biochem J 1997 321 265 279 9020855
Xing Y Xu Q Lee C Widespread production of novel soluble protein isoforms by alternative splicing removal of transmembrane anchoring domains. FEBS Lett 2003 555 572 578 14675776 10.1016/S0014-5793(03)01354-1
Cline MS Shigeta R Wheeler RL Siani-Rose MA Kulp D Loraine AE The effects of alternative splicing on transmembrane proteins in the mouse genome. Pacific Symposium on Biocomputing: January 6-10 2004; Hawaii 2004 17 28
Minneman KP Splice Variants of G protein-coupled receptors. Mol Interv 2001 1 108 116 14993330
Garcia J Gerber SH Sugita S Sudhof TC Rizo J A conformational switch in the Piccolo C(2)A domain regulated by alternative splicing. Nat Struct Mol Biol 2004 11 45 53 14718922 10.1038/nsmb707
Kamatkar S Radha V Nambirajan S Reddy RS Swarup G Two splice variants of a tyrosine phosphatase differ in substrate specificity, DNA binding, and subcellular location. J Biol Chem 1996 271 26755 26761 8900155 10.1074/jbc.271.43.26755
Krogh A Larsson B Heijne Gv 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
Stamm S Signals and their transduction pathways regulating alternative splicing: a new dimension of the human genome. Hum Mol Genet 2002 11 2409 2416 12351576 10.1093/hmg/11.20.2409
Gupta R Birch H Rapacki K Brunak S Hansen JE O-GLYCBASE version 4.0: a revised database of O-glycosylated proteins. Nucleic Acids Res 1999 27 370 372 9847232 10.1093/nar/27.1.370
Diella F Cameron S Gemund C Linding R Via A Kuster B Sicheritz-Ponten T Blom N Gibson TJ Phospho.ELM: a database of experimentally verified phosphorylation sites in eukaryotic proteins. BMC Bioinformatics 2004 5 79 15212693 10.1186/1471-2105-5-79
Hansen JE Lund O Tolstrup N Gooley AA Williams KL Brunak S NetOglyc: prediction of mucin type O-glycosylation sites based on sequence context and surface accessibility. Glycoconj J 1998 15 115 130 9557871 10.1023/A:1006960004440
Blom N Gammeltoft S Brunak S Sequence and structure-based prediction of eukaryotic protein phosphorylation sites. J Mol Biol 1999 294 1351 1362 10600390 10.1006/jmbi.1999.3310
Bateman A Coin L Durbin R Finn RD Hollich V Griffiths-Jones S Khanna A Marshall M Moxon S Sonnhammer ELL The Pfam protein families database. Nucleic Acids Res 2004 32 Database issue D138 D141 14681378 10.1093/nar/gkh121
Hiller M Backofen R Heymann S Busch A Glaesser TM Freytag J-C Efficient prediction of alternative splice forms using protein domain homology. In Silico Biol 2004 4 195 208 15107023
Sorek R Shamir R Ast G How prevalent is functional alternative splicing in the human genome? Trends Genet 2004 20 68 71 14746986 10.1016/j.tig.2003.12.004
Xu Q Lee C Discovery of novel splice forms and functional analysis of cancer-specific alternative splicing in human expressed sequences. Nucleic Acids Res 2003 31 5635 5643 14500827 10.1093/nar/gkg786
Puntervoll P Linding R Gemund C Chabanis-Davidson S Mattingsdal M Cameron S Martin DMA Ausiello G Brannetti B Costantini A ELM server: a new resource for investigating short functional sites in modular eukaryotic proteins. Nucleic Acids Res 2003 31 3625 3630 12824381 10.1093/nar/gkg545
Sakurai Y Onishi Y Tanimoto Y Kizaki H Novel protein kinase C delta isoform insensitive to caspase-3. Biol Pharm Bull 2001 24 973 977 11558579 10.1248/bpb.24.973
Wang Z Rolish ME Yeo G Tung V Mawson M Burge CB Systematic identification and analysis of exonic splicing silencers. Cell 2004 119 831 845 15607979 10.1016/j.cell.2004.11.010
Newton DC Bevan SC Choi S Robb GB Millar A Wang Y Marsden PA Translational regulation of human neuronal nitric-oxide synthase by an alternatively spliced 5'-untranslated region leader exon. J Biol Chem 2003 278 636 644 12403769 10.1074/jbc.M209988200
Sorek R Ast G Graur D Alu-containing exons are alternatively spliced. Genome Res 2002 12 1060 1067 12097342 10.1101/gr.229302
Kondrashov FA Koonin EV Evolution of alternative splicing: deletions, insertions and origin of functional parts of proteins from intron sequences. Trends Genet 2003 19 115 119 12615001 10.1016/S0168-9525(02)00029-X
Modrek B Lee CJ Alternative splicing in the human, mouse and rat genomes is associated with an increased frequency of exon creation and/or loss. Nat Genet 2003 34 177 180 12730695 10.1038/ng1159
Ast G How did alternative splicing evolve? Nat Rev Genet 2004 5 773 782 15510168 10.1038/nrg1451
Sorek R Shemesh R Cohen Y Basechess O Ast G Shamir R A Non-EST-based method for exon-skipping prediction. Genome Res 2004 14 1617 1623 15289480 10.1101/gr.2572604
Huang H-D Horng J-T Lin F-M Chang Y-C Huang C-C SpliceInfo: an information repository for mRNA alternative splicing in human genome. Nucleic Acids Res 2005 33 Database Issue D80 D85 15608290 10.1093/nar/gki129
Human RefSeq Database
Human Fraction of dbEST
Moller S Croning MD Apweiler R Evaluation of methods for the prediction of membrane spanning regions. Bioinformatics 2001 17 646 653 11448883 10.1093/bioinformatics/17.7.646
SwissProt and RefSeq IDs 2001
| 15998447 | PMC1175989 | CC BY | 2021-01-04 16:05:40 | no | Genome Biol. 2005 Jun 22; 6(7):R58 | utf-8 | Genome Biol | 2,005 | 10.1186/gb-2005-6-7-r58 | oa_comm |
==== Front
Genome BiolGenome Biology1465-69061465-6914BioMed Central London gb-2005-6-7-r591599844810.1186/gb-2005-6-7-r59ResearchGenomic analysis of metabolic pathway gene expression in mice Ghazalpour Anatole [email protected] Sudheer [email protected] Sonal S [email protected] Leslie A [email protected] Eric E [email protected] Aldons J [email protected] Thomas A [email protected] Department of Human Genetics, Department of Medicine and Department of Microbiology, Immunology and Molecular Genetics, and Molecular Biology Institute, University of California, Los Angeles, CA 90095-1679, USA2 Department of Pathology and Laboratory Medicine, University of California, Los Angeles, CA 90095-1732, USA3 Rosetta Inpharmatics LLC, Kirkland, WA 98034, USA2005 1 7 2005 6 7 R59 R59 8 12 2004 1 2 2005 8 6 2005 Copyright © 2005 Ghazalpour 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 evaluate metabolic pathways associated with obesity, global gene-expression data were integrated with phenotypic and genetic segregation analyses, identifying 13 metabolic pathways the genes of which are coordinately regulated in association with obesity. Four genomic regions were found to control the coordinated expression of these pathways and novel genes potentially associated with the identified pathways were identified.
Background
A segregating population of (C57BL/6J × DBA/2J)F2 intercross mice was studied for obesity-related traits and for global gene expression in liver. Quantitative trait locus analyses were applied to the subcutaneous fat-mass trait and all gene-expression data. These data were then used to identify gene sets that are differentially perturbed in lean and obese mice.
Results
We integrated global gene-expression data with phenotypic and genetic segregation analyses to evaluate metabolic pathways associated with obesity. Using two approaches we identified 13 metabolic pathways whose genes are coordinately regulated in association with obesity. Four genomic regions on chromosomes 3, 6, 16, and 19 were found to control the coordinated expression of these pathways. Using criteria that included trait correlation, differential gene expression, and linkage to genomic regions, we identified novel genes potentially associated with the identified pathways.
Conclusion
This study demonstrates that genetic and gene-expression data can be integrated to identify pathways associated with clinical traits and their underlying genetic determinants.
==== Body
Background
Comprehensive high-throughput analytical techniques such as expression microarrays have brought about a rethinking of the possibility of understanding biology and disease at a global ('systems biology') level [1,2]. These techniques have been applied widely to the study of the time course of events in specific cells or organisms, and to different conditions for a given cell type or organism. More recently, there has been an appreciation of the possibility of using naturally occurring genetic variation as a means of generating perturbations, with the advantage of allowing identification of the individual genetic factors affecting the trait of interest in the segregating population [3]. We and others have begun to apply this approach to various model organisms where tracking of genetic segregation is feasible [4-10].
Traditional quantitative trait locus (QTL) analyses of complex traits in model organisms have focused on the identification of specific causative genes that differ between the originating strains and that are directly responsible for the variation in trait expression [11]. The availability of genome-wide expression data (or proteomic, metabolomic, or other such global analyses) to complement the measurements of the physiologic trait opens new opportunities for identifying specific biologic processes and genes that are involved in trait expression. As well as providing a means of evaluating many of the potential candidate genes responsible for a specific QTL, such data allow the identification of pathways and genes that have a role in the expression of the phenotype, either as intermediate players between the causative gene and the phenotype, or as being responsive to the trait [12].
We have been interested in using these approaches in mice to understand the pathogenesis of obesity and vascular disease, among other related conditions [6,13]. Metabolic dysregulation has been recognized to be an important element in the pathogenesis of these poorly understood conditions. An important dataset for these purposes is the (C57BL/6 × DBA/2)F2 (BxD) intercross presented in Schadt et al. [6]. This set comprises the microarray data from the liver of BxD F2 female mice fed an atherogenic diet for 4 months beginning at 12 months of age. In this study, we integrate the global gene-expression data with phenotypic and genetic segregation analyses to evaluate metabolic pathways associated with obesity in the BxD set. We show that this approach allows identification of specific pathways whose gene expression is coordinately regulated in association with obesity, defined genomic regulatory loci controlling the expression of these genes, and novel genes that are functionally associated with the identified pathways.
Results
Identification of gene sets/pathways associated with the subcutaneous fat pad mass trait
To identify functionally related gene sets that are differentially perturbed in lean and obese mice of the BXD cross, two methods were applied: gene set enrichment analysis (GSEA) [14] and over-representation analysis using Fisher's exact test. For both of these, we began by selecting a set of 4,670 genes that were differentially expressed in the liver, filtered as described in the Materials and methods section. In both analyses, we used the same 378 gene sets compiled mainly from the Kyoto Encyclopedia of Genes and Genomes (KEGG) [15] and Biocarta [16] sources. GSEA was implemented as previously described [14]. To apply GSEA, the 4,670 filtered genes were ranked on the basis of significance of differential expression between the obese and lean groups of mice (identified as those mice in the upper and lower 15th percentiles for the subcutaneous fat-pad mass trait, respectively). The Kolmogorov-Smirnov (KS) test was then applied as described in [14] to test the null hypothesis that members of a given gene set are randomly ranked. This procedure generates an enrichment score (ES) for each gene set. The significance of the ES statistic was determined empirically by performing the analysis after random permutation of the grouping assignment of the mice, with the probability of falsely rejecting the null hypothesis determined by repeating the permutation procedure 1,000 times. This established that ESs greater than 114 allowed rejection of the null hypothesis at a level of 0.05. Using this empirically determined threshold for the ES statistic, we identified 13 gene sets that were differentially regulated between mice that have high and low subcutaneous fat-pad mass (Figure 1).
As a second approach, we used an analysis of over-representation of gene set members by applying the Fisher's exact test, as implemented by the Expression Analysis Systematic Explorer (EASE) software [17]. The results are shown in Table 1 and are comparable to those of the GSEA analysis. Ten of the 13 pathways identified as significantly differentially expressed using the GSEA approach were also identified as significantly differentially expressed by the EASE analysis after application of the Hochberg adjustment for multiple comparisons. Two of the other three gene sets identified by GSEA were the 12th- and 13th-ranked sets by the EASE analysis. One gene set (tetrachloroethene degradation) was identified as significant by the Fisher's exact analysis but not by GSEA.
Relationships among pathway sets, gene expression, and the fat-mass phenotype
The gene sets identified as being differentially regulated between obese and lean mice represented metabolic pathways involved in energy and lipid metabolism. Note that genes can contribute to multiple pathways, as different pathways can be biologically interrelated. The majority of pathways feed into the tricarboxylic acid (TCA) cycle (Figure 2). The aggregate gene set from all 13 fat-mass-associated pathways comprises 170 genes, 150 of which were represented on the array. An examination of the correlation between relative gene-expression levels and the fat-mass trait showed that 68 of the 150 genes (45%) in the aggregate set were correlated at p < 0.05 (not corrected for multiple comparisons). A large majority of these genes (61/68) showed a significant positive correlation (that is, they were upregulated in obese compared with lean mice), with relatively few being negatively correlated (10%). Within individual pathway sets, between 42% and 63% of the genes were significantly correlated. The majority of correlated genes in a given set were always positively correlated. The magnitude of the correlations was comparable across the 13 pathways.
Identification of genetic loci controlling differentially regulated pathway genes
We hypothesized that the composite set of pathways represented a functionally interrelated gene set responding to common genetic controls (referred to subsequently as the obesity-associated pathway set). Two approaches were taken to test for such common genetic controls. The first was to test for the enrichment of pathway gene-expression quantitative trait loci (eQTLs) along genomic intervals, and the second was principal component analysis (PCA) of the gene set, followed by QTL mapping of the derived principal components.
When eQTLs for the 4,670 most differentially expressed genes were mapped we found that there were 15,262 eQTLs with log10 of the odds (LOD) score greater than 2.0. Of these eQTLs, 278 corresponded to 77 genes represented in the 13 differentially regulated pathways identified above by the GSEA procedure. To examine whether there was an enrichment of pathway gene eQTLs among the eQTLs of the 4,670 most differentially expressed genes along the genomic intervals, we first divided the genome into sequential overlapping 20 centimorgan (cM) intervals, beginning a new interval every 10 cM. We then counted how many of the eQTLs overall for the 4,670 differentially expressed genes and how many of the eQTLs specific for the 77 pathway genes defined above mapped to each interval. In general, assuming under the null hypothesis that the genomic location of genes influencing the transcript levels of other genes in common pathways is independent, the expected number of pathway gene eQTLs mapping to each interval is (M/B) × K where B is 4,670, M is the subset from the 4,670 genes that have eQTLs mapping to the interval, and K is 77. To compute the statistics we applied Fisher's exact test. The p-values obtained for the statistics are shown in Table 2. For these calculations we applied a total of 180 tests overall, corresponding to 180 bins that comprise the entire genome. The significance levels of these tests are reported in Table 2 under 'adjusted p-value', which reflects the p-values after applying the Bonferroni correction criteria (multiplying the p-values by 180, the number of tests).
After correcting for multiple comparisons, significant over-representation of pathway gene eQTLs was obtained for four loci located on chromosomes 3 (80-100 cM), 6 (30-50 cM), 16 (1-20 cM), and 19 (20-50 cM) (Figure 3, Table 2). All pathways were represented at each locus and there were no apparent subgroups associated with some loci and not others. We also found that about half of the genes had eQTLs at two or more of the four loci.
An alternative approach that has been proposed to test for coordinated regulation of the genes in individual pathways is PCA [7,18]. PCA was applied using the gene sets from all 13 significant pathways identified by GSEA, and the aggregate set of 150 genes comprising these 13 pathways. The mean log ratios of the gene expression for all genes within these pathways across the F2 mice were subject to the analysis. Principal components (PC) that explained more than 5% of the variance of transcript levels of the genes within each pathway were then mapped as quantitative traits. Only three components showed significant LOD scores (> 4.3). These included the third PC for fatty acid metabolism-1 pathway (LOD 21.4), the fourth PC for γ-hexachlorocyclohexane degradation (LOD 14.8), and the fourth PC of the tryptophan degradation pathway (LOD 8.1). All these linkages mapped to the same locus at 36 cM on chromosome 19. The percentage of variance explained by each of these PCs was 8%, 8%, and 7%, respectively. QTL analyses of individual genes in the fatty-acid metabolism and γ-hexachlorocyclohexane degradation pathway revealed the presence of two cis-eQTLs at the chromosome 19 locus with very significant LOD scores: Cyp2c40 (LOD = 20.7) and Cyp2c39 (LOD = 12.4). To see whether these two genes were the major contributors to mapping of the PCs to chromosome 19, we carried out the analysis without these two genes for each of the pathways separately. The resulting components showed no evidence of significant linkage genome-wide.
Relationship of pathway regulatory regions to the subcutaneous fat-mass trait
The loci identified above are genomic regions with apparent regulatory control of pathway genes that are differentially perturbed between obese and lean mice. The locus allele effects on the subcutaneous fat-mass trait, using the nearest markers at each locus, are shown in Figure 4. The loci on chromosomes 6 and 19, which involve the greatest number of pathway genes, showed the strongest effects and in fact had been identified as being linked to adiposity in previous studies [6,19]. Consistent allele effects were observed among the subset of genes from the aggregate pathway set that had eQTLs at both the chromosome 6 and chromosome 19 loci (that is, a gene that showed a positive correlation with fat mass showed greater expression with the B6 versus DBA homozygous mice at the chromosome 19 locus, and the reverse at the chromosome 6 locus).
Use of expression data for prioritization of candidate gene selection at the chromosome 19 locus
The chromosome 19 locus had the greatest impact on the expression of pathway set gene expression, and it was of interest to identify the candidate responsible genes for further study. For candidate genes regulated at the level of expression (that is, the gene controls its own transcript abundance), an eQTL will be present proximal to the physical location of the gene; we have referred to such an eQTL as a cis-eQTL [6]. Cis-eQTLs were defined as any gene with eQTLs mapping to 20 Mb (approximately 10 cM) on either side of the gene's physical location. This definition reflects the relative accuracy of the 13 cM dense BxD QTL map. The decision to choose such a definition for cis-eQTLs is supported by the work of Doss et al. [20], who showed that this definition is reasonable by validating the cis nature of such defined eQTLs and by showing the extreme bias in the mapping of these cis eQTLs to regions which are not identical by descent between the B6 and DBA strains of mice.
We identified genes with significant cis-acting eQTL (LOD score > 4.3) that map to chromosome 19 between 20 and 50 cM, and determined the genomic location of each gene using the University of California Santa Cruz (UCSC) Genome Browser (build 33, mm5, May 2004). Of 249 eQTLs mapping to the chromosome 19 locus, 19 overlapped the physical location of their respective gene (Table 3), thus establishing them as primary candidate genes. Of these, eight were significantly correlated with the fat-mass trait, making them particularly attractive candidates.
Integration of trait, expression, and mapping data to identify novel genes potentially related to the obesity-associated gene set
The above analyses were restricted to the genes that had been assigned to gene sets in the starting databases such as KEGG. To identify novel genes that may belong to the same aggregate group of obesity-associated pathway genes, we filtered the 4,670 differentially regulated genes first for correlation with the fat-mass trait at p < 0.05 (uncorrected, corresponding to a correlation coefficient of approximately 0.25) and subsequently for the presence of eQTLs falling within the identified regulatory regions for both chromosome 6 (30 to 50 cM interval) and 19 (20 to 50 cM interval). A total of 117 genes were identified that were not members of one of the 13 obesity-associated pathways. These include 20 genes with defined relationships to the starting pathways (based on published literature), but also 28 expressed sequence tags (ESTs) with no functional annotation. A complete list is provided in Additional data file 1.
Discussion
This study shows that the incorporation of genome-wide gene-expression data with traditional phenotypic trait data in a mouse intercross population enables the identification of regulated metabolic pathways and the genomic regions that control the expression of the constituent pathway genes. In contrast to studies with genetically altered mice, we have studied a heterogeneous population where the genetic variations leading to the phenotypes and the changes in gene expression are all naturally occurring. In this work, we identified metabolic pathway gene sets with altered gene expression in association with the subcutaneous fat-pad mass trait, but the concept is applicable to other phenotypic traits where the target organ examined would be expected to show changes in relevant pathway gene expression. While gene-expression changes are not sufficient to imply actual metabolite flux through a pathway [21], changes in gene expression are likely to reflect pathway involvement in a given condition.
Two approaches were used to identify gene sets whose transcripts were altered in association with the fat-mass trait: the GSEA method and Fisher's exact test. These yielded comparable results, although the GSEA appeared to be more sensitive. Central to both is the availability or construction of predefined gene sets to test. We used two primary sources, the KEGG database, composed primarily of traditional metabolic pathways, and the Biocarta database, composed of gene sets of varying type. The total number of genes represented in these was relatively low (just under 2,000). There are certain to be biologically important gene sets as yet unrecognized, as well as currently uncategorized genes that belong in the existing sets. Our approach allows tentative functional assignment to such uncategorized genes when they share properties of correlation of expression levels with the phenotypic trait and co-localization of eQTL with the trait QTL.
Consistent with an obesity trait, the gene sets identified were of pathways directly or indirectly related to energy metabolism. Within any given pathway, only about half of the genes showed a correlation between transcript levels and fat mass, consistent with the concept that regulation of a metabolic pathway requires the control of only selected elements rather than all constituents. It is of interest that the pathways are at least loosely interconnected, and several of the genes examined have been shown to influence adipose tissue mass in other studies. For example, of the total of 150 pathway genes represented on the microarrays, the gene with the highest differential expression between obese and lean mice was amine oxidase copper-containing 3 (Aoc3) (12.5-fold change in expression level). This observation is consistent with the study in which overexpression of human AOC3 in mice resulted in an increase in weight and subcutaneous white adipose tissue when the animals were fed an atherogenic diet for 15 weeks [22]. Ten of the 13 pathways are connected through the generation of products involved in the TCA cycle (Figure 2). The other three pathways (bile acid biosynthesis, androgen/estrogen metabolism, and γ-hexachlorocyclohexane degradation) are involved in cholesterol metabolism. Coordinated control of cholesterol and fatty-acid synthesis has been recognized for some time [23]. The production of estrogen and androgen requires cholesterol as a precursor and the activity of the cytochrome P450 family, members of which were significantly upregulated in the obese F2 mice. In the bile acid pathway, cholesterol serves as a substrate that is converted to bile acids through the activation of 7-α-hydroxylase, the transcript for which is also significantly elevated in obese mice. Lastly, studies show that γ-hexachlorocyclohexane administration has been shown to increase serum cholesterol and very low-density lipoprotein (VLDL) levels [24]. Several genes occur multiple times in these 13 pathways (for example, the aldehyde dehydrogenase gene family (Aldh1a1, Aldh2, and Aldh3a2)). These enzymes are involved in the generation of acetate, which can be converted to acetyl CoA. The family is also involved in ascorbate/aldarate metabolism, valine, leucine, and isoleucine degradation, propanoate metabolism, fatty-acid metabolism, glycerolipid metabolism, β-alanine metabolism, bile acid biosynthesis, and pyruvate metabolism pathways (eight out of the 13 pathways). The increased expression of the aldehyde dehydrogenases such as Aldh3a2 in mice with higher subcutaneous fat-pad mass is consistent with recent studies that correlate hyperinsulinemia with increased hepatic expression of Aldh3a2 [25].
We were able to identify specific genomic regions on chromosomes 3, 6, 16, and 19 that regulated these pathways. Gene transcript levels are themselves quantitative traits for which genetic control regions can be identified by QTL analysis (eQTLs), just as for a traditional physiologic trait [4-6,26]. Loci that coordinately regulate pathway genes were identified by analyzing for over-representation of identified pathway-associated gene eQTLs among eQTLs of the set of most differentially expressed genes in a given genomic region. A possible criticism of this approach is the use of eQTLs with low LOD scores. While a given eQTL with a low LOD score cannot be considered significant in the genome-wide context on its own, the co-localization of multiple related eQTLs with LOD scores that may not meet the genome-wide significance criteria is highly unlikely to have occurred by chance, as we show in Table 2. It is unlikely that such eQTL clustering results from correlated traits linking to a particular marker by chance, given that the percent variation explained by any such marker is modest (typically less than 10%) and given that the correlation between traits is also modest (typically explaining less than 30% of the variation between the traits). Therefore, because the covariance between the two traits is small, it is unlikely that the covariance between the two traits captures the by-chance covariance between the traits and marker. The other approach that we used to identify pathway regulatory loci - PCA - proved less useful. When used on random or less correlated groups of genes, PCA spreads the variance components of the dataset over several different component vectors, thereby diluting the very signal one is trying to capture. Other techniques used for the decomposition of multivariate data such as non-negative matrix factorization [27] are being explored and may offer a more effective approach for this type of analysis.
The loci on chromosomes 6 and 19 had the greatest number of eQTLs, and were the two loci identified in previous studies to be significantly linked with the subcutaneous fat-mass trait. It is of interest that the most significant locus controlling this fat-mass trait is located on distal chromosome 2 [19], which was not represented among the loci we identified here. In the analysis reported in Schadt et al. [6], the loci on chromosomes 2 and 19 were associated with distinct subsets of mice, identified by differing expression profiles of the most differentially regulated genes. The gene sets we identified were associated with the chromosome 19 locus (and others) but not the chromosome 2 locus. There are several possible explanations (not mutually exclusive). One is that the responsible 'pathways' or gene sets involved with the chromosome 2 locus are not represented in our sets. A second is that too few genes of an associated pathway are differentially regulated to detect over-representation. A third is that the associated pathways primarily involve tissues that were not arrayed. That is, because only liver expression data were available, expression differences in other relevant tissues such as adipose, muscle, brain, or gut could be primarily related to the chromosome 2 QTL, but were not detected as these tissues were not examined.
We show that there are hotspots of eQTL activity for pathway gene members along the genome. One explanation for the presence of these hotspots is the presence of a major regulator gene within each locus that coordinately regulates obesity-related pathways. Alternatively, these loci may contain several such regulators that are closely linked. It is well known that clusters of duplicated genes with related functions occur commonly in the genome. Other clusters with diverse genes but related functions, such as the H-2 locus, are also known. Yet another possible explanation is that the genes with eQTL mapping to a common hotspot are in linkage disequilibrium (LD). Two or more genes that differ genetically between the two strains and that are in LD will show correlated expression levels [20]. This spurious correlation could lead to a more convoluted trans-eQTL hotspot for these genes which is more difficult to interpret [20]. We determined whether this situation might be occurring in our data and saw that in the 77 active pathway genes only 13 have cis-eQTLs for which there is a minimal clustering of physical location (data not shown). On this basis we conclude that LD is unlikely to explain the hotspots identified in this study.
An approach to prioritizing candidate genes using eQTL information is presented for the chromosome 19 locus. This uses the concept that for genes regulated at the level of transcription, cis-acting regulation can be identified by the coincidence of an eQTL for a gene and its physical location. We have experimentally verified that the majority of such putative cis-eQTLs are real [20]. This approach relies, however, on variations affecting transcript levels and will miss candidate genes with variations affecting protein function or post-transcriptional modification, as well as variants associated with alternative splicing. In addition, the list of candidate genes is limited to genes represented on the microarray chip.
Conclusion
In conclusion, this study presents one approach for using gene-expression microarray data in conjunction with genetic segregation to identify metabolic pathways that are altered in association with obesity, and to identify specific genomic regions that exert regulatory control over these pathways. These findings contribute, along with other approaches developing genetic regulatory networks [12], to the ultimate goal of understanding disease pathophysiology at the systems biology level.
Materials and methods
Animals, tissue collection, and gene-expression profiling
This study, along with others previously reported [6,19,28], was based on data originating from an F2 intercross carried out between mouse strains C57BL/6J and DBA/2J. This intercross encompassed a two-phase study as described in Colinayo et al. [28], using 142 female F2 mice with genome-wide genotyping initially, followed by an additional 144 female mice to confirm specific QTLs (with genotyping limited to selected loci). Subsequently, genome-wide expression microarray analyses were obtained on portions of stored liver samples; 111 of these were from the initial 142 mouse set that was reported by Schadt et al. [6], and an additional 44 were from a second set that lacked genome-wide genotype data.
All F2 progeny were fed a chow diet until 12 months of age and then put on a high-fat high-cholesterol atherogenic diet containing 0.5% cholic acid for 4 months (diet 90221; Harlan Teklad). At 16 months of age these mice were euthanized and liver and other organs were removed and frozen at -70°C for subsequent RNA and DNA extraction. Body weight was determined before organ removal and adipose tissue depots were removed and weighed for a subset of 96 animals in the phase 1 group.
RNA extraction and expression microarray profiling were carried out as described using a custom ink-jet microarray (Agilent Technologies) containing 23,574 60mer oligonucleotide non-control probes for profiling gene expression in liver of the F2 mice [6]. Other tissues were not available for analysis. The microarray data discussed in this publication have been deposited in the Gene Expression Omnibus [29] of the National Center for Biotechnology Information (NCBI) and are accessible through GEO Series accession number [GSE:2008]. The genotype and the phenotype of F2 mice are given in Additional data file 2. Overall, 155 F2 mice had liver array data, 111 had both array and genome-wide genotype information, and 69 had array, genotype, and fat-mass trait data available.
Linkage analysis
QTL analysis was performed using mean log10 expression ratio values (averaged over fluor-reverse pairs) as quantitative traits for the 111 mice that were part of the phase 1 group described above. The linkage map was constructed using microsatellite markers at an average density of 13 cM using MapManager QTX version 0.30 and QTL Cartographer version 2.0 [30-32]. For each trait, the LOD score was calculated at 1 cM intervals using standard interval mapping.
Gene set and pathway assembly
A total of 378 gene sets were developed for gene set enrichment analysis (GSEA). Gene sets (118) representing metabolic pathways were obtained from KEGG [15,33,34]. Ten gene sets comprising genes that are highly correlated across 46 tissues in mice were selected [35]. These data were obtained using the SOURCE database [36,37]. Seven pathways were manually curated from the published literature for genes involved in insulin signaling. Two hundred and forty-four gene sets were obtained by querying all the pathways available at Biocarta [16]. The latter set comprises both signaling and metabolic pathways submitted to the Biocarta server by independent scientists. Gene sets composed of more than 50 members were divided into smaller sets to avoid bias related to gene set size in the GSEA analysis [38].
Microarray data filtering and ranking
A subset of genes represented on the microarray were selected on the basis of differential expression across all the 155 mice arrayed, irrespective of phenotype. This was done to ensure a starting gene set that was active and differentially expressed in liver among mice. To filter for these genes, the expression ratio and the corresponding p-value for differential expression (compared to the control pool) for each gene were calculated as described [6,39]. Genes with p-values of 0.05 or less in 10 or more F2 mice were defined as differentially expressed, which yielded a set of 4,670 genes. To study gene subsets (within these 4,670 genes) that are associated with obesity, the mice comprising the upper and lower 15th percentiles of the subcutaneous fat-pad mass trait were selected. The average and standard deviation of the mean log ratio of expression of each differentially expressed gene for these two groups of mice were calculated and used for the GSEA and the Fisher's exact test analyses described below.
Gene set enrichment analysis
Gene sets that are differentially regulated in relation to the subcutaneous fat-pad mass trait were identified using the GSEA method as described in [14]. The groups of mice analyzed were obese and lean F2 mice as defined above. To calculate the significance of the enrichment score (ES) assigned to each pathway, class labels of lean and obese mice were randomly permuted and ESs were recalculated 1,000 times. The cutoff for significance of ESs in the original analysis was defined as the score above the 50th highest score calculated in the permutation, corresponding to an empirical p-value of 0.05.
Fisher's exact test analysis
The test subset was composed of transcripts identified as being the 20% most differentially regulated between obese and lean groups of mice. Fisher's exact test was applied as implemented in the Expression Analysis Systemic Explorer (EASE) program [17], downloaded from the Database for Annotation, Visualization, and Integrated Discovery (DAVID) [40,41]. The pathway set we compiled was incorporated as a file into the program for use in the analysis (LocusLink and pathway identifier for each gene associated with each pathway). Genes identified as being differentially regulated as described above, but not belonging to an identified pathway or gene set were included with a designation 'unclassified', so that the EASE program would perform the Fisher's exact analysis using the full set of differentially regulated genes, rather than just those associated with an identified pathway or gene set. The set of all analyzed transcripts on the array for which unique LocusLink identifiers were available was entered as the population set in the analysis, although in the EASE analysis this set is limited to the genes that are represented in the pathway being evaluated.
cis-eQTL identification
We adopted the same cis-eQTL definition as described in [20]. An eQTL was defined as a proximal or cis-eQTL when the eQTL mapped within a 20 megabase (Mb) window on either side of the physical location of the gene. This corresponds to an approximately 20 cM segment of the genetic map, which is the approximate level of accuracy for a QTL given a genetic map with an average inter-marker distance of 13 cM. The following approach was taken to equate cM locations on the genetic map (constructed by the QTL software for this specific cross) with the Mb locations of the physical map, in order to be able to define eQTL as cis or trans. The physical location of the gene was obtained from the UCSC Genome Browser (build 33, mm5, May 2004). To map the genetic map location of the eQTL to the approximate equivalent location on the physical map, the physical map locations of the microsatellite markers immediately adjacent (proximal and distal) to the eQTL were identified from the UCSC genome browser (build 33, mm5, May 2004). Then the Mb 'location' of the eQTL was determined by interpolation based on the relative location of the eQTL between the markers on the genetic map.
Principal components analysis
PCA was performed as implemented in the R statistical software version 1.8.1 [42] using the 'base' package. The principal components and eQTL were mapped using the standard interval mapping procedures implemented in MapManager QTX version 0.30 and composite interval mapping implemented in the Windows QTL Cartographer software version 2.0 [30-32].
Additional data files
The following additional data are available with the online version of this paper. Additional data file 1 is a table listing genes that were not included in any tested gene set but are significantly correlated with subcutaneous fat mass and have eQTLs on chromosomes 6 and 19 (ordered alphabetically by gene name). Additional data file 2 is a table listing genotype and fat-mass trait data for the mice used in this study.
Supplementary Material
Additional File 1
A table listing genes that were not included in any tested gene set but are significantly correlated with subcutaneous fat mass and have eQTLs on chromosomes 6 and 19 (ordered alphabetically by gene name).
Click here for file
Additional File 2
A table listing genotype and fat-mass trait data for the mice used in this study.
Click here for file
Acknowledgements
This work was supported by grants HL28481 and HL70526 (T.A.D. and A.J.L.) and HL30568 (A.J.L.) from the National Institutes of Health, USPHS National Research Service Award GM07104 (A.G.), and the Bristol Meyers Squibb unrestricted research grant (A.J.L.).
Figures and Tables
Figure 1 Significantly differentially regulated pathways identified by GSEA analysis. The histogram represents the frequency distribution of maximum ES scores obtained with GSEA analysis after 1,000 random permutations of class assignment (obese or lean). The location of the gray arrow corresponds to p = 0.05 (ES = 114.0), and pathways with scores greater than this (indicated by the black arrows) are listed.
Figure 2 Relationship of the 13 differentially regulated pathways identified by GSEA. Nine pathways were related via the tricarboxylic acid (TCA) cycle and three other pathways are linked through cholesterol metabolism. These pathways contain a total of 170 genes, some of which are present in multiple pathways.
Figure 3 Identification of genomic regions controlling obesity-associated pathway gene expression. The x-axis represents sequential 10-cM intervals across the genome and the y-axis represents the negative log10 of adjusted Fisher's exact test p-value for enrichment of eQTLs in overlapping 20-cM bins. The dashed line corresponds to the p-value 0.05 after correcting for multiple comparison. The four significantly enriched regions on chromosomes 3, 6, 16, and 19 are indicated by arrows.
Figure 4 Homozygous allele effects on fat-pad mass of mice at the four loci identified to regulate obesity associated pathway gene members. (a) Chromosome 3 locus (peak marker at D3Mit86); (b) chromosome 6 locus (peak marker at D6Mit149); (c) chromosome 16 locus (peak marker at D16Mit100), and (d) chromosome 19 locus (peak marker at D19Mit63). B6, C57BL/6J; DBA, DBA/2J. *p-value < 0.05 calculated using two-sample t-test.
Table 1 Top 13 differentially regulated pathways identified by EASE analysis ranked by their corresponding p-values
Gene category p-value Adjusted p-value
Fatty-acid metabolism 6.91 E-08 1.48 E-05
Tryptophan metabolism 3.72 E-06 2.94 E-04
Butanoate metabolism 4.12 E-06 2.94 E-04
Glycerolipid metabolism 7.45 E-06 3.99 E-04
Valine, leucine, isoleucine degradation 4.22 E-05 1.60 E-03
Gamma-hexachlorocyclohexane degradation 4.48 E-05 1.60 E-03
Beta-alanine metabolism 6.10 E-05 1.86 E-03
Bile acid biosynthesis 1.54 E-04 4.13 E-03
Ascorbate and aldarate metabolism 1.54 E-03 3.33 E-02
Tetrachloroethene degradation 1.56 E-03 3.33 E-02
Androgen and estrogen metabolism 1.87 E-03 3.64 E-02
Pyruvate metabolism 3.72 E-03 7.56 E-01
Propanoate metabolism 6.12 E-03 9.93 E-01
Table 2 Chromosomal loci with significant enrichment for obesity-associated pathway gene eQTLs relative to eQTLs of all 4,670 most differentially expressed genes
Chromosome (cM) Total number of genes with eQTL Expected number of pathway genes with eQTL Observed number of pathway genes with eQTL p-value Adjusted p-value
3 (80-100) 25 1 9 9.24 E-11 1.68 E-08
6 (30-50) 836 14 28 8.05 E-05 1.47 E-02
16 (0-20) 431 7 21 3.44 E-06 6.27 E-04
19 (20-40) 891 15 38 1.45 E-09 2.63 E-07
19 (30-50) 709 12 31 6.77 E-08 1.23 E-05
Table 3 Nineteen candidate genes with cis-eQTL in chromosome 19 locus
cis-QTL gene symbol Physical location of the gene (Mb) Estimated physical location of the eQTL (Mb) Correlation with subcutaneous fat mass
RIKEN cDNA 9530025L08 29.5 48.6 -0.29*
Cyp2c40† 39.1 40.6 0.24
Col17a1 47.0 40.6 -0.25*
Scd2 43.6 40.6 -0.19
Slk 46.9 41.8 -0.07
Cyp2c37 39.3 37.0 0.32*
Cyp2c39† 38.8 40.6 0.03
Gsto1 47.2 41.8 0.26*
Rnf134 46.4 48.6 0.05
Cutc 43.1 39.4 0.25*
Cpn1 43.3 39.4 0.28*
Cnnm1 42.8 46.2 0.27*
Acsl5 54.6 37.0 0.30*
RIKEN cDNA 1810018P12 32.0 35.7 0.24
Pik3ap1 40.6 39.4 -0.07
RIKEN cDNA 0610010D20 41.4 41.8 0.10
RIKEN cDNA 4833409A17 42.1 40.6 -0.21
RIKEN cDNA 4930538D17 45.7 38.2 -0.22
RIKEN cDNA 5730455O13 37.4 40.6 0.14
*Significant correlation of transcript levels with subcutaneous fat mass (p < 0.05). †The two genes that solely contributed to the significant LOD score of the fatty-acid metabolism and γ-hexachlorocyclohexane degredation principal components to chromosome 19 (see text for details).
==== Refs
Kitano H Computational systems biology. Nature 2002 420 206 210 12432404 10.1038/nature01254
Hood L Heath JR Phelps ME Lin B Systems biology and new technologies enable predictive and preventative medicine. Science 2004 306 640 643 15499008 10.1126/science.1104635
Jansen RC Studying complex biological systems using multifactorial perturbation. Nat Rev Genet 2003 4 145 151 12560811 10.1038/nrg996
Brem RB Yvert G Clinton R Kruglyak L Genetic dissection of transcriptional regulation in budding yeast. Science 2002 296 752 755 11923494 10.1126/science.1069516
Wayne ML McIntyre LM Combining mapping and arraying: an approach to candidate gene identification. Proc Natl Acad Sci USA 2002 99 14903 14906 12415114 10.1073/pnas.222549199
Schadt EE Monks SA Drake TA Lusis AJ Che N Colinayo V Ruff TG Milligan SB Lamb JR Cavet G Genetics of gene expression surveyed in maize, mouse and man. Nature 2003 422 297 302 12646919 10.1038/nature01434
Lan H Stoehr JP Nadler ST Schueler KL Yandell BS Attie AD Dimension reduction for mapping mRNA abundance as quantitative traits. Genetics 2003 164 1607 1614 12930764
Klose J Nock C Herrmann M Stuhler K Marcus K Bluggel M Krause E Schalkwyk LC Rastan S Brown SD Genetic analysis of the mouse brain proteome. Nat Genet 2002 30 385 393 11912495 10.1038/ng861
Stoll M Cowley AW JrTonellato PJ Greene AS Kaldunski ML Roman RJ Dumas P Schork NJ Wang Z Jacob HJ A genomic-systems biology map for cardiovascular function. Science 2001 294 1723 1726 11721057 10.1126/science.1062117
Nadeau JH Burrage LC Restivo J Pao YH Churchill G Hoit BD Pleiotropy, homeostasis, and functional networks based on assays of cardiovascular traits in genetically randomized populations. Genome Res 2003 13 2082 2091 12952877 10.1101/gr.1186603
Abiola O Angel JM Avner P Bachmanov AA Belknap JK Bennett B Blankenhorn EP Blizard DA Bolivar V Brockmann GA The nature and identification of quantitative trait loci: a community's view. Nat Rev Genet 2003 4 911 916 14634638
Zhu J Lum PY Lamb J GuhaThakurta D Edwards SW Thieringer R Berger JP Wu MS Thompson J Sachs AB An integrative genomics approach to the reconstruction of gene networks in segregating populations. Cytogenet Genome Res 2004 105 363 374 15237224 10.1159/000078209
Ghazalpour A Doss S Yang X Aten J Toomey EM Van Nas A Wang S Drake TA Lusis AJ Thematic review series: the pathogenesis of atherosclerosis. Toward a biological network for atherosclerosis. J Lipid Res 2004 45 1793 1805 15292376 10.1194/jlr.R400006-JLR200
Mootha VK Lindgren CM Eriksson KF Subramanian A Sihag S Lehar J Puigserver P Carlsson E Ridderstrale M Laurila E PGC-1alpha-responsive genes involved in oxidative phosphorylation are coordinately downregulated in human diabetes. Nat Genet 2003 34 267 273 12808457 10.1038/ng1180
Kanehisa M Goto S Kawashima S Okuno Y Hattori M The KEGG resource for deciphering the genome. Nucleic Acids Res 2004 32 Database D277 D280 14681412 10.1093/nar/gkh063
Biocarta Pathways Database
Hosack DA Dennis G JrSherman BT Lane HC Lempicki RA Identifying biological themes within lists of genes with EASE. Genome Biol 2003 4 R70 14519205 10.1186/gb-2003-4-10-r70
Montooth KL Marden JH Clark AG Mapping determinants of variation in energy metabolism, respiration and flight in Drosophila. Genetics 2003 165 623 635 14573475
Drake TA Schadt E Hannani K Kabo JM Krass K Colinayo V Greaser LE 3rdGoldin J Lusis AJ Genetic loci determining bone density in mice with diet-induced atherosclerosis. Physiol Genomics 2001 5 205 215 11328966
Doss S Schadt E Drake T Lusis A Cis-acting expression quantitative trait loci in mice. Genome Res 2005 15 681 691 15837804 10.1101/gr.3216905
Hellerstein MK In vivo measurement of fluxes through metabolic pathways: the missing link in functional genomics and pharmaceutical research. Annu Rev Nutr 2003 23 379 402 12704218 10.1146/annurev.nutr.23.011702.073045
Stolen CM Madanat R Marti L Kari S Yegutkin GG Sariola H Zorzano A Jalkanen S Semicarbazide sensitive amine oxidase overexpression has dual consequences: insulin mimicry and diabetes-like complications. FASEB J 2004 18 702 704 14977883
Gibbons GF Regulation of fatty acid and cholesterol synthesis: co-operation or competition? Prog Lipid Res 2003 42 479 497 14559068 10.1016/S0163-7827(03)00034-1
Grajewski O Oberdisse E Increase of serum very low density lipoproteins in rats after administration of alpha-hexachlorocyclohexane. Naunyn Schmiedebergs Arch Pharmacol 1977 298 129 135 69995 10.1007/BF00508619
Demozay D Rocchi S Mas JC Grillo S Pirola L Chavey C Van Obberghen E Fatty aldehyde dehydrogenase: potential role in oxidative stress protection and regulation of its gene expression by insulin. J Biol Chem 2004 279 6261 6270 14638678 10.1074/jbc.M312062200
Morley M Molony CM Weber TM Devlin JL Ewens KG Spielman RS Cheung VG Genetic analysis of genome-wide variation in human gene expression. Nature 2004 430 743 747 15269782 10.1038/nature02797
Lee DD Seung HS Learning the parts of objects by non-negative matrix factorization. Nature 1999 401 788 791 10548103 10.1038/44565
Colinayo VV Qiao JH Wang X Krass KL Schadt E Lusis AJ Drake TA Genetic loci for diet-induced atherosclerotic lesions and plasma lipids in mice. Mamm Genome 2003 14 464 471 12925895 10.1007/s00335-002-2187-2
GEO: Gene Expression Omnibus
Manly KF Cudmore RH JrMeer JM Map Manager QTX, cross-platform software for genetic mapping. Mamm Genome 2001 12 930 932 11707780 10.1007/s00335-001-1016-3
Zeng ZB Kao CH Basten CJ Estimating the genetic architecture of quantitative traits. Genet Res 1999 74 279 289 10689805 10.1017/S0016672399004255
Basten CJWB Zeng Z-B QTL Cartographer Version 115 2001 Raleigh, NC: Department of Statistics, North Carolina State University
KEGG: Kyoto Encyclopedia of Genes and Genomes
Kanehisa M Goto S Kawashima S Nakaya A The KEGG databases at GenomeNet. Nucleic Acids Res 2002 30 42 46 11752249 10.1093/nar/30.1.42
Su AI Cooke MP Ching KA Hakak Y Walker JR Wiltshire T Orth AP Vega RG Sapinoso LM Moqrich A Large-scale analysis of the human and mouse transcriptomes. Proc Natl Acad Sci USA 2002 99 4465 4470 11904358 10.1073/pnas.012025199
SOURCE
Diehn M Sherlock G Binkley G Jin H Matese JC Hernandez-Boussard T Rees CA Cherry JM Botstein D Brown PO SOURCE: a unified genomic resource of functional annotations, ontologies, and gene expression data. Nucleic Acids Res 2003 31 219 223 12519986 10.1093/nar/gkg014
Damian D Gorfine M Statistical concerns about the GSEA procedure. Nat Genet 2004 36 663 author reply 663 15226741 10.1038/ng0704-663a
He YD Dai H Schadt EE Cavet G Edwards SW Stepaniants SB Duenwald S Kleinhanz R Jones AR Shoemaker DD Stoughton RB Microarray standard data set and figures of merit for comparing data processing methods and experiment designs. Bioinformatics 2003 19 956 965 12761058 10.1093/bioinformatics/btg126
Database for Annotation, Visualization, and Integrated Discovery (DAVID)
Dennis G JrSherman BT Hosack DA Yang J Gao W Lane HC Lempicki RA DAVID: Database for Annotation, Visualization, and Integrated Discovery. Genome Biol 2003 4 P3 12734009 10.1186/gb-2003-4-5-p3
R statistical software
| 15998448 | PMC1175990 | CC BY | 2021-01-04 16:05:40 | no | Genome Biol. 2005 Jul 1; 6(7):R59 | utf-8 | Genome Biol | 2,005 | 10.1186/gb-2005-6-7-r59 | oa_comm |
==== Front
Genome BiolGenome Biology1465-69061465-6914BioMed Central London gb-2005-6-7-r601599844910.1186/gb-2005-6-7-r60MethodAnalysis of the Macaca mulatta transcriptome and the sequence divergence between Macaca and human Magness Charles L 1Fellin P Campion 1Thomas Matthew J 2Korth Marcus J 2Agy Michael B 3Proll Sean C 2Fitzgibbon Matthew 2Scherer Christina A 1Miner Douglas G 1Katze Michael G 23Iadonato Shawn P [email protected] Illumigen Biosciences Inc., Suite 450, 2203 Airport Way South, Seattle, WA 98134, USA2 Department of Microbiology, University of Washington, Seattle, WA 98195-8070, USA3 Washington National Primate Research Center, University of Washington, Seattle, WA 98195-8070, USA2005 30 6 2005 6 7 R60 R60 18 1 2005 4 4 2005 23 5 2005 Copyright © 2005 Magness et al.; licensee BioMed Central Ltd.2005Magness 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.
Putative Macaca mulata orthologs for over 6,000 human genes have been sequenced from eleven tissues and three species of macaque. Macaque inter- and intraspecific nucleotide diversity is also reported.
We report the initial sequencing and comparative analysis of the Macaca mulatta transcriptome. Cloned sequences from 11 tissues, nine animals, and three species (M. mulatta, M. fascicularis, and M. nemestrina) were sampled, resulting in the generation of 48,642 sequence reads. These data represent an initial sampling of the putative rhesus orthologs for 6,216 human genes. Mean nucleotide diversity within M. mulatta and sequence divergence among M. fascicularis, M. nemestrina, and M. mulatta are also reported.
==== Body
Background
The sequencing of genes and genomes has become a hallmark of modern molecular biology. The resulting wealth of nucleotide sequence information has fostered advances in gene discovery, the development of genome-based technologies to study gene expression and function, and a growing interest in comparative genomics. The comparison of the human genome with the genomes of closely related species has particular appeal, and there is considerable interest in identifying genomic traits that set humans apart from other primate species [1-4]. The recent growth in sequence information for the chimpanzee has fueled this interest [4]. However, beyond that generated for chimpanzee, there has been remarkably little sequence information developed for other nonhuman primate species.
The rhesus macaque (Macaca mulatta) is a widely used small primate model of human disease, development, and behavior. Throughout the United States, National Institutes of Health (NIH)-supported facilities house more than 25,000 nonhuman primates, including more than 15,000 rhesus macaques [5]. Each year, approximately 13,000 nonhuman primates are used for NIH-funded research, 65% of which are rhesus [5]. These animals are used principally for infectious disease, pharmacology, and neuroscience research [6]. In particular, the rhesus model is an essential tool for acquired immunodeficiency syndrome (AIDS) research and for the development of new drugs and vaccines against human immunodeficiency virus (HIV) [7,8].
We report here on our initial efforts to sequence the rhesus macaque transcriptome. The close evolutionary relationship between rhesus and human, and its widespread use as a model for human reproduction, development, and disease, make it an ideal candidate for cDNA and genome sequencing. We have constructed cDNA libraries from a selection of diverse macaque tissues and multiple animals, and we have performed single-pass sequencing on 48,642 independent clones. This sequence information has been used to generate a rhesus macaque oligonucleotide microarray and to perform comparative analyses with human.
Results
Sequence data collection and preliminary analysis
We prepared cloned cDNA libraries from 11 M. mulatta tissues derived from nine separate animals. In addition, the liver was independently sampled from one animal each of the M. mulatta, M. nemestrina, and M. fascicularis species. cDNA libraries were prepared by directional lambda-based cloning into Escherichia coli and sequenced using standard fluorescent dye-terminator chemistry. Sequencing was performed from the vector-insert junction distal to the polyadenylate sequence.
A preliminary dataset of 48,642 independent clone sequences were collected as described in Table 1. We screened and analyzed these data as described in Materials and methods. Sequence data quality was assessed using the phred algorithm [9], with a mean of 539 high-quality base-pairs per read over the entire dataset. High-quality sequence bases are defined as those with a computed phred quality value of 20 or greater (Q ≥ 20) and an expected error rate of less than 1%. Of the cloned sequences, 9,219 contain a mammalian polyadenylation consensus sequence followed by a polyadenosine tail [10]. Data meeting minimum quality criteria (n = 36,921) have been submitted to GenBank and contribute to all subsequent analyses. Project data and associated information are also publicly available on the project website [11].
Table 1 Data-collection summary
Tissue Sequence reads
Placenta 12,033
Brain 10,511
PBMC 7,056
Spleen 6,658
Jejunum 3,840
Liver 3,744
Ileum 2,112
Lung 1,152
Ovary 672
Testis 480
Duodenum 384
M. mulatta 46,626
M. nemestrina 1,152
M. fascicularis 864
Total 48,642
We compared each macaque sequence to the mRNA RefSeq [12] component of GenBank using the MEGABLAST algorithm [13]. The most similar human sequence was identified as that reference sequence with the most significant match by bit score. In some cases, this method will identify matches between macaque and human sequences that are not orthologs, and so should be interpreted with caution. For all subsequent analyses, those macaque sequences with equally probable matches to more than one distinct human UniGene cluster have been excluded [14]. The entire dataset taken together provides a sampling of the putative macaque orthologs for 6,216 human genes (unique human LocusLink IDs), representing approximately 25% of the human gene content by recent estimate [15].
Although libraries were constructed from poly(dT)-primed cDNAs, the dataset includes a significant amount of coding sequence. Of the 6,216 unique human LocusLink IDs that were sampled in macaque, 69.3% include coding sequence (mean aligned coding length = 602 bp), whereas 30.7% include only 5' or 3' untranslated region (UTR) sequence (mean aligned UTR length = 485 bp). Of those 69.3% of genes with sampled coding sequence, the average extent of coding sequence coverage in the macaque database is 49.9% (data not shown).
Similarity of Macaca transcripts with human
We used the initial alignment information from the above data to define a subset of sequences whose alignment with their best human match extended 150 bp in each direction around a well defined stop codon. This dataset was used to compute the distribution of sequence similarity between macaque and human as represented by the histograms in Figure 1. The use of this constrained dataset permitted a direct comparison between the distributions for coding and noncoding sequence in the vicinity of the stop codon. Data for 1,180 macaque-human alignments are included in this analysis. Sequence-similarity distributions are not normal, with a modest tail toward lower values. The average degree of similarity for coding sequence is 97.79 ± 1.78% and 95.10 ± 4.15% for the 3' UTR. This analysis excludes data where the macaque stop codon was either mutated or in a different location relative to the human reference sequence. This analysis uses the 3' UTR proximal to the stop codon as a surrogate for all untranslated sequences. However, human-chimp comparative analysis suggests that the 5' UTR may be more divergent between species than other gene regions [16]. We did not have a sufficiently sized dataset to locate and independently test conservation of the 5' UTR.
Figure 1 Distribution of coding and noncoding sequence similarity between macaque and human. A histogram showing the degree of nucleotide sequence similarity between macaque and human for coding (blue) and noncoding (3' UTR, yellow) transcribed sequence. Sequences (n = 1,180) were selected that cross a well defined stop codon and that provide concurrent sampling of 150 bp of sequence both proximal and distal to the stop. The best human match for each macaque sequence was identified using MEGABLAST. The high-quality subset of these data (composed only of contiguous stretches of phred Q ≥ 20 bp, n = 633) is plotted for both coding (squares) and noncoding (diamonds) sequence.
In order to determine if local regions of poor data quality contribute to biases in the computed degree of sequence similarity, we recomputed the histogram using alignments composed of only high-quality (Q ≥ 20) sequence. Constraining the dataset to include only high-quality bases (n = 633 sequences) did not result in significant differences in either the shape or the mean of the distributions (Figure 1).
To provide a reference dataset with which to evaluate the current results, we computed the degree of sequence similarity between human and Pan troglodytes (chimpanzee) using the same method as above. This analysis was performed using chimpanzee expressed sequence tag (EST) and cDNA sequences, as most currently available chimpanzee reference sequences are computationally predicted and therefore lack data from the 3' UTR. However, our chimpanzee-human analysis was hampered by the relative paucity of chimpanzee full-length cDNA and EST sequence in the public databases. There are currently only 209 full-length chimpanzee cDNA sequences and 6,930 EST sequences of varying quality in GenBank.
These data together provide a sampling of the 150 bp proximal and distal to the stop codon for only 134 human genes. On the basis of this small dataset, the degree of nucleotide identity between human and chimpanzee for coding and 3' UTR sequences is 98.3 ± 3.0% and 97.65 ± 3.2% respectively (Additional data file 1). As expected, the distribution of sequence similarity is strongly biased toward larger values, with 59.0% of sampled chimpanzee coding sequences and 46.3% of 3' UTR sequences identical to their best human match over the 150-bp window. The distribution of sequence identity between human and chimpanzee is presented in Additional data file 2.
We expect that most observed nucleotide substitutions between macaque and human within coding sequence will be conservative. To evaluate the degree of similarity between human and macaque at the amino-acid level, we analyzed macaque sequences that overlapped with their best-matching human reference sequence by at least the terminal 450 bp proximal to the stop codon. Data from the terminal 450 bases were favored for this analysis in order to include more of the overall dataset and to be directly comparable to our previous nucleotide-based analysis. We also constrained the dataset to again include only high-quality bases. The distribution of amino-acid similarity was as expected, given the distribution of nucleotide similarity, with a bias toward higher values (Figure 2). The mean similarity between macaque and human protein sequences over the aligned window is 96.83 ± 4.95%. A relaxation of data quality constraints resulted in a broadening of the distribution toward lower values (data not shown).
Figure 2 Distribution of amino-acid sequence similarity between human and macaque. Sequencing reads containing the terminal 150 amino acids of each macaque gene were compared to their best human match using MEGABLAST. Only sequences composed of contiguous high-quality bases (phred Q ≥ 20 bp, n = 320) throughout the terminal 150 amino acids are included. Of these sequences, 5% show less than 88% nucleotide similarity to their best-matching human homolog.
We identified 21 high-quality macaque sequences with very weak amino-acid similarity (< 90%) to their best-matching human reference sequence (Table 2). Of these, 15 are either highly expressed in placenta or immune tissue (peripheral blood mononuclear cells (PBMCs) or spleen mononuclear lymphocytes) and/or are associated with pregnancy or the immune response. The observation of poor sequence identity for immune genes is not surprising, as increased divergence and evidence for positive selection have previously been reported for members of this group [17,18]. The most interesting example of divergence from our study is APOBEC3C, a member of the cytidine deaminase family. Rhesus macaque APOBEC3C is only approximately 85% identical to its putative human ortholog. Members of the APOBEC family are important mediators of lentivirus infection [19], and accelerated evolution has been reported for several members of this gene family [20].
Table 2 Macaque sequences showing weak identity with best human match
Gene Name RefSeq ID* Amino-acid identity (%)† Unigene ID* LocusLink/ Gene ID*
PSG11 Pregnancy specific beta-1-glycoprotein 11 NM_203287.1 68.04 Hs.502097 5680
PSG5 Pregnancy specific beta-1-glycoprotein 5 NM_002781.2 73.71 Hs.534030 5673
ANG Angiogenin, ribonuclease, RNase A family, 5 NM_001145.2 75.17 Hs.283749 283
PIP Prolactin-induced protein NM_002652.2 75.86 Hs.99949 5304
GNLY Granulysin NM_006433.2 76.55 Hs.105806 10578
LAIR2 Leukocyte-associated Ig-like receptor 2 NM_002288.3 80.13 Hs.43803 3904
CRYL1 Crystallin, lambda 1 NM_015974.1 80.31 Hs.370703 51084
ARP10 ARP10 protein NM_181773.2 82.58 Hs.440515 164668
LOC151174 Hypothetical protein LOC151174 XM_371605.1 83.04 Hs.424165 151174
GH2 Growth hormone 2 NM_022558.2 84.56 Hs.406754 2689
APOBEC3C Apolipoprotein B mRNA editing enzyme, catalytic polypeptide-like 3C NM_014508.2 85.26 Hs.441124 27350
NDUFC2 NADH dehydrogenase (ubiquinone) 1, subcomplex unknown, 2 NM_004549.3 85.71 Hs.407860 4718
SAA4 Serum amyloid A4 NM_006512.1 85.94 Hs.512677 6291
SEPP1 Selenoprotein P, plasma, 1 NM_005410.1 86.07 Hs.275775 6414
GZMB Granzyme B (cytotoxic T-lymphocyte-associated serine esterase 1) NM_004131.3 86.64 Hs.1051 3002
IFITM1 Interferon induced transmembrane protein 1 NM_003641.2 87.2 Hs.458414 8519
GH1 Growth hormone 1 NM_000515.3 87.56 Hs.500468 2688
TMEM14B Transmembrane protein 14B NM_030969.2 87.72 Hs.273077 81853
PRG2 Proteoglycan 2 NM_002728.4 88.35 Hs.512633 5553
MRPL40 Mitochondrial ribosomal protein L40 NM_003776.2 88.94 Hs.431307 64976
GKN1 Gastrokine 1 NM_019617.2 89.07 Hs.69319 56287
*GenBank identifiers for best matching human homolog. †Amino-acid sequence identity between macaque and human.
We also identified ten placentally expressed pregnancy-related transcripts with very weak similarity to their putative human ortholog. Prominent among these are the pregnancy-specific glycoproteins (PSG5 and PSG11). For example, the best macaque match to human PSG11 shows only 68% identity and is not better matched to any other member of the human PSG family. Other placentally expressed weak orthologs include the growth mediators angiogenin (ANG) and growth hormone 1 and 2 (GH1 and GH2). Episodic accelerated evolution has previously been reported for both angiogenin and the growth hormones, although its biological and developmental implications are not well understood [21,22].
We compiled amino-acid similarity data into gene functional groupings using the 'biological process' classifications from the Gene Ontology (GO) Consortium [23] (Table 3). Data are shown for only those classes containing three or more entries. The data reveal a wide degree of variation in class-specific values of sequence similarity between human and macaque. Highly conserved classes include those involved in intracellular signaling, small GTPase-mediated signal transduction, translation initiation, and protein biosynthesis and folding. Poorly conserved biological process groups include pregnancy and immune and inflammatory response. We note that the small size of the dataset is reflected in large standard deviations for several classes of genes.
Table 3 Mean amino-acid identity by GO ontology
Biological process group Mean identity (%)* Standard deviation
Pregnancy 80.8 11.7
Cell proliferation 92.7 3
Immune response 93.9 4.1
Negative regulation of cell proliferation 94 6.2
Regulation of cell cycle 94.3 5.9
Response to oxidative stress 94.3 6.7
Inflammatory response 94.4 4.3
Transport 95 3.1
Cell-cell signaling 95.5 3.6
Apoptosis 95.6 5.3
Proteolysis and peptidolysis 96.1 4.5
Positive regulation of cell proliferation 96.2 2.4
G-protein coupled receptor protein signaling pathway 96.3 5.5
Electron transport 96.3 2.3
Development 96.4 4.1
Carbohydrate metabolism 96.7 3.2
Metabolism 96.9 2.4
Signal transduction 97 4
Cell growth and/or maintenance 97.2 3.8
Angiogenesis 97.3 2
Regulation of transcription from Pol II promoter 97.7 2.4
Mitosis 97.7 1.8
Ubiquitin cycle 97.7 3.8
Antimicrobial humoral response (sensu Vertebrata) 97.8 1.9
Ribosome biogenesis 98 1.9
Ion transport 98.1 0.6
Cell adhesion 98.3 2.9
Anti-apoptosis 98.5 1.7
Ubiquitin-dependent protein catabolism 98.7 1.3
Regulation of transcription, DNA-dependent 98.7 1.4
Protein folding 98.8 1.5
Translational initiation 99 1.7
Protein biosynthesis 99.1 1.7
Response to stress 99.4 0.5
Intracellular protein transport 99.4 0.7
Glycolysis 99.6 0.3
Nuclear mrna splicing, via spliceosome 99.6 0.3
Small gtpase mediated signal transduction 99.7 0.5
Protein transport 99.9 0.3
Intracellular signaling cascade 100 0
*Mean identity between group members and their best matching human homologs.
These data share similarity with recent comparative analyses between human and chimpanzee [4,24]. For example in chimpanzee, a high degree of sequence conservation and low rates of nonsynonymous substitution were found for several biological classes, including protein transport, small GTPase-mediated signal transduction, regulation of DNA-dependent transcription, intracellular signaling, and glycolysis. However, not all biological functional groups demonstrate consistent conservation among the three species. For example, the signal transduction biological class is highly conserved between chimpanzee and human, whereas its conservation between macaque and human does not significantly deviate from the mean over all classes.
Sequence divergence within and among macaque species
Our dataset includes sequence data from nine M. mulatta, one M. fascicularis, and one M. nemestrina. The breadth of the dataset provides an opportunity to conduct a preliminary analysis of the polymorphism frequency within M. mulatta and the degree of nucleotide divergence between macaque species. We estimated the polymorphism frequency within M. mulatta by assembling sequencing reads from multiple animals for the same gene using phrap [9]. Polymorphisms were identified by a modified version of phred that calls two alleles at each base in the assembly and assigns each allele a quality score based on combined phred quality values (C.M., unpublished work). High-scoring polymorphisms were manually verified and are presented in Table 4 for a sample of 24 genes. This analysis includes both coding and noncoding transcribed sequences. The average nucleotide diversity (π) for this gene set in M. mulatta is 15.8 ± 12.5 × 10-4 [25]. A large standard deviation in nucleotide diversity across genes is consistent with reports from other primate species [26-28]. The animals included in this analysis were primarily bred from wild-caught parents of Indian origin. A more comprehensive determination of nucleotide diversity will require sequence data from a greater number of genes and animals from multiple geographic locations.
Table 4 Estimate of Macaca mulatta nucleotide diversity
Gene Comparative length Number of animals Nucleotide diversity (π)
ACTB 1,067 5 0.00110
ACTG1 708 6 0.00290
APOA1 746 2 0.00200
APOA2 431 2 0.00350
ATF4 469 4 0.00160
B2M 439 7 0.00000
C15orf15 860 3 0.00117
CAP1 667 5 0.00127
CCNI 547 4 0.00000
CDC10 693 3 0.00190
CTSB 967 4 0.00078
EEF1A1 865 7 0.00235
EEF1G 771 6 0.00000
ENO1 793 5 0.00100
FTH1 775 5 0.00100
PPID 891 3 0.00148
RPL14 657 5 0.00457
RPL15 631 4 0.00264
RPL3 796 6 0.00339
RPS20 483 4 0.00155
SLC25A5 749 3 0.00088
TPT1 740 7 0.00000
TXNIP 614 3 0.00000
UBC 824 6 0.00281
Mean 0.00158
SD 0.00125
We were also able to evaluate the degree of nucleotide sequence divergence between the three macaque species for a sample of 21 genes in this dataset (Table 5). Phred and phrap were again used to assemble overlapping sequences from multiple species and to identify species-specific variants that were then manually confirmed. Given the high degree of nucleotide similarity among the species and the small sample size, the three species did not differ beyond the measured standard deviations. However, M. mulatta and M. fascicularis appear more closely related to each other than either is to M. nemestrina, with an average sequence divergence between the two of 0.380 ± 0.380%. The degree of sequence divergence between M. mulatta and M. nemestrina is 0.588 ± 0.438% and 0.522 ± 0.419% between M. fascicularis and M. nemestrina. However, the dataset is not large enough for any of these pairwise differences to reach statistical significance.
Table 5 Interspecies substitution rates
Gene Alignment length Number of reads Frequency per kilobase
M f. M.m. M.n. m vs n* m vs f* n vs f*
ADH1B 819 8 5 14 0.00 0.00 2.44
AFP 537 1 3 1 11.17 7.45 3.72
ALB 2047 > 20 > 20 > 20 0.00 0.49 0.49
AMBP 731 1 5 4 4.10 1.37 5.47
ANGPTL3 371 1 1 2 2.70 0.00 2.70
APOA1 746 10 76 2 10.72 4.02 5.36
APOA2 431 3 4 5 6.96 4.64 2.32
APOC4 312 3 1 1 16.03 12.82 16.03
APOE 217 2 2 2 4.61 4.61 0.00
APOH 1007 4 14 14 2.98 0.00 2.98
B2M 587 1 90 1 0.00 11.93 11.93
EEF1A1 920 7 >20 7 0.00 0.00 0.00
FGA 379 3 6 1 7.92 0.00 7.92
FGB 407 3 43 11 2.46 0.00 2.46
FGG 694 3 24 9 1.44 1.44 2.88
HPR 567 2 20 12 3.53 1.76 1.76
RPL9 680 1 35 1 4.41 4.41 0.00
SERPINC1 787 1 3 1 1.27 2.54 1.27
TTR 599 5 2 6 5.01 3.34 5.01
UBC 460 1 40 1 0.00 0.00 0.00
UGT2B7 228 1 13 1 8.77 0.00 8.77
Mean 5.88 3.80 5.22
Median 3.53 1.44 2.70
SD 4.38 3.80 4.19
*Pair wise interspecies substitution frequencies computed on a gene-by-gene basis M.f., Macaca fascicularis; M.m., M. mulatta; M.n., M. nemestrina.
Putative rhesus sequences without human orthologs
Analysis of the entire dataset revealed a small number of transcribed macaque sequences that had little or no sequence similarity to any human cDNA or genomic sequence (Table 6). We speculate that some of these macaque sequences are without orthologs in the human genome. The observation of species-specific transcribed sequences among the primates is consistent with recent comparative analysis between human and chimpanzee [4,29]. Although an absolute determination of species specificity will require a complete macaque genome sequence, we conducted preliminary computational and PCR-based analyses to test the presence or absence of these sequences in the human and other primate genomes.
Table 6 Macaque sequences without apparent human ortholog
Class GenBank Accession Ortholog by MEGABLAST* PCR product length† PCR‡
Human genome Human EST Macaque genome Human genome
I CX078602 Yes/93%§ No 98 + -
I CX078592 No No 111 + +
I CX078596 Yes/93%§ No 123 + +
I CB552301 No No 107 Indeterminate Indeterminate
II CX078598 No No 103 + -
II CX078591 No No 111 + -
III CB555845 No No 127 + Indeterminate
III CB552531 No No 90 + -
*Defined as identity greater than three standard deviations below the mean. †Primer sequences are available in Materials and methods. ‡Tested under a variety of thermal cycling conditions and annealing temperatures. §Borderline identity values are displayed.
As above, we used MEGABLAST to test each macaque nucleotide sequence for one or more significant hits to the human EST or genome databases. The absence of an orthologous human sequence was defined as either no significant MEGABLAST hit in the human subset of GenBank or hits with sequence identity less than three standard deviations below the mean as measured over the entire dataset (Figure 1). Because the data were not normally distributed, the identity cutoff (approximately 92.2%) was computed using the geometric mean, which relies on a logarithmic transformation of the data. All sequences meeting this cutoff definition were also outliers based on Tukey's test [30].
We selected eight of the resulting macaque sequences for PCR-based analysis using a number of primate and human genomes (Table 6, Figure 2). The purpose of this analysis was simply to verify the presence or absence of the observed sequences in a panel of primate genomes. Selected primers had an average computed annealing temperature of 59.6 ± 0.9°C with an average amplified length of 108 ± 12 bp (Materials and methods). For each primer pair, PCR analysis was conducted at several annealing temperatures between 55 and 60°C. Genomic DNA was selected from independent M. nemestrina and M. mulatta animals in order to confirm the presence of these sequences in multiple independent genomes. Of the eight tested primer pairs, two resulted in amplification of consistent bands in both human and macaque genomic DNA, two were indeterminate in human but present in the macaques, and four, while obviously present in the macaque genomes, resulted in no consistent human-specific product under any cycling conditions.
The eight tested sequences fall generally into three categories: those with weak sequence similarity to the human genome or human-derived ESTs (class I), those with weak sequence similarity only to genes and proteins from nonhuman species (class II), and those with no significant amino-acid or nucleotide sequence similarity to any GenBank nucleic acid or protein sequence (class III).
Those with weak similarity to human sequences (class I) include CX078602, a 657-bp cDNA sequence derived from macaque liver with 79-87% nucleotide sequence identity to CYP2C18 from several mammalian species. Its closest matches to human are two regions of 86-93% identity to human chromosome 10, one of which contains four cytochrome P450 2C genes. PCR-based analysis failed to amplify a consistent band from any primate species other than M. nemestrina, M. mulatta, and Lagothrix lagotricha (woolly monkey) (Figure 3a).
Figure 3 PCR analysis of putative macaque-specific sequences. PCR primers were developed from high-quality macaque cDNA sequences - (a) CX078602, (b) CX078598, and (c) CB555845 - and used to test for the presence or absence of the resulting amplicons in genomic DNA from 12 primate genomes, including two separate humans. Amplification conditions were the same as in Materials and methods, except that annealing was performed at 55°C. Expected product sizes are as in Table 6. (d) Amplification primers from exon 4 of the human oligoadenylate synthetase 1 gene (OAS1 ) are included as a positive control, resulting in the expected 648-bp product from most primate species.
Likewise, CX078592 from brain demonstrated 88-90% nucleotide similarity to the IL15RA gene and other immune-derived transcripts, as well as to a region of human chromosome 10 containing IL15RA. PCR primers derived from this sequence amplified multiple specific products from macaque, human, and other primates (data not shown). Similarly, CX078596 from placenta, although having no significant match to any human EST, demonstrated significant similarity to a region of human chromosome 22. CX078596 contained a clear mammalian polyadenylation signal and poly(A) tail, and primers derived from this sequence amplified an appropriately sized product from macaque. Alignment of this sequence with human chromosome 22 revealed a 284-bp insertion in human relative to macaque, which was reflected by amplification of a proportionately larger product in two human genomic DNA samples (data not shown). Finally, although CB552301 from spleen demonstrated significant sequence identity to regions of human chromosomes 4 and 15 and multiple ESTs from UniGene cluster Hs.459311, we failed to amplify a specific product from any primate species using primers derived from this sequence (data not shown).
The second class of sequences (class II) in Table 6 had no identified human match, while demonstrating weak sequence identity to nucleic acid or protein sequences from other species. For example, CX078598, a 670-bp transcript from PBMCs, demonstrated weak amino-acid identity (67%) to the endogenous retrovirus (ERV)-BabFcenv envelop polyprotein, a member of the ERV-F/H family of primate retroviruses [31]. PCR with primers derived from CX078598 under a variety of thermal cycling conditions resulted in the consistent amplification of a product of expected size from only M. mulatta and M. nemestrina (Figure 2b). Similarly, CX078591 from macaque brain demonstrated weak amino-acid identity (20-45%) to ariadne homolog 2 (ARIH2/TRIAD1) from rodents and to two unnamed proteins from the puffer fish Tetraodon nigroviridis. Primers derived from this sequence amplified the appropriately sized product only from macaque genomic DNA (data not shown).
The last class of sequences (class III) in Table 6 demonstrated no significant similarity to any protein or nucleotide sequence in GenBank (represented by CB555845 and CB552531). Both showed evidence of a mammalian polyadenylation consensus sequence near their 3' terminus, with CB552531 additionally demonstrating a clear poly(A) tail. CB555845, a 485-bp sequence from spleen, amplified expected products from both M. nemestrina and M. mulatta. However, this clone was ultimately scored as indeterminate because of its consistently weak amplification of a discrete product from all hominids including human (Figure 2c). CB552531 amplified products of the expected size from macaque species and from Ateles geoffroyi and Lemur catta, but not from human (data not shown).
It is important to note that PCR-based analysis of divergent sequences is subject to a variety of influences and may result in different conclusions under different conditions. Furthermore, we cannot rule out the possibility that one or more of the sequences in Table 6 are alternatively spliced relative to human, pseudogenes, or genomic DNA contamination. However, each clone sequence in Table 6 demonstrated similarity to known expressed sequences or a polyadenylation consensus sequence and poly(A) tail at their 3' terminus upon complete sequencing of the clones.
Development of a macaque-specific expression microarray resource
Genome-based technologies such as DNA microarrays are now commonplace in human biomedical research. Similarly, species-specific arrays exist for model organisms such as the mouse and rat, for which a considerable amount of genome information is available. In contrast, researchers wishing to carry out gene-expression analyses on nonhuman primate cells or tissues are currently forced to use human DNA microarrays. As part of our effort to bring genome-based technologies to researchers using nonhuman primates, we have used ESTs generated by this project to construct a rhesus macaque-specific oligonucleotide microarray.
Oligonucleotides were designed as described in Materials and methods and arrayed onto glass slides by Agilent Technologies. Briefly, macaque cDNA sequences were assembled into 9,344 distinct clusters using The Institute for Genome Research (TIGR) clustering tools [32]. From these, 7,973 macaque-specific oligonucleotide probes were identified for inclusion on the array. These probes represent the putative macaque equivalent of 3,519 unique human UniGene clusters [14] and 3,045 unique human RefSeqs [12]. To quality control the microarray, we measured tissue-specific differences in gene expression as a means of evaluating whether the oligonucleotides were successfully binding target sequences. For these experiments, we hybridized the microarray with probes derived from RNA isolated from various rhesus macaque tissues. Probes were paired in different combinations and two dye-flipped technical replicates were performed for each pair of samples. Of the 7,973 rhesus macaque oligonucleotides present on the microarray, 6,215 showed differential expression (equal or greater than twofold; P ≤ 0.01) in at least one of the three experiments.
Plots of the log-transformed ratios for genes in each experiment that showed an equal to or greater than twofold difference in expression between two tissues are shown in Figure 4. In each plot, points are colored according to the source library of the sequence used to derive the corresponding oligonucleotide. From this analysis, it is apparent that the majority of genes that were more highly expressed in the spleen correspond to sequences derived from the spleen cDNA library. Similarly, the majority of genes that were more highly expressed in the brain correspond to sequences derived from the brain cDNA library. These results show that a majority of the oligonucleotides were successfully binding target sequence. In addition, it is likely that many of the oligonucleotides that did not measure differential gene expression in these experiments are also successfully binding target sequences, as not all genes would be expected to be expressed in all tissues or to show differential levels of expression between the tissues analyzed.
Figure 4 Tissue-specific gene-expression patterns measured using the rhesus macaque oligonucleotide microarray. To evaluate whether arrayed oligonucleotides were binding target sequences, microarrays were hybridized with probes derived from RNA isolated from spleen, brain, or placenta. Probes were paired in different combinations as indicated: (a) spleen vs brain; (b) brain vs placenta; (c) spleen vs placenta. Oligonucleotides that detected a difference in gene expression of twofold or more (P ≤ 0.01) between two tissues are indicated as colored points and are displayed across the x-axis to facilitate visualization. The y-axis represents the log ratio for differentially expressed genes in each tissue comparison. Points corresponding to oligonucleotides derived from spleen ESTs are colored blue, those from brain are magenta, and those from placenta are orange. Thus, in the comparison depicted in (a), genes more highly expressed in the spleen are indicated by points in the upper portion of the panel (and are predominantly sequences derived from the spleen cDNA library) and genes more highly expressed in the brain are indicated by points in the lower portion on the panel (and are predominantly sequences derived from the brain cDNA library). Plots were prepared using Spotfire.
Discussion
Primate models are essential to the study of human biology and disease and to the development of new pharmaceutical products, many of which require primate testing before approval for use in humans. The closest living primate relatives to human are the chimpanzee and other great apes [33]. Human and chimp lineages diverged from a common ancestor 5-7 million years ago (Mya) and the genomes of the two species are highly conserved [4,24,34-36]. Experimental research using chimpanzees and other great apes is, however, significantly hampered by their size, maintenance costs, and endangered species status. The human-like qualities of the chimpanzee also make research using this animal generally unacceptable for ethical reasons. For the most part, chimpanzees are rarely used for invasive studies except, for example, when investigating diseases for which there is no other animal model (for example, hepatitis C infection) [37].
Old World monkeys, a group that includes macaque, baboon, and African green monkey, are our closest non-ape relatives. Old World monkeys and humans shared a common ancestor around 25 Mya, and the genomes of these organisms are highly conserved with human [33,35,38]. Furthermore, the biology of these organisms is such that they are an appropriate primate model for human physiology and disease. For this and other reasons, Old World monkeys are widely used in biomedical research, with members of the Macaca genus most frequently used [6].
We report here on the first phase of a study to sequence the rhesus macaque transcriptome. Our group has collected sequence data from 48,642 cDNA clones from nine animals and 11 tissues. For the current study, standard cDNA sequencing methods were used, with an emphasis on large clone-inserts and long sequence read lengths. Alternative methods could have been used for data collection that would have resulted in less 3'-end bias (for example, ORESTES [39]) or reduced redundancy in the collected data (for example, library normalization [40]).
We determined the average sequence divergence between human and macaque to be 2.21% for coding and 4.90% for noncoding sequence. An identical analysis of transcribed chimpanzee sequences demonstrated divergences of 1.70% and 2.35% for coding and noncoding sequence respectively. This is in comparison to a recently reported mean 1.44% divergence between human chromosome 21 and chimpanzee chromosome 22 over their entire length [4]. The continued analysis of sequence divergence between the macaque and human species will be important for translating data collected in this primate model to human biology. Recent evidence suggests that even minor inter-species sequence variation can result in large phenotypic differences between macaque models and human disease [8,41,42].
In addition, we have identified gene functional groups with higher than average sequence divergence at the amino-acid level. In one example, we observe 15% amino-acid sequence divergence between putative human and macaque orthologs of the cytidine deaminase APOBEC3C. Consistent with this observation, Sawyer et al. have reported evidence for accelerated evolution of the primate APOBEC gene family, probably under the selective pressure of viruses [20]. Members of this family (for example, APOBEC3G) have antiviral activity against lentiviruses and specifically against HIV [19]. APOBEC3G is packaged into nascent virions and delivered together with the viral genome into newly infected host cells. The cytidine deaminase cargo results in hypermutation of the replicating virus in target cells, thereby inhibiting virus infection. The Vif proteins of HIV and other lentiviruses bind APOBEC3G and inhibit its antiviral activity. However, the interaction between Vif and APOBEC3G is highly species and virus specific. HIV Vif can inhibit the function of human but not simian APOBEC3G [42]. Likewise, Yu and colleagues have recently reported that human APOBEC3B and APOBEC3C can inhibit SIV but not HIV-1 infection of human cells [43]. Our observation of poor sequence conservation between macaque and human APOBEC3C is consistent with a model of accelerated evolution under selective pressure for this gene family.
This dataset has further enabled us to conduct a preliminary analysis of nucleotide diversity within the M. mulatta species and the degree of divergence among M. nemestrina, M. fascicularis, and M. mulatta. Mean nucleotide divergence computed over 24 genes is 15.8 ± 12.5 × 10-4, approximately twofold greater than that computed for human transcribed sequences by several recent comprehensive studies [26,44]. Excess nucleotide diversity in macaque versus human is consistent with observations from other primate species. In general, numerous groups have observed increased nucleotide diversity in mitochondrial [45-47], sex chromosome [48-51], and autosomal DNA [28,52,53] sequences from chimpanzee, bonobo, and gorilla. Consistent with other primate species, this observation is likely to reflect a larger effective population size for macaque throughout evolution relative to human. Our analysis also confirms a high degree of sequence similarity among macaque species, with pairwise divergence estimates (0.380-0.588%) exceeding intraspecies heterozygosity. M. mulatta and M. fascicularis appear more closely related to each other than either one is to M. nemestrina, although these differences did not reach significance.
We describe a small number of macaque sequences without apparent human ortholog. Confirmation of this observation will require a complete sequence of the rhesus genome, but these preliminary data are consistent with recent comparative human-chimpanzee analyses demonstrating many small insertions/deletions and rearrangements between these species, some of which contain open reading frames or expressed sequences [4,29,54].
Finally, we report on the development of a first-generation rhesus-specific oligonucleotide microarray to support gene expression analyses of cells and tissues from this animal. Previously, investigators have used human DNA microarrays to measure gene expression changes in macaque tissues. Although the high degree of nucleotide sequence identity between humans and macaques makes this cross-species hybridization feasible, it is not clear to what extent sequence divergence between these species may affect gene expression measurements. Our observation of a small number of macaque sequences without apparent human ortholog also suggests the importance of using species-specific arrays. The rhesus microarray should therefore facilitate the use of the macaque model for future gene expression profiling experiments and may also be useful for studying similarities and differences in gene expression between macaque and human tissues [55]. To this end, we have included on the microarray 1,014 human oligonucleotide sequences, many of which were chosen because they are orthologs of macaque sequences also present on the array. In addition, because we anticipate this array will be widely used for infectious disease research, many of the human sequences have relevance to cytokine signaling, apoptosis, or the immune response, and we have included oligonucleotides corresponding to genes from 20 different viruses.
Conclusion
While the macaque species are widely used primate models of human physiology and disease, there are few species-specific genomic resources available to the research community. Furthermore, the applicability of the macaque model to human disease will be highly dependent on the degree of sequence divergence between macaque and human, among the macaque species, and among animals of divergent geographic origin. Comprehensive genome-wide analysis has begun to characterize inter-species differences and to provide resources, such as the rhesus-specific microarray, that will enable a more efficient use of this model organism in the future.
Materials and methods
Animal tissues
Animal tissues and blood were provided by the Tissue Distribution Programs of the Washington and Oregon National Primate Research Centers. All M. mulatta animals were of Indian origin and had wild-caught parents. No mitochondrial DNA or major histocompatibility complex (MHC) typing was performed.
RNA isolation
Rhesus macaque tissues used for RNA isolation were harvested at necropsy and immediately placed in RNAlater stabilization and storage solution (Ambion). Tissues were then homogenized in Solution D [56] either by hand or mechanically using a Polytron tissue homogenizer, and total RNA was isolated by guanidinium isothiocyanate-phenol-chloroform extraction and further purified using RNeasy purification columns (Qiagen). Extraction of mRNA was performed using the FastTrack 2.0 mRNA extraction kit (Invitrogen). RNA quality and quantity were determined by spectrophotometry and by capillary electrophoresis using an Agilent Technologies BioAnalyzer.
cDNA library construction and sequencing
cDNA libraries were constructed using two alternative methods. Spleen mononuclear lymphocyte, brain, lung, activated PBMC, and two placental libraries (from male and female fetuses) were constructed with the Uni-ZAP cDNA library construction kit (Stratagene) using 3-5 μg high-quality mRNA for each library. Clones were isolated by ampicillin resistance and grown in 96-well plates containing LB-ampicillin medium. Liver, duodenum, ileum, jejunum, testes, ovary, and activated PBMC libraries were constructed with the CloneMiner cDNA construction kit (Invitrogen), again using 3-5 μg high-quality mRNA for each library. Clones were isolated by kanamycin resistance and grown in 96-well plates containing LB-kanamycin medium. All libraries were constructed with size-fractionated RNA, resulting in a mean insert size of approximately 1.5 kbp for each library as determined by PCR. Clone inserts were sequenced from the vector-insert junction distal to the poly(A) tail such that most resulting sequences do not include a poly(A) tail at their 3' terminus. For each clone, inserts were amplified by PCR directly from 0.2 μl of the glycerol stock using the following primers: for the Stratagene pBluescript SK (+/-) vector: 5' -CCCTCACTAAAGGGAACAAAA (the sequencing primer) and 5' -CACTATAGGGCGAATTGGGTA; for the Invitrogen pDONR222 vector: 5' -GACGTTGTAAAACGACGGC (the sequencing primer) and 5' -GCCAGGAAACAGCTATGACC. PCR products were sequenced using standard fluorescent dye-terminator chemistries on an Applied Biosystems 3700 capillary sequencer.
Sequence data analysis
cDNA sequences were first base-called using a modified version of the phred algorithm ([9,57] and C.M., unpublished data) and then screened for cloning vector, lambda-phage, and E. coli contamination using the program cross_match [9]. Sequences exhibiting multiple cloning sites or any contamination with lambda-phage or E. coli sequence were removed from further analysis. Leading- and trailing-cloning vector sequence was masked from all remaining sequences. Putative polyadenylation was identified by the presence of a consensus mammalian polyadenylation signal [10] followed by an (A)10 tract within 50 bp. The remaining sequences were analyzed using MEGABLAST [13] against rhesus mitochondrial sequence (GenBank accession AY612638.1) and against the human mRNA RefSeq collection [12] to identify putative human orthologs. Sequences with a significant match to any putative human ortholog were selected for GenBank submission. Of these sequences, 36,921 met minimum quality criteria for submission to GenBank. Each sequence was assigned a putative human ortholog if there was a unique maximally-scoring match by MEGABLAST bit score comparison; sequences with multiple maximal matches were not assigned an ortholog. Sequences with no significant RefSeq match were further analyzed by similar MEGABLAST comparisons against EST, genomic, and protein databases. Sequences were also analyzed for human repetitive sequence families using cross_match.
Rhesus-human similarity analyses
For the nucleotide comparisons, macaque sequences were selected for inclusion that were assigned a putative human ortholog and that spanned the human ortholog's final coding nucleotide by at least 150 nucleotides in each direction (as determined by initial MEGABLAST results). Selected sequences were then realigned independently against two subsequences of the corresponding human ortholog: one containing the final 150 coding nucleotides and the other containing the first 150 nucleotides in the 3' UTR. Results across all sequences were grouped by ortholog and the maximal bit score match in each region selected. For the amino-acid comparisons, macaque sequences were selected that had been assigned a putative human ortholog and contained at least 450 high-quality bases spanning the 3' end of the putative coding region (as determined by initial MEGABLAST results). Selected sequences were realigned independently (by translated MEGABLAST) against the protein sequence corresponding to the assigned ortholog. Results from all sequences were grouped by ortholog and by the maximal bit-score match selected. Grouping by protein classes was completed by cross-reference of each orthology against its GO biological process assignment.
Genomic PCR analysis of macaque sequences
PCR reactions (10 μl) included 0.132 U of Platinum Taq polymerase (PerkinElmer), 0.5 μM each primer, 0.132 μM each dNTP, 13 mM Tris-HCl (pH 8.4), 33 mM KCl, and 1 mM MgCl2. Thermal cycling was conducted in a PerkinElmer 9700 as follows: 95°C for 5 min (one cycle); 95°C for 30 sec, 55°C for 30 sec, 72°C for 1 min (40 cycles); and 72°C for 1 min (one cycle). Amplifications were evaluated under a variety of annealing temperatures between 55-60°C. Primer sequences are as follows (target/forward/reverse):
CX078591: 5'-GGAGAATCCAGTTAACGGCT-3', 5'-CTCTCATCCAGCCTAACGTG-3';
CX078602: 5'-GTTTTCAAAGAGCCCAGCAA-3', 5'-CTTTGGCATAGCTTCGGTTC-3';
CX078598: 5'-GGCAACAAGTGGGAATCAAC-3', 5'-GAGGAATCGGGATGGTCATA-3';
CB552301: 5'-CCTCCTTGGACTTGGACCTT-3', 5'-AGGACAGGAGTCTTGCCAAA-3';
CB555845: 5'-GTCAACAGGCTGGCATTTTA-3', 5'-CAATTATTGACCCCAAGGCTA-3';
CX078592: 5'-CAAAGCCATCAGACAGCAGA-3', 5'-GAGACCAGGAAAGTCGAAGG-3';
CB552531: 5'-CTGGAATAAGGCCAGAAGCA-3', 5'-ATTCCTCAGGTCTGGTGGAG-3';
CX078596: 5'-CCTCATGGTGTGGCTATGTG-3', 5'-ACACAAGGCGAGCTCTGGTA-3';
OAS1: 5'-GAGCCAAGAAGTACAGATGC-3', 5'-AGGACAGAGCTGTCCAATAG-3'.
Oligonucleotide microarray design
To design sequences for a rhesus macaque oligonucleotide microarray, we began with over 20,000 EST reads from clones derived from six cDNA libraries (spleen mononuclear lymphocyte, brain, lung, PBMC, and male and female placenta). After base-calling and quality filtering, sequences were processed using TIGR clustering tools [32] and compared by BLASTN with Human UniGene cluster representatives (build 167). High-quality reads that had at least one strong hit to Human UniGene were carried forward for oligonucleotide design. An additional 584 rhesus macaque sequences were provided by Robert Norgren (University of Nebraska) and Eliot Spindel (Oregon National Primate Research Center).
Oligonucleotides based on these sequences were designed by Agilent Technologies. Repeat sequences were identified, masked, and excluded. Candidate oligonucleotides were selected from the 3' end of each target sequence, filtered according to optimal base-composition profiles, and screened on the basis of predicted hybridization properties and potential cross-hybridization with other sequences. Four 60-mer oligonucleotides were initially designed for each target sequence that passed quality-control checks. To estimate specificity, each oligonucleotide was compared with Human UniGene build 167. Oligonucleotides with strong similarity to more than one UniGene cluster were then manually checked for cross-hybridization against the July 2003 assembly of the human genome (hg16) using the University of California Santa Cruz Genome Browser [58]. Oligonucleotides that hit more than one region of the human genome were discarded as ambiguous.
Because target sequences were not filtered by annotation before oligonucleotide design, multiple oligonucleotides were often designed to different regions of the same gene. Oligonucleotides were therefore mapped to UniGene cluster sequences and two high-scoring oligonucleotides were selected for each underlying transcript represented. This resulted in a final set of 7,973 macaque oligonucleotides representing approximately 3,944 unique genes. In addition to the macaque oligonucleotides, 1,014 oligonucleotides corresponding to 894 human genes and 96 oligonucleotides corresponding to genes from 20 different viruses were also selected for inclusion on the microarray. Duplicate 60-mer oligonucleotides were arrayed onto glass slides by Agilent Technologies. The array is commercially available from Agilent Technologies, Agilent Microarray Design Identification (AMADID) Number 012650 [59].
Labeled probe synthesis and microarray hybridization
For microarray analysis, total RNA was extracted from spleen, brain, and placental tissues. Each tissue was obtained from a different animal. RNA quality and quantity were determined by spectrophotometry and by capillary electrophoresis using an Agilent Technologies BioAnalyzer. Labeled cRNA probes were generated using the Low Input RNA Probe Synthesis Kit (Agilent Technologies) according to the manufacturer's protocol for 11K postage-stamp oligonucleotide microarrays. The probes were hybridized in replicate to the rhesus macaque oligonucleotide microarray according to the manufacturer's protocol. Slides were scanned with an Agilent DNA microarray scannerand image analysis was performed using Agilent feature extraction software. All data were entered into a custom-designed gene-expression database, Expression Array Manager, and then uploaded into Resolver 4.0 (Rosetta Biosoftware) and DecisionSite for Functional Genomics (Spotfire) for analysis.
Data submission and databases
There are 36,921 GenBank accession numbers associated with this manuscript. They are cross-referenced and publicly available at the project website [11]. Expression microarray data have been submitted to EBI ArrayExpress, accession number E-TABM-9.
Additional data files
The following additional data are available with the online version of this paper. Additional data file 1 is a table of chimpanzee versus human sequence identity. Additional data file 2 is a figure showing the distribution of coding and noncoding sequence similarity between chimpanzee and human.
Supplementary Material
Additional File 1
Chimpanzee versus human sequence identity. Chimpanzee cDNA and EST sequences in GenBank were compared to human RefSeqs to provide a dataset comparable to the macaque-human results described herein. Due to the paucity of expressed chimpanzee sequence data in GenBank, the analysis could only be performed for 134 human-chimp pairs. Data for sequence identity between chimpanzee and human for these genes are noted in the table. Macaque-human identities are also reported for the same genes where available.
Click here for file
Additional File 2
Distribution of coding and noncoding sequence similarity between chimpanzee and human. A histogram showing the degree of nucleotide sequence similarity between chimpanzee and human for coding (blue) and noncoding (3' UTR, yellow) transcribed sequence. EST and cDNA sequences (N = 134, Additional data file 1) were selected that cross a well defined stop codon and that provide concurrent sampling of 150bp of sequence both proximal and distal to the stop. The best human match for each chimpanzee sequence was identified and compared using MegaBlast.
Click here for file
Acknowledgements
This work was supported by Public Health Service grants P51RR00166 and R24RR16354 from the National Center for Research Resources. The Oregon National Primate Research Center Tissue Distribution Program is supported by Public Health Service grant P51RR00163. We thank Robert Norgren and Eliot Spindel for supplying additional rhesus macaque sequences; Barney Saunders, Charlie Nelson, and Christopher Hopkins at Agilent Technologies for oligonucleotide design; and Lianne Okada and Katie Woodard for excellent technical assistance. We also thank Roger Bumgarner and Ted Holzman for assistance with early phases of nucleotide sequence data development and analysis.
==== Refs
Carroll SB Genetics and the making of Homo sapiens . Nature 2003 422 849 12712196 10.1038/nature01495
Enard W Paabo S Comparative primate genomics. Annu Rev Genomics Hum Genet 2004 5 351 378 15485353 10.1146/annurev.genom.5.061903.180040
Olson MV Varki A Sequencing the chimpanzee genome: insights into human evolution and disease. Nat Rev Genet 2003 4 20 28 12509750 10.1038/nrg981
Watanabe H Fujiyama A Hattori M Taylor TD Toyoda A Kuroki Y Noguchi H BenKahla A Lehrach H Sudbrak R DNA sequence and comparative analysis of chimpanzee chromosome 22. Nature 2004 429 382 388 15164055 10.1038/nature02564
Rhesus monkey demands in biomedical research: a workshop report. Workshop on Rhesus Monkey Demands In Biomedical Research 2002 Washington DC: National Academy of Sciences 31
Carlsson HE Schapiro SJ Farah I Hau J Use of primates in research: a global overview. Am J Primatol 2004 63 225 237 15300710 10.1002/ajp.20054
Nath BM Schumann KE Boyer JD The chimpanzee and other nonhuman-primate models in HIV-1 vaccine research. Trends Microbiol 2000 8 426 431 10989311 10.1016/S0966-842X(00)01816-3
Sauermann U Making the animal model for AIDS research more precise: the impact of major histocompatibility complex (MHC) genes on pathogenesis and disease progression in SIV-infected monkeys. Curr Mol Med 2001 1 515 522 11899095 10.2174/1566524013363555
Ewing B Green P Base-calling of automated sequencer traces using phred. II. Error probabilities. Genome Res 1998 8 186 194 9521922
Fitzgerald M Shenk T The sequence 5'-AAUAAA-3'forms parts of the recognition site for polyadenylation of late SV40 mRNAs. Cell 1981 24 251 260 6113054 10.1016/0092-8674(81)90521-3
Macaque.org
Pruitt KD Maglott DR RefSeq and LocusLink: NCBI gene-centered resources. Nucleic Acids Res 2001 29 137 140 11125071 10.1093/nar/29.1.137
Zhang Z Schwartz S Wagner L Miller W A greedy algorithm for aligning DNA sequences. J Comput Biol 2000 7 203 214 10890397 10.1089/10665270050081478
Wheeler DL Church DM Federhen S Lash AE Madden TL Pontius JU Schuler GD Schriml LM Sequeira E Tatusova TA Database resources of the National Center for Biotechnology Information. Nucleic Acids Res 2003 31 28 33 12519941 10.1093/nar/gkg033
International Human Genome Sequencing Consortium Finishing the euchromatic sequence of the human genome. Nature 2004 431 931 945 15496913 10.1038/nature03001
Hellmann I Zollner S Enard W Ebersberger I Nickel B Paabo S Selection on human genes as revealed by comparisons to chimpanzee cDNA. Genome Res 2003 13 831 837 12727903 10.1101/gr.944903
Geraghty DE Genetic diversity and genomics of the immune response. Immunol Rev 2002 190 5 8 12493002 10.1034/j.1600-065X.2002.19001.x
Klein J Satta Y O'HUigin C Takahata N The molecular descent of the major histocompatibility complex. Annu Rev Immunol 1993 11 269 295 8476562 10.1146/annurev.iy.11.040193.001413
Sheehy AM Gaddis NC Choi JD Malim MH Isolation of a human gene that inhibits HIV-1 infection and is suppressed by the viral Vif protein. Nature 2002 418 646 650 12167863 10.1038/nature00939
Sawyer SL Emerman M Malik HS Ancient adaptive evolution of the primate antiviral DNA-editing enzyme APOBEC3G. PLoS Biol 2004 2 E275 15269786 10.1371/journal.pbio.0020275
Forsyth IA Wallis M Growth hormone and prolactin - molecular and functional evolution. J Mammary Gland Biol Neoplasia 2002 7 291 312 12751893 10.1023/A:1022804817104
Zhang J Rosenberg HF Diversifying selection of the tumor-growth promoter angiogenin in primate evolution. Mol Biol Evol 2002 19 438 445 11919285
Ashburner M Ball CA Blake JA Botstein D Butler H Cherry JM Davis AP Dolinski K Dwight SS Eppig JT Gene ontology: tool for the unification of biology. The Gene Ontology Consortium. Nat Genet 2000 25 25 29 10802651 10.1038/75556
Deloukas P Earthrowl ME Grafham DV Rubenfield M French L Steward CA Sims SK Jones MC Searle S Scott C The DNA sequence and comparative analysis of human chromosome 10. Nature 2004 429 375 381 15164054 10.1038/nature02462
Nei M Li WH Mathematical model for studying genetic variation in terms of restriction endonucleases. Proc Natl Acad Sci USA 1979 76 5269 5273 291943 10.1073/pnas.76.10.5269
Livingston RJ von Niederhausern A Jegga AG Crawford DC Carlson CS Rieder MJ Gowrisankar S Aronow BJ Weiss RB Nickerson DA Pattern of sequence variation across 213 environmental response genes. Genome Res 2004 14 1821 1831 15364900 10.1101/gr.2730004
Sachidanandam R Weissman D Schmidt SC Kakol JM Stein LD Marth G Sherry S Mullikin JC Mortimore BJ Willey DL A map of human genome sequence variation containing 1.42 million single nucleotide polymorphisms. Nature 2001 409 928 933 11237013 10.1038/35057149
Yu N Jensen-Seaman MI Chemnick L Ryder O Li WH Nucleotide diversity in gorillas. Genetics 2004 166 1375 1383 15082556 10.1534/genetics.166.3.1375
Frazer KA Chen X Hinds DA Pant PV Patil N Cox DR Genomic DNA insertions and deletions occur frequently between humans and nonhuman primates. Genome Res 2003 13 341 346 12618364 10.1101/gr.554603
Tukey J Exploratory Data Analysis 1977 29 Reading, MA: Addison-Wesley
Benit L Calteau A Heidmann T Characterization of the low-copy HERV-Fc family: evidence for recent integrations in primates of elements with coding envelope genes. Virology 2003 312 159 168 12890629 10.1016/S0042-6822(03)00163-6
Pertea G Huang X Liang F Antonescu V Sultana R Karamycheva S Lee Y White J Cheung F Parvizi B TIGR Gene Indices clustering tools (TGICL): a software system for fast clustering of large EST datasets. Bioinformatics 2003 19 651 12651724 10.1093/bioinformatics/btg034
Stewart CB Disotell TR Primate evolution - in and out of Africa. Curr Biol 1998 8 R582 588 9707399 10.1016/S0960-9822(07)00367-3
Crouau-Roy B Service S Slatkin M Freimer N A fine-scale comparison of the human and chimpanzee genomes: linkage, linkage disequilibrium and sequence analysis. Hum Mol Genet 1996 5 1131 1137 8842732 10.1093/hmg/5.8.1131
Page SL Goodman M Catarrhine phylogeny: noncoding DNA evidence for a diphyletic origin of the mangabeys and for a human-chimpanzee clade. Mol Phylogenet Evol 2001 18 14 25 11161738 10.1006/mpev.2000.0895
Schmutz J Martin J Terry A Couronne O Grimwood J Lowry S Gordon LA Scott D Xie G Huang W The DNA sequence and comparative analysis of human chromosome 5. Nature 2004 431 268 274 15372022 10.1038/nature02919
Bukh J A critical role for the chimpanzee model in the study of hepatitis C. Hepatology 2004 39 1469 1475 15185284 10.1002/hep.20268
Kumar S Hedges SB A molecular timescale for vertebrate evolution. Nature 1998 392 917 920 9582070 10.1038/31927
de Souza SJ Camargo AA Briones MR Costa FF Nagai MA Verjovski-Almeida S Zago MA Andrade LE Carrer H El-Dorry HF Identification of human chromosome 22 transcribed sequences with ORF expressed sequence tags. Proc Natl Acad Sci USA 2000 97 12690 12693 11070084 10.1073/pnas.97.23.12690
Soares MB Bonaldo MF Jelene P Su L Lawton L Efstratiadis A Construction and characterization of a normalized cDNA library. Proc Natl Acad Sci USA 1994 91 9228 9232 7937745 10.1073/pnas.91.20.9228
Billick E Seibert C Pugach P Ketas T Trkola A Endres MJ Murgolo NJ Coates E Reyes GR Baroudy BM The differential sensitivity of human and rhesus macaque CCR5 to small-molecule inhibitors of human immunodeficiency virus type 1 entry is explained by a single amino acid difference and suggests a mechanism of action for these inhibitors. J Virol 2004 78 4134 4144 15047829 10.1128/JVI.78.8.4134-4144.2004
Bogerd HP Doehle BP Wiegand HL Cullen BR A single amino acid difference in the host APOBEC3G protein controls the primate species specificity of HIV type 1 virion infectivity factor. Proc Natl Acad Sci USA 2004 101 3770 3774 14999100 10.1073/pnas.0307713101
Yu Q Chen D Konig R Mariani R Unutmaz D Landau NR APOBEC3B and APOBEC3C are potent inhibitors of simian immunodeficiency virus replication. J Biol Chem 2004 279 53379 53386 15466872 10.1074/jbc.M408802200
Cargill M Altshuler D Ireland J Sklar P Ardlie K Patil N Shaw N Lane CR Lim EP Kalyanaraman N Characterization of single-nucleotide polymorphisms in coding regions of human genes. Nat Genet 1999 22 231 238 10391209 10.1038/10290
Ferris SD Brown WM Davidson WS Wilson AC Extensive polymorphism in the mitochondrial DNA of apes. Proc Natl Acad Sci USA 1981 78 6319 6323 6273863 10.1073/pnas.78.10.6319
Garner KJ Ryder OA Mitochondrial DNA diversity in gorillas. Mol Phylogenet Evol 1996 6 39 48 8812304 10.1006/mpev.1996.0056
Jensen-Seaman MI Kidd KK Mitochondrial DNA variation and biogeography of eastern gorillas. Mol Ecol 2001 10 2241 2247 11555266 10.1046/j.0962-1083.2001.01365.x
Gagneux P The genus Pan : population genetics of an endangered outgroup Trends Genet 2002 18 327 330 12127763 10.1016/S0168-9525(02)02695-1
Kaessmann H Wiebe V Paabo S Extensive nuclear DNA sequence diversity among chimpanzees. Science 1999 286 1159 1162 10550054 10.1126/science.286.5442.1159
Kitano T Schwarz C Nickel B Paabo S Gene diversity patterns at 10 X-chromosomal loci in humans and chimpanzees. Mol Biol Evol 2003 20 1281 1289 12777533 10.1093/molbev/msg134
Stone AC Griffiths RC Zegura SL Hammer MF High levels of Y-chromosome nucleotide diversity in the genus Pan Proc Natl Acad Sci USA 2002 99 43 48 11756656 10.1073/pnas.012364999
Deinard AS Kidd K Identifying conservation units within captive chimpanzee populations. Am J Phys Anthropol 2000 111 25 44 10618587 10.1002/(SICI)1096-8644(200001)111:1<25::AID-AJPA3>3.3.CO;2-I
Morin PA Moore JJ Chakraborty R Jin L Goodall J Woodruff DS Kin selection, social structure, gene flow, and the evolution of chimpanzees. Science 1994 265 1193 1201 7915048 10.1126/science.7915048
Ruvolo M Comparative primate genomics: the year of the chimpanzee. Curr Opin Genet Dev 2004 14 650 656 15531160 10.1016/j.gde.2004.08.007
Khaitovich P Muetzel B She X Lachmann M Hellmann I Dietzsch J Steigele S Do HH Weiss G Enard W Regional patterns of gene expression in human and chimpanzee brains. Genome Res 2004 14 1462 1473 15289471 10.1101/gr.2538704
Chomczynski P Sacchi N Single-step method of RNA isolation by guanidinium thiocyanate-phenol-chloroform extraction. Analyt Biochem 1987 162 156 2440339 10.1016/0003-2697(87)90021-2
Ewing B Hillier L Wendl MC Green P Base-calling of automated sequencer traces using phred. I. Accuracy assessment. Genome Res 1998 8 175 185 9521921
Kent WJ Sugnet CW Furey TS Roskin KM Pringle TH Zahler AM Haussler D The human genome browser at UCSC. Genome Res 2002 12 996 12045153 10.1101/gr.229102. Article published online before print in May 2002
Agilent Technologies Inc
| 15998449 | PMC1175991 | CC BY | 2021-01-04 16:35:49 | no | Genome Biol. 2005 Jun 30; 6(7):R60 | utf-8 | Genome Biol | 2,005 | 10.1186/gb-2005-6-7-r60 | oa_comm |
==== Front
Genome BiolGenome Biology1465-69061465-6914BioMed Central London gb-2005-6-7-r611599845010.1186/gb-2005-6-7-r61MethodTranscript copy number estimation using a mouse whole-genome oligonucleotide microarray Carter Mark G [email protected] Alexei A [email protected] Vincent [email protected] Dawood B [email protected] Condie E [email protected] Charlie [email protected] Minoru SH [email protected] Developmental Genomics and Aging Section, Laboratory of Genetics, National Institute on Aging, National Institutes of Health, 333 Cassell Drive, Baltimore, MD 21224, USA2 Agilent Technologies, Deer Creek Rd, Palo Alto, CA 94304, USA2005 30 6 2005 6 7 R61 R61 31 12 2004 27 4 2005 25 5 2005 Copyright © 2005 Carter et al.; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permitsunrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
An in-situ-synthesized 60-mer oligonucleotide microarray designed to detect transcripts from all mouse genes is presented. Exogenous RNA controls derived from yeast allow quantitative estimation of absolute endogenous transcript abundance
The ability to quantitatively measure the expression of all genes in a given tissue or cell with a single assay is an exciting promise of gene-expression profiling technology. An in situ-synthesized 60-mer oligonucleotide microarray designed to detect transcripts from all mouse genes was validated, as well as a set of exogenous RNA controls derived from the yeast genome (made freely available without restriction), which allow quantitative estimation of absolute endogenous transcript abundance.
==== Body
Background
One of the most tantalizing promises of gene-expression profiling technology has been to develop assays that measure expression of all genes in a given species [1]. This is especially important for the mouse, which is a standard model for various human diseases. The early and rapid development of murine bioinformatics resources such as the draft genome assembly [2] and numerous expressed sequence tag (EST) projects have bolstered the feasibility of developing such microarray platforms for the mouse. However, because it has been difficult to identify all murine genes and correctly group genomic and expressed sequences into genes and transcripts, microarray platforms intended to cover all mouse genes are only now being made widely available, long after the draft assembly was released.
Relatively recent microarray technologies, which require sequence information instead of clones as input, allow investigators to design microarray platforms to detect genes without having to obtain clones, including genes which have yet to be cloned or confirmed as an expressed transcript [3]. Platforms that utilize long oligonucleotides give high sensitivity, with the potential for transcript specificity sufficient to distinguish transcripts from the same locus or closely related gene-family members [4,5].
While microarray-based methods can provide very accurate relative (ratio-based) expression measurements, they usually do not provide absolute expression measurements (that is, transcript copy number). One notable exception described in the literature does provide absolute expression measurements in yeast, but not as copy numbers [6]. That method relies on labeled oligonucleotides complementary to common sequence in each cDNA probe, which are hybridized against each slide as the reference target. In the case of long-oligonucleotide-based microarrays, there is no sequence common to all probes, so such a strategy is not feasible. An appropriate approach for such microarray platforms is to monitor the hybridization behavior of a few spiked-in RNA controls with sequence derived from yeast or other genomes. Control transcript probe intensity data can be used to create a generalized dose-signal model and applied to endogenous transcript intensity data to give transcript abundance estimates. Not only would such absolute expression measurements from microarrays help determine what level of sensitivity is required for downstream validation methods, but they would also allow direct comparison of expression data generated using different methods, as well as a valuable mechanism to compare performance between slides, platforms, or experiments [7]. Most importantly, global absolute expression measurements can be used to more fully describe a given transcriptome, perhaps identifying mRNAs present at less than one copy per cell as candidates for heterogeneous or cell-type-specific expression, or subdividing groups of genes in Gene Ontology (GO) nodes [8] based on transcript abundance.
The work described here is focused on two goals, aimed at facilitating standardization and comparison among mouse microarray studies: first, to create a long-oligonucleotide-based microarray platform covering all identified mouse genes, which can be made widely available; and second, to develop exogenous RNA controls which will allow quantitative estimation of absolute endogenous transcript abundance. The microarray will be made available to the community through Agilent Technologies and exogenous control plasmid vectors will be available upon request from the authors and the American Type Culture Collection (ATCC) (ATCC MBA-201 to -207) without restriction, to be used with the design presented here or incorporated into any non-yeast microarray platform.
Results and discussion
The development of a mouse whole-genome microarray in our laboratory has been an ongoing effort, and each new design has been derived in part from its predecessor (see Additional data files 1 and 2 and Materials and methods for details) [9]. Development of the National Institute on Aging (NIA) Mouse Gene Index [10] facilitated more complete, less redundant microarray design than EST clustering alone for the following reasons. First, clustering was mapped to the genome assembly, improving consolidation of transcriptional units. Second, transcript selection is no longer restricted to library contents, allowing genes absent from NIA cDNA clone collections [11] to be included from other public sequence collections. Finally, all potential splice variants were solved from EST alignments with genomic sequence, so that probes can be designed to common regions in a transcript family, minimizing the effect of differential splicing. Therefore the index has been the basis of gene/transcript identification and sequence selection for all oligonucleotide array designs subsequent to the NIA Mouse 22K Microarray v1.1. During the preparation of this paper, assembly of a long-oligonucleotide microarray platform with full coverage of the mouse genome was reported by Zhang et al. [12] using a sequence selection protocol that incorporated all National Center for Biotechnology Information (NCBI) RefSeq entries, including all mRNA transcripts based solely on prediction algorithms, without experimental evidence of expression (XM sequences). In contrast, our protocol included only a minority of the XM sequences (only those annotated as an identified gene).
As our oligonucleotide probe design and selection process differed slightly from protocols previously used with ink-jet microarrays, we first established that our oligonucleotide probes perform as well as or better than those designed with standard protocols [5,9,13]. To assess the overall performance of the oligonucleotide probes, we carried out a mixing experiment, combining total RNA from E12.5 mouse embryos and placentas to produce a range of gene-expression ratios for each transcript, using a preliminary microarray design (NIA Mouse 22K Microarray v2.0, see Additional data files 1 and 2 for details). In a comparison of E12.5 mouse embryo and placental RNA, statistically significant differential expression was detected for 8,461 of the test array's 21,044 oligonucleotide probes. These differential targets were then examined in the mixtures to calculate observed placental RNA fractions. Figure 1 shows that the distributions of the observed placental RNA fractions at each input level were closely matched with the input placental RNA fractions (median observed fraction = input fraction ± 0.075), and the boundaries of 95% confidence regions were 0.121 to 0.405 from the median. These distributions were consistent with, although narrower than, those seen in a similar study [13] using standard oligonucleotide design procedures, suggesting that our design protocol produces comparable results. More importantly, these data suggest that the oligonucleotide probes are capable of highly quantitative, proportional measurements of transcript abundance, a property required for transcript abundance estimation.
Exogenous RNA control transcripts were developed from Saccharomyces cerevisiae intronic and intergenic sequences [14,15]. A total of 11 candidate sequences were cloned and tested against multiple oligonucleotide probes in preliminary microarray hybridizations (data not shown). After assessing which target/probe pairs produced the best dynamic responses to abundance with the lowest noise, seven control transcripts and corresponding oligonucleotide probes (Tables 1 and 2) were selected for use in the control set. As a result, the NIA Mouse 44K Microarray v2.0 contains all 63 oligonucleotide probes considered as controls, while version 2.1, the final version which will be made available to the community, contains only the seven selected for use, spotted ten times each at different locations on the slide. Loading of each control transcript into total RNA was confirmed as accurate within 2.6-fold by quantitative real-time RT-PCR (qPCR) (Figure 2a), with a very tight correlation (r2 ≥ 0.99) between expected and measured values over seven orders of magnitude.
One basic assumption made in our experimental design is that amplification efficiencies are approximately equal between endogenous mouse transcripts and exogenous yeast control transcripts. To test this, transcript abundances were determined by qPCR for cDNA pools synthesized from total RNA with spike-in controls added, as well as labeled cRNA target mixtures amplified from the same total RNA/spike-in control mixtures, and transcript abundances were determined by qPCR. After linear amplification, individual ratios of each control transcript to the endogenous transcript Dnchc1 (Table 3) were within 3.5-fold (average = 1.98-fold) of those prior to amplification (Figure 3), and the slopes of regression lines for pre- and post-amplification datasets were 0.967 and 0.992, respectively. Results were consistent whether using amplification yield versus input or the increase in Dnchc1 transcripts as measured by qPCR to calculate the fold amplification and fraction of the original sample represented by each qPCR well. The stability of the relationship holds over seven orders of magnitude, suggesting that amplification of transcripts during cRNA microarray target synthesis is not a source of significant bias. In previous attempts using control transcripts with short (20-40 nucleotides) vector-derived poly(A) tails, exogenous controls amplified one or two orders of magnitude less efficiently than endogenous messages (data not shown), indicating that sufficient polyadenylation of controls is critical for efficient amplification.
Microarray expression profiles were generated for three distinct samples each of total RNA from E12.5 whole embryos (EM), E12.5 placenta (PL), R1 embryonic stem cells (ES), and GFP-Exe trophoblast stem cells (TS) [16]. For each microarray, linear regression analysis on mean normalized log10[intensity] values for seven yeast spike-in control probes was used to define a standard curve relating signal intensity to copy number (Figure 2b) for estimation of endogenous transcript abundances. Correlations were very strong between log10[intensity] and log10[input copy number], with r2 ≥ 0.95.
To test the accuracy of estimating transcript abundance in this way, we compared the results with qPCR measurements for a panel of 13 endogenous transcripts (Figure 4). Most (36 of 52, or 69.2%) of the microarray-based transcript copy-number estimates for a panel of 13 endogenous genes were within fivefold of qPCR measurements. Furthermore, trending for each transcript across the four tissue types was consistent between the two methods for all ten non-housekeeping genes showing differential expression.
Many factors are likely to affect the accuracy of transcript abundance estimates. Measurements at or near the microarray's detection limit, but still above that of qPCR assays (Figure 4, Lpl and Axl in TS, filled arrows), tend to overestimate transcript abundance, and these data suggest that the lower limit of microarray-based transcript abundance measurement is approximately 0.05 to 0.06 copies per cell in this experiment. Differential transcript splicing can also have an effect: note that for Ank, H19, Hand1, and Igf2bp3 (Figure 4, open arrows), only one tissue out of four shows greater than a tenfold discrepancy, whereas the other measurement pairs are more closely matched. Given the preceding discussion, we present this method as a way to estimate transcript abundances for groups of genes. Accuracy of the estimates for each gene/probe may be further improved in the future by studying the effects of various probe-selection parameters on measured fluorescence intensity.
Using conservative estimates of the total RNA content recovered from mammalian cells (2.0-3.0 pg/cell in this case, see Materials and methods), transcript abundances were expressed on a copies-per-cell basis (Figure 5). The analysis revealed two striking properties of these transcript-abundance distributions. First, mRNA populations in mammalian tissues are highly complex, which is consistent with previous observations [17,18]. Many transcripts were measured at less than one copy per cell in each tissue (EM = 40.1 ± 0.6%, PL = 46.9 ± 1.3%, ES = 48.2 ± 1.9%, TS = 47.4 ± 3.4%) (Figure 5). A log10[intensity] value of 2.5 was used as a lower cutoff, which corresponds to about one copy in 26 cells, so it appears that measured values from 0.038 to one copy per cell represent transcripts present at very low measurable copy numbers, rather than nonexpressed transcripts. Indeed, quantitative RT-PCR studies in yeast have shown that many genes, particularly transcription factors, are expressed at less than one copy per cell [19]. Furthermore, our estimates of numbers of expressed genes/transcripts and mRNA message content per cell (519,688 to 851,087 mRNAs per cell, 8,357 to 12,739 transcripts, expressed from 8,101 to 11,360 genes, Table 4) compare well with previous estimates ranging from 200,000 to 600,000 mRNAs per cell [20,21], consisting of 11,500 to 15,000 diverse mRNA species [18,20], transcribed from as many or more genes up to 17,000 [18,20,22]. Second, a majority of transcripts expressed in one tissue or cell type are commonly expressed in other diverse cell and tissue types. The number of expressed genes in each tissue was estimated by counting the number of microarray features measuring absolute expression of at least one copy per cell, and converting this set of microarray probes to U-clusters (loci) and transcripts via the NIA Mouse Gene Index (Table 4). Examination of the overlap between each cell type's roster of expressed genes and transcripts reveals that the majority are expressed in common (Tables 4 and 5), as suggested by previous assessments of mRNA complexity [18,20,22]. For example, 93% of expressed placental transcripts are also expressed in embryo, and this group represents 72% of the expressed transcripts in embryo (Table 5). The same relationship holds true for pairings of cultured cells with embryo, with 95% of expressed transcripts in cultured cells also found in embryo, covering 69% of embryonic transcripts.
When comparing frequency distributions for complex, in vivo samples and less complex in vitro cultured cells, we might expect to see large differences, particularly in the case of genes expressed at less than one copy per cell. Transcripts present at less than one copy per cell cannot be present in every cell, and therefore must be expressed heterogeneously. As might be expected, whole embryos had the most distinctive frequency distribution of the four samples examined: embryos had significantly fewer transcripts in the range log10[copies per cell] = -1.0 (0.1 copies per cell), but significantly more in the 0-2 (1 to 100 copies per cell) range. This difference, combined with the higher estimate of total transcripts per cell for whole embryos (Table 4), may reflect the activation, within the context of the very high transcriptional activity present in developing embryos, of many developmental pathways that are normally inactive or minimally active.
In contrast, the high degree of similarity between the frequency distributions for placenta, ES, and TS cells (Figure 5) suggests that levels of expression heterogeneity can be similar for complex tissues and cultured cells. In fact, there is evidence in ES cells that gene expression within a culture is not as uniform as previously supposed, and even key differentiation markers such as Oct4 and cKit are expressed in cellular subpopulations within cultures [23]. Taken together, these observations suggest that cultured ES and TS cells, although clonally isolated, are quite heterogeneous in terms of their gene-expression patterns, with a transcriptional complexity similar to that of E12.5 placenta. Further study, perhaps using in situ hybridization or single-cell RT-PCR methods, will be required to address this issue, but it does beg the question of whether or not this heterogeneity is common to all cultured cells, or a feature specific to pluripotent stem cells.
Conclusion
Here we present an oligonucleotide microarray for gene-expression profiling with representation of the entire mouse genome, according to the NIA Mouse Gene Index version 2.0 [24]. An integral feature of this new whole-genome microarray design is a set of probes detecting yeast spike-in control transcripts, which will be available to the community without restriction. Using qPCR, we have shown that this control system allows the reproducible estimation of absolute transcript levels. A valuable tool for the mammalian functional genomics community, this system is a step towards standardization of microarray results by using exogenous RNA control systems that are compatible with multiple microarray platforms and model organisms.
Materials and methods
Microarray design: target sequence selection
The NIA Mouse 44K Microarray v2.0 (Whole Genome 60-mer Oligo) design was based on the NIA Mouse Gene Index v2.0 [24]. Like the first version of the NIA Mouse Gene Index [10], it combines data from multiple transcript databases (RefSeq, Ensembl, Riken, GenBank, and NIA) to construct gene/transcript models which represent all possible transcripts. Briefly, 249,200 ESTs developed at NIA were clustered using clustering tools from The Institute for Genome Reserach (TIGR) [25], generating 58,713 consensus and singleton sequences which were then combined with the other datasets. The major difference in version 2 from version 1 is the use of a clustering method based on genome alignments rather than sequence homology between NIA EST clusters and public sequences. Individual sequences were aligned to the mouse genome [2] using BLAT [26], then clustered by an algorithm similar to the one described by Eyras et al. [27], to be published elsewhere. Our assembly included 30,796 primary genes and 1,318 gene copies or pseudogenes, as well as 28,928 clusters that did not match our criteria for high-confidence genes (open reading frame (ORF) of more than 100 amino acids or multiple exons). There were 65,477 transcripts associated with primary genes. Because transcripts were built from sequence alignments to the mouse genome, they match published genomic sequences [2] (February 2003 edition) exactly.
Microarray design: oligonucleotide probe design and selection
In designing a mouse whole-genome microarray, we began by examining existing designs - the NIA Mouse 22K Microarray v1.1 (Development 60-mer Oligo) [9], which became commercially available from Agilent as the Agilent Mouse (Development) Oligonucleotide Microarray (see Additional data files 1 and 2), and the National Institute of Environmental Health Sciences (NIEHS) Toxicogenomics Consortium mouse array (Agilent Mouse Microarray). Criteria for selecting previously designed probes included a good match to the target gene's major transcript with the longest ORF, minimum predicted cross-reactivity with other expressed sequences, and nonredundancy. Although a perfect match of all 60 base-pairs (bp) of the oligonucleotide was preferred, we also accepted up to two mismatches to the genome if the oligonucleotide matched perfectly to the RefSeq sequence, and oligonucleotide sequences that did not match 100% to the RefSeq entry were corrected. An oligonucleotide was considered cross-reactive if its last 43 bp (solution end) matched to a non-target gene with less than five mismatches. Deletion placement studies using in-situ synthesized 60-mer oligonucleotide probes suggest that the 17 bp at the support surface have a negligible effect on hybridization intensity [5]; thus only the external 43 bp were considered important. While the cross-reactivity criterion is easily satisfied for unique genes with low similarity to other genes, many gene families had high sequence similarity between member transcripts, and it was impossible to find regions with low predicted cross-reactivity. In this case we considered the whole gene family as a target; then the oligonucleotide was considered cross-reactive only if it matched to genes outside the family. Gene families were assembled using a 30% transcript length alignment as a threshold of similarity; alignments for each pair of transcripts were generated using BLAT [26]. According to the nonredundancy criterion, we left only one oligonucleotide that matched to each gene or gene family, and when probes from both the NIA Mouse 22K v1.1 and NIEHS Toxicogenomics arrays matched well to the same gene, preference was given to the NIA oligonucleotide.
After filtering with the above criteria, we obtained 6,563 probes from the NIA Mouse 22K Microarray v1.1 and 9,551 probes from the NIEHS Toxicogenomics array. Among these oligonucleotides, 3,327 did not match the target gene's major transcript with the longest ORF, so we generated an additional 3,327 probes for major transcripts of the same genes. Then we generated 22,850 probes for the best transcripts of primary genes in the gene index that were not represented in the NIA Mouse 22K Microarray v1.1 (Development 60-mer Oligo) and NIEHS Toxicogenomics arrays, for a total of 42,291 non-control oligonucleotide probes (see Additional data file 2). For each transcript we generated ten probes using ArrayOligoSelector [28], then selected the best oligonucleotide on the basis of minimum predicted cross-reactivity, proximity to the 3' end, and degree of matching to RefSeq or GenBank sequences. The latter criterion was important only in cases of mismatches between genomic sequence and RefSeq or GenBank.
All microarray data described in this report were generated using the NIA Mouse 44K Microarray v2.1 (Whole Genome 60-mer Oligo) and NIA Mouse 22K Microarray v2.0 (Development 60-mer Oligo). We have slightly modified the probe content of the NIA Mouse 44K v2.0 array by including Agilent's standard QC probe set, removing candidate spike-in control probes which were not used, and including additional probes for known genes that have existing probes with poor performance or ambiguous targeting. The updated version (NIA Mouse 44K Microarray v2.1 (Whole Genome 60-mer Oligo) will be made available to the community (see Additional data file 1).
Yeast spike-in controls
Yeast (S. cerevisiae) sequences were selected from public repositories [14,15] to produce exogenous RNA control transcripts, commonly referred to as 'spike-in' controls. Fourteen candidates (ten intergenic and four intronic) were selected on the basis of sequence length and the absence of restriction endonuclease cleavage sites important for our cloning strategy. Sequences with significant matches to transcripts in the NIA mouse Gene Index v2.0 [10] were discarded, and ten of the 14 remaining candidates were successfully cloned from genomic DNA, with one sequence divided into two clones for a total of 11 potential controls. Yeast sequences were amplified with added 5' SalI and 3' XbaI sites from S. cerevisiae genomic DNA (ATCC 2601D) using Sigma RedTaq, and cloned directly into pCR4-TOPO (Invitrogen). TA-TOPO clones were verified by sequencing on an Applied Biosystems 3100 capillary DNA sequencer, and inserts were directionally subcloned into pSP64 Poly(A) (Promega Catalog number P1241) using the introduced SalI and XbaI sites. A total of 63 60-mer oligonucleotide 'sense-strand' probes were selected for the 14 candidate sequences using both ArrayOligoSelector software [28] and arbitrary manual selection. Oligonucleotide probes were compared to NIA Gene Index transcripts, and no significant matches were found. Control probes were spotted ten times each in various locations throughout the slides.
Spike-in RNA was transcribed, polyadenylated, and purified using Ambion mMessage mMachine, poly(A) tailing, and MegaClear kits, then sized and quantitated by RNA 6000 Nano assay on an Agilent Bioanalyzer 2100. Spike-in RNAs were pooled to create tenfold concentration differences, from 104 to 1010 copies per microliter (Table 1). Before preparation of microarray targets, 1 μl of this control transcript mixture was added to 5-μg aliquots of each total RNA sample, including the reference RNA. A separate pool with all yeast control transcripts present at the same copy number was added to reference RNA and converted to cDNA for use as a standard in qPCR assays.
RNA collection/preparation
Total RNA was prepared using TriZol reagent (Invitrogen) from E12.5 C57BL/6J embryos, pooled by litter, and corresponding E12.5 C57BL/6J placenta pools [9]. Total RNA was also prepared from R1 ES cells passaged briefly on gelatin to remove feeder cells, and GFP-Exe TS cells grown on plastic in conditioned medium as previously described [16]. Total RNA quantity and quality were assessed by RNA 6000 Nano assay. For oligonucleotide signal linearity testing, E12.5 embryo and placenta total RNA were pooled, based on this quantitation, to produce duplicate samples with 0, 25, 50, 75, and 100% placental RNA content.
cRNA target labeling
Fluorescently labeled microarray targets were prepared from 2.5 μg aliquots of total RNA samples with yeast sequence control mixtures added as described above, using a Low RNA Input Fluorescent Linear Amplification Kit (Agilent). A reference target (Cy5-CTP-labeled) was produced from Stratagene Universal Mouse Reference RNA, and all other targets were labeled with Cy3-CTP. Targets were purified using an RNeasy Mini Kit (Qiagen) as directed by Agilent's clean-up protocol, and quantitated on a NanoDrop scanning spectrophotometer (NanoDrop Technologies).
Microarray hybridization
All hybridizations compared one Cy3-CTP-labeled experimental target to the single Cy5-CTP-labeled reference target. Microarrays were hybridized and washed according to Agilent protocol G4140-90030 (Agilent 60-mer oligo microarray processing protocol - SSC Wash, v1.0). Slides were scanned on an Agilent DNA Microarray Scanner, using standard settings, including automatic PMT adjustment.
Real-time quantitative RT-PCR
Primer sets were designed and tested for SYBR Green chemistry using an established in-house protocol [9]. Total RNA was used to prepare cDNA as described previously [9]. Because the microarray targets were oligo(dT) primed, all cDNA synthesis reactions were oligo(dT) primed as well, and qPCR primer sets were designed so that amplicons were upstream of 60-mer oligonucleotide probes when possible, or less than 650 bp downstream. These steps were taken to minimize the effects of 3' end-labeling bias from microarray target synthesis. Yeast spike-in standard curve cDNA was prepared by mixing equal copy numbers of each synthetic yeast RNA with Mouse Universal Reference total RNA, followed by cDNA synthesis. A standard for copy-number measurement of endogenous mouse genes was prepared by transcribing cDNA clones and adding these transcripts in equal numbers to yeast total RNA, followed by cDNA synthesis. A BioMek 2000 liquid-handling system (Beckman) was used to aliquot cDNA into 96- and 384-well plates, then assemble and aliquot PCR master mix into 20-25 μl reactions. Plates were run on ABI 7700 or ABI 7900 HT Sequence Detection Systems using the default cycling program, and data was processed using SDS 1.9 or SDS 2.2 software (Applied Biosystems) and Microsoft Excel.
Data analysis
Microarray images were processed with Agilent Feature Extractor A.7.5.1 software to generate normalized, background-subtracted feature intensities. Dye normalization was performed by applying a LOWESS algorithm to all significant, non-control and non-outlier features. Analysis of variance (ANOVA) and replicate averaging was performed as previously described [9] using NIA Array Analysis Tool software [29], which normalizes each probe according to reference RNA signals.
For each probe identified as differentially expressed in mixing experiments (false discovery rate < 0.05) [9], linear regressions of ratios against pure placental RNA across the five levels of placental RNA content were calculated, and observed ratios were back-calculated for population analysis as
where Poi is the observed fraction placental RNA content calculated from a given probe i, Ipi and I100i are the normalized log10[intensity] values for the probe i at placental RNA percentages p and 100, respectively, and ai and bi are the intercept and slope of the ratios versus the input placental RNA fraction for probe i. For the population of observed fractions at each input placental RNA fraction, the mean and median were calculated, along with the 2.5, 25, 75, and 97.5 percentile boundaries (Figure 1).
For endogenous transcript abundance estimation experiments, linear regression analysis was performed on seven yeast spike-in probe mean normalized log10[intensity] values for each microarray and the results were used to back-calculate estimated copy numbers for endogenous transcripts as
where Chmi is the microarray-estimated number of copies per hybridization for probe i, Ii is the normalized log10[intensity] for probe i, and a and b are the intercept and slope of spike-in control probe microarray signal intensities versus. input spike-in transcript copy numbers. Dividing these values by the estimated number of cells represented in each hybridization,
converts them to estimates of transcript copies per cell. Amounts of total RNA extracted per cell for the four tissue types (EM 3.0 pg/cell, PL 2.0 pg/cell, ES 2.3 pg/cell, TS 3.0 pg/cell) were estimated from cell counts, RNA yields, and in the case of E12.5 embryo and placenta, our estimate that the average cell volume in these tissues is approximately 1.5 × 10-9 cm3 per cell (data not shown).
For measurement of abundances of mouse endogenous gene and spiked-in yeast transcripts in total RNA and labeled/amplified target mixtures by qPCR, linear regression of threshold cycle (Ct) values versus input spike-in transcript copy numbers in a standard was used to back-calculate copy numbers per well of the transcripts in the total RNA samples and labeled/amplified target mixtures. These results were converted to copies per cell as follows:
In the case of endogenous mouse transcript measurements, results from both the microarray and qPCR were normalized to Gapd expression.
All microarray data will be deposited to the public repositories Gene Expression Omnibus at NCBI [30,31] and ArrayExpress at EBI [32,33] as soon as possible.
Additional data files
The following additional data are available with the online version of this paper. Additional data file 1 is a table containing a standardized naming scheme for NIA oligonucleotide microarray platforms. Additional data file 2 is a table containing additional information on previous NIA microarray platforms and how they relate to that presented in this work. Additional data file 3 contains annotation of all probes in the NIA 44K Mouse Microarray v2.1.
Supplementary Material
Additional File 1
A standardized naming scheme for NIA oligonucleotide microarray platforms.
Click here for file
Additional File 2
Additional information on previous NIA microarray platforms and how they relate to that presented in this work.
Click here for file
Additional File 3
Annotation of all probes in the NIA 44K Mouse Microarray v2.1
Click here for file
Acknowledgements
The authors thank Peter Webb at Agilent Technologies for his assistance in preparing the microarray design for production, and his colleague Paul Wolber for advice in constructing the yeast spike-in control constructs. Yong Qian of the NIA provided invaluable bioinformatics and computational support for many aspects of this work. We also thank Janet Rossant and Tilo Kunath for providing ES and TS cell RNA. DNA microarrays produced according to NIA designs are available commercially from Agilent Technologies. However, The National Institutes on Health and The National Institute on Aging do not endorse these products or make any claims or guarantees as to their quality or performance.
Figures and Tables
Figure 1 60-mer oligonucleotide probe linearity testing. To test the performance of 21,044 60-mer oligonucleotide probes, E12.5 embryo RNA and placenta RNA were combined to form five pairs of duplicate samples containing from 0 to 100% placental RNA. Box-plot distribution data for each placental RNA input level is shown above, with median values labeled. The boxes show the 25-75 percentile range, with the mean and median indicated by the central straight line and diamond, respectively. Upper and lower bars show the 2.5 to 97.5 percentile range. Observed fraction medians are within 0.075 of input values, and 95% of values are within 0.405 of input values.
Figure 2 Relating yeast spike-in RNA control copy number to qPCR measurements and microarray signal intensity. (a) To verify abundances of yeast sequence RNA transcripts in a control mixture, cDNA was transcribed from the control mixture alone (open boxes), as well as E12.5 whole-mouse embryo total RNA (open diamonds) and Universal Mouse RNA (filled triangles) with added spike-in control mixture. The cDNA was used as template for real-time PCR quantitation of each yeast sequence RNA, using a separately prepared standard of cDNA transcribed from the yeast sequences. Expected and measured copy numbers are closely matched (r2 ≥ 0.99), with maximum measured/observed ratios of 1.5, 1.5, and 2.6, respectively. (b) Expression profiles were generated for triplicate total RNA samples from E12.5 embryo (filled circles), E12.5 placenta (open circles), ES cells (filled boxes), and TS cells (open boxes) with yeast sequence control transcripts spiked-in prior to target labeling. For the seven control transcripts, mean log10[intensity] is shown for each tissue type, as well as the mean across all samples (filled triangles), and these data were used to perform linear regression analysis and relate signal intensity to transcript copy number, allowing abundance estimation for endogenous transcripts. The regression line for the average of all tissues (dashed line) and its equation is shown. Intensity-copy number correlations for individual tissues were very strong, with r2 values of 0.98 - 0.99.
Figure 3 Exogenous control and endogenous transcript amplification rates are closely matched over seven orders of magnitude. Transcript abundance of each spike-in control transcript was measured by qPCR before and after linear amplification labeling, and compared to amounts of the exogenous transcript Dnchc1. After amplification, individual ratios of each control transcript to the endogenous transcript were within 3.5-fold (average = 1.98-fold) of those prior to amplification. Blue diamonds = log10[ratio mean control/Dnchc1 transcripts] of three E12.5 embryo and three E12.5 placenta samples before amplification. Red boxes, green triangles = log10[ratio mean control/Dnchc1 transcripts] for the same samples after amplification, using yield versus input (red boxes) or the increase in Dnchc1 transcripts as measured by qPCR (green triangles) to calculate the fraction of the original sample represented by each qPCR well.
Figure 4 Validation of transcript abundance estimation for endogenous transcripts. qPCR primer sets were designed for selected genes so that amplicons were upstream of 60-mer oligonucleotide probes when possible, or less than 650 bp downstream, and copy number was estimated using serial dilutions of RNA, in vitro transcribed from mouse cDNAs, at known copy numbers as standards. Error bars represent one standard deviation across three replicate samples for each tissue. Dotted diagonal lines represent five- and tenfold differences between the two datasets. Each gene's official symbol, along with the unique identifier for the 60-mer oligonucleotide probe it was measured with, are listed in the key. Data was normalized to Gapd expression for both methods. EM = E12.5 embryo, PL = E12.5 placenta, ES = embryonic stem cells, TS = trophoblast stem cells.
Figure 5 Distribution of mouse transcript abundances in E12.5 embryo and placenta, and cultured ES and TS cells. Transcript abundances are expressed as log10[copies per cell], varying over six orders of magnitude. The distributions are highly similar, despite the significant differences between the four tissues (for example, monolayer culture versus tissue, placenta versus embryo), suggesting that such distributions are not heavily skewed according to tissue structure or function. The percentage of transcripts present at less than one copy per cell ranged from 40.1 to 48.2% in the four tissues. Bins were centered on indicated values, and the dotted lines indicate values corresponding to mean upper and lower signal intensity reliability limits of one copy per 26 cells to 2,188 copies per cell. For definitions of tissue type see Figure 4 legend.
Table 1 Yeast controls used in this study with corresponding qPCR primers
Yeast intronic/intergenic control transcript Vector name ATCC number GenBank Accession Insert size (bp) Copies spiked/5 μg total RNA Forward/reverse qPCR oligo sequence Optimal concentration Amplicon Intron spanned?
Size Tm
YPL075W_16_412249_415357_INTRON_9_759 pNIAysic-1 MBA-201 DQ023287 630 1.00E+04 5'-CCTACTTGATAAAGCCACATACCTCTACCTCTTCTATTAG-3' 5'-TTGCGTTACTCTATTAATAATCCATAGTTGGAAC-3' 300 nM
50 nM 134 bp 73.4°C No
YPL081W_16_404945_406039_INTRON_8_508 pNIAysic-2 MBA-202 DQ023288 400 1.00E+05 5'-CGACACTTCAGGTAAAGCGTTCCGAAGTAATTCAAC-3' 5'-TCTCAAACCTAACACATTTCTGTATTAAGCCTAG-3' 300 nM
300 nM 129 bp 75.8°C No
NOT:D_1493031-1494574_553-1543 pNIAysic-3 MBA-203 DQ023289 997 1.00E+06 5'-TTACCATTCACTCCATGATGTCGTACCTGTTACACTAC-3' 5'-CGGTACATGTTATTACCAGAAAAAGATGTATATCC-3' 300 nM
300 nM 145 bp 79.8°C No
YER133W_5_432491_433954_INTRON_178_702 pNIAysic-4 MBA-204 DQ023290 428 1.00E+07 5'-GTCGAGATAGCCGAGATAATGTGTGTG-3' 5'-GCAAGGGGGATTTTTCTGAATATGG-3' 300 nM
300 nM 136 bp 76.5°C No
YNL162W_14_331319_332151_INTRON_5_516 pNIAysic-5 MBA-205 DQ023291 367 1.00E+08 5'-TGCAGCAACAGAGTATCATATGCATGG-3' 5'-CACTGCACAATCTGAAGATAGCGAGG-3' 300 nM
300 nM 145 bp 77.7°C No
YNL302C_14_62942_61957_INTRON_21_571 pNIAysic-6 MBA-206 DQ023292 416 1.00E+09 5'-ATTTCCCATTACCTGATAAATTGAAGTTCATC-3' 5'-TTTGTATAGTTGGCTCAAAATATTCTCTCCAC-3' 900 nM
300 nM 100 bp 73.8°C No
YBL087C_2_60732_59815_INTRON_43_546 pNIAysic-7 MBA-207 DQ023293 436 1.00E+010 5'-GCAGATGAAGTGATACCTGTCAATATTCATG-3' 5'-AGAAATAACATTTCGATGGTTATCCATTAGTATG-3' 300 nM
300 nM 128 bp 76.2°C No
Table 2 Yeast controls with corresponding in situ-synthesized 60-mer oligonucleotide probes
Control transcript NIA probe ID 60-mer oligonucleotide microarray probe sequence
NIA yeast control 1 Z10000036-1 5'-TTCAAGGGACAAATAACAGGATAAAACGTAATGTCAGGACACAAAGTGTGCCATCAACTT-3'
NIA yeast control 2 Z10000039-1 5'-TCTTCATAGAATACTTTTTTTTTCGGAGAAAACCTTTACACTGAACTCCCGACACTTCAG-3'
NIA yeast control 3 Z10000041-1 5'-TTTAATTATTCTTATTTCGCTTTTTTTCTCAAGGTGACCTGTTGTATCACGTTAGCTGAA-3'
NIA yeast control 4 Z10000020-1 5'-TCATCCGGCCGGCGCCTCCCATATTCAGAAAAATCCCCCTTGCTCACACTAAAAAAAGAA-3'
NIA yeast control 5 Z10000021-1 5'-TCAGATTGTGCAGTGATATTCTTTGAGGAAGGAAACGTAGAGGGGATAAGTTGGATAACT-3'
NIA yeast control 6 Z10000026-1 5'-CATTTACCGAACGAATGAGTTAAACTATTATGATATAATTGCTGTAATTGTGGAGAGAAT-3'
NIA yeast control 7 Z10000002-1 5'-AAAGTAAAGTTCCAAGATTTCATTTTGCTGGGTACAACAGAATTAAACAGAGGTTTAAAA-3'
Table 3 qPCR primer pairs used to quantitate endogenous transcripts in this study
Gene symbol Forward/reverse qPCR oligo sequence Optimal concentration Amplicon Intron spanned?
Size Tm
Ank 5'-AGTACCATAGTACACTCGGTTACCTGTCCTG-3' 900 nM 114 bp 78.8°C Yes
5'-GCAAAGCTTTAAGTCGTAATCTAGCATCC-3' 50 nM
Axl/Ufo 5'-CGACTACCTGCGTCAAGGAAATCG-3' 300 nM 112 bp 82.8°C Yes
5'-AAAACTTGGCCGGTCTCGAGG-3' 300 nM
Cd34 5'-TGCTCTGGAATCCGAGAAGTGAGG-3' 300 nM 140 bp 78.0°C Yes
5'-TCAGCCTCAGCCTCCTCCTTTTC-3' 300 nM
Dnchc1 5'-AACTAAACCCAGCCATTCGGCC-3' 300 nM 98 bp 84.3°C No
5'-TTGCGTTGGCGGGTGACAG-3' 900 nM
Gap43 5'-GAGAAGGGAAGGAGAGAAGGCAGG-3' 900 nM 131 bp 79.5°C Yes
5'-TCCGGCTTGACACCATCTTGTTC-3' 900 nM
Gapd 5'-CGGAGTCAACGGATTTGGTCGTAT-3' 900 nM 214 bp 82.6°C Yes
5'-GAAGATGGTGATGGGCTTCC-3' 300 nM
H19 5'-AGCTAACACTTCTCTGCTGCTCTCTGG-3' 300 nM 144 bp 81.4°C Yes
5'-ATCTTCTTGATTCAGAACGAGACGGAC-3' 900 nM
Hand1 5'-GAGATGTATACCTGAGAGCAACAGGCATGATAGGTAG-3' 300 nM 113 bp 75.1°C No
5'-CTTCTCCTTCATTTCTTTCCTTTTCCTTC-3' 900 nM
Hif1a 5'-GTCAGCAGTACATGGTAGCCACAATTG-3' 900 nM 139 bp 74.4°C No
5'-GATCCAGGCTTAACAATTCCATAGGC-3' 300 nM
Hmga1 5'-AATTCAGGAGGATGAACATCTGACGC-3' 900 nM 114 bp 77.3°C No
5'-TCTGTTCACAAACTACCTCTGGACGG-3' 50 nM
Hprt1 5'-AACAATGCAAACTTTGCTTTCCCTG-3' 300 nM 123 bp 80.1°C Yes
5'-TCAAATCCAACAAAGTCTGGCCTG-3' 300 nM
Igf2bp3 5'-AAGTATACATTCTCACAGAGACAGGATCGAGTGACTG-3' 900 nM 126 bp 81.5°C No
5'-AAAGACAGATTTGCTTAACCAACAGACG-3' 900 nM
Lpl 5'-TTTCCAGCCAGGATGCAACATTG-3' 300 nM 105 bp 82.3°C No
5'-TGAATGGAGCGCTCATGCGAG-3' 900 nM
Myo1b 5'-AATACACACCTTGTACCAATCAGCTCTCTC-3' 900 nM 143 bp 76.1°C No
5'-TGATAAGAAGAGGCTGAGAGCCGTTC-3' 900 nM
Table 4 Expressed genes and transcripts in developing mouse tissues and cultured stem cells
EM PL ES TS Any tissue All tissues
mRNAs/cell 851,087 519,688 400,045 568,196
Features ≥ 1 CpC 13,718 10,559 9,667 9,840 14,908 8,073
U-clusters ≥ 1 CpC 11,360 8,828 8,101 8,271 12,264 6,838
Transcripts ≥ 1 CpC 11,762 9,108 8,357 8,534 12,739 7,037
Mean copies per cell 1.09 0.63 0.51 0.56
Median copies per cell 0.79 0.45 0.36 0.40
U-clusters and transcripts from the NIA mouse gene index were considered expressed if microarray features measured absolute expression estimated at one copy per cell or more. Copy-number estimates from expressed transcripts were summed to estimate the number of mRNA molecules per cell for each tissue, as well as the mean and median copy numbers. Microarray features corresponding to expressed genes and transcripts were mapped to the NIA Gene Index to calculate the number of U-clusters (loci) and transcripts expressed in each tissue.
Table 5 Pairwise comparison of expressed transcript sets in developing mouse tissues and cultured cells
Total expressed features Overlapping features EM PL ES TS
13,718 EM 9,840 9,212 9,314
10,559 PL 8,508 8,881
9,667 ES 8,816
9,840 TS
Total expressed U-clusters Overlapping U-clusters EM PL ES TS
11,360 EM 8,271 7,749 7,853
8,828 PL 7,181 7,492
8,101 ES 7,435
8,271 TS
Total expressed transcripts Overlapping transcripts EM PL ES TS
11,762 EM 8,516 7,980 8,090
9,108 PL 7,386 7,718
8,357 ES 7,657
8,534 TS
Sets of microarray features measuring expressed genes (≥ 1 copy per cell) were compared pairwise to calculate the number of members common to each pair. By matching microarray features to the NIA Gene Index, numbers of U-clusters (loci) and transcripts expressed in common were derived for each pairwise comparison. Signal intensities which were lower than those for all spike-in controls, as well as saturated signals, were not converted to copy number estimates (see Materials and methods), so these calculations may underestimate the number of expressed genes.
==== Refs
Ko MS An 'equalized cDNA library' by the reassociation of short double-stranded cDNAs. Nucleic Acids Res 1990 18 5705 5711 2216762
Waterston RH Lindblad-Toh K Birney E Rogers J Abril JF Agarwal P Agarwala R Ainscough R Alexandersson M An P Initial sequencing and comparative analysis of the mouse genome. Nature 2002 420 520 562 12466850 10.1038/nature01262
Schadt EE Edwards SW GuhaThakurta D Holder D Ying L Svetnik V Leonardson A Hart KW Russell A Li G A comprehensive transcript index of the human genome generated using microarrays and computational approaches. Genome Biol 2004 5 R73 15461792 10.1186/gb-2004-5-10-r73
Singh-Gasson S Green RD Yue Y Nelson C Blattner F Sussman MR Cerrina F Maskless fabrication of light-directed oligonucleotide microarrays using a digital micromirror array. Nat Biotechnol 1999 17 974 978 10504697 10.1038/13664
Hughes TR Mao M Jones AR Burchard J Marton MJ Shannon KW Lefkowitz SM Ziman M Schelter JM Meyer MR Expression profiling using microarrays fabricated by an ink-jet oligonucleotide synthesizer. Nat Biotechnol 2001 19 342 347 11283592 10.1038/86730
Dudley AM Aach J Steffen MA Church GM Measuring absolute expression with microarrays with a calibrated reference sample and an extended signal intensity range. Proc Natl Acad Sci USA 2002 99 7554 7559 12032321 10.1073/pnas.112683499
van Bakel H Holstege FC In control: systematic assessment of microarray performance. EMBO Rep 2004 5 964 969 15459748 10.1038/sj.embor.7400253
Ashburner M Ball CA Blake JA Botstein D Butler H Cherry JM Davis AP Dolinski K Dwight SS Eppig JT Gene ontology: tool for the unification of biology. The Gene Ontology Consortium. Nat Genet 2000 25 25 29 10802651 10.1038/75556
Carter MG Hamatani T Sharov AA Carmack CE Qian Y Aiba K Ko NT Dudekula DB Brzoska PM Hwang SS Ko MS In situ-synthesized novel microarray optimized for mouse stem cell and early developmental expression profiling. Genome Res 2003 13 1011 1021 12727912 10.1101/gr.878903
Sharov AA Piao Y Matoba R Dudekula DB Qian Y VanBuren V Falco G Martin PR Stagg CA Bassey UC Transcriptome analysis of mouse stem cells and early embryos. PLoS Biol 2003 1 E74 14691545 10.1371/journal.pbio.0000074
Carter MG Piao Y Dudekula DB Qian Y VanBuren V Sharov AA Tanaka TS Martin PR Bassey UC Stagg CA The NIA cDNA project in mouse stem cells and early embryos. C R Biol 2003 326 931 940 14744099
Zhang W Morris QD Chang R Shai O Bakowski MA Mitsakakis N Mohammad N Robinson MD Zirngibl R Somogyi E The functional landscape of mouse gene expression. J Biol 2004 3 21 15588312 10.1186/jbiol16
Parrish ML Wei N Duenwald S Tokiwa GY Wang Y Holder D Dai H Zhang X Wright C Hodor P A microarray platform comparison for neuroscience applications. J Neurosci Methods 2004 132 57 68 14687675 10.1016/j.jneumeth.2003.09.013
Grate L Ares M Jr Searching yeast intron data at Ares lab web site. Methods Enzymol 2002 350 380 392 12073325
Ares Lab Yeast Intron Database
Tanaka S Kunath T Hadjantonakis AK Nagy A Rossant J Promotion of trophoblast stem cell proliferation by FGF4. Science 1998 282 2072 2075 9851926 10.1126/science.282.5396.2072
Van Ness J Hahn WE Sequence complexity of cDNA transcribed from a diverse mRNA population. Nucleic Acids Res 1980 8 4259 4270 6159588
Axel R Feigelson P Schutz G Analysis of the complexity and diversity of mRNA from chicken liver and oviduct. Cell 1976 7 247 254 954082 10.1016/0092-8674(76)90024-6
Holland MJ Transcript abundance in yeast varies over six orders of magnitude. J Biol Chem 2002 277 14363 14366 11882647 10.1074/jbc.C200101200
Hastie ND Bishop JO The expression of three abundance classes of messenger RNA in mouse tissues. Cell 1976 9 761 774 1017013 10.1016/0092-8674(76)90139-2
Bishop JO Morton JG Rosbash M Richardson M Three abundance classes in HeLa cell messenger RNA. Nature 1974 250 199 204 4855195
Jongeneel CV Iseli C Stevenson BJ Riggins GJ Lal A Mackay A Harris RA O'Hare MJ Neville AM Simpson AJ Strausberg RL Comprehensive sampling of gene expression in human cell lines with massively parallel signature sequencing. Proc Natl Acad Sci USA 2003 100 4702 4705 12671075 10.1073/pnas.0831040100
Hubner K Fuhrmann G Christenson LK Kehler J Reinbold R De La Fuente R Wood J Strauss JF 3rdBoiani M Scholer HR Derivation of oocytes from mouse embryonic stem cells. Science 2003 300 1251 1256 12730498 10.1126/science.1083452
NIA Mouse Gene Index 2.0
Pertea G Huang X Liang F Antonescu V Sultana R Karamycheva S Lee Y White J Cheung F Parvizi B TIGR Gene Indices clustering tools (TGICL): a software system for fast clustering of large EST datasets. Bioinformatics 2003 19 651 652 12651724 10.1093/bioinformatics/btg034
Kent WJ BLAT - the BLAST-like alignment tool. Genome Res 2002 12 656 664 11932250 10.1101/gr.229202. Article published online before March 2002
Eyras E Caccamo M Curwen V Clamp M ESTGenes: alternative splicing from ESTs in Ensembl. Genome Res 2004 14 976 987 15123595 10.1101/gr.1862204
Bozdech Z Zhu J Joachimiak MP Cohen FE Pulliam B DeRisi JL Expression profiling of the schizont and trophozoite stages of Plasmodium falciparum with a long-oligonucleotide microarray. Genome Biol 2003 4 R9 12620119 10.1186/gb-2003-4-2-r9
NIA Array Analysis Tool
Gene Expression Omnibus
Edgar R Domrachev M Lash AE Gene Expression Omnibus: NCBI gene expression and hybridization array data repository. Nucleic Acids Res 2002 30 207 210 11752295 10.1093/nar/30.1.207
ArrayExpress
Brazma A Parkinson H Sarkans U Shojatalab M Vilo J Abeygunawardena N Holloway E Kapushesky M Kemmeren P Lara GG ArrayExpress - a public repository for microarray gene expression data at the EBI. Nucleic Acids Res 2003 31 68 71 12519949 10.1093/nar/gkg091
| 15998450 | PMC1175992 | CC BY | 2021-01-04 16:05:40 | no | Genome Biol. 2005 Jun 30; 6(7):R61 | utf-8 | Genome Biol | 2,005 | 10.1186/gb-2005-6-7-r61 | oa_comm |
==== Front
Genome BiolGenome Biology1465-69061465-6914BioMed Central London gb-2005-6-7-r621599845110.1186/gb-2005-6-7-r62MethodValidation and refinement of gene-regulatory pathways on a network of physical interactions Yeang Chen-Hsiang [email protected] H Craig [email protected] Scott [email protected] Christopher [email protected] Tommi [email protected] Trey [email protected] Center for Biomolecular Science and Engineering, Baskin School of Engineering, University of California at Santa Cruz, Santa Cruz, CA 95064, USA2 Department of Bioengineering, University of California at San Diego, La Jolla, CA 92093, USA3 Computer Science and Artificial Intelligence Laboratory, Massachusetts Institute of Technology, Cambridge, MA 02139, USA2005 1 7 2005 6 7 R62 R62 9 3 2005 3 5 2005 3 6 2005 Copyright © 2005 Yeang 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.
A new automated procedure for prioritizing genetic perturbations was used to evaluate 38 candidate regulatory networks in yeast. Further analysis of four high-priority gene knockout experiments provided new insights into two regulatory pathways
As genome-scale measurements lead to increasingly complex models of gene regulation, systematic approaches are needed to validate and refine these models. Towards this goal, we describe an automated procedure for prioritizing genetic perturbations in order to discriminate optimally between alternative models of a gene-regulatory network. Using this procedure, we evaluate 38 candidate regulatory networks in yeast and perform four high-priority gene knockout experiments. The refined networks support previously unknown regulatory mechanisms downstream of SOK2 and SWI4.
==== Body
Background
Recent advances in genomics and computational biology are enabling construction of large-scale models of gene-regulatory networks. High-throughput technologies such as automated sequencing [1], gene-expression arrays [2], chromatin immunoprecipitation [3], and yeast two-hybrid assays [4], each probe different aspects of the gene-regulatory system through genome-wide datasets. These data have spawned a variety of methods to infer the structure of gene-regulatory networks or to study their high-level properties, as recently reviewed [5].
Regulatory network models generated thus far in Escherichia coli and budding yeast (Saccharomyces cerevisiae) have been most often validated against functional databases or previous literature [6,7]. In contrast, only a few studies have attempted to validate or refine models systematically [8-11]. However, if we are to accurately model large gene networks in complex organisms, including fly, worm, mouse, and human, automated procedures will be essential for analyzing the network, choosing the best new experiments to test the model, conducting the experiments, and integrating the resulting data.
The problem of choosing the best experiments to estimate a model, termed 'experimental design' or 'active learning', has been a significant area of research in statistics and machine learning [12-14]. Automating the experimental design process can greatly accelerate data collection and model building, leading to substantial savings in time, materials, and human effort. For these reasons, many industries such as electronic circuit fabrication and airplane manufacturing incorporate experimental design as an integral step in the design process [15,16]. A promising application of experimental design for biological systems was presented by King et al. [17], who integrated computational modeling and experimental design to reconstruct a small, well studied metabolic pathway. Whether automated experimental design can be useful in a large and poorly characterized biological system with noisy data remains an open question.
We recently reported a procedure for inferring gene-regulatory network models by integrating gene-expression profiles with high-throughput measurements of protein interactions [18]. Here we extend this procedure to incorporate automated design of new experiments. First, we use the previously described modeling procedure to generate a library of models corresponding to different gene-regulatory systems in yeast. Many of these models contain transcriptional interactions for which the regulatory effects (inducer versus repressor) are ambiguous and cannot be determined from publicly available expression profiles. Next, to address these ambiguities we implement a score function that ranks possible genetic perturbation experiments on the basis of their projected information content over the models. We perform four of the highest-ranking perturbations experimentally and integrate the data back into the model. The new data support two out of three novel regulatory pathways predicted to mediate expression changes downstream of the yeast transcriptional regulator SWI4.
Results
Summary of physical regulatory models
We applied a previously described network-modeling procedure [18] to integrate three complementary sources of gene-regulatory information in yeast: 5,558 promoter-binding interactions for 106 transcription factors measured using chromatin immunoprecipitation followed by microarray chip hybridization (ChIP-chip) [3]; the set of all 15,116 pairwise protein-protein interactions recorded in the Database of Interacting Proteins as of April 2004 [19]; and a panel of mRNA expression profiles for 273 individual gene-deletion experiments [20]. Software for performing the network-modeling procedure is available as a plug-in to the Cytoscape package [21,22] on our supplementary website [23].
For each gene-deletion experiment, the modeling procedure identified the most probable paths of protein-protein and promoter-binding interactions that connect the deleted gene (the perturbation) to genes that were differentially expressed in response to the deletion (the effects of perturbation). Thus, a path represented one possible physical explanation by which a deleted gene regulates a second gene downstream. From the expression data, each interaction on a path was annotated with its probable direction of information flow and its probable regulatory effect as an inducer or repressor.
For example, the model in Figure 1a (top center) includes a path from GLN3 through GCN4 to a block of downstream affected genes. This model integrates evidence that: Gln3p binds the promoter of GCN4 with high significance in a ChIP-chip assay [3] (p ≤ 8 × 10-4); Gcn4p binds the promoters of many genes in the ChIP-chip assay (RIB5, YJL200C, and others in the downstream block); and a significant number of genes in the block are upregulated in a gln3Δ knockout but downregulated in a gcn4Δ knockout [20]. Together, this evidence confirms Gcn4p as an activator of downstream genes [24] and leads to a (novel) annotation that Gln3p is likely to regulate GCN4 via transcriptional repression.
In total, the modeling process generated 4,836 paths, each explaining expression changes for a particular gene in one or more knockout experiments. Of the 965 interactions covered by paths, 194 had regulatory effects that were uniquely determined by the data, while regulatory effects of the remaining 771 interactions were ambiguous. For example, Figure 1b includes ambiguous interaction paths through SWI4, SOK2, and MSN4, explaining the observation that many genes for which the promoters are bound by Msn4p are upregulated in a swi4Δ knockout. This observation can be explained by several alternative annotations: one scenario is that SWI4 activates SOK2 and SOK2 represses MSN4 (Figure 1b), whereas another is that SWI4 represses SOK2 and SOK2 activates MSN4 (Figure 1c). These regulatory annotations could be uniquely determined by measuring the expression changes of genes downstream of MSN4 in the model in response to a sok2Δ deletion and an msn4Δ deletion (see below).
Paths with ambiguous interactions were partitioned into 37 independent network models (numbered 1-37), where each model contained a distinct region of the physical network (see Materials and methods and Additional data file 1). The remaining non-ambiguous paths were grouped into a single model (Model 0). As shown in Table 1, 21 of the models (55%) contained pathways that are well documented in the literature or are significantly enriched for genes belonging to specific Munich Information Center for Protein Sequences (MIPS) [25] functional categories. Of 132 protein-DNA interactions incorporated into Model 0, we found that 50 had been confirmed in classical (low-throughput) assays as reported in the Proteome BioKnowledge Library [26]. Moreover, the inferred regulatory roles (induction or repression) for 48 out of 50 of these interactions agreed with their experimentally determined roles (96%, binomial p-value < 1.22 × 10-7). Wiring diagrams for Models 0 and 1 are given in Figure 1; diagrams for all other regulatory network models are provided in Additional data file 1 and at [23].
Experiment selection
As shown in Figure 2, we implemented an information-theoretic approach to discriminate between ambiguous model annotations using the fewest additional gene-expression experiments. All non-lethal single-gene knockout experiments were ranked by their projected information content based on the inferred models (see Materials and methods). Table 2 reports the list of top-ranking experiments. This list coincides roughly with biological intuition, in the sense that informative target genes typically encode proteins that are network 'hubs', each having a large number of regulatory interactions with downstream genes in the models. However, as discussed later, knocking out hubs only is not as effective as using the information-theoretic criteria.
Among the highest-priority experiments, Model 1 (Figure 1b) was the most often targeted, containing three of the top 10 highest-scoring genetic perturbations: sok2Δ, yap6Δ, and msn4Δ. A fourth perturbation to Model 1, hap4Δ, was also highly ranked (rank 34). Therefore, Model 1 was chosen for further experimentation.
Model validation
Knockout strains corresponding to the high-ranking perturbations sok2Δ, yap6Δ, hap4Δ, and msn4Δ were grown in quadruplicate under conditions identical to those for the initial 273 knockouts by Hughes et al. [20]. Gene-expression profiles were obtained for each knockout culture versus wild type using yeast genome microarrays. We sought to test the three regulatory cascades leading from SWI4 to SOK2 to either MSN4, HAP4, or YAP6 (Figure 1b). To verify these cascades independently of the model, we analyzed the expression patterns of gene sets known to be directly regulated by MSN4, HAP4, or YAP6 (obtained from the Proteome BioKnowledge Library [26]; see Additional data file 1). To normalize between our microarray procedures and those of Hughes et al., we also repeated the original swi4Δ expression profile, and filtered the above sets to select only those genes with expression changes that were reproducible (that is, same direction of change) between the Hughes et al. swi4Δ profile and our new profile. Expression changes were reproducible for 28 of 42 Msn4p-regulated genes, 11 of 29 Hap4p-regulated genes, and 64 of 119 Yap6p-regulated genes. Expression similarity among the genes in each filtered set was captured formally in a measure called 'coherence'; details of the computation of expression coherence and the selection of the gene sets are described further in Materials and methods and [23].
As shown in Figure 3a, the gene set downstream of MSN4 showed coherent upregulation in the swi4Δ (p ≤ 10-4) and sok2Δ (p ≤ 10-4) knockouts, but downregulation in the msn4Δ (p ≤ 8 × 10-4) knockout. This result supports the existence of a regulatory cascade leading from SWI4 to SOK2 to MSN4. Furthermore, in the context of the present regulatory cascade, MSN4 appears to be an inducer as its downstream gene set was downregulated in the msn4Δ experiment. In contrast, SOK2 appears to be a repressor of MSN4 as a sok2Δ deletion experiment upregulates the same set of genes. Finally, SWI4 appears to be an inducer of SOK2 as the swi4Δ knockout has the same effect as sok2Δ (that is, upregulation).
Results were qualitatively similar for the HAP4 pathway (Figure 3b). The gene set downstream of HAP4 was upregulated in the swi4Δ (p ≤ 10-2) and sok2Δ (p ≤ 9 × 10-4) knockouts but downregulated in hap4Δ (p ≤ 10-4). These results suggest that swi4Δ, sok2Δ, and hap4Δ deletions affect the set of genes immediately downstream of HAP4, supporting the SWI4-SOK2-HAP4 regulatory pathway hypothesis. In contrast to the MSN4 and HAP4 pathways, the gene set downstream of YAP6 had insignificant responses to all follow-up knockout experiments (Figure 3c). Thus, the existence of the SWI4-SOK2-YAP6 regulatory pathway was not supported by our validation experiments.
Automated model refinement
We used our modeling procedure to construct a new physical network model using the original 273 knockout gene-expression experiments of Hughes et al. combined with the new sok2Δ, hap4Δ, msn4Δ, and yap6Δ profiles. Overall, 60 protein-DNA interactions were disambiguated by our data: 50 interactions were resolved as definite inducers or repressors, whereas ten interactions were removed from the model because the expression of downstream genes did not change as a result of the knockout. In the updated Model 1, MSN4 and HAP4 were unambiguously annotated as inducers of downstream genes, SOK2 was annotated as a repressor of MSN4 and HAP4, and SWI4 was annotated as an inducer of SOK2 (Figure 3e). These results agree with our previous manually derived annotations (see 'Model validation' above).
Learning-curve analysis
We quantified the efficiency of our information-based approach by comparing it to two other methods of prioritizing gene knockout experiments: prioritizing hubs and prioritizing genes randomly. First, we generated a 'reference' model by fixing each ambiguous interaction in Models 1-37 to be an inducer or repressor. Assignments were chosen arbitrarily from the set of annotations that were consistent with the original knockout data. Next, we used each method (information, hub, or random) to iteratively 'learn' these assignments. In each iteration, we selected the highest-priority knockout experiment, simulated the resulting expression changes (up/down) using the reference model, updated the inferred model, and recorded the number of ambiguous interactions that were resolved. This iterative learning procedure was repeated 100 times.
As shown in Figure 4, the mutual information criterion significantly outperformed hub-based and random selection. The learning curves also provide an estimate of the number of additional experiments needed to reduce model ambiguity below a given level. For example, using the information-based score, ten knockout experiments are needed to reduce the number of ambiguous interactions by 50%. In contrast, over 25 experiments are needed according to the hub-based method. Figure 4 suggests that performing 40 additional experiments selected using the information-based score will clarify the regulatory roles of about 70% of the ambiguous interactions. The learning rate of the final 30% becomes very slow because these interactions are isolated in the physical network, unconnected to others, and thus require separate knockouts to decipher each of them.
Discussion
We have used global expression profiles to validate models of transcriptional regulation inferred from protein-protein interactions, genome-wide location analysis, and expression data. A previously described network inference algorithm [18] identifies probable paths of physical interactions connecting a gene knockout to genes that are differentially expressed as a result of that knockout. The proposed validation strategy uses information gain as a criterion for choosing optimal knockouts to profile using microarray experiments. This strategy agrees with intuition, in that optimal knockouts typically target intermediate genes along the pathways under consideration. If an intermediate gene knockout fails to affect downstream genes in a pathway, that pathway is removed from the model.
The validated pathways point to a combination of previously documented and novel findings. First, in agreement with previous literature, we confirm that MSN4 and HAP4 are inducers [27,28] and that SOK2 is a repressor [29]. For instance, SOK2 is known to act downstream of protein kinase A (PKA) to repress genes involved in stress response, glycogen storage, and pseudohyphal growth [29]. However, although SOK2 is thought to control these pathways via a transcriptional cascade, the components of this cascade have remained unclear. Here, we provide evidence for a model in which SOK2 acts as a negative regulator upstream of MSN4 and HAP4. Interestingly, MSN4 has been shown to activate stress-response genes [28], and HAP4 has been shown to activate genes involved in energy conservation and oxidative carbohydrate metabolism [27]. Thus, we have identified a candidate model for the transcriptional cascade downstream of PKA signaling that mediates stress response. This model includes two novel regulatory pathways from SWI4 to SOK2 to MSN4 and from SWI4 to SOK2 to HAP4. The validation experiments do not support the third predicted pathway from SWI4 to SOK2 to YAP6.
In model simulations, choosing new gene knockout experiments with an information-theoretic approach significantly outperformed both random and hub-based selection. It also outperformed the observed experimental results: approximately 280 interactions were disambiguated after four simulated knockouts (Figure 4), whereas only 60 interactions were resolved due to the four actual knockouts sok2Δ, hap4Δ, msn4Δ, and yap6Δ. This difference in performance stems from key differences between the simulated and actual scenarios. In simulation, the four experiments are performed independently and iteratively, selecting the absolute highest-ranking knockout each time. In the actual study, four high-ranking experiments (but not the highest) are chosen to interrogate and maximally resolve a single pathway model, resulting in experiments that are highly co-dependent and performed simultaneously without intervening rounds of inference and experimental design. In addition, the simulation assumes that all interactions in the model are correct, along with one of the initial sets of inducer/repressor annotations. It therefore isolates the process of learning regulatory role annotations, whereas the actual procedure also serves to distinguish interactions as true versus false positives. Nevertheless, the simulation provides a useful comparison of experimental design methods relative to each other.
An important limitation of the single-gene knockout approach is that single perturbations do not identify pathway intermediates for which loss of function can be compensated by another gene. Furthermore, our approach may not identify regulatory pathways in which several transcription factors independently activate gene expression. Applying knockouts in combination may prove fruitful in these cases. For instance, approximately 4,000 double knockouts have been reported in yeast that lead to synthetic lethality: that is, a lethal phenotype that is not observed in either of the single knockouts individually [30]. These interactions suggest regulatory relationships which could be incorporated into future work.
Conclusion
Scientific discovery is an iterative process of building models to explain experimental observations and validating models with new experiments [31]. Experimental design is the essential link between these two aspects. Here we have explored a framework for modeling transcriptional networks in which experimental design and validation are central features. This framework is based on computational analysis and expression microarrays, both of which are amenable to automation, suggesting a high-throughput strategy for mapping gene-regulatory pathways.
Materials and methods
Model building and inference
Physical mechanisms of transcriptional regulation were modeled using an approach described previously [18]. Briefly, we postulated that the regulatory effects of deleting a gene are propagated along paths of physical interactions (protein-protein and protein-DNA). We formalized the properties of these paths and interactions using a factor graph [32] and found the most probable set of paths using the max-product algorithm [32]. The resulting set of paths was partitioned into independent network models, also as described previously [18]. The raw data used in the modeling procedure included 5,558 promoter-binding interactions (at p-value < 0.001) for 106 transcription factors [3], the set of all 15,166 pairwise protein-protein interactions recorded in the Database of Interacting Proteins as of April 2004 [19], and mRNA expression profiles for 273 individual gene deletion experiments [20]. Expression changes with a p-value < 0.02 were considered significant.
Experiment scoring
We calculated the expected information gain for each of the 4,756 possible non-lethal single-gene deletion experiments that were not included in the set of 273 deletions used to generate our network models. Intuitively, information gain measures (the logarithm of) the number of ambiguous annotations in the model that are likely to be determined after generating a yeast-genome expression profile in response to a particular gene deletion under consideration. Each gene-deletion experiment predicts a distinct expression profile given a particular configuration of model annotations. Experiments with high information gain are those for which the predicted expression profiles are highly variable over the set of possible annotations. In these cases, only one (or at most a few) of the predicted profiles will match the true observed profile, efficiently constraining the space of possible model annotations.
The information gain discussed above arises from the expected value of information calculations in statistical decision theory [12]. Here we describe the score more directly in terms of reduction of model entropy. The entropy of a set of ambiguous model annotations is given by:
The expected information gain is the difference between the entropies before and after a hypothetical experiment:
where Ye denotes the vector of predicted expression changes for each gene in the model under experiment e. The conditional entropy H(M|Ye) requires us to consider all possible models and corresponding outcomes resulting from experiment e. Direct enumeration of all values of M and Ye is impractical; instead, we make several simplifying approximations as described at [23].
Expression profiling
Expression profiling experiments were based on the wild-type diploid BY4743 and homozygous gene knockout strains derived from this parent [33] (Invitrogen), with cultures grown identically to those of Hughes et al. [20]. Labeled cDNA from each gene knockout strain was co-hybridized versus wild type cDNA in quadruplicate two-color microarray hybridizations. Total RNA was isolated by hot acid phenol extraction, purified to mRNA (Ambion PolyAPure kits), and labeled with Cy3 or Cy5 by direct incorporation (Amersham CyScribe First-Strand cDNA Labeling Kit). DNA microarrays were spotted from the Yeast Genome Oligo Set v1.1 (Qiagen) on Corning UltraGAPS slides using an OmniGrid 100 robot (Genomic Solutions). Lyophilized Cy3- and Cy5-labeled samples were resuspended in 50 μl buffer (5× SSC, 0.1% SDS, 1× Denhardt's solution, 25% formamide) and co-hybridized at 42°C beneath a coverslip for 15 h. Arrays were imaged at 10 μm resolution using a ScanArray Lite instrument (PerkinElmer). Raw quantitated background intensities were smoothed using a 7 × 7 median filter, separately for the Cy3 and Cy5 channels, and data were corrected for cyanine-dye dependent bias using a Qspline normalization [34]. The VERA/SAM package [34] was used to assign a log-likelihood statistic λ with each gene, indicating its significance of differential expression in each experiment. Microarray expression data are deposited in the ArrayExpress database [35] under accession numbers A-MEXP-217 (Arrays) and E-MEXP-351 (Experiments).
Expression coherence
The expression coherence of a set of genes measures whether the expression levels of these genes behave similarly in a particular experiment. Each gene i in gene-deletion experiment e has an expression ratio rie (versus wild type) and associated p-value pie of differential expression. First, we filter out insignificant expression changes with a p-value > 0.5. Then, we use the inverse Gaussian cumulative distribution function, Φ-1, to convert each remaining p-value into a z-score [36,37]:
zie = Φ-1 (1 - pie)
Next, we compute a 'signed z-score' by multiplying z by +1 if the expression level is increasing and by -1 if it is decreasing. The average signed z-score for a gene subset of size N is computed as:
Gene sets with expression changes that are significant and in the same direction result in large Z-values. A distribution of Z values obtained from random gene sets of size N was used to determine a p-value for each expression coherence score.
Additional data files
Additonal data is available with the online version of this paper. Additional data file 1 contains Tables S1-S4 and wiring diagram illustrations for Models 0-44. Table S1 gives the internal validation for 17 out of 24 restricted network models; Table S2 lists the correlations between swi4δ and gcn4δ data and Rosetta and the new experiments; Table S3 gives the restricted subsets used to evaluate the reproducibility; and Table S4 gives the gene sets for external validation.
Supplementary Material
Additional File 1
Table S1 gives the internal validation for 17 out of 24 restricted network models; Table S2 lists the correlations between swi4δ and gcn4δ data and Rosetta and the new experiments; Table S3 gives the restricted subsets used to evaluate the reproducibility; and Table S4 gives the gene sets for external validation.
Click here for file
Acknowledgements
We are grateful to Owen Ozier, Ryan Kelley, and Rowan Christmas for their valuable assistance with model visualization, and to Julia Zeitlinger for commenting on the manuscript. C.M., C.W., and S.M. were supported by NIGMS grant GM070743-01 and NSF grant CCF-0425926. T.I. was supported by a David and Lucille Packard Fellowship award. C.Y. and T.J. were supported in part by NIH grant(s) GM68762 and GM69676.
Figures and Tables
Figure 1 Wiring diagrams for example network models. (a) Model 0, showing regulatory pathways that have unique functional annotations. (b,c) Model 1, showing regulatory pathways downstream of SWI4 and SOK2 with ambiguous functional annotations (several would be consistent with the observed expression responses: two possibilities are shown in (b) and (c), respectively). In the models, a connection from gene a to b represents the experimental observation that the proteins encoded by a and b physically interact in a protein-protein interaction (dotted links), or that the protein encoded by a binds the promoter of b (solid links). Each gene is either defined by an original knockout (red nodes), a differentially expressed effect (yellow nodes), or a signal transducer that was chosen for follow-up perturbation (gray nodes). Functional annotations (edge colors) are uniquely determined in (a) whereas multiple annotations are possible in (b) and (c) based on the available data. Diagram layout is performed automatically using the Cytoscape package [21].
Figure 2 Schematic of the experimental design approach. The input to the approach is a set of alternative representations of a gene-regulatory model, each of which is equally likely given current expression data. In the present work, the alternatives arise as a result of ambiguities in the regulatory roles of interactions in the model as inducers or repressors of downstream genes. Next, a scoring procedure is used to rank candidate perturbations according to their expected information gain over the model alternatives. High-ranking perturbations are applied to the system and characterized using gene-expression microarrays. The resulting expression profiles validate or invalidate particular connections in the model and reduce the set of model alternatives to those that are consistent with both old and new expression measurements.
Figure 3 Validation and refinement of Swi4 transcriptional cascades. Yeast genome microarrays were used to explore three transcriptional cascades from Model 1 involving the transcriptional regulators Swi4p, Sok2p, and either (a) Msn4p, (b) Hap4p, or (c) Yap6p. Bar charts show the expression coherence of genes regulated by Msn4p, Hap4p, or Yap6p in knockout strains swi4Δ, sok2Δ, msn4Δ, hap4Δ, and yap6Δ. Coherence scores more extreme than ± 0.7 are significant (p < 0.01, dotted lines). (d) Results are also shown for genes bound by Msn1p as representative of an unrelated model not targeted by these perturbations. This analysis provides validation for the Msn4 and Hap4 pathways and disambiguates the role of each pathway interaction as activating (Swi4 interactions) or repressing (Sok2 interactions) downstream genes (e). The Yap6 pathway hypothesis is not supported by this analysis.
Figure 4 Simulated learning curves of three experimental design methods. Three different methods of selecting experiments are compared: mutual information scores (triangles), hub selection (circles), and random selection (squares). We performed 100 simulated trials and show the average number of ambiguous interactions remaining in the inferred model after each simulated knockout experiment. Vertical bars indicate the standard deviations for the random selection method. The standard deviations for the information and hub selection curves are less than five and are not shown for clarity.
Table 1 Internal validation for 21 of the 38 inferred models
Model Number of genes Number of variants Validated literature pathway Enriched MIPS functions
0 130 1 Kss1/Fus3-Ste12 (mating response and filamentous growth) Cell fate (1.48 × 10-7); metabolism (0.0067)
1 69 8 Sok2-Msn4 (PKA pathway)
2 63 16 Tup1-Hhf1 (histone regulation) Protein synthesis (7.13 × 10-8)
3 44 2 Tup1/Ssn6-Nrg1 (glucose metabolism) Transport (1.05 x 10.5); metabolism (5.41 × 10-4)
4 58 8 Tup1/Ssn6-α2/Mcm1 (mating response) Cell fate (1.12 × 10-5);
5 58 4 Rpd3-Abf1 (histone modification)
6 44 2 Swi4-Ndd1-Ace2 (cell cycle)
7 26 4 Cell cycle (0.035)
8 36 8 Slt2-Rlm1/Swi4 (PKC pathway)
10 45 16 Med2-Gal4/Gcn4 (general transcription)
15 13 4 Cmd1-Cna1-Skn7 (calcium signaling)
19 9 4 Cell defense (6.33 × 10-6)
23 17 2 Metabolism (1.49 × 10-6);energy (0.04)
26 8 8 Cell defense (9.62 × 10-5)
29 9 4 Yap1-Cad1 (metal response)
30 12 4 Med2-Srb6-Gal4 (general transcription)
32 12 4 Med2-Gal11-Gal4 (general transcription)
33 12 4 Med2-Srb5-Gal4 (general transcription)
34 9 4 Ste12-Mcm1 (mating response) Cell fate (4.55 × 10-8); homeostasis (0.0012); cell communication (0.0345)
36 7 4 Metabolism (0.0258)
40 5 2 Metabolism (0.0017)
The number of genes and variants are shown for each model along with the results of our preliminary validations. Each variant corresponds to a distinct set of functional annotations on the interactions in the model (directions and regulatory effects, see text). For Model 0, the expression data implied a unique set of annotations; for all other models multiple sets of annotations were possible. Each model was validated if its pathways were (wholly or partially) cited in previous studies or its downstream genes were significantly enriched for MIPS functional categories (p ≤ 0.05; hypergeometric test with Bonferroni correction).
Table 2 Top-ranking knock-out experiments proposed for model discrimination
Gene Function Score Downstream genes Rank Model
HHF1 Histone 52.1429 74 1 2
SOK2* Regulator for meiosis and PKA pathway 45.0279 64 2 1
CKA1 Protein kinase of cell cycle 45.0075 64 3 5
A2 Mating response 40.9023 58 4 4
YAP6* Stress response regulator 35.1652 50 5 1, 3
NRG1 Regulator of glucose dependent genes 31.6501 45 6 3
FKH1 Regulator of cell cycle 29.1194 41 7 2
FKH2 Regulator of cell cycle 26.7131 38 8 7
SLT2 Protein kinase of cell wall integrity pathway 23.4727 31 9 8
MSN4* Regulator of stress response 21.8224 31 10 1
HAP4* Regulator of cellular respiration 6.3310 9 34 1
Each proposed target gene is reported, along with its function, mutual information score, rank, and the model(s) it informs. All target genes are non-lethal in rich media. *Gene knockouts selected in this study.
==== Refs
Hood LE Hunkapiller MW Smith LM Automated DNA sequencing and analysis of the human genome. Genomics 1987 1 201 212 3328736 10.1016/0888-7543(87)90046-2
Lockhart D Winzeler E Genomics, gene expression and DNA arrays. Nature 2000 405 827 836 10866209 10.1038/35015701
Lee TI Rinaldi NJ Robert F Odom DT Bar-Joseph Z Gerber GK Hannett NM Harbison CT Thompson CM Simon I Transcriptional regulatory networks in Saccharomyces cerevisiae. Science 2002 298 799 804 12399584 10.1126/science.1075090
Uetz P Giot L Cagney G Mansfield TA Judson RS Knight JR Lockshon D Narayan V Srinivasan M Pochart P A comprehensive analysis of protein-protein interactions in Saccharomyces cerevisiae. Nature 2000 403 623 627 10688190 10.1038/35001009
de Jong H Modeling and simulation of genetic regulatory systems: a literature review. J Comput Biol 2002 9 67 103 11911796 10.1089/10665270252833208
Eisen MB Spellman PT Brown PO Botstein D Cluster analysis and display of genome-wide expression patterns. Proc Natl Acad Sci USA 1998 95 14863 14868 9843981 10.1073/pnas.95.25.14863
Segal E Shapira M Regev A Pe'er D Botstein D Koller D Friedman N Module networks: identifying regulatory modules and their condition-specific regulators from gene expression data. Nat Genet 2003 34 166 176 12740579
Akutsu T Kuhara S Maruyama O Miyano S A system for identifying genetic networks in gene expression patterns produced by gene disruptions and overexpressions. Genome Inform Ser Workshop 1998 9 151 160
Wagner A How to reconstruct a large genetic network from n gene perturbations in fewer than n(2) easy steps. Bioinformatics 2001 17 1183 1197 11751227 10.1093/bioinformatics/17.12.1183
Ideker T Thorsson V Karp RM Discovery of regulatory interactions through perturbation: inference and experimental design. Pac Symp Biocomput 2000 305 316 10902179
Ideker T Thorsson V Ranish JA Christmas R Buhler J Eng JK Bumgarner R Goodlett DR Aebersold R Hood L Integrated genomic and proteomic analysis of a systematically perturbed metabolic network. Science 2001 292 929 934 11340206 10.1126/science.292.5518.929
Raiffa H Shlaifer R Applied Statistical Decision Theory 1962 Cambridge, MA: MIT Press
Fedorov FF Theory of Optimal Experimental Design 1972 New York: Academic Press
Tong S Koller D Active learning for parameter estimation in Bayesian networks. Proc 13th Conf Neural Information Processing 2000 Tübingen: Neural Information Processing Systems 647 563
Abadir MS Ferguson J Kirkland TE Logic design verification via test generation. IEEE Trans Computer-Aided Design 1988 7 138 148 10.1109/43.3141
Rea C Settle MA An automated test approach for US Air Force fighter engines. IEEE Aerospace Electron Syst Mag 1996 11 24 28 10.1109/62.486733
King RD Whelan KE Jones FM Reiser PG Bryant CH Muggleton SH Kell DB Oliver SG Functional genomic hypothesis generation and experimentation by a robot scientist. Nature 2004 427 247 252 14724639 10.1038/nature02236
Yeang CH Ideker T Jaakkola T Physical network models. J Comput Biol 2004 11 243 262 15285891 10.1089/1066527041410382
Deane CM Salwinski L Xenarios I Eisenberg D Protein interactions: two methods for assessment of the reliability of high throughput observations. Mol Cell Proteomics 2002 1 349 356 12118076 10.1074/mcp.M100037-MCP200
Hughes TR Marton MJ Jones AR Roberts CJ Stoughton R Armour CD Bennett HA Coffey E Dai H He YD Functional discovery via a compendium of expression profiles. Cell 2000 102 109 126 10929718 10.1016/S0092-8674(00)00015-5
Shannon P Markiel A Ozier O Baliga NS Wang JT Ramage D Amin N Schwikowski B Ideker T Cytoscape: a software environment for integrated models of biomolecular interaction networks. Genome Res 2003 13 2498 2504 14597658 10.1101/gr.1239303
Cytoscape
Cell Circuits Pathway Database
Natarajan K Meyer MR Jackson BM Slade D Roberts C Hinnebusch AG Marton MJ Transcriptional profiling shows that Gcn4p is a master regulator of gene expression during amino acid starvation in yeast. Mol Cell Biol 2001 21 4347 4368 11390663 10.1128/MCB.21.13.4347-4368.2001
Munich Information Center for Protein Sequences
Proteome BioKnowledge Library
Blom J De Mattos MJ Grivell LA Redirection of the respiro-fermentative flux distribution in Saccharomyces cerevisiae by overexpression of the transcription factor Hap4p. Appl Environ Microbiol 2000 66 1970 1973 10788368 10.1128/AEM.66.5.1970-1973.2000
Smith A Ward MP Garrett S Yeast PKA represses Msn2p/Msn4p-dependent gene expression to regulate growth, stress response and glycogen accumulation. EMBO J 1998 17 3556 3564 9649426 10.1093/emboj/17.13.3556
Ward MP Gimeno CJ Fink GR Garrett S SOK2 may regulate cyclic AMP-dependent protein kinase-stimulated growth and pseudohyphal development by repressing transcription. Mol Cell Biol 1995 15 6854 6863 8524252
Tong AH Lesage G Bader GD Ding H Xu H Xin X Young J Berriz GF Brost RL Chang M Global mapping of the yeast genetic interaction network. Science 2004 303 808 813 14764870 10.1126/science.1091317
Popper K The Logic of Scientific Discovery 1959 New York: Basic Books
Kschischang F Frey B Loeliger H Factor graphs and the sum-product algorithm. IEEE Trans Inform Theory 2001 47 498 519 10.1109/18.910572
Winzeler EA Shoemaker DD Astromoff A Liang H Anderson K Andre B Bangham R Benito R Boeke JD Bussey H Functional characterization of the S. cerevisiae genome by gene deletion and parallel analysis. Science 1999 285 901 906 10436161 10.1126/science.285.5429.901
Workman C Jensen LJ Jarmer H Berka R Gautier L Nielser HB Saxild HH Nielsen C Brunak S Knudsen S A new non-linear normalization method for reducing variability in DNA microarray experiments. Genome Biol 2002 3 research 0048.1 0048.16 12225587 10.1186/gb-2002-3-9-research0048
ArrayExpress Gene Expression Database
Ideker T Thorsson V Siegel A Hood L Testing for differentially-expressed genes by maximum likelihood analysis of microarray data. J Comput Biol 2000 7 805 817 11382363 10.1089/10665270050514945
Ideker T Ozier O Schwikowski B Siegel AF Discovering regulatory and signalling circuits in molecular interaction networks. Bioinformatics 2002 18 Suppl 1 S233 S240 12169552
| 15998451 | PMC1175993 | CC BY | 2021-01-04 16:05:40 | no | Genome Biol. 2005 Jul 1; 6(7):R62 | utf-8 | Genome Biol | 2,005 | 10.1186/gb-2005-6-7-r62 | oa_comm |
==== Front
Genome BiolGenome Biology1465-69061465-6914BioMed Central London gb-2005-6-7-r631599845210.1186/gb-2005-6-7-r63MethodGenomic analysis of heat-shock factor targets in Drosophila Birch-Machin Ian [email protected] Shan [email protected] David [email protected] Richard [email protected] Robert AH [email protected] Steven [email protected] Department of Anatomy, University of Cambridge, Downing Street, Cambridge, CB2 3EH, UK2 Department of Genetics, University of Cambridge, Downing Street, Cambridge, CB2 3EH, UK2005 10 6 2005 6 7 R63 R63 31 1 2005 7 4 2005 10 5 2005 Copyright © 2005 Birch-Machin 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.
A ChIP-assay approach is used for the whole-genome transcription factor target mapping in vivo using an intact whole organism. Using this ChIP-array 141 genes are identified in a genomic search for Hsf targets in Drosophila.
We have used a chromatin immunoprecipitation-microarray (ChIP-array) approach to investigate the in vivo targets of heat-shock factor (Hsf) in Drosophila embryos. We show that this method identifies Hsf target sites with high fidelity and resolution. Using cDNA arrays in a genomic search for Hsf targets, we identified 141 genes with highly significant ChIP enrichment. This study firmly establishes the potential of ChIP-array for whole-genome transcription factor target mapping in vivo using intact whole organisms.
==== Body
Background
Chromatin immunoprecipitation or, more correctly, immunopurification (ChIP) has emerged as a valuable approach for identifying the in vivo binding sites of transcription factors [1-6]. Before the availability of complete genome sequence the use of this approach for identifying transcription targets on a genome-wide scale was, however, limited. Over the past few years, a number of laboratories have successfully used high-density DNA microarrays to identify sequences enriched by chromatin immunopurification (the ChIP-array approach). In the yeast Saccharomyces cerevisiae, microarrays containing virtually all of the intergenic sequences from the genome have been used to identify the binding sites of a large number of transcription factors [7,8]. In principle, the same techniques can be applied to higher eukaryotes, but the complexity of their genomes presents a challenge for the construction of full genomic microarrays.
Despite such difficulties, several studies have shown the feasibility of the ChIP-array approach with small regions of complex eukaryotic genomes using tissue culture systems. In cultured mammalian cells, for example, the binding sites for several transcription factors have been mapped using microarrays composed of specific promoter regions or enriched for promoter sequences with CpG arrays [9-11]. Although such studies are valuable in identifying some of the targets of particular transcription factors, they are limited because the microarray designs restrict the analysis to proximal promoter elements of a subset of genes. It would be preferable to examine binding sites in an unbiased fashion by constructing tiling arrays composed of all possible binding targets. Such tiling arrays have been constructed on a small scale with microarrays containing a series of 1-kb fragments from the β-globin locus [12], or on a large scale with oligonucleotide arrays containing elements that detect all the unique sequences of human chromosomes 21 and 22 [13]. These studies indicate that the DNA-binding patterns of regulatory molecules in large eukaryotic genomes are complex and highlight the need for a comprehensive approach to understand how transcription factors interact with DNA in vivo.
Drosophila melanogaster, with a genome complexity intermediate between that of yeast and human, provides a powerful system for investigating transcription factor targets and regulatory networks in a complex multicellular eukaryote. Recently, the principle of using Drosophila genome tile arrays to identify transcription factor binding sites in tissue culture cells has been demonstrated. Using a technique employing fusions between DNA-binding proteins and the Escherichia coli DNA adenine methyltransferase (DamID; [14]) the binding locations for the GAGA transcription factor and the heterochromatin protein HP1 were mapped within a 3-Mb region of the Drosophila genome in a tissue culture system [15]. Other studies have used this method to map proximal binding sites with cDNA arrays [16]. While this elegant technique has the advantage that high-quality antibodies against particular transcription factors are not required, and a recent study indicates that it may be possible to transfer from a tissue culture system to the intact organism [17], it clearly has limitations, as in vivo the DAM-tagged transcription factor is not expressed in its normal developmental context. It is therefore desirable to develop methods that allow the mapping of native transcription factors in their correct in vivo context within the organism.
Here we adapt chromatin immunopurification techniques using intact Drosophila embryos and demonstrate the reliable identification of in vivo binding sites for the heat-shock transcription factor Hsf on both genome tile and cDNA arrays. The response of most organisms to heat stress involves the rapid induction of a set of heat-shock proteins (Hsps), including several chaperone molecules that assist in protecting the cell from the deleterious effects of heat [18-21]. Several direct targets of the Hsf transcription factor are already well characterized. In higher eukaryotes, including Drosophila and mammals, heat stress results in the trimerization of Hsf monomers, which then bind with high affinity to regulatory elements (heat-shock elements, HSE) close to the transcriptional start sites of Hsp genes [22,23]. The Drosophila heat-shock system has been characterized at several levels, from the cytological mapping of Hsf-binding sites on polytene chromosomes [22] to the detailed molecular and biochemical analysis of transcriptional regulation at individual Hsp genes [24-26]. In this study we extend the analysis of the Drosophila heat-shock response by demonstrating that chromatin immunopurification from embryos can accurately map in vivo Hsf-binding sites on genome tile microarrays and identify new potential in vivo HSEs. In addition, using microarrays containing full-length cDNA clones for over 5,000 Drosophila genes we identify almost 200 genes that are reproducibly bound by Hsf upon heat shock in Drosophila embryos. The targets correspond well with previously identified cytological locations of Hsf binding on salivary gland polytene chromosomes, thus providing direct target genes associated with the low-resolution cytological analysis. A comparison with studies using S. cerevisiae Hsf [27,28] suggest that a set of conserved genes are regulated by Hsf in both organisms. Overall, this study presents the strong potential of this approach for in vivo genome-wide mapping of transcription factor binding sites in higher eukaryotes using the whole organism.
Results and discussion
Immunopurification of Hsf-bound chromatin
To test the effectiveness of ChIP-array and assess the possibility of using genome tile arrays to map the in vivo location of transcription factor binding sites with intact whole organisms, we used the well characterized transcription factor Hsf, the mediator of the heat-shock response in Drosophila. Formaldehyde-crosslinked chromatin from Drosophila embryos was used as the input for immunopurifications with either anti-Hsf antisera or preimmune sera. After immunopurification and washing, the formaldehyde crosslinks were reversed by heating and the DNA purified. This DNA was initially analyzed for the enrichment of known Hsf targets by quantitative real-time PCR assays using a series of specific primers. We assayed the Hsp26 and Hsp70A genes with primers that amplify fragments spanning either the 5' HSE or a control 3' untranslated region (UTR) fragment of each gene. As shown in Table 1, the chromatin immunopurification shows both good enrichment and high specificity. With both Hsp26 and Hsp70A we observe over 100-fold enrichment of HSE fragments with anti-Hsf versus preimmune serum and a similar enrichment of HSE versus 3' ends with the anti-Hsf sera.
Because many of the published ChIP-array studies employ a ligation-mediated PCR step (LM-PCR) to amplify the enriched DNA, we assayed whether LM-PCR amplification of the DNA prepared from anti-Hsf immunopurifications maintained the enrichments we observe with unamplified material. We find that the enrichment of Hsp gene HSEs, as measured by quantitative PCR, is similar between amplified and unamplified material, demonstrating, at least with respect to the Hsp genes we examined, the validity of using LM-PCR amplification of ChIP-enriched DNA (data not shown). During the course of our experiments we tested embryos that had not been subjected to a heat shock but were processed in the same way as heat-shocked embryos. We found significant enrichment by quantitative real-time PCR (between 25- and 90-fold enrichment of HSEs in three independent experiments). Because considerable evidence indicates that Hsf is not specifically bound to HSEs in unstressed Drosophila cells [20], our observation suggests that the preparation of the embryos may have induced the stress response, possibly during the dechorionation step in bleach.
Genome tile arrays
We assayed the effectiveness of using genome tile arrays to identify in vivo Hsf-binding sites. We constructed microarrays containing a total of 3,444 PCR products. These include 3,092 fragments representing 2.9 Mb of chromosome arm 2L, from kuzbanian to cactus, 96 fragments representing the regulatory regions for a set of early segmentation genes (even-skipped, hairy, runt and Dichaete) and a set of 95 products spanning fragments identified in a previous immunopurification experiment with anti-Ubx [2]. The fragments ranged in size from 282 to 1,380 bp with an average size of 930 bp (SD ± 53 bp). In addition to these we produced 162 fragments encompassing five different Hsp gene loci; regions of approximately 10 kb encompassing Hsp68 at 95D11, Hsp83 at 63B11, Hsp60 at 10A and Hsp70A at 87A2 along with a 22-kb region from 67B1 containing Hsp67Bc, Hsp67Ba, CG32041, Hsp23, Hsp26 and Hsp27. The Hsp gene regions were represented in two fragment sets: a set of 1-kb fragments overlapping by 500 bp and a set of 2-kb fragments overlapping by 1 kb. Finally, 480 elements were spotted with sheared Drosophila DNA to give a microarray containing 3,924 elements.
We prepared chromatin from heat-shocked embryos, performed immunopurification in parallel with anti-Hsf and preimmune sera and amplified the resulting purified DNA by LM-PCR. Each sample was independently labeled with a fluorescent dye, the labeled anti-Hsf and preimmune samples were mixed and then co-hybridized to the tiling path microarrays. We performed dye-swap experiments to assess any bias in the incorporation of the fluorescent dyes. We used three independent biological replicates and for each preparation performed technical replicates, in total carrying out 11 separate hybridizations (see Additional data file 1 for the full data).
After normalization, we calculated the ratio of anti-Hsf signal to the preimmune signal. Ratios for each technical replicate were averaged and the average ratios used to calculate a probability score for each spot using Cyber-T [29]. The 480 sheared genomic DNA fragments were distributed evenly across the slide and allowed us to evaluate the consistency of input DNA samples; these had an average asinh ratio of -0.13 ± 0.09 (standard error = 0.004, variance = 0.009) indicating no significant overall difference between the samples. Of the 3,444 elements containing PCR-amplified fragments of Drosophila DNA, 59 showed a greater than 1.6-fold enrichment (up to 10-fold enrichment) with the DNA purified with anti-Hsf sera at p-values better than 10-3. Of these elements, 53 (88%) correspond to fragments from Hsp gene loci, five from the Adh region and one from the putative Ubx target set. Plotting the average ratio for each array element with respect to the order of the fragments on the genome (Figure 1), we observe a striking distribution of signal; the fragments derived from the Adh region and the segmentation genes show little signal above asinh ratios of 0.5, with only four fragments showing more than twofold enrichment. In contrast, many fragments from the Hsp gene regions show substantial enrichment. Of the 162 fragments from the Hsp gene loci, 46 show greater than twofold enrichment with the anti-Hsf sample. The results are highly reproducible; comparing the ratios obtained with the 162 Hsp fragments from each of the replicate slides, the correlation between any two slides ranged from 0.7 to 0.98, with an average correlation of 0.84.
The distribution of the signals across the Hsp genes shows excellent agreement with the known location of HSEs at the 5' end of the transcription units and, in addition, show a monotonic signal distribution centered on the fragments containing HSEs. This is best exemplified by the 20-kb region, which encompasses the eight known or putative Hsp genes in the 67B region (Hsp67Bc, the bicistronic CG32041, CG4461, Hsp26, Hsp67Ba, Hsp23 and Hsp27) where we observe strong enrichment of fragments close to the 5' ends of heat-inducible genes and negligible signals in between (Figure 2). Five clear peaks of fragment enrichment are observed and there is good overlap with the known locations of Hsf-binding sites [30]. A major peak 5' to Hsp26 encompasses the characterized Hsf-binding sites at -349 and -56. Three further peaks cover the regions of the 5' ends of Hsp67Ba, Hsp23 and Hsp27, including the known HSEs upstream of Hsp23 (-391 and -119) and Hsp27 (-366, -328 and -270). Finally, a fifth peak overlaps the 5' ends of the divergent transcription units of Hsp67Bc and CG32041, the latter being a dicistronic gene encoding Hsp22 and Hsp67Bb. There appears to be no substantial enrichment covering the 5' end of the Hsp20-like CG4461; however, it is not known if this gene is Hsf-inducible. Thus seven out of the eight Hsp genes in the region have 5' regions enriched by our assay. Fragments including known HSEs show the highest enrichments (more than 3.5-fold), whereas nearby fragments show no significant signal over the background. This region demonstrates the potential for high-resolution mapping of in vivo DNA binding and suggests that even gene-dense regions can be accurately mapped using the ChIP-array technique with 1-kb tiling paths.
The other Hsp gene loci show similar distributions of fragment enrichment (Figure 3). With Hsp70, three fragments show greater than twofold enrichment with the two fragments (Hsp-130 and Hsp-114) encompassing the known Hsp70A regulatory elements, several HSEs between -252 and -46 bp [30], showing the greatest enrichment (Figure 3a). In the case of Hsp83 we see a different organization, and Hsf binding is not restricted to the immediate 5' region (Figure 3b). We observe two strong peaks of signal enrichment. One centers on the area immediately 5' to the start of Hsp83 expression where HSEs have been mapped between -88 and -49 [30]. However, the ChIP also reveals a second peak at the 3' of Hsp83 extending to cover CG14966 (a gene of unknown function) and 3' to CG32276, a predicted chaperone. This additional signal contains matches with an Hsf consensus binding sequence, suggesting that it represents a bona fide Hsf-binding site. It has previously been noted that Hsp83 stands out from other Hsp genes in the dynamics of its response to heat shock [24] and this may be linked to the distinct arrangement of Hsf-binding sites we find.
With Hsp68 we find that two overlapping fragments show greater than fourfold enrichment (Hsp-117 and Hsp-131) and these correspond to the region immediately 5' to the start of Hsp68 transcription; the fragments flanking these are also detected with lower ratios (Figure 3c). Although there are no reports of mapping Hsf-binding sites in the Hsp68 region, we find three perfect matches to a consensus Hsf-binding site 160 bp upstream of the mRNA start site, consistent with the fragment enrichment we observe. Finally, with the Hsp60 gene we observe moderate but clear enrichment with fragments encompassing the first intron of the gene, and also find a match to a consensus HSE sequence in this region (Figure 3d, see below). Hsp60 is reported not to be induced by heat shock in Drosophila and previous studies have failed to find HSE sequences 5' to the start of Hsp60 transcription [31]. In mammals and yeast, however, Hsp60 homologs are heat inducible [32,33] and our data indicate conservation of Hsf binding.
As well as the Hsp genes, we observe a greater than twofold enrichment with two fragments in the Adh region (Figure 1). One fragment maps between the divergently transcribed genes l(2)35Bg and Su(H) suggesting that either of these genes could be regulated by Hsf. Supporting this suggestion, we find that l(2)35Bg gives a strong positive signal when independent anti-Hsf immunopurifications are used to interrogate the cDNA arrays described below. In the second case, we observe a twofold enrichment of a fragment overlapping the 5' end of the longest transcript from the PRL-1 gene and we also observe a weak enrichment (1.2-fold) of a fragment overlapping a second transcription start-site 5 kb downstream (data not shown). Interestingly, the PRL-1 gene was identified by Sun et al. [15] as a candidate GAGA-factor (Gaf)-regulated gene in their DamID analysis of the Adh region. In some cases, most notably Hsp70A and Hsp26, Hsf- and Gaf-binding sites are located in close proximity and are both involved in transcriptional regulation of Hsp genes [34].
In addition to the fragments showing greater than twofold enrichment, we find a further eight fragments showing greater than 1.5-fold enrichment with the anti-Hsf immunopurification. Some of these may represent weak Hsf-binding sites. For two of these regions (CG4500 and CG3793) we detect enrichment in the experiments with the cDNA arrays described below, suggesting that they may represent bona fide Hsf-binding sites in the genome.
To try and assess the validity of the fragments identified on the array and relate the degree of enrichment with the presence of HSE, we used the informatics tool MEME [35] to examine the enriched fragments for the presence of consensus Hsf-binding sites. As noted above, we find predicted Hsf-binding sequences in the regions enriched upstream of Hsp68, downstream of Hsp83 and in the intron of Hsp60. We also find potential Hsf-binding sequences within the fragments enriched from the Adh -region, indicating that enrichment on the tiling arrays corresponds to the location of some Hsf-binding sites. Taken together, the experiments and analysis described above demonstrate that chromatin immunopurification used in tandem with tiling DNA microarrays can successfully identify genuine in vivo transcription factor binding sites at the level of the whole organism. Our mapping suggests locations for new HSE elements regulating Hsp83, Hsp68 and Hsp60.
Genome-wide search for HSF target genes
Since much previous work, along with the observations presented above, indicates that the binding sites for Hsf tend to be located close to the transcriptional start of responsive genes [24], we reasoned that we could identify new genes with Hsf-binding sites by performing a ChIP-array analysis using arrays containing cDNA clones. To this end we utilized a microarray containing 5,372 full-length cDNA clones representing 5,073 genes, prepared from the Drosophila Gene Collection V1.0 [36]. We performed immunopurifications using anti-Hsf and preimmune sera on chromatin isolated from three independent biological preparations. In addition, to assess reproducibility, we performed independent immunopurification reactions with two of the chromatin preparations. With chromatin A we performed four separate immunopurifications (1-4); the first two of these were technically replicated as well as dye-swapped and the second two were dye-swapped only. From chromatin B we performed two independent immunopurifications and each of these were dye-swapped. With chromatin C we performed a single immunopurification and dye-swap (full data in Additional data file 2). In total we performed 18 hybridizations to the cDNA arrays. The average correlation between each technical replicate was very high (> 0.85) and after generating an average ratio for each technical replicate we used the CyberT algorithm to generate p-values from the average ratios for each independent immunopurification.
We identified 188 genes that showed greater than 1.6-fold enrichment. While we recognize that defining an enrichment cutoff in the absence of other data is somewhat arbitrary, we selected a 1.6-fold value based on the enrichments observed on the genome tiling arrays with known Hsf-binding sites. We note however that this criterion may underestimate the Hsf-binding targets as the cDNA array elements will only detect binding sites close to the 5' end of the cDNA. Genes that bind Hsf at more distant sites will be expected to generate weaker signals on the array that will escape detection owing to noise issues with low signals. To validate the Hsf targets we selected 11 genes distributed across the ranking from 1 to 188, and tested for enrichment of the 5' genomic DNA upstream of each gene in a standard ChIP assay along with 5' and 3' end of hsp26 as a control. As shown in Figure 4, all 11 genes tested showed clear enrichment when DNA derived from anti-Hsf sera and preimmune sera are compared. Thus the microarray assay is in excellent agreement with standard PCR assays and suggests that, at least with the enrichments we observe, the ChIP-array data is highly reliable. Of the 188 genes with the selected 1.6-fold enrichment, 141 were enriched with p-values of 9 × 10-3 or better. Enrichments as high as eightfold were reproducibly observed and, reassuringly, enriched genes include a number of Hsp genes along with other predicted chaperone-encoding genes such as DnaJ-1, CG32041 and CG32649 (Table 2). Using the stringent p-value cutoff, our analysis indicates that approximately 3% of the genes in the Drosophila genome (around 400) may be direct targets of Hsf, a figure that is in remarkable agreement with a recent analysis of Hsf binding in S. cerevisiae [28].
In general, the agreement between the independent immunopurifications and the different chromatin samples was very good, however we noticed that each immunopurification identified a set of genes that showed no significant enrichment in other samples. These 'IP-specific' signals were consistent within the technical replicates and showed high enrichments (up to sevenfold). They did not, however, correlate with a particular chromatin preparation, since there was no similarity between the different immunopurifications performed from the same chromatin. We assume that these artifacts reflect the inherent noisiness of the system and emphasize the need to perform replicate immunopurifications from particular biological samples in order to identify consistently positive signals.
We determined the predicted cytological location of the all 188 top Hsf target genes and compared this list to the cytological mapping of Hsf-binding sites on polytene chromosomes, which is, of course, quite low resolution [22]. Of these genes, 82 are predicted to map to the same cytological band as an Hsf site (50%) and a further 40 are predicted to map within a lettered division of a site mapped by Westwood et al. [22] (Figure 5). Thus from the 164 cytological sites reported to bind Hsf immediately after heat shock, we have identified 122 (75%) candidate genes as Hsf targets in these locations with our survey of approximately 40% of the predicted genes in the genome.
We examined the expression of the cDNAs on the array by hybridizing with labeled cDNA prepared from heat-shocked embryos compared to unshocked controls; 16 of the top 188 genes showed induction greater than 1.7-fold (Table 2) with known heat-shock response genes being robustly induced; for example, over 30-fold increases in Hsp26 and Hsp27 expression. A further two genes are repressed more than twofold. We examined the only other reported Drosophila array data, obtained from custom oligonucleotide arrays hybridized with RNA derived from heat-shocked and non-heat-shocked embryos [37]. Of the genes represented on the custom array, 21 are found in our top 188 Hsf-binding genes; of these, seven genes (Hsp26, 27 and 23, DnaJ-1, Hsc70-5, CG3488 and Cct-gamma) show induction and one (cyclophilin 1; Cyp1) is repressed, according to the quality criteria used by the authors. In general the data are in reasonable agreement; however, we find no evidence with our cDNA array for induction of Cct-gamma and CG3488 or repression of Cyp1. These discrepancies may reflect strain differences, platform-specific signals or experimental noise. We conclude that only a minority of the Hsf targets that we have identified show clear evidence of direct induction or repression using our heat-shock regimes and sampling times.
In a recent Hsf1 ChIP study of mammalian cell lines, approximately 50% of the 94 identified Hsf1-bound promoters did not directly produce heat-induced transcripts [38], leading to the interpretation that Hsf binding alone may not confer heat-inducibility. Indeed it is clear that even in the well characterized Hsp gene regulatory regions, Hsf collaborates with other transcription factors [39]. In contrast, Hahn et al. [28] were able to use the extensive expression data available in yeast to determine what fraction of the 165 Hsf targets they identified by ChIP showed evidence of induction by heat shock. Only 7% of the putative Hsf targets did not show evidence of heat-shock induction. In multicellular eukaryotes, with the possibilities of considerable developmental and tissue-specific effects on gene expression, more extensive expression analyses will be required to enable us to address the question of how many of the Hsf target sites are associated with Hsf-mediated regulation of expression.
We used the Gene Ontology (GO) annotation to classify the gene products represented by the 188 Hsf-bound genes (Figure 6). As would be predicted, proteins annotated with chaperone or chaperone ATPase activity are well represented; we find 17 chaperones among the Hsf target genes. Using GeneMerge to assess enrichment of GO terms in the Hsf targets compared to all of the genes on the array, we find highly significant enrichment of genes with chaperone or heat-shock protein activity (p < 8 × 10-6) functional annotation. In terms of biological processes, response to heat or temperature are over-represented (p < 2 × 10-4) (Figure 5). In addition, we find 18 genes involved in basic metabolism, in protein modification or degradation, 12 genes associated with the cell cycle or programmed cell death and, interestingly, 14 genes associated with gene expression. Of this latter class, eight are documented as showing changes in expression in response to dietary changes or oxidative stress [40,41] and this suggests a link between downstream components of different stress responses. Of particular interest are four genes (Taf7, CG33097, TfIIEα and Trap36) that encode core components of the RNA polymerase II transcription machinery. Trap36 is a component of the Mediator complex, which has been shown to play a vital role in transcriptional induction by Hsf at the Hsp70A promoter [42]. These data suggest that part of Hsf function may be to regulate components of the core transcriptional machinery necessary for the stress response in order to modulate or temporally control the response.
As noted above, in some cases heat-shock responsive genes may be regulated by both Hsf and Gaf. A recent study identified potential binding targets of Gaf by the Dam-ID technique using cDNA arrays very similar to those used here [16]. We therefore examined the overlap between the sets of genes binding both factors. Of the 188 Hsf-binding genes, 39 were identified as being potential Gaf targets (>1.4-fold enrichment p < 10-3, Table 2). Of these we find, as expected, the chaperones Hsp22, Hsp23, Hsp26, Hsp27 and DnaJ-1. There is no obvious correlation between high expression and binding of both Hsf and Gaf. Although the highly expressed chaperones discussed above appear to be targets of both Hsf and Gaf, four other chaperones (CG7945, Hsc70Cb, Hsc70-5 and CG32649), which are induced by heat shock, bind only Hsf and not Gaf. Of interest in the set of genes bound by both factors is the TGFβ receptor thick veins, as well as three annotated transcriptional regulators (Taf7, CG6792 and GATAd). This suggests that a complex secondary response to stress may involve co-regulation of key transcriptional and signaling regulators by both Hsf and Gaf.
We next sought to determine whether the sequences upstream of the top Hsf-binding genes were enriched for potential Hsf-binding sites. We used standard pattern matching software to look for matches to a consensus Hsf-binding site TTCnnGAAnnTTC [43] in the 1 kb immediately upstream of the top-ranked 188 Hsf-binding genes. As a control we examined the 1-kb regions upstream of the 5,000 genes on the array that showed no enrichment with Hsf. Plotting the number of predicted Hsf sites against the number of genes shows that for both the anti-Hsf enriched and the non-enriched sequences there is a broadly similar distribution for upstream regions containing five or fewer matches to the consensus (Figure 7a). However, in the case of the anti-Hsf enriched fragments we find an over-representation of upstream regions that contain six or more consensus Hsf sites. These include, as expected, the known heat-shock genes (Hsp23, Hsp26 and Hsp27) but also genes for transcription factors (TfIIEα and CG6197) and genes of unknown function. In most of these cases we find that predicted Hsf sites are clustered and preferentially located within 500 bp upstream of the transcription start (Figure 7b). This supports the view that the sites we have identified represent genuine HSEs. We also observe that the number of predicted Hsf sites is not related to the fold enrichment we observe on the microarrays, suggesting either that fragment enrichment is not an accurate measure of Hsf 'binding affinity' or that simple binding site prediction is not a reliable way of identifying genuine HSEs.
Comparative analysis
Two genome-wide studies in the budding yeast S. cerevisiae have mapped the location of HSF1 by ChIP-array. In one case, Hsf binding was determined using unstressed cultures and over 100 potential targets identified with significant p-values according to the error model used by the authors [27]. In a second study, using both unstressed and heat shocked cells, 165 genes were identified with Hsf1-binding sites that showed enrichment above a threshold set by consideration of heat-inducible expression [28]. We compared our data from the Drosophila cDNA array with this yeast data to look for similarities in the sets of genes potentially regulated by Hsf in both organisms. Taking the protein sequences of the top hits from the cDNA array, we looked for yeast genes encoding proteins with BLAST matches better than 1e-10 and identified 83 genes. We then examined their enrichment in the yeast Hsf-binding datasets. These data are summarized in Table 3. Using the cutoff criteria employed by Hahn et al. [28] we find 11 yeast genes that are predicted to bind Hsf and another gene (CG4800) just below their threshold. A further 11 genes, identified in the Lee et al. data [27] with p-values better than 1 × 10-2 are also conserved. This set of 23 Hsf target genes conserved between fly and yeast not only includes characterized heat-shock response genes (DnaJ-1 and Hsc70-5) but also seven other putative stress-response genes (including Hsp60), 12 genes with other functions and two genes of unknown function. This clearly represents a minimal set as it is limited by identification of homologous genes and by cross-comparability of the datasets. For example, the small Hsp20-like chaperones are not conserved in sequence between fly and yeast, although proteins with similar functions are clearly bound by Hsf in both organisms. Since we have only surveyed approximately 40% of the Drosophila genome, it suggests a minimal core of over 50 genes as a conserved set of eukaryotic Hsf targets.
Along with chaperone-encoding genes, we find other genes whose products are suggested to be implicated in stress responses; CG4800, a putative microtubule-binding protein associated with the defense response and cyclophilin1. CG1416 encodes a protein with a possible Hsp90 interaction domain, which, according to the data from a Drosophila gene-expression time course [44], is coexpressed with two genes: foraging, encoding a cyclic-nucleotide dependent protein kinase [45], and effete, a predicted ubiquitin-conjugating enzyme [46]. We find that both these genes are bound by Hsf in Drosophila, albeit with lower enrichments than CG1416 (1.5- and 1.6-fold) and homologous genes are also bound in yeast (the protein kinase TPK2, with a modest 2.8-fold enrichment and UBC4, a ubiquitin-conjugating enzyme with a highly significant enrichment (p = 1.1e-4). This suggests the possibility that these proteins may interact in a common stress-response pathway.
Among the remaining genes, l(2)35Bg represents a highly conserved protein found throughout eukaryotes. While the function of this protein is unknown, mutations in yeast and Drosophila are lethal, in the latter case lethal in embryos. Our findings suggest that l(2)35Bg encodes a conserved factor involved in the stress response.
Of particular interest among the conserved Hsf targets is the helix-turn-helix containing transcription coactivator multiprotein bridging factor 1 (Mbf-1). This protein has been shown to mediate the interaction between nuclear hormone receptors and TATA-binding protein (TBP) in both Drosophila and mammalian systems [47,48] and plays a similar role in yeast, where it is involved in mediating the interaction between TBP and the leucine-zipper transcription factor GCN4. Null mutants in yeast are viable but sensitive to amino-acid deprivation [49]. In Drosophila the gene is strongly induced by oxidative stress (paraquat treatment [40]), moderately induced by heat shock (this paper) and repressed under starvation conditions [41]. Recent reports suggest that mbf-1 mutants are also viable in Drosophila but are sensitive to oxidative stress [50]. This report further suggests that Mbf-1 interacts with the c-Jun/c-Fos AP-1 dimer to mediate AP-1 stress-response activity. These observations suggest that there may be an underlying link between different types of stress response (heat, oxidation and nutritional) and that Mbf-1 may be intimately involved in the transcriptional response to environmental conditions, playing a vital role in coordinating the interaction of different stress-response transcription factors with the core RNA polymerase II complex.
Conclusions
We have used chromatin immunopurification in conjunction with genome tiling and cDNA microarrays to map the in vivo binding sites of the heat-shock factor Hsf. Our results demonstrate the potential for mapping bona fide transcription factor binding sites at a genome-wide scale in complex multicellular eukaryotes. We find that the technique is highly reproducible and, with appropriate experimental replication, can identify binding regions with high fidelity. We further demonstrate a core set of Hsf targets conserved between fly and yeast that may represent a evolutionarily conserved regulatory network. The response of an organism or cell to stress is highly complex and necessitates direct control of physiological processes as well as modulation of gene transcription. The set of Hsf targets we identify includes many metabolic enzymes, which may be candidates for control points directly controlling metabolic and physiological processes in times of stress. The finding that several genes encoding transcriptional regulators are bound by Hsf, in particular components of the core RNA polymerase complex, suggests that one of the roles of Hsf may be in initiating or establishing a transcriptional state necessary for recovery from heat stress as well as its more traditional role in activating immediate stress-response genes. In both flies and mammalian systems Hsf target genes are not all immediately transcriptionally induced, suggesting that the heat response may be more complex than simply activating chaperones. In addition, the observation that Hsf may be regulating genes implicated in other stress responses suggests that responses to different stresses may involve underlying similarities. The extension of these studies to full genome coverage in Drosophila as well as other tractable model systems such as Caenorhabditis elegans, offers the prospect of understanding the regulatory response underpinning a fundamental cellular process.
Materials and methods
Anti-Hsf antiserum
We generated specific rabbit polyclonal antisera against a bacterially expressed Drosophila Hsf (CG5748). Briefly, we used a construct (MBP-dHsf [25], kindly provided by J. Lis, Cornell University) to produce a fusion protein containing the first 691 amino acids of Hsf fused to maltose-binding protein. After excision from an SDS-polyacrylamide gel, the gel slice containing the fusion protein was used as antigen in rabbits (approx 100 μg per rabbit per immunization) to produce a high titer antiserum (Eurogentec, Seraing, Belgium). The specificity of the antiserum for Hsf was confirmed by western blots of Drosophila nuclear extracts, where a band of approximately 110 kDa is recognized, as expected for Drosophila Hsf [22]. In addition, immunolabeling of Drosophila embryos with the anti-Hsf antiserum gives the expected ubiquitous nuclear staining, which is absent from embryos labeled with the preimmune serum and from hsf-null embryos labeled with the anti-Hsf antiserum (two Hsf null conditions were tested; hsf1and Df(2R)ED3610 homozygotes).
Chromatin immunopurification from Drosophila embryos
Embryos (1-2 g) were collected over a 16 h period and then heat shocked for 15 min at 37°C. After the embryos were dechorionated in weak bleach (5% w/w available chlorine) for 3 min they were washed in H2O and then in PBS/0.01% Triton (PBST). The embryos were then centrifuged (1 min at 500 g) and resuspended in 10 ml crosslinking solution (50 mM HEPES pH 8.0, 1 mM EDTA.Na2, 0.5 mM EGTA, 100 mM NaCl) containing formaldehyde (1.95%) and 30 ml n-heptane. This was incubated at room temperature with vigorous shaking for 15 min. The fixed embryos were centrifuged (1 min at 500 g), resuspended in PBST-glycine (PBST, 125 mM glycine) and allowed to sediment. After the embryos were washed with ice-cold PBST, they were again allowed to sediment. The supernatant was removed and the embryos were resuspended in 15 ml ice-cold PBST containing protease inhibitors. After douncing using a Wheaton Dounce Tissue Grinder (pestle B), and centrifugation at 400 g at 4°C for 1 min, the supernatant was removed and centrifuged at 1,100 g for 10 min at 4°C. The pellet was resuspended in 15 ml ice-cold cell lysis buffer (5 mM PIPES pH 8, 85 mM KCl, 0.5% Nonidet P-40) containing protease inhibitors and dounced using a Wheaton Dounce Tissue Grinder (pestle A). Then the extract was centrifuged at 2,000 g for 4 min at 4°C and the pelleted nuclei were resuspended in 3 ml ice-cold nuclear lysis buffer (50 mM Tris.HCl pH 8.1, 10 mM EDTA.Na2, 1% SDS) including protease inhibitors. After 20 min at 4°C, 0.3 g acid-washed glass beads (Sigma, 212-300 μm diameter) were added and the extract was sonicated using a heat systems ultrasonic liquid processor XL sonicator with a microtip attached. The extract was exposed to a 1 × 30 sec burst at level 3, and 5 × 30 sec bursts at level 4 with 90 sec resting on ice between bursts. Fragment sizes between 0.5 and 1 kb were produced. The chromatin extract was clarified by centrifugation at 14,000 rpm for 10 min at 4°C, flash frozen in liquid nitrogen and stored at -80°C.
Chromatin immunopurification was performed according to the method of Oberley et al., [51]. Briefly, the chromatin solution was diluted with IP dilution buffer (16.7 mM Tris.HCl pH 8, 167 mM NaCl, 1.2 mM EDTA.Na2, 1.1% Triton X-100, 0.01% SDS) and precleared with fixed and killed Staphylococcus aureus Protein A-positive strain cells (SAC) for 15 min. The precleared diluted chromatin sample was incubated with 1 μl of either preimmune serum or anti-Hsf serum overnight at 4°C. To capture the antibody-chromatin complexes, SAC were added and the samples were incubated for 15 min at room temperature. The SAC were washed twice in IP dialysis buffer (50 mM Tris.HCl pH 8, 2 mM EDTA.Na2, 0.2% sarkosyl) and four times in IP wash buffer (100 mM Tris.HCl pH 9, 500 mM LiCl, 1% deoxycholic acid, 1% Nonidet P-40). The immunopurified material was eluted from the SAC by vigorously vortexing for 15 min in elution buffer (50 mM NaHCO3, 1% SDS). RNase A was then added (33.3 μg/ml) and NaCl to 0.3 M. To reverse the crosslinks the material was incubated for 5 h at 67°C and then precipitated with ethanol. Proteinase K (0.6 units/ml) was added and the samples were incubated at 45°C for 2 h and purified with one extraction with phenol/chloroform/isoamyl alcohol followed by a chloroform extraction. After precipitation with ethanol in the presence of glycogen the DNA was resuspended in TE buffer.
Quantitive real-time PCR
Quantitive real-time PCR experiments were performed with a Corbett Research RotorGene utilizing SYBR Green fluorescence. Reactions were carried out in 15 μl using SYBR Green PCR master mix according to the manufacturer's protocol (Qiagen) with 2.4 μl DNA. Cycling was for 15 min at 95°C, followed by 40 cycles of 94°C, 60 sec; 60°C, 30 sec and 72°C, 60 sec. The primer pairs to amplify heat-shock element and 3' ends of the genes for heat-shock proteins 26 (3' UTR) and 70 (5' HSE and 3' UTR) were as described in Andrulis et al. [25] except the primer pair (5'-GCTGTTTCTTTTGCGCTCTT and 5'-TTGTTTGACTTGTAAGCAAAGGTT) for the heat-shock element of heat-shock protein 26 (5' HSE). Serial dilutions of genomic DNA (100-0.3125 pg/μl) were used to produce a calibration curve. A no-template control was also used. All samples, controls and standards were performed in triplicate.
Standard PCR
Positives from the cDNA microarray were validated in standard PCR assays as follows: To 3 μl of immunopurified DNA, 1 μl of 100 pmol/μl primers, 1.5 μl 10x buffer IV (Abgene), 1.5 μl 10 mM dNTPs, 1.2 μl 25 mM MgCl2, 1 μl Thermo-Start Taq DNA polymerase (Abgene, 5 units/μl) and H2O to 15 μl were added. PCR reactions were carried out for 5 min at 95°C, 35 cycles of 1 min at 95°C, 1 min at 57°C and 1 min at 72°C, followed by 10 min at 72°C. The primers used are listed in Table 4.
Sample labeling
Concentrations of anti-Hsf and preimmune IP DNA samples were determined using a NanoDrop spectrophotometer (Nanodrop Technologies). Fifty nanograms of each IP sample was incubated with 1 unit T4 DNA polymerase (Promega) in a total volume of 50 μl manufacturer's buffer for 5 min at 37°C. The reaction was stopped by adding 2 μl of 0.5 M EDTA and the DNA purified with MinElute PCR purification columns (Qiagen). Ten nanograms of purified DNA was combined with 1 μM of annealed linkers (Linker 1, 5'AGAAGCTTGAATTCGAGCAGTCAG3': Linker 2, 5'CTGCTCGAATTCAAGCTTCT 3') and incubated overnight at 4°C with 1 unit of DNA ligase (Invitrogen) in standard ligase buffer.
PCR amplification was carried out directly without further DNA purification in a reaction volume of 100 μl containing 0.2 mM dNTPs, 15 mM MgCl2, 5 U Thermo-Start DNA polymerase (Abgene) and 100 pg of linker 2 using the following conditions: 1 cycle of 55°C 2 min, 72°C 5 min, 94°C 5 min; 24 cycles of 94°C 1 min, 55°C 1 min, 72°C 1 min; 1 cycle of 72°C 5 min, 4°C hold. PCR products were purified with MinElute columns (Qiagen).
Labeling
Purified PCR products were labeled with a Bioprime random priming labeling system with 0.1 mM each dATP, dGTP and dTTP, 0.04 mM dCTP and 0.06 mM Cy3 or Cy5-conjugated dCTP (Amersham Biosciences) at 37°C for 2 h. 5% of the reaction was checked by agarose gel electrophoresis for an expected smear of product from 200-600 bp and the remainder purified with Sephadex G50 minicolumns.
Tiling path microarrays
A total of 3,091 fragments of 1 kb average length were amplified with primers designed across the Adh region by PCR (coordinates chr2L:13488459-16409825; these primers were generously donated by P. Spellman and G. Rubin, University of California at Berkeley). All sequence coordinates are from release 3.1 of the Drosophila genome sequence [52]. The primer design was generated against release 1 of the genome sequence and we have remapped the fragments onto release 3.1 of the sequence using the UCSC genome browser [53,54]. In addition we synthesized three sets of primers to amplify the following loci. The first was a set of 1-kb and 2-kb overlapping fragments covering several Hsp gene loci: Hsp70A (chr3R:7,776,000-7,830,000), Hsp83 (chr3L:3,170,043-3,180,013), Hsp67Ba-Hsp27 (chr3L:9,326,084-9,348,399), Hsp68 (chr3R:19,868,471-19,878,621) and Hsp60 (chrX:10,847,030-10,857,109). The second was a set of 1-kb fragments covering a set of segmentation genes: Eve (chr2R:5,035,032-5,051,166), Dichaete (chr3L:14,096,994-14,125,069), Hairy (chr3L:8,619,968-8,637,146) and Runt (chrX:20,349,976-20,380,172). The third was a set of 81 1-kb fragments corresponding to regions previously identified by ChIP with an anti-Ubx antibody [2]. All primer sequences are given in Additional data file 4.
All PCR amplifications were performed in 96-well plate format and each product was assayed by agarose gel electrophoresis. In total, 3,444 fragments were amplified and spotted, along with 480 samples of sonicated Drosophila genomic DNA (250 ng/μl), onto FMB-cDNA glass microarray slides with a Biorobotics Microgrid II arrayer. cDNA arrays were constructed from PCR amplified inserts from the Drosophila Gene Collection V.1 [36] and are described at the FlyChip website [55].
Slides were treated as described on the FlyChip website [55]. Slide hybridization and washing was carried out with a Genomic Solutions GenTAC hybridization station. Slides were scanned with an Applied Precision ArrayWoRx CCD scanner and the data processed using a custom implementation of the VSN normalization method of Huber et al. [56]. VSN performs an asinh transform with the microarray ratio data rather than the more traditional log2 transformation; in most cases the two are equivalent. The CyberT framework and website was used to assess the significance of the microarray results [29,57]. Yeast data were obtained from Lee et al. [27] and Hahn et al. [28] and the mammalian data from Trinklein et al. [38].
Expression analysis comparing RNA from heat-shock treated embryos and unshocked embryos was carried out as described on the FlyChip website [55]. Four independent samples were labeled and hybridized to the cDNA arrays. Data processing, normalization and statistical analysis were as described above for the ChIP-array studies.
Binding site distribution
The 1 kb of sequence immediately upstream of all transcripts of genes within Release 1 of the Drosophila Gene Collection (DGC1) was obtained by filtering the corresponding sequences recovered from Ensembl version 24.3b.1 (BDGP Release 3.1) via EnsMart against a list of DGC1 genes obtained from FlyBase. This subset was then searched for partial matches (mismatches <2) to an Hsf consensus sequence (GAAnnTTCnnGAA [43]) on both strands and statistics on the total number of hits for each of the DGC1 fragments calculated with custom-written BioJava-based software. The distribution of hits against the 137 Hsf-binding genes and the remaining non-Hsf-binding genes were then plotted against increasing hit counts.
Data availability
Raw and processed microarray data are available from the national Center for Biotechnology Information (NCBI) Gene Expression Omnibus site [58] with the following series accession numbers: genome tile arrays, GSE2423; cDNA arrays, GSE2398.
Additional data files
The following additional data are available with the online version of this paper. Additional data file 1 contains a tab-delimited table containing the ratios and CyberT statistics for the genome tile array. Additional data file 2 contains a tab-delimited file of data from the cDNA arrays. Additional data file 3 contains compiled data for the top 188 cDNA clones. Additional data file 4 contains sequences of the primer pairs used to amplify each of the fragments on the genome tile array.
Supplementary Material
Additional Data File 1
Tab delimited table containing the ratios and CyberT statistics for the genome tile array. For each fragment (Clone_ID) the VSN-normalized ratio (average of the dye swaps) of Hsf-enriched v control for each independent immunopurification is given. The CyberT statistics are: #obs = number of measurements accepted, Mn = mean ratio, SD = standard deviation of the mean, t = t-test statistic calculated using the standard deviation,p = p-value associated with the t-test.
Click here for file
Additional Data File 2
Tab-delimited file of data from the cDNA arrays. The average ratio for each of the 7 independent immunopurifications from 3 separate chromatin preparations is given (N = number of technical replicates used to calculate the average). For each clone the FlyBase ID, gene symbol and BDGP clone number is given. The headers indicate the chromatin batch (A, B and C) and immunopurification (1-4).
Click here for file
Additional Data File 3
Compiled data for the top 188 cDNA clones. For each gene the Fbgn and FlyBase symbol is given followed by the summarized CyberT statistics, mean ratio (Mn) and p value. The remaining data are: Custom Affymetrix array (Affy), expression ratio and p-value from cDNA arrays (cDNA and cDNA P), GAGA-factor DAM-ID experiment (GAGA, GAGA P), number of predicted Hsf sites in the 1kb upstream fragment, predicted cytological location from FlyBase, yeast homologue, expression and percentile ranking from Hahn et al. [28] (MaxExp, Non-HS, HS), p value and ratios from Lee et al. [27] and the blast score of Drosophila vs yeast protein matches.
Click here for file
Additional Data File 4
Sequences of the primer pairs used to amplify each of the fragments on the genome tile array. Clone_ID - identifier as in Additional data file 1, 5' and 3' - primer sequences, size - PCR amplicon size.
Click here for file
Acknowledgements
We are grateful to John Lis for his help in providing reagents to initiate this project and Paul Spellman and Gerry Rubin for the primers for the Adh region. We thank Peggy Farnham for advice on chromatin immunopurification and Vishy Iyer for help with the yeast data. We are indebted to Richard Auburn and FlyChip for help with the array construction, Gos Micklem for informatics input and to other members of the Russell and White labs for advice and support. This work was funded by the UK Biotechnology and Biological Sciences Research Council.
Figures and Tables
Figure 1 Distribution of fragment enrichment with anti-Hsf immunopurified chromatin on the genomic tiling array. The y-axis plots the asinh transformation (approximately equivalent to the log2 scale) of the ratio of anti-Hsf versus preimmune sera. The x-axis represents each of the 3,444 PCR products, the Adh region, Hsp gene and segmentation gene (Seg) sequences are indicated below the x-axis. Strong enrichment of fragments from the Hsp genes is indicated by their high ratio. The signals from l(2)35Bg and PRL-1 in the Adh region are indicated.
Figure 2 Graphical representation derived with the University of California at Santa Cruz (UCSC) genome browser of fragment enrichments in the 67B region containing eight putative Hsp genes (CG32041 encodes a dicistronic transcript). The blue fragments represent the 1-kb and 2-kb tiling fragments with the intensity of the blue color reflecting the degree of enrichment (asinh ratio); selected regions have been labeled with fold enrichments. The direction of transcription for each of the Hsp genes is indicated by the red arrow. The black triangles at the bottom indicate the locations of known HSEs.
Figure 3 Graphical representation of fragment enrichments for four Hsp gene regions derived with the UCSC genome browser. Details as for Figure 2; gray triangles represent predicted Hsf-binding sites. See text for details. (a) Hsp70A; (b) Hsp83, note the enrichment both 5' and 3' to the gene; (c) Hsp68, enriched fragments 5' to the gene contain predicted Hsf-binding sites; (d) Hsp60, the enriched fragments within the intron contain predicted Hsf sites.
Figure 4 PCR validation of selected positives from the cDNA arrays. Agarose gels showing the products generated by specific PCRs for each of the indicated genes using preimmune purified (-) or anti-Hsf purified (+) chromatin as an input.
Figure 5 Representation of the predicted cytological location of the top 188 Hsf-binding genes. Those identified with our cDNA array are indicated by blue triangles and the mapping of Hsf sites on polytene chromosomes reported by Westwood et. al. [22] is shown by red triangles. Filled triangles represent matches between the two studies and open triangles represent unmatched mapping.
Figure 6 Gene ontology classification of the top 188 genes identified from the cDNA array. Percentage representations are given for the prominent categories.
Figure 7 Predicted Hsf-binding sequences in the 1-kb region upstream of Hsf-binding genes. (a) Plot of the distribution of the number of predicted sites as a proportion of the population of anti-Hsf-enriched (Heat shock) or non-enriched (Control). (b) The relative position of predicted Hsf sites for each of the genes containing eight or more sites. The annotated gene start is on the right. Red triangles, perfect match; purple, one mismatch; light blue, two mismatches. Gray boxes represent the known HSEs upstream of Hsp23, Hsp26 and Hsp27.
Table 1 Enrichment of HSE with anti-Hsf ChIP as measured by quantitative real-time PCR
Hsp Primer pairs used Fold enrichment
Hsp26 5' HSE 110
Hsp26 3' UTR < 0.1
Hsp70A 5' HSE 103
Hsp70A 3' UTR 3.5
DNA was analyzed by quantitative real-time PCR as described in Materials and methods using primer pairs specific for the 5' HSE and 3' UTR regions of Hsp26 and Hsp70A. Fold enrichment is based on the comparison between amplifications with DNA from ChIP using anti-Hsf or preimmune antiseum.
Table 2 Top 50 cDNA clones identified by anti-HSF ChIP on cDNA arrays
FlyBase gene Mean ratio p-value Gene chip cDNA DAM GAGA GAGA p-value HSF sites Cytology
CG32041 3.043 2.02E-05 - 15 1.305 1.25E-05 5 67B1
CG1416 2.793 2.19E-04 - 2.4 -0.024 8.44E-01 1 40A2
CG9705 2.674 4.91E-05 - 1.5 0.118 4.86E-01 8 73C4
CG3428 2.428 6.53E-05 - 2.1 -0.086 4.76E-01 3 67B8
DnaJ-1 2.375 5.25E-04 6.13 4.4 0.489 4.23E-03 1 64E5
FKBP59 2.321 3.88E-04 - 2.4 -0.047 6.19E-01 1 30E1
CG1553 2.179 3.11E-05 - 2.4 0.368 1.47E-02 2 43E17
Hsc70Cb 2.164 2.90E-04 - 2.5 0.189 2.89E-01 1 70C15
Taf7 2.128 3.06E-06 - 1.2 0.462 5.95E-03 1 84E5
CG10286 2.128 6.95E-06 - 1.2 0.226 1.01E-01 5 83E4
CG2182 2.080 1.85E-05 - 1.1 0.188 1.26E-01 5 83B8
MESR6 2.079 9.11E-06 - 1.6 0.104 3.14E-01 4 75F7
Fer1HCH 1.986 4.23E-05 -1.09 0 1.793 6.62E-06 6 99F2
CG8258 1.962 3.83E-05 - 1.4 0.215 1.32E-01 4 44F5
CG11455 1.954 1.47E-03 - 0 0.100 5.55E-01 4 21B1
EP2237 1.928 3.25E-04 - 1.4 0.258 4.45E-02 0 21D6
alphaCop 1.926 4.35E-04 - -0.7 0.820 5.90E-01 5 62A9
Trap36 1.919 1.58E-04 - -2 -0.208 7.55E-02 2 65F2
Sir2 1.917 1.16E-04 - 1.4 0.280 4.32E-02 9 34A7
CG11791 1.906 5.08E-06 - 1.3 0.490 4.34E-03 3 96B19
CG32649 1.836 7.90E-04 - 2 0.064 5.98E-01 5 11D1
l(1)G0331 1.833 1.13E-04 - 1.3 0.143 1.77E-01 3 7B1
Cyp1 1.805 9.67E-05 -1.13 0 0.109 3.59E-01 1 14B12
RNaseX25 1.803 6.23E-05 - 1.1 -0.310 1.87E-02 2 66A21
l(2)08717 1.794 7.56E-04 - 0 1.624 1.49E-07 2 55F3
CG10576 1.724 2.14E-04 - 1.3 -0.329 4.11E-03 6 64E6
Xbp1 1.710 2.23E-04 - 1.5 0.108 3.20E-01 6 57C3
Pgi 1.708 1.65E-03 2.01 1.4 -0.011 9.08E-01 2 44F6
Hsc70-5 1.686 1.76E-04 1.44 2 0.019 8.58E-01 3 50E6
sgl 1.667 1.74E-07 1.84 1.6 0.172 2.51E-01 0 64D4
Hsp23 1.665 7.44E-04 10.11 21 0.786 2.56E-04 14 67B1
Arf79F 1.651 7.42E-04 1.08 0 0.277 7.76E-02 2 80B2
CG8297 1.623 1.95E-03 - 1.9 -0.208 1.77E-01 5 52D2
dmt 1.623 1.39E-03 - 1.2 -0.175 1.19E-01 2 85E5
l(1)G0022 1.591 1.16E-03 - 1.2 -0.110 3.57E-01 3 13E14
CG7945 1.581 9.89E-05 - -2.6 0.034 7.40E-01 5 71D4
CG31536 1.579 1.06E-04 - 0 -0.045 7.04E-01 1 82E2
Hsp27 1.568 6.82E-04 12.42 32 1.001 4.92E-05 9 67B1
Lrr47 1.560 9.19E-04 - 1.1 -0.252 1.77E-02 1 50E6
CG1103 1.551 7.79E-04 - -1.1 -0.310 1.16E-02 5 82A4
CG10600 1.539 9.21E-05 - 1.2 -0.145 1.51E-01 5 37B1
CG10973 1.532 7.99E-03 - 2 -0.148 1.52E-01 4 69E1
CG12744 1.496 5.10E-03 - 0 0.693 2.52E-03 3 46C1
sra 1.476 1.79E-04 - 2.2 -0.110 3.25E-01 6 89B12
Rpn6 1.469 8.39E-05 - 1.4 -0.237 4.20E-02 3 51C1-2
CG3488 1.466 8.40E-04 3.2 1.3 -0.023 9.13E-01 3 23D4
sktl 1.462 2.79E-03 1.14 1.1 -0.090 4.45E-01 5 57B3
Actr13E 1.447 1.04E-03 -1.27 -1.1 -0.288 2.35E-02 6 13E12
CG17294 1.447 1.81E-03 - 1.4 -0.241 2.30E-02 7 29B3
CG33111 1.426 1.27E-04 - 0 NA NA 9 95B7
The FlyBase gene symbol, corresponding to the cDNA clone on the array, is given along with the mean asinh ratio and p-values derived from Cyber-T. Expression data is given from custom Affymetrix GeneChips and from the cDNA arrays with RNA extracted from heat-shocked embryos; bold indicates significant expression (p better than 10-3). The mean ratios and p-values from a GAGA-factor DamID experiment are listed for each gene; bold indicates significant ratios. Hsf sites indicates the number of predicted Hsf sites found 1 kb upstream of each gene and the column heading cytology indicates the predicted cytological location; matches with the polytene chromosome studies are in bold. See text for details. The full list of 188 genes with associated data is given in Additional data file 3.
Table 3 Genes binding Hsf in both Drosophila and S. cerevisiae
FlyBase Fly ratio Yeast Yeast express NO-HS HS Lee ratio Lee p BLAST Yeast GO function
Acon 0.968 ACO1 2.876 0.833 0.516 1.86 6.90E-02 4.00E-274 Aconitate hydratase
Hsc70-5 1.686 SSC1 1.98 0.999 0.999 7.05 1.50E-04 6.60E-201 Protein transporter*
Hsp60 1.039 HSP60 2.608 0.997 0.997 2.33 1.50E-02 1.00E-177 Single-stranded DNA binding*
Cctgamma 1.275 CCT3 -0.28 0.646 0.343 1.85 7.80E-02 4.00E-162 Unfolded protein binding*
l(1)G0022 1.591 CCT6 0.063 0.714 0.921 1.73 8.60E-02 1.00E-136 Unfolded protein binding*
Hsc70Cb 2.164 SSE1 1.771 0.998 0.999 4.48 1.50E-03 1.70E-115 Unfolded protein binding*
CG8863 0.818 YDJ1 1.529 0.999 0.997 6.62 2.40E-04 2.00E-74 Chaperone regulator*
Cyp1 1.805 CPR1 1.251 0.996 0.995 10.83 3.40E-05 3.00E-65 Peptidyl-prolyl cis-trans isomerase
CG8258 1.962 CCT4 NA 0.385 0.055 1.85 7.80E-02 3.00E-55 Unfolded protein binding*
CG2918 1.100 SSA3 4.479 0.934 0.995 1.19 4.80E-01 7.00E-50 ATPase
DnaJ-1 2.375 SIS1 3.237 0.999 0.999 11.31 3.50E-05 7.00E-46 Unfolded protein binding*
Rab35 1.416 YPT32 0.338 0.269 0.343 1.98 4.30E-02 4.00E-40 GTPase activity
sktl 1.462 MSS4 1.042 0.712 0.879 1.91 4.80E-02 7.00E-35 Phosphatidylinositol kinase
CG4800 1.424 RBF18 0.03 0.872 0.964 2.13 1.50E-01 1.00E-33 Unknown
Cyt-c-d 0.932 CYC1 1.738 0.959 0.942 2.53 1.10E-02 4.70E-33 Electron carrier
mbf1 0.706 MBF1 NA 0.999 0.998 9.4 1.80E-04 2.00E-29 Transcription coactivator
CG1416 2.793 AHA1 3.288 0.999 0.999 9.14 5.60E-04 3.00E-28 Chaperone activator*
CG4500 1.37 FAA1 3.889 0.796 0.943 2.61 8.00E-03 9.00E-23 Fatty-acid-CoA ligase
CG32920 1.081 AHP1 1.547 0.983 0.984 3.81 1.10E-03 1.10E-19 Thioredoxin peroxidase
SH3PX1 0.877 SNX4 1.26 0.824 0.372 2.52 1.30E-02 6.30E-13 lipid binding
CG10973 1.532 FES1 3.541 0.997 0.993 NA NA 1.10E-10 Adenyl-nucleotide exchange factor
l(2)35Bg 0.755 DRE2 0.304 0.843 0.832 3.36 3.70E-03 5.00E-10 Unknown
CG12200 0.786 CST9 1.022 0.346 0.467 1.69 9.40E-02 8.70E-10 DNA binding
The FlyBase gene ID is given, along with the average asinh ratio of Hsf enrichment. Yeast homologs are indicated by their Saccharomyces genome database (SGD) common names. The average heat-induced expression and ranking of Hsf binding in non-heat shocked (NO-HS) and heat shocked (HS) cells are from [28], the Lee ratio and Lee p are from [27]. BLAST scores are derived from searches at the SGD using the Drosophila sequences as probes. Gene Ontology (GO) functional classifications are for the yeast proteins; asterisks indicate stress-response proteins.
Table 4 Primers used in the standard PCR analysis
Gene 5' primer 3' primer Product (bp)
CG3273 ACCTGGCGGAATATCACAGA ACCCCAATGTCGGATGTAGA 421
CG9746 GCGAAAACCAATCGATGTTA CGAAGCAAGATGACCTTTCC 403
CG10077 CGACCCAAAAACCAAAGTGT GATATCGGTTTTCGCCTTCA 444
CG11166 GGCCTGCGAGGAAAAGTTAT GTCGATCCCAACAGCTACAA 414
CG12876 TTTTTATTACTAACATGAACCGGTAA GCCGTTGGTTTCTCCACTT 408
CG33111 TACGCAGCGAATATCGATTG TTCTGCACGAGGGGTAGTCT 417
CG33144 CCCAATTGGAAATGAGTGCT GAATTTCCTAAATTTTGCAAGGA 441
dmt ACCATCCCCCGATCTCTAAG GCAGGCAGGAAAATCACAAT 404
EP2237 GAAAAAGGCAAAGCCATTCA CTCGGAAAAGATGGCAACAT 451
MBF1 CCAGATGGTTAAACGGCAAT GGCTCAAGGAGCTACTGAAAAA 405
veg AATTCTCGTTGCTCTCGAACT TGGAGTTCTTCTTGGCCACT 409
==== Refs
Solomon MJ Varshavsky A Formaldehyde-mediated DNA-protein crosslinking: a probe for in vivo chromatin structures. Proc Natl Acad Sci USA 1985 82 6470 6474 2995966
Gould AP Brookman JJ Strutt DI White RA Targets of homeotic gene control in Drosophila. Nature 1990 348 308 312 1979146 10.1038/348308a0
Orlando V Paro R Mapping Polycomb-repressed domains in the bithorax complex using in vivo formaldehyde cross-linked chromatin. Cell 1993 75 1187 1198 7903220 10.1016/0092-8674(93)90328-N
Orlando V Strutt H Paro R Analysis of chromatin structure by in vivo formaldehyde cross-linking. Methods 1997 11 205 214 8993033 10.1006/meth.1996.0407
Walter J Biggin MD Measurement of in vivo DNA binding by sequence-specific transcription factors using UV cross-linking. Methods 1997 11 215 224 8993034 10.1006/meth.1996.0408
Boyd KE Farnham PJ Coexamination of site-specific transcription factor binding and promoter activity in living cells. Mol Cell Biol 1999 19 8393 8399 10567564
Ren B Robert F Wyrick JJ Aparicio O Jennings EG Simon I Zeitlinger J Schreiber J Hannett N Kanin E Genome-wide location and function of DNA binding proteins. Science 2000 290 2306 2309 11125145 10.1126/science.290.5500.2306
Iyer VR Horak CE Scafe CS Botstein D Snyder M Brown PO Genomic binding sites of the yeast cell-cycle transcription factors SBF and MBF. Nature 2001 409 533 538 11206552 10.1038/35054095
Yan PS Chen CM Shi H Rahmatpanah F Wei SH Caldwell CW Huang TH Dissecting complex epigenetic alterations in breast cancer using CpG island microarrays. Cancer Res 2001 61 8375 8380 11731411
Ren B Cam H Takahashi Y Volkert T Terragni J Young RA Dynlacht BD E2F integrates cell cycle progression with DNA repair, replication, and G(2)/M checkpoints. Genes Dev 2002 16 245 256 11799067 10.1101/gad.949802
Weinmann AS Yan PS Oberley MJ Huang TH Farnham PJ Isolating human transcription factor targets by coupling chromatin immunoprecipitation and CpG island microarray analysis. Genes Dev 2002 16 235 244 11799066 10.1101/gad.943102
Horak CE Mahajan MC Luscombe NM Gerstein M Weissman SM Snyder M GATA-1 binding sites mapped in the beta-globin locus by using mammalian chIp-chip analysis. Proc Natl Acad Sci USA 2002 99 2924 2929 11867748 10.1073/pnas.052706999
Cawley S Bekiranov S Ng HH Kapranov P Sekinger EA Kampa D Piccolboni A Sementchenko V Cheng J Williams AJ Unbiased mapping of transcription factor binding sites along human chromosomes 21 and 22 points to widespread regulation of noncoding RNAs. Cell 2004 116 499 509 14980218 10.1016/S0092-8674(04)00127-8
van Steensel B Delrow J Henikoff S Chromatin profiling using targeted DNA adenine methyltransferase. Nat Genet 2001 27 304 308 11242113 10.1038/85871
Sun LV Chen L Greil F Negre N Li TR Cavalli G Zhao H Van Steensel B White KP Protein-DNA interaction mapping using genomic tiling path microarrays in Drosophila. Proc Natl Acad Sci USA 2003 100 9428 9433 12876199 10.1073/pnas.1533393100
van Steensel B Delrow J Bussemaker HJ Genomewide analysis of Drosophila GAGA factor target genes reveals context-dependent DNA binding. Proc Natl Acad Sci USA 2003 100 2580 2585 12601174 10.1073/pnas.0438000100
Bianchi-Frias D Orian A Delrow JJ Vazquez J Rosales-Nieves AE Parkhurst SM Hairy transcriptional repression targets and cofactor recruitment in Drosophila. PLoS Biol 2004 2 E178 15252443 10.1371/journal.pbio.0020178
Morimoto RI Regulation of the heat shock transcriptional response: cross talk between a family of heat shock factors, molecular chaperones, and negative regulators. Genes Dev 1998 12 3788 3796 9869631
O'Brien T Lis JT Rapid changes in Drosophila transcription after an instantaneous heat shock. Mol Cell Biol 1993 13 3456 3463 8497261
Wu C Heat shock transcription factors: structure and regulation. Annu Rev Cell Dev Biol 1995 11 441 469 8689565 10.1146/annurev.cb.11.110195.002301
Liu XD Liu PC Santoro N Thiele DJ Conservation of a stress response: human heat shock transcription factors functionally substitute for yeast HSF. EMBO J 1997 16 6466 6477 9351828 10.1093/emboj/16.21.6466
Westwood JT Clos J Wu C Stress-induced oligomerization and chromosomal relocalization of heat-shock factor. Nature 1991 353 822 827 1944557 10.1038/353822a0
Mason PB JrLis JT Cooperative and competitive protein interactions at the hsp70 promoter. J Biol Chem 1997 272 33227 33233 9407112 10.1074/jbc.272.52.33227
Shopland LS Lis JT HSF recruitment and loss at most Drosophila heat shock loci is coordinated and depends on proximal promoter sequences. Chromosoma 1996 105 158 171 8781184 10.1007/s004120050171
Andrulis ED Guzman E Doring P Werner J Lis JT High-resolution localization of Drosophila Spt5 and Spt6 at heat shock genes in vivo: roles in promoter proximal pausing and transcription elongation. Genes Dev 2000 14 2635 2649 11040217 10.1101/gad.844200
Lis JT Mason P Peng J Price DH Werner J P-TEFb kinase recruitment and function at heat shock loci. Genes Dev 2000 14 792 803 10766736
Lee TI Rinaldi NJ Robert F Odom DT Bar-Joseph Z Gerber GK Hannett NM Harbison CT Thompson CM Simon I Transcriptional regulatory networks in Saccharomyces cerevisiae. Science 2002 298 799 804 12399584 10.1126/science.1075090
Hahn JS Hu Z Thiele DJ Iyer VR Genome-wide analysis of the biology of stress responses through heat shock transcription factor. Mol Cell Biol 2004 24 5249 5256 15169889 10.1128/MCB.24.12.5249-5256.2004
Baldi P Long AD A Bayesian framework for the analysis of microarray expression data: regularized t -test and statistical inferences of gene changes. Bioinformatics 2001 17 509 519 11395427 10.1093/bioinformatics/17.6.509
Fernandes M Xiao H Lis JT Binding of heat shock factor to and transcriptional activation of heat shock genes in Drosophila. Nucleic Acids Res 1995 23 4799 4804 8532521
Perezgasga L Segovia L Zurita M Molecular characterization of the 5' control region and of two lethal alleles affecting the hsp60 gene in Drosophila melanogaster. FEBS Lett 1999 456 269 273 10456322 10.1016/S0014-5793(99)00963-1
Patriarca EJ Maresca B Acquired thermotolerance following heat shock protein synthesis prevents impairment of mitochondrial ATPase activity at elevated temperatures in Saccharomyces cerevisiae. Exp Cell Res 1990 190 57 64 2143732 10.1016/0014-4827(90)90143-X
Ferm MT Soderstrom K Jindal S Gronberg A Ivanyi J Young R Kiessling R Induction of human hsp60 expression in monocytic cell lines. Int Immunol 1992 4 305 311 1567788
Lis J Promoter-associated pausing in promoter architecture and postinitiation transcriptional regulation. Cold Spring Harb Symp Quant Biol 1998 63 347 356 10384299 10.1101/sqb.1998.63.347
Bailey TL Elkan C Fitting a mixture model by expectation maximization to discover motifs in biopolymers. Proc Int Conf Intell Syst Mol Biol 1994 2 28 36 7584402
Rubin GM Hong L Brokstein P Evans-Holm M Frise E Stapleton M Harvey DA A Drosophila complementary DNA resource. Science 2000 287 2222 2224 10731138 10.1126/science.287.5461.2222
Leemans R Egger B Loop T Kammermeier L He H Hartmann B Certa U Hirth F Reichert H Quantitative transcript imaging in normal and heat-shocked Drosophila embryos by using high-density oligonucleotide arrays. Proc Natl Acad Sci USA 2000 97 12138 12143 11035778 10.1073/pnas.210066997
Trinklein ND Murray JI Hartman SJ Botstein D Myers RM The role of heat shock transcription factor 1 in the genome-wide regulation of the mammalian heat shock response. Mol Biol Cell 2004 15 1254 1261 14668476 10.1091/mbc.E03-10-0738
Boehm AK Saunders A Werner J Lis JT Transcription factor and polymerase recruitment, modification, and movement on dhsp70 in vivo in the minutes following heat shock. Mol Cell Biol 2003 23 7628 7637 14560008 10.1128/MCB.23.21.7628-7637.2003
Zou S Meadows S Sharp L Jan LY Jan YN Genome-wide study of aging and oxidative stress response in Drosophila melanogaster Proc Natl Acad Sci USA 2000 97 13726 13731 11095759 10.1073/pnas.260496697
Zinke I Schutz CS Katzenberger JD Bauer M Pankratz MJ Nutrient control of gene expression in Drosophila : microarray analysis of starvation and sugar-dependent response. EMBO J 2002 21 6162 6173 12426388 10.1093/emboj/cdf600
Park JM Werner J Kim JM Lis JT Kim YJ Mediator, not holoenzyme, is directly recruited to the heat shock promoter by HSF upon heat shock. Mol Cell 2001 8 9 19 11511356 10.1016/S1097-2765(01)00296-9
Xiao H Lis JT Germline transformation used to define key features of heat-shock response elements. Science 1988 239 1139 1142 3125608
Arbeitman MN Furlong EE Imam F Johnson E Null BH Baker BS Krasnow MA Scott MP Davis RW White KP Gene expression during the life cycle of Drosophila melanogaster. Science 2002 297 2270 2275 12351791 10.1126/science.1072152
Osborne KA Robichon A Burgess E Butland S Shaw RA Coulthard A Pereira HS Greenspan RJ Sokolowski MB Natural behavior polymorphism due to a cGMP-dependent protein kinase of Drosophila. Science 1997 277 834 836 9242616 10.1126/science.277.5327.834
Matuschewski K Hauser HP Treier M Jentsch S Identification of a novel family of ubiquitin-conjugating enzymes with distinct amino-terminal extensions. J Biol Chem 1996 271 2789 2794 8576256 10.1074/jbc.271.5.2789
Takemaru K Li FQ Ueda H Hirose S Multiprotein bridging factor 1 (MBF1) is an evolutionarily conserved transcriptional coactivator that connects a regulatory factor and TATA element-binding protein. Proc Natl Acad Sci USA 1997 94 7251 7256 9207077 10.1073/pnas.94.14.7251
Brendel C Gelman L Auwerx J Multiprotein bridging factor-1 (MBF-1) is a cofactor for nuclear receptors that regulate lipid metabolism. Mol Endocrinol 2002 16 1367 1377 12040021 10.1210/me.16.6.1367
Takemaru K Harashima S Ueda H Hirose S Yeast coactivator MBF1 mediates GCN4-dependent transcriptional activation. Mol Cell Biol 1998 18 4971 4976 9710580
Jindra M Gaziova I Uhlirova M Okabe M Hiromi Y Hirose S Coactivator MBF1 preserves the redox-dependent AP-1 activity during oxidative stress in Drosophila. Embo J 2004 23 3538 3547 15306851 10.1038/sj.emboj.7600356
Oberley MJ Tsao J Yau P Farnham PJ High-throughput screening of chromatin immunoprecipitates using CpG-island microarrays. Methods Enzymol 2004 376 315 334 14975315
Celniker SE Wheeler DA Kronmiller B Carlson JW Halpern A Patel S Adams M Champe M Dugan SP Frise E Finishing a whole-genome shotgun: release 3 of the Drosophila melanogaster euchromatic genome sequence. Genome Biol 2002 3 research0079 12537568 10.1186/gb-2002-3-12-research0079
Kent WJ Sugnet CW Furey TS Roskin KM Pringle TH Zahler AM Haussler D The human genome browser at UCSC. Genome Res 2002 12 996 1006 12045153 10.1101/gr.229102. Article published online before print in May 2002
UCSC Genome Browser
FlyChip Cambridge Microarray Facility
Huber W von Heydebreck A Sultmann H Poustka A Vingron M Variance stabilization applied to microarray data calibration and to the quantification of differential expression. Bioinformatics 2002 18 Suppl 1 S96 S104 12169536
CyberT microarray statistical analysis
NCBI Gene Expression Omnibus
| 15998452 | PMC1175994 | CC BY | 2021-01-04 16:05:40 | no | Genome Biol. 2005 Jun 10; 6(7):R63 | utf-8 | Genome Biol | 2,005 | 10.1186/gb-2005-6-7-r63 | oa_comm |
==== Front
J BiolJournal of Biology1478-58541475-4924BioMed Central London jbiol231598240810.1186/jbiol23Research ArticleMotifs, themes and thematic maps of an integrated Saccharomyces cerevisiae interaction network Zhang Lan V [email protected] Oliver D [email protected] Sharyl L [email protected] Debra S [email protected] Amy HY [email protected] Guillaume [email protected] Brenda [email protected] Howard [email protected] Charles [email protected] Frederick P [email protected] Department of Biological Chemistry and Molecular Pharmacology, Harvard Medical School, Boston, MA 02115 USA2 Banting and Best Department of Medical Research and Department of Medical Genetics and Microbiology, University of Toronto, Toronto ON M5G 1L6, Canada3 Department of Biology, McGill University, Montreal PQ H3A 1B1, Canada2005 1 6 2005 4 2 6 6 17 11 2004 21 2 2005 8 4 2005 Copyright © 2005 Zhang 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
Large-scale studies have revealed networks of various biological interaction types, such as protein-protein interaction, genetic interaction, transcriptional regulation, sequence homology, and expression correlation. Recurring patterns of interconnection, or 'network motifs', have revealed biological insights for networks containing either one or two types of interaction.
Results
To study more complex relationships involving multiple biological interaction types, we assembled an integrated Saccharomyces cerevisiae network in which nodes represent genes (or their protein products) and differently colored links represent the aforementioned five biological interaction types. We examined three- and four-node interconnection patterns containing multiple interaction types and found many enriched multi-color network motifs. Furthermore, we showed that most of the motifs form 'network themes' – classes of higher-order recurring interconnection patterns that encompass multiple occurrences of network motifs. Network themes can be tied to specific biological phenomena and may represent more fundamental network design principles. Examples of network themes include a pair of protein complexes with many inter-complex genetic interactions – the 'compensatory complexes' theme. Thematic maps – networks rendered in terms of such themes – can simplify an otherwise confusing tangle of biological relationships. We show this by mapping the S. cerevisiae network in terms of two specific network themes.
Conclusion
Significantly enriched motifs in an integrated S. cerevisiae interaction network are often signatures of network themes, higher-order network structures that correspond to biological phenomena. Representing networks in terms of network themes provides a useful simplification of complex biological relationships.
==== Body
Background
A cellular system can be described as a web of relationships amongst genes, proteins, and other macromolecules. Proteins can interact via direct or indirect physical contact (referred to as protein-protein interactions). They can also interact genetically; for example, if a combination of mutations in two genes causes a more severe fitness defect (or death) than either mutation alone, the two genes have a synthetic sick or lethal (SSL) genetic interaction. In addition, two genes can relate to each other by transcriptional regulation, sequence homology, or expression correlation. Overlaps between different types of biological interaction have been noted previously. For example, interacting proteins are more likely to have similar expression patterns [1,2]; genes with correlated expression are more likely to be controlled by a common transcription factor [3]; and synthetic genetic interactions are more likely to occur between homologous genes [4]. These represent pairwise relationships between various types of biological interaction, however, understanding how they are organized in an integrated network remains a challenging task.
The concept of network motifs (referred to simply as 'motifs' hereafter) has been developed to describe simple patterns of interconnection in networks that occur more frequently than expected in randomized networks [5,6]. It has been proposed that network motifs represent the basic building blocks of complex networks [5-7]. Different types of network exhibit different motif profiles, providing a means for network classification [8]. The network motif concept is extensible to an integrated network of many interaction types (that is, a 'multi-color network', with interactions of each type represented by a different color). Multi-color network motifs characterize relationships between different biological interaction types within local network neighborhoods. A recent study examined network motifs in integrated cellular networks of two interaction types – transcriptional regulation and protein-protein interaction [9]. Other gene-pair relationships are also important. Correlated expression profiles may reflect common regulation or a cellular requirement for contemporaneous action. Sequence homology suggests descent from a common ancestor and therefore an increased likelihood of performing a related function. Genetic interactions describe synergistic or antagonistic consequences of mutations in two or more genes. For example, a recent systematic study [4] identified a large number of SSL interactions, revealing gene pairs in which one gene compensates for loss of the other, suggesting a functional relationship between the two gene products. Here, we describe network motifs discovered from a Saccharomyces cerevisiae network that integrates five types of biological interactions or relationships: protein-protein interactions, genetic interactions, transcriptional regulation, sequence homology, and expression correlation.
It has been shown for the Escherichia coli and Caenorhabditis elegans transcriptional network that subgraphs matching two types of transcriptional regulatory circuit motif – feed-forward and bi-fan – overlap with one another and form large clusters [6,10,11]. This suggests that instead of representing network "building blocks", motifs should in some cases be viewed as signatures of more fundamental higher-order structures. Here, we describe 'network themes' - recurring higher-order interconnection patterns that encompass multiple occurrences of network motifs and reflect a common organizational principle. We show that most network motifs found in the integrated S. cerevisiae network can be understood in terms of only a few network themes. Network themes can be tied to specific biological phenomena and may represent more fundamental network design principles. They also suggest a natural simplification of the otherwise complex set of relationships in an integrated network. We demonstrate this by providing two thematic maps of the integrated S. cerevisiae network.
Results
An integrated S. cerevisiae network
We constructed an integrated S. cerevisiae network by combining five types of biological interaction. Nodes in the network represent genes or proteins, and differently colored links represent different biological interaction types. These include: 3,060 SSL interactions derived from synthetic genetic array (SGA) analysis [4]; 40,438 protein sequence homology relationships from a genome-wide BLAST search [12]; 57,367 correlated mRNA expression relationships derived from microarray data [13]; 49,537 stable protein interactions defined by shared membership in a protein complex [14-16]; and 4,357 transcriptional regulatory interactions from a genome-wide chromatin immuno-precipitation (ChIP) study [7]. This collection of data resulted in a single integrated network involving 5,831 nodes and 154,759 links in total (for a full list see Additional data file 1).
Three-node network motifs and corresponding themes in the integrated network
Networks of protein-protein and synthetic genetic interaction have been reported to be scale-free and 'small-world' [4,17,18]. Being a small-world network implies neighborhood clustering, where neighbors of a given node tend to interact with one another, resulting in an abundance of three-node interconnection patterns – that is, 'triangles'. In addition, relationships such as sequence homology and correlated expression are often transitive (that is, if gene A is homologous to gene B, and gene B is homologous to gene C, then gene A is often homologous to gene C). Thus, a triangle motif for each of these component subnetworks is expected. In order to find additional motifs involving multiple interaction types, we looked for frequently occurring patterns of interconnection in the integrated network, assessing their significance by comparing the observed network with appropriately randomized networks.
We first exhaustively tested all three-node interconnection patterns defined by a single type of link between each pair of nodes (there are 50 such patterns; for a full list see Additional data file 2). Shown in Figure 1 is a list of enriched three-node network motifs, each describing a significantly (p < 0.001) enriched topological relationship among biological interactions of varying types in the integrated S. cerevisiae network. We found that most motifs can be explained in terms of higher-order structures, or network themes, which are representative of the underlying biological phenomena. We classified these motifs into seven sets (Figure 1a–g) according to the themes discussed below. There are five additional motifs that we could not classify into themes (Figure 1h). These are addressed further in the Discussion.
The first motif set contains the transcriptional feed-forward motif (Figure 1a), which has been characterized in several earlier studies of single-color networks of transcriptional regulation [5-7,11]. Because transcriptional regulation links often overlap co-expression links, we added to this set another motif composed of two genes with correlated expression that are also indirectly connected by transcriptional regulatory links through an intermediate gene. We noticed that gene triads matching the feed-forward motif in the S. cerevisiae network often overlap with one another to form large clusters, as in the E. coli and C. elegans transcriptional regulatory networks [6,10,11]. For example, Swi4 and its transcriptional activator Mcm1 together regulate a number of cell-cycle-related genes (Figure 1a) [19-21]. Most gene triads matching the feed-forward motif belong to such clusters, leading us to note a 'feed-forward' theme – a pair of transcription factors, one regulating the other, and both regulating a common set of target genes that are often involved in the same biological process.
The next set contains 'co-pointing' motifs, in which a target gene is regulated by two transcription factors that interact physically or share sequence homology (Figure 1b). These co-pointing motifs reflect the fact that two transcription factors regulating the same target gene are often derived from the same ancestral gene, or function as a protein complex. We found that these motifs also overlap extensively, forming a co-pointing theme, in which multiple transcription factors, connected to one another by physical interaction or sequence homology, regulate a common set of target genes. Figure 1b shows one such example, where Hap2, Hap3, Hap4 and Hap5 form the CCAAT-binding factor complex [22] which regulates common target genes, many of which are involved in carbohydrate metabolism [23].
A third set of motifs contains two targets of the same transcription factor bridged by a link of correlated expression, protein-protein interaction, or sequence homology (Figure 1c). These motifs indicate that transcriptional co-regulation is often accompanied by co-expression, membership in the same protein complex, or descent from a common ancestor [3,24], and suggest a 'regulonic complex' theme in which co-regulated proteins are often components of a complex or related by gene duplication and divergence. Illustrating this theme, six members of the histone octamer, Hhf1, Hhf2, Hht1, Hht2, Hta1 and Htb1 are all regulated by Hir1 and Hir2, histone transcriptional co-repressors that are required for periodic repression of the histone genes (Figure 1c) [25].
The fourth motif set consists of four three-node motifs each containing protein-protein interactions or correlated expression links (Figure 1d). Protein-protein interaction is known to correlate positively with co-expression [1,2], and proteins corresponding to these motifs often reside in the same complex. Thus, motifs within this set are likely to be signatures of a 'protein complex' theme. One of the many examples is the ATP synthase complex [26,27], whose members are linked extensively to one another by protein-protein interaction and correlated expression (Figure 1d).
The fifth motif set contains three-node motifs linked by SSL interaction or by sequence homology (Figure 1e). In the SSL network, neighbors of the same gene often interact with one another [4]. This translates into a triangle motif of three SSL links. Furthermore, homology relationships are often transitive (that is, if gene A is homologous to gene B, and gene B is homologous to gene C, then gene A is often homologous to gene C). These phenomena, combined with the fact that genes sharing sequence homology have an increased tendency to show SSL interaction, suggest an underlying theme of the neighborhood clustering in the integrated SSL/homology network: SSL or homology neighbors of one node tend to be linked to one another by SSL interaction or sequence homology. This theme is exemplified by Myo2 and a number of genes connected to Myo2 by SSL interaction or sequence homology (Figure 1e) [4,28,29].
The sixth motif set describes network motifs containing two nodes linked either by SSL interaction or by sequence homology, with a third node connected to each of them through protein-protein interaction or through correlated expression (Figure 1f). All three proteins (a, b and c, as in the schematic diagram in Figure 1f) may be members of the same complex, with either b or c being sufficient to support the essential function of the complex. Proteins b and c may either reside in the complex at the same time, or be mutually exclusive (that is, competing for the same docking position in the complex). This can be generalized to a network theme of a protein complex with partially redundant or compensatory members. As one instance of this theme, both Ssn8 and Cdc73 associate with the RNA polymerase II complex [30,31], and only one of them is required for viability (Figure 1f) [4].
We found the seventh motif set particularly interesting. Motifs in this set contain two nodes linked by protein-protein interaction or correlated expression, with a third node connected to both either by SSL interaction or by sequence homology (Figure 1g). Considering previously observed correlations between protein-protein interaction and co-expression [1,2] and between SSL interaction and sequence homology [4], these motifs indicate that members of a given protein complex or biological process often have common synthetic genetic interaction partner(s) (Figure 1g). For instance, four out of the five Gim complex proteins [32] exhibit synthetic lethality with Sec72 (Figure 1g) [4]. A 'compensatory protein and complex/process' theme, in which a protein and a distinct protein complex or biological process have compensatory function, results in synthetic sickness or lethality between the protein and any member of the complex/process essential to the function of that complex/process. It is also possible for the single protein to be part of another complex/process, so that these motifs may in turn be signatures of a larger 'compensatory complexes/processes' theme, which we examine further below.
In addition to the motif sets described above, there are five motifs that we did not categorize (Figure 1h). These are especially interesting, as they may represent unknown biological phenomena (described further in the Discussion).
Four-node network motifs corresponding to the 'compensatory complexes/processes' theme in the integrated network
There are over 5,000 different connected four-node interconnection patterns with each pair of nodes bridged by at most one link type. Here, we have focused on a subset of four-node patterns of particular interest. Recalling the 'compensatory protein and complex/process' theme (Figure 1g), in which a protein has compensatory function with other proteins in a complex or a process, we wondered whether there also exists a network theme corresponding to a pair of complexes/processes with compensatory function (connected to each other by many links of SSL interaction or sequence homology). We searched for all four-node interconnection patterns that would fit this 'compensatory complexes/processes' theme (there are a total of 66 such patterns – for a full list see Additional data file 3). Each pattern is composed of two pairs of nodes such that a protein-protein interaction or correlated expression link exists within each pair and SSL or sequence homology links extend between the two pairs (Figure 2). Using one thousand randomized networks to assess significance, 48 out of the 66 patterns corresponding to this theme were found to be network motifs defined by significant enrichment (p < 0.001) in the real network (see Figure 2 for a few examples and Additional data file 3 for a full list). This supports our hypothesis that compensatory pairs of complexes or processes are a theme in the integrated S. cerevisiae network. The endoplasmic reticulum (ER) protein-translocation subcomplex [33] and the Gim complex [32], connected by many SSL interactions [4], together illustrate this theme. This example also encompasses the 'compensatory protein and complex/process' theme depicted in Figure 1g, wherein multiple SSL or homology links connect Sec72 and the Gim complex.
A thematic map of compensatory complexes
In order to identify additional pairs of protein complexes with overlapping or compensatory function, we rendered a map of the network in terms of the 'compensatory complexes' theme. This map can also serve as a guide to 'redundant systems' within the integrated S. cerevisiae network, wherein two complexes provide the organism with robustness with respect to random mutation when each complex acts as a 'failsafe mechanism' for the other. To generate a thematic map of compensatory complexes, we searched for pairs of protein complexes with many inter-complex SSL interactions. For this purpose, we only considered links of protein-protein interaction and SSL interaction and reduced the original network to one in which nodes are complexes and links are SSL interactions (with multiple links allowed between a pair of 'collapsed' nodes). For each pair of protein complexes, we calculated the number of links between them and assessed the significance of enrichment (see the Materials and methods section for details). Among the 72 complexes examined (for a list of complexes see Additional data file 1), we found 21 pairs of complexes (involving 26 complexes; listed in Additional data file 4) showing significant enrichment (p ≤ 0.05) for inter-complex SSL interactions. These compensatory complexes can be visualized as a thematic map in which each node represents a protein complex and each link bridges a pair of complexes connected by a significant number of SSL interactions (Figure 3).
A thematic map of regulonic complexes
Other themes depicted in Figure 1 that might be usefully exploited to generate a simplified thematic map include the 'regulonic complex' theme (Figure 1c), wherein one transcription factor (TF) regulates multiple members of a given protein complex. Such a phenomenon has been observed previously [34]. Here, we provide an automated procedure for drawing the map in terms of this network theme. To this end, we examined all possible pairings of a transcription factor with a particular protein complex (together, a 'TF-complex pair'). We reduced the integrated network of stable protein-protein interactions and transcriptional regulations to one in which nodes are either transcription factors or complexes and links indicate transcriptional regulation (with multiple links allowed between a pair of nodes). For each TF-complex pair, we calculated the number of links between them, and assessed the significance according to the probability of obtaining at least the observed number of links if each transcription factor were to choose its regulatory targets randomly. A total of 91 TF-complex pairs showed significant enrichment (p ≤ 0.05) for transcriptional regulation links. These significant TF-complex relationships can also be viewed as a network whose nodes are transcription factors or complexes and whose links represent TF-complex pairs with significantly enriched transcriptional regulation (Figure 4a). Judging from experimental evidence, many of the links connect transcription factors and protein complexes involved in the same biological process, and complexes of related function are often connected to the same transcription factor (Figure 4b).
Discussion
Network motifs have previously been sought in simple networks [5-7,10,11] and recently in an integrated network of transcriptional regulation and protein-protein interaction [9]. In this study, we sought network motifs in an integrated S. cerevisiae network with five types of biological interaction. We identified many significantly enriched motifs, which fall into several classes with distinct biological implications, revealing the interplay of different types of biological interaction in local network neighborhoods. Previously, motifs have been described as elementary building blocks of complex networks [5-7,9,11]. Here, we describe network themes – recurring higher-order interconnection patterns that encompass multiple occurrences of network motifs. We show that the abundance of most motifs in the integrated S. cerevisiae network can be explained in terms of a network theme.
Network themes represent a more fundamental level of abstraction that may often be preferable to network motifs for several reasons. Network motifs have been defined with artificial restrictions on the number of nodes and the specific interconnection patterns, and gene triads or tetrads corresponding to these motifs often do not exist in isolation in the network. Rather, they often overlap extensively with one another to form higher-order structures corresponding in many cases to known biological phenomena; this is supported by observations from other studies [9,10]. This phenomenon suggests that motifs are often not 'atomic' elements of the network, but are instead signatures or symptoms of more fundamental higher-order structures, or network themes. Although many motifs can be explained in terms of higher-order themes, some network motifs have an elemental function that is preserved even when that motif is embedded within a larger theme. This was demonstrated, for example, by Alon and colleagues for the coherent feed-forward loop [35].
In addition to the network themes and motifs depicted in Figure 1a–g, there are five motifs that we did not categorize (Figure 1h). Each of these motifs contains: a transcriptional regulation link, with a third node connecting to the transcription factor and its target via two stable physical interactions (motif H1); two sequence homology links (motif H2); one correlated expression link and one homology link, respectively (motif H3); one homology link and one correlated expression link, respectively (motif H4), or two correlated expression links (motif H5). Given that physical interaction links are mostly transitive, motif H1 indicates that transcription factors often co-complex with the target proteins they regulate, and suggests a mechanism of feedback regulation for transcription through protein-protein interaction. Motif H2 implies sequence homology between a transcription factor and its target, given the near transitivity of homology links. Such homology may seem unexpected but can be explained if there is frequent serial regulation of one transcription factor by another, since transcriptional factors often share homology, for example in their DNA binding domains. Motif H5 may be due simply to the overlap between transcriptional regulation links and correlated expression links, and the near transitivity of correlated expression links. The implications of motifs H3 and H4 are unclear to us; they might represent currently unknown trends in transcriptional regulatory mechanism. We hope to address some of these questions in the future by investigating the roles of genes in the subnetworks corresponding to the motifs (for example, whether the target gene in motif H2 is often a transcription factor).
Both network motifs and themes represent network characteristics that can be exploited to predict individual interactions given sometimes-uncertain experimental evidence. As has recently been shown, integration of multiple evidence types [22,36-38] can be successfully used to predict protein-protein interactions and synthetic genetic interactions, or to stratify them by confidence. In addition, the dense local neighborhood characteristic of the protein-protein interaction network can be exploited to predict protein-protein interactions [39-42]. This idea, extended to multi-color network motifs, allows us to make predictions based on topological relationships involving multiple types of links. In particular, we may predict a certain type of link between a given pair of nodes if its addition would complete a structure matching an enriched network motif. For example, two genes with a common SSL interaction partner may have increased probability of protein-protein interaction, because the addition of a protein-protein interaction link between these two genes results in a match to motif G1 (Figure 1g). Similarly, an SSL link between two genes can complete a match to motif G1 if the two genes are connected to a third gene by a protein-protein interaction link and an SSL link, respectively (Figure 1g). Such a 'two-hop physical-SSL' relationship has been recently shown to be a strong predictor of SSL interaction [38]. An interaction can also be predicted if its addition fits into a recurring network theme. For instance, there are significantly enriched SSL interactions between the ER protein-translocation subcomplex and the Gim complex (Figure 2). However, no SSL interactions have been observed between Sec62 or Sec63, two members of the ER protein-translocation subcomplex and any protein in the Gim complex because Sec62 and Sec63 were not used as queries in the SGA analysis [4]. We therefore hypothesize that Sec62 or Sec63 has SSL interactions with many members of the Gim complex.
In addition, since themes represent the network organization at the functional level, they can also be used to predict functions for genes involved in a specific theme. For example, in the feed-forward theme depicted in Figure 1a, most of the genes regulated by both Mcm1 and Swi4 are involved in control or execution of the cell cycle. We therefore hypothesize that Yor315w, a protein of unknown function, is involved in the cell cycle. More refined hypotheses can be achieved by incorporating other information such as sequence data and expression profiles. Predictions based on network themes may be robust with respect to errors in the input data, since they depend on connectivity patterns in extended network neighborhoods instead of one or very few links.
To assess whether SSL interactions involving essential genes are enriched in subgraphs matching the motifs, we counted, for each motif containing an SSL link, the fraction of subgraphs with at least one SSL interaction involving an essential gene. The results are summarized in Additional data file 2. In the SGA analysis, 11 of the 132 query genes are essential. Among the 3,060 SSL interactions, 322 of them (10.5%) involve an essential gene. Results for the network motifs are mostly consistent with this frequency of essentiality: for most motifs (E1, E2, E3, G1, G4 and G5), approximately 10% of the matching subgraphs contain SSL interactions involving an essential gene (see Additional data file 2). It is interesting, however, that subgraphs matching motifs F1 and F3 are particularly enriched with SSL interactions involving essential genes (36.4% and 24.4%, respectively). This suggests that SSL interactions within a protein complex may often involve essential genes.
Each network theme has a different biological implication, and each permits a natural simplification of the integrated network. To demonstrate this, we produced thematic maps of compensatory complexes and of regulonic complexes. The map of compensatory complexes identifies specific protein complexes with overlapping or compensatory function. Many of the links connect functionally related complexes, as supported by previous experimental evidence. For example, the replication complex, is 'genetically connected' to the Mre11/Rad50/Xrs2 complex [43], the Rad54-Rad51 complex [44], and the Rad17/Mec3/Ddc1 complex [45]. The first two function in the repair of double-strand DNA breaks [44,46] and the third is required for cell-cycle checkpoint control after DNA damage [47], both of which are associated with DNA replication. The histone deacetylase B (HDB) complex [48,49] is linked to the SAGA complex [50]; both of these affect histone acetylation and are important components of transcriptional regulation [51]. There are also some unverified but intriguing links, such as the one between the Gim complex [32] and the CCAAT-binding factor [22], which connects two seemingly unrelated complexes (Figure 3). The potential functional relationship between these complexes awaits further experimental validation.
Novel predictions for synthetic sick or lethal interactions can be made from the thematic map of compensatory complexes. Specifically, we can predict any two proteins to have an SSL interaction if they are members of two separate complexes bridged by a link in the map. There were 1,134 such protein pairs that had not been previously tested by the SGA study used to derive the compensatory complex map. We sought independent validation of these predictions among published smaller-scale studies of genetic interaction. We conservatively estimate that 10% of these pairs will have been examined for genetic interaction (note that Tong et al. [4], the largest systematic study to date, examined only approximately 4% of all gene pairs). Therefore, we might only hope to find approximately 113 validated pairs (10% of 1,134 predictions). Tong et al. [4] observed the baseline rate of SSL interaction to be 0.5%, so by chance we might expect to find fewer than one SSL interaction (0.5% of 10% of 1,134). Our literature search revealed ten gene pairs with known SSL interactions among the predictions: Arp2-Myo1 [52], Vrp1-Myo1 [53], Las17-Myo1 [54], Bem1-Myo1 [54], Rvs167-Myo1 [55], Rvs167-Myo2 [55], Smy1-Pfy1 [56], Rad50-Cdc2 [57,58], Rad54-Cdc2 [57], and Rad51-Cdc2 [58]. From this we conservatively estimate a success rate of around 9%, demonstrating the value of the thematic map in predicting new SSL interactions. Our use of the thematic map to predict genetic interactions differs from the previous prediction approach based on two-hop physical-SSL interactions [38] in that here we required a greater abundance of SSL interactions between two protein complexes than would be expected by chance, whereas previous work did not exploit the number of observed two-hop physical-SSL interactions. Furthermore, the thematic map approach has the potential to predict genetic interaction between two genes even if neither gene has any previously known SSL interactions.
In producing the thematic map of compensatory complexes, the statistical power was limited because only 4% of yeast gene pairs have been examined for synthetic genetic interactions [4]. Many compensatory complex pairs have escaped detection because too few inter-complex protein pairs have been tested for SSL to achieve statistical significance. We expect this map to grow substantially as large-scale studies of genetic interaction proceed [59]. In higher organisms for which exhaustive determination of genetic interaction is a more distant goal, we may advance our understanding more rapidly by choosing a 'scaffold' set of genes such that each known or hypothesized protein complex or pathway is represented by at least one query gene in an SSL screen.
Materials and methods
Constructing an integrated S. cerevisiae network
Synthetic genetic interactions between 132 query genes and about 5,000 array genes were obtained from a recent large-scale SGA analysis in S. cerevisiae [4]. Genome-wide BLAST [12] was performed using all yeast protein sequences. Pairs of proteins with E values of less than 10-3 were considered homologous. Pearson correlation coefficients were calculated for all pairs of yeast proteins based on the Rosetta compendium microarray dataset [13]. Protein pairs with correlation coefficients larger than 0.6 were regarded as having correlated expression. Protein complexes were obtained from the MIPS [14] database and two large-scale affinity purification studies [15,16]. All pairs of proteins residing in the same complex were treated as having stable protein-protein interactions. Transcriptional regulation was inferred from the genome-wide ChIP studies of 106 yeast transcription factors [7]. If transcription factor A binds to the promoter region of gene B with a p value of less than 0.001, then a directed transcriptional regulatory link is assigned from A to B.
Detecting network motifs
We enumerated all connected three-node subgraphs in the network as previously described [5]. For each interconnection pattern defined by one link between each pair of nodes, we recorded the number of subgraphs matching this pattern in the real network as well as in all randomized networks. A subgraph is considered a 'match' to the pattern if the subgraph can be transformed to the pattern by any combination of node identity permutations or link removals. The p value for the enrichment of an interconnection pattern is defined by the fraction of randomized networks having at least the number of matching subgraphs as the real network.
Generating randomized networks
Different types of interactions in the integrated network were randomized independently, and then overlaid to generate a randomized multi-color network. For each interaction type, we applied a previously described method [60] to sample from an ensemble of random networks with the property that the expected degree of each node is the same as its degree in the real network. Such a method uniformly samples networks with the same degree sequence. The fugacities – parameters controlling the expected degree for each node [60] – were obtained using the multidimensional Newton-Raphson method.
Links in the network of transcription regulation are directional, originating from the transcriptional regulator and ending at the target gene. We distinguished two types of degree for each node – the in-degree (the number of links that end at the node) and the out-degree (the number of links that originate from the node). We then sampled from an ensemble of random networks [60] such that the expected in-degree and out-degree of each node in the ensemble are the same as the corresponding in-degree and out-degree, respectively, in the real network. Such a randomization procedure preserved the directionality of the transcriptional regulatory links.
Nodes in the SSL network can be divided into three mutually-exclusive categories – genes that were used as both query and array genes in the SGA analysis (denoted as 'query/array' genes), genes that were used only as query genes (denoted as 'query-only' genes), and genes that were used only as array genes (denoted as 'array-only' genes) [4]. Since an SSL link can only exist between a query gene (that is, either a 'query/array' gene or a 'query-only' gene) and an array gene (that is, either a 'query/array' gene or an 'array-only' gene) [4], we decomposed the SSL network into three sub-networks – a 'query/array↔query/array' sub-network containing only links between 'query/array' genes, a 'query-only↔query/array' sub-network containing only links between 'query-only' genes and 'query/array' genes, and a 'query↔array-only' sub-network containing only links between 'query' genes (that is, either 'query/array' or 'query-only' genes) and 'array-only' genes. When randomizing each of the three sub-networks, only links between the specified gene groups were allowed (for example, in the 'query↔array-only' sub-network, only links between 'query' genes and 'array-only' genes were allowed in the randomized network). A randomized SSL network was then generated by overlaying three such random sub-networks, one from each type. The above procedure preserved the inspection bias of the SGA method, and prohibited any link that could never be observed based on the experiment design.
Creating the thematic map of compensatory complexes
To generate a thematic map of compensatory complexes, the integrated protein network containing SSL interaction links from the SGA analysis [4] and stable protein-protein interaction links from the MIPS complex catalog [14] was transformed to a network of protein complexes by merging multiple nodes belonging to the same protein complex into a single node. Nodes that do not belong to any known protein complex were removed, along with their associated SSL links. A few mistakes in the MIPS complex catalog were corrected, and some redundantly listed complexes were merged (for the final list of complexes see Additional data file 1). This generated a multi-graph in which multiple links are allowed between two nodes. For each pair of complexes, we recorded the number of links between them, and calculated the probability of obtaining an equal or greater number of links if each protein were to choose its SSL interaction partners randomly from all eligible proteins. Here two proteins are eligible interaction partners for each other if the pair has been tested by the SGA method [4] and both have at least one observed SSL partner in the transformed network. The nature of the SSL network, introduced from the SGA experiments, complicates the analysis, because interactions were tested only between 'query' genes and each of the 5,000 or so 'array' genes [4]. For each complex, therefore, some links originate with a query gene in the complex and end with a query gene outside the complex, some links connect a query gene within the complex and a non-query gene outside the complex, while others link a non-query gene within the complex and a query gene outside the complex. Hence, each complex has three different degree types, and the total number of links between two complexes follows a distribution corresponding to the sum of three hypergeometric distributions. The p values were calculated based on this composite distribution. A pair of complexes is connected in the map if the p value is less than 0.05 and there are two or more inter-complex SSL links.
Creating the thematic map of regulonic complexes
The integrated protein network containing directed transcriptional regulation links from the genome-wide ChIP study (with a p value threshold of 0.005) [7], and stable protein-protein interaction links from the MIPS complex catalog was transformed to a network of transcription factors and protein complexes by collapsing nodes belonging to the same protein complex into a single node. Pairs of complexes that overlap by more than 50% were merged. This generates a multigraph in which multiple links are allowed between two nodes. For each TF-complex pair, we recorded the number of links between them, and calculated the probability of obtaining at least the same number of links if each node chose its interaction partners randomly. We calculated p values according to the cumulative hypergeometric distribution. A TF-complex pair is connected in the map if the p value is less than 0.05 and there are two or more regulatory links between the TF and the complex.
Additional data files
The following supplementary tables of motifs and protein complexes are provided as Additional data files: Additional data file 1 is a zipped archive containing the five types of biological interactions in the integrated S. cerevisiae network as well as lists of MIPS complexes used to generated Figure 3 and Figure 4; Additional data file 2 lists all three-node interconnection patterns examined; Additional data file 3 lists all four-node interconnection patterns examined; Additional data file 4 lists all complexes in Figure 3; Additional data file 5 lists all the transcription factors in Figure 4; Additional data file 6 lists all protein complexes in Figure 4.
Supplementary Material
Additional data file 1
A zipped archive containing the five types of biological interactions in the integrated S. cerevisiae network as well as lists of MIPS complexes used to generated Figure 3 and Figure 4
Click here for additional data file
Additional data file 2
All three-node interconnection patterns examined
Click here for additional data file
Additional data file 3
All four-node interconnection patterns examined
Click here for additional data file
Additional data file 4
All complexes in Figure 3
Click here for additional data file
Additional data file 5
All the transcription factors in Figure 4
Click here for additional data file
Additional data file 6
All protein complexes in Figure 4
Click here for additional data file
Acknowledgements
We thank G. Berriz, F. Gibbons, M. Umbarger and Z. Wunderlich for critical comments of the manuscript. L.V.Z. was supported by Fu and Ryan Fellowships. O.D.K., S.L.W., and D.S.G. were supported by NRSA (from NHGRI), Ryan, and NSF Fellowships, respectively. In addition, this work was supported by an institutional grant from HHMI (F.P.R.), the Milton Fund of Harvard University (S.L.W. and F.P.R.), and grants from the CIHR (B.A. and C.B.), Genome Canada (B.A., C.B. and H.B.), Genome Ontario (B.A. and C.B), and Genome Quebec (H.B.).
Figures and Tables
Figure 1 Three-node motifs and corresponding themes in the integrated S. cerevisiae network. (a) A motif corresponding to the 'feed-forward' theme; (b) motifs corresponding to the 'co-pointing' theme; (c) motifs corresponding to the 'regulonic complex' theme; (d) motifs corresponding to the 'protein complex' theme; (e) motifs corresponding to the theme of neighborhood clustering of the integrated SSL/homology network; (f) motifs corresponding to the 'compensatory complex members' theme; (g) motifs corresponding to the 'compensatory protein and complex/process' theme; (h) other unclassified motifs. Each of (a-g), from left to right, shows a schematic diagram unifying the collection of motifs in that set, the list of motifs with the motif statistics, a specific example of a subgraph matching one or more of these motifs, and a larger structure corresponding to the network theme. Each colored link represents one of the five interaction types according to the color scheme (bottom right). For a given motif, Nreal is the number of corresponding subgraphs in the real network, and Nrand describes the number of corresponding subgraphs in a randomized network, represented by the average and the standard deviation. A node labeled 'etc.' signifies that the structure contains more nodes with connectivity similar to the labeled node.
Figure 2 Four-node network motifs corresponding to the 'compensatory complexes/processes' theme. (a) A schematic diagram unifying the collection of four-node motifs corresponding to the 'compensatory complexes/processes' theme; (b) examples of specific four-node motifs together with the motif statistics; (c) a specific example of a four-node subgraph matching a few of these motifs; (d) the larger structure corresponding to the network theme. Each colored link represents one of the four interaction types according to the color scheme (see key). For a given motif, Nreal is the number of corresponding subgraphs in the real network, and Nrand describes the number of corresponding subgraphs in a randomized network, represented by the average and the standard deviation.
Figure 3 A thematic map of compensatory complexes. Here, nodes represent protein complexes, and a link is drawn between two nodes if there is a significantly large number of inter-complex SSL interactions. Links between compensatory complexes are labeled with the numbers of supporting SSL interactions.
Figure 4 A thematic map of regulonic complexes. (a) Here, blue nodes represent transcription factors, red nodes represent protein complexes, and a link is drawn between a transcription factor and a protein complex if the promoters of a significantly large number of complex members are bound by the transcription factor. (b) An enlarged region of the regulonic complex map in (a). Links between transcription factors and the complexes they regulate are labeled with the numbers of supporting interactions in the transcription regulation network. For lists of transcription factors and complexes in the map see Additional data files 5 and 6.
==== Refs
Ge H Liu Z Church GM Vidal M Correlation between transcriptome and interactome mapping data from Saccharomyces cerevisiae Nat Genet 2001 29 482 486 11694880 10.1038/ng776
Jansen R Greenbaum D Gerstein M Relating whole-genome expression data with protein-protein interactions Genome Res 2002 12 37 46 11779829 10.1101/gr.205602
Yu H Luscombe NM Qian J Gerstein M Genomic analysis of gene expression relationships in transcriptional regulatory networks Trends Genet 2003 19 422 427 12902159 10.1016/S0168-9525(03)00175-6
Tong AH Lesage G Bader GD Ding H Xu H Xin X Young J Berriz GF Brost RL Chang M Global mapping of the yeast genetic interaction network Science 2004 303 808 813 14764870 10.1126/science.1091317
Milo R Shen-Orr S Itzkovitz S Kashtan N Chklovskii D Alon U Network motifs: simple building blocks of complex networks Science 2002 298 824 827 12399590 10.1126/science.298.5594.824
Shen-Orr SS Milo R Mangan S Alon U Network motifs in the transcriptional regulation network of Escherichia coli Nat Genet 2002 31 64 68 11967538 10.1038/ng881
Lee TI Rinaldi NJ Robert F Odom DT Bar-Joseph Z Gerber GK Hannett NM Harbison CT Thompson CM Simon I Transcriptional regulatory networks in Saccharomyces cerevisiae Science 2002 298 799 804 12399584 10.1126/science.1075090
Milo R Itzkovitz S Kashtan N Levitt R Shen-Orr S Ayzenshtat I Sheffer M Alon U Superfamilies of evolved and designed networks Science 2004 303 1538 1542 15001784 10.1126/science.1089167
Yeger-Lotem E Sattath S Kashtan N Itzkovitz S Milo R Pinter RY Alon U Margalit H Network motifs in integrated cellular networks of transcription-regulation and protein-protein interaction Proc Natl Acad Sci USA 2004 101 5934 5939 15079056 10.1073/pnas.0306752101
Dobrin R Beg QK Barabasi AL Oltvai ZN Aggregation of topological motifs in the Escherichia coli transcriptional regulatory network BMC Bioinformatics 2004 5 10 15018656 10.1186/1471-2105-5-10
Kashtan N Itzkovitz S Milo R Alon U Topological generalizations of network motifs Phys Rev E Stat Nonlin Soft Matter Phys 2004 70 031909 15524551
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 Nucleic Acids Res 1997 25 3389 3402 9254694 10.1093/nar/25.17.3389
Hughes TR Marton MJ Jones AR Roberts CJ Stoughton R Armour CD Bennett HA Coffey E Dai H He YD Functional discovery via a compendium of expression profiles Cell 2000 102 109 126 10929718 10.1016/S0092-8674(00)00015-5
Mewes HW Frishman D Guldener U Mannhaupt G Mayer K Mokrejs M Morgenstern B Munsterkotter M Rudd S Weil B MIPS: a database for genomes and protein sequences Nucleic Acids Res 2002 30 31 34 11752246 10.1093/nar/30.1.31
Gavin AC Bosche M Krause R Grandi P Marzioch M Bauer A Schultz J Rick JM Michon AM Cruciat CM Functional organization of the yeast proteome by systematic analysis of protein complexes Nature 2002 415 141 147 11805826 10.1038/415141a
Ho Y Gruhler A Heilbut A Bader GD Moore L Adams SL Millar A Taylor P Bennett K Boutilier K Systematic identification of protein complexes in Saccharomyces cerevisiae by mass spectrometry Nature 2002 415 180 183 11805837 10.1038/415180a
Wagner A The yeast protein interaction network evolves rapidly and contains few redundant duplicate genes Mol Biol Evol 2001 18 1283 1292 11420367
Solé RV Pastor-Satorras R Smith E Kepler TB A model of large-scale proteome evolution Adv Complex Systems 2002 5 43 54 10.1142/S021952590200047X
Mai B Miles S Breeden LL Characterization of the ECB binding complex responsible for the M/G(1)-specific transcription of CLN3 and SWI4 Mol Cell Biol 2002 22 430 441 11756540 10.1128/MCB.22.2.430-441.2002
Fitch MJ Donato JJ Tye BK Mcm7, a subunit of the presumptive MCM helicase, modulates its own expression in conjunction with Mcm1 J Biol Chem 2003 278 25408 25416 12738768 10.1074/jbc.M300699200
Simon I Barnett J Hannett N Harbison CT Rinaldi NJ Volkert TL Wyrick JJ Zeitlinger J Gifford DK Jaakkola TS Serial regulation of transcriptional regulators in the yeast cell cycle Cell 2001 106 697 708 11572776 10.1016/S0092-8674(01)00494-9
McNabb DS Xing Y Guarente L Cloning of yeast HAP5: a novel subunit of a heterotrimeric complex required for CCAAT binding Genes Dev 1995 9 47 58 7828851
Gancedo JM Yeast carbon catabolite repression Microbiol Mol Biol Rev 1998 62 334 361 9618445
Zhang LV Wong SL King OD Roth FP Predicting co-complexed protein pairs using genomic and proteomic data integration BMC Bioinformatics 2004 5 38 15090078 10.1186/1471-2105-5-38
Spector MS Raff A DeSilva H Lee K Osley MA Hir1p and Hir2p function as transcriptional corepressors to regulate histone gene transcription in the Saccharomyces cerevisiae cell cycle Mol Cell Biol 1997 17 545 552 9001207
Arnold I Pfeiffer K Neupert W Stuart RA Schagger H ATP synthase of yeast mitochondria. Isolation of subunit j and disruption of the ATP18 gene J Biol Chem 1999 274 36 40 9867807 10.1074/jbc.274.1.36
Arnold I Pfeiffer K Neupert W Stuart RA Schagger H Yeast mitochondrial F1F0-ATP synthase exists as a dimer: identification of three dimer-specific subunits EMBO J 1998 17 7170 7178 9857174 10.1093/emboj/17.24.7170
Liu H Bretscher A Characterization of TPM1 disrupted yeast cells indicates an involvement of tropomyosin in directed vesicular transport J Cell Biol 1992 118 285 299 1629236 10.1083/jcb.118.2.285
Wang T Bretscher A The rho-GAP encoded by BEM2 regulates cytoskeletal structure in budding yeast Mol Biol Cell 1995 6 1011 1024 7579704
Myer VE Young RA RNA polymerase II holoenzymes and subcomplexes J Biol Chem 1998 273 27757 27760 9774381 10.1074/jbc.273.43.27757
Mueller CL Jaehning JA Ctr9, Rtf1, and Leo1 are components of the Paf1/RNA polymerase II complex Mol Cell Biol 2002 22 1971 1980 11884586 10.1128/MCB.22.7.1971-1980.2002
Geissler S Siegers K Schiebel E A novel protein complex promoting formation of functional alpha- and gamma-tubulin EMBO J 1998 17 952 966 9463374 10.1093/emboj/17.4.952
Hanein D Matlack KE Jungnickel B Plath K Kalies KU Miller KR Rapoport TA Akey CW Oligomeric rings of the Sec61p complex induced by ligands required for protein translocation Cell 1996 87 721 732 8929540 10.1016/S0092-8674(00)81391-4
Simonis N van Helden J Cohen GN Wodak SJ Transcriptional regulation of protein complexes in yeast Genome Biol 2004 5 R33 15128447 10.1186/gb-2004-5-5-r33
Mangan S Zaslaver A Alon U The coherent feedforward loop serves as a sign-sensitive delay element in transcription networks J Mol Biol 2003 334 197 204 14607112 10.1016/j.jmb.2003.09.049
Jansen R Yu H Greenbaum D Kluger Y Krogan NJ Chung S Emili A Snyder M Greenblatt JF Gerstein M A Bayesian networks approach for predicting protein-protein interactions from genomic data Science 2003 302 449 453 14564010 10.1126/science.1087361
Jansen R Lan N Qian J Gerstein M Integration of genomic datasets to predict protein complexes in yeast J Struct and Funct Genomics 2002 2 71 81 12836664 10.1023/A:1020495201615
Wong SL Zhang LV Tong AH Li Z Goldberg DS King OD Lesage G Vidal M Andrews B Bussey H Combining biological networks to predict genetic interactions Proc Natl Acad Sci USA 2004 101 15682 15687 15496468 10.1073/pnas.0406614101
Goldberg DS Roth FP Assessing experimentally derived interactions in a small world Proc Natl Acad Sci USA 2003 100 4372 4376 12676999 10.1073/pnas.0735871100
Bader GD Hogue CW An automated method for finding molecular complexes in large protein interaction networks BMC Bioinformatics 2003 4 2 12525261 10.1186/1471-2105-4-2
Albert I Albert R Conserved network motifs allow protein-protein interaction prediction Bioinformatics 2004 20 3346 3352 15247093
King AD Przulj N Jurisica I Protein complex prediction via cost-based clustering Bioinformatics 2004 20 3013 3020 15180928 10.1093/bioinformatics/bth351
Usui T Ohta T Oshiumi H Tomizawa J Ogawa H Ogawa T Complex formation and functional versatility of Mre11 of budding yeast in recombination Cell 1998 95 705 716 9845372 10.1016/S0092-8674(00)81640-2
Solinger JA Lutz G Sugiyama T Kowalczykowski SC Heyer WD Rad54 protein stimulates heteroduplex DNA formation in the synaptic phase of DNA strand exchange via specific interactions with the presynaptic Rad51 nucleoprotein filament J Mol Biol 2001 307 1207 1221 11292336 10.1006/jmbi.2001.4555
Majka J Burgers PM Yeast Rad17/Mec3/Ddc1: a sliding clamp for the DNA damage checkpoint Proc Natl Acad Sci USA 2003 100 2249 2254 12604797 10.1073/pnas.0437148100
Lobachev KS Gordenin DA Resnick MA The Mre11 complex is required for repair of hairpin-capped double-strand breaks and prevention of chromosome rearrangements Cell 2002 108 183 193 11832209 10.1016/S0092-8674(02)00614-1
Kondo T Matsumoto K Sugimoto K Role of a complex containing Rad17, Mec3, and Ddc1 in the yeast DNA damage checkpoint pathway Mol Cell Biol 1999 19 1136 1143 9891048
Kadosh D Struhl K Repression by Ume6 involves recruitment of a complex containing Sin3 corepressor and Rpd3 histone deacetylase to target promoters Cell 1997 89 365 371 9150136 10.1016/S0092-8674(00)80217-2
Zhang Y Sun ZW Iratni R Erdjument-Bromage H Tempst P Hampsey M Reinberg D SAP30, a novel protein conserved between human and yeast, is a component of a histone deacetylase complex Mol Cell 1998 1 1021 1031 9651585 10.1016/S1097-2765(00)80102-1
Grant PA Schieltz D Pray-Grant MG Steger DJ Reese JC Yates JR 3rdWorkman JL A subset of TAF(II)s are integral components of the SAGA complex required for nucleosome acetylation and transcriptional stimulation Cell 1998 94 45 53 9674426 10.1016/S0092-8674(00)81220-9
Kadonaga JT Eukaryotic transcription: an interlaced network of transcription factors and chromatin-modifying machines Cell 1998 92 307 313 9476891 10.1016/S0092-8674(00)80924-1
Norden C Liakopoulos D Barral Y Dissection of septin actin interactions using actin overexpression in Saccharomyces cerevisiae Mol Microbiol 2004 53 469 483 15228528 10.1111/j.1365-2958.2004.04148.x
Roumanie O Peypouquet MF Bonneu M Thoraval D Doignon F Crouzet M Evidence for the genetic interaction between the actin-binding protein Vrp1 and the RhoGAP Rgd1 mediated through Rho3p and Rho4p in Saccharomyces cerevisiae Mol Microbiol 2000 36 1403 1414 10931290 10.1046/j.1365-2958.2000.01958.x
Roumanie O Peypouquet MF Thoraval D Doignon F Crouzet M Functional interactions between the VRP1-LAS17 and RHO3-RHO4 genes involved in actin cytoskeleton organization in Saccharomyces cerevisiae Curr Genet 2002 40 317 325 11935222 10.1007/s00294-001-0268-5
Breton AM Aigle M Genetic and functional relationship between Rvsp, myosin and actin in Saccharomyces cerevisiae Curr Genet 1998 34 280 286 9799361 10.1007/s002940050397
Marcoux N Cloutier S Zakrzewska E Charest PM Bourbonnais Y Pallotta D Suppression of the profilin-deficient phenotype by the RHO2 signaling pathway in Saccharomyces cerevisiae Genetics 2000 156 579 592 11014808
Giot L Chanet R Simon M Facca C Faye G Involvement of the yeast DNA polymerase delta in DNA repair in vivo Genetics 1997 146 1239 1251 9258670
Chanet R Heude M Characterization of mutations that are synthetic lethal with pol3-13, a mutated allele of DNA polymerase delta in Saccharomyces cerevisiae Curr Genet 2003 43 337 350 12759774 10.1007/s00294-003-0407-2
Ooi SL Shoemaker DD Boeke JD DNA helicase gene interaction network defined using synthetic lethality analyzed by microarray Nat Genet 2003 35 277 286 14566339 10.1038/ng1258
Park J Newman MEJ Origin of degree correlations in the Internet and other networks Phys Rev E Stat Nonlin Soft Matter Phys 2003 68 026112 14525054
| 15982408 | PMC1175995 | CC BY | 2021-01-04 16:05:45 | no | J Biol. 2005 Jun 1; 4(2):6 | utf-8 | J Biol | 2,005 | 10.1186/jbiol23 | oa_comm |
==== Front
J BiolJournal of Biology1478-58541475-4924BioMed Central London jbiol241598515010.1186/jbiol24Research ArticleGuanine-nucleotide exchange on ribosome-bound elongation factor G initiates the translocation of tRNAs Zavialov Andrey V 1Hauryliuk Vasili V 1Ehrenberg Måns [email protected] Department of Cell and Molecular Biology, Biomedical Center, Uppsala University, SE-75124 Uppsala, Sweden2005 27 6 2005 4 2 9 9 17 1 2005 23 3 2005 19 4 2005 Copyright © 2005 Zavialov 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 the translation of mRNA into polypeptide, elongation factor G (EF-G) catalyzes the translocation of peptidyl-tRNA from the A site to the P site of the ribosome. According to the 'classical' model, EF-G in the GTP-bound form promotes translocation, while hydrolysis of the bound GTP promotes dissociation of the factor from the post-translocation ribosome. According to a more recent model, EF-G operates like a 'motor protein' and drives translocation of the peptidyl-tRNA after GTP hydrolysis. In both the classical and motor protein models, GDP-to-GTP exchange is assumed to occur spontaneously on 'free' EF-G even in the absence of a guanine-nucleotide exchange factor (GEF).
Results
We have made a number of findings that challenge both models. First, free EF-G in the cell is likely to be in the GDP-bound form. Second, the ribosome acts as the GEF for EF-G. Third, after guanine-nucleotide exchange, EF-G in the GTP-bound form moves the tRNA2-mRNA complex to an intermediate translocation state in which the mRNA is partially translocated. Fourth, subsequent accommodation of the tRNA2-mRNA complex in the post-translocation state requires GTP hydrolysis.
Conclusion
These results, in conjunction with previously published cryo-electron microscopy reconstructions of the ribosome in various functional states, suggest a novel mechanism for translocation of tRNAs on the ribosome by EF-G. Our observations suggest that the ribosome is a universal guanosine-nucleotide exchange factor for EF-G as previously shown for the class-II peptide-release factor 3.
==== Body
Background
During the translation of protein, in every peptide elongation cycle, one aminoacyl-tRNA arrives at and binds to the A site of the ribosome. Then, peptidyl transfer brings the ribosome to its pre-translocation (preT) state, with a peptidyl-tRNA in the A site (Figure 1a,b). Subsequent translocation of the complex comprising two charged tRNAs and the mRNA – the tRNA2-mRNA complex – to the post-translocation state (postT) (Figure 1c) completes the elongation cycle. In bacteria, translocation of peptidyl-tRNA from the A site to the P site of the ribosome is catalyzed by elongation factor EF-G (Figure 1b,c). Like the ribosomal GTPases RF3, EF-Tu and IF2, EF-G belongs to the family of small GTPases [1]. Conserved features of the GTP-binding domain of these protein factors are responsible for their function as molecular switches [2]. In the active GTP-bound conformation, the GTPases bind tightly to their targets. After GTP hydrolysis, they adopt an inactive GDP-bound conformation, and dissociate rapidly from their targets [1]. Such GTPases usually require a guanine-nucleotide exchange factor (GEF), which catalyzes the exchange of GDP to GTP, and a GTPase-activating protein (GAP), which stimulates GTP hydrolysis [2]. In the case of EF-G, the role of GAP has been ascribed to the ribosomal L7/L12 stalk [3]. No GEF has so far been identified for EF-G, however, and it has been postulated that rapid and extensive exchange of GDP to GTP occurs spontaneously on free EF-G [3]. Accordingly, it has been assumed that EF-G is in the GTP-bound form as it enters the ribosome, although this structure has eluded detection in solution [4], and has only been observed in ribosomal complexes [5].
According to the 'classical' model, the binding of EF-G•GTP to the preT ribosome complex (Figure 1b) promotes translocation of the peptidyl-tRNA from the A to the P site. Then, GTP hydrolysis removes the EF-G from the postT ribosome [4,6]. Recent experiments, suggesting that GTP hydrolysis on EF-G precedes translocation and that EF-G together with GDP can promote rapid translocation, have led to the contrasting suggestion that EF-G is in fact a 'motor protein' that drives translocation with the energy liberated by GTP hydrolysis [7]. Previously, we showed that the postT ribosome complex has low affinity for EF-G•GTP [8], presumably as a result of the inability of a peptidyl-tRNA to be accommodated in a hybrid P/E tRNA site, where the CCA-end of the tRNA is in the E site of the large ribosomal subunit, and the anticodon-end of the tRNA is in the P site of the small ribosomal subunit. This effectively prevents formation of the 'twisted' ribosome conformation [5] with a high affinity for the GTP form of EF-G. These results also show that translocation cannot be carried out by EF-G and GDP, in line with the notion that EF-G, like other small GTPases, has an active GTP- and inactive GDP-bound form [1].
In this study, we challenge current ideas about the mechanism of translocation. The paradigm shift we propose follows from our observations that intracellular EF-G is likely to be in the GDP-bound form, that the GDP form of the factor can rapidly enter the preT ribosome complex, and that the preT ribosome acts as the GEF for EF-G, similar to the way that the post-termination ribosome acts as the GEF for the peptide-release factor RF3 [9,10]. Our results, partially based on the use of A-site-specific cleavage of mRNA by the bacterial toxin RelE [11] to monitor the position of the mRNA at various translocation steps, show that the exchange of GDP for the non-cleavable GTP analog GDPNP on EF-G bound to the preT complex drives the ribosome into an intermediate translocation state (transT*), wherein the tRNA2-mRNA complex has moved in relation to the 30S subunit. The removal of EF-G•GDPNP from a transT* ribosome by addition of excess GDP brings the ribosome back to its preT state, while GTP addition brings it to the postT state. From these and previous biochemical data [8], in conjunction with cryo-electron microscopy (cryo-EM) reconstructions of functional ribosomal complexes [5], we provide a mechanistic reinterpretation of the major steps of translocation.
Results
The ribosome is the missing guanine-nucleotide exchange factor for EF-G
It has been reported that EF-G from Escherichia coli binds to GTP with ten-fold lower affinity than it binds to GDP [12]. On the assumption that there is a ten-fold excess of GTP over GDP in the cytoplasm and rapid nucleotide exchange on free EF-G, it was suggested that the rate-limiting step of guanine-nucleotide exchange in the EF-G cycle is the dissociation of EF-G•GDP from the postT ribosome [3].
Our earlier data, showing the ribosome to be a GEF for RF3 [9], prompted us to re-check the binding of free EF-G to GDP or GTP. The dissociation constant (KD) for the EF-G•GDP complex was about 9 μM (Figure 2a), close to an earlier estimate of 4 μM [12]. Results from experiments in which [3H]-GDP in complex with EF-G was chased with unlabeled and further purified GTP [9] (see below and Figure 4a for purification details), however, show a 60-fold larger effective KD-value for the binding of EF-G to GTP than to GDP (Figure 2b). This factor of 60 provides a lower boundary to the correct value, because purified GTP solutions do contain some fraction of GDP from the hydrolysis of GTP. The intracellular GTP:GDP ratio has been estimated as 7:1 for Salmonella enterica serovar Typhimurium [13], and is probably similar in E. coli. This suggests that a major fraction of free EF-G in E. coli is bound to GDP.
If binding of EF-G to the pre-translocation (preT) ribosome required the factor to be bound to GTP, this would significantly reduce the rate of association of EF-G with the ribosome. This problem would, however, be eliminated if EF-G in the GDP-bound form associated rapidly with the ribosome and GDP-to-GTP exchange took place on, rather than off, the ribosome. To test the latter two hypotheses, we prepared preT ribosomes with fMet-Ile-tRNAIle and its corresponding codon in the A site and a UAA stop codon immediately downstream from the Ile codon. Translocation was catalyzed by EF-G at such a small concentration that each EF-G molecule had to cycle many times to obtain a significant fraction of translocated ribosomes. The concentration of GTP was fixed at 0.5 mM during incubations with varying concentrations of GDP, and the ribosome concentration chosen was sufficiently low that the rate of translocation per ribosome was approximated by the concentration of free EF-G multiplied by its effective association rate constant (kcat/Km) for ribosome binding (see Materials and methods). Because translocation brought the stop codon UAA into the ribosomal A site, the extent of translocation was conveniently quantified as the fraction of fMet-Ile peptide that could be rapidly released by RF2, when RF2 was added to a concentration in excess of that of the ribosomes at varying incubation times (Figure 2c).
We obtained 50% inhibition of the rate of EF-G recycling at 0.25 mM GDP, at which concentration the concentration of EF-G•GDP (KD = 9 μM) must have been at least 30 times larger than the concentration of EF-G•GTP (KD > 0.6 mM). If entry of EF-G to the ribosome had required the EF-G•GTP complex, this would have led to a 30-fold, rather than the observed two-fold, inhibition of translocation at 0.25 mM GDP (see Materials and methods). This implies that EF-G must have entered the ribosome in complex with GDP, and that the exchange of GDP for GTP must have taken place on, rather than off, the ribosome. The parameters that determine how the kcat/Km value for the entry of EF-G to the preT ribosome complex depends on varying ratios of GDP to GTP are defined in Materials and methods for a particular kinetic scheme.
The preT ribosome contains a deacylated tRNA in the P site (Figure 1b), which may be important for the GDP-to-GTP exchange reaction. This is suggested by experiments on guanine-nucleotide binding to EF-G in another type of ribosome complex. Here, EF-G was incubated with [3H]-GDPNP and either post-termination (postTerm) or naked ribosomes at varying concentrations of unlabeled GDP (Figure 2d). The postTerm ribosome has a deacylated tRNA in the P site and an empty A site programmed with a stop codon (Figure 1d), while the naked ribosome lacks ligands. The fraction of [3H]-GDPNP retained on a nitrocellulose filter, corresponding to ribosome-bound EF-G• [3H]-GDPNP, was reduced to 50% at a 160-fold excess of GDP in the postTerm case, or a 13-fold excess for the naked ribosomes. This implies that EF-G, bound to either type of ribosome, had much higher affinity for GDPNP than for GDP, and that the difference was more pronounced for postTerm than for naked ribosomes (Figure 2d,e). Accordingly, the presence of a deacylated tRNA in the P site of the preT ribosome led to more stable binding of EF-G• [3H]-GDPNP to this complex than to the naked ribosome. A corresponding stabilization of the EF-G•GTP complex on preT ribosomes by the P-site tRNA is expected, and would contribute to efficient guanine-nucleotide exchange (Figure 2c).
So far, we have not addressed the question of whether formation of a complex between EF-G•GDP and the preT ribosome leads directly to guanosine exchange, or whether the exchange reaction is preceded by a change in conformation of the ribosome. This problem is addressed in the next section.
EF-G•GDP drives the preT ribosome into a state that has hybrid tRNA sites
EF-G•GTP binds poorly to the pre-termination (preTerm) ribosome with a peptidyl-tRNA in the P site and an empty A site programmed with a stop codon (Figure 1c), but binds with high affinity to the postTerm ribosome with a deacylated tRNA in the P site [8] (Figure 1d). In the latter case, cryo-EM results show the postTerm ribosome in a ratcheted state with the P-site tRNA in the hybrid P/E site [5]. This suggests that high-affinity binding of EF-G•GTP to the ribosome requires the ratcheted state with hybrid tRNA sites; this state cannot be formed when there is peptidyl-tRNA in the P site. It is likely that the ratcheted ribosome conformation appears also in the translocation process, suggesting that EF-G•GDP can move the preT ribosome from the relaxed state, with three full binding sites for the tRNAs [5], to the ratcheted state, with no E site binding and only two binding sites for tRNA [14]. This would facilitate rapid GDP-to-GTP exchange on EF-G, and we have tested one of the predictions that emerges from this hypothesis, namely that the apparent affinity of a deacylated tRNA for the E site of the preT ribosome will be reduced by the addition of EF-G•GDP. This prediction was confirmed by an experiment showing that the affinity of tRNAfMet for the E site of the preT ribosome was successively reduced by increasing amounts of EF-G in the presence of GDP (Table 1, set 1).
In order to monitor the translocation events that follow guanine-exchange on EF-G on the preT ribosome, we used A-site-specific cleavage of the mRNA by the bacterial toxin RelE, and this is described next.
Translocation events monitored by RelE cleavage of the A-site codon
RelE cuts mRNA specifically within the ribosomal A site [11], and we used this activity to monitor ribosome movement along mRNA in the translocation steps (Figure 1). An initiation complex (Init; Figure 1a) with fMet-tRNAfMet in the P site was constituted by incubating ribosomes in the presence of initiation factors IF1, IF2 and IF3, fMet-tRNAfMet, and 33P-end-labeled mRNA encoding the dipeptide Met-Ile-stop (AUG AUU UAA). Exposure of this complex to RelE led to unique cleavage of the A-site codon to AU*U (Figure 3a, lane 2). The Init complex (Figure 1a) was then converted to the preT complex (Figure 1b) by addition of the ternary EF-Tu•GTP•Ile-tRNAIle complex. The resulting presence of fMet-Ile-tRNAIle in the A site blocked the entry of RelE to the A site and reduced the rate of cleavage of the AUU codon (Figure 3a, lane 3). Addition of EF-G•GTP to the preT complex catalyzed rapid translocation of fMet-Ile-tRNAIle from the A to the P site, generating the postT complex (Figure 1c), and moved the stop codon into the A site of the postT complex, where it was rapidly cleaved by RelE (Figure 3a, line 4).
Complete translocation requires GTP and GTP hydrolysis
In order to study further the guanine-nucleotide dependence of the translocation steps, the ribosomal preT complex was first separated from all other components of the translation mixture [8]. RelE cleavage of the A-site codon was monitored after addition of EF-G to the purified preT complex in the presence of GTP, GDP or the non-cleavable GTP analog GDPNP (Figure 3b). In one type of experiment, the preT complex was first incubated with EF-G and either GTP or GDP for 10, 25 or 40 min and then the ribosomes were exposed to RelE for 5 min. In the presence of GTP, there was extensive cleavage by RelE of the stop codon (Figure 3b, + GTP), meaning that a major fraction of the ribosomes had moved from the preT to the postT state.
In the presence of GDP, there was no significant RelE-dependent cleavage of the stop codon in the A site, even during the longest incubation time of 45 minutes (Figure 3b, + GDP), meaning that the ribosomes had remained in their preT state during the whole incubation period. This implies that EF-G and GDP were unable to promote translocation, in apparent contradiction to previous results, showing rapid translocation by EF-G and GDP [7]. We have noted that GTP contamination, common in commercial preparations of GDP, can have profound effects on the GTPases of protein synthesis. A typical elution profile (Figure 4a) shows such a GDP preparation to contain between 1 and 2% GTP, and the effect of this low level of contamination was studied in an experiment in which translocation of fMet-[14C]-Ile-tRNA from the A site to the P site was probed by the fraction of peptide that could be rapidly released by RF2. The rate of translocation was insignificant with purified GDP, intermediate with unpurified GDP or with purified GDP + 2% GTP and fast with GTP (Figure 4b). Similarly, no translocation with pure GDP was detected by assessing the RelE-dependent cleavage of the mRNA (Figure 4c). Our nucleotide preparations were further purified by ion exchange chromatography on a MonoQ column [9], while those of Rodnina et al. [7] were not. This suggests that their 'GDP-dependent translocation' was, in fact, due to contaminating GTP. At such a large excess of GDP, the guanine-exchange reaction on the preT ribosome is expected to be the rate-limiting step for translocation, and this will lead to slow, monophasic translocation, exactly as they observed (see Materials and methods) [7].
In the presence of GDPNP, about 11% of the stop codons were cleaved after addition of RelE, irrespective of the time of exposure of preT ribosomes to EF-G and GDPNP (Figure 3b, + GDPNP 2). In a similar experiment, modified so that RelE was present from the start of the incubation of preT ribosomes with EF-G and GDPNP, the fraction of cleaved stop codons increased slowly with time (Figure 3b, + GDPNP 1). This means that EF-G and GDPNP drove the ribosomes to a state that remained stable during the 45 min incubation in the absence of RelE (Figure 3b, + GDPNP 1). In this state, the stop codon was partially available for RelE-mediated cleavage in the A site, resulting in very slow truncation of the mRNA (Figure 3b, + GDPNP 2). A priori, this ribosomal state could be the postT state of the ribosome or a novel transition state ('transT*') in the translocation process where, in both cases, RelE-mediated cleavage of the stop codon in the A site was inhibited by ribosome-bound EF-G•GDPNP. An experiment in which the rates of RelE cleavage in the A-site codons of ribosomes in the putatively new state and postT ribosomes were compared at the same concentrations of EF-G and GDPNP (Figure 3c) showed that RelE cleaved the mRNA in the postT complex much faster than the mRNA on the ribosomes in the unknown state complex, proving that the ribosomal complexes could not have been the same. This means that the unknown state was transT*, and in the next section we characterize these complexes with respect to tRNA-exchangeability.
Exchangeability of tRNAfMet in preT, transT* and postT ribosomes
We characterized the transT* state with respect to the exchangeability of its deacylated tRNAfMet. First, we used nitrocellulose filtration to study dissociation of [33P]-tRNAfMet, originally in the P site of the preT complex (Figure 1b), from ribosomes incubated with EF-G together with GDP, GTP or GDPNP. In one type of experiment, the fraction of ribosome-bound [33P]-tRNAfMet was monitored as a function of time in the presence of either unlabeled tRNAfMet or tRNAPhe at fixed concentrations (Figure 5a). In another type of experiment, the fraction of ribosome-bound [33P]-tRNAfMet was monitored at a fixed time while varying the concentrations of unlabeled tRNAfMet or tRNAPhe (Figure 5b).
In the GDP experiment in which no translocation occurred (Figure 3b, + GDP), there was no significant removal of [33P]-tRNAfMet from the ribosome during 6 min in the presence of any unlabeled tRNA, as would be expected for ribosomes with deacylated tRNAfMet stably bound to the P site after peptidyl transfer (Figure 5a,b). In the GTP case, in which there was rapid translocation (Figure 3b, + GTP), there was fast dissociation of [33P]-tRNAfMet in the presence of either tRNAfMet or tRNAPhe (Figure 5a). The titration experiment (Figure 5b) shows that one fraction of [33P]-tRNAfMet dissociated from the postT ribosomes in the absence of chasing tRNAs, and that the remaining fraction could be titrated out with either tRNAfMet or tRNAPhe. These results reflect the comparatively low affinity of [33P]-tRNAfMetfor the E site and the lack of codon specificity for the E-site-bound tRNAs ([15]; see also below).
In the case of GDPNP, [33P]-tRNAfMet dissociated slowly in the presence of tRNAfMet, but there was no dissociation in the presence of tRNAPhe, suggesting high affinity for [33P]-tRNAfMet and retained codon-specificity for deacylated tRNAs (Figure 5a). In line with this, the titration experiment (Figure 5b) shows that [33P]-tRNAfMet could be exchanged with unlabeled tRNAfMet but not with unlabeled tRNAPhe.
In a third type of experiment, [33P]-tRNAfMet was chased with unlabeled tRNAfMet from preT ribosomes incubated for a fixed amount of time in the presence of EF-G at a constant concentration and GDPNP at varying concentrations (Figure 5c). The fraction of ribosomes lacking [33P]-tRNAfMet increased from 0 to 50% when GDPNP was varied from 0 to 40 μM and increased further to almost 100% at 250 μM GDP. This result shows that the affinity of EF-G•GDPNP for the transT* ribosome, containing one deacylated and one peptidyl tRNA, was approximately 100 times weaker than the affinity of EF-G•GDPNP for the postTerm ribosome, containing only one deacylated tRNA (see below)[8].
Another experiment (Figure 5d) shows that tRNAPhe could not replace [33P]-tRNAfMet in transT* ribosomes, either with intact mRNA or with mRNA that had been cleaved by RelE. This means that the transT* ribosomes did not move to the postT state as a result of the mRNA cleavage, since that would have resulted in weak, non-selective E-site binding of the deacylated tRNAs (as shown in Table 1 and Figure 7).
Addition of GDP to transT* ribosomes brings them back to the preT state
When GDP was added to transT* ribosomes, on which we have observed RelE-mediated cleavage of the stop codon to UA*A (Figure 3b, + GDPNP 1; and Figure 6a, + GDPNP), stop codon cleavage was completely eliminated and replaced by cleavage of the AUU codon (Figure 6a, + GDPNP + GDP). The latter cleavage reaction was typical for the preT ribosome and occurred when the peptidyl-tRNA dissociated from the A site (Figure 3a). When, in contrast, GDP was added to postT ribosomes that were incubated in the presence of EF-G and GDPNP, the ribosomes remained in the postT state and there was rapid cleavage of the UAA codon (data not shown). These results strongly suggest that addition of GDP to the transT* ribosome brought it back to the preT state, providing further evidence that the transT* state is different from the postT state of the ribosome.
In line with previous results [8], addition of EF-G•GDPNP to preT ribosomes brought them to a puromycin-reactive state (Figure 6b); puromycin mimics an aminoacyl tRNA and removes a nascent peptide from the ribosome by acting as a receptor in peptidyl-transfer. When GDP was also included, however, the puromycin-reactivity of the ribosomes was lost (Figure 6b), again showing that the resulting state could not have been the postT ribosome, which is fully reactive to puromycin [9].
A deacylated [33P]-tRNAfMet in the transT* ribosome could readily be chased with unlabeled tRNAfMet, but its exchange rate in the preT ribosome was almost zero (Figure 5a,b). If GDP addition brought the transT* ribosome back to the preT state, one would therefore expect the exchange rate of the tRNAfMet to drop drastically. This prediction was nicely confirmed by experiments showing that addition of GDP to transT* ribosomes did indeed prevent exchange of [33P]-tRNAfMet with tRNAfMet (Figure 6d).
When release factor RF2 was added to transT* ribosomes, there was slow release of peptide (Figure 6c), suggesting that there was partial availability of the UAA stop codon in the A site, a necessary condition for termination by class-1 release factors [8]. Addition of GDP to transT* ribosomes made them non-reactive not only to puromycin (Figure 6b), but also to peptide release induction by RF2 (Figure 6c). These mRNA cleavage results (Figure 6a), along with those for puromycin (Figure 6b), RF2 (Figure 6c) and tRNA exchange (Figure 6d) show that removal of EF-G•GDPNP from the transT* ribosome by the addition of GDP brought the ribosome back to the preT state with peptidyl-tRNA in the A site. This confirms that the transT* state cannot be identical to the postT state of the ribosome, and corroborates that transT* is a transition state in the translocation process, in which rapid hydrolysis of native GTP on EF-G normally occurs. When EF-G dissociated from the transT* ribosome, the mRNA rapidly slipped back to its preT position, but there was a short time during which RelE could cleave and RF1 could interact with the stop codon exposed in an EF-G-free A site.
Deacylated tRNAs bind to the ribosomal E site with low codon specificity
We showed above that [33P]-tRNAfMet could be chased by tRNAfMet but not by tRNAPhe in transT* (Figure 5d). This contrasts with E-site binding of deacylated tRNA, as follows. We designed experiments to obtain dissociation constants for the binding of deacylated tRNAfMet or tRNAPhe to the E site of postT ribosomes, programmed with Met (AUG), Phe (UUU) or Thr (ACG) codons. The binding of [33P]-tRNAfMet to the E site was assayed by nitrocellulose filtration, and a representative experiment with the Thr (ACG) codon in the E site is shown in Figure 7a. Dissociation constants for the binding of tRNAPhe or tRNAThr to the differently programmed E sites of postT ribosomes were obtained as I50 values in competition experiments with a constant and almost saturating concentration of [33P]-tRNAfMet and varying concentrations of unlabeled tRNAPhe or tRNAThr (Figure 7b). The outcome of typical experiments, probing the binding of tRNAPhe to E sites programmed with Met, Phe or Thr codons, is shown in Figure 7b, and all data are collected in Table 1. The results show that tRNAfMet and tRNAPhe bound to postT ribosomes with similar affinities and weak codon specificity. In similar experiments, we also found that the affinity of tRNAfMet for the E site was similar for postT ribosomes, postT ribosomes with RelE-mediated cleavage of the mRNA in the A site, and preT ribosomes, all with the same codon in the E site (Table 1).
Discussion
In this work, we present experimental results that redefine the roles of guanine-nucleotide exchange and GTP hydrolysis on the elongation factor EF-G during translocation of tRNAs on the ribosome. Taking advantage of an in vitro system with pure components and high activity, we characterized the different states of the ribosome during translocation of tRNAs and mRNA as catalyzed by EF-G. On the basis of these experimental results we propose a novel mechanism for translocation, a process that is conserved throughout all kingdoms of life.
The ribosome is a GEF for EF-G
On the basis of measurements of the affinities of GDP and GTP for free EF-G [12], it has been assumed that EF-G in the GTP-bound form binds to the pre-translocation (preT) ribosome, and that exchange of GDP for GTP occurs after release of EF-G•GDP from the ribosome [1,3]. In this work, however, we found the dissociation constant for the EF-G•GDP complex (KGDP = 9 μM) to be more than 60 times smaller than the dissociation constant for the EF-G•GTP complex (KGTP > 0.6 mM) (Figure 2a,b). Given that the GTP:GDP ratio in the living cell is only about 7:1 [13], a major fraction of EF-G is expected to be bound to GDP in vivo (Figure 2a,b). We also demonstrated that EF-G in complex with GDP can rapidly enter the preT ribosome, and that guanine-nucleotide exchange on EF-G occurs on, rather than off, the ribosome (Figure 2c and Materials and methods). This defines the preT ribosome as a previously unknown GEF for EF-G.
The crystal structures of guanine-nucleotide-free and GDP-bound EF-G are virtually identical [16], and it has been difficult to obtain the crystal structure of GTP-bound EF-G. Regarding solution structures, a study with small-angle X-ray scattering (SAXS) was performed by Czworkowski and Moore [4] on EF-G from Thermus thermophilus. They found the scattering data obtained from guanine-nucleotide-free, GDP-bound, GTP-bound and GDPCP-bound EF-G to be virtually identical and very close to the virtual scattering curve calculated from the crystal structure of GDP-bound T. thermophilus EF-G. They estimated the dissociation equilibrium constants for the binding of GDP, GTP and the non-cleavable GTP analog GDPCP as 0.92, 14.1 and 790 μM, respectively. Unfortunately, the GTP-binding experiments were performed without further purification of the GTP solution, meaning that it contained an unknown fraction of GDP. Accordingly, they may have significantly overestimated the affinity of EF-G for GTP, which appears likely in light of our new results (Figure 2a,b). Even though their SAXS experiments were performed in the presence of an energy-regeneration system, pumping GDP back to GTP, the fraction of GDP was significant (2–5%) [4]. It could therefore be that the scattering data reflected EF-G molecules primarily bound to GDP. Our data (Figure 2a,b) do not exclude the possibility that a small fraction of EF-G in the cell is GTP-bound. As shown in Additional data file 1 with the online version of this article, this does not necessarily mean that the conformation of EF-G has switched to its GTP-bound form, in line with the close similarity of a very recent crystal structure of an EF-G mutant in complex with GDPNP [17] or the crystal structure of EF-G bound to GDP [18,19]. Furthermore, the mechanism we propose with ribosome-dependent guanine-nucleotide exchange makes all EF-G molecules active in ribosome binding independent of their guanine-nucleotide content, and effectively prevents idling GTPase activity of EF-G [8].
Guanine-nucleotide exchange requires the 'twisted' ribosome conformation
Cryo-EM studies [5] have shown that the deacylated tRNA in the P site of the post-termination (postTerm) ribosome moves to the hybrid P/E site when the ribosome forms a high-affinity complex with EF-G•GDPNP (Figure 8d). During this movement, the 30S subunit is rotated (twisted) in relation to the 50S subunit. The affinity of EF-G•GDPNP for the post-translocation (postT) ribosome with a peptidyl-tRNA in the P site (Figure 8a,k) is weak [8]. A peptidyl-tRNA, in contrast to a deacylated tRNA, cannot adapt to the P/E hybrid site, and it was therefore concluded that high-affinity binding of EF-G•GDPNP (or EF-G•GTP) to the ribosome requires that the ribosome is in the twisted conformation [5,8]. This suggests that guanine-nucleotide exchange on EF-G and the concomitant conformational switch from its GDP- to its GTP-bound form during translocation must take place when the ribosome is in the twisted conformation with hybrid sites for its tRNAs. If so, this would be analogous to the mechanism by which class-1 release factors are recycled by RF3. In that case, it was found that the GEF for RF3 is the postTerm ribosome (Figure 8b) with a deacylated tRNA in the P site and a class-I peptide-release factor in the A site [9,10]. Before discussing the experimental evidence supporting this hypothesis, we will discuss what appears to be conflicting experimental evidence regarding the conformation of the preT ribosome.
Moazed and Noller [14] found that, after peptide-bond formation, a peptidyl-tRNA originally bound in the A site moves to the hybrid A/P site (Figure 8i). It was further suggested that the 30S ribosome must rotate relative to the 50S subunit to preserve the structure of the tRNA [20]. At the same time, reconstruction of the ribosomal preT complex using cryo-EM shows a relaxed state of the ribosome with three tRNAs in 'classical' A, P and E sites [5]; (Figure 8e). This apparent contradiction between two different sets of experimental results can, we suggest, be explained by different experimental conditions. The twisted preT state [14] was observed in the absence of free deacylated tRNAs (D. Moazed, personal communication), while the relaxed preT state was seen in their presence [5,8]. Because filling of the E site is incompatible with the twisted state (Figure 8i), the presence of deacylated tRNA in excess can drive the ribosome from its twisted [14] to its relaxed [5] preT state (Figure 8e).
In the present work, we demonstrate that deacylated tRNA binds to the E site of preT as well as postT ribosomes with low codon-anticodon specificity, and with affinities favored by low temperature (Figures 5a,b and 7a,b,e). These data can explain the presence of E-site-bound tRNA in the cryo-EM reconstructions of the relaxed preT (Figure 8e; [5]) and postTerm (Figure 8b; [21]) ribosome. Recent results [22] indicate that the tRNAs on the preT ribosome fluctuate between 'classical' A-P (Figure 8e) and hybrid A/P-P/E (Figure 8i) sites in an equilibrium that may depend on the occupancy of the E site, and therefore on the concentration of free deacylated tRNA.
We have now found that addition of EF-G•GDP to preT ribosomes reduces the affinities of deacylated tRNAs to the E site (Table 1). This supports the hypothesis that EF-G•GDP can drive a preT ribosome from its relaxed [5] to its twisted conformation [5,20] (Figure 8e,f), as this would reduce the apparent affinity of deacylated tRNAs for the E site. From this we propose that when EF-G•GDP encounters a relaxed preT ribosome, it induces a twisted ribosome conformation in which the exchange of GDP to GTP on EF-G takes place.
EF-G in the GTP-bound form drives the ribosome into a transition state
When GDP is exchanged for GDPNP on EF-G in the preT ribosome, EF-G changes conformation and the ribosome moves from the preT state to the transition state transT* (Figure 8j). The transT* structure has not been observed directly, but some of its characteristics can be guessed from the cryo-EM reconstruction of the complex between EF-G•GDPNP and the postTerm ribosome (Figure 8d; [5]). Here, domain IV, located on the tip of EF-G and mimicking the anticodon end of tRNA [23], is positioned in the A site of the 30S subunit. The shape of EF-G•GDPNP is reminiscent of previous observations of the EF-Tu•GDP•aminoacyl-tRNA ternary complex, stalled in the A/T site of the postT ribosome with the antibiotic kirromycin [24]. In the A/T site, the anticodon end of the aminoacyl-tRNA is bound to the A site of the small ribosomal subunit, while its CCA end is bound to EF-Tu. We propose that in order to adopt a similar conformation on the preT ribosome, EF-G must displace the peptidyl-tRNA from the A site of the 30S subunit (Figure 8g). The transT* state is formed, we suggest, by translocation of the tRNA2-mRNA complex in relation to the 30S subunit. This would allow ribosomal RNA helix h44 and ribosomal protein S12 to interact with domain IV of EF-G [5], which could prevent slippage of the tRNAs back to their preT positions. The fundamental role that we propose for domain IV of EF-G is supported by previous observations that truncation of, or single point mutations in, domain IV inhibit translocation [25]. Furthermore, the finding that the affinity of EF-G•GDPNP to the transT* ribosome (Figures 5c,8j) is much lower than for the postTerm ribosome [8] can be explained by the proposed competition for binding to the A site between domain IV of EF-G•GDPNP and the anticodon part of peptidyl-tRNA in the former case (Figure 8j) but not in latter case (Figure 8d).
We have shown previously that peptidyl-tRNA, situated in the A site of the preT ribosome, becomes puromycin-reactive in the ribosomal transT* state, but that the rate of this reaction is much slower than for peptidyl-tRNA in the postT state (Figure 6b; [8]). The class-1 release factor RF1 or RF2, which normally binds to the stop codon of the preTerm state, here appearing in an identical conformation to the postT state (Figure 8k), also induces slow peptide release from the peptidyl-tRNA in the transT* ribosome (Figure 6c). This means that the stop codon, which in the preT ribosome is downstream from and adjacent to the A-site codon (Figure 8e), must be at least intermittently present in the A site of the transT* ribosome (Figure 8j). At the same time, dissociation of the peptide from the transT* ribosome in the presence of RF2 is considerably slower than peptidyl transfer to puromycin (Figure 6b,c), and much slower than RF2-dependent peptide release from the postT (Figure 8k) or preTerm (Figure 8a) ribosome [9,10]. These results suggest that EF-G•GDPNP blocks access of the release factor to the A site (Figure 3c). It is only when EF-G•GDPNP dissociates from the ribosome that RF2 can bind to the A site and induce peptide release (Figure 6c; [8]).
Further evidence that the mRNA moves one codon in relation to the 30S subunit when the ribosome switches from preT to transT* state comes from experiments with RelE. These reveal a cut in the stop codon of transT* ribosomes, implying that it must have moved at least intermittently into the A site (Figures 3b,6a). The mRNA cleavage is slow, suggesting that EF-G•GDPNP must dissociate before RelE can enter the ribosome and cut the mRNA (Figure 8j). After RelE cleavage of the mRNA, the ribosome remains in the transT* state (Figure 5d).
We have also demonstrated here that the deacylated tRNA, which was originally in the P site (Figure 8e), remains in contact with the mRNA in the transT* state and thus can be exchanged only with a deacylated tRNA of the same type (Figure 5a,b). This further indicates that the tRNAs in the transT* state have not been completely accommodated in the postT state where a deacylated tRNA bound to the E site would lack codon specificity (Figure 7a,b and Table 1).
Ribosome movement from the transition state to the post-termination state
To understand how the ribosome moves from the transT* to the postT state, one must understand the role of GTP hydrolysis in the translocation process.
A striking finding of the work described here is that removal of EF-G•GDPNP from the transT* ribosome by addition of excess GDP does not lead to complete translocation, as would follow from the 'classical' model [4,6], but instead brings the ribosome back to the preT state (Figure 6). This means that, after dissociation of EF-G•GDPNP, the transT* structure is spontaneously pulled back to the preT, rather than the postT, structure of the ribosome (Figure 8e,i,j). The driving force for this movement is provided by the affinity of the anticodon end of the peptidyl-tRNA for the decoding center of the 16S rRNA, which is missing in the transT* and present in the preT state. Subsequent return to the transT* state by the action of EF-G is prevented by the presence of GDP (Figure 8i,j). When, in contrast, GTP is added to a transT* ribosome that is stabilized by EF-G•GDPNP, translocation is completed (Figure 8g,h,j).
Remarkably, there is another pathway to complete translocation. That is, when RF2 is added to a ribosome in the transT* state with EF-G•GDPNP which has a UAA stop codon in the A site of its postT state, the ribosome eventually ends up in the postTerm state with RF2 in the A site and deacylated tRNA in the P site (Figure 8k). This means that the driving force, generated by the strong affinity of RF2 for the relaxed postTerm ribosome [10], is sufficient to overcome the counteracting force provided by the affinity of the anticodon end of the peptidyl-tRNA for the decoding center in the preT ribosome. When GTP is hydrolyzed on EF-G in the transT* ribosome (Figure 8g,h,k), the elongation factor must adopt a conformation that can do the same trick as RF2.
This conformation of EF-G in Figure 8h cannot be the structure of EF-G•GDPNP [5] because the latter has very low affinity for the postT ribosome [8]. Instead, we suggest that GTP hydrolysis on EF-G brings EF-G•GDPNP from its GTP-bound form, with high affinity for the twisted transT* ribosome and low affinity for the relaxed postT ribosome [8], to an EF-G•GDP•Pi form that has high affinity for the relaxed postT ribosome. During this conformational change, the 30S subunit is pulled by domain IV of EF-G into the relaxed conformation with docking of the tRNA2-mRNA complex in the postT state (Figure 8k). In this step, the deacylated tRNA in the hybrid P/E site of the transT* ribosome loses contact with the mRNA and moves into the E/E site (Figures 5a,b and 7; [15]). It is possible that the GDP•Pi form of EF-G is similar to the EF-G•GDP form found in postT ribosomes in the presence of fusidic acid [5]. When inorganic phosphate is released from EF-G in the absence of fusidic acid, EF-G adopts the free GDP-bound conformation with low affinity to the postT state, and rapidly dissociates from the ribosome. When, in contrast, fusidic acid is present, EF-G remains in the EF-G•GDP•Pi form, stably bound to the postT ribosome.
Comparison with previous models
Our proposal that EF-G in the GTP form drives the ribosome into a transition state (transT*), with properties distinct from those of the preT and postT ribosome, contrasts with the 'motor' mechanism proposed by Rodnina et al. [7], in which the action of EF-G comes after GTP hydrolysis. We have also identified the GTP-bound conformation of EF-G in the transition state with the cryo-EM reconstruction of EF-G•GDPNP on the postTerm ribosome [5].
Our proposal differs further from the classical model [4,6] as well as from the model proposed by Rodnina et al. [7] in that movement of the ribosome to the postT state requires both GTP and GTP hydrolysis, and cannot occur with EF-G in the presence of a GTP analog or GDP. The present finding that complete translocation to the postT ribosome requires GTP hydrolysis is compatible with the observation that GTP hydrolysis precedes translocation [7]. We suggest further that the GDP-dependent translocation observed by Rodnina et al. [7] originates in a GTP-contaminated GDP solution, which would result in exactly the slow, monophasic translocation they observe (see Materials and methods for further details).
Our model proposes roles for the EF-G•GDP•Pi structure and release of inorganic phosphate that are distinct from previous suggestions [7]. That is, we suggest that GTP hydrolysis on EF-G in the ribosomal transT* state results in a conformation of EF-G•GDP•Pi that catalyzes ribosome movement from the transition state to the postT state. Then, dissociation of EF-G from the ribosome requires release of inorganic phosphate (Pi), which favors formation of the GDP-bound structure of EF-G [16] with low affinity for the ribosome [8]. If it is assumed that the conformation of EF-G in the EF-G•GDP•Pi complex is the same as the conformation of EF-G in the EF-G•GDP•fusidic acid complex [5], then the mechanistic action of fusidic acid could be rationalized as a freezing of EF-G in the conformation that catalyzes the second major step of translocation, bringing the ribosome from the transT* state to the postT state. In contrast, Rodnina and collaborators [26] associate release of inorganic phosphate with a 'relocking' conformational change of the ribosome, which leads to the postT state.
Conclusion
The mechanism of translocation in eubacteria that we have suggested differs radically from all previous models [4,6,7,26,27] in that we propose the following: first, that EF-G enters the preT ribosome in the GDP-favoring form [16]; second, that EF-G•GDP drives the preT ribosome from its relaxed state with full binding sites for three tRNAs [5] to a twisted conformation with hybrid sites for two tRNAs [5]; and third, that exchange of GDP for GTP and an accompanying switch of the EF-G conformation from its GDP-bound to its GTP-bound structure occur on, rather than off, the ribosome.
Our results suggest that the ribosome plays a previously unidentified dual role of both guanine-nucleotide exchange factor and GTPase-activating protein for at least two translation factors in eubacteria, EF-G and RF3. This may be rationalized by the requirement that ribosomal GTPase activity be strictly controlled [8] and that each of the translation factors must selectively target a particular state of the ribosome.
Materials and methods
E. coli components for protein synthesis in vitro
Ribosomes of high activity, polymix buffer, initiation factors IF1, IF2 and IF3, translation factors EF-Tu, EF-Ts, EF-G and RF2, tRNA bulk, tRNAPhe, tRNAIle and overexpressed fMet-tRNAfMet, PheRS, ThrRS and IleRS were prepared as described [9,28]. The tRNAs were further purified by HPLC according to Cayama et al. [29] to aminoacylation activities greater than 80%. [3H]-GDPNP (Amersham Bioscience, Uppsala, Sweden) and other nucleotides (Sigma, Stockholm, Sweden) were further purified on a MonoQ column (Amersham Bioscience) as described ([9]; Figure 4a). They were bound to the MonoQ in Buffer A (20 mM Tris-HCl) and eluted with NaCl in buffer B (20 mM Tris-HCl and 1 M NaCl). The Met-Ile-encoding mRNA used to make translocation complexes had the sequence gggcccuuguuaacaauuaaggagguauxxx AUG AUU UAA uugcag(a)21. It contained a strong ribosome-binding site, an open reading frame encoding a Met-Ile dipeptide (capital letters), a stop codon UAA and a poly(A) tail for purification on a poly(dT) column; xxx was either a Thr (acu) or a Phe (uuu) codon. The mRNA encoding the Met-Phe-Thr-Ile tetrapeptide had the open reading frame AUG UUU ACG AUU. The mRNAs were synthesized by T7 RNA polymerase transcription of DNA oligonucleotides. [33P]-labeling of tRNAfMet and mRNAs followed standard protocols (Amersham Bioscience).
Ribosomal complexes
Init and preT complexes were prepared according to Zavialov and Ehrenberg [8]. PostT and preTerm complexes were obtained according to Zavialov et al. [9]. In some experiments either [33P]-labeled mRNA [11] or [33P]-tRNAfMet was used for complex preparation. The fraction of peptidyl-tRNA in the complexes was more than 80% (for Init, postT and preTerm complexes) or 70% (for preT).
Binding of GDP, GTP and GDPNP to EF-G
For the experiment shown in Figure 2a, 5 μM EF-G was incubated for 2 min at 37°C with [3H]-GDP (3–18 μM) in 50 μl. The tubes with EF-G and GDP were then put on ice. After 15 min, 45 μl of each sample was nitrocellulose-filtered and washed with 500 μl ice-cold polymix buffer. All operations were done in a cold room (4°C). The amount of EF-G• [3H]-GDP retained on the filter was quantified by scintillation counting [9]. For the experiment shown in Figure 2b, 5 μM EF-G was incubated with 45 μM [3H]-GDP and unlabeled GTP (0–2.5 mM) or GDP (0–1.2 mM). Other conditions were the same as in Figure 2a. For the experiment shown in Figure 2d, 2 μM EF-G, 92 nM 70S ribosomes with 1 μM Met-Phe-Thr-Ile mRNA or 92 nM preTerm complexes with 0.4 mM puromycin were incubated with 1 μM [3H]-GDPNP in the presence of cold GDP (0–2 mM) for 7 min at 37°C in 50 μl. Then, 45 μl of each sample was nitrocellulose-filtered. For the experiment shown in Figure 2e, 2 μM EF-G, 92 nM 70S ribosomes with 1 μM Met-Phe-Thr-Ile mRNA or preTerm complexes with 0.4 mM puromycin were pre-incubated with 1 μM [3H]-GDPNP for 4 min at 37°C and then 2 mM GDP (final concentration) was added. After GDP addition 45 μl aliquots were nitrocellulose-filtered. For Figure 2f, EF-G was either incubated for 4 min at 37°C with preTerm complexes, puromycin and [3H]-GDPNP and then buffer or RF2 was added, or alternatively, [3H]-GDPNP was added to EF-G pre-incubated with preTerm complexes, puromycin and RF2. Then, 50 μl aliquots were nitrocellulose-filtered to determine the amount of EF-G• [3H]-GDPNP bound to the postTerm complex. The final concentrations of components in the mixtures were: 92 nM preTerm, 0.4 mM puromycin, 2 μM EF-G, 1 μM RF2 and 1 μM [3H]-GDPNP.
Toxin inducible cleavage of mRNA (TICOM)
For the experiment shown in Figure 3a, 0.12 μM RelE was incubated for 1 min at 37°C with 0.2 μM initiation complex and 0.12 μM preT complex or with 0.24 μM preT complex, 2 μM EF-G and GTP in 50 μl. Then, 10 μl of 'kill' mix (1.25% SDS and 0.25 M EDTA pH 8.1) was added to stop the reaction. The samples were phenol-extracted, precipitated with ethanol and the RNA fragments were run in a sequencing gel (10% polyacrylamide, 8 M urea; [11]). The amount of mRNA cleaved by RelE, normalized to the total intensity of RNA bands in the lane, was determined using the "ImageQuant5.0" program (Amersham Bioscience). The background hydrolysis of mRNA was used as a nucleotide ladder. For Figure 3b, either RelE was added to preT ribosomes incubated with EF-G and one of the nucleotides at different times and then the resulting mix was incubated 5 min more before quenching, or RelE was always present in the mix. The final concentrations of components in the mix (50 μl) were: 0.1 μM RelE, 2 μM EF-G, 0.15 μM preT and 1 mM of each nucleotide. In all experiments cleavage in the UAA codon in the presence of EF-G and GTP was set at 100%. For Figure 3c, 120 nM RelE was incubated with 0.3 μM postT or preT complex, 2 μM EF-G and 0.6 mM GDPNP at 37°C, and 50 μl aliquots were removed and analyzed at different time points (0–4 min). At 4 min, 100 mM GTP was added to obtain 1 mM final concentration and the incubation continued for 1 min before quenching.
For Figure 4c, 80 nM RelE was incubated for 3 min at 37°C in 50 μl with 2 μM EF-G, 0.15 μM preT and nucleotides as indicated in the figure.
For Figure 6a, EF-G, GDPNP and preT were incubated at 37°C for 3 min. Then, the mix was divided into two tubes with GDP in one and buffer in the other. The mixes were then incubated with RelE at 37°C and 50 μl aliquots were analyzed at 15 and 28 min of incubation. After 29 min GTP was added and the incubation continued for 1 min before quenching. The final concentrations of the components were 2 μM EF-G, 40 μM GDPNP, 0.4 μM preT, 166 nM RelE, 2 mM GDP and 1 mM GTP.
Release of [33P]-tRNAfMet from the pre-translocation complex
For the experiment shown in Figure 5a, EF-G and a nucleotide were pre-incubated with preT complex containing [33P]-tRNAfMet in the P site for 3 min at 37°C. Then, cold tRNAfMet or tRNAPhe was added and 45 μl aliquots were nitrocellulose-filtered at different times to determine the fraction of [33P]-tRNAfMet on the ribosome. The concentrations of components in the mix were: 70 nM preT, 2 μM EF-G, 1 μM cold tRNAfMet or tRNAPhe and 1 mM nucleotide. For Figure 5b, EF-G, a nucleotide, preT and 0–2 μM tRNAfMet or tRNAPhe were pre-incubated for 9 min at 37°C. Then, 45 μl aliquots were nitrocellulose-filtered and washed with 1 ml polymix buffer. For Figure 5c, 2 μM EF-G was incubated for 7 min at 37°C with 88 nM preT, 2 μM tRNAfMet and GDPNP (0–240 μM). Then, 45 μl aliquots were nitrocellulose-filtered. For Figure 5d, preT complexes with [33P]-tRNAfMet in the P site were pre-incubated with EF-G and GDPNP, with or without RelE, for 3 min at 37°C. Then, cold tRNAfMet or tRNAPhe was added and 45 μl aliquots were nitrocellulose-filtered. After 27.5 min incubation GTP was added to the mix, which then contained 78 nM preT, 2 μM EF-G, 0.4 mM GDPNP, 2 μM tRNAfMet or tRNAPhe, 1 mM GTP and 80 nM RelE.
For Figure 6d, EF-G was pre-incubated for 3 min at 37°C in polymix buffer with or without preT and GDPNP. Then, GDP or polymix buffer was added. After 1 min incubation, tRNAfMet was added and then 45 μl aliquots were nitrocellulose-filtered after different times. The final mix contained 88 nM preT,2 μM EF-G, 100 μM GDPNP, 1 μM tRNAfMet and 2 mM GDP.
Release of [3H]-fMet-[14C]-Ile peptide by RF2 or puromycin
For the experiment shown in Figure 4b, mix A, containing 1 μM EF-G, 23 nM preT complex and 0.4 μM RF2, was pre-incubated for 3 min at 37°C. Then mix B, containing nucleotides, was added to mix A, and 45 μl aliquots were removed at different time points for determination of the amount of released peptide according to Zavialov and Ehrenberg [8]. The resulting mix contained 1 μM EF-G, 23 nM preT, 0.4 μM RF2 and nucleotides as indicated in the figure.
For Figure 2c, mix A containing EF-G, GTP, RF2, and varying GDP concentrations was pre-incubated for 3 min at 37°C. Then, mix B containing preT complex was added and the amount of released peptide in 45 μl aliquots was determined at different time points. The resulting mix contained 10 nM EF-G, 23 nM preT, 0.4 μM RF2, 0.5 mM GTP and 0–0.8 mM GDP.
For Figure 6b and c, EF-G was pre-incubated for 3 min at 37°C with preT complex and GDPNP or polymix buffer. Then, GDP or buffer was added, the incubation was continued for 1 min, and then puromycin (Figure 6b) or RF2 (Figure 6c) was added and the amount of released peptide in 45 μl aliquots was determined at different time points. The final mix contained 2 μM EF-G, 46 nM preT, 0.5 μM RF2 or 0.4 mM puromycin and nucleotides as indicated on the figure.
Binding of tRNAs to the E site
Typically, a ribosomal complex (10–30 nM) was incubated with [33P]-tRNAfMet dilutions for 5 min at 37°C and then 45 μl of the mix was nitrocellulose-filtered in order to estimate the amount of ribosome-bound [33P]-tRNAfMet. To observe the effect of A-site cleavage of mRNA, the incubation was carried out with 100 nM RelE. To measure the binding of [33P]-tRNAfMet to the E site of postT ribosomes at 0°C, tubes with the mix were incubated on ice for 20 min before the filtration. To measure the binding (exchange) of [33P]-tRNAfMet to preT ribosomes in the presence of EF-G, 3 μM EF-G and 1 mM GDPNP, GTP or GDP were added to the reaction mix. To determine the dissociation constant (KI) for the binding of tRNAPhe to the E site from the inhibition curves in Figure 7b, a ribosomal complex was incubated with 0.8 μM [33P]-tRNAfMet and increasing concentrations of unlabeled tRNAPhe KI values were obtained from KI = I50/(1 + [tRNA fMet]/Kd), where I50 is [tRNAPhe] at 50% inhibition of tRNAfMet binding (dissociation constant Kd).
Translocation kinetics with mixtures of GDP and GTP
Experimental analysis of EF-G•GDP binding to the preT ribosome
In the experiment shown in Figure 2c, EF-G is recycling from ribosome to ribosome, catalyzing translocation, and the ratio between GDP and GTP is varied. Under this condition of ribosomes in excess, one can formally describe EF-G as an enzyme and the ribosome as its substrate and apply text-book enzyme kinetics [30].
If only EF-G in the GTP-bound form can bind the ribosome, and if we only consider EF-G bound to either GTP or GDP, neglecting nucleotide-free EF-G, then the steady state flow of translocation [30] can be written as
Here, [EF-G0] and [70S0] are total concentrations of EF-G and 70S ribosomes, respectively; kcat is the maximal rate of the EF-G cycle at saturating ribosome concentration; Km is the Km value of this cycle in the absence of GDP; KGTP and KGDP are dissociation constants for the binding of GTP or GDP to free EF-G, respectively. As written in the main text, this experiment was performed under conditions where the flow j is a linear function of the total ribosome concentration, meaning that [70S0] <<Km, so that j can be written
We know that the ratio between KGTP and KGDP is larger than 60 (Figure 2b), from which it follows that inhibition of the recycling rate should be larger than 30 at a [GDP]: [GTP] ratio of 0.5. But given that the observed inhibition at this nucleotide ratio is only a factor of two, it follows that EF-G in the GDP-bound form must be able to enter the ribosome. There is, in fact, no other way to explain the combination of results in Figure 2a–c.
If, in contrast, EF-G can enter the ribosome in the GDP-bound form, and guanine-nucleotide exchange takes place on rather than off the ribosome, the experimental results in Figure 2a–c are readily explained. One such scheme is presented in the next subsection (equation 3).
Ribosome kinetics with guanine-nucleotide exchange on the ribosome
A simple kinetic scheme for translocation with guanine-nucleotide exchange on the ribosome is
EF-G enters the preT ribosome in complex with GDP with an association rate constant ka. In the next step, EF-G•GDP may dissociate from the ribosome (rate constant kd) or, alternatively, GDP dissociates from EF-G (rate constant ). Ribosome-bound guanine-nucleotide-free EF-G can either bind GDP (compounded rate constant [GDP]) or GTP (compounded rate constant [GTP]). GTP-bound EF-G then promotes translocation (rate constant kT). To simplify, we have neglected dissociation of guanine-nucleotide free EF-G from the ribosome. The Michaelis-Menten parameter kcat/Km for this scheme is given by
Interpreted in terms of this scheme, the two-fold reduction in the rate of cycling of EF-G at 0.5 mM GTP and 0.25 mM GDP shown in Figure 2c, simply means that
The kcat value for the scheme, corresponding to the rate of translocation at saturating concentration of EF-G (single-round kinetics) is given by
The first term on the right side of this equation is the average time for GDP to dissociate from EF-G (1/kdGDP ) multiplied by the average number of dissociations of GDP per successful translocation.
Rodnina et al. [7] found monophasic kinetics when they measured translocation in the presence of an unpurified solution of GDP that is likely to contain a trace amount of GTP. They monitored translocation with EF-G in excess over the ribosome, in which case our model predicts monophasic kinetics with a much slower rate than when GTP is in large excess over GDP. To see this, we first note that when the [GDP]: [GTP] ratio is large, the kcat value for translocation simplifies to
To show that in this limiting case translocation occurs with a single first-order rate constant equal to kcat, we denote the concentration of the preT ribosome in complex with EF-G•GDP by c1, in complex with guanine-nucleotide-free EF-G by c2, and set c = c1 + c2. At saturating free concentration of EF-G•GDP we have c = c1 + c2 ≈ c1 and
Furthermore,
This shows that single-round translocation in a system with a large concentration of GDP and with a small contamination of GTP is expected to be approximated by a single exponential
as observed in the experiment carried out by Rodnina et al. [7].
Additional data files
The following is provided as an additional data file with the online version of this article. Additional data file 1, showing that EF-G in solution can change from a GDP to a GTP conformation.
Supplementary Material
Additional File 1
An additional data file showing that EF-G in solution can change from a GDP to a GTP conformation
Click here for file
Acknowledgements
We thank J. Frank and M. Valle for valuable suggestions and F. Darfeuille for technical assistance. This work was supported by grants from the Swedish Foundation for Strategic Research and the Swedish Research Council.
Figures and Tables
Figure 1 Schematic representation of (a) initiation, (b) pre-translocation, (c) post-translocation, and (d) post-termination complexes, referred to as Init, preT and postT, and postTerm, respectively. A, amino-acyl tRNA site on the ribosome; P, peptidyl-tRNA site; E, exit site; L1, ribosomal protein. The large subunit of the ribosome is shown in yellow and the small subunit in blue. The colored ribbons represent tRNAs and the colored balls represent amino acids in aminoacyl- or peptidyl-tRNA. The purple arrow represents RelE, which cleaves the codon shown at the *. The mauve padlock in (d) illustrates a state of the ribosome in which the mRNA is locked, and cannot move in relation to the small subunit. The figure represents a special case in which the postT ribosome has a stop codon (UAA) in the A site, and is therefore also a pre-termination (preTerm) ribosome. For further details see text and Figure 8.
Figure 2 Ribosome-dependent exchange of GDP to GTP on EF-G. (a) Scatchard plot from a nitrocellulose-filtration experiment to obtain the dissociation constant for the binding of [3H]-GDP to free EF-G. (b) Chase of [3H]-GDP from free EF-G by unlabeled GTP or, as a control, GDP. The dissociation constant for GTP binding to free EF-G was obtained from the corresponding constant for GDP binding in (a) and from the inhibition of [3H]-GDP binding to EF-G by GTP addition. The figure shows the results of two independent experiments (1 and 2). (c) Time-dependent release of fMet-Ile by 0.5 μM RF2 after translocation of fMet-Ile-tRNAIle from the A to the P site by a catalytic amount of EF-G (10 nM) added to 23 nM preT ribosomes together with 0.5 mM GTP and 0–0.8 mM GDP. CM(GTP) is the GTP concentration and I50 is the GDP concentration at which the rate of translocation is reduced to half-maximal value. (d) Inhibition of EF-G•GDPNP binding to post-termination (PostTerm) complexes or naked 70S ribosomes (Nakedribo) in the presence of 1 μM [3H]-GDPNP and 0–2 mM unlabeled GDP. (e) Fraction of [3H]-GDPNP (total concentration 1 μM) bound to EF-G• [3H]-GDPNP in postTerm complexes or in naked ribosomes as a function of time after addition of unlabeled GDP to a concentration of 2 mM. (f) Time-dependence of EF-G• [3H]-GDPNP binding to postTerm ribosomes in the presence of 1 μM [3H]-GDPNP: in the absence of RF2 (control), after addition of [3H]-GDPNP to EF-G pre-incubated with RF2 and postTerm ribosomes, or after addition of RF2 to EF-G pre-incubated with [3H]-GDPNP and postTerm ribosomes.
Figure 3 RelE cleavage of mRNA in the A site of ribosomal complexes. (a) The mRNA fragments resulting from RelE cleavage in the A site of the three ribosomal complexes Init (see Figure 1a), preT (see Figure 1b) and postT (see Figure 1c), separated on a 10% sequencing gel. The amount of radioactivity in the postT lane was doubled to make the AUU cleavage visible. (b) Time-dependent cleavage of mRNA by RelE; preT ribosomes were incubated with EF-G together with GDPNP (+ GDPNP 2) or GTP (+ GTP) or GDP (+ GDP). RelE was added after 10, 25 or 40 min, and the reaction was in each case quenched 5 min after RelE addition. Alternatively, preT ribosomes were incubated together with EF-G, RelE and GDPNP and the reaction quenched after 15, 30 or 45 min (+ GDPNP 1). (c) Time-dependent cleavage of mRNA by 120 nM RelE in the A site of 0.3 μM postT or preT ribosome complexes incubated with 2 μM EF-G and 0.6 mM GDPNP. As a control, in the last two lanes 1 mM GTP was added to postT or preT ribosomes at the end of the incubation.
Figure 4 Contamination of GDP preparations with GTP strongly stimulates translocation by EF-G. (a) Elution profile of commercially available GDP from a MonoQ column showing the GTP and GMP contaminations. %B is the percentage of buffer B (20 mM Tris-HCl, 1 M NaCl) in the buffer A (20 mM Tris-HCl) + B mixture. (b) Time-dependent release of peptide by 0.4 μM RF2 after translocation of fMet-Ile-tRNA (23 nM total) from the A site to the P site by 1 μM EF-G in the presence of 1 mM purified GDP, unpurified GDP, purified GDP containing 20 μM GTP (2%), or 20 μM GTP. (c) Cleavage of mRNA by RelE incubated with 0.15 μM preT, 2 μM EF-G and nucleotides. Lanes: (1) no GDP; (2) 1 mM purified GDP; (3) 1 mM unpurified GDP; (4) 1 mM purified GDP containing 2% GTP; (5) 20 μM GTP.
Figure 5 Properties of the transition state. (a) Time-dependent exchange of [33P]-tRNAfMet bound to the P site of 70 nM preT complex with 1 μM unlabeled tRNAfMet or tRNAPhe after the addition of 2 μM EF-G and 1 mM nucleotide. (b) The fraction of [33P]-tRNAfMet exchanged with tRNAfMet or tRNAPhe after 9 min incubation of 70 nM preT with 2 μM EF-G, 1 mM nucleotide and 0–2 μM tRNAfMet or tRNAPhe. (c) Fraction of [33P]-tRNAfMet on 88 nM preT ribosomes exchanged after 7 min incubation with 2 μM unlabeled tRNAfMet, 2 μM EF-G and 0–240 μM GDPNP to estimate the fraction of ribosomes containing EF-G•GDPNP. (d) Exchange of [33P]-tRNAfMet with 2 μM tRNAfMet or tRNAPhe added to 78 nM preT incubated with 2 μM EF-G, 0.4 nM GDPNP with or without 80 nM RelE. At 27.5 min, 1 mM GTP was added to translocate [33P]-tRNA fMet to the E site.
Figure 6 Removal of EF-G•GDPNP from the transition state with GDP. (a) Time-dependent cleavage of mRNA by 166 nM RelE in transT* complex in the presence of 2 μM EF-G and 0.32 mM GDPNP (GDPNP case) or after further addition of GDP to a concentration of 1 mM to remove EF-G from the ribosome (GDPNP + GDP case). In each case, GTP was added to a final concentration of 1 mM at 29 min to show the fraction of ribosomes that was active in translocation (lanes 3 and 6). (b,c) Time-dependent release of fMet-Ile by (b) 0.4 mM puromycin or (c) 0.5 μM RF2; 2 μM EF-G was pre-incubated with 46 nM preT complex and 40 μM GDPNP or polymix buffer for 3 min at 37°C. Then, buffer or 2 mM GDP was added and the incubation was continued for 1 min. Finally, (b) 0.4 mM puromycin or (c) 0.5 μM RF2 was added and the extent of peptide release was observed over time. (d) Exchange of [33P]-tRNAfMet on 88 nM preT complex, pre-incubated with 2 μM EF-G and 100 μM GDPNP or with buffer, with 2 μM tRNAfMet in the presence or absence of 2 mM GDP.
Figure 7 Binding of deacylated tRNA to the E site. (a) Binding of [33P]-tRNAfMet to the postT complex with fMet-Phe-Ile-tRNAIlein the P site and a Thr codon (ACG) in the E site. Insert: Scatchard plot to obtain the dissociation constant. (b) Chase of [33P]-tRNAfMet from the E site of postT complexes containing Met (AUG), Phe (UUU) or Thr (ACG) codons with unlabeled tRNAPhe. Dissociation constants for [33P] tRNAfMet were obtained as in (a) and the dissociation constants for tRNAPhe were calculated from the 50% chase (I50) concentrations of tRNAPhe.
Figure 8 Interaction between the ribosome and EF-G; a proposal for the whole mechanism of translocation. (a-c) Release of peptide from the postTerm ribosome allows the intersubunit rotation. (d) EF-G•GDPNP stabilizes the twisted form of the ribosome as observed by cryo-EM [5]. (e-h,k) A model explaining the mechanism of translocation. (i,j) A GTP-analog pathway. The model is explained in detail in the text and symbols are as in Figure 1. L7/12 is a complex of ribosomal proteins thought to activate GTP hydrolysis on ribosomal GTPases. The mauve padlock illustrates states of the ribosome in which the mRNA is locked, and cannot move in relation to the small subunit. Domain IV of EF-G is suggested in the main text to play an important role in translocation.
Table 1 Dissociation constants for the binding of tRNAfMet and tRNAPhe to different ribosomal complexes
Set State Peptide Codon in the E site tRNA Dissociation constant KD (nM) Additional factors; temperature
1 PreT fMI Thr (ACG) fMet 144 ± 7 37°C
179 ± 14 EF-G + GTP; 37°C
770 ± 80 EF-G + GDP; 37°C
40 ± 2 EF-G + GDPNP; 37°C
2 PostT fMI Met (AUG) fMet 153 ± 9 37°C
Phe 250 ± 30
fM Phe (UUU) fMet 295 ± 26
Phe 84 ± 11
fMFTI Thr (ACG) fMet 162 ± 5
Phe 134 ± 20
3 PostT fMI Met (AUG) fMet 155 ± 10 + RelE (A-site cut); 37°C
4 PostT fMI Met (AUG) fMet 24.8 ± 1.2 0°C
Different conditions were used for measuring the dissociation constants for the different combinations of tRNA and ribosomal complexes as in
Figure 7, and are shown in the last column.
==== Refs
Bourne HR Sanders DA McCormick F The GTPase superfamily: conserved structure and molecular mechanism Nature 1991 349 117 127 1898771 10.1038/349117a0
Sprang SR G protein mechanisms: insights from structural analysis Annu Rev Biochem 1997 66 639 678 9242920 10.1146/annurev.biochem.66.1.639
Savelsbergh A Mohr D Wilden B Wintermeyer W Rodnina MV Stimulation of the GTPase activity of translation elongation factor G by ribosomal protein L7/12 J Biol Chem 2000 275 890 894 10625623 10.1074/jbc.275.2.890
Czworkowski J Moore PB The conformational properties of elongation factor G and the mechanism of translocation Biochemistry 1997 36 10327 10334 9254632 10.1021/bi970610k
Valle M Zavialov A Sengupta J Rawat U Ehrenberg M Frank J Locking and unlocking of ribosomal motions Cell 2003 114 123 134 12859903 10.1016/S0092-8674(03)00476-8
Inoue-Yokosawa N Ishikawa C Kaziro Y The role of guanosine triphosphate in translocation reaction catalyzed by elongation factor G J Biol Chem 1974 249 4321 4323 4605331
Rodnina MV Savelsbergh A Katunin VI Wintermeyer W Hydrolysis of GTP by elongation factor G drives tRNA movement on the ribosome Nature 1997 385 37 41 8985244 10.1038/385037a0
Zavialov AV Ehrenberg M Peptidyl-tRNA regulates the GTPase activity of translation factors Cell 2003 114 113 122 12859902 10.1016/S0092-8674(03)00478-1
Zavialov AV Buckingham RH Ehrenberg M A posttermination ribosomal complex is the guanine nucleotide exchange factor for peptide release factor RF3 Cell 2001 107 115 124 11595190 10.1016/S0092-8674(01)00508-6
Zavialov AV Mora L Buckingham RH Ehrenberg M Release of peptide promoted by the GGQ motif of class 1 release factors regulates the GTPase activity of RF3 Mol Cell 2002 10 789 798 12419223 10.1016/S1097-2765(02)00691-3
Pedersen K Zavialov AV Pavlov MY Elf J Gerdes K Ehrenberg M The bacterial toxin RelE displays codon-specific cleavage of mRNAs in the ribosomal A site Cell 2003 112 131 140 12526800 10.1016/S0092-8674(02)01248-5
Baca OG Rohrbach MS Bodley JW Equilibrium measurements of the interactions of guanine nucleotides with Escherichia coli elongation factor G and the ribosome Biochemistry 1976 15 4570 4574 788779 10.1021/bi00666a004
Neuhard J Nygaard P Neidhardt FC Purines and pyrimidines Escherichia coli and Salmonella typhimurium: Cellular and Molecular Biology 1987 Washington, DC: American Society for Microbiology 445 473
Moazed D Noller HF Intermediate states in the movement of transfer RNA in the ribosome Nature 1989 342 142 148 2682263 10.1038/342142a0
Lill R Robertson JM Wintermeyer W Binding of the 3' terminus of tRNA to 23S rRNA in the ribosomal exit site actively promotes translocation EMBO J 1989 8 3933 3938 2583120
al-Karadaghi S Æevarsson A Garber M Zheltonosova J Liljas A The structure of elongation factor G in complex with GDP: conformational flexibility and nucleotide exchange Structure 1996 4 555 565 8736554 10.1016/S0969-2126(96)00061-5
Hansson S Singh R Gudkov AT Liljas A Logan DT Crystal structure of a mutant elongation factor G trapped with a GTP analog FEBS Lett
Ævarsson A Brazhnikov E Garber M Zheltonosova J Chirgadze Y al-Karadaghi S Svensson LA Liljas A Three-dimensional structure of the ribosomal translocase: elongation factor G from Thermus thermophilus EMBO J 1994 13 3669 3677 8070397
Czworkowski J Wang J Steitz TA Moore PB The crystal structure of elongation factor G complexed with GDP, at 2.7 Å resolution EMBO J 1994 13 3661 3668 8070396
Noller HF Yusupov MM Yusupova GZ Baucom A Cate JH Translocation of tRNA during protein synthesis FEBS Lett 2002 514 11 16 11904173 10.1016/S0014-5793(02)02327-X
Rawat UB Zavialov AV Sengupta J Valle M Grassucci RA Linde J Vestergaard B Ehrenberg M Frank J A cryo-electron microscopic study of ribosome-bound termination factor RF2 Nature 2003 421 87 90 12511960 10.1038/nature01224
Blanchard SC Kim HD Gonzalez RL Jr Puglisi JD Chu S tRNA dynamics on the ribosome during translation Proc Natl Acad Sci USA 2004 101 12893 12898 15317937 10.1073/pnas.0403884101
Nissen P Kjeldgaard M Thirup S Polekhina G Reshetnikova L Clark BF Nyborg J Crystal structure of the ternary complex of Phe-tRNAPhe, EF-Tu, and a GTP analog Science 1995 270 1464 1472 7491491
Valle M Zavialov A Li W Stagg SM Sengupta J Nielsen RC Nissen P Harvey SC Ehrenberg M Frank J Incorporation of aminoacyl-tRNA into the ribosome as seen by cryo-electron microscopy Nat Struct Biol 2003 10 899 906 14566331 10.1038/nsb1003
Savelsbergh A Matassova NB Rodnina MV Wintermeyer W Role of domains 4 and 5 in elongation factor G functions on the ribosome J Mol Biol 2000 300 951 961 10891280 10.1006/jmbi.2000.3886
Savelsbergh A Katunin VI Mohr D Peske F Rodnina MV Wintermeyer W An elongation factor G-induced ribosome rearrangement precedes tRNA-mRNA translocation Mol Cell 2003 11 1517 1523 12820965 10.1016/S1097-2765(03)00230-2
Wilson DN Nierhaus KH The ribosome through the looking glass Angew Chem Int Ed Engl 2003 42 3464 3486 12900959 10.1002/anie.200200544
Antoun A Pavlov MY Tenson T Ehrenberg M Ribosome formation from subunits studied by stopped-flow and Rayleigh light scattering Biol Proced Online 2004 6 35 54 15103398 10.1251/bpo71
Cayama E Yepez A Rotondo F Bandeira E Ferreras AC Triana-Alonso FJ New chromatographic and biochemical strategies for quick preparative isolation of tRNA Nucleic Acids Res 2000 28 E64 10871387 10.1093/nar/28.12.e64
Fersht A Structure and Mechanism in Protein Science: A Guide to Enzyme Catalysis and Protein Folding 1998 New York: WH Freeman
| 15985150 | PMC1175996 | CC BY | 2021-01-04 16:05:45 | no | J Biol. 2005 Jun 27; 4(2):9 | utf-8 | J Biol | 2,005 | 10.1186/jbiol24 | oa_comm |
==== Front
PLoS MedPLoS MedpbioplosbiolPLoS Medicine1549-12771549-1676Public Library of Science San Francisco, USA 1601392110.1371/journal.pmed.0020182Research ArticleBioengineeringBiotechnologyInfectious DiseasesHIV/AIDSInfectious DiseasesHIV Infection/AIDSMedicine in Developing CountriesA Microchip CD4 Counting Method for HIV Monitoring in Resource-Poor Settings Microchip CD4 CountingRodriguez William R [email protected]
1
2
3
Christodoulides Nicolaos
4
Floriano Pierre N
4
Graham Susan
3
Mohanty Sanghamitra
4
Dixon Meredith
1
Hsiang Mina
1
Peter Trevor
5
Zavahir Shabnam
5
Thior Ibou
5
Romanovicz Dwight
4
Bernard Bruce
4
Goodey Adrian P
4
Walker Bruce D
1
2
McDevitt John T [email protected]
4
1 Partners AIDS Research Center, Massachusetts General Hospital, Charlestown, Massachusetts, United States of America,2 Division of AIDS, Harvard Medical School, Boston, Massachusetts, United States of America,3 Brigham and Women's Hospital, Boston, Massachusetts, United States of America,4 Department of Chemistry and Biochemistry, University of Texas, Austin, Texas, United States of America,5 Botswana–Harvard AIDS Institute Partnership, Princess Marina Hospital, Gaborone, BotswanaBentwich Zvi Academic EditorRosetta GenomicsIsrael
Competing Interests: WRR, NC, PNF, BDW, and JTM have applied for a patent for the application described here.
Author Contributions: WRR, NC, PF, SG, BDW, and JTM designed the study. WRR, NC, PNF, SG, MD, SM, ST, IB, TP, MH, DR, BB, APG, BDW, and JTM collected and analyzed the data. WRR, BDW, NC, PNF, and JTM prepared the manuscript.
7 2005 19 7 2005 2 7 e18231 1 2005 26 4 2005 Copyright: © 2005 Rodriguez 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.
CD4 Measurements in Patients with HIV: Are They Feasible for Poor Settings?
Test Your Knowledge: HIV Infection Epidemiology and Diagnosis
Towards a Cheap and Easy Way to Monitor HIV/AIDS
Background
More than 35 million people in developing countries are living with HIV infection. An enormous global effort is now underway to bring antiretroviral treatment to at least 3 million of those infected. While drug prices have dropped considerably, the cost and technical complexity of laboratory tests essential for the management of HIV disease, such as CD4 cell counts, remain prohibitive. New, simple, and affordable methods for measuring CD4 cells that can be implemented in resource-scarce settings are urgently needed.
Methods and Findings
Here we describe the development of a prototype for a simple, rapid, and affordable method for counting CD4 lymphocytes. Microliter volumes of blood without further sample preparation are stained with fluorescent antibodies, captured on a membrane within a miniaturized flow cell and imaged through microscope optics with the type of charge-coupled device developed for digital camera technology. An associated computer algorithm converts the raw digital image into absolute CD4 counts and CD4 percentages in real time. The accuracy of this prototype system was validated through testing in the United States and Botswana, and showed close agreement with standard flow cytometry (r = 0.95) over a range of absolute CD4 counts, and the ability to discriminate clinically relevant CD4 count thresholds with high sensitivity and specificity.
Conclusion
Advances in the adaptation of new technologies to biomedical detection systems, such as the one described here, promise to make complex diagnostics for HIV and other infectious diseases a practical global reality.
A rapid affordable method for counting CD4 cells may be useful for management of HIV in resource poor settings.
==== Body
Introduction
More than 35 million HIV-infected people live in developing countries with significant resource limitations. Although 6 million people living in developing countries are in urgent need of antiretroviral therapy, only 700,000 currently receive effective treatment [1]. Global treatment efforts, including the World Health Organization's “3 by 5” Initiative, aim to extend therapy to several million people over the next few years [2]. While the cost of antiretroviral medications has dropped considerably, other obstacles, including the cost, technical, and operational requirements of CD4 counts, viral loads, and other sophisticated diagnostic tests used to initiate and monitor HIV treatment, remain to be addressed.
In particular, measurements of CD4+ T lymphocytes are essential for staging HIV-infected patients, determining their need for antiretroviral medications, and monitoring the course of their infection [3]. The CD4 count—expressed in adults as the absolute number of CD4 cells per microliter of blood, and in children as a percentage of total lymphocytes or total T lymphocytes—has enormous prognostic and therapeutic implications, and forms the basis for most HIV treatment decisions [4–6]. In developed countries, CD4 counts are typically performed every three to six months for each patient using the method of flow cytometry. Flow cytometers use lasers to excite fluorescent antibody probes specific for CD4 and other cell surface markers, to distinguish one type of lymphocyte from another. Several factors—including the cost of a flow cytometer (which ranges from $30,000 to $150,000), technical and operational complexity, the need for reliable electricity, and the high cost of reagents—have made these instruments impractical and/or difficult to sustain in resource-scarce settings. The urgent need for affordable and technically simple CD4 diagnostics is widely recognized [7–11].
Several efforts have been made to develop alternative, affordable CD4 counting methods for resource-poor settings. Single-purpose flow cytometers have been designed solely for counting CD4 cells, such as the Becton Dickinson FACSCount, the Partec CyFlow, and desktop instruments from Guava and PointCare Technologies. Although these newer versions make flow cytometry more affordable in some settings, reagent costs remain high, and the instruments remain expensive and in most cases, technically complex [7–13]. Low-cost microbead separation of CD4 cells from other blood cells, followed by standard manual cell counting techniques using a light microscope, offers significantly lower reagent costs than flow cytometry. These methods, however, are low throughput and extremely labor intensive, and appear to be less accurate than traditional flow cytometry; thus, they have not been widely adopted [13–18].
Less expensive CD4 counting methods that capitalize on low-cost microfabrication, efficient light sources, and affordable microelectronics and digital imaging hardware have been conceptualized, but never realized [19,20]. One of us (JTM) has previously reported the development of a novel microchip-based detection system for measuring analytes such as acids, bases, electrolytes, and proteins in solution phase [21–23]. This electronic taste chip (ETC) system carries out chemical and immunological reactions on microspheres positioned in the inverted pyramidal microchamber wells of a silicon or plastic microchip, which is housed in a miniature flow cell. Microfluidic channels deliver a series of small-volume reagents and washes to the flow cell, and hence to the chip and to each one of the microspheres. Optical signals generated by the reactions on the microspheres are visualized and captured on a charge-coupled device (CCD) with the use of transfer optics and a digital video chip. Using the ETC system, complex immunological assays, such as the ones developed to quantify cardiac risk factors in serum, can be performed with small sample volumes, short analysis times, and markedly reduced reagent costs [22].
Further development of the ETC system has shown that it could be adapted to the detection of bacteria, spores, and living cells [24]. We hypothesized that additional modifications could be made to provide accurate, low-cost CD4 counts to monitor HIV infection in resource-constrained settings. We show that a microchip-based system can perform CD4 counts from 16.5 μl of whole blood rapidly, simply, and with a high degree of accuracy compared to flow cytometry, particularly for patients with CD4 counts below 500 cells/μl. We suggest how this prototype system can be readily developed as a low-cost, portable device for use in resource-poor settings.
Methods
Flow Cell
The ETC system was originally designed for microsphere-based assays [21–23]. The modified version of the flow cell (see Figure 1) is enclosed within a three-piece metal casing with a flat platform permanently affixed to a circular vertical support, which is in turn connected to a screw-on cap. Within the metal casing there are top and bottom plastic inserts made from PMMA. Fluids are introduced to and drained out of the flow cell through integrated stainless steel tubing within the inserts. The bottom PMMA insert also features a plastic screen disc that acts as a support for a 3-μm Nuclepore polycarbonate, track-etch filter (Whatman, Florham Park, New Jersey, United States), which serves as a lymphocyte capture and red blood cell separation membrane. A gasket between the membrane and the top insert prevents leaks and ensures that the entire sample is delivered into the flow cell and filtered through the membrane. The top outlet is used with lateral flow for the removal of air bubbles.
Figure 1 Components of the ETC System
A fluid delivery system is used to introduce sample containing fluorescently stained lymphocytes in whole blood and wash buffer to a capture flow cell. Lymphocytes captured within the flow cell are visualized with a fluorescence imaging station using a mercury pressure lamp as a light source, and a CCD for image collection. Raw data images are then processed and analyzed using an automated algorithm run by an attached computer. The flow cell includes a polymer membrane supported on a chip and two transparent polymethylmethacrylate inserts that allow for the optical evaluation of captured lymphocytes.
Fluid Delivery System
In initial studies, we used a single peristaltic pump to deliver sample and washes to the flow cell. Subsequently, a partially automated fluid delivery system was developed. This functional adaptation uses two miniature OEM peristaltic pumps, each in conjunction with a pinch valve, and 0.031-in. (0.79-mm) silicone tubing capable of delivering flow rates of 46–920 μl/min to the flow cell. Integrated software (LabVIEW, National Instruments, Austin, Texas, United States) directs delivery of whole blood samples and washes to the flow cell using the appropriate pumps and valves. Sample filtrate, including red blood cells, is captured in a waste reservoir.
Optical Station and Image Capture
The flow cell was positioned on the stage of a modified BX2 Olympus (Tokyo, Japan) compound microscope equipped with a 10× objective lens and a high-pressure 100 W mercury burner arc lamp as a light source. Focusing was maintained on a fixed plane throughout the duration of the assay. Visualization of AlexaFluor-647-stained lymphocytes was achieved using a Cy5 filter cube (620 nm excitation, 660 nm long-pass beam splitter dichroic mirror, and 700 nm emission), while AlexaFluor-488-stained lymphocytes were visualized with a fluoroisothiocyanate (FITC) filter cube (480 nm excitation, 505 nm long-pass beam splitter dichroic mirror, and 535 ± 25 nm emission). For each study participant, images were obtained from each of five nonoverlapping regions of the lymphocyte capture membrane in the flow cell, using a 12-bit CCD digital camera (DVC, Austin, Texas, United States) mounted on the microscope. Each imaged region represented 0.18 μl of whole blood, so that for each assay, cells were counted from a total volume of 0.9 μl of blood. Each region was imaged serially with both filter cubes. The corresponding images were stored separately as monochromatic eight-bit images for subsequent digital image analysis and automated cell counting.
Image Analysis
Images were analyzed using a custom algorithm supported by Image-Pro Plus (Media Cybernetics, Silver Spring, Maryland, United States) processing software. An iterative approach allowed for flexible analysis of data acquired under different conditions of illumination, focus, and sampling. For each iteration, an upper and lower value defined a range of green or red intensities that were then used to segment the image. Pixels whose intensity values fell within the defined range were reassigned values of one, while all others were set to zero. The process yielded a binary version of the original eight-bit image. A lymphocyte selection algorithm was then applied. Objects (i.e., lymphocytes) were defined as contiguous groups of pixels with values of one. Object selection was refined by a lymphocyte profile (defined by size, aspect ratio, and uniformity); objects not fitting the profile were not counted. The number of counted objects was recorded for each iteration. From one iteration to the next, the upper and lower intensity limits used to segment the image were both increased by a single intensity count. The final cell count per image was the maximum object count over 256 iterations (upper intensity limits 1→255) for which the average object roundness fell below a threshold value. In this manner, the software algorithm determined the optimal analysis parameters for each image individually, greatly relaxing the stringency of image capture requirements. Cell counts were recorded in a spreadsheet as numbers of CD4+CD3−, CD4+CD3+, CD4−CD3+, CD8+CD3−, CD8+CD3+, CD8−CD3+, and CD4+CD8+ cells, depending on the combination of antibodies used. Absolute CD4 counts were recorded as the summed number of CD4+CD3+ cells counted over five images, normalized per microliter of imaged blood. CD4:CD8 ratios were recoded as the ratio of CD4+CD3+ cells to CD8+CD3+ cells counted over five images. Relative CD4 abundance as a percentage of total T lymphocytes was recorded as 100 times the ratio of CD4+CD3+ cells to total CD3+ cells, with cells counted over five images.
Lymphocyte Staining and Delivery
Antibodies utilized in these studies were stored at 4 °C and centrifuged to remove precipitated material prior to use. This process ensured removal of fluorescent particulate matter that could be captured by the membrane and might interfere with imaging. For the initial dilution control studies, CD4 cells were purified by immunomagnetic separation from donor buffy coats. CD4 cells labeled with AlexaFluor-488-conjugated anti-CD4 antibodies (A21335, clone 289–14120, Molecular Probes, Eugene, Oregon, United States) were introduced to the flow cell in amounts ranging from zero to 200,000 cells, and washed with phosphate buffered saline (PBS). For whole blood studies, 33 μl of whole blood collected by venipuncture was incubated at ambient temperature (20–25 °C) with 3 μl of AlexaFluor-488- and AlexaFluor-647-conjugated antibodies to CD4 and CD3 (A21332, clone 289–13801, Molecular Probes), respectively, and allowed to react for 8 min. Similarly, for CD8 enumeration, 33 μl of whole blood with 3 μl of AlexaFluor-488- and AlexaFluor-647-conjugated antibodies to CD8 (A21340, clone 289–13804, Molecular Probes) and CD3, respectively, was allowed to react for 8 min at ambient temperature. Stained blood samples were brought up to 1,000 μl with PBS, half of which was introduced directly into the flow cell (representing 16.5 μl of the original sample of blood) and then washed with 1 ml of PBS. Because red blood cells are mechanically separated from white blood cells, red blood cell lysis is not necessary. Images of labeled cells captured on the membrane were obtained and analyzed as described above. For SEM (scanning electron microscopy), a fixative (2% paraformaldehyde/2.5% glutaraldehyde) was added into the flow cell and rinsed with PBS. The filter was removed from the flow cell, fixed for 90 s with OsO4 vapor, and then dehydrated with EtOH/HMDS. The same SEM protocol was applied to a drop of whole blood on a glass slide.
Study Participants and Comparison to Flow Cytometry
Blood was obtained from HIV-1-uninfected control participants and HIV-infected participants at the Massachusetts General Hospital in Boston, Massachusetts, United States, and from HIV-infected participants at the Botswana–Harvard AIDS Institute HIV Reference Laboratory in Gaborone, Botswana. The Botswana samples originated from a study of HIV-infected pregnant women attending maternal–child health clinics in Gaborone or three nearby villages, Molepolole, Mochudi, and Lobatse. Six infants were also included in the study. Three milliliters of venous whole blood was collected from each participant (in EDTA anticoagulant). All samples were run on the microchip on the day of blood collection. Parallel samples were processed using standard four-color flow cytometry on a Becton Dickinson FACSCalibur, using the MultiTEST reagents and TruCOUNT beads, and analyzed using MultiSET software. All samples were processed by flow cytometry according to standard operating procedure in the HIV reference laboratory in Botswana. The majority were processed within 24 h of blood collection, and all were processed and analyzed within 72 h of blood collection. A total of 70 participants were enrolled, including 64 adults and six infants. Three adults did not have flow cytometry results available, leaving 67 participants for analysis. The study was approved by the institutional review boards of the participating institutions. For a preliminary assessment of assay variability, blood from a single study participant was assayed as described above 20 separate times over the course of a single afternoon by a single operator.
Statistical Methods
The accuracy of the microchip-based CD4 counting system was determined by comparing results directly to parallel samples processed by flow cytometry using Passing–Bablok regression analysis and the Bland–Altman methods comparisons approach [25,26]. For assay reproducibility, a coefficient of variance was calculated from 20 replicates of a single participant. Data were analyzed and processed using Analyse-It software (Analyse-It Software, Leeds, United Kingdom).
Results
In initial experiments using the original ETC system [21–23], microspheres were coated with monoclonal antibodies to the lymphocyte surface markers CD3, CD4, or CD8, followed by microfluidic delivery of fluorescently labeled lymphocytes from whole blood obtained from non-HIV-infected participants. Although lymphocytes were readily captured, precise quantification of cell numbers and CD4 cell counts were not possible using the microsphere as a surface for lymphocyte capture (data not shown). We next modified the flow cells with a disposable, microporous membrane filter for lymphocyte capture. A single polycarbonate, track-etch membrane with 3-μm pores was immobilized and secured within the flow cell, creating a lymphocyte capture surface with a surface area of 80 mm2. Whole blood samples were delivered to the flow cell from a sample reservoir tube, and the membrane within the flow cell was washed with PBS from a second reservoir. As in the original ETC system, cells were imaged under fluorescence optics using a mercury arc lamp light source and a CCD camera (Figure 1).
To confirm that cells could be adequately captured, 33 μl of unprocessed whole blood from non-HIV-infected participants was incubated for 8 min with fluorophore-conjugated anti-CD4 antibodies, and delivered by a peristaltic pump to the modified microfluidics chip. Red blood cells passed readily through the pores under appropriate fluid flow conditions. In contrast, the majority of white blood cells were captured onto a single imaging focal plane (Figure 2). This mechanical separation of autofluorescent red blood cells allows for the imaging and counting of white blood cells from unprocessed whole blood without additional sample processing, such as centrifugation or red blood cell lysis. Using the digital imaging system originally developed for microsphere-based capture in the ETC system, fluorescently labeled white blood cells can then be imaged directly on the chip and counted.
Figure 2 The Membrane Flow Cell Selectively Captures Lymphocytes and Provides for the Removal of Red Blood Cells without Sample Processing
(A) A whole blood sample collected atop a glass slide and imaged by a scanning electron microscope reveals the overabundance of red blood cells in the sample.
(B) A whole blood sample processed through the flow cell reveals that lymphocytes are captured on the membrane support while red blood cells are largely excluded from within the cell. Arrows indicate red blood cells passing through the membrane.
(C) Fluorescent antibody stain specific for a lymphocyte marker is used to visualize captured lymphocytes within the flow cell in a representative single-color data image.
To assess the analytical validity of the membrane-based microchip system, we first performed a dilution control study to evaluate the correlation between total fluorescence intensity and the absolute number of purified CD4 cells from non-HIV-infected participants (labeled with fluorophore-conjugated anti-CD4 antibody) captured in the microchamber. The results show a linear correlation between the number of cells in the sample and the intensity of light emitted from the membrane filter (R
2 = 0.999) for a range of CD4 cell counts relevant to advanced HIV disease (0–200 CD4 cells/μl blood) (Figure 3). This dose–response study established proof of the concept that a modified microfluidic flow cell and a digital image analysis system can accurately detect and measure populations of whole blood lymphocytes labeled with fluorescent markers.
Figure 3 CD4 Lymphocyte Dose Response
Purified CD4 cells were labeled with AlexaFluor-488-conjugated anti-CD4 antibodies, introduced to the flow cell in amounts ranging from zero to 200,000 cells and imaged. There is a linear correlation between the number of cells in the sample and the intensity of light emitted from the membrane filter (R
2 = 0.999).
We next quantified the percentages of CD3, CD4, and CD8 cells in whole blood samples from healthy control participants using this system. Prior to delivery to the flow cell, we labeled a 33-μl whole blood sample with 3 μl of fluorophore-conjugated anti-CD3 and anti-CD4 antibodies for 8 min off chip, then diluted the sample with 961 μl of PBS, and delivered 500 μl of the resulting sample (containing 16.5 μl of blood) to the flow cell using a fluidics controller. Digital images from one region of the lymphocyte capture membrane were obtained with two different emission filters, one specific for the AlexaFluor-488-conjugated antibody used to stain CD4+ T lymphocytes green (Figure 4A), and the other specific for the AlexaFluor-647-conjugated antibody used to stain CD3+ T lymphocytes red (Figure 4B). Automated digital merging of the two images and image processing allowed the system to distinguish the CD3+CD4+ T lymphocytes of interest (i.e., “CD4 cells”), which appear yellow, from the CD4+CD3− monocytes (green) and the CD3+CD4− T lymphocytes (red) (Figure 4C).
Figure 4 Data Collection and Processing for Digital Images Obtained from a Single Diluted Whole Blood Specimen from an HIV-Infected Participant
A total of 16.5 μl of whole blood stained with antibodies specific for CD4 and CD3 markers is delivered to the flow cell after 8 min, and an image of the same region of the membrane is obtained with two different emission filters.
(A) AlexaFluor-488-conjugated anti-CD4 antibody stains CD4+ cells (T lymphocytes and monocytes) green.
(B) AlexaFluor-647-conjugated anti-CD3 antibody stains CD3+ T lymphocytes red.
(C) By digitally merging the two images, CD3+CD4+ T lymphocytes (i.e., “CD4 cells”) appear yellow and are distinguished from CD4+CD3− monocytes (green) and CD3+CD4− T lymphocytes (red).
(D) A lymphocyte selection algorithm is applied to the merged image, based on a lymphocyte profile as defined by size, shape, and uniformity. Objects not fitting the lymphocyte profile are deleted while remaining objects are selected and ultimately counted. A similar protocol to count CD8 cells is used in each participant.
Boxed region indicates two CD4+ cells (yellow in [C]) in the original (A and B), merged (C), and processed (D) images. Large green and red objects seen in some images represent aggregates of fluorescent antibody.
We next developed a custom algorithm for translating these digital images into accurate CD4 and CD8 T cell counts using pixel analysis with the aid of a commercial image processing package. Automated counting of the three subsets of cells was based on object size, aspect ratio, and uniformity, iterated across the range of color intensity levels. As shown in Figure 4D, a binary mask first removes the unwanted cell types, and residual objects representing CD4 T cells are counted. A similar protocol was applied to a second aliquot of blood stained with AlexaFluor-647-conjugated CD3-specific antibody and AlexaFluor-488-conjugated CD8-specific antibody to visualize and count CD3+CD8+ T lymphocytes.
In order to calculate an absolute CD4 count with standard flow cytometry, one of two measures must be undertaken to calculate a concentration in cells per microliter. Either a standardized reference reagent, such as calibration beads at a known concentration, can be added to the assay (“single-platform” flow cytometry), or an absolute total lymphocyte count in cells per microliter can be obtained on a hematology analyzer (“dual-platform” flow cytometry). The microchip assay we describe here uses a direct volumetric method and functions as a single-platform approach. By delivering a consistent volume of blood to the flow chamber (16.5 μl of stained whole blood, diluted to a total volume of 500 μl of PBS), and calculating the unit volume of blood per digital image (0.18 μl), we were able to count the total number of CD4+CD3+ cells in 0.9 μl of blood, and determine the absolute CD4 count per microliter.
We next tested this rapid, whole blood microchip assay in a series of samples acquired in an HIV reference laboratory in Botswana. Seventy consecutive HIV-infected participants presenting to the HIV reference laboratory for standard CD4 counting as part of a vertical transmission study were enrolled, of whom 64 were adult women and six were infants. Parallel samples were processed by standard four-color flow cytometry on a Becton Dickinson FACSCalibur. The time from blood collection to complete analysis and results reporting using the chip-based assay was approximately 15 min per sample. Three adult participants did not have valid flow cytometry results available, leaving 61 adults and six infants for analysis.
Representative processed data images from three participants, two adult women and one infant, are shown in Figure 5. Figure 5A shows a 31-y-old woman with an absolute CD4 count by flow cytometry of 83 cells/μl. While numerous CD3+ T cells (red) are present as well as scattered monocytes (green), her low CD4 count is reflected in the few double-labeled CD3+CD4+ T cells (yellow) seen in the image. Similar representative data images from a young woman with a CD4 count of 271 cells/μl by flow cytometry and a 5-mo-old infant with a CD4 percentage of T lymphocytes of 0.39 by flow cytometry are also shown in Figure 5B and 5C, respectively. These images illustrate the dynamic range of the membrane capture and digital image analysis system, including the ability to quantify both absolute CD4 counts and CD4 percentages.
Figure 5 Representative Processed Data Images from Three Participants in Botswana
The participants included (A) a 31-y-old woman with a CD4 count of 83 cells/μl by flow cytometry; (B) a 33-y-old woman with a CD4 count of 271 cells/μl by flow cytometry; and (C) a 5-mo-old infant with an absolute CD4 count of 2,098 cells/μl and a CD4:CD8 ratio of 1.80 by flow cytometry. In these images, CD3+CD8+ T cells appear red, monocytes appear green, and CD3+CD4+ T cells appear yellow. Each image reflects 0.18 μl of whole blood.
We compared results from our microchip assay with results available from flow cytometry, the latter obtained on a FACSCalibur through standard clinical laboratory operating procedures. The data for adult absolute CD4 counts are plotted in the Bland–Altman methods comparison plot shown in Figure 6. For 61 adult participants with CD4 counts ranging from 35 to 1,087 cells/μl (mean, 372 cells/μl) by flow cytometry, results show a good correlation between absolute CD4 counts measured by our microchip assay and those measured by flow cytometry. Bland–Altman methods comparison analysis shows a bias of −50 cells/μl (95% confidence interval, −81 to −20 cells/μl), and good 95% limits of agreement (Figure 6). Several of the results from participants at the higher end of absolute CD4 counts fall outside the 95% limits. For these participants, individual lymphocytes may overlap in the digital images (as seen in Figure 5C), which can interfere with the accuracy of the lymphocyte counting algorithm. In resource-limited settings, the primary use of CD4 counts is as a trigger to initiate antiretroviral therapy, which typically occurs at a CD4 count of 200 cells/μl. Higher CD4 count thresholds of 350 and 500 cells/μl are also used to increase the intensity of monitoring. For these values, the sensitivity and specificity of our method are: CD4 < 250, sensitivity = 0.86, specificity = 0.81; CD4 < 350, sensitivity = 0.97, specificity = 0.83; and CD4 < 500, sensitivity = 0.96, specificity = 0.85.
Figure 6 Methods Comparison and Correlation Studies for Absolute CD4 Counts in 61 Adults in Botswana
Bland–Altman methods comparison plot comparing absolute CD4 cells per microliter of whole blood obtained by the microchip system as compared to standard four-color flow cytometry processed in parallel on a FACSCalibur in 61 HIV-infected adult participants. There is a proportional bias of −50 cells/μl relative to flow cytometry. Grey line indicates zero bias. Red lines indicate upper and lower 95% limits of agreement.
One important application of our method is in pediatric HIV monitoring. The wide range of normal absolute CD4 counts in infants and children requires the use of CD4:CD8 ratios or CD4 percentages in pediatric infection. Results for CD4:CD8 ratios and CD4 percentages of T lymphocytes for all 67 participants (61 adults and six infants) are shown in Figure 7. Agreement, bias, and correlations between the microchip method and flow cytometry are excellent for both CD4 percentages of T lymphocytes (Figure 7A and 7B) and CD4:CD8 ratios (Figure 7C and 7D). Bland–Altman plots for both CD4 percentages of T lymphocytes and CD4:CD8 ratios show low proportional bias, with tight 95% limits of agreement. Correlations are excellent for both CD4 percentages of T lymphocytes (r = 0.98, p < 0.0001) and CD4:CD8 ratios (r = 0.98, p < 0.0001). Overall, the data show that all three approaches to measuring CD4 cell counts can be accurately quantified using the microchip method, and that both adult and pediatric CD4 results can be obtained.
Figure 7 Methods Comparison and Correlation Studies for CD4 Percentages of Total T Cells and CD4:CD8 Ratios in 67 Human Subjects
(A and B) CD4 percentages of total T cells and (C and D) CD4:CD8 ratios in 67 human participants, including 61 adults and six children. In Passing–Bablok correlation plots (A and C), solid black lines indicate identity, blue lines indicate the observed correlations, and dashed black lines indicate 95% confidence limits. Correlations for CD4 percentages of total T cells (r = 0.98, p < 0.0001) and CD4:CD8 ratios (r = 0.98, p < 0.0001) are shown. For Bland–Altman methods comparison plots (B and D), notations are as described in Figure 6 caption.
To determine assay variability, we examined 20 replicate samples of blood from a single participant over the course of one day, using the established basic protocol. We determined that the coefficient of variance was 12% (data not shown), which is similar to other methods of CD4 counting [27]. Although the assay described here introduced 16.5 μl of blood into the system, the actual volume of blood analyzed by digital image analysis is only 0.90 μl. We have conducted preliminary studies that suggest that we can accurately measure CD4 counts from less than 5 μl of blood obtained via fingerstick (data not shown); additional studies will be required to assess the correlation between CD4 counts obtained by fingerstick and by venipuncture.
Discussion
Our results provide proof of principle that low-cost microfluidic structures combined with fluorescence imaging and digital image analysis can be successfully applied to the measurement of CD4 cell counts, which are critical to the clinical management of HIV infection. The method described here can deliver both absolute CD4 counts for adult monitoring, and CD4 percentages or CD4:CD8 ratios for pediatric monitoring. Most importantly, the rapid and accurate CD4 assessments obtained with this method, together with its anticipated low cost relative to flow cytometry, may make this type of approach ideal for resource-scarce settings. As our results show, this method may be less accurate at the higher range of CD4 counts, where cells may be more likely to overlap in our digital images. While this may limit its applicability, our method is accurate at CD4 counts below 500 cells/μl, which represent the clinically relevant CD4 levels in resource-poor settings. In addition, both the bias in the method described here (−50 cells) and the accuracy at higher CD4 counts are likely to be improved significantly by the further development of a disposable microfluidic cartridge, where the volume of distribution of the sample will be much smaller, and more accurate volumetric control will be possible.
Our study was designed to evaluate the accuracy of our method against the gold standard in a population of adults. During enrollment, a small number of pediatric samples were made available to us by the staff at Princess Marina Hospital in Botswana. We chose to include these samples in the data presented here to provide proof of principle that pediatric CD4 percentages can also be assessed with this method. Although only six pediatric samples were available, limiting claims of statistical significance, we believe the issue of pediatric CD4 count monitoring to be of such importance that the data merited inclusion. Excluding the six pediatric samples does not affect the analysis.
The results presented here were obtained with a stationary, tabletop monitoring system using a standard epifluorescence microscope and commercial image processing software. While the methods we described provide the basis for a highly portable and flexible miniaturized CD4 counting system, it should be emphasized that a number of additional developments are required to enable the widespread use of this approach in resource-limited settings. With additional engineering of optics, electronics, and mechanical components along with advancements in integrated microfluidic systems, it should be possible to develop a point-of-care instrument that is battery-powered, uses simple light emitting diodes (LEDs), and secures analyzable digital images with affordable video imaging chips. When combined with an embedded microprocessor and disposable assay cartridges for both adult and pediatric monitoring manufactured from injection-molded plastic, it should be possible to create a functional CD4 counting device that can be used at the point of care. Further trials in a larger, more diverse cohort of patients, including adult men and children, will be necessary to confirm the accuracy of the method, including an assessment of assay bias and reproducibility. Such a device is currently in commercial development, and may be available by early 2006. While it is too early to provide an accurate cost estimate for a portable instrument and disposable plastic CD4 assay, we expect the equipment cost would be substantially lower than for flow cytometry, and the assay cost would be similar to assays using existing methods (Table 1).
Table 1 A Comparison of Methods for CD4 Determination
Although several CD4 counting systems are now used in resource-limited settings, they remain suboptimal to meet the needs of HIV care and treatment scale-up. None can truly be used at the point of care beyond a district hospital or similar facility, and either the capital and operating costs remain high, or throughput is low, or both (Table 1). Pediatric monitoring using CD4 percentages also remains largely unavailable. The method we describe here addresses several of the limitations of performing diagnostic assays in resource-limited settings. First, sample volumes are minimal, so that tests can be performed on fingerstick samples of blood, circumventing the need for venipuncture, and minimizing both medical waste and operator exposure to biohazardous material. Second, reagent use is minimized in the microchip system, reducing reagent costs by as much as 90%. Third, labor- and equipment-intensive sample preparation is eliminated. Fourth, the microchip CD4 assay is extremely rapid. CD4 results in the prototype system described here are available in less than 15 min from the time of blood collection. In a mature microfluidic device with push-button operation, results should be available in less than 10 min, and thus can be used to make real-time clinical decisions at the point of care. Fifth, the assay is technically simple, analogous to a portable glucometer, and ultimately will be useable by a health-care worker in remote settings with minimal training, extending the reach of CD4 assays to district hospitals and remote clinics, and reducing labor costs. Sixth, both adult and pediatric monitoring are possible.
We believe that the future of low-cost diagnostics for use in the developing world lies in the development of new lab-on-a-chip technologies that integrate sample preparation and sample measurement systems into miniaturized devices with minimal power requirements. Preliminary cost estimates for the instrumentation here described suggest, at a minimum, a 10-fold reduction in the cost for the associated measurement system. Further, reagent consumption for the microchip system can be reduced by a similar factor relative to flow cytometry, while sample storage and shipping costs are expected to be reduced dramatically by virtue of the point-of-care capabilities of this new lab-on-a-chip structure. The importance of microtechnologies to the realities of laboratory infrastructure worldwide has been recognized previously [28–30]. Although CD4 counting represents the most urgent need in HIV diagnostics for resource-poor settings, the microchip platform is adaptable to other important assays. Through the interface of the lymphocyte capture membrane described here with the previously reported microchip arrays, cellular assays like CD4 counts can be multiplexed with other molecular biomarker measurements (i.e., proteins and nucleic acids) on a single miniaturized chip. The rapid extension of the chip-based CD4 counting method described here to HIV RNA measurements, diagnostics for opportunistic infections, liver enzymes, and other biochemical markers of interest in infectious disease is feasible.
Patient Summary
Background
Most HIV-infected people don't develop AIDS right away, because their immune systems can keep the virus in check for months and sometimes years. In general, doctors don't recommend that infected people start taking HIV medications while their immune system is still healthy. Doctors know whether or not a patient's immune system is healthy-and therefore whether or not to start treatment-by measuring the "CD4 count." This is the number of CD4 cells in a sample of blood. CD4 cells, also called CD4+ T cells, are a type of white blood cell that fights infection. HIV destroys CD4 cells, weakening the body's immune system and ultimately causing AIDS. CD4 counts should be determined before a patient receives antiretroviral therapy and then measured regularly while the patient is on therapy.
Why Was This Study Done?
Most tools available to count CD4 cells are large and expensive to buy, and every actual count is also expensive and difficult to carry out. These tools are therefore unsuitable for many low-income countries. The researchers wanted to develop a tool that allows easier and cheaper measurement of CD4 cells, and is small and simple enough that health-care workers can take it to patients when they visit them in remote areas.
What Did the Researchers Do?
They built a prototype for a new tool that counts CD4 cells in a simpler and cheaper way. They then took samples from 61 adults and six children and compared the results when they used both the standard technologies and their prototype counter.
What Did They Find?
They found that their prototype works well and is indeed cheaper and easier to use. It also appears to be just as reliable as the large and more complicated machines in helping doctors make decisions about when to start therapy and when to change therapies. They have done enough samples from adults infected with HIV to be confident about that. It looks like it might work for infected children as well, but they haven't done enough child samples yet to be certain.
What Does This Mean?
This suggests that with some additional work, it should be possible to develop a handheld CD4 counter that is cheap, easy, and transportable. This could make a big difference for the care of HIV patients in developing countries and other remote areas.
What Next?
There is still more development work to do to get from the prototype to a handheld counter, and the researchers should also study more samples from children to see whether the new test is equally reliable for pediatric patients.
Additional Online Resources
Information about CD4 monitoring for HIV/AIDS patients can be found at the following Web sites.
World Health Organization (search for CD4 count):
http://www.who.int/en/
The World Health Organization's specific review of CD4 counting technologies:
http://www.who.int/3by5/amds/Suzanne_Crowe.pdf
The Well Project:
http://www.thewellproject.org/
The Well Project's specific page on CD4 counts:
http://www.thewellproject.org/Treatment_and_Trials/First_Things_First/Understanding_CD4_and_CD8_Cells.jsp
Find Diagnostics, an organization focused on affordable diagnostics for infectious diseases worldwide:
http://www.finddiagnostics.org/
The Forum for Collaborative HIV Research, a consortium with an advocacy and funding role for HIV/AIDS research priorities, including low-cost CD4 and viral load tests:
http://www.hivforum.org/
Project Inform (search for CD4 and monitoring):
http://www.projectinform.org/
The authors would like to thank Drs. Max Essex, Joel Katz, Shahin Lockman, Ric Marlink, and Roger Shapiro for their assistance with field testing of the microchip system in Botswana. This work was supported by grants from the National Institutes of Health, the Bill and Melinda Gates Foundation, and the Doris Duke Charitable Foundation. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Citation: Rodriguez WR, Christodoulides N, Floriano PN, Graham S, Mohanty S, et al. (2005) A microchip CD4 counting method for HIV monitoring in resource-poor settings. PLoS Med 2(7): e182.
Abbreviations
CCDcharge-coupled device
ETCelectronic taste chip
PBSphosphate buffered saline
==== Refs
References
Joint United Nations Programme on HIV/AIDS, World Health Organization AIDS epidemic update: 2004 2004 December Available: http://whqlibdoc.who.int/unaids/2004/9291733903.pdf . Accessed 9 March 2005
World Health Organization 2004 World Health Organization (2004) 3 by 5 progress report December 2003 through June 2004 Available: http://www.who.int/3by5/en/Progressreport.pdf . Accessed 15 January 2005
Department of Health and Human Services Guidelines for the use of antiretroviral agents in HIV-1-infected adults and adolescents 2005 April 7 Available: http://aidsinfo.nih.gov/guidelines/adult/AA_040705.pdf . Accessed 9 June 2005
O'Brien WA Hartigan PM Martin D Esinhart J Hill A Changes in plasma HIV-1 RNA and CD4+ lymphocyte counts and the risk of progression to AIDS. Veterans Affairs Cooperative Study Group on AIDS N Engl J Med 1996 334 426 431 8552144
Mellors JW Munoz A Giorgi JV Margolick JB Tassoni CJ Plasma viral load and CD4+ lymphocytes as prognostic markers of HIV-1 infection Ann Intern Med 1997 126 946 954 9182471
O'Brien WA Hartigan PM Daar ES Simberkoff MS Hamilton JD Changes in plasma HIV RNA levels and CD4+ lymphocyte counts predict both response to antiretroviral therapy and therapeutic failure. VA Cooperative Study Group on AIDS Ann Intern Med 1997 126 939 945 9182470
Glencross DK Stevens G Scott LE Mendelow BV Stevens W The challenge of laboratory monitoring of HIV S Afr Med J 2002 92 248
Glencross D Scott LE Jani IV Barnett D Janossy G CD45-assisted PanLeucogating for accurate, cost-effective dual-platform CD4+ T-cell enumeration Cytometry 2002 50 69 77 12116348
Huff B Low cost diagnostics GMHC Treat Issues 2001 15 1 4
Bartholomew C Prices of CD4 assays and viral load tests must be reduced for developing countries BMJ 2001 323 809 810
Sherman GG Galpin JS Patel JM Mendelow BV Glencross DK CD4+ T cell enumeration in HIV infection with limited resources J Immunol Methods 1999 222 209 217 10022387
Janossy G Jani IV Bradley NJ Bikoue A Pitfield T Affordable CD4+ -T-cell counting by flow cytometry: CD45 gating for volumetric analysis Clin Diagn Lab Immunol 2002 9 1085 1094 12204964
Kannangai R Ramalingam S Jesudason MV Vijayakumar TS Abraham OC Correlation of CD4+ T-Cell counts estimated by an immunocapture technique (Capcellia) with viral loads in human immunodeficiency virus-seropositive individuals Clin Diagn Lab Immunol 2001 8 1286 1288 11687479
Kannangai R Prakash KJ Ramalingam S Abraham OC Mathews KP Peripheral CD4+ /CD8+ T-lymphocyte counts estimated by an immunocapture method in the normal healthy south Indian adults and HIV seropositive individuals J Clin Virol 2000 17 101 108 10942090
Diagbouga S Durand G Sanou PT Dahourou H Ledru E Evaluation of a quantitative determination of CD4 and CD8 molecules as an alternative to CD4+ and CD8+ T lymphocyte counts in Africans Trop Med Int Health 1999 4 79 84 10206260
Landay A Ho JL Hom D Russell T Zwerner R A rapid manual method for CD4+ T-cell quantitation for use in developing countries AIDS 1993 7 1565 1568 7904449
Didier JM Katachkine MD Demouchy C Moat C Diagbouga S Comparative assessment of five alternative methods for CD4+ T-lymphocyte enumeration for implementation in developing countries J Acquir Immune Defic Syndr 2001 26 193 195 11242190
Lyamuya EF Kagoma C Mbena EC Urassa WK Pallangyo K Evaluation of the FACScount, TRAx CD4 and Dynabeads methods for CD4 lymphocyte determination J Immunol Methods 1996 195 103 112 8814325
Shapiro H “Cellular astronomy”—A foreseeable future in cytometry Cytometry A 2004 60 115 124 15290712
Janossy G Jani IV Kahan M Barnett D Mandy F Precise CD4 T-cell counting using red diode laser excitation: For richer, for poorer Cytometry B Clin Cytom 2002 50 78 85
Goodey AP McDevitt JT Multishell microspheres with integrated chromatographic and detection layers for use in array sensors J Am Chem Soc 2003 125 2870 2871 12617642
Christodoulides N Tran M Floriano PN Rodriguez M Goodey A A microchip-based multianalyte assay system for the assessment of cardiac risk Anal Chem 2002 74 3030 3036 12141661
Goodey A Lavigne JJ Savoy SM Rodriguez MD Curey T Development of multianalyte sensor arrays composed of chemically derivatized polymeric microspheres localized in micromachined cavities J Am Chem Soc 2001 123 2559 2570 11456925
Floriano PN Christodoulides N Romanovicz D Bernard B Simmons GW Membrane-based on-line optical analysis system for rapid detection of bacteria and spores Biosens Bioelectron 2005 20 2079 2088 15741078
Bland JM Altman DG Measuring agreement in methods comparison studies Stat Methods Med Res 1999 8 135 160 10501650
Passing H Bablok W A new biometrical procedure for testing the equality of measurements from different analytical methods. Application of linear regression procedures for method comparison studies in clinical chemistry J Clin Chem Clin Biochem 1983 21 709 720 6655447
Whitby L Granger V Storie I Goodfellow K Sawle A Quality control of CD4+ T-lymphocyte enumeration: results from the last 9 years of the United Kingdom National External Quality Assessment Scheme for Immune Monitoring (1993–2001) Cytometry 2002 15 102 110
Daar AS Thorsteinsdóttir H Martin DK Smith AC Nast S Top ten biotechnologies for improving health in developing countries Nat Genet 2002 32 229 232 12355081
Harris E Developing essential scientific capability in countries with limited resources Nat Med 1996 2 737 739 8673911
Harris E Tanner M Health technology transfer BMJ 2000 321 817 820 11009526
| 16013921 | PMC1176233 | CC BY | 2021-01-05 10:40:14 | no | PLoS Med. 2005 Jul 19; 2(7):e182 | utf-8 | PLoS Med | 2,005 | 10.1371/journal.pmed.0020182 | oa_comm |
==== Front
PLoS MedPLoS MedpbioplosbiolPLoS Medicine1549-12771549-1676Public Library of Science San Francisco, USA 1600850010.1371/journal.pmed.0020214PerspectivesHIV/AIDSHIV Infection/AIDSMedicine in Developing CountriesCD4 Measurements in Patients with HIV: Are They Feasible for Poor Settings? PerspectivesBentwich Zvi Zvi Bentwich is Professor of Medicine (Emeritus), Hebrew University Medical School, Professor in the Department of Microbiology and Immunology, Ben Gurion University Health Sciences Faculty, Beer Sheba, and Chief Scientist, Rosetta Genomics, Rehovot, Israel. E-mail: [email protected]
Competing Interests: The author declares that he has no competing interests. Rosetta Genomics has not developed any devices for HIV evaluation in resource-poor settings.
7 2005 19 7 2005 2 7 e214Copyright: © 2005 Zvi Bentwich.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.
A Microchip CD4 Counting Method for HIV Monitoring in Resource-Poor Settings
Test Your Knowledge: HIV Infection Epidemiology and Diagnosis
Towards a Cheap and Easy Way to Monitor HIV/AIDS
The lack of facilities to measure CD4 counts in poor countries impedes instituting rational and effective antiretroviral therapy in these countries. Can a new microchip counting technique help to solve the problem?
==== Body
Measurement of peripheral blood CD4 T lymphocytes is probably the most important laboratory assay for evaluation and monitoring of patients with HIV. The CD4 count is critical for determining the clinical stage of HIV infection, for deciding when to start antiretroviral therapy (ART), for evaluating the efficacy of treatment, and for changing the medications when necessary. Most HIV treatment decisions are therefore based upon the CD4 count [1–3].
Flow Cytometry
The most common technique for measuring CD4 counts in developed country settings is flow cytometry. Flow cytometers use lasers to excite fluorescent antibody probes specific for various cell surface markers, such as CD3, CD4, and CD8, which distinguish one type of lymphocyte from another.
As Rodriguez et al. point out in their study in this issue of PLoS Medicine [4], the cost of a flow cytometer ranges from $30,000 to $150,000, and the reagents needed for determining the lymphocyte surface markers by this method are very costly. In addition, use of flow cytometry requires technical and operational expertise as well as a reliable electricity source. Considering all these factors together, it is no surprise that CD4 measurements cannot be widely applied in developing world settings.
Why CD4 Counts Matter in Developing Countries
This grim reality—the lack of facilities to measure CD4 counts in poor countries—stands in sharp contrast to the urgent need for instituting rational and effective ART in these countries. The absence of tools to measure CD4 counts clearly jeopardizes the success of the recently launched global campaigns to fight AIDS, such as those of the World Health Organization and the Global Fund to Fight AIDS, Tuberculosis, and Malaria. These campaigns aim to distribute ART to millions of people with HIV, mostly living in developing countries. Regretfully, it is highly likely that these major efforts will fail, unless improved and widely used means for counting CD4 cells become available and can be applied where they are most needed. Since at least 35 million people are infected with HIV, and several million of them are in need of urgent lifesaving ART, the issue of CD4 monitoring has become a crucial one.
Rodriguez et al. point out that several efforts have been made to develop alternative, affordable CD4 counting methods for resource-poor settings [4]. These include improved flow cytometric approaches and microbead capture/separation of CD4 cells followed by manual cell counting [5–8]. Also, new single-purpose flow cytometers have been designed that perform the test at a much lower price. Though all these assays are indeed cheaper than regular flow cytometry, they suffer from decreased accuracy and, most importantly, they are all of low throughput.
A New Method for Counting CD4
Rodriguez et al. describe a novel method for counting CD4 in resource-poor settings (Figure 1) [4]. The method is based on a novel microchip detection system for measuring various analytes in very small volumes. A series of chemical and immunological reactions carried out on microspheres are visualized and captured on a charge-coupled device (developed for digital camera technology). This method allows for accurate measurement of CD4, CD8, and CD4/CD8. The prototype used for demonstration of the new apparatus shows extremely good agreement with currently used flow cytometry. Most importantly, the investigators claim that the cost of each assay is much lower than that for flow cytometry.
Figure 1 CD4 Cell Measurement Using a Prototype Microchip Counting Method This is a digital image of whole blood from a five-month-old male infant from Botswana with an absolute CD4 count of 2,098 cells/ml and a CD4 percentage of T cells of 0.39, obtained using a prototype method for low-cost CD4 count monitoring. CD4+ T cells are yellow. Also visualized are monocytes (green) and CD8+ T cells (red). (Photo courtesy of the authors of [8])
There are, however, a number of unresolved issues in this study that need further clarification before the assay can meet the expectations for becoming a widely used tool in resource-poor settings. Firstly, the study was performed with a prototype apparatus, tailored to meet the requirements of the study, but not yet representing a commercially established and viable production line. Secondly, though the authors state that the price of each CD4 determination will become much cheaper, it is not clear how much each assay will cost in the end, and whether the final cost is realistic in the context of developing countries. It is clear, though, that the actual price of the assay will change once it is widely and consistently used on a large scale. Thirdly, although a few children were tested (six infants in total), the results in this small group remain questionable, and therefore the application of the test to pediatric populations needs further testing. It may well be that application to pediatric patients will require an improved apparatus or improved handling.
Conclusion
Despite these reservations, the authors of this study should be commended for addressing an extremely important issue and developing this novel approach for counting CD4 in patients with HIV. Their study may lead to further development of such an apparatus, which is sorely needed for the global fight against AIDS. Such efforts will hopefully be noticed by public funding agencies, leading to the improvement of tools for measuring CD4 counts.
Citation: Bentwich Z (2005) CD4 measurements in patients with HIV: Are they feasible for poor settings? PLoS Med 2(7): e214.
Abbreviation
ARTantiretroviral therapy
==== Refs
References
Mellors JW Munoz A Giorgi JV Margolick JB Tassoni CJ Plasma viral load and CD4+ lymphocytes as prognostic markers of HIV-1 infection Ann Intern Med 1997 126 946 954 9182471
Fahey J Taylor JM Manna B Nishanian P Aziz N Prognostic significance of plasma markers of immune activation, HIV viral load, and CD4 T cell measurements AIDS 1998 14 1581 1590
Didier JM Kazatchkine MD Demouchy C Moat C Diagbouga S Comparative assessment of five alternative methods for CD4+ T-lymphocyte enumeration for implementation in developing countries J Acquir Immune Defic Syndr 2001 26 193 195 11242190
Rodriguez WR Christodoulides N Floriano PN Graham S Mohanty S A microchip CD4 counting method for HIV monitoring in resource-poor settings PLoS Med 2005 2 e182 10.1371/journal.pmed.0020182 16013921
Landay A Ho JL Hom D Russell T Zwerner R A rapid manual method for CD4+ T-cell quantitation for use in developing countries AIDS 1993 7 1565 1568 7904449
Janossy G Jani IV Bradley NJ Bikoue A Pitfield T Affordable CD4(+)-T-cell counting by flow cytometry: CD45 gating for volumetric analysis Clin Diagn Lab Immunol 2002 9 1085 1094 12204964
Kannangai R Ramalingam S Jesudason MV Vijayakumar TS Abraham OC Correlation of CD4(+) T-cell counts estimated by an immunocapture technique (Capcellia) with viral loads in human immunodeficiency virus individuals Clin Diagn Lab Immunol 2001 8 1286 1288 11687479
Sherman GG Galpin JS Patel JM Mendelow BV Glencross DK CD4+ T cell enumeration in HIV infection with limited resources J Immunol Methods 1999 222 209 217 10022387
| 16008500 | PMC1176234 | CC BY | 2021-01-05 11:13:38 | no | PLoS Med. 2005 Jul 19; 2(7):e214 | utf-8 | PLoS Med | 2,005 | 10.1371/journal.pmed.0020214 | oa_comm |
==== Front
PLoS MedPLoS MedpbioplosbiolPLoS Medicine1549-12771549-1676Public Library of Science San Francisco, USA 1600850510.1371/journal.pmed.0020216EssayEpidemiology/Public HealthHIV/AIDSHIV Infection/AIDSMedicine in Developing CountriesEpidemiologyPublic HealthHIV in Nepal: Is the Violent Conflict Fuelling the Epidemic? EssaySingh Sonal [email protected] Edward Honeyman Steven Krishna Suvedi Bal Prasad Pant Nur Sonal Singh is an MPH student at the Bloomberg School of Public Health, Johns Hopkins University, Baltimore, Maryland, United States of America and a resident physician at the Department of Medicine, Unity Health System, Rochester, New York, United States of America. Edward Mills is at the Department of Clinical Epidemiology and Biostatistics, McMaster University, Hamilton, Ontario, Canada. Steven Honeyman is Country Representative, Population Services International, Budhanilkantha, Nepal. Bal Krishna Suvedi is Senior Medical Officer, Ministry of Health, Kathmandu, Nepal. Nur Prasad Pant is an HIV/AIDS specialist, CARE-Nepal, Lalitpur, Nepal.
Competing Interests: The authors declare that they have no competing interests.
8 2005 19 7 2005 2 8 e216Copyright: © 2005 Singh 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.Singh and colleagues discuss the HIV epidemic in Nepal, including the local and international response from the health and development community.
==== Body
HIV/AIDS has reached alarming proportions in Southeast Asia [1]. The magnitude of the epidemic is projected to exceed that of sub-Saharan Africa in the 21st century [2]. More than 7 million South Asians are currently infected with HIV [3], nearly 5 million of whom are in India [3].
Nepal has had a comparatively lower prevalence of HIV/AIDS compared to other countries in Southeast Asia. Seasonal migration and sexual trafficking across a porous Indian border [4], fuelled by a bloody Maoist conflict (see Sidebar), has raised Nepal's HIV prevalence to the second highest in the region after India.
In this essay, we characterize the current HIV epidemic in Nepal, look at the ways in which the conflict may be fuelling the infection rate, and discuss the current local and international response from the health and development community. We argue that there is currently a short window of opportunity to take action to control the HIV epidemic in Nepal—and that inaction will lead to HIV becoming the biggest killer of the young and middle-aged in the next decade.
The Current HIV Epidemic in Nepal
The first case of AIDS was reported in Nepal in 1998 [5]. Most cases of HIV infection in Nepal are HIV-1, although HIV-2 was also recently reported [6]. As of February 2005, the National Center for AIDS and STD Control in Nepal reported that there were 4,755 HIV-positive people and 856 confirmed cases of AIDS in Nepal [7]. However, because of the poor surveillance systems and the lack of access to quality voluntary counselling and testing services coupled with antiretroviral treatment, these prevalence figures are likely to be a gross under-estimate [2].
Violent Conflict in Nepal
Nepal is one of the poorest countries in the world. Rebels, led by the Communist Party of Nepal (Maoist), have been fighting a bloody war with the government forces, led by the Royal Nepalese Army, since 1996. Originating from the western heartlands of Nepal, which were the least economically developed and most inaccessible regions of the country, the conflict has spread to nearly all of the 75 districts of Nepal. It has claimed more than 10,000 lives, led to widespread destruction of infrastructure, and disrupted the flow of essential supplies. Both sides have been accused of serious human rights violations. Violent clashes have escalated since the king seized power in February 2005, ousting his appointed prime minister and imposing a state of emergency [43].
A dramatically higher estimate comes from UNAIDS, which estimated that 62,000 out of a population of 24.1 million in Nepal were living with HIV/AIDS in 2003 [8]. One in 200 (0.5%) people aged 15–49 years are living with HIV/AIDS in Nepal [8]. About 30% of those infected are female [8]. WHO/UNAIDS estimates that 940 children are living with HIV and that nearly 13,000 children were orphaned in Nepal due to AIDS at the end of 2003 [5]. The prevalence in the general population may still be low, but it masks an increasing prevalence in several risk groups—the prevalence of HIV/AIDS consistently exceeds 5% in injecting drug users, commercial sex workers, and migrant workers [5].
High-Risk Groups
Among injecting drug users—estimated to be about 30,000 in Nepal—about 40% are HIV-positive [9]. Needle sharing and risky sexual behaviour is common in this group. The figures are particularly alarming in Kathmandu, the capital city, where nearly 68% of injecting drug users are HIV-positive [10]. A subtype C virus with restricted genetic diversity is thought to have caused this epidemic in Kathmandu [11]. Concomitant hepatitis C infection is a contributing factor to the rapidity and severity of disease progression in injecting drug users, and 94% of users in Kathmandu have tested positive for hepatitis C [12].
The predominant mode of transmission of HIV in Nepal is heterosexual contact with commercial sex workers. HIV prevalence rates are about 4% among sex workers in the Terai regions of Nepal and about 1.5% in their clients (which is more than five times the national average prevalence) [13]. In Kathmandu, nearly 17% of sex workers are HIV-positive [8]. There are about 25,000 sex workers in Kathmandu [4] and an estimated 200,000 Nepalese women working in Indian brothels. About 5,000 to 10,000 Nepalese sex workers are trafficked every year [4], numbers that are likely to increase as a result of the conflict. One striking estimate is that nearly 70% of sex workers returning from India are HIV-positive [8].
As in India, a major contributor to the spread of HIV in Nepal has been Nepal's mobile population, including truckers and migrant workers. About half of the estimated cases of HIV are in 29 highway districts of Nepal [14]. A 1999 survey indicated that 75% of the truckers surveyed had had sex with a sex worker and that only 70% of the truckers had used a condom during their last sexual encounter [15].
Men who have sex with men still account for only a small proportion of those affected with HIV in Nepal. But the recent detention and subsequent release of 39 male transvestite members of the Blue Diamond Society, a local organization that provides sexual health, HIV/AIDS, and advocacy services to sexual minorities, highlights the challenges faced by sexual minorities in Nepal [16].
The Role of Conflict in Fuelling the Epidemic
Several factors contribute to the propagation of HIV in times of conflict (Box 1) [17]. While accurate numbers are hard to come by, the recent conflict may have contributed to the propagation of HIV/AIDS by fuelling displacement [18]. As the insurgency drags on, seasonal and long-term migration of labourers to neighbouring countries, such as India, is becoming critical to the economic survival of many households. Young men have left the country “en masse” for fear of execution and migrated to the high-prevalence areas in India. UNAIDS estimates at least 10% of the 2 million to 3 million Nepalese migrant workers in India are HIV-positive [19]. These men are now infecting spouses and others in many parts of the country. By pushing rural residents from war-torn areas to the capital, Kathmandu, the conflict may have helped spread HIV/AIDS.
Box 1. Factors Contributing to the Propagation of HIV in Times of Conflict
Increased levels of commercial sex
Breakdown of vital services in health and education
Decreased availability and use of reproductive health services
Decreased use of means to prevent HIV transmission
Increased population mixing following displacement, which may promote high-risk behaviour
Fragmentation of families
Sexual violence
Stigmatization and discrimination related to HIV
Prayer flags in the valley of Katmandu, Nepal
Some 200,000 to 400,000 people may have been displaced in Nepal since the beginning of the conflict [20]. Migrations within the country and internationally, coupled with a general lack of awareness and knowledge about risk factors, is likely to have contributed to the propagation of the epidemic. The far western regions of the country, which are less economically developed and which are the hotbed of the current insurgency, are more prone to migration. These regions have one of the fastest growth rates of HIV in South Asia.
Migrants are more likely to practice high-risk sexual behaviour [21]. About 8% of migrant male workers returning from Mumbai and examined in western Nepal were infected with HIV [22]. The epidemiological impact on women in migrant communities has yet to be realized, as only 0.5% of 900 women in the Kailali district in western Nepal were HIV-positive, although more than 30% had some form of STD [23].
Sex traffickers have shifted their trade from Sindhupalchowk and Nuwakot in central Nepal to Rukum and Rolpa (the hotbeds of the insurgency) in the mid-west, taking undue advantage of the socio-economic conditions borne by conflict and violence. The low socio-economic status of women along with the current conflict makes the women in this region particularly vulnerable [24]. Save the Children Norway's recent study, Impacts of Armed Conflict Pushing Women and Girls Into Sexual Abuse and Sex Trade, revealed that about 19% of female sex workers stated they had entered the sex trade directly because of the conflict [25].
While the number of people infected has risen, HIV prevention and awareness work has declined in Nepal as a result of the conflict. Save the Children's work in the Accham district of western Nepal has been hindered by fighting between the Maoist rebels and government forces since early 2002 [26]. Offices of nongovernmental organizations (NGOs) have been burned, and volunteers are afraid to work [26]. Médecins Sans Frontières was forced to curtail its activities last year in Jumla, one of the poorest districts in the Midwest, due to the conflict. Local activists fear that the situation will continue to deteriorate if NGOs are not allowed to work in Jumla. People living with HIV/AIDS in rural Nepal are desperate for care and support.
Condom Use
In a study of young men in Nepal in 2001, only 42% of men used a condom consistently [27]. The vast majority of young men, about 80%, who had sex with non-regular partners felt they were not at risk of contracting any form of sexually transmitted disease or HIV [27]. The reasons for non-use included fear of losing sexual pleasure, embarrassment over buying condoms, and a belief that careful selection of partners offers sufficient protection [27]. Monitoring of behavioural trends has shown that condom use during the last sexual encounter as reported by female sex workers has markedly increased from 63% in 1998 to 90% in 2002 [28].
However, consistent use of condoms as reported by female sex workers (54%) still needs improvement [28]. Many female sex workers are young children and lack the power to negotiate safe sex [29]. Many of the young girls are considered free of disease by clients who see no need to use condoms. These girls command a high price as “virgins” for brothel owners who do not want to disturb the situation by requiring condom use [29]. Further, the weakening of traditional socio-cultural constraints in times of conflict makes the girls more prone to sexual abuse and exploitation and more likely to engage in high-risk sexual behaviour.
The Local and International Response to the Epidemic
Nepal's National Center for AIDS and STD Control has received support from WHO, UNAIDS, UNDP, and USAID. But it lacks substantive funding to complete the necessary studies and interventions that are the key to HIV control. National efforts aimed at awareness of HIV are complicated by the ethnic diversity of Nepal, where some 75 different ethnic groups exist, speaking more than 50 languages.
A network of about 1,600 NGOs is now working on HIV/AIDS in Nepal. Local organisations are doing their best to reach out to those living with HIV/AIDS, but due to the urban-centric nature of most funding, funds available for the rural areas are scarce. Most donor representatives lack direct knowledge of the rural areas and rely on the instructions provided by NGOs in Kathmandu to dispense funds.
Foreign aid, which accounts for nearly 60% of Nepal's development budget, may have paradoxically contributed to lopsided development in Nepal. While aid money has favoured urban development and centralized power, the rural–urban gap has widened over the years. In Nepal, weak linkages between urban and rural areas, and lack of roads, communications, infrastructure, and appropriate skills among the rural poor mean that this urban bias has led to centralization of effective power on the one hand and maintenance of the economic, social, and political status quo on the other [30]. Urban biases inevitably play a deterrent role, discouraging poor patients from seeking help. The poor see very little of the aid money, since most of it is used for prevention, information, and awareness in urban centres rather than for care and support in rural areas.
In our experience, some donors consider HIV care and support to be too expensive to fund, arguing that Nepal lacks the kind of infrastructure—clinics, district hospitals, and distribution units—needed to provide effective antiretroviral treatment, and that antiretrovirals are a priority only in countries with a high prevalence of HIV/AIDS (1% or more) in the general population [31].
Harm-reduction interventions have been shown to slow the course of HIV among intravenous drug users (IDUs) in many developed countries [32], but in Nepal the concept of harm reduction is still new. Few harm-reduction programs are government-supported or integrated into mainstream service delivery. Organisations such as the Lifesaving and Lifegiving Society, a street-based NGO established in Nepal in 1991, have been providing education, counselling, and primary health care—as well as bleach, sterile water, condoms, and new needles and syringes—to IDUs to lower their risk of acquiring blood-borne diseases [33]. The prevalence of HIV infection among IDUs who were in regular contact with the program from 1991 to 1994 was low, at 1.6% [33].
However successful, these programs have not reached the border zones with India where HIV infection has risen dramatically among IDUs. This dramatic rise is not surprising since research has shown that cross-border drug-use patterns in areas of Nepal bordering India are particularly conducive to risky needle sharing [34]. Unlike IDUs in other Nepalese towns, very few of the IDUs in border towns belong to stable “injecting groups.” Sharing of contaminated injecting equipment in border towns is widespread, in part because of the makeshift arrangements in which the cross-border injecting takes place. Users often share their small amounts of money to buy drugs. Sexual intercourse with casual partners occurs, with inconsistent condom use [34]. Effective intervention would therefore require complex cross-border collaborative efforts [34].
People living with HIV/AIDS are stigmatized and face discrimination at all societal levels—in the community, at health facilities, and, most importantly, within the family [35]. In a recent survey by CARE-Nepal, almost half of those who came to the voluntary counselling and testing centre at the Doti District Hospital in Silgadhi, a conflict-affected area, during June–July 2004 tested positive for HIV, and almost all of those tested positive were widows in their twenties and thirties [35]. About 60% of them were breast-feeding their infants. These young widows faced rejection from their families, discrimination at work, and difficulty in coping with their life circumstances [35].
One important development in recent years has been within the area of condom promotion for HIV/AIDS prevention. A national research study found that 76.6% of retail outlets surveyed had never sold a condom in 2002 [36]. Population Services International began a national condom promotion program in early 2002 [37] using a three-pronged approach: (1) a national media campaign promoting condoms (Figure 1); (2) increased widespread condom availability within the private sector to destigmatize condom use; and (3) targeted condom promotion to high-risk groups such as sex workers [38]. The impact of such efforts on condom sales has been dramatic. Total sale of condoms jumped from 11.9 million units in 2002 to 23.1 million units in 2004, and sales continue to climb [39]. Creating long-term behaviour change and making condoms accessible in the private sector and affordable through subsidy to high-risk groups such as female sex workers appears to be increasing the uptake of condoms in Nepal.
Figure 1 Advertisement for “Number One” Condoms Marketed in Nepal by Population Services International
In April 2003, Nepal launched Number One condoms with support from USAID. In January 2004, Population Services International, a non-profit organization that uses social marketing to encourage condom use, launched Number One single-pack condoms for high-risk target groups.
A Window of Opportunity
There is a window of opportunity to combat HIV/AIDS in Nepal. If trends continue, AIDS will be the leading cause of death among 15–49-year-olds in the next ten years [8]. 100,000 to 200,000 young adults could be infected, and 10,000 to 15,000 annual AIDS deaths could occur in the next ten years [8]. What steps must be taken to prevent Nepal from being devastated by the infection?
First, we must gather better data on HIV seroprevalence, on sexual behaviour in the general population, and on sexual networking [40]. We have argued in this essay that the extent of the AIDS epidemic in Nepal will depend upon rates at which sexual partners are exchanged by commercial sex workers and the men who regularly visit them, as well as the proportion of the general population that has multiple and concurrent sexual partners. There are extensive migration patterns both within the country and internationally, fuelled by the recent conflict, which provide the potential for considerable sexual networking. Better knowledge of this networking is crucial for our HIV control efforts.
Second, the government must recognize and acknowledge the needs of high-risk groups—drug users, commercial sex workers, migrant workers, and men who have sex with men. Plans to create more behaviour changes are needed within these groups. The 2002–2006 HIV/AIDS strategy proposed by the government [41], which adopts a multisectoral approach focusing on prevention among vulnerable groups, on control, care and support, and on voluntary counselling, hopes to address some of these issues.
Third, information campaigns should focus on changing attitudes that create barriers to regular use of condoms. Condoms should be readily available, and research should focus on the impact of mass media on perceptions of risk, negative attitudes toward condoms, and risky behaviour [27]. A culturally specific approach to HIV prevention is needed that includes education of clients and brothel owners about condoms.
Fourth, the various players involved in addressing HIV and who are working in closely adjacent fields need to interact more and partner with each other, and there needs to be greater inclusion of civil society in campaigns to raise HIV/AIDS awareness and reduce stigma.
Finally, successful national programs that have the scale to alter the course of the epidemic in Nepal should be expanded.
Conclusion
HIV/AIDS is no longer only a health issue, it is also a development issue. Tackling the epidemic will require not only prevention and control of HIV infection among vulnerable and risk groups, but a multi-sectoral approach addressing the lack of access by risk groups to health care and education and recognition of the populations at risk [42]. People living with HIV and AIDS should be brought to the forefront in the fight against HIV/AIDS [35]. Family members, local communities, civil society organizations, donors, and the government all have their own important role to play [35]. The status of women must change so that they are considered autonomous individuals who can make their own decisions [29].
As with all international declarations on HIV/AIDS, there is an absolute need to take a strong human rights approach to combating the epidemic. This approach includes recognizing fundamental rights such as access to health care and information, addressing gender equity, and a concerted effort to reduce sex trafficking. It further requires addressing the root causes of poverty and inequality, which give rise to the phenomenon of migration and trafficking as well as propagate violent uprisings. Only such efforts will prevail in mitigating the effects of both HIV and conflict in Nepal.
Citation: Singh S, Mills E, Honeyman S, Suvedi BK, Pant NP (2005) HIV in Nepal: Is the violent conflict fuelling the epidemic? PLoS Med 2(8): e216.
Abbreviations
IDUintravenous drug user
NGOnon-governmental organization
==== Refs
References
Bharadwaj A Biswas R Shetty KJ HIV in Nepal: Is it rare or the tip of an iceberg? Trop Doct 2001 31 211 213 11676055
Nagelkerke NJ Jha P de Vlas SJ Korenromp EL Moses S Modelling HIV/AIDS epidemics in Botswana and India: Impact of interventions to prevent transmission Bull World Health Organ 2002 80 89 96 11953786
American Foundation for AIDS Research HIV AIDS in Asia 2004 July 8 Available: http://www.amfar.org/cgi-bin/iowa/asia/aids/?record=4 . Accessed 13 June 2005
Seddon D HIV-AIDS in Nepal: The coming crisis Bull Concern Asian Sch 1998 30 35 45 12295918
The World Bank Group Issue Brief—South Asian Region Nepal 2004 June Available: http://siteresources.worldbank.org/INTNEPAL/Resources/Nepal_Final_AIDS_Brief.pdf . Accessed 14 December 2004
Chandler A Pahoa VK Evidence for HIV-2 infection in Nepal Nepal Med Coll J 2004 6 60 61 15449658
Ministry of Health, NCASC (National Center for AIDS and STD Control) Cumulative HIV/AIDS situation of Nepal 2005 February
UNDP YouandAids The HIV portal for South Asia—Nepal at a glance 2005 Available: http://www.youandaids.org/Asia%20Pacific%20at%20a%20Glance/Nepal/ . Accessed 14 December 2004
Karki BB Rapid assessment among drug users in Nepal Aids Watch 2000 5 3 5
New Era, STD Aids Counseling and Training Service Behavioral and sero prevalence survey among IDUs in Kathmandu. Kathmandu: Family Health International 2002 December Available: http://www.un.org.np/uploads/reports/UNAIDS/2002/behavioral_sero_ktm.pdf . Accessed 14 June 2005
Oelrichs RB Shrestha IL Anderson DA Deacon NJ The explosive human immunodeficiency virus type 1 epidemic among injecting drug users of Kathmandu, Nepal, is caused by a subtype C virus of restricted genetic diversity J Virol 2000 74 1149 1157 10627525
Centre for Harm Reduction Nepal 2003 Available: http://www.chr.asn.au/projects/southasia/nepal . Accessed 13 December 2004
Family Health International STD and HIV prevalence survey among female sex workers and truckers on highway routes in the Terai 2000 Available: http://www.fhi.org/en/HIVAIDS/pub/survreports/nepalstdprev.htm . Accessed 14 December 2004
Family Health International Behavioral surveillance survey in the highway route of Nepal: Round 4, May 2002 2002 Available: http://www.fhi.org/en/HIVAIDS/pub/survreports/sexworkersnepal.htm . Accessed 15 December 2004
Family Health International Behavioral surveillance survey in the highway route of Nepal: Round 1, May 1999 1999 Available: http://www.fhi.org/en/HIVAIDS/pub/survreports/bssnepalround1.htm . Accessed 31 March 2005
British Broadcasting Corporation Human Rights Watch Nepal gay plea 2005 August 12 Available: http://news.bbc.co.uk/2/hi/south_asia/3557794.stm . Accessed 15 March 2005
Mock NB Duale S Brown LF Mathys E Omaonaigh HC Conflict and HIV: A framework for risk assessment to prevent HIV in conflict-affected settings in Africa Emerg Themes Epidemiol 2004 1 6 Available: http://www.ete-online.com/content/1/1/6 . Accessed 9 March 2005 15679919
Dhakal P Insurgency contributing to propagation of HIV/AIDS 2004 November 30 Kathmandu Nepal News Available: http://www.nepalnews.com.np/archive/2004/nov/nov30/nov_30_news15.php . Accessed 14 June 2005
Irin News Network Reuters News alert Nepal: Focus on the plight of rural people living with HIV/AIDS 2005 January 24 Available: http://www.irinnews.org/report.asp?ReportID=45194&SelectRegion=Central_Asia . Accessed 13 June 2005
Global IDP Project, Norwegian Refugee Council Nepal. Up to 200,000 people displaced by fighting remain largely unassisted 2004 September 8 Available: http://www.db.idpproject.org/Sites/idpSurvey.nsf/wCountries/Nepal . Accessed 13 December 2004
Poudel KC Jimba M Okumura J Joshi AB Wakai S Migrants' risky sexual behaviours in India and at home in far western Nepal Trop Med Int Health 2004 9 897 903 15303995
Poudel KC Jimba M Okumura J Sharma M Poudel Tandukar K Migration in far western Nepal: A time bomb for a future HIV/AIDS epidemic Trop Doct 2004 34 30 31 14959972
Family Health International STD prevalence study among women in migrant communities of Kailali district, Nepal 2001 Available: http://www.fhi.org/en/HIVAIDS/pub/survreports/migrantwomen.htm . Accessed 14 December 2004
The Rising Nepal Girl traffickers find new fertile ground in Rukum, Rolpa 2004 Available: http://www.gorkhapatra.org.np/pageloader.php?file=2004/03/29/topstories/main8 . Accessed 12 January 2005
Rijal S Insurgency fuelling flesh trade The Kathmandu Post. Available 2005 March 29 http://www.kantipuronline.com/kolnews.php?&nid=35647 . Accessed 14 June 2005
Lawday A HIV and conflict: A double emergency 2002 London Save the Children UK Available: http://www.savethechildren.org.uk/temp/scuk/cache/cmsattach/212_hivconflict.pdf . Accessed 14 June 2005
Tamang A Nepal B Puri M Shrestha D Sexual behaviour and risk perception among young men in border towns of Nepal Asia Pac Popul J 2001 16 195 210 Available: http://www.un.org.np/uploads/reports/UNAIDS/2001/sexualbehavior.pdf . Accessed 13 June 2005
Tuladhar SM Ross J Basnyat A Mills S Acharya L HIV behavioral surveillance survey (BSS) of female sex workers (FSWs) and male subpopulation groups in the highway groups of Nepal—A trend analysis [abstract] Med Gen Med 2004 6 The XV International AIDS Conference 2004 ThPeD7643 Available: http://www.iasociety.org/ejias/show.asp?abstract_id=2174311 . Accessed 15 June 2005
Reinfeld MR Nepal: A cultural prostitution Child Worldw 1993 20 45 46 12179312
Francis MG Nepal's woes—Instability, inequality, insurgency and the IMF-WB 2002 October 18 New Delhi International Development Economics Associates Available: http://www.networkideas.org/themes/inequality/oct2002/ie18_Nepal.htm . Accessed 17 June 2005
Newar N PLWHA left in lurch by donor agencies 2005 Available: http://www.youandaids.org/Features/PanosNepal5Oct.asp . Accessed 13 June 2005
Crofts N Costigan G Narayanan P Gray J Dorabjee J Harm reduction in Asia: A successful response to hidden epidemics. The Asian Harm Reduction Network AIDS 1998 12 S109 115
Singh M Harm reduction and street-based program: Looking into Nepal Subst Use Misuse 1998 33 1069 1074 9596376
Nepal B Pelto PJ Acharya LB Tamang A Shrestha M Factors behind the high prevalence of HIV among drug injectors in a border town of Nepal [abstract] 2004 APHA meeting 2004 November 6–10 Washington, D.C. Available: http://apha.confex.com/apha/132am/techprogram/paper_88558.htm . Accessed 13 June 2005
CARE-Nepal A pilot initiative on understanding people living with HIV/AIDS and exploring the ways forward for their support in Nepal 2004 December 27 Available: http://www.carenepal.org/Care_nepal_Library/our_Projects_current_HIV-AIDS.htm . Accessed 13 June 2005
Pant PD KC K Das AKL Adhikary KN Condom distribution market study 2002 October Research report presented and distributed at USAID public stakeholders meeting 200 November 21 Kathmandu, Nepal. Kathmandu: PSI Nepal 95
Chapman S Evaluating health promotion: Practice and methods 2004 New York Oxford University Press 186
Nielsen AC Media recall study for communication and advocacy campaign 2003 January Kathmandu Population Services International Available: http://www.stratecommhost.com/aidsmark/web/files/PSI1_advc_fin2.doc . Accessed 15 June 2005
Honeyman SW Bajracharya D Valentine D Charman N “Dashboard marketing” of condoms for HIV/AIDS prevention produces exceptional results in Nepal [abstract] 2005 Seventh International Congress on AIDS in Asia and the Pacific (ICAAP) 2005 July Kobe, Japan . Abstract #1916
Furber AS Newell JN Lubben MM A systematic review of current knowledge of HIV epidemiology and of sexual behaviour in Nepal Trop Med Int Health 2002 7 140 148 11841704
Family Planning Association of Nepal HIV/AIDS strategy 2002–2006 2004 Available: http://www.fpan.org/demo/hivaids%20strategy.html . Accessed 14 December 2004
Smith-Estelle A Gruskin S Vulnerability to HIV/STIs among rural women from migrant communities in Nepal: A health and human rights framework Reprod Health Matters 2003 11 142 151 14708405
Singh S Impact of long-term political conflict on population health in Nepal CMAJ 2004 171 1499 1501 Available: http://www.cmaj.ca/cgi/content/full/171/12/1499 . Accessed 14 June 2005 15583204
| 16008505 | PMC1176235 | CC BY | 2021-01-05 11:13:39 | no | PLoS Med. 2005 Aug 19; 2(8):e216 | utf-8 | PLoS Med | 2,005 | 10.1371/journal.pmed.0020216 | oa_comm |
==== Front
PLoS MedPLoS MedpbioplosbiolPLoS Medicine1549-12771549-1676Public Library of Science San Francisco, USA 1600850610.1371/journal.pmed.0020227Health in ActionHealth EconomicsHealth PolicyHIV/AIDSHIV Infection/AIDSHealth EconomicsHealth PolicyInternational healthResource allocation and rationingThe “Free by 5” Campaign for Universal, Free Antiretroviral Therapy Health in ActionWhiteside Alan [email protected] Sabrina Alan Whiteside is the Director of the Health Economics and HIV/AIDS Research Division (HEARD) at the University of KwaZulu-Natal in Durban (http://www.heard.org.za). Sabrina Lee is a graduate of the Development Studies Masters programme at the University of KwaZulu-Natal and a research assistant at HEARD.
Competing Interests: The authors declare that they have no competing interests. AW's time in preparing this article was supported by a British Department for International Development Knowledge Programme. The views expressed are the authors' alone.
8 2005 19 7 2005 2 8 e227Copyright: © 2005 Whiteside and Lee.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 Free by 5 declaration, launched in November 2004, aims to achieve free access for all individuals with HIV to a comprehensive medical package including antiretroviral treatment.
User fees pose a significant barrier to achievement of the “3 by 5” strategy
==== Body
The Free by 5 declaration (http://www.ukzn.ac.za/heard/freeby5/freeby5.htm), launched in November 2004, is a campaign to achieve free access for all individuals with HIV to a comprehensive minimum medical package including antiretroviral treatment (ART). The declaration was developed in response to the World Health Organization (WHO) and the Joint United Nations Programme on HIV/AIDS (UNAIDS) “3 by 5” strategy, which aims to scale up access to ART to ensure that 3 million people have access to ART by the end of 2005 [1]. We believe that the declaration made an important contribution to the debate on provision of ART in resource-poor settings.
A number of debates still surround the 3 by 5 strategy, including how it will be operationalised and funded [1]. Questions remain over how the poorest and most vulnerable groups can be reached and how a suitable level of adherence to the drugs can be achieved to avoid drug resistance [2].
The Problem with User Fees
Currently, fewer than 8% of people living with HIV/AIDS in Africa are on ART [3]. While user fees represent only one of the barriers to access to essential ART, the Free by 5 initiative firmly states that unless treatment is provided free of charge to all in developing countries, 3 by 5 will struggle to meet its ambitious target. Many countries still impose user fees for ART and associated tests, and as AIDS is an impoverishing disease, this means treatment is unaffordable for most [4].
We must state clearly that we acknowledge that providing treatment free of charge is not the only precondition for the scale-up of treatment programmes. In most developing countries, the availability and efficiency of health infrastructure is the dominant obstacle to effective health care. Adequate primary health-care infrastructure and staff is fundamental to the provision of treatment programmes. However, we argue that where drugs and services are administered, providing treatment for free would assist patients to gain greater access to, remain adherent to, and avoid instability in treatment regimens. Simply stated, the Free by 5 campaign maintains that user fees are an additional and unnecessary obstacle to treatment access and the efficiency and equity of treatment programmes in the context of this major health crisis. Furthermore, removal of patient fees as a significant barrier to access is realistic and feasible in resource-poor settings
Antiretroviral drugs are unaffordable to most patients in Africa (Photo: Copyright WHO/Eric Miller)]
This article outlines the Free by 5 campaign and its objectives. It sketches the arguments for free treatment and describes how the initiative influenced the debate surrounding HIV/AIDS treatment among donors, international institutions, non-government organisations, professionals in the fields of public health and economics, and the concerned public.
Origins and Objectives of Free by 5
The Free by 5 initiative began as a challenge to one of us (Alan Whiteside) at a Commission on HIV/AIDS and Governance in Africa meeting in Maputo from Veronique Collard, an academic and activist, who asked, “What do economists believe?” This challenge led to the idea that a declaration, giving the economists' position, could be adopted by the International AIDS Economics Network meeting in Bangkok. A first draft was prepared and presented at this meeting, but the reception from the assembled economists was underwhelming.
Despite this, we decided to press ahead. Veronique gave freely of her time in developing the initiative and was guided by many people, particularly Bernard Taverne of Institut de Recherche pour le Développement, Senegal; Alice Desclaux of University of Paul Cezanne, France; Gorik Ooms of Médecins Sans Frontières (MSF) in Belgium; and Alan Whiteside. We at HEARD contributed to the costs involved and provided a secretariat. Countless others in the field gave their support and enthusiasm, without which the initiative could not have happened.
The overall goal of the initiative was to gather support from professionals and organisations to promote universal free access to a minimum health-care package, including ART, for people with HIV. It aimed to lobby international institutions, including UNAIDS, WHO, and donors— such as the Global Fund, the World Bank, and the US President's Emergency Plan for AIDS Relief—to adopt guidelines and actively promote the principle and implementation of free treatment. Donors were urged to pledge additional resources to ART through long-term commitments. Finally, the initiative sought to provide economic and public health evidence to inform the decisions of policy makers and governments and assist activists in their advocacy efforts.
What Do WHO and 3 by 5 Say about Free Treatment?
The Free by 5 declaration makes the point that the 3 by 5 strategy is unrealistic. Although the WHO seeks increased access to ART for all people living with HIV/AIDS, it does not address the costs at the patient level. The Free by 5 campaign believes that free treatment is an absolute prerequisite to the scale-up of treatment programmes and universal access to treatment.
User fees inhibit patient adherence.
Despite clear indications that patient fees inhibit access to treatment programmes, as outlined in the following section, the WHO strategy documents do not address this issue. Instead, the WHO treatment guidelines make frequent references to “affordability.” The WHO strategy published in 2003 recommended making antiretrovirals affordable and providing them free of charge to the poor [5]. A revision of the guidelines recommends providing “medication free of charge to those who can least afford treatment through subsidized or other financing strategies” [6]. A 2004 Consultation Report stated that if “cost recovery schemes prove inefficient or obstructive to access, free delivery to all should be considered” [7]. It should be noted that much of this debate is confined to ART. In the Free by 5 declaration, we recognise that treatment includes testing, laboratory examinations, and associated drugs.
Arguments for and against Free Treatment
What affordability means and who is poor is not defined in these WHO guidelines. Defining the poor by income level is problematic in countries where the informal economy (that is, unregistered, unrecognised, and unsupported employment) dominates and income records are poor.
The process of implementing exemptions based on income is a waste of scarce financial and human resources as systems are costly to put in place and administer. Exemptions or waivers rarely reach those who are eligible to receive them [8]. While the WHO maintains that “free treatment would be difficult to implement in many health systems” [9], the Free by 5 declaration states that it will be easier and more cost-effective to provide treatment to all patients free of charge.
Countries set their own criteria for access, and these vary. In addition, perceptions of equity vary among and between governments, donors, and activists. These variations are difficult to manage from a clinical perspective and prevent equity from being attained at a national and international level [10]. Existing criteria for access are inequitable: a first-come, first-served basis favours the rich, more educated, and urban people. Universal free treatment is necessary to achieve equity in access and to avoid exclusion of the most susceptible and vulnerable groups.
The Free by 5 declaration details a number of arguments that have been made against free treatment and gives evidence that counter these views.
There are claims that patients should pay in order to give value to treatment and remain adherent to the drugs. Studies in Senegal have shown that user fees inhibit patient adherence and cause frequent interruptions in therapy [11], and in Kenya user fees have led to the discontinuation of treatment and delays in health-seeking behaviour [12]. The negative relationship between end-user costs and adherence has also been echoed in data from Uganda [13], Nigeria [14], Botswana [15], and the Côte d'Ivoire [16]. When ART must be paid for, patients are also more likely to misuse drugs and purchase them on the informal market. This ultimately leads to drug resistance. In some cases, the costs of laboratory tests deter people from joining treatment programmes. Providing treatment for free reduces delays in seeking care and improves adherence and may influence the quality of care.
A major argument against free treatment is that of AIDS exceptionalism. In other words, why should AIDS be treated for free when others diseases are not? There are three simple arguments countering this view.
First, AIDS is the worst epidemic humanity has ever faced, which has devastating long-term social, economic, and personal impacts and is a major obstacle to development. The exceptionalism of the disease requires exceptionalism in the response. At the UN General Assembly meeting on HIV/AIDS in New York, on 22 September 2003, WHO Director General Jong-Wook Lee described the lack of ART as a global health emergency. Second, other diseases are treated free where there is a public health reason to do so. Third, it is feasible to implement free HIV/AIDS treatment in resource-poor settings. Given the nature of the AIDS epidemic, providing free treatment should be an imperative even if it can not be applied to all diseases or all in need.
It is argued that patient fees are necessary to ensure the sustainability of treatment programmes. However, in Senegal fees amount to little more than 10% of the cost of drugs [11]. Patient contributions do not cover other costs such as staff, training, and social services. Sustainability can be achieved only through long-term commitments from donors and governments. The WHO and UNAIDS estimate that the total cost of providing treatment through the 3 by 5 initiative for 2005 is $3.8 billion, and this will increase to $6.7 billion in 2007 [17]. The contribution made by fee-paying patients is negligible. In Ghana, for instance, user fees amount to no more than a tiny fraction of the Ministry of Health aggregate budget [18]. Therefore, providing treatment for free will have virtually no effect on global resource needs. Significant resources still need to be mobilised.
AIDS is the worst epidemic humanity has ever faced.
The Impact of Free by 5
The Free by 5 declaration, which is available in French, English, and Spanish, was disseminated worldwide through global MSF offices, universities, schools of public health, and NGO networks. It was signed by more than 600 people, many of whom are respected public health professionals, economists, policy makers, and key activists. Some of the influential signatories are shown in Box 1. The declaration was also signed by a number of organisations, shown in Box 2.
Box 1. Some of the Influential Signatories of Free by 5
Stephen Lewis, UN Special Envoy on HIV/AIDS
Helene Rossier-Blavier, Director General of AIDES France and Vice-President of the Global Fund
Hoosen Coovadia, Victor Daitz Chair in HIV/AIDS Research, University of KwaZulu-Natal
Philippe Douste Blazy, French Minister of Health
Bernard Kouchner, Founder of Médecins Sans Frontières, former French Health Minister
Omar Kabbaj, President of the African Development Bank, Tunis
Max Essex, Chair, Harvard School Public Health AIDS Initiative
Box 2. Some of the Organizations That Signed the Free by 5 Declaration
Norwegian Council for Africa
International HIV/AIDS Alliance
Treatment Action Campaign
National Agency for AIDS Research (ANRS) France
Réseau Accès aux Médicaments Essentiels (RAME), Burkina Faso
Fondation Femme Plus, DRC
The initiative sparked extensive debate among Internet-based development and public health fora. It culminated in a media release, which was disseminated widely among the global press. The initiative was picked up by British, French, South African, and Kenyan national newspapers as well as the UN IRIN Plus News. It was also featured on a number of Web sites.
It is difficult to asses the impact of the declaration over such a short time frame. After the media launch, the declaration and list of signatures were sent to UNAIDS, WHO, and the World Bank. We urged these organisations to give unambiguous support to the implementation of free treatment and take a lead in raising awareness about the issue. We encouraged all governmental and nongovernmental actors to adopt universal free treatment and actively promote its implementation. We asked that the issue be included on the agenda of technical meetings and political forums planned in the framework of the 3 by 5 initiative, and reflected clearly in all WHO/UNAIDS guidelines.
At the very least, the Free by 5 declaration publicised the issue and the importance of universal free access to treatment and created a debate surrounding free treatment in public health, health economics, and human rights circles throughout the world. It is an issue that has been placed on the international agenda for discussion.
For instance, the Free by 5 campaign played a role in inspiring the recent WHO/UNAIDS/World Bank meeting entitled “Ensuring universal access: User fees and free care polices in the context of HIV treatment” (21–23 March 2005). The meeting acknowledged that “for many individuals in poor countries, affordability poses an insurmountable obstacle” that depresses uptake of AIDS treatment programmes and decreases adherence of those enrolled. The meeting also noted that user fees contribute very little to overall sustainability and that if AIDS care and treatment programmes are to be scaled up, a broad shift to other financing models is required [19].
Conclusion
There are many lessons from this initiative, but two are particularly important. The first is that a good idea can, with the right support, be turned into something concrete. The second is that goodwill and good sense are as important as money in shaping the policy environment. It was a small initiative with a big impact. We hope that eventually we will see the fruits of our efforts: universal free access to ART to all who need them. We feel privileged to have been part of this important campaign.
Citation: Whiteside A, Lee S (2005) The “Free by 5” campaign for universal, free antiretroviral therapy. PLoS Med 2(8): e227.
Abbreviations
ARTantiretroviral treatment
UNAIDSJoint United Nations Programme on HIV/AIDS
WHOWorld Health Organization
==== Refs
References
Kim JY Ammann A Is the “3 by 5” initiative the best approach to tackling the HIV pandemic? PLoS Med 2004 1 e37 15578107
Holmes W 3 by 5, but at what cost? Lancet 2004 363 1072 1073 15065565
WHO/UNAIDS 3 by 5 progress report, December 2004 2004 Geneva World Health Organization
Yazdanpanah Y Costs associated with combination antiretroviral therapy in HIV-infected patients J Antimicrob Chemother 2004 53 558 561 14985277
WHO Treating three million by 2005: Making it happen, WHO strategy 2003 Geneva World Health Organization
WHO Scaling up ARV therapy in resource-limited settings: Treatment guidelines for a public health approach, WHO strategy 2003 Geneva World Health Organization
WHO/UNAIDS Consultation on ethics and equitable access to treatment and care for HIV/AIDS 2004 Geneva World Health Organization
Desclaux A Castro A Singer M Equity in access to AIDS treatment in Africa: Pitfalls among achievements Unhealthy health policy: A critical anthropological examination 2004 Walnut Creek (California) Altamira Press 387
WHO/UNAIDS Emergency scale-up of anti-retroviral therapy in resource-limited settings: Technical and operational recommendations to achieve three by five 2003 Report of the WHO/UNAIDS meeting 18 November 2003 Lusaka, Zambia
Furber AS Hodgson IJ Desclaux A Mukasa DS Barriers to care for people with AIDS in developing countries BMJ 2004 329 1281 1283 15564262
Laniece I Ciss M Desclaux A Diop K Mbodj F Adherence to HAART and its principal determinants in a cohort of Senegalese adults AIDS 2003 17 S103 S108
African Women and Children Feature Services AIDS patients quitting treatment 2004 September 30 Nairobi African Women and Children Feature Services
Byakika-Tusiime J Oyugi JH Tumwikikirize WA Katabira ET Mugyenyi PN Adherence to HIV antiretroviral therapy in HIV+ Ugandan patients purchasing therapy Int J STD AIDS 2005 16 38 41 15705271
Daniel OJ Ogun SA Odusoga OL Falola RL Ogundahunsi OA Adherence pattern to ARV drugs among AIDS patients on self-purchased drugs and those on free medications in Sagamu, Nigeria [abstract] 2004 XV International AIDS Conference 2004 July 11–16 Bangkok
Weiser S Wolfe W Bangsberg D Thior I Gilbert P Barriers to anti-retroviral adherence for patients living with HIV infection and AIDS in Botswana J Acquir Immune Defic Syndr 2003 34 281 288 14600572
Laguide R Elengua N Fassinou P Moatti JP Coriat B Souteyrand Y Barnett T Dumoulin J Direct costs of medical care for HIV-infected children before and during HAART in Abidjan, Cote D'Ivoire, 2000–2002 Economics of AIDS and access to HIV/AIDS care in developing countries. Issues and challenges 2003 311 327
UNAIDS Financing the expanded response to AIDS 2004 Geneva UNAIDS 11
McGreevey W Borda M Stover J Public financing for HIV care and treatment: What will it cost in six countries, draft 21; March 2005 2005 Washington, D.C Futures Group
WHO/UNAIDS Ensuring universal access: Free HIV care and treatment at point of service (background and proposed principles), draft WHO/UNAIDS policy position; March 2005 2005 Geneva UNAIDS
| 16008506 | PMC1176236 | CC BY | 2021-01-05 11:13:39 | no | PLoS Med. 2005 Aug 19; 2(8):e227 | utf-8 | PLoS Med | 2,005 | 10.1371/journal.pmed.0020227 | oa_comm |
==== Front
PLoS MedPLoS MedpbioplosbiolPLoS Medicine1549-12771549-1676Public Library of Science San Francisco, USA 1600850710.1371/journal.pmed.0020234EssayEpidemiology/Public HealthHIV/AIDSHIV Infection/AIDSInfectious DiseasesSexual HealthSexually transmitted diseasesHealth education (including prevention and promotion)The Abandoned Trials of Pre-Exposure Prophylaxis for HIV: What Went Wrong? EssaySingh Jerome A Mills Edward J [email protected] A. Singh is at the Centre for the AIDS Programme of Research in South Africa and the Howard College School of Law, University of KwaZulu–Natal, Durban, South Africa, and the University of Toronto Joint Centre for Bioethics, Toronto, Ontario, Canada. Edward J. Mills is at the Centre for International Human Rights Law, University of Oxford, United Kingdom, and the Department of Clinical Epidemiology and Biostatistics, McMaster University, Hamilton, Ontario, Canada.
Competing Interests: The authors declare that they have no competing interests.
9 2005 19 7 2005 2 9 e234Copyright: © 2005 Singh and Mills.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 Must Not Let Protestors Derail Trials of Pre-Exposure Prophylaxis for HIV
The dramatic protests that shut down trials of tenofovir as pre-exposure prophylaxis against HIV may prove damaging in our fight against the HIV pandemic.
==== Body
New approaches to HIV/AIDS prevention are urgently needed to stem the estimated 5 million new infections that occur worldwide each year. One such promising, novel intervention has been the proposed use of the oral antiretroviral drug tenofovir (Viread) as a pre-exposure prophylaxis (PREP) in high-risk groups (for example, uninfected women who have high-risk commercial sex). However, emerging opposition has halted the progress of at least two important clinical trials of tenofovir as PREP and brought negative attention to tenofovir, somewhat similar to that visited on thalidomide more than four decades ago. This could prove damaging in the long term.
If tenofovir is someday proven to be clinically efficacious as a PREP, today's irresponsible reporting and activism surrounding tenofovir could cause those in need to snub the drug if, or when, it becomes licensed for use as a PREP. This unfortunate prospect raises questions about responsible media reporting, responsible conduct on the part of investigators and activists, and what should be done to avert or repair damaging trial-related disputes in the future.
Protests against Trials of PREP
In July 2004, increasing pressure from activist groups and affiliated non-governmental organizations persuaded the Cambodian Prime Minister to halt the initiation of a PREP trial of tenofovir among Cambodian commercial sex workers [1], a trial funded by the United States National Institutes of Health and the Bill and Melinda Gates Foundation. The Cambodian Ministry of Health has provided no official statement for its decision to halt the trial.
The dramatic protest against the Cambodian trial at the XV International AIDS Conference in Bangkok, Thailand, caught the world's media attention [2] and brought tenofovir to the forefront of the public's attention (Figures 1 and 2). The primary reasons cited for the demonstrations included alleged inadequate prevention counseling by the study investigators, a lack of pre- and post-test HIV counseling, and the nonprovision of medical services and insurance for those who seroconverted during the study or experienced adverse events related to the trial drug [3]. Participant activist groups also argue that the safety of tenofovir for long-term use by individuals who are HIV negative has not been established, and that there was limited involvement of the target communities in the trial design [1,3,4]. The activist groups representing the participants argue that the participants take all of the risks and get little of the benefits.
Figure 1 The Dramatic Protest at the Gilead Booth at the International AIDS Conference 2004 Caught the World's Attention
(Photo: Act Up–Paris)
Figure 2 The Protest Closed the Gilead Booth at the International AIDS Conference 2004
(Photo: Act Up–Paris)
More recently, in February 2005, a similar trial in Cameroon, led by Family Health International (FHI), was halted by the Minister of Public Health. The activist group Act Up–Paris (www.actupparis.org) has led protests against both the Cambodian and Cameroonian trials, and has attracted French media attention. Act Up–Paris is supportive of tenofovir trials in general but has strong concerns about the current trials. In a documentary aired on French TV-2 and subsequent media reports, activists alleged that the FHI investigators intentionally allowed participants to become infected and provided inadequate counseling by having only five counselors for 400 participants [5].
Protest should be carried out in a responsible manner.
An independent inquiry, commissioned by the Cameroon Ministry of Public Health, reported on February 23, 2005, that they now require more regular reporting and a formal study site accreditation for a satellite hospital clinic [6]. The committee also confirmed that, contrary to popular belief and widespread reports, participants in the trial were not injected with HIV, and the study's tablets did not contain HIV. The inquiry recommended that the clinical trials could resume if the sponsors rectified certain administrative issues that the commission identified. The enrolled participants are being followed up, but neither tenofovir nor placebo is being dispensed.
On March 11, 2005, FHI made the announcement that the Nigerian arm of the tenofovir PREP trial will discontinue prematurely. FHI closed the trial voluntarily, because it determined that the study team was unable to comply with the required operational and laboratory procedures at the level necessary for conducting this study. The ability to meet these standards is critical for ensuring the safety of participants and the quality of the data from the study. This decision was made in conjunction with FHI's external, independent Data and Safety Monitoring Committee. More than 100 participants had been randomized. The announcement came as a disconcerting blow to the already fragile network of trials.
The Protestors' Concerns
Activists and ethicists argue about the contentious issue of the standard of care in randomized trials [7]. According to the FHI protocols, participants who seroconverted during the trial would be provided with state-of-the-art antiretroviral therapy, if indicated (according to the World Health Organization's criteria for AIDS), with the possibility of extending treatment after the trial ended. Activist groups argue that treatment should be initiated in the same manner as would be provided in developed countries. Gilead, the maker of tenofovir, has promised to provide the drug at cost to the participating countries.
Guideline 29 of the Helsinki Declaration states that interventions should be tested against the best prophylactic interventions available [8]. In participants with a high risk of infection through sexual behaviour, this entails the provision of safe-sex education and condoms. The trial sites appear to have provided adequate counseling and male condom provision, but should also ensure female condom provision. Activists argue that if the primary endpoint is infection, counseling on safe sexual behaviour reduces the likelihood of finding an effect. They allege that investigators have a conflict of interest between meeting standards of human rights and obtaining scientific data.
In response, investigators claim that the tenofovir PREP trials were developed collaboratively with the host countries to meet relevant ethical standards. In West Africa, formative research studies with the community of participants helped to design the informed-consent instrument, to identify the preferred sites of receiving health care, and to identify sources of stigma, which the investigators tried to reduce.
Act Up–Paris and other activist groups report that they plan to continue protests against other tenofovir trials taking place elsewhere in the world (Table 1). However, while freedom of expression is a cherished ideal, we believe that protest should be carried out in a responsible manner. Important risks exist in all clinical trials, and the protection of research participants is of utmost concern. Speculation, unwarranted criticism, overreaction, or sensationalizing facts risk stigmatizing tenofovir and could jeopardize future attempts to find an efficacious PREP. This is in nobody's interest.
Table 1 Current Tenofovir PREP Studies
NIH, National Institutes of Health.
Ongoing Threats to PREP Trials
As of March 2005, there are at least six ongoing or planned human clinical trials of tenofovir as PREP (Table 1). The repercussions of the aforementioned trial closures will conceivably influence other current or proposed tenofovir PREP trials and could also detract from the use of tenofovir as a treatment agent. Act Up–Paris contends that the US-based trial will only examine safety, while the trials in developing countries examine efficacy.
Recent protests by the Thai Drug Users Network and other Thai AIDS advocacy groups, opposing the recent approval of a tenofovir prophylaxis trial funded by the US Centers for Disease Control and Prevention (CDC) among intravenous drug users (IDUs), should be setting off alarm bells for the trial investigators concerned. Thai activists cite ethical flaws in the trial design and a lack of community involvement on the part of the trialists. The activists argue that withholding the provision of clean injection equipment is an ethical violation, saying that clean equipment is a standard prevention tool akin to condoms, which are offered to trial participants in other countries who are chosen for their high-risk sexual behaviours.
But according to a CDC fact sheet, participating IDUs will be offered follow-up in a methadone drug-treatment program, and will receive bleach and instructions on how to use it to clean needles [9]. Consistent with Thai government policy and US national policy, sterile syringes will not be provided, although they are widely available in Thailand without a prescription and at low cost. Thailand has no harm-reduction policies for IDUs, and the government's war on drugs has been widely blamed for widespread human rights abuses against IDUs [10].
There Is No Time to Waste
HIV/AIDS advocacy and other activist groups have openly criticized those involved in protesting these trials for what was seen as callous behaviour on their part. Most investigators and advocates for patients with HIV/AIDS laud the past work of activist groups in attracting the world's attention to the HIV epidemic. Indeed, activism can undoubtedly play an important role in ensuring that researchers and sponsors maintain ethical standards. However, activism should be based on informed opinion and communication.
Activism should be based on informed opinion.
In May 2005, the Bill and Melinda Gates Foundation brought together activist, advocacy, and research groups to discuss future tenofovir trials. In an effort to engage stakeholders, the meeting sought resolution and clarifications to the standards of care and prophylaxis in planned provision. The meeting identified the many rumors and miscommunications that had existed in the reporting of the closed trials. It exemplified the necessity for early interventions to promote communication, in order to prevent partisan behaviour among stakeholders.
The rapidly collapsing tenofovir trial network shows that a lack of communication between activists, participants, and researchers can lead to suspicion, speculation, and, ultimately, damaging outcomes. While clinical trial investigators have increasingly begun to involve stakeholder groups in medical decision-making and trial planning, this gesture must go beyond mere tokenism. Investigators should engage in pre-trial “preventative diplomacy”. This celebrated dispute resolution mechanism is often used in political contexts but could find useful application in the research arena, too. While anything intended to keep a conflict from worsening might be described as “preventative”, preventative diplomacy involves “proaction” rather than reaction and emphasizes that crises can be better addressed before or as they emerge rather than when they have already deepened and widened.
Instruments that may be used to prevent the emergence or escalation of disputes culminating in trial suspension include the following: (1) the establishment of early warning mechanisms (such as a community liaison officer), (2) fact finding missions (these may establish that the operational reality does not resonate with protocol schedule), (3) confidence-building measures (such as the inclusion of activist groups in community advisory boards), (4) engaging the media, and (5) education (particularly on important issues such as therapeutic misconception, compensation for study-related injuries, and post-trial benefits). The development of a forum for identifying mutual interests and concerns, while still invoking reciprocity and transparency, may identify early concerns.
The investigators from the PREP trials report that they did involve activist and advocacy groups in designing the trials, but say that they were unsuccessful in addressing the wider activist community. Investigators should not merely encourage the involvement of activist groups in future prevention trials—they must make a genuine attempt to address their concerns. Such an approach may have averted the trial closures.
The world desperately needs efficacious HIV prevention and therapy. If tenofovir is shown to be an effective PREP agent, it will become a powerful tool in the fight against AIDS, but it will need to be delivered alongside behavioural interventions, condoms, clean needles, HIV testing, and access to HIV treatment. The ethics of the aforementioned Cambodian and Cameroonian tenofovir trials will likely remain forever contentious. The operations of the Nigerian trial remain a continuing issue, concerning how to assure research quality meets international standards in resource-poor settings. However, the important lesson to be learned from these experiences is that investigators, sponsors, participants, members of the study community, government authorities, and activist groups must actively engage at all stages of a trial to ensure that the study is conducted in a manner that is beneficial to, and respectful of, the participants, while remaining scientifically sound.
Stakeholders must rise above ideological differences and keep their eye on the ultimate goal: combating AIDS. If disputes arise, they must meaningfully commit themselves to addressing the issues expeditiously and in a manner that is conducive to ongoing dialogue and a sustainable relationship. A failure to do otherwise will frustrate attempts to combat AIDS and needlessly prolong the suffering of those in need.
The authors are indebted to Dr. Ross Upshur, Dr. Jim Lavery, Beth Robinson, Lori Heise, Dr. Kim Shafer, Prof. Bebe Loff, Mitchell Warren, Dr. Ward Cates, and Act Up–Paris. The opinions expressed in this paper are solely those of the authors. This work is based on an ongoing policy analysis of the tenofovir trial closures.
Citation: Singh JA, Mills EJ (2005) The abandoned trials of pre-exposure prophylaxis for HIV: What went wrong? PLoS Med 2(9): e234.
Abbreviations
CDCCenters for Disease Control and Prevention
FHIFamily Health International
IDUintravenous drug user
PREPpre-exposure prophylaxis
==== Refs
References
Cohen J AIDS research. Cambodian leader throws novel prevention trial into limbo Science 2004 305 1092
Chase M Key AIDS study in Cambodia now in jeopardy The Wall Street Journal 2004 August 12 1 Sect B
Ahmad K Trial of antiretroviral for HIV prevention on hold Lancet Infect Dis 2004 4 597 15470760
James JS Cambodia stops important tenofovir prevention trial AIDS Treat News 2004 July 23 4 5
Integrated Regional Information Networks Cameroon: Clinical trial of anti-HIV drug on sex workers in question 2005 Available: http://www.irinnews.org/report.asp?ReportID=45252&SelectRegion=West_Africa . Accessed 13 June 2005
Atatah C Douala AIDS drug controversy: Medical council says trials violated ethical norms The Post 2005 February 24 Available: http://www.postnewsline.com/2005/02/strongdouala_ai.html . Accessed 13 June 2005
Singh JA Standards of care in the antiretroviral rollout world Lancet 2004 364 920 922 15364176
World Medical Association World Medical Association Declaration of Helsinki: Ethical principles for medical research involving human subjects 1964 Available: http://www.wma.net/e/policy/b3.htm . Accessed 13 June 2005
Centers for Disease Control and Prevention CDC trials of daily oral tenofovir for preventing HIV infection Phase II and III clinical trials in Botswana, Thailand, and the United States 2004 Available: http://www.cdc.gov/hiv/pubs/TenofovirFactSheet.pdf . Accessed 13 June 2005
Human Rights Watch Thailand: Anti-drug campaign reaches new low 2000 Available: http://hrw.org/english/docs/2004/10/05/thaila9445.htm . Accessed 13 June 2005
| 16008507 | PMC1176237 | CC BY | 2021-01-05 10:40:31 | no | PLoS Med. 2005 Sep 19; 2(9):e234 | utf-8 | PLoS Med | 2,005 | 10.1371/journal.pmed.0020234 | oa_comm |
==== Front
PLoS MedPLoS MedpmedplosmedPLoS Medicine1549-12771549-1676Public Library of Science San Francisco, USA 10.1371/journal.pmed.0020242SynopsisBioengineeringBiotechnologyInfectious DiseasesHIV/AIDSTowards a Cheap and Easy Way to Monitor HIV/AIDS Synopsis7 2005 19 7 2005 2 7 e242Copyright: © 2005 Public Library of Science.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.
A Microchip CD4 Counting Method for HIV Monitoring in Resource-Poor Settings
Test Your Knowledge: HIV Infection Epidemiology and Diagnosis
CD4 Measurements in Patients with HIV: Are They Feasible for Poor Settings?
==== Body
It doesn't take a trained physician to know that a disease needs to be diagnosed to be treated. And it doesn't take an economist to know that a disease cannot be diagnosed if the required tools are unaffordable or impractical. The absence of early diagnosis and treatment is particularly problematic for infectious disease, where the lack of early treatment or isolation can result in an epidemic.
Given the technological requirements, diagnosis and monitoring of HIV infection is problematic in resource-poor areas. The advent of rapid tests for diagnosing HIV infection represents one part of the solution. Less clear is how patients diagnosed with HIV infection will be monitored, given the importance of CD4 cell counts. A decrease in CD4+ T lymphocytes—a critical immune cell infected by HIV—is one of the hallmarks of HIV disease, and CD4+ cell number is a key factor in determining disease progression and monitoring treatment. The methods for determining CD4+ cell numbers are technically complex, expensive, and not easily transportable. These factors severely limit the ability to monitor HIV disease in locations where resources, training, and mobility are limited.
Digital image of whole blood obtained using the prototype device: CD4+ T cells shown in yellow, monocytes in green, CD8+ T cells in red
Lymphocytes are characterized by cell surface markers; thus, CD4+ lymphocytes express the CD4 marker on their surface. Antibody probes that specifically recognize this and other cell surface markers (such as CD8, which distinguishes that lymphocyte population from CD4+ lymphocytes, and CD3, which is a marker for all T lymphocytes) are used to count and differentiate various cell populations. By labeling cells with fluorescently tagged antibodies that recognize one or more cell surface molecules, the relative and absolute numbers of specific cells can be determined by a technique called flow cytometry. The labeled cells are passed through the flow cytometer, where the fluorescent probes are activated by lasers in a manner that can be read by specific detectors. The CD4+ cell number is directly correlated with the resultant fluorescent intensity and other light scatter properties. The problem is that flow cytometry requires costly reagents and substantial technical expertise—factors that limit its use in less developed areas.
Taking advantage of advances in microfluidics, digital imaging, and cell analysis, William Rodriguez and colleagues now report on a way to count CD4+ cells in a relatively quick, easy, and affordable manner. Small volumes of blood (an amount that could be obtained by a finger prick as opposed to drawing blood from a vein) are labeled as in flow cytometry, but with far less of the expensive reagents. Microfiltration allows the labeled CD4+ cells to be captured and separated from red blood cells, another simplification relative to flow cytometry. Digital images of the labeled cells, obtained by digital fluorescence microscopy, are then analyzed by newly developed software that can distinguish the CD4, CD8, and CD3 labels, thus allowing determination of absolute CD4+ counts, CD4+ percentages, and CD4+:CD8+ lymphocyte ratios.
Rodriguez et al. found that this new method was less accurate than flow cytometry for determining absolute CD4+ lymphocyte counts above 500 cells/mm3 (levels that are typically not relevant for monitoring HIV-infected individuals). But the method was as accurate as flow cytometry at clinically relevant levels of CD4+ cells for HIV-infected adult individuals. Although only a small number of pediatric patients were examined (and thus statistical significance could not be ascertained), the method appears to be also effective in determining CD4+ lymphocyte percentages in children.
The detection system used in the present report is a tabletop instrument that serves as a prototype for a fully portable handheld model, which is now under development. After some modest training, such a tool should allow a variety of health-care workers in remote areas to accurately analyze the CD4+ status of HIV-infected patients (the basis for treatment decisions) locally. In an accompanying Perspective discussing this new tool (DOI: 10.1371/journal.pmed.0020214), Zvi Bentwich argues that before it is ready for widespread use, several issues still need to be resolved, such as its final cost and its applicability to pediatric patients. “Despite these reservations,” he says, “the authors of this study should be commended for addressing an extremely important issue and developing this novel approach for counting CD4 in patients with HIV.”
| 0 | PMC1176238 | CC0 | 2021-01-05 10:40:16 | no | PLoS Med. 2005 Jul 19; 2(7):e242 | utf-8 | PLoS Med | 2,005 | 10.1371/journal.pmed.0020242 | oa_comm |
==== Front
PLoS MedPLoS MedpbioplosbiolPLoS Medicine1549-12771549-1676Public Library of Science San Francisco, USA 1602256410.1371/journal.pmed.0020246Research ArticleEpidemiology/Public HealthHealth PolicyHIV/AIDSMedical EthicsMedical LawEthicsHIV Infection/AIDSHealth PolicyQuality of health careHuman rightsDiscriminatory Attitudes and Practices by Health Workers toward Patients with HIV/AIDS in Nigeria HIV/AIDS and Clinician AttitudesReis Chen [email protected]
1
Heisler Michele
1
2
Amowitz Lynn L
1
3
Moreland R. Scott
4
Mafeni Jerome O
4
Anyamele Chukwuemeka
5
Iacopino Vincent
1
1Physicians for Human RightsBoston, MassachusettsUnited States of America2Department of Internal Medicine, School of Medicine, and Ann Arbor Veterans' Affairs Health System, University of MichiganAnn Arbor, MichiganUnited States of America3Brigham and Women's Hospital and Harvard Medical SchoolBoston, MassachusettsUnited States of America4Policy ProjectMaitamaNigeria5Center for the Right to HealthShomoluNigeriaBenatar Solomon Academic EditorUniversity of Cape TownSouth Africa
Competing Interests: The authors have declared that no competing interests exist.
Author Contributions: CR, LLA, MH, RSM, JOM, CA, and VI designed the study. CR, LLA, JOM, and CA implemented the study. CR and MH analyzed the data. CR, LLA, MH, RSM, JOM, CA, and VI contributed to writing the paper.
8 2005 19 7 2005 2 8 e24627 11 2004 13 6 2005 Copyright: © 2005 Reis 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.
How Do Nigeria's Health-Care Personnel Treat Patients with HIV/AIDS?
Achieving Gold Standards in Ethics and Human Rights in Medical Practice
Background
Nigeria has an estimated 3.6 million people with HIV/AIDS and is home to one out of every 11 people with HIV/AIDS worldwide. This study is the first population-based assessment of discrimination against people living with HIV/AIDS in the health sector of a country. The purpose of this study was to characterize the nature and extent of discriminatory practices and attitudes in the health sector and indicate possible contributing factors and intervention strategies. The study involved a cross-sectional survey of 1,021 Nigerian health-care professionals (including 324 physicians, 541 nurses, and 133 midwives identified by profession) in 111 health-care facilities in four Nigerian states.
Methods and Findings
Fifty-four percent of the health-care professionals (550/1,021) were sampled from public tertiary care facilities. Nine percent of professionals reported refusing to care for an HIV/AIDS patient, and 9% indicated that they had refused an HIV/AIDS patient admission to a hospital. Fifty-nine percent agreed that people with HIV/AIDS should be on a separate ward, and 40% believed a person's HIV status could be determined by his or her appearance. Ninety-one percent agreed that staff and health-care professionals should be informed when a patient is HIV-positive so they can protect themselves. Forty percent believed that health-care professionals with HIV/AIDS should not be allowed to work in any area of health-care that requires patient contact. Twenty percent agreed that many with HIV/AIDS behaved immorally and deserve the disease. Basic materials needed for treatment and prevention of HIV were not adequately available. Twelve percent agreed that treatment of opportunistic infections in HIV/AIDS patients wastes resources, and 8% indicated that treating someone with HIV/AIDS is a waste of precious resources. Providers who reported working in facilities that did not always practice universal precautions were more likely to favor restrictive policies toward people with HIV/AIDS. Providers who reported less adequate training in HIV treatment and ethics were also more likely to report negative attitudes toward patients with HIV/AIDS. There was no consistent pattern of differences in negative attitudes and practices across the different health specialties surveyed.
Conclusion
While most health-care professionals surveyed reported being in compliance with their ethical obligations despite the lack of resources, discriminatory behavior and attitudes toward patients with HIV/AIDS exist among a significant proportion of health-care professionals in the surveyed states. Inadequate education about HIV/AIDS and a lack of protective and treatment materials appear to contribute to these practices and attitudes.
Results of a survey of attitudes and actions of over 1000 Nigerian health care professionals (doctors, nurses, and midwives).
==== Body
Introduction
With an estimated 3.5 million people with HIV/AIDS, Nigeria is home to one of every 11 of the 40 million people with HIV/AIDS worldwide [1]. The HIV prevalence among adults in Nigeria has increased from 1.8% in 1991 to an estimated 5.8% in 2001 [1]. Prevalence ranges from 2% to 14.9% in the country's 36 states and Federal Capital Territory [2]. According to official estimates, Nigeria has an estimated 3.6 million people with HIV/AIDS and approximately 310,000 AIDS deaths this year alone [3], and these numbers are projected to increase each year. In 1999, with the election of President Olusegun Obasanjo, Nigeria emerged from approximately 20 years of military dictatorship in which little governmental attention or funding was directed at addressing HIV/AIDS [4].
People living with HIV/AIDS (PLWA) in Nigeria have been found to be subject to discrimination and stigmatization in the work place, and by family and communities [5,6]. PLWA may also face discrimination from those employed in the health-care sector. [5]. Discriminatory or unethical behavior by health-care professionals against PLWA, as documented in other countries [7–11], may create an atmosphere that interferes with effective prevention and treatment by discouraging individuals from being tested or seeking information on how to protect themselves and others from HIV/AIDS [12–14]. Furthermore, discriminatory practices and violations of international principles of medical ethics may serve to legitimize other forms of discrimination against people living with HIV/AIDS.
Anecdotal information suggests that health-care professionals in Nigeria may engage in discrimination against and stigmatization of PLWA [6,15]. The prevalence, character of, and factors contributing to these practices are, however, largely undocumented. To address this, Physicians for Human Rights (PHR), Policy Project Nigeria, and the Center for the Right to Health conducted a survey of health professionals in four sites in Nigeria. The study was designed to answer three research questions. (1) Are there discriminatory practices in the health sector that affect the health and well-being of people with HIV/AIDS in Nigeria? (2) How receptive are health workers and institutions to treating people with HIV/AIDS? And (3) what underlying factors may contribute to any discriminatory practices? The study was intended to inform ongoing policy discussions and development of effective interventions.
Methods
Sampling
At the time of the study, approximately 120,000,000 people were living in the 36 states and Federal Capital Territory of Nigeria [16]. We conducted the study in four states: Abia, Gombe, Kano, and Oyo. These sites were selected by dividing the country's six geopolitical zones into two sections—north and south—in order to capture geographical and other differences and then randomly selecting two of three zones from each section. Within the four selected zones, using health-care facility lists compiled by Nigeria's Federal Ministry of Health [17], we identified states that have a tertiary care institution and randomly selected one of these states from each zone. To obtain a representative sample of health-care professionals, we proportionally sampled doctors, nurses, and midwives from the tertiary facility and systematically selected public and private secondary and primary health-care facilities in the four states. Fifty-four percent of the health-care professionals were sampled from tertiary care facilities. We determined the sample size based on local activists' estimates that 10% of clinicians have discriminatory behavior and attitudes, a margin of error of ± 0.01% and a 90% confidence (10% significance) level. The sample size required given these constraints was 301 health-care professionals. However, our sample design included several levels of clustering, and we therefore assumed a design effect of three and thus the sample size needed was calculated to be approximately 1,000. Eligible facilities were medical facilities included in the published federal government database, which indicated that there were 2,585 health facilities in the four states [17]. In each health-care facility, we systematically sampled from all doctors, nurses, and midwives to acquire information about their knowledge, attitudes, and behavior. Eligible professionals were physicians or certified nurses or midwives working in positions with direct patient contact. Data on the number of health-care professionals were derived from Federal Ministry of Health data, which indicated that these four states have a total of nearly 4,500 health-care professionals who serve a population of approximately 17.8 million people [18].
Survey Questionnaire
The 104-item health-care professional survey included questions on respondent demographics; practices regarding informed consent, testing, and disclosure; treatment and care of patients with HIV/AIDS; and attitudes and beliefs about treatment and care of patients with HIV/AIDS including informed consent, testing, and disclosure.
Treatment and care practices of patients with HIV/AIDS were assessed using Likert-type scales (e.g., the possible answers were “always,” “most of the time,” “sometimes,” “rarely,” and “never”). Attitudes and beliefs were assessed by a response of “agree” or “disagree” with statements regarding testing, treatment, and care of patients with HIV/AIDS.
Using a separate 103-item survey instrument, we obtained information about each facility's capacity, resources, and policies from the person in charge of the facility. The questionnaires were written and interviews conducted in English. Seven regional, human rights, and medical experts reviewed the questionnaires for content validity. The instruments were pilot tested among 20 participants in Lagos and suggestions regarding clarity and cultural appropriateness were incorporated.
Interviewers
After completing an intensive training program, 24 Nigerian surveyors conducted the survey interviews. Interviewer training consisted of 5 d of classroom teaching and role-playing followed by several days of field observation and ongoing supervision by PHR and Nigerian researchers.
All interviews were conducted over 5 wk in October and November 2002. Interviews lasted approximately 20–30 min and were conducted in the most private setting possible within each health-care facility. All questionnaires were reviewed for completeness and for correctness of recording after the interview by the interviewers themselves, by the Nigerian research team leaders, and by PHR field supervisors at the end of each day.
Definitions
In the surveys, informed consent was defined as ensuring that a patient who is competent to make decisions is informed and consulted about his or her care. Respondents were informed that this included the responsibility of the clinician to let the patient know about any procedure or medical decision, reasonable alternatives to it, and the risks, benefits, uncertainties, and possible consequences related to each alternative. The clinician must carry out the discussion in layperson's terms, assess the patient's understanding along the way, and ensure that the patient understands the information and consents to it voluntarily [19]. Universal precautions were defined as the use of protective barriers such as gloves, gowns, aprons, masks, or protective eyewear, which can reduce the risk of exposure to potentially infective materials at all times regardless of a patient's HIV or other status [20].
Human Subjects Protection
This study was reviewed and approved by an independent ethics review board of individuals with expertise in clinical medicine, public health, bioethics, and international HIV/AIDS and human rights research developed for this research project by PHR. In reviewing the research, the review board was guided by the relevant provisions of Title 45 of the US Code of Federal Regulations [21], and complied with the Declaration of Helsinki, as revised in 2000 [22]. The study was also reviewed for ethical and cultural appropriateness by a panel convened in Nigeria by Policy Project Nigeria. In addition, permission for the study and access to facilities was granted by the Nigerian Federal Ministry of Health, state and local government authorities, and facility directors. There were no limitations placed on movement or surveying. Verbal informed consent was obtained from all participants, their names were not recorded, and only minimal identifying information was taken in order to preserve the anonymity of their responses. Participants did not receive any compensation.
Statistical Analysis
The data were analyzed using Stata 7 [23]. To control for clustering and design effect, the sample was weighted by the number of states selected with a tertiary facility from each of six selected geopolitical zones, the number of local government areas per location, the number of facilities selected from each local government area, and the response rate in each location. The study's principal objective was to describe health-care professional practices and attitudes towards people with HIV/AIDS, rather than to conduct comparisons between professionals or explore associations between professional characteristics and different outcomes. However, we conducted bivariate analyses using chi-square analyses and simple logistic regression to compare negative practices and attitudes among the three health specialties surveyed (doctors, nurses, and midwives) and to test for associations between reported facility resources and providers' reported adequacy of AIDS training and reported negative practices and attitudes about HIV/AIDS.
Results
Characteristics of Facilities
Of the 163 facilities sampled, 20 were no longer operational; for ten, contact could not be established after two attempts at the time of sampling; and 15 were not eligible. Of the 118 eligible facilities where contact was established, 111 participated in the study (78% of operational facilities). Over half of the facilities were general hospitals (54%) and 23% were primary health centers.
Eighty-four percent of facility directors reported not having antiretroviral medications in their facility. Moreover, the availability of other medications and dietary supplements was limited, and protective materials and other supplies and utilities were not always available (Table 1).
Table 1 Characteristics of 111 Participating Facilities
Characteristics of Respondents
Of the 1,103 professionals sampled, 23 were not eligible, five were not available after two attempts at the time of sampling, eight were interrupted during the course of the interview, and 46 refused to participate. Consequently, 1,021 professionals participated in the study (93% response rate). Although we did not gather information on nonrespondents, most of the 46 nonrespondents who refused to participate cited lack of time as the reason they were unable to participate (36), with other nonrespondents citing other obligations (four), fear of reprisal (one), and opposition to the study (three).
Sociodemographic characteristics
Professionals were predominantly female (67%) with a mean age of 36 y. Fifty-six percent were nurses, 31% were physicians, and 12% were certified midwives (Table 2).
Table 2 Demographic Characteristics among 1,021 Respondents
HIV/AIDS training
Most professionals reported having some training on HIV/AIDS (Table 2). Current literature (69%), conferences (56%), and courses as a student (52%) were most frequently reported by professionals as the sources of this training. Seven percent reported having no training on HIV/AIDS at all.
Testing and Consent
Practices
Seventeen percent of surveyed health-care professionals reported that their facility had a written HIV testing policy (Table 3). Respondents indicated that the policies included requirements for informed consent (58%), pre-test counseling (53%), post-test counseling (52%), and post-test referral (29%).
Table 3 HIV/AIDS Testing, and Consent, Practices
Over 50% of professionals reported obtaining informed consent of patients for HIV tests half of the time or less, including 14% who reported never obtaining consent for HIV tests (Table 3). Fifty-four percent of respondents reported that, regardless of consent, routine HIV testing of all patients scheduled for surgery always took place at their facilities, and 50% reported such routine HIV testing of all women attending antenatal care clinics. Providers who reported that they lacked adequate training in HIV/AIDS treatment and ethics had 50% higher odds of reporting that they failed to obtain informed consent for HIV tests (more than 50% of the time) compared to providers reporting adequate training in these areas (odds ratio (OR) 1.53, 95% confidence interval [CI] 1.17–2.01).
Attitudes
Ninety-one percent of professionals agreed that staff and health-care professionals should be informed when a patient is HIV-positive so they can protect themselves (Table 4). Over three-quarters of respondents (78%) agreed that there are circumstances when it is appropriate to test a patient without his or her knowledge or permission. Fifty-seven percent of participants believed that relatives and sexual partners of patients with HIV/AIDS should be notified of the patient's status even without the patient's consent. Forty-six percent of professionals thought that the charts or beds of patients with HIV should be marked so that health facility workers know the patient's status.
Table 4 Provider Attitudes and Beliefs about HIV Testing, Consent, and Disclosure
Forty percent believed that health-care professionals with HIV/AIDS should not be working in any area of the health professions that requires patient contact. Twenty percent of respondents agreed that many of those who have HIV/AIDS behaved immorally and deserve the disease (Table 4).
Providers working in facilities that did not always practice universal precautions (65% citing lack of sufficient materials as the reason) were significantly more likely than those working in facilities that always observed universal precautions to agree that people with HIV/AIDS should not be employed in the health field (OR 1.43, 95% CI 1.09–1.74) and should not work in areas that require patient contact. They also had higher odds of agreeing that under certain circumstances, patients could be tested for HIV without their knowledge or permission (OR 1.63, 95% CI 1.14–2.33) Working in a facility that did not always practice universal precautions, being a nurse or midwife, and reporting inadequate training in HIV/AIDS treatment were all associated with agreeing that patients with HIV/AIDS should be on a separate ward in a hospital or clinic. Nurses and midwives both had more than five times the odds of agreeing that people with HIV/AIDS should not be employed in the health field than doctors and that the charts or beds of HIV patients should be marked. Nurses and midwives also had almost twice the odds of physicians of agreeing that under certain circumstances it is acceptable to test patients for HIV without their consent or knowledge.
Treatment and Care
Practices
Among health-care professionals, the three most important concerns about treating patients with HIV/AIDS were fear of becoming contaminated (81%), contamination of facility, materials, or instruments (17%), and not having materials needed to treat them (10%) (Table 5). Seventy-two percent of respondents reported that universal precautions were always practiced in the facilities in which they worked. Lack of materials—reported by 65% of professionals—was cited as the main reason for non-practice of universal precautions (Table 5).
Table 5 HIV/AIDS Treatment and Care Practices
Nine percent of professionals reported refusing to care for a patient with HIV/AIDS, and 9% indicated that they had refused a patient with HIV/AIDS admission to a hospital (Table 6). Sixty-six percent had observed other health-care professionals refusing to care for a patient with HIV/AIDS, and 43% had observed others refusing a patient with HIV/AIDS admission to a hospital. While less than one percent of professionals reported verbally mistreating a patient with HIV/AIDS, 27% of respondents reported seeing others verbally mistreat patients with HIV/AIDS.
Table 6 Assessment of Practices toward Patients with HIV/AIDS
Thirty-eight percent of professionals reported giving confidential information to a patient's family member without the patient's consent, and 53% had observed this behavior. Twelve percent of professionals reported giving confidential information to a person not related to a patient without consent, and 22% had observed this behavior (Table 6).
Providers who reported inadequate training in HIV/AIDS treatment and in ethics were significantly more likely to have refused to treat a patient with HIV/AIDS than those reporting adequate training in those two areas (OR 2.06, 95% CI 1.31–3.22). Providers working in facilities that did not always practice universal precautions were not more likely to have refused care to a patient themselves but were significantly more likely to report having observed other providers refuse to care for a patient with HIV/AIDS (OR 1.09, 95% CI 1.01–1.45). There were no differences among specialties in reporting having refused to care for a patient with HIV/AIDS.
Attitudes
To prevent discrimination by health-care professionals against patients with HIV/AIDS, most participants (87%) indicated that health-care professionals who engage in discriminatory practices should be educated and counseled. Health facility policies against discrimination were cited as solutions by 19% of professionals, and stronger laws against discrimination were suggested by 11% (Table 6).
Ninety-four percent indicated that medications to treat opportunistic infections may prolong the life of a patient who is HIV-positive (Table 7). Over half (59%) of professionals agreed that people with HIV/AIDS should be on a separate ward in a hospital or clinic. Forty-eight percent of participants expressed their belief that a person with HIV/AIDS cannot be treated effectively in their facility. Forty percent of health-care professionals reported that it is possible to determine a person's HIV status by looking at him or her, and 21% agreed that they could refuse to treat a patient with HIV/AIDS to protect themselves and their family. Twelve percent expressed agreement with the statement that treatment of opportunistic infections in patients with HIV/AIDS wastes resources, and 8% agreed that treating someone with HIV/AIDS is a waste of precious resources.
Table 7 Provider Attitudes and Beliefs regarding Treatment and Care of Patients with HIV/AIDS
Nurses had higher odds than physicians of agreeing that treating opportunistic infections in patients with HIV/AIDS is a waste of resources (OR 2.14, 95% CI 1.35–3.40), but physicians were 50% more likely than nurses to agree that they could refuse to treat a patient with HIV/AIDS to protect themselves and their family. Respondents who reported inadequate training in HIV/AIDS treatment also were significantly more likely to agree it was acceptable to refuse to treat a patient for these reasons (OR 1.34, 95% CI 1.31–3.22). Physicians had significantly higher odds than either nurses or midwives of agreeing that there were circumstances under which it was appropriate to reveal a person's HIV status to others without the patient's knowledge or permission.
Discussion
Most health-care professionals in the four states where the study was conducted appeared to be providing care to patients who were HIV-positive and complying with their ethical responsibilities despite their lack of training on HIV/AIDS and their having insufficient supplies of materials needed for treatment and prevention in the facilities where they work. A significant number, however, reported engaging in discriminatory and/or unethical behavior. These practices are corrosive to the health professions as they taint all health professionals and erode trust in them. They also represent missed opportunities for prevention, positive living education, and treatment, thereby undermining Nigeria's concerted national efforts to address the HIV/AIDS epidemic. Our study findings suggest that there are several factors that may contribute to such discriminatory and/or unethical behavior by health-care professionals against people with HIV/AIDS in Nigeria.
The vast majority of professionals expressed an interest in additional information and suggested education as a way to address discriminatory behaviors by their colleagues. An immediate investment to ensure the education of all existing clinical staff about HIV/AIDS, including modes of transmission, universal precautions, and the rights of PLWA would likely reduce the number of discriminatory practices towards PLWA and may improve these patients' care and access to health services. This assertion is supported by previous studies that demonstrate the effect of HIV/AIDS education of nurses and other health workers on their attitudes and behavior towards patients who are HIV-positive in Nigeria and elsewhere [24–26]. These studies also suggest that education about scientific matters is not likely to be sufficient to achieve change in practice and that educational programs may also need to address attitudes and cultural beliefs.
This study further suggests that the lack of protective and other materials needed to treat and prevent the spread of HIV and related conditions contributes to discriminatory behavior. While the issue of access to affordable antiretroviral treatment is the subject of much debate in Nigeria [13,27], many of the facilities in this study did not even have sufficient stocks of basic antibiotics to treat opportunistic infections. The lack of protective materials, documented in the health facility survey and cited also by professionals as the main reason for not applying universal precautions, contributes to discriminatory behavior in two ways. First, professionals lacking adequate protection may come to fear PLWA and fear may lead to discrimination [28–30]. Second, lack of resources also results in differential treatment practices that may contribute to stigmatization of PLWA.
In order to do their jobs safely and effectively, health professionals must be provided with adequate supplies of essential protective materials. Further, the lack of basic medications hampers the ability of health professionals to provide appropriate treatment. Without these materials, it is unlikely that education of health professionals and implementation of anti-discrimination policies alone will have the desired impact on practice.
It is likely that in other low-resource contexts, the absence of medications needed to treat HIV/AIDS-related illnesses, a lack of materials needed for protection of health personnel, and insufficient knowledge of health personnel about HIV/AIDS may contribute to discriminatory behavior towards people with HIV/AIDS. The role of these factors should be investigated. While addressing these factors may not eliminate all discriminatory behavior, these basic investments in the health-care sector are likely to result in improvements.
HIV infection is both a product of and a factor contributing to human rights violations [12]. The documented marginalization of certain groups, and their increased risk for infection with HIV in Nigeria [1], must be considered in light of this study. Misconceptions must be taken into account when developing education and training programs for professionals and the public. Nigerian health professionals are members of their society, one in which stigma and moral judgment appear to be attached to HIV/AIDS [6]. Twenty percent of respondents agreed that many of those who have HIV/AIDS behaved immorally and deserve the disease. As such, it is likely that governmental and facility policies and monitoring to reduce discriminatory practices in the health-care sector will be an important aspect of addressing these practices.
Numerous international and regional human rights instruments, to which Nigeria is a party [31], protect the rights of PLWA. These include [32–37] the African Charter on Human and People's rights [36], the Convention on Elimination of All Forms of Discrimination against Women [34], the Convention on the Rights of the Child [35], the International Convention on Elimination of All Forms of Racial Discrimination [37], the International Covenant on Economic, Social, and Cultural Rights [32], and the International Covenant on Civil and Political Rights [33]. Of these, only the African Charter and the Convention on the Rights of the Child have been incorporated into the domestic law of Nigeria [4,38]. Nigeria is also a signatory to the Universal Declaration of Human Rights [39].
The above instruments set out Nigeria's obligations to protect the rights of PLWA including the right to life [33], the right to education [36], the right to marry and found a family [33], the right to nondiscrimination [36], the right to share in the benefits of scientific advancements [32], the right to privacy [33], and the right to freedom of association [36].
Several of the instruments to which Nigeria is a party include the right to health [34–37,40]. The right to health, was first elaborated in the International Covenant on Economic, Social, and Cultural Rights, Article 12 [32], which states:
1. The States Parties to the present Covenant recognize the right of everyone to the enjoyment of the highest attainable standard of physical and mental health.
2. The steps to be taken by the States Parties to the present Covenant to achieve the full realization of this right shall include those necessary for:
(a) The provision for the reduction of the stillbirth-rate and of infant mortality and for the healthy development of the child;
(b) The improvement of all aspects of environmental and industrial hygiene;
(c) The prevention, treatment and control of epidemic, endemic, occupational and other diseases;
(d) The creation of conditions which would assure to all medical service and medical attention in the event of sickness.
In 2000, the Committee on Economic Social and Cultural Rights (ESCR Committee), responsible for interpretation and monitoring of the International Covenant on Economic, Social, and Cultural Rights, published General Comment 14 on the Right to the Highest Attainable Standard of Health [40]. The ESCR Committee determined that fulfillment of the right to health means that access to health services must not be limited based on discrimination on a prohibited ground, including HIV status .
In General Comment 14, the ESCR Committee also set out the core obligations of a state party to protect the right to health, which include ensuring “the right of access to health facilities, goods and services on a non-discriminatory basis, especially for vulnerable or marginalized groups,” the provision of essential drugs “as from time to time defined by WHO's Action Programme on Essential Drugs,” and ensuring “equitable distribution of all health facilities, goods and services.” In addition to these and other core obligations, the ESCR Committee also set out “obligations of comparable priority” , including a state party's obligation “to take measures to prevent, treat and control epidemic and endemic diseases,” “to provide education and access to information concerning the main health problems in the community, including methods of preventing and controlling them,” and “to provide appropriate training for health personnel, including education on health and human rights.”
The ESCR Committee also stated in General Comment 14 that “any person or group who is a victim of a violation of the right to health should have access to effective judicial or other appropriate remedies at both national and international levels.”As a state party, Nigeria is bound by the provisions of the International Covenant on Economic, Social, and Cultural Rights and the authoritative interpretations of the ESCR Committee. This study finds that some PLWA have been excluded from access to health care because of their HIV status and that, at this time, PLWA have no access to judicial or other remedial processes to address this. The data further suggest that inadequate education of health personnel about HIV/AIDS along with a lack of protective and treatment materials likely contribute to these behaviors by health professionals. It is therefore likely that Nigeria has not met its core obligations to fulfill and protect the right to health. The findings of this study suggest that, in order to fulfill its obligations, the government of Nigeria should continue to address gaps in policy and legislation and work together with the international community to ensure that health professionals receive the training, protective materials, and medications they need to treat PLWA.
International principles of medical ethics and Nigerian codes of conduct clearly provide for patient autonomy, i.e., the right to informed consent and confidentiality of patient information. In addition to representing violations of human rights, the denial of treatment and breaches of informed consent and confidentiality detailed in this paper contravene international principles of medical ethics and Nigerian health professional codes of conduct. The Rules of Professional Conduct for Medical and Dental Practitioners in Nigeria [41] states that “a doctor shall preserve absolute secrecy on all he knows about his patient even after the patient has died, because of the confidence entrusted to him.” The binding rules also state that “practitioners…must always obtain consent of the patient or the competent relatives...before embarking on any special treatment procedures with determinable risks.”
Nigerian medical practitioners also have a duty under these rules to report any unethical conduct by their peers to the Medical and Dental Council of Nigeria. According to the rules, “every doctor or dentist must be his brother's keeper, with regard to the observance and indeed the enforcement of the rules and regulations which guide the profession. Doctors and dentists should expose without fear or favour, before the Medical and Dental Council of Nigeria either directly or through the Nigerian Medical Association, any corrupt, dishonest, unprofessional or criminal act or omission on the part of any doctor or dentist.” There is no indication that this may have happened in the case of the breaches documented in this study. At the time of publication, no specific medical ethics principles on HIV and AIDS have been articulated by the Medical and Dental Council of Nigeria.
Limitations
The study was conducted in four states in Nigeria with a total population of 17 million [2]. It is possible that these sites, though chosen at random from states with tertiary care facilities, may differ significantly from others in terms of resources and training provided to health-care providers. Although sampled systematically, it is possible that sampled facilities and health-care professionals may differ significantly from those that were not sampled in the four study states.
Although the findings of this study can not be generalized to Nigeria as a whole, it is likely that, depending on resources and training available to the health-care sector, the level of discriminatory behavior may differ in other parts of the country.
The apparent discrepancy between reported and observed behavior may indicate under- or overreporting of discriminatory behavior or may result from health-care professionals within the same institution having observed the same incidents.
While this study focused on HIV/AIDS, it is possible that health-care professionals also engage in inappropriate behavior toward or breach the confidentiality of people with other conditions. The health-care system in Nigeria is underfunded and suffers from fundamental problems including material scarcity and inadequacies in infrastructure, which may contribute to this behavior overall [4,42,43]. We did not specifically ask clinicians to compare their treatment of patients positive for HIV with that of other patients. Even if health-care professionals engage in breaches of confidentiality and other inappropriate behavior toward patients with other conditions, however, it is likely that the consequences of such actions may be worse for patients positive for HIV than for patients with other conditions.
Despite efforts to ensure privacy during interviews, the lack of privacy, or concern about job status, may have resulted in an underreporting of discriminatory behavior and/or an overreporting of “correct” practices or attitudes. Although interviewers were careful to explain that there would be no material gain or penalty to the respondent or his or her facility from participation in the study, the responses may have been inaccurate if respondents judged it in their material or political interest to exaggerate or conceal certain behaviors.
Conclusion
Despite these limitations, the study documents a significant proportion of health professionals in four states in Nigeria as reporting discriminatory attitudes and engaging in discriminatory and unethical behavior toward patients with HIV or AIDS, including denial of care, breach of confidentiality, and non-consented HIV testing. The breaches of confidentiality and testing for HIV without informed consent reported by participants are in contravention of international principles of medical ethics [44], and are also breaches of the Nigerian physician code of conduct [41]. The study identifies four factors that may contribute to this behavior: lack of correct information and education about HIV/AIDS and prevention of infection, lack of protective materials needed for the practice of universal precautions, lack of materials needed to care for and treat patients with HIV/AIDS, and prevailing attitudes about PLWA. This study suggests that adequately addressing these discriminatory practices and attitudes requires targeted education of health professionals and provision of adequate resources to health-care facilities combined with instituting and enforcing anti-discrimination policies.
Patient Summary
Background
People living with HIV/AIDS experience discrimination all over the world, often in both their private and professional lives, and sometimes by health-care personnel, including doctors and nurses.
Why Was This Study Done?
The researchers, who are part of a not-for-profit organization called Physicians for Human Rights, wanted to find out whether discrimination existed among health workers in Nigeria. Nigeria is home to one in 11 people living with HIV/AIDS, and antiretroviral drugs to treat patients are not widely available in the country.
What Did the Researchers Do?
They developed a survey and trained interviewers to get answers from 1,021 Nigerian health-care professionals—including 324 physicians, 541 nurses, and 133 midwives—about their attitudes and actions towards patients with HIV/AIDS. These health-care professionals worked in 111 different facilities in four of Nigeria's 36 states. The survey included questions such as “have you refused to care for an HIV/AIDS patient?” and “do you believe that a person's HIV status can be determined by his/her appearance?”
What Did They Find?
They found that while most of the interviewed people said that they treated people with HIV/AIDS in accordance with ethical and medical guidelines, a significant number of them reported attitudes and behavior that the researchers found worrisome. For example, 20% agreed that many individuals with HIV/AIDS had behaved immorally and deserved their infection, and 8% felt that treating someone with HIV/AIDS was a waste of resources. It didn't seem that there were big differences between the three groups (doctors, nurses, and midwives). Negative attitudes were higher among people from facilities that were not always able to take precautions against HIV infection (owing to lack of supplies). The same was true for facilities that did not have antiretroviral drugs to treat patients.
What Does This Mean?
It means that quite a few of the professionals surveyed—whose job it is to care for patients with HIV/AIDS—have negative attitudes towards them, and some of them have behaved in discriminating ways.
What Next?
It is not clear how representative the attitudes and behaviors reported by the participants in this study are, and future studies in Nigeria and in other countries are necessary to answer that question. Despite its limitations, this study suggests three things might help to reduce discrimination: educating health-care workers about HIV/AIDS and ethics, making sure that all facilities take appropriate precautions against HIV transmission, and making sure that all facilities can provide adequate care for patients with HIV/AIDS, including antiretroviral drugs.
Additional Online Resources
Information about stigmatization of people living with AIDS can be found at the following sources.
Joint United Nations Program on HIV/AIDS:
http://www.unaids.org/en/in+focus/hiv_aids_human_rights/stigma_discrimination.asp
United Kingdom National AIDS Trust:
http://www.areyouhivprejudiced.org/AboutUs.aspx
An article from the Indian Journal of Medical Ethics:
http://www.issuesinmedicalethics.org/082mi060.html
Physicians for Human Rights home page: http://www.phrusa.org/
Physicians for Human Rights Web page on HIV and human rights:
http://www.phrusa.org/campaigns/aids/aidsandhr.html
WHO work on HIV and human rights:
http://www.who.int/hhr/activities/publications/en/index.html
We would like to thank Frank Davidoff, MD, Ronald Waldman, MD, MPH, Jana Asher, MS, Leonard Rubenstein, JD, and Barbara Ayotte for their incisive comments and editing. We would also like to thank Ifeanyi Okekearu of the Center for the Right to Health who assisted in the supervision of data collection, Oluwole Fajemisin of Policy Project Nigeria, and the interviewers who assisted in data collection. We are especially grateful to all who participated in this study. PHR is a nongovernmental organization that uses medical and scientific knowledge to document violations of international human rights and humanitarian law. Policy Project Nigeria of the Futures Group International works in Nigeria to improve the policy environment for HIV/AIDS programs. The Center for the Right to Health is a Nigerian nongovernmental organization that monitors and documents policies and practices that influence the right to health in Nigeria.
This research was supported by grants from the United States Agency for International Development under USAID Contract No. HRN-C-00–00–00006–00. The views expressed herein are those of the authors only. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Citation: Reis C, Heisler M, Amowitz LL, Moreland RS, Mafeni JO, et al. (2005) Discriminatory attitudes and practices by health workers toward patients with HIV/AIDS in Nigeria. PLoS Med 2(8): e246.
Abbreviations
CIconfidence interval
ESCR CommitteeCommittee on Economic Social and Cultural Rights
ORodds ratio
PHRPhysicians for Human Rights
PLWApeople living with HIV/AIDS
==== Refs
References
Joint United Nations Programme on HIV/AIDS Report on the global HIV/AIDS epidemic 2002 Geneva Joint United Nations Programme on HIV/AIDS 229
Nigeria Federal Ministry of Health A technical report: The 2001 national HIV/syphilis sentinel survey among pregnant women attending ante-natal clinics in Nigeria 2001 Abuja (Nigeria) Federal Ministry of Health Available: http://rrh.deakin.edu.au/articles/subvieweuro.asp?ArticleID=273 . Accessed 20 June 2005
Joint United Nations Programme on HIV/AIDS Epidemiological fact sheets on HIV/AIDS and sexually transmitted infections: 2004 update 2004 Available: http://www.who.int/GlobalAtlas/PDFFactory/HIV/EFS_PDFs/EFS2004_NG.pdf . Accessed 20 June 2005
Unied Nations Common country assessment. Lagos (Nigeria): United Nations 2001 Available: http://www.worldvolunteerweb.org/dynamic/ infobase/pdf/2003/031201_EBCR_NGA_nigeria_report2.pdf . Accessed 20 June 2005
Center for the Right to Health Human rights and HIV/AIDS: Experiences of people living with HIV/AIDS in Nigeria 2001 Lagos (Nigeria) Center for the Right to Health Available: http://www.crhonline.org/pubdetail.php?pubid=2 . Accessed 20 June 2005
Alubo O Zwandor A Jolayemi T Omudo E Acceptance and stigmatization of PLWA in Nigeria AIDS Care 2002 14 117 126 11798411
Joint United Nations Programme on HIV/AIDS India: HIV and AIDS-related discrimination, stigmatization, and denial 2001 Geneva Joint United Nations Programme on HIV/AIDS 66
Danziger R Discrimination against people with HIV and AIDS in Poland BMJ 1994 308 1145 1147 8173458
Tirelli U Accurso V Spina M Vaccher E HIV and discrimination BMJ 1991 303 582
Devroey D Van Casteren V Sasse A Wallyn S Non-consented HIV testing by Belgian general practitioners AIDS 2003 17 641 642 12598792
Richter M Nature and extent of discrimination against PLWAs in South Africa: Interviews and a study of AIDS Law Project client files 1993–2001 2001 Johannesburg AIDS Law Project 46
Mann J Gostin L Gruskin S Brennan T Lazzarini Z Health and human rights Health Hum Rights 1994 1 6 23 10395709
Integrated Regional Information Networks Nigeria: Antiretroviral scheme draws poor response 2002 Available: http://www.irinnews.org/AIDSreport.asp?ReportID=1331&SelectRegion=West_Africa&SelectCountry=NIGERIA . Accessed 3April 2003
Parker R Aggleton P HIV/AIDS-related stigma and discrimination: A conceptual framework and an agenda for action 2002 New York Population Council Available: http://www.popcouncil.org/pdfs/horizons/sdcncptlfrmwrk.pdf . Accessed 15 June 2005
Adelekan ML Jolayemi SO Ndom RJE Adegboye J Babatunde S Caring for people with AIDS in a Nigerian teaching hospital: Staff attitudes and knowledge AIDS Care 1995 7 S63 S72 7632786
United Nations Population Fund State of the world population 2002 New York United Nations Population Fund Available: http://www.unfpa.org/swp/2002/english/ch1/index.htm . Accessed 15 June 2005
Nigeria Federal Ministry of Health Health facilities data base in Nigeria 2000 Abuja (Nigeria) Federal Ministry of Health 430
Nigeria Federal Ministry of Health Department of Research Planning and Statistics Nigeria health profile 1996 Abuja (Nigeria) Federal Ministry of Health
Beauchamp TL Childress JF Principles of biomedical ethics 1989 New York Oxford University Press 470
US Centers for Disease Control and Prevention Universal precautions for the prevention of transmission of HIV and other bloodborne infections 1987 Available: http://www.cdc.gov/ncidod/hip/Blood/UNIVERSA.HTM . Accessed 3 April 2003
United States Department of Health and Human Services Title 45 CFR Part 46: Protection of human subjects 2001 Bethesda (Maryland) National Institutes of Health Available: http://ohsr.od.nih.gov/guidelines/45cfr46.html . Accessed 20 June 2005
World Medical Association Declaration of Helsinki: Ethical principles for medical research involving human subjects 2004 Ferney-Voltaire (France) World Medical Association Available: http://www.wma.net/e/policy/b3.htm . Accessed 20 June 2005
Stata Stata 7.0 (intercooled) for Windows [computer program] 2002 College Station (Texas) Stata
Uwakwe CBU Systematized HIV/AIDS education for student nurses at the University of Ibadan, Nigeria: Impact on knowledge, attitudes and compliance with universal precautions J Adv Nurs 2000 32 416 424 10964190
McCann TV Sharkey RJ Educational intervention with international nurses and changes in knowledge, attitudes and willingness to provide care to patients with HIV/AIDS J Adv Nurs 1998 27 267 273 9515635
Ezedinachi EN Ross MW Meremiku M Essien EJ Edem CB The impact of an intervention to change health workers' HIV/AIDS attitudes and knowledge in Nigeria: A controlled trial Public Health 2002 116 106 112 11961679
Akanni O Problems dog Nigeria's ARV programme. Journalists against AIDS Nigeria 2002 April 4 Available: http://www.nigeria-aids.org/MsgRead.cfm?ID=591 . Accessed 3 April 2003
Bermingham S Kippax S HIV-related discrimination: A survey of New South Wales general practitioners Aust N Z J Public Health 1998 22 92 97 9599859
Essien EJ Ross MW Ezedinachi ENU Meremikwu M Cross-national HIV infection control practices and fear of AIDS: A comparison between Nigeria and the USA Int J STD AIDS 1997 8 764 771 9433951
Link RN Feingold AR Charap MH Freeman K Shelov SP Concerns of medical and pediatric house officers about acquiring AIDS from their patients Am J Public Health 1988 78 455 459 3348474
For the record 2002: The United Nations human rights system. Nigeria—Treaties: ratifications and reservations 2002 Available: http://www.hri.ca/fortherecord2002/vol2/nigeriarr.htm . Accessed 31 March 2003
United Nations General Assembly International covenant on economic, social and cultural rights. General Assembly Resolution 2200 (XXI), 21st session, Supp. (No. 16) 49, UN Doc A/6316 1966 New York United Nations General Assembly Official Records
United Nations General Assembly International covenant on civil and political rights. United Nations General Assembly Resolution 2200a (XXI), 21st session, Supp. No. 16, UN Doc A/6316 1967 New York United Nations General Assembly Official Records
United Nations General Assembly Convention on the elimination of all forms of discrimination against women. General Assembly Resolution 34/180, 34th session, Supp. No. 46, UN Doc A/34/46 1976 New York United Nations General Assembly Official Records
United Nations General Assembly Convention on the rights of the child. General Assembly Resolution 44/25, 44th session, Supp No. 49, UN Doc A/44/49 1989 New York United Nations General Assembly Official Records
Organization of African Unity African (Banjul) charter on human and people's rights 1981 OAU Doc CAB/LEG/67/3 rev.5, 21 I.L.M. 58 (1982). Available: http://www.africa-union.org/Official_documents/Treaties_%20Conventions_%20Protocols/Banjul%20Charter.pdf . Accessed 20 June 2005
United Nations General Assembly International convention on elimination of all forms of racial discrimination. General Assembly Resolution 2106A (XX) 1965 New York United Nations General Assembly Official Records
Oluduru O Legal issues raised by HIV/AIDS HIV/AIDS and human rights: Role of the judiciary 2001 Lagos (Nigeria) Center for the Right to Health Available: http://www.crhonline.org/pubdetail.php?pubid=3 . Accessed 20 June 2005
United Nations General Assembly Universal declaration of human rights General Assembly Resolution 217A (III), UN Doc A/810 1948 New York United Nations General Assembly Official Records
Committee on Economic Social and Cultural Rights General comment No. 14: The right to the highest attainable standard of health 2000 Available: http://www.unhchr.ch/tbs/doc.nsf/(Symbol)/40d009901358b0e2c1256915005090be?Opendocument . Accessed 20 June 2005
Medical and Dental Council of Nigeria Rules of professional conduct for medical and dental practitioners in Nigeria 1995 Abuja (Nigeria) Medical and Dental Council of Nigeria 42
Hargreaves S Time to right the wrongs: Improving basic health care in Nigeria Lancet 2002 359 2030 2035 12076574
Nigeria Federal Office of Statistics Core welfare indicators questionnaire survey, Benue State 2001 Abuja (Nigeria) Federal Office of Statistics 60
World Medical Association International code of medical ethics 1983 Available: http://www.wma.net/e/policy/c8.htm . Accessed 20 June 2005
| 16022564 | PMC1176239 | CC BY | 2021-01-05 10:40:22 | no | PLoS Med. 2005 Aug 19; 2(8):e246 | utf-8 | PLoS Med | 2,005 | 10.1371/journal.pmed.0020246 | oa_comm |
==== Front
PLoS MedPLoS MedpbioplosbiolPLoS Medicine1549-12771549-1676Public Library of Science San Francisco, USA 1600850810.1371/journal.pmed.0020247EssayInfectious DiseasesHealth PolicyHIV/AIDSHIV Infection/AIDSInfectious DiseasesMedicine in Developing CountriesHealth PolicyHealth education (including prevention and promotion)The Stigma of Being HIV-Positive in Africa EssayRankin William W [email protected] Sean Schell Ellen Laviwa Jones Rankin Sally H William W. Rankin is President, Global AIDS Interfaith Alliance, San Francisco, California, United States (http://www.thegaia.org). Sean Brennan is Associate Specialist, University of California Berkeley School of Public Health, California, United States. Ellen Schell is International Programs Director, Global AIDS Interfaith Alliance, San Francisco, California, United States. Jones Laviwa is Executive Director, Churches Action in Relief and Development, Blantyre, Malawi, Central Africa. Sally H. Rankin is Professor, School of Nursing, University of California San Francisco, California, United States.
Competing Interests: The authors have declared that no competing interests exist.
8 2005 19 7 2005 2 8 e247Copyright: © 2005 Rankin 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.Rankin and colleagues argue that HIV-related stigma is fueling the epidemic, and disempowering women even further.
HIV-related stigma is fueling the epidemic, and disempowering women even further
==== Body
Jonathan Mann, founder of the World Health Organization's Global Program on AIDS and untiring advocate for justice for people with HIV/AIDS, addressed the United Nations General Assembly in 1987 [1]. His speech characterized the three major phases of an HIV/AIDS epidemic. After the initial silent spread of virus came the outbreak of ill health. The final stage, he said—the stage of social impact—is marked by stigma, grinding down its victims with shame and isolation.
Mann's tragically short life was devoted to protecting all who stood to be diminished by illness-related stigma and the erosion of elemental human rights [2]. Why were stigma and human rights so essential to the work of a medical doctor fighting an infectious disease?
Fear of Stigma Fuels the HIV Epidemic
Stigma is of utmost concern because it is both the cause and effect of secrecy and denial, which are both catalysts for HIV transmission. Fear of stigma limits the efficacy of HIV-testing programs across sub-Saharan Africa, because in most villages everyone knows—sooner or later—who visits test sites [3,4]. While in some places the advent of free and accessible antiretroviral therapy has offered hope and encouraged people to go for testing, [5] stigma remains a barrier to testing even where treatment is available [6]. Without HIV testing, an essential first step to treatment, years may go by while people who are infected transmit the virus to others. When individuals finally become ill and seek care, treatment as a prevention strategy has lost much of its potential effectiveness.
Fear of stigma can cause pregnant women to avoid HIV testing, the first step in reducing mother-to-child transmission [7–9]. It may force mothers to expose babies to HIV infection through breast-feeding because the mothers do not want to arouse suspicion of their HIV status by using alternative feeding methods [10,11]. Fear of stigma, and the resulting denial, may even inhibit condom use in HIV discordant couples. Further evidence of how stigma leads to denial is the way in which newspaper obituaries avoid mentioning HIV/AIDS as a cause of death.
HIV-related stigma directly hurts people, who lose community support due to their real or supposed HIV infection. Individuals may be isolated within their family, hidden away from visitors, or made to eat alone [3]. These repercussions may or may not be simple acts of heartlessness. They may be a well-intentioned but ignorant attempt to preserve the family. In the community, the entire family may be sanctioned because one member is ill; in an impoverished society with no safety net of public services, this can be ominous for everyone [3].
In many African villages, an individual's, and a family's, life is closely intertwined with others. The same people have lived closely together for several generations, and there are few secrets. Inside families, caregivers may be largely concerned about contracting HIV through casual contact, and outside they fear the gossip that can greatly affect everyone's social standing. Neighbors and other customers, for instance, may refuse to purchase vegetables or poultry from someone associated with HIV [12]. In impoverished areas, this can devastate a family's chances of economic survival.
The language used to describe people living with HIV (such as “she is an HIV,” “he is a walking corpse”) clearly conveys stigmatizing attitudes. A particularly powerful example of stigmatizing language is found in parts of Tanzania, where an HIV-positive person is called nyambizi, or submarine, implying that the HIV-positive person is stealthy, menacing, and deadly [3,13].
People living with HIV can also experience a form of internalized stigma (Figure 1). Even without the burden of externally imposed social opprobrium, those living with a serious illness can face an enormous and painful inner struggle. They may eventually cease to be who they were, instead becoming a unitary “person with an illness” or—more damning—an “ill person,” a thing in which personhood and illness have completely fused. The philosopher Simone Weil characterized this assault of illness upon the self with the classical Greek notion of the soul—Malheur (affliction) stamps the soul to its very depths with scorn and disgust [14].
Figure 1 Kgalalelo Ntsepe, Who Was Named Miss HIV Stigma Free in 2003
In 2003, in Gaborone, Botswana, 14 women competed to become Miss HIV Stigma Free. The contest was won by Kgalalelo Ntsepe, who said: “It took a long time before I accepted my HIV status. At first, I almost wanted to kill myself. Eventually, I overcame my fears, even though my family and friends deserted me. But my church and my belief helped me to find a meaning in life again. I am Miss HIV Stigma Free. It's my responsibility to give strength to others. There's a life with HIV. There's life with AIDS.”
(Photo: Copyright WORLD VISION/Sönke C. Weiss)
The combination of external stigma and internal oppression of the self may impose a heavy burden. In our experience of working with people with HIV in Africa, the result of this burden is often a downward spiral marked by fatalism, self-loathing, and isolation from others. And by shaming and silencing the very people who could credibly speak for HIV prevention and provide care for HIV-positive others, stigma fuels the HIV epidemic, consigning more people to suffering and death.
Stigma in Society
Stigma is part of the attitudes and social structures that set people against each other. It impedes any countervailing forces for social equality. Certainly since Erving Goffman's seminal work on stigma in the early 1960s, stigma (plural stigmata) has been recognized as “an attribute that is significantly discrediting,” and it is known as a potent and painful force in individual lives [15]. Fueled by prejudice and appealing to it, stigma functions to diminish the person or group being targeted. Some commentators since Goffman have particularly examined stigma's broader social functioning. They have noted that while subordinating individuals or groups in society, the stigmatizing process also reinforces hierarchical patterns of privilege, where those at the top of a stratified society are pre-eminent over, and sometimes predatory upon, others at lower levels [16].
To see this perhaps more clearly, think of certain religious settings where punishment theories of illness causation are in force [17–19]. One such outlook presumes an aroused deity or ancestor bringing illness upon a person in retribution for an offense. This notion stigmatizes people struggling with their illness. It blames their sickness upon misbehaviors, while at the same time it rationalizes privileging the well over the ill. Punishment theories authorize communities to isolate or purge the “impure”—people whose illness or imagined “sinfulness” would contaminate the whole—while reassuring that virtue and social status will protect the righteous.
Clergy and other religious leaders are as susceptible as any to the temptation to exercise power over others. This imbalance of power is facilitated by such structured inequalities within churches as the preeminence of clergy over laity, of men over women, and even by the presumed superiority of the more “spiritual” over the less so. Under the influence of western missionaries, many African Christian organizations still promote evangelical formulae in which, it is taught, creation was originally good, but then the “fall” of humankind occurred, which is bad, and finally, redemption is available only for the chosen. This theological approach warrants valorizing or stigmatizing people as “saved” or “sinner,” “pure” or “impure,” “us” or “them,” and it strengthens the broader social stratifications within which stigma flourishes. What is weakened is the opportunity to apply healing insights from the rich Christian legacy of compassion, liberation, and hope [20].
Gender and HIV
In much of sub-Saharan Africa, women are a subordinate group who are expected to become pregnant, bear children, and fulfill the sexual desires of their husbands without hesitation [20]. Such traditional assumptions, sometimes reinforced by the missionary religions, greatly benefit men while predisposing women to HIV infection. Often husbands carry HIV, while barrier methods of disease prevention, such as condoms, are proscribed, perhaps most vigorously by male-dominant religious organizations.
In addition to women's subordinate status in many societies, they are also frequently stigmatized as the vectors of HIV transmission, despite overwhelming evidence to the contrary. In Malawi, the term for a sexually transmitted disease, regardless of its origin, is “woman's disease.” [21] Husbands have beaten and/or abandoned wives thought to be HIV-positive, despite the fact that many women contract the virus from their husbands. Some women are subject to violence if they refuse a sexual overture, ask their husband to use a condom, or request an HIV test. If a husband should die, the wife's in-laws may seize her inheritance [22]. A woman exhibiting the independence needed to protect her health and self-esteem risks the disapprobation of her family and of the community.
Men are the clear winners of this arrangement in both social and economic terms, and many widows and their children, dispossessed or not, struggle against enormous odds simply to survive. Public attitudes, stigma among them, help to sustain the entire unjust system.
Stigma and Human Rights
We marvel at the prescience and lucidity of Mann, a doctor who was dedicated to treating the whole person—both the physical ills and the emotional distress attendant upon these ills, including the stigma inherited from or imposed by societies where the oppression of some fortifies the privilege of others. Mann respected the healing potential of social justice in general, and of human rights in particular. He knew that a society in which multiple injustices routinely occur is itself not well, and he knew that widespread respect of human rights made less room for stigma and its harmful consequences.
Respect of human rights makes less room for stigma
The way to tackle social oppression of any kind is to introduce strategies that address underlying conditions of poverty, racism, and sexism that support such oppression [5]. This approach should be bolstered by sufficient legal and policy mechanisms to protect people subject to stigma and the erosion of human rights in general [5,23]. The same mechanisms should be functional and accessible to all.
To be effective, all HIV interventions should include an analysis of how stigma functions, how it enhances dominance and subordination in society, how it is that some win and others lose in the pernicious struggle for pre-eminence, and why it is that such a social scheme perversely flourishes in the first place [16].
Enlightened HIV prevention and care interventions (Figure 2) will empower the stigmatized through health education that lifts self-blame and shifts opprobrium to external, self-serving forces. While teaching respect for all through a more just society, these interventions will help people who are stigmatized to critique unjust societal dynamics and challenge assumptions and warrants of privilege.
Figure 2 Village Caregivers
This photograph shows some of the local women who work with Global AIDS Interfaith Alliance village-level projects in Africa, funded by the Bill and Melinda Gates Foundation. These women teach people about HIV/AIDS, help care for orphaned children, and visit and care for each person ill with AIDS every day. In doing so, they have helped to break down stigma in their villages.
(Photo: Global AIDS Interfaith Alliance)
A tall order? Maybe, but Mann asked all of us—those struggling with illness and the presumably healthy—to make societies as healthy as their individual members.
William W. Rankin acknowledges the kindness of the Rockefeller Foundation in enabling his research on HIV/AIDS-related stigma in Africa while a resident at the Foundation's Study and Conference Center in Bellagio, Italy, in April and May of 2004.
Citation: Rankin WW, Brennan S, Schell E, Laviwa J, Rankin SH (2005) The stigma of being HIV-positive in Africa. PLoS Med 2(8): e247.
==== Refs
References
Mann J Statement at an informal briefing on AIDS to the 42nd session of the United Nations General Assembly 1987 10 20 New York
Mann J The future of the global AIDS movement Harv AIDS Rev 1999 18 21 12295157
Nyblade L Pande R Mathur S MacQuarrie K Kidd R Disentangling HIV and AIDS stigma in Ethiopia, Tanzania and Zambia 2003 Washington, DC International Center for Research on Women 53
Cameron E The deafening silence of AIDS Health Hum Rights 2000 5 7 24 11154521
Castro A Farmer P Understanding and addressing AIDS-related stigma: From anthropological theory to clinical practice in Haiti Am J Public Health 2005 95 53 59 15623859
Green C Access to treatment of HIV/AIDS and other diseases: An overview Healthlink Worldwide 2003 Available: http://www.healthlink.org.uk/pubs/access.html . Accessed 25 March 2005
Etiebet MA Fransman D Forsyth B Coetzee N Hussey G Integrating prevention of mother-to-child HIV transmission into antenatal care: Learning from the experiences of women in South Africa AIDS Care 2004 16 37 46 14660142
Ekanem EE Gbadegesin A Voluntary counselling and testing (VCT) for Human Immunodeficiency Virus: A study on acceptability by Nigerian women attending antenatal clinics Afr J Reprod Health 2004 8 91 100 15623124
Thorne C Newell ML Mother-to-child transmission of HIV infection and its prevention Curr HIV Res 2003 1 447 462 15049430
Shapiro RL Lockman S Thior I Stocking L Kebaabetswe P Low adherence to recommended infant feeding strategies among HIV-infected women: Results from the pilot phase of a randomized trial to prevent mother-to-child transmission in Botswana AIDS Educ Prev 2003 15 221 230 12866834
Harvard AIDS Institute Update Trying to prevent mother-to-child transmission of HIV 2001 Harvard AIDS Institute researchers seek new methods. Harv AIDS Inst Update. Available: http://www.researchmatters.harvard.edu/story.php?article_id=334 . Accessed 14 March 2005
Muyinda H Seeley J Pickering H Barton T Social aspect of AIDS-related stigma in rural Uganda Health Place 1997 3 143 147 10670965
International Center for Research on Women Understanding HIV-related stigma and resulting discrimination in Sub-Saharan Africa: Emerging themes from early data collection in Ethiopia, Tanzania and Zambia ICRW Res Update 2002 Available: http://www.icrw.org/docs/Stigma_ResearchUpdate_062502.pdf . Accessed 25 March 2005
Weil S Waiting for God. E. Crawford, translator 1973 New York Harper and Row 227
Goffman R Stigma: Notes on the management of spoiled identity 1963 Englewood Cliffs (New Jersey) Prentice-Hall 147
Parker R Aggleton P HIV and AIDS-related stigma and discrimination: A conceptual framework and implications for action Soc Sci Med 2003 57 13 24 12753813
Allen PL The wages of sin: Sex and disease, past and present 2000 Chicago The University of Chicago Press 202
Avolos H Health care and the rise of Christianity 1999 Peabody (Massachusetts) Hendrickson 166
Rankin WW Cracking the monolith: The struggle for the soul of America 1994 New York Crossroad 155
Messer DE Breaking the conspiracy of silence: Christian churches and the global AIDS crisis 2004 Minneapolis Fortress 192
Rankin SH Lindgren T Rankin WW Ng'oma J Donkey work: Women, religion, and HIV/AIDS in Malawi Health Care Women Int 2005 26 4 16 15764458
Alumbo O Zwandor A Jolayemi T Omudu E Acceptance and stigmatization of PLWA in Nigeria AIDS Care 2002 14 117 126 11798411
Goldin CS Stigmatization and AIDS: Critical issues in public health Soc Sci Med 1994 39 1359 1366 7801171
| 16008508 | PMC1176240 | CC BY | 2021-01-05 10:40:32 | no | PLoS Med. 2005 Aug 19; 2(8):e247 | utf-8 | PLoS Med | 2,005 | 10.1371/journal.pmed.0020247 | oa_comm |
==== Front
PLoS MedPLoS MedpbioplosbiolPLoS Medicine1549-12771549-1676Public Library of Science San Francisco, USA 1600850110.1371/journal.pmed.0020248PerspectivesInfectious DiseasesHealth PolicyHIV/AIDSHIV Infection/AIDSInfectious DiseasesMedicine in Developing CountriesClinical trialsHealth PolicyWe Must Not Let Protestors Derail Trials of Pre-Exposure Prophylaxis for HIV PerspectivesLange Joep M. A Joep M. A. Lange is Director of the Center for Poverty-Related Communicable Diseases, Academic Medical Center, University of Amsterdam, Netherlands. He is the former president of the International AIDS Society (2002–2004). E-mail: [email protected]
Competing Interests: JMAL serves or has served on scientific (HIV) advisory boards and/or spoken at symposia of and/or received research grants from the following pharmaceutical companies: Abbott, Boehringer Ingelheim, Bristol-Myers Squibb, Gilead, GlaxoSmithKline, Idenix, Merck, Panacos, Pfizer, Roche, and Tibotec. He declares that he does not own any pharmaceutical company stock. He serves on the scientific advisory boards of the International AIDS Vaccine Initiative and the International Partnership for Microbicides, on the Global HIV Prevention Working Group of the Bill and Melinda Gates Foundation, and on the Strategic and Technical Advisory Committee for HIV/AIDS of the World Health Organization.
9 2005 19 7 2005 2 9 e248Copyright: © 2005 Joep M. A. Lange.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 Abandoned Trials of Pre-Exposure Prophylaxis for HIV: What Went Wrong?
The international AIDS community is being held hostage by a small number of activist groups, argues Lange.
==== Body
One of the great tragedies of our times is the extent to which HIV prevention efforts are falling short. In 2004, more new HIV infections occurred than in any previous year: close to 14,000 a day, 570 per hour, almost ten per minute. The greater part of new infections occurs in young people, over half in persons between 15 and 24 years of age, and over half in women. The increasing feminization of the HIV/AIDS epidemic reflects the vulnerable position of women in many societies [1]. HIV is a virus, but inequity is at the roots of most of its spread.
Condoms are highly effective at preventing sexual transmission of HIV, but only if they are available and used [2]. Even if the former is the case, women are often in a difficult position to negotiate use by their male partners [3]; this applies to female as well as male condoms.
In the absence of an effective preventive HIV vaccine [4], which is felt to be the only tool that can definitively break the epidemic, there is thus great need for alternative prevention technologies, especially those that can be “female-controlled”, i.e., use of which does not require consent of the male partner. Other vulnerable populations might equally profit from the availability of such interventions. The urgent need to develop female-controlled prevention methods explains the thrust to develop vaginal microbicides [5,6], and the more recent interest in using an antiretroviral oral pre-exposure prophylaxis (PREP), for which proof of concept has come from preventing mother-to-child transmission via breastfeeding [7].
Mills and Singh have written an informative and thoughtful essay for PLoS Medicine on the recent interruptions of several tenofovir PREP trials, instigated predominantly by activist groups, including Act Up–Paris [8]. They properly describe what has happened in the countries concerned, and, at least initially in their essay, they do not seem to be afraid of identifying culprits, albeit in an indirect way. For example, they say, “while freedom of expression is a cherished ideal, protest should be carried out in a responsible manner”, and “speculation, unwarranted criticism, overreaction, or sensationalizing facts risks stigmatizing tenofovir”. But I feel that in the end, the authors could have gone further in their criticism of the protestors who derailed the tenofovir trials and in their support for the trial investigators and sponsors.
The fear or trepidation about making a correct diagnosis, or being outspoken about the correct diagnosis, is in this case, as in most other circumstances, counterproductive and may even lead to more harm being done. The fact of the matter is that the investigators of the tenofovir trials did consult intensely with community groups concerned, but this consultative process obviously did not include every activist group in the world.
As Co-Chair of last year's XVth International AIDS Conference in Bangkok, Thailand, I became painfully aware of a structural flaw in the system of dealing with the activist community. The governing body of the conference, the Conference Organizing Committee, included three international and one local Community Co-Organizers, who were supposed to have a mandate to speak on behalf of the national and international AIDS community. In the end, though, it turned out that agreements on limiting disruption of speeches and destruction of property, which took endless hours of discussion and which were communicated to attendants, governments, speakers, and sponsors, could not be honored because certain activist groups simply ignored them. Likewise, in clinical research, investigators may end up in situations in which they may have had an intense dialogue and come to an agreement with what they have identified to be the relevant community organizations, and yet a day later are put on the stand by yet another activist group.
Life would be simpler if there were an umbrella organization with undisputed leadership among AIDS activists that was mandated to speak and act on behalf of this community; somewhat like labor unions in their heydays. But we are far from that.
Activist groups have now managed to derail several PREP trials, arguably the most important studies for those at high risk of acquiring HIV infection around the globe. Similarly, activist groups have endangered the funding and therefore the continuity of the International AIDS Conference (the only global forum about HIV/AIDS where researchers, health-care workers, community, activists, drug manufacturers, politicians, and leaders from all walks of life meet). And lately activist groups have prevented clinical trials with the promising and highly needed new class of CCR5-receptor-blocking antiretrovirals from proceeding in several European countries.
The methods of these specific activist groups are uninformed demagogy, intimidation, and “AIDS exceptionalism”, the last in the sense that they exploit their HIV-positive status to get away with behavior that would not be accepted from others. Within the international AIDS community, such form of activism is only practiced by a tiny minority, but it has taken us hostage. Those who will suffer the most from the misguided ethical imperialism that derailed the PREP trials do not live in Paris, but as usual in Nairobi, Johannesburg, Phnom Penh, and Calcutta.
There is no other area of medicine where activism has been so strong and has accomplished so much as in the HIV/AIDS field. Let's be just a little brave, and stand up to protect that legacy.
Citation: Lange JMA (2005) We must not let protestors derail trials of pre-exposure prophylaxis for HIV. PLoS Med 2(9): e248.
Abbreviations
PREPpre-exposure prophylaxis
==== Refs
References
Joint United Nations Programme on HIV/AIDS AIDS epidemic update 2004 2004 December Available: http://www.unaids.org/wad2004/report_pdf.html . Accessed 17 June 2005
Weller SC A meta-analysis of condom effectiveness in reducing sexually transmitted HIV Soc Sci Med 1993 36 1644 1653
Allen S Tice J Van de Perre P Serufilira A Effect of serotesting with counseling on condom use and seroconversion among HIV discordant couples in Africa BMJ 1992 304 1605 1609 1628088
Coordinating Committee of the Global HIV/AIDS Vaccine Enterprise The Global HIV/AIDS Vaccine Enterprise: Scientific strategic plan PLoS Med 2005 2 e25 10.1371/journal.pmed.0020025 15740411
Shattuck RA Moore JP Inhibiting sexual transmission of HIV-1 infection Nat Rev Microbiol 2003 1 25 34 15040177
Weber J Desai K Darbyshire J on behalf of the Microbicides Development Programme The development of vaginal microbicides for the prevention of HIV transmission PLoS Med 2005 2 e142 10.1371/journal.pmed.0020142 15916473
Vyankadondera J Luchters S Hassink E Pakker N Mmiro F Reducing risk of HIV-1 transmission through breastfeeding using antiretroviral prophylaxis in infants (SIMBA study) [abstract] 2003 Second IAS Conference on HIV Pathogenesis and Treatment 2003 July 13–16 Paris, France Abstract LB7
Singh JA Mills EJ The abandoned trials of pre-exposure prophylaxis for HIV: What went wrong? PLoS Med 2005 2 e234 10.1371/journal.pmed.0020234 16008507
| 16008501 | PMC1176241 | CC BY | 2021-01-05 10:40:32 | no | PLoS Med. 2005 Sep 19; 2(9):e248 | utf-8 | PLoS Med | 2,005 | 10.1371/journal.pmed.0020248 | oa_comm |
==== Front
PLoS MedPLoS MedpbioplosbiolPLoS Medicine1549-12771549-1676Public Library of Science San Francisco, USA 1601872110.1371/journal.pmed.0020249Research ArticleBioinformatics/Computational BiologyImmunologyInfectious DiseasesMicrobiologyVirologyEpidemiology/Public HealthHIV/AIDSStatisticsVaccinesInfectious DiseasesHIV Infection/AIDSImmunology and allergyPreclinical Assessment of HIV Vaccines and Microbicides by Repeated Low-Dose Virus Challenges Low-Dose Virus ChallengesRegoes Roland R [email protected]
1
Longini Ira M Jr
2
Feinberg Mark B
3
4
Staprans Silvija I
3
1Department of Biology, Emory UniversityAtlanta, GeorgiaUnited States of America2Department of Biostatistics, Rollins School of Public Health, Emory UniversityAtlanta, GeorgiaUnited States of America3Departments of Medicine and Microbiology and Immunology, Emory University School of Medicine and the Emory Vaccine CenterAtlanta, GeorgiaUnited States of America4Merck Vaccine DivisionMerck, West Point, PennsylvaniaUnited States of AmericaLipsitch Marc Academic EditorHarvard School of Public HealthUnited States of America
Competing Interests: The authors have declared that no competing interests exist.
Author Contributions: RRR performed the analysis. RRR, IML, MBF, and SIS designed the study and wrote the paper.
8 2005 19 7 2005 2 8 e24911 4 2005 13 6 2005 Copyright: © 2005 Regoes 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.
Modeling HIV Vaccine Strategy in Animals
Background
Trials in macaque models play an essential role in the evaluation of biomedical interventions that aim to prevent HIV infection, such as vaccines, microbicides, and systemic chemoprophylaxis. These trials are usually conducted with very high virus challenge doses that result in infection with certainty. However, these high challenge doses do not realistically reflect the low probability of HIV transmission in humans, and thus may rule out preventive interventions that could protect against “real life” exposures. The belief that experiments involving realistically low challenge doses require large numbers of animals has so far prevented the development of alternatives to using high challenge doses.
Methods and Findings
Using statistical power analysis, we investigate how many animals would be needed to conduct preclinical trials using low virus challenge doses. We show that experimental designs in which animals are repeatedly challenged with low doses do not require unfeasibly large numbers of animals to assess vaccine or microbicide success.
Conclusion
Preclinical trials using repeated low-dose challenges represent a promising alternative approach to identify potential preventive interventions.
Trials of HIV vaccines in animals using repeated low- dose challenges of the virus are feasible and may be more true to life.
==== Body
Introduction
Worldwide approximately 40 million people are infected with HIV, and more than 3 million people died of AIDS last year alone [1]. Unfortunately, numerous obstacles to providing effective antiretroviral treatment to the majority of infected individuals in resource-poor countries exist. The development of a vaccine or other preventive biomedical intervention therefore bears the greatest hope to curb the rampant HIV epidemic [2].
Research on HIV vaccines and prevention relies strongly on preclinical studies in macaque models for the identification and evaluation of potential vaccines or prophylactic treatment strategies [3]. Initially, the goal was to use animal trials to screen for preventive interventions that induce sterilizing immunity (i.e., protection against infection) since this would clearly be the most effective way to contain the AIDS pandemic. Unfortunately, most of the vaccine approaches assessed to date in animal studies have failed to induce sterilizing immunity [4–7], although some prophylactic approaches were found to reduce susceptibility to infection [8–12]. As a result of this shortcoming, vaccine candidates are at present primarily examined with regard to their effects on set point viremia, disease progression, and their general immunogenicity, rather than with regard to the degree of protection against infection they confer. However, the inference as to the degree of sterilizing immunity from the level of immunogenicity is limited by our lack of knowledge about the mechanisms of protection against infection as such [13].
The inability of most vaccine candidates to induce protection against infection in animal studies may be due, at least in part, to unintended consequences of the design of the animal trials, rather than to problems inherent in the vaccination approaches themselves. In most animal studies that seek to test the efficacy of a given preventive intervention, very high challenge doses are used, typically of approximately 10–100 times the infectious dose at which 50% of the animals become infected (ID
50). The motivation for using such high challenge doses is mostly practical: the experimenter wants to ascertain infection success in unvaccinated/untreated animals, which can then be compared to the hopefully lower infection success in vaccinated/treated animals. There are, however, concerns with using high challenge doses. Firstly, the extremely high probability of infection in high-dose challenge studies conflicts with the low transmission rate of HIV per sex act [14–17]. Although it has been argued that transmission rates may be higher under some circumstances (such as during acute infection or when other infections of the genital tract are present) than the estimates obtained from discordant couple studies suggest (e.g., the recent study by Pilcher et al. [18]), transmission of HIV during one sex act surely does not occur with certainty. Secondly, protection against high-dose virus challenges may be more difficult to achieve because the use of high challenge doses makes stochastic extinctions that can play an important role in early control of the infection [19] very unlikely. Thus, standard high-dose challenge studies may rule out preventive intervention strategies that could protect against infections following “real life” exposures.
The problems of using high virus doses in animal studies can be illustrated by the discrepancy between the protection zidovudine (AZT) confers in animals and humans. Whereas macaques [20,21] and mice [22] were not protected from infection with high challenge doses by zidovudine (a relatively weak antiretroviral drug when used in monotherapy), clinical studies surprisingly showed that two-thirds of perinatal infections (i.e., mother-to-child transmissions during birth) can be prevented by zidovudine administration [23]. It is important to note that the use of zidovudine to prevent perinatal HIV infection is a biomedical intervention aiming to protect from infection, whereas zidovudine is most commonly used as a therapeutic agent after infection. This example suggests that there is a need for experimental designs that allow the assessment of the protection against infection with lower, and thus more realistic, challenge doses.
The belief that experiments involving realistically low challenge doses require unfeasibly large numbers of animals has prevented the development of low-dose challenge models. In this theoretical study, we show that, contrary to this widely held belief, low-dose challenge experiments can be designed such that they do not require large numbers of animals. Using statistical power analysis, we compare two experimental designs (see Figure 1): (i) a single low-dose challenge design in which each animal is challenged only once, and (ii) a repeated low-dose challenge design in which each animal is challenged until it is infected or a predetermined maximum number of challenges is reached. We find that the repeated low-dose challenge design does not require unfeasibly large numbers of animals.
Figure 1 Single and Repeated Low-Dose Challenge Designs
Figure shows designs for single (A) and repeated (B) low-dose challenge designs. Small arrows denote challenges, and white and red symbols denote uninfected and infected animals, respectively.
In the following, we are going to discuss the case of assessing whether a vaccine candidate induces sterilizing immunity. All the considerations in this article, however, apply equally to other preventive interventions, such as microbicides.
Methods
To assess the quality of the single and the repeated low-dose challenge designs, we conducted a statistical power analysis. The statistical power of an experimental design is defined as the probability that an effective vaccine or treatment is correctly determined to be effective. This analysis consists of simulating the experiments, evaluating them, and then repeating this procedure thousands of times to estimate the statistical power of a given experimental design.
Simulation of Single Low-Dose Challenge Experiments
In our simulations of the single low-dose challenge experiments, we assume that we have n unvaccinated control animals and n vaccinated animals.
In the control group, we simulate single challenges of each animal with the ID
50 by performing n Bernoulli trials with a probability of success of pc = 0.5. The probability of success corresponds to the probability with which an animal becomes infected after a single challenge. (By assuming the same probability pc for each animal, we ignore potential between-animal variation of the susceptibility to infection. This assumption will be relaxed below.) The results of these trials can be written as a vector x
c, the entries of which were either zero (uninfected) or one (infected):
By summing over the elements of x
c, we obtain the number of infected animals in the control group, ιc:
In the vaccinated group, we simulate single challenges with the ID
50 similarly to the control group by performing Bernoulli trials. However, we assume that, because of vaccination, the probability of infection (or success) in the vaccinated group, pv, is lower than that in the control animals, pc. The relation of pv to the effect of the vaccine on the susceptibility of the host, VES, is given by:
The results of these Bernoulli trials can again be written as a vector x
v, and summing the elements of x
v we obtain the number of infected animals in the vaccinated group, ιv.
The outcome of the simulated experiment can then be summarized in a contingency table as shown in Table 1. On this contingency table, we perform a standard one-tailed Fisher's exact test [24] to assess whether the fraction of infected animals in the vaccinated group is significantly lower than that in the control group.
Table 1 Contingency Table of a Single Low-Dose Challenge Experiment
Simulation of Repeated Low-Dose Challenge Experiments
In our simulations of the repeated low-dose challenge experiments, we once more assume that we have n unvaccinated control animals and n vaccinated animals.
We again simulate challenges of each control animal with the ID
50 by performing Bernoulli trials with a probability of success of pc = 0.5. Unlike in the simulations of the single low-dose challenge experiments, however, we now repeatedly challenge each animal until it is infected or until a maximum number of challenges, Cmax, has been performed. We assume that the probability of infection pc is independent of how often an animal has been challenged before. The results of these repeated Bernoulli trials can be written as two vectors, y
c, which contains the number of challenges that have been performed on each animal:
and s
c, which contains information on whether a given animal is uninfected (zero) or infected (one):
By summing over y
c, we obtain the total number of challenges performed in the control group, ηc:
And, by summing over s
c, we obtain the number of infected animals in the control group, ιc:
To simulate repeated low-dose challenges in the vaccinated group, we perform repeated Bernoulli trials with a probability of success pv. For a given vaccine efficacy VES,
pv is determined by equation 3. Analogously to the control group, the results of these repeated Bernoulli trials can be written as two vectors, y
v and s
v, and summing the elements of these two vectors yields the total number of challenges performed in the vaccinated group, ηv, and, the number of infected animals in the vaccinated group, ιv.
As in the case of the single low-dose challenge design, the outcome of the simulated experiment can be summarized in a contingency table (Table 2). To assess whether the fraction of infected animals in the vaccinated group is significantly lower than that in the control group, we again perform a one-tailed Fisher's exact test [24]. In general, the number of challenges, ηc and ηv, is larger than the number of animals per group, n. This increase of numbers in the contingency table leads to increased statistical power of the repeated low-dose challenge design. To analyze the outcome of the simulated repeated low-dose challenge experiments, we chose to use Fisher's exact test rather than a more obvious Cox proportional hazards model because the latter depends on large sample asymptotics while we were interested in cases of small numbers of experimental animals.
Table 2 Contingency Table of a Repeated Low-Dose Challenge Experiment
Heterogeneity in Infection Probabilities
In our mml:mathematical description of challenge experiments, we have assumed that animals within each group have equal infection probabilities—pc and pv, for the control and vaccinated groups, respectively. To simulate potential animal-to-animal variation in susceptibility to infection, we relaxed this assumption and assigned individual infection probabilities to each animal.
The individual infection probabilities are drawn from a β-distribution, which is often used as a prior distribution for binomial proportions. The β-distribution has two shape parameters, α and β. Its probability density is given by
and its mean and variance are
We assume that μ = pc in the control group and μ = pv = (1 − VES)pc in the vaccinated group. Further, we assume that the coefficients of variation, CV = σ/μ, of the distributions in the two groups are equal. With these assumptions, we can rewrite the two shape parameters of the β-distribution, α and β, in terms of the infection probability, p, and the coefficient of variation, CV:
Hereby, p = pc for the control group and p = pv = (1 − VES)pc for the vaccinated group.
To incorporate potential heterogeneity in susceptibility into the virtual low-dose challenge experiments, we replaced the probability of success in the Bernoulli trials (see above) with the individual infection probabilities.
Power Analysis
To calculate the statistical power of the single and the repeated low-dose challenge designs, we performed 100,000 such simulated experiments for a given number, n, of animals per group. The statistical power can be estimated as the fraction of simulated experiments in which the vaccine is found to be significantly efficacious (significance level α = 0.05). We estimated the statistical power for the number of animals per group, n, ranging from one to 20, and for vaccine efficacies VES = 0.67, 0.8, and 0.9. The power analysis outlined above was implemented in the R Language of Statistical Computing [25]. An R-script that performs the power analysis presented here is provided as Protocol S1.
For large numbers of animals per group, n, the statistical power can be approximated using asymptotic theory. For the single low-dose challenge design the power is approximately (e.g., [26], p. 240):
Hereby, Φ denotes the cumulative normal distribution,
and z
α is the standard normal deviate associated with the one-tailed probability α (the significance level). Furthermore, pc and pv denote the infection probabilities of animals in the control and vaccinated groups, respectively, and n the number of animals per group. The term 1/n in the numerator is the continuity correction [27,28].
For the repeated low-dose challenge design, the number of challenges is not the same as the number of animals, n, but is a random variable. The number of challenges for each individual is geometrically distributed with a maximum of Cmax. The expected number of challenges in the control group, , and the vaccinated group, are
and
Substituting the expected number of challenges for the actual number, we can approximate the statistical power of the repeated low-dose challenge design as
Hereby, γ = (1/〈ηc〉 + 1/〈ηv〉)/2 is the continuity correction. For Cmax = 1, equation 17 reduces to equation 13. Because the approximation in equation 17 involves the substitution of a random variable with its expectation, it is less accurate than the approximation for the power of the single low-dose challenge design in equation 13. The R-script provided as Protocol S1 also contains a function that calculates the statistical power using equation 17.
Results
Single Low-Dose Challenge Design Requires Large Numbers of Animals
How would we measure protection against infection in a low-dose challenge model? The most straight-forward design would involve a large number of hosts, some vaccinated and some unvaccinated. After challenge with a low dose, one would determine the fraction of infected hosts in vaccinated and unvaccinated groups, and assess whether there is a statistically significant difference in the fractions (see Figure 1A).
To assess how many animals would be required in a single low-dose challenge experiment, we performed a statistical power analysis (see Methods). The statistical power of an experimental design is defined as the probability that, in an experiment with an effective vaccine, the vaccine is correctly determined to be effective. Obviously the power depends on the efficacy of the vaccine (which is called the “effect size” in the context of power analysis) and the number of host animals used in the experiment. In the power analysis we performed, we assumed that we had equal numbers of unvaccinated and vaccinated animals, and that all animals within a group were equally susceptible to infection. Lastly, we assumed that the vaccine was “leaky” [29,30], i.e., that the susceptibility of vaccinated animals was by a constant factor lower than the susceptibility of the unvaccinated control animals.
In virtual experiments, we then challenged each (virtual) animal once with a challenge dose of one ID
50, the dose at which on average 50% of the unvaccinated animals become infected after a single challenge. Using a one-sided Fisher's exact test, we tested whether the fraction of infected animals in the vaccinated group was significantly lower than in the control group. Performing 100,000 such virtual experiments for a given number n animals per group, we estimated the statistical power as the fraction of virtual experiments that yielded significant results (significance level α = 0.05).
The result of this power analysis is shown by the green curves in Figure 2. We calculated the power for vaccine efficacies of 67%, 80%, and 90%. We found that, even for the highest vaccine efficacy of 90%, the single low-dose challenge design required more than 20 animals per group to reach a statistical power of 95%. Thus, the single low-dose challenge design is not feasible, or at least not practical, to assess the efficacy of a vaccine or other preventive interventions in animals.
Figure 2 Power Analysis for the Repeated Low-Dose Challenge Design and the Single Low-Dose Challenge Design
In our virtual experiments, we set the challenge dose equal to the ID
50, and assumed that the vaccine efficacy was 67% (dotted lines), 80% (dashed lines), or 90% (solid lines). The graph shows the statistical power of the repeated low-dose challenge design (black lines) and the single low-dose challenge design (green lines) for a given number of animals per group as determined from 100,000 virtual experiments. If the vaccine is 90% effective, the statistical power of the repeated low-dose challenge design is higher than 95% with only five animals per group, as compared to only 15% for the single low-dose challenge design.
Repeated Low-Dose Challenge Design Does Not Require Large Numbers of Animals
We propose an alternative design involving repeated challenges of individual animals with low doses, which circumvents the disadvantage of the single low-dose challenge design that large numbers of host individuals are required. Repeated challenges effectively “recycle” host animals, thus increasing the statistical power of the experiment. In addition to increasing the statistical power of the experimental design, repeated challenges recapitulate much more realistically the circumstances of human exposure than single challenges. In this alternative design, the efficacy of a vaccine can be estimated by measuring the difference in the number of challenges needed to infect vaccinated versus unvaccinated hosts (see Figure 1B).
To show that this alternative design does not require unfeasibly large numbers of animals, we performed a statistical power analysis (see Methods). As for the single low-dose design, we assumed that we had equal numbers of unvaccinated and vaccinated animals, and that all animals within a group were equally susceptible to infection. We further made the important assumption that the susceptibility of an individual animal was independent of how often the animal was unsuccessfully challenged previously. This assumption is commonly adopted in statistical models that are used to estimate the transmission rate of HIV [14–17]. By making this assumption, we ignored that an unsuccessful challenge may induce some degree of immunity against subsequent challenges. We would like to emphasize, however, that this assumption is not crucial for our argument, unless the degree of induced immunity is very high. Lastly, we again assumed that the vaccine was leaky [29,30].
In virtual experiments, we then challenged the (virtual) animals repeatedly with a challenge dose of one ID
50. We allowed for a maximum number of 20 challenges of each individual animal. Table 3 shows the outcome of one such virtual experiment. We analyzed the outcome of the virtual experiments with a one-tailed Fisher's exact test (see Methods). We again estimated the statistical power by performing 100,000 such virtual experiments for a given number n animals per group.
Table 3 Outcome of One Virtual Repeated Low-Dose Challenge Experiment
Figure 2 shows the statistical power of the repeated low-dose challenge design as a function of the number of animals per group for varying vaccine efficacies (black lines), and compares it to the statistical power of the single low-dose challenge design (green lines). The statistical power achieved with the repeated low-dose challenge design is generally higher than that achieved with the single low-dose challenge design. If the vaccine is 90% effective (VES = 0.9), i.e., it reduces the susceptibility by a factor of ten, the number of animals per group could be as low as five to achieve more than 95% statistical power. In contrast, in single low-dose challenge experiments with the same number of animals per group the statistical power is only 15%. Thus, repeated low-dose challenge experiments are expected to require far fewer animals than single low-dose challenge experiments.
How Often Should Virus Challenges Be Repeated?
To investigate how the maximum number of challenges affected the statistical power, we plotted the power against Cmax for trials involving six and 12 animals per group (Figure 3). We found that the power increases with Cmax, but for high Cmax the returns diminished considerably. The lower the number of animals per group, n, the higher the maximum number of challenges, Cmax, for which the power effectively saturated. Even for low numbers of animals per group, n, however, the maximum number of challenges, Cmax, needed to unfold the full potential of the repeated low-dose challenge design was in a feasible range.
Figure 3 Impact of the Maximum Number of Challenges, Cmax, on the Statistical Power
For this plot we assumed trials with vaccine efficacies of VES = 0.67 (dotted line), VES = 0.8 (dashed line), and VES = 0.9 (solid line). In (A) we calculated the statistical power for six animals per group, n = 6, and in (B) for 12 animals per group, n = 12.
Impact of Animal-to-Animal Variation in Susceptibility
To study how potential heterogeneity in susceptibility affected the power of low-dose challenge trials, we simulated experiments in which each animal was assigned an individual infection probability (see Methods). In these simulations, the degree of heterogeneity was measured by the coefficient of variation, CV, of the susceptibility distributions. Figure 4A shows susceptibility distributions for three different values of CV.
Figure 4 Impact of Heterogeneity in Susceptibility on the Statistical Power
(A) Susceptibility distributions for different levels of heterogeneity, measured by the coefficient of variation, CV, of the susceptibility distribution. The vaccine is assumed to be 80% effective, VES = 0.8.
(B) The statistical power depends on the coefficient of variation, CV, for the repeated low-dose challenge design (black lines) and the single low-dose challenge design (green lines). For these plots we assumed trials with six and 12 animals per group and vaccine efficacies of VES = 0.67 (dotted lines), VES = 0.8 (dashed lines), and VES = 0.9 (solid lines).
We extended our power analysis by considering the impact of the heterogeneity parameter CV on the statistical power (Figure 4B). We found that the statistical power of the single low-dose challenge design was almost unaffected by animal-to-animal variation in infection probability, whereas, for the repeated low-dose challenge design, the power decreased with increasing heterogeneity. Importantly, however, the power did not decrease linearly with heterogeneity: it was sufficiently stable in the range 0 < CV < 0.3 and dropped mainly for CV > 0.3. Thus, over a wide range of potential animal-to-animal variation in susceptibility, low-dose challenge designs are sufficiently powered, and the power of the repeated low-dose experiments is superior to that of single low-dose challenge experiments.
Discussion
Preclinical studies assessing the efficacy of potential vaccines, microbicides, or systemic chemoprophylaxis are usually conducted with very high virus challenge doses, which result in infection with certainty. Since these high challenge doses do not reflect the low probability of HIV transmission in humans, vaccines or prophylactic treatment strategies that are effective against “real life” exposures may go undetected in high-dose challenge experiments. For example, zidovudine was found to prevent a large fraction of perinatal HIV infections [23], even though studies in animal models, conducted with high challenge doses, could not establish any protection against infection by zidovudine [20–22].
In this paper, we investigated how efficacy trials of vaccines and preventive treatment could be conducted with low challenge doses in animal models. We showed that the repeated low-dose challenge design is expected to require far fewer experimental animals than commonly believed. It may therefore be feasible to conduct trials with low challenge doses, which more realistically simulate exposures of humans to HIV, allowing us to more directly and sensitively assess vaccine or treatment efficacy than with high-dose challenge experiments.
Owing to the concerns with high challenge doses, several research groups, including our own, have started to develop low-dose challenge models [31–34]. In these preliminary studies, infection could be achieved by challenging macaques intra-rectally [31], intra-vaginally [32,34], or orally [33].
Since adopting low-dose challenge approaches has far-reaching consequences for the design of efficacy trials of vaccines or preventive treatment in animal models, we would like to discuss how some important aspects of trial design, such as transient infections, the challenge schedule, the route of infection, and the phenotype and dose of the challenge strain, should be dealt with and could be optimized.
Using virus challenge doses that do not give rise to infection with certainty, one has to carefully define what one means by successful infection. This question is of particular importance in the repeated low-dose challenge design, because the efficacy of a preventive intervention is estimated on the basis of the number of challenges needed to infect an individual animal. Low-dose challenges have been observed to give rise to transiently detectable viremia [32–34]. Since transient infection is much more likely to lead to immunization [35], thus leading to lower probabilities of infection in subsequent challenges, we suggest considering transient viremia as successful infection and not to re-challenge animals that were transiently infected.
The time interval between challenges is also an essential parameter in the design of repeated low-dose challenge experiments. In the four ongoing repeated low-dose challenge studies [31–34], different approaches have been taken, with time intervals ranging from hours to a week. There may be logistical reasons for choosing short time intervals between challenges, but from a statistical standpoint, the time intervals should be large enough to allow the identification of the challenge that gives rise to infection. Otherwise, the statistical power of the experimental design will be suboptimal and a beneficial effect of the vaccine candidate may be missed.
In parallel to using more realistic, lower challenge doses, other crucial parameters of the experimental infection process, such as the route of transmission and the coreceptor usage of the challenge virus, should also be chosen to be as realistic as possible. Thus, we propose infecting intra-vaginally or intra-rectally in experiments that aim to assess a vaccine or prophylactic treatment against sexual transmission of HIV. Further, we suggest using challenge viruses that utilize CCR5 as coreceptor, such as for example SHIV-SF162P [36], rather than the standard strain SHIV89.6P, which has been found to use mainly CXCR4 [37,38]. These more realistic choices of the route of infection and coreceptor usage will permit the assessment of the efficacy of the preventive intervention in a setting that more accurately reflects HIV exposures of humans, and will enable us to carefully investigate the processes that give rise to infection.
The challenge dose in a low-dose challenge study is another parameter of crucial importance. Although the most realistic choice would be a challenge dose that gives rise to infection with a probability of approximately 0.0005–0.10 [14–17], such extremely low doses would require unfeasibly large numbers of repeated challenges per animal. Moreover, there is substantial variation in transmission rates due to differences in factors such as virus load or the presence of other infections of the genital tract [15–18], and theoretical studies suggest that preventing the transmission events that occur with higher probability would have a disproportionately large effect on controlling the epidemic [39]. To maximize their epidemiological relevance, low-dose challenge experiments should therefore involve challenge doses that reflect transmission probabilities at the upper end of the spectrum. As a compromise between the practicality of high doses and the sensitivity associated with realistically low doses, we propose the ID
50. The ID
50 can be estimated using well-established nonparametric methods like Spearman-Kärber [40] or single-parameter methods [41], and there is software available, such as a freely distributed package called ID50 developed by John Spouge (http://www.ncbi.nlm.nih.gov/CBBresearch/Spouge/Virology/, which allows an automated estimation of the ID
50 from data generated in titration experiments.
The inability to detect sterilizing immunity in high-dose challenge experiments led to a shift of focus towards indirect effects of vaccine candidates on the pathogenicity of the infection and the infectiousness of the vaccinee. This shift of focus required the development of novel statistical models that allowed the estimation of these indirect effects [42,43]. Will the estimation of vaccine efficacy in repeated low-dose challenge studies also require the development of novel statistical techniques? The answer to this question depends on how much the realities of the infection process deviate from our idealized model. There are three potential deviations. First, we assumed in large parts of this study that the susceptibilities to infection were equal for all animals within each group. This is almost certainly not the case. Although we have shown that low-dose challenge experiments are sufficiently powered even if there is substantial animal-to-animal variation in susceptibility, we did not develop the statistical techniques that would allow the estimation of this variation. The extent of animal-to-animal variation in susceptibility can, in principle, be estimated, but this will probably require larger numbers of animals than the estimation of vaccine efficacy. Second, the vaccine may affect the susceptibility of individual animals differently. While we assumed in the present study that the vaccine is leaky, i.e., that the susceptibility is reduced by a constant factor in each animal, other modes of action of a vaccine are possible. In particular, some animals could be completely protected by vaccination, while others may remain completely susceptible. This mode of action is referred to as all-or-none [29,30]. Statistical methods based on maximum likelihood approaches exist that allow the determination of the mode of action of a given vaccine. However, these methods are based on large sample asymptotics, and exact methods will have to be developed to analyze the outcome of low-dose challenge experiments that involve small numbers of animals. Last, it will have to be determined whether the probability of infection changes with the number of challenges performed in a given animal, or, to put it differently, whether the animal has a “memory” of previous challenges. In our analysis, we assumed that the susceptibility of an animal did not change from challenge to challenge. If the probability of infection changes significantly with the number of challenges, however, the development of novel statistical models that take such changes into account will be necessary to adequately estimate vaccine efficacy.
In addition to the potential to assess the vaccine or microbicide efficacy more sensitively and in a more realistic setting, a low-dose challenge approach may enable us to answer questions that cannot even be asked in high-dose challenge models. Some of the most relevant of these questions relate to the effect of challenges that do not lead to infection. If a low-dose challenge does not give rise to infection, where was the virus blocked? Did the virus fail to establish an infection at all? Or did it replicate transiently, but was cleared by the host's immunity? And, very importantly, is an unsuccessfully challenged animal partially immunized against further challenges, or, alternatively, do unsuccessful challenges facilitate future infection by “seeding” animals with defective proviruses that may recombine with complementing viruses upon subsequent exposures [44]?
The answers to these questions would greatly enhance our understanding of HIV transmission and pathogenesis, and thus would provide further guidance toward an effective vaccine or microbicide. Furthermore, by assessing the protection against infection directly, we may be able to discern the specific types and levels of vaccine-induced cellular and humoral immune responses associated with sterilizing immunity [13]. This would provide important benchmarks by which to judge new vaccine candidates, and could also allow retrospective analysis of vaccine candidates evaluated earlier in high-dose challenge studies.
In conclusion, the repeated low-dose challenge approach may enable us to assess the potential efficacy of vaccines and prophylactic treatment strategies more realistically, and more sensitively than the standard high-dose challenge approach. The increased sensitivity may allow us to more rapidly identify interventions that significantly reduce the transmission of low-dose infections that characterize the natural spread of HIV.
Supporting Information
Protocol S1 R-Function for the Calculation of the Statistical Power of Low-Dose Challenge Experiments
(8 KB TXT).
Click here for additional data file.
Patient Summary
Background
Before trials of medicines or vaccines are done in humans, most are tested in animals. There are many controversies about these animal trials, including whether they mimic the human disease accurately. In testing vaccines for HIV, animals are mostly given high doses of the virus, whereas in real life people are often repeatedly exposed to small amounts of the virus. No vaccine that has been tested against HIV prevents infection in animals. It is possible that some of this lack of success may be due to the design of the vaccine trials rather than the vaccine itself.
What Did the Authors Do?
They wanted to look at experimental designs that allowed assessment of protection against infection with lower, and thus more realistic, doses of virus. Previously, researchers had suggested that many animals would be needed for this type of study. The authors wanted to see whether this was correct. They developed a model to test how well single and multiple low-dose experiments performed. They did this by simulating the experiments with doses of virus, assessing the results, and then repeating this procedure 100,000 times to estimate how valid a given experimental design was.
Their modeling showed that by repeatedly giving animals low doses of virus, it was possible to use a smaller number of animals than was needed for trials with a single low dose.
What Do These Results Mean?
It may be possible to use these results to plan trials of vaccines in animals that mimic more closely the way that humans are exposed to HIV, and hence the results may be more reliable for human disease.
Where Can I Get More Information?
MedlinePlus has a great deal of information on HIV:
http://www.nlm.nih.gov/medlineplus/aids.html
The Body has information targeted to both patients and health professionals:
http://www.thebody.com/index.shtml
We thank Rustom Antia, Steven Self, Mark Tanaka, and Andrew Yates for discussion. RRR was supported by the Deutsche Forschungsgemeinschaft grant number Re 1618/1–2 and the National Institutes of Health (NIH) grant AI-49334. The support of the NIH National Institute of Allergy and Infectious Disease grant 1 R21 AI54260 is gratefully acknowledged. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Citation: Regoes RR, Longini IM Jr, Feinberg MB, Staprans SI (2005) Preclinical assessment of HIV vaccines and microbicides by repeated low-dose virus challenges. PLoS Med 2(8): e249.
Abbreviations
ID50infectious dose at which 50% of the animals become infected
==== Refs
References
Joint United Nations Programme on HIV/AIDS AIDS epidemic update 2004 2004 December Geneva Joint United Nations Programme on HIV/AIDS Available: http://www.unaids.org/wad2004/report.html . Accessed 24 June 2005
Garber DA Silvestri G Feinberg MB Prospects for an AIDS vaccine: Three big questions, no easy answers Lancet 2004 4 397 413 15219551
Staprans SI Feinberg MB The roles of nonhuman primates in the preclinical evaluation of candidate AIDS vaccines Expert Rev Vaccines 2004 3 S5 32 15285703
Amara RR Villinger F Altman JD Lydy SL O'Neil SP Control of a mucosal challenge and prevention of AIDS by a multiprotein DNA/MVA vaccine Science 2001 292 69 74 11393868
Barouch DH Santra S Schmitz JE Kuroda MJ Fu TM Control of viremia and prevention of clinical AIDS in rhesus monkeys by cytokine-augmented DNA vaccination Science 2000 290 486 492 11039923
Rose NF Marx PA Luckay A Nixon DF Moretto WJ An effective AIDS vaccine based on live attenuated vesicular stomatitis virus recombinants Cell 2001 106 539 549 11551502
Shiver JW Fu TM Chen L Casimiro DR Davies ME Replication-incompetent adenoviral vaccine vector elicits effective anti-immunodeficiency-virus immunity Nature 2002 415 331 335 11797011
Lederman MM Veazey RS Offord R Mosier DE Dufour J Prevention of vaginal SHIV transmission in rhesus macaques through inhibition of CCR5 Science 2004 306 485 487 15486300
Tsai CC Emau P Follis KE Beck TW Benveniste RE Effectiveness of postinoculation (r)-9-(2-phosphonylmethoxypropyl) adenine treatment for prevention of persistent simian immunodeficiency virus SIVmne infection depends critically on timing of initiation and duration of treatment J Virol 1998 72 4265 4273 9557716
Tsai CC Follis KE Sabo A Beck TW Grant RF Prevention of SIV infection in macaques by (R)-9-(2-phosphonylmethoxypropyl)adenine Science 1995 270 1197 1199 7502044
Veazey RS Klasse PJ Ketas TJ Reeves JD Piatak M Use of a small molecule CCR5 inhibitor in macaques to treat simian immunodeficiency virus infection or prevent simian-human immunodeficiency virus infection J Exp Med 2003 198 1551 1562 14623909
Veazey RS Shattock RJ Pope M Kirijan JC Jones J Prevention of virus transmission to macaque monkeys by a vaginally applied monoclonal antibody to HIV-1 gp120 Nat Med 2003 9 343 346 12579198
Johnson RP Mechanisms of protection against simian immunodeficiency virus infection Vaccine 2002 20 1985 1987 11983259
Brookmeyer R Gail MH AIDS epidemiology: A quantitative approach 1994 Oxford Oxford University Press 354
Gray RH Wawer MJ Brookmeyer R Sewankambo NK Serwadda D Probability of HIV-1 transmission per coital act in monogamous, heterosexual, HIV-1-discordant couples in Rakai, Uganda Lancet 2001 357 1149 1153 11323041
Mastro TD Kitayaporn D HIV type 1 transmission probabilities: Estimates from epidemiological studies AIDS Res Hum Retroviruses 1998 14 S223 S227 9814947
Wawer MJ Gray RH Sewankambo NK Serwadda D Li X Rates of HIV-1 transmission per coital act, by stage of HIV-1 infection, in Rakai, Uganda J Infect Dis 2005 191 1403 1409 15809897
Pilcher CD Tien HC Eron JJ Vernazza PL Leu SY Brief but efficient: Acute HIV infection and the sexual transmission of HIV J Infect Dis 2004 189 1785 1792 15122514
Wick D Self SG Early HIV infection in vivo: Branching-process model for studying timing of immune responses and drug therapy Math Biosci 2000 165 115 134 10854614
Fazely F Haseltine WA Rodger RF Ruprecht RM Postexposure chemoprophylaxis with ZDV or ZDV combined with interferon-α—Failure after inoculating rhesus-monkeys with a high-dose of SIV retrovirology J Acquir Immune Defic Syndr 1991 4 1093 1097 1753337
Legrand R Clayette P Noack O Vaslin B Theodoro F An animal-model for antilentiviral therapy—Effect of zidovudine on viral load during acute infection after exposure of macaques to simian immunodeficiency virus AIDS Res Hum Retroviruses 1994 10 1279 1287 7848683
Ruprecht RM Bronson R Chemoprevention of retroviral infection—Success is determined by virus inoculum strength and cellular-immunity DNA Cell Biol 1994 13 59 66 8286040
Mofenson LM Reducing the risk of perinatal HIV-1 transmission with zidovudine: Results and implications of AIDS Clinical Trials Group protocol 076 Acta Paediatr Suppl 1997 421 89 96 9240866
Fisher RA The logic of inductive inference J R Stat Soc A 1935 98 39 82
R Development Core Team The R project for statistical computing 2005 Vienna R Foundation for Statistical Computing Available: http://www.R-project.org . Accessed 24 June 2005
Agresti A Categorical data analysis 1990 New York Wiley 558
Fleiss JL Statistical methods for rates and proportions, 2nd ed 1981 New York Wiley 321
Newcombe RG Interval estimation for the difference between independent proportions: Comparison of eleven methods Stat Med 1998 17 873 890 9595617
Smith PG Rodrigues LC Fine PE Assessment of the protective efficacy of vaccines against common diseases using case-control and cohort studies Int J Epidemiol 1984 13 87 93 6698708
Halloran ME Longini IM Struchiner CJ Design and interpretation of vaccine field studies Epidemiol Rev 1999 21 73 88 10520474
McDermott AB Mitchen J Piaskowski S De Souza I Yant LJ Repeated low-dose mucosal simian immunodeficiency virus SIVmac239 challenge results in the same viral and immunological kinetics as high-dose challenge: A model for the evaluation of vaccine efficacy in nonhuman primates J Virol 2004 78 3140 3144 14990733
Otten RA Adams DR Kim CN Jackson E Pullium JK Multiple vaginal exposures to low doses of R5 simian-human immunodeficiency virus: Strategy to study HIV preclinical interventions in nonhuman primates J Infect Dis 2005 191 164 173 15609225
Van Rompay KK Abel K Lawson JR Singh RP Schmidt KA Attenuated poxvirus-based simian immunodeficiency virus (SIV) vaccines given in infancy partially protect infant and juvenile macaques against repeated oral challenge with virulent SIV J Acquir Immune Defic Syndr 2005 38 124 134 15671796
Ma ZM Abel K Rourke T Wang Y Miller CJ A period of transient viremia and occult infection precedes persistent viremia and antiviral immune responses during multiple low-dose intravaginal simian immunodeficiency virus inoculations J Virol 2004 78 14048 14052 15564513
McChesney MB Collins JR Lu D Lu X Torten J Occult systemic infection and persistent simian immunodeficiency virus (SIV)-specific CD4(+)-T-cell proliferative responses in rhesus macaques that were transiently viremic after intravaginal inoculation of SIV J Virol 1998 72 10029 10035 9811741
Harouse JM Gettie A Eshetu T Tan RC Bohm R Mucosal transmission and induction of simian AIDS by CCR5-specific simian/human immunodeficiency virus SHIV(SF162P3) J Virol 2001 75 1990 1995 11160699
Zhang YJ Lou B Lal RB Gettie A Marx PA Use of inhibitors to evaluate coreceptor usage by simian and simian/human immunodeficiency viruses and human immunodeficiency virus type 2 in primary cells J Virol 2000 74 6893 6910 10888629
Feinberg MB Moore JP AIDS vaccine models: Challenging challenge viruses Nat Med 2002 8 207 210 11875482
May RM Anderson RM The transmission dynamics of human immunodeficiency virus (HIV) Philos Trans R Soc Lond B Biol Sci 1988 321 565 607 2907158
Miller RG Nonparametric estimators of the mean tolerance in bioassay Biometrika 1973 60 535 542
Spouge JL Statistical-analysis of sparse infection data and its implications for retroviral treatment trials in primates Proc Natl Acad Sci U S A 1992 89 7581 7585 1323844
Longini IM Hudgens MG Halloran ME Sagatelian K A Markov model for measuring vaccine efficacy for both susceptibility to infection and reduction in infectiousness for prophylactic HIV vaccines Stat Med 1999 18 53 68 9990692
Gilbert PB DeGruttola VG Hudgens MG Self SG Hammer SM What constitutes efficacy for a human immunodeficiency virus vaccine that ameliorates viremia: Issues involving surrogate end points in phase 3 trials J Infect Dis 2003 188 179 193 12854072
Kim EY Busch M Abel K Fritts L Bustamante P Retroviral recombination in vivo: Viral replication patterns and genetic structure of simian immunodeficiency virus (SIV) populations in rhesus macaques after simultaneous or sequential intravaginal inoculation with SIVmac239Δvpx/Δvpr and SIVmac239Δnef J Virol 2005 79 4886 4895 15795274
| 16018721 | PMC1176242 | CC BY | 2021-01-05 10:40:30 | no | PLoS Med. 2005 Aug 19; 2(8):e249 | utf-8 | PLoS Med | 2,005 | 10.1371/journal.pmed.0020249 | oa_comm |
==== Front
PLoS MedPLoS MedpmedplosmedPLoS Medicine1549-12771549-1676Public Library of Science San Francisco, USA 10.1371/journal.pmed.0020257SynopsisEpidemiology/Public HealthHealth PolicyHIV/AIDSMedical EthicsMedical LawEthicsHealth PolicyHIV Infection/AIDSHuman rightsQuality of health careHow Do Nigeria's Health-Care Personnel Treat Patients with HIV/AIDS? Synopsis8 2005 19 8 2005 2 8 e257Copyright: © 2005 Public Library of Science.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.
Assessment of Discriminatory Health Worker Attitudes and Practices toward Patients with HIV/AIDS in Nigeria
The Discriminatory Attitudes of Health Workers against People Living with HIV
Achieving Gold Standards in Ethics and Human Rights in Medical Practice
==== Body
People living with HIV/AIDS (PLWA) face many forms of stigma and discrimination. This is the case in whichever country they may live, as has been shown in a number of previous research studies. In addition to experiencing unfair treatment in their families, communities, and places of work, PLWA may encounter discrimination from health-care professionals. This can interfere with effective prevention and treatment. Discriminatory practices in the health-care sector may also appear to legitimize other forms of discrimination against PLWA.
Vincent Iacopino and colleagues from the organization Physicians for Human Rights, in collaboration with researchers from Policy Project–Nigeria and the Center for the Right to Health (also in Nigeria) investigated the problem in Nigeria. With a population of roughly 130 million, Nigeria is home to one in 11 of the 40 million PLWA worldwide. Around 6% of adult Nigerians are thought to be HIV-positive, and there will be an estimated 310,000 AIDS deaths this year. The indications are that infection rates will increase. Until now, little has been known about the nature and extent of discrimination against patients with HIV/AIDS in Nigeria.
Trained interviewers conducted a cross-sectional questionnaire survey of 1,021 Nigerian health-care professionals in 111 health-care facilities in four of Nigeria's 36 states. Those sampled were 324 physicians, 541 nurses, and 133 midwives, and 23 health-care workers of unknown profession. Fifty-four percent of them worked in public tertiary care facilities. Many of the survey's results are worrying. Nine percent of professionals reported refusing to care for a patient with HIV/AIDS, and 9% said they had refused a patient with HIV/AIDS admission to a hospital. Fifty-nine percent agreed that PLWA should be on a separate ward, and 40% believed a person's HIV status could be determined by their appearance. Ninety-one percent agreed that staff should be informed when a patient was HIV-positive in order to protect themselves. Forty percent believed health-care professionals with HIV/AIDS should not be allowed to work in any area of health-care requiring patient contact. Twenty percent agreed that many with HIV/AIDS had behaved immorally and deserved their infection. Eight percent felt that treating someone with HIV/AIDS was a waste of resources.
Providers who reported working in facilities that did not always practice universal precautions against HIV transmission were more likely to favor restrictive policies towards PLWA. In general, basic materials needed for treatment and prevention of HIV infection were not sufficiently available. Providers who reported less adequate training in HIV/AIDS treatment and in ethics were more likely to report negative attitudes towards patients with HIV/AIDS. There was no consistent pattern of differences in negative attitudes and practices across the different professions surveyed.
Training of interviewers for the study included 5 days of classroom teaching and role-playing
The researchers concluded that, while most health-care professionals surveyed reported being in compliance with their ethical obligations, discriminatory behavior and attitudes towards patients with HIV/AIDS existed among a significant proportion. Inadequate education about HIV/AIDS and a lack of protective and treatment materials appear to favor these practices and attitudes. The findings of the study, in just four states, cannot be generalized to Nigeria as a whole and, although sampled systematically, it is possible that sampled facilities and health-care professionals may differ significantly from those that were not sampled in the study states. Concerns over a perceived lack of privacy in the interviews or about job status may have resulted in an underreporting of discriminatory behavior and/or an overreporting of “correct” practices or attitudes. The authors note that the health-care system in Nigeria is underfunded and suffers from fundamental problems, including material scarcity and inadequacies in infrastructure, both of which may contribute to discriminatory behavior. They call for targeted education of health-care professionals and provision of adequate resources to health-care facilities, and for the introduction and enforcement of anti-discrimination policies.
| 0 | PMC1176243 | CC0 | 2021-01-05 10:40:31 | no | PLoS Med. 2005 Aug 19; 2(8):e257 | utf-8 | PLoS Med | 2,005 | 10.1371/journal.pmed.0020257 | oa_comm |
==== Front
PLoS MedPLoS MedpmedplosmedPLoS Medicine1549-12771549-1676Public Library of Science San Francisco, USA 10.1371/journal.pmed.0020258SynopsisBioinformatics/Computational BiologyImmunologyInfectious DiseasesMicrobiologyVirologyEpidemiology/Public HealthHIV/AIDSStatisticsVaccinesHIV Infection/AIDSImmunology and allergyInfectious DiseasesModeling HIV Vaccine Strategy in Animals Synopsis8 2005 19 7 2005 2 8 e258Copyright: © 2005 Public Library of Science.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.
Preclinical Assessment of HIV Vaccines and Microbicides by Repeated Low-Dose Virus Challenges
==== Body
Animal models can play an essential role in guiding preclinical vaccine development, including in studies of preclinical vaccine safety, vaccine toxicity, and vaccine immunogenicity. Appropriate pathogen challenge models can also provide the opportunity to perform preclinical tests of vaccine efficacy. Preclinical tests of HIV vaccine efficacy are usually performed by exposing macaques to simian immunodeficiency virus (SIV), a virus that is closely related to HIV. However, the viral inoculum sizes used to infect macaques with SIV vastly exceed the amounts of HIV that humans are exposed to during a given exposure. Typically animals are exposed to 10–100 times the infectious dose at which 50% of the animals become infected (ID50). These excessive doses may not provide realistic preclinical tests of vaccine efficacy. Indeed, no vaccine has been shown to be effective in preventing infection by SIV (so-called sterilizing immunity) in such high viral inoculum trials. Now, in a paper published in PLoS Medicine, Roland Regoes and colleagues speculate that an alternative approach to trials in animals not only can mimic the human patterns of repeated low-dose exposure, but also can remove one concern for animal researchers—the need to use very large numbers of animals in experiments.
What the researchers did was use statistical power analysis to compare a single low-dose challenge design, in which each animal is challenged only once, and a repeated low-dose challenge design, in which each animal is challenged until it is infected or a predetermined maximum number of challenges is reached. The statistical power of an experimental design—a measure of the statistical quality—was assessed by simulating experiments, evaluating them, and then repeating the procedure thousands of times.
What they found was that the experimental design using a single low dose of virus in each animal required unfeasibly large numbers of animals; even for the highest modeled vaccine efficacy of 90% the single low-dose challenge design required more than 20 animals per group to reach a statistical power of 95%. However, when the researchers modeled a protocol of repeatedly challenging the (virtual) animals with a challenge dose of one ID
50, and allowing for a maximum number of 20 challenges of each individual animal, as few as five animals were required to achieve more than 95% of statistical power.
Where do these results leave the design of HIV trials? To begin with, the results should encourage researchers to develop animal models that reflect, to the fullest extent possible, what is known about the natural history and pathogenesis of the disease in humans, rather than designing trials to fit the animal models that are available. The authors have made available the programming script of their analysis so anyone can repeat it; it would be interesting to know whether preclinical trials assessing vaccines or treatments against infections by other pathogens could be usefully modeled in this way as well.
| 0 | PMC1176244 | CC0 | 2021-01-05 11:13:40 | no | PLoS Med. 2005 Aug 19; 2(8):e258 | utf-8 | PLoS Med | 2,005 | 10.1371/journal.pmed.0020258 | oa_comm |
==== Front
PLoS MedPLoS MedpbioplosbiolPLoS Medicine1549-12771549-1676Public Library of Science San Francisco, USA 1602255510.1371/journal.pmed.0020260PerspectivesHIV/AIDSMedical EthicsEthicsHIV Infection/AIDSMedicine in Developing CountriesAchieving Gold Standards in Ethics and Human Rights in Medical Practice PerspectivesBenatar Solomon R Solomon R. Benatar is Professor of Medicine and Director of the Bioethics Centre, University of Cape Town, Cape Town, South Africa, and Visiting Professor in Medicine and Public Health Sciences at the University of Toronto, Toronto, Ontario, Canada. E-mail: [email protected]
Competing Interests: The author is on the editorial board of PLoS Medicine. He declares that he has no competing interests.
8 2005 19 7 2005 2 8 e260Copyright: © 2005 Solomon R. Benatar.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.
Assessment of Discriminatory Health Worker Attitudes and Practices toward Patients with HIV/AIDS in Nigeria
How Do Nigeria's Health-Care Personnel Treat Patients with HIV/AIDS?
A new study suggests that some physicians in Nigeria fall short in their ethical conduct. Benatar urges caution in interpreting the study, saying that "comparisons should be made with other countries."
==== Body
Chen Reis and colleagues' study in this month's PLoS Medicine showed that most Nigerian physicians commendably appeared to be providing appropriate care for HIV/AIDS patients [1]. However, 9% refused to care for such patients, 9% admitted they had refused a patient with HIV/AIDS admission to hospital, and 20% felt that many of these patients had behaved immorally and deserved the disease. The authors also noted the adverse impact of limited health care resources upon ethical practice and protection of human rights (this impact is not surprising in a very poor country with a per-capita GDP of US$290 [2], less than 1% of the United States per-capita GDP). They conclude that discriminatory behaviour and breaches of ethical codes could be addressed effectively through education, enforcement of anti-discrimination policies, increasing resources for health care, and attempts to change attitudes and cultural beliefs.
Presumably, their motivations for this study were (1) to better understand how well physicians in Nigeria respect human rights and meet universal ethical standards of medical practice in caring for patients with HIV/AIDS, and (2) to make recommendations that could improve professional practice. Their findings would be more convincing if they could compare their data with similar studies done elsewhere in the world, including the US (home to some of the authors of the study). In particular, it would be valuable to have comparative international data on ethics and human rights standards achieved in medical practice, and on health professionals' attitudes to patients with HIV/AIDS and other stigmatised conditions. However, Reis and colleagues' study raises several important questions.
What Are the Gold Standards?
What are the gold standards against which to evaluate the ethics and human rights standards that physicians are expected to meet in practice? Presumably these are the standards set out in the declarations, codes, and guidelines quoted by Reis et al., and presumably nothing less than 100% compliance is acceptable. These standards are then applicable to all physicians everywhere—especially to those from rich countries, where the resources available for medical care and continuing education outstrip by orders of magnitude those available in very low income countries like Nigeria.
In order to fully understand the significance of the failure of some Nigerian physicians to meet these gold standards, substantive comparisons should be made with other countries. To do so would entail systematic studies of the extent to which physicians from the US and other wealthy countries meet the requirements of international and local codes of ethics and human rights. Military physicians in the US fall short of gold standards [3,4], and a third of US scientists have engaged in serious research misconduct in the past three years [5]. There is clearly a need for comparative studies in everyday practice.
Do We Know What Causes Discriminatory Practices?
Reis and colleagues' study raises several questions about the basis for discriminatory practices—questions that the study itself cannot fully answer. In particular, how can we judge whether ethical shortcomings in physicians' behaviour (in Nigeria and elsewhere) are due to lack of appropriate health care facilities, inadequate education, lack of enforcement of anti-discriminatory policies, or cultural and other socially determined discriminatory attitudes that might persist despite adequate education and health care facilities?
In order to be able to explore these questions about causality, the authors would need to look at whether there is any evidence that standards of ethics and human rights achieved in practice correlate significantly with health care facilities, education about ethics and human rights, mechanisms for enforcing anti-discrimination policies, and cultural attitudes. Is there any comparative data, for example, on whether higher ethical and human rights standards are achieved in medical practice in wealthy industrialised countries as compared with poor countries, or in countries with universal access to health care as compared with privatised medicine?
It would also be valuable to consider whether health professionals' cultural attitudes about illness elsewhere in the world (not only in Nigeria) may contribute to stigma, discrimination, and worse patient care. Such attitudes could include the belief that medical care is a commodity that should be most accessible to those who can pay, that HIV/AIDS is punishment from God, and that patients who smoke, are obese, or have an unwanted pregnancy are all morally blameworthy and deserve their conditions. Surely such cultural attitudes, wherever they exist, are also worthy of study?
How Can We Reduce Discriminatory Practices?
Reis and colleagues argue that there need to be more resources for HIV care in Nigeria. Their plea should be considered within the broader context of the perverse cultural attitudes that drive the global political economy, promote continuing extraction of human and material resources from developing countries, and sustain poverty [6,7]. Such cultural attitudes and practices that undermine health globally may in fact have a more widely corrosive effect on human rights and professional practices than anything Nigerians do.
If discriminatory practices could be reduced in Nigeria by attention to the rights of people living with AIDS, this implies that such attention could also improve the rights of patients anywhere. For example, in middle-income and wealthy countries, such attention might reduce stigmatisation and moral blaming, improve the rights of many who suffer from chronic “lifestyle diseases” (such as chronic lung diseases related to smoking), and prevent discrimination in access to health care for those who lack insurance cover. It would be valuable to compare Reis and colleagues' study with any studies done in middle-income and wealthy countries that examine whether attention to patient rights improves discriminatory practices.
Conclusions
It is unlikely that any group of physicians anywhere in the world fully meets all of the ethical and human rights standards set in international guidelines. Studies of physicians' shortcomings should be universal. What should be avoided is the previous colonial mentality of wanting to study and improve others [8] while oblivious of the need to address the more sophisticated and covert faults of Western researchers' own societies. The desire to improve the behaviour of others should also be associated with awareness that one's own exemplary moral behaviour might be more effective in promoting ethical behaviour and respect for human rights [9,10] than exhortation, “education” and attempts to change the cultural attitudes of others while neglecting our own adverse cultural attitudes.
This work was supported in part by the University of Toronto and by a grant made by the US National Institutes of Health's Fogarty International Center to the University of Cape Town's capacity-building program in International Research Ethics in southern Africa (for which the author is the program director).
Citation: Benatar SR (2005) Achieving gold standards in ethics and human rights in medical practice. PLoS Med 2(8): e260.
==== Refs
References
Reis C Amowitz LL Heisler M Moreland RS Mafeni JO Assessment of discriminatory clinician attitudes and practices toward patients with HIV/AIDS in Nigeria PLoS Med 2005 2 e246 10.1371/journal.pmed.0020246 16022564
US Department of State. Bureau of African Affairs Background note: Nigeria 2005 Available: http://www.state.gov/r/pa/ei/bgn/2836.htm . Accessed 15 June 2005
Singh JA American physicians and dual loyalty obligations in the “war on terror” 2003 Available: http://www.biomedcentral.com/1472-6939/4/4 . Accessed 23 June 2005
Miles SH Abu Ghraib: Its legacy for military medicine Lancet 2004 364 725 729 15328601
Martinson BC Anderson MS de Vries R Scientists behaving badly Nature 2005 435 737 738 15944677
Pogge T World poverty and human rights: Cosmopolitan responsibilities and reforms 2002 Cambridge Polity Press 296
Dovlo D Taking more than a fair share? The migration of health professionals from poor to rich countries PLoS Med 2005 2 e109 10.1371/journal.pmed.0020109 15916458
Hochschild A King Leopold's ghost: A story of greed, terror, and heroism in colonial Africa 1998 New York Houghton Mifflin 384
Barrigar DL Flagel D Upshur REG Hepatitis B virus infected physicians and disclosure of transmission risk to patients: A critical analysis 2001 Available: http://www.biomedcentral.com/1472-6939/2/4 . Accessed 24 June 2005
Donnelly J Human rights: A new standard of civilization? Int Aff 1998 74 1 24
| 16022555 | PMC1176245 | CC BY | 2021-01-05 10:40:31 | no | PLoS Med. 2005 Aug 19; 2(8):e260 | utf-8 | PLoS Med | 2,005 | 10.1371/journal.pmed.0020260 | oa_comm |
==== Front
PLoS MedPLoS MedpbioplosbiolPLoS Medicine1549-12771549-1676Public Library of Science San Francisco, USA 1602255610.1371/journal.pmed.002026105-PLME-P-0311PerspectivesHealth PolicyHIV/AIDSMedical EducationMedical EthicsHIV Infection/AIDSEthicsTreatment refusalConfidentialityHealth PolicyThe Discriminatory Attitudes of Health Workers against People Living with HIV PerspectivesLetamo Gobopamang Gobopamang Letamo is an assistant representative at United Nations Population Fund, Botswana Country Office, Gaborone, Botswana. E-mail: [email protected], E-mail: [email protected]
Competing Interests: The author declares that no competing interests exist.
8 2005 19 7 2005 2 8 e261Copyright: © 2005 Gobopamang Letamo.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.
Assessment of Discriminatory Health Worker Attitudes and Practices toward Patients with HIV/AIDS in Nigeria
How Do Nigeria's Health-Care Personnel Treat Patients with HIV/AIDS?
Letamo discusses the implications of a new study suggesting that some health professionals discriminate against people with HIV.
==== Body
Anecdotal evidence suggests that health-care professionals in Nigeria may discriminate against and stigmatise people living with HIV/AIDS(PLWA). In a study in the August issue of PLoS Medicine, Reis and colleagues set out to characterise the nature and extent of discriminatory practices and attitudes in the health sector, and indicate possible contributing factors and intervention strategies [1]. The study was specifically designed to answer three research questions: (1) are there discriminatory practices in the health sector that affect the health and well-being of PLWA in Nigeria, (2) how receptive are health workers and institutions to treating PLWA, and (3) what underlying factors may contribute to any discriminatory practices?
In order to answer these questions, a representative sample of 1,103 health-care professionals (doctors, nurses, and midwives) working directly with patients with HIV/AIDS were selected from four states in Nigeria and asked to participate in a study. The response rate was 93% (i.e., 1,021 surveyed professionals participated). A survey questionnaire was administered to respondents to collect information about their knowledge, attitudes, and behaviour. The study was reviewed and approved by an independent ethics review board of individuals with expertise in clinical medicine, public health, bioethics, and international HIV/AIDS and human rights research.
Main Findings of the Study
The results suggest that some health-care professionals discriminate against and stigmatise PLWA. For instance, 9% of professionals reported refusing to care for a patient with HIV/AIDS, and 9% reported that they refused a patient with HIV/AIDS admission to hospital. Two-thirds reported observing other health professionals refusing to care for a patient with HIV/AIDS, and 43% observed others refusing a patient with HIV/AIDS admission to hospital.
The study suggests that a significant number of health-care professionals engage in discriminatory and unethical behaviour. Some professionals reported giving confidential information to other people (family members and unrelated individuals) without the patient's consent. Despite these discriminatory attitudes, an optimistic finding is that most health-care professionals expressed an interest in additional information and suggested education and counselling as a way to address discriminatory behaviours by their colleagues.
Myths and misconceptions about HIV transmission play a role in promoting discrimination.
The study concludes that all clinical staff should be educated about HIV/AIDS, modes of transmission of the virus, universal precautions, and the rights of PLWA. Such education is likely to reduce discriminatory practices towards PLWA and may improve these patients' care and access to health services. The study also asserts that a lack of protective materials and other materials needed to treat and prevent the spread of HIV, documented in several health facilities and reported by professionals themselves, contributes to discriminatory behaviour among health professionals.
Implications of the Study
The study raises the possibility (although does not prove) that patients with HIV/AIDS may not fully utilise health-care services because they are denied access by health-care providers who discriminate against them. The fact that some health workers have discriminated against PLWA in the past suggests that health-care professionals serving PLWA should urgently be educated about HIV/AIDS so that they fully understand how HIV can and cannot be transmitted.
It is clear from Reis and colleagues' study and others, including my own research in Botswana [2–4], that myths and misconceptions about how HIV can and cannot be transmitted play an important role in promoting discrimination. The implication of these findings is clear—there is a dire need to strengthen the information, education, and communication component of HIV/AIDS prevention efforts in order to dispel misconceptions that people tend to hold.
In my own research in Botswana, I found that although the prevalence of discriminatory attitudes towards PLWA may be high, respondents tend to be less discriminating when the HIV-positive person happens to be a family member [2]. People are reluctant to make the serostatus of their relatives public when their relatives are HIV positive, for fear of discrimination.
In order to further our understanding of the root causes of discrimination and stigmatisation of PLWA, qualitative research is needed to understand cognitive processes that lead one to discriminate. It would also be interesting to investigate how PLWA feel that they are perceived by people around them. In order to adequately address issues of discrimination, we must involve PLWA and find out what they feel needs to be done to address stigma and discrimination. We also need to investigate the extent to which researchers are able to measure what they purport to measure with the current indicators of discrimination and stigmatisation. It is possible that the prevalence of stigma and discrimination against PLWA is not being adequately measured with the research instruments currently in use.
Citation: Letamo G (2005) The discriminatory attitudes of health workers against people living with HIV. PLoS Med 2(8): e261.
Abbreviation
PLWApeople living with HIV/AIDS
==== Refs
References
Reis C Amowitz LL Heisler M Moreland RS Mafeni JO Discriminatory attitudes and practices by health workers toward patients with HIV/AIDS in Nigeria PLoS Med 2005 2 e246 10.1371/journal.pmed.0020246 16022564
Letamo G Prevalence of, and factors associated with, HIV/AIDS-related stigma and discriminatory attitudes in Botswana J Health Popul Nutr 2003 21 347 356 15038590
Letamo G HIV/AIDS-related stigma and discrimination among adolescents in Botswana Afr Popul Stud 2004 19 191 203
Letamo G Gender dimensions in misconceptions about HIV/AIDS prevention and transmission in Botswana [poster] 2005 25th International Union for the Scientific Study in Population Conference 2005 18–23 July Tours, France
| 16022556 | PMC1176246 | CC BY | 2021-01-05 10:40:30 | no | PLoS Med. 2005 Aug 19; 2(8):e261 | utf-8 | PLoS Med | 2,005 | 10.1371/journal.pmed.0020261 | oa_comm |
==== Front
BMC Complement Altern MedBMC Complementary and Alternative Medicine1472-6882BioMed Central London 1472-6882-5-121595525410.1186/1472-6882-5-12Research ArticleA systematic review of how homeopathy is represented in conventional and CAM peer reviewed journals Caulfield Timothy [email protected] Suzanne [email protected] Health Law Institute, University of Alberta, Canada2005 14 6 2005 5 12 12 15 3 2005 14 6 2005 Copyright © 2005 Caulfield and DeBow; licensee BioMed Central Ltd.2005Caulfield and DeBow; 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
Growing popularity of complementary and alternative medicine (CAM) in the public sector is reflected in the scientific community by an increased number of research articles assessing its therapeutic effects. Some suggest that publication biases occur in mainstream medicine, and may also occur in CAM. Homeopathy is one of the most widespread and most controversial forms of CAM. The purpose of this study was to compare the representation of homeopathic clinical trials published in traditional science and CAM journals.
Methods
Literature searches were performed using Medline (PubMed), AMED and Embase computer databases. Search terms included "homeo-pathy, -path, and -pathic" and "clinical" and "trial". All articles published in English over the past 10 years were included. Our search yielded 251 articles overall, of which 46 systematically examined the efficacy of homeopathic treatment. We categorized the overall results of each paper as having either "positive" or "negative" outcomes depending upon the reported effects of homeopathy. We also examined and compared 15 meta-analyses and review articles on homeopathy to ensure our collection of clinical trials was reasonably comprehensive. These articles were found by inserting the term "review" instead of "clinical" and "trial".
Results
Forty-six peer-reviewed articles published in a total of 23 different journals were compared (26 in CAM journals and 20 in conventional journals). Of those in conventional journals, 69% reported negative findings compared to only 30% in CAM journals. Very few articles were found to be presented in a "negative" tone, and most were presented using "neutral" or unbiased language.
Conclusion
A considerable difference exists between the number of clinical trials showing positive results published in CAM journals compared with traditional journals. We found only 30% of those articles published in CAM journals presented negative findings, whereas over twice that amount were published in traditional journals. These results suggest a publication bias against homeopathy exists in mainstream journals. Conversely, the same type of publication bias does not appear to exist between review and meta-analysis articles published in the two types of journals.
==== Body
Background
Growing popularity of complementary and alternative medicine (CAM) in the public sector is reflected in the scientific community by an increased number of research articles assessing efficacy and therapeutic effects. At the same time, there is increasing concern about how research results are disseminated and communicated to both professionals and the public (AMA) [1]. Indeed, there is some suggestion that publication bias occurs in both mainstream medicine and with CAM [2]. Given the impact that the scientific literature has on the perception of treatments, particularly within the conventional health care community, it is worthwhile to gain an understanding of how various forms of scientific investigation are represented in the relevant literature. This seems particularly so in the context of CAM, where there has been some criticism about how the conventional scientific community has addressed many CAM treatments [3]. As such, in this systematic review, we seek to explore the difference between how "conventional" and "CAM" peer reviewed journals represent CAM research.
Homeopathy is one of the most widespread and most controversial forms of CAM. A reasonable amount of clinical research has been done on homeopathy (though the quality of the research varies substantially) and the efficacy of many homeopathic remedies remains unclear [4,5]. Moreover, despite its relative popularity among the general public, it has (rightly or not) struggled to gain legitimacy in the medical establishment – particularly in North America [6]. Because homeopathy is a treatment that is likely to evoke a spectrum of responses (ranging from acceptance to deep skepticism), it seems an ideal CAM treatment to study in this context. Indeed, it is an area of study where balance is particularly important. As noted in one paper: "An unbiased conclusion is of utmost importance in this domain because it is a scientific, emotional and political issue in many areas of the world." [7].
Methods
Relevant homeopathic papers reporting on clinical trials were collected. Literature searches were performed using Medline (PubMed), AMED and Embase computer databases. Search terms included "homeo-pathy, -path, and -pathic" and "clinical" and "trial." All articles published in English over the past 10 years were included (1994 through to August 2004). Our search yielded 251 articles overall, of which 46 systematically examined the efficacy of homeopathic treatment. An examination of 15 meta-analyses on homeopathic efficacy was also performed. Review and meta-analysis articles were found by inserting the term "review" instead of "clinical" and "trial". These review articles, all of which analyzed and commented on homeopathic clinical trials, were also used to ensure that our collection of studies was reasonably comprehensive.
If the study or review was published in a journal that specialized in CAM generally (e.g., The Journal of Alternative and Complementary Medicine) or homeopathy (e.g., Homeopathy and the British Homeopathic Journal) it was classified as appearing in a "CAM journal." If the study or review appeared in a general medical journal (e.g., Lancet) or a non-CAM specialty journal (e.g., European Journal of Clinical Pharmacology), it was classified as appearing in a "conventional journal." We categorized the general conclusions of the study (positive or negative results). In addition, we did an analysis of the general tone of the articles. If the article used language that seemed explicitly negative against homeopathy (e.g., condemnation of the general area), it would be categorized as a "negative tone." If the piece used language that was optimistic or encouraging of the area, it was categorized as "positive."
Results
Clinical trials
We examined 46 peer-reviewed articles published in a total of 23 different journals. Twenty-six experiments published in conventional journals were compared with twenty articles published in CAM journals. Of those in conventional journals, 69% (18/26) reported negative findings compared with only 30% (6/20) of CAM journals reporting negative findings.
In our analysis of the general tone of the articles, we found only a few that were clearly negative. However, those that were negative appeared in conventional journals and used relatively harsh language. For example, in one study it was stated "homeopathy is pure quackery" [8]. That said, most of the research papers presented the results in a relatively balanced fashion. We categorized 10 of the 26 conventional journal articles and 8 of the 20 CAM journal articles as having a positive tone. The rest were deemed to be "neutral." Many of the papers categorized as neutral suggested possible clinical benefit despite a negative result. For example, one study published in a conventional journal suggested that " [a]lthough the overall results of the study were negative, they do not rule out the possibility that individual patients may benefit from this homeopathic treatment" [9].
Review articles
A total of the 15 meta-analyses/systematic reviews were examined (10 conventional, 5 CAM). The most common conclusions in the reviews are that the existing evidence remains inconclusive and that more high quality research is required. The conclusions of the reviews were ambiguous enough to make it difficult to categorize them as either a strictly "positive" or "negative" finding. Nevertheless, we can say that none of these reviews had an overtly negative tone (e.g., none of the papers completely dismissed homeopathy, even in the face of negative conclusions). Six articles in the conventional journals and two in the CAM journals had a neutral tone. The rest of the review articles were somewhat positive, encouraging further research and/or suggesting possible clinical benefit. For example, some of the systematic review articles were encouraging of the area, despite skepticism about the possible biological mechanism underlying homeopathy. This occurred in both the conventional and CAM journals. Indeed, we felt that most of the review articles sought to present a balanced picture of the available evidence associated with homeopathy. That said, there were a number of notable differences between the papers published in the conventional and CAM journals.
Almost all of the systematic reviews in conventional journals start on a skeptical note. Indeed, 9 out of 10 of the articles begin with a statement that questions the scientific plausibility of homeopathy. Some of the articles use relatively strong language to make the point. For example, Ernst and Pittler suggest that it is the use of "highly diluted material that overtly flies in the face of science and has caused homeopathy to be regarded as placebo therapy at best and quackery at worst" [10]. Others merely raise the point that there are scientific issues associated with homeopathy and then provide an opposing perspective. McCarney et al., summarize the debate as follows: "The molecules contained in a homeopathic remedy are diluted beyond Avogadro's number. This has led some investigators to question whether homeopathic therapy could have any effect over placebo. However, proponents of homeopathy claim that the remedies act through biophysical pathways, and all include the idea of some form of information transfer from the diluted substance to the diluting agent" [11].
Of the systematic reviews published in CAM journals, only one mentions the issue of "scientific implausibility" and goes on to suggest that "homeopathy's possible mechanisms of action remain intangible theories, and it will be important ultimately to substantiate these" [12]. The other four review articles published in CAM journals begin with defensive statements, noting the benefits of homeopathy (e.g., low side effects) and the challenges of using conventional clinical trial and systematic review methods to assess homeopathic treatments [13].
Despite these differences in approach, all of the systematic reviews are relatively evenhanded, noting arguments on either side of the debates surrounding efficacy and research methodology.
Discussion
Some scholars have speculated about a possible trend toward publishing negative results in conventional journals. For example, Cucherat, et al., have suggested "homeopathy trials with 'negative' results might be more readily accepted by non-homeopathic journals, since the lack of efficacy of homeopathy is in accordance with the belief of many non-homeopathic physicians." [5]. The results of this study provide some preliminary evidence to support this claim. While a small study with clear limitations, there was a stark difference between the numbers of studies that were negative in the conventional journals (69%) as compared to the CAM journals (30%).
That said, publication bias – that is, a journal favoring the publication of positive or negative results – is only one possible explanation of this apparent trend. There may also be a submission bias. For instance, are studies with a negative result submitted to conventional journals and those with a positive to CAM journals? Without access to submission patterns, it will be difficult to analyze this issue. However, it is worth noting that an examination of the affiliations of the first authors of the clinical trials revealed that there was no clear pattern regarding where medical doctors and homeopathic experts publish. In other words, the apparent discipline (and this was not always clear) and home institution of the first author was not a predictor of where an article is published.
Our analysis of the systematic reviews also has some interesting implications. Though there is some evidence of a possible bias in the publication of clinical trials (toward the negative in the conventional journals and toward the positive in the CAM journals), there does not appear to be a similar trend with reviews. In addition, these articles seemed relatively balanced in the presentation of results. Given that there is often significant opportunity to editorialize in systematic reviews, this conclusion demonstrates that the authors (and the editors of journals) are striving to explore this controversial area in a relatively impartial manner. This conclusion does not necessarily conflict with our data that suggests a publication bias. A publication bias would likely be an inadvertent systemic problem (i.e., not an explicit policy) whereas the tone of the articles would reflect apparent writing and editorial decisions.
Conclusion
This small study has a number of clear limitations. For example, we only considered articles that were published in English and we did not critique the quality of the study (might higher quality studies be published in more well known conventional journals?). In addition, aside from the data on whether the conclusions were positive or negative, much of the analysis was subjective. Nevertheless, is does provide some preliminary evidence of a possible publication bias and insight on the general tone of publications involving homeopathy.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
Both Caulfield and Debow contributed to the design of the study, the analysis of the data and the writing and editing of the manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
We would like to thank Edna Einsiedel, Tania Bubela and Heather Boon (and her University of Toronto team) for their input and suggestions and the Advanced Food and Materials Network (AFMNet) and the AHFMR for the funding support.
==== Refs
Council on Scientific Affairs, American Medical Association, Influence of Funding Source on Outcome, Validity, and Reliability of Pharmaceutical Research (presented as CSA Report 10 at the 2004 AMA Annual Meeting, June 2004)
Resch K Ernst Garrow J A randomized controlled study of reviewer bias against an unconventional therapy Journal of the Royal Society of Medicine 2000 93 164 167 10844878
Steuter K Pedalling Skepticism: Media Representations of Homeopathy as "Junk Science" Journal of American and Comparative Cultures 2001 24 1 19 10.1111/j.1537-4726.2001.2403_1.x
Kleijnen J Knipschild P ter Riet G Clinical Trials of Homeopathy BMJ 1991 302 316 23 1825800
Ernst E A systematic review of systematic reviews of homeopathy Br J Clin Pharmacol 2001 54 577 582 12492603 10.1046/j.1365-2125.2002.01699.x
Steuter E Contesting the rule(s) of medicine: homeopathy's battle for legitimacy Journal of Canadian Studies (Special Issue on Science and Politics in Canada) 2002 37 1 20
Cucherat M Evidence of clinical efficacy of homeopathy Eur J Clin Pharmacol 2000 56 27 33 10853874 10.1007/s002280050716
Smolle J A double-blind, controlled clinical trial of homeopathy and an analysis of lunar phases and postoperative outcome Arch Dermatologica 1998 134 1368 1370 10.1001/archderm.134.11.1368
Stevinson C Homeopathic arnica for prevention of pain and bruising: randomized placebo-controlled trial in hand surgery J R Soc Med 2003 96 60 65 12562974 10.1258/jrsm.96.2.60
Ernst E Pittler M Efficacy of Homeopathy Arnica: A systematic review of placebo-controlled clinical trials Arch Surg 1998 133 1187 1190 9820349 10.1001/archsurg.133.11.1187
McCarney RW Lasserson TJ Linde K Brinkhaus B An overview of two Cochrane Systematic Reviews of Complementary treatment for Chronic Asthma: Acupuncture and Homeopathy Respiratory Medicine 2004 98 687 696 15303632 10.1016/j.rmed.2004.05.005
Mathie RT The research evidence base for homeopathy: a fresh assessment of the literature Homeopathy 2003 92 84 91 12725250 10.1016/S1475-4916(03)00006-7
Jonas W Anderson R Crawford C Lyons J A systematic review of the quality of homeopathic clinical trials BMC Complementary and Alternative Medicine 2001 1 1 10 11346455 10.1186/1472-6882-1-12
| 15955254 | PMC1177924 | CC BY | 2021-01-04 16:31:45 | no | BMC Complement Altern Med. 2005 Jun 14; 5:12 | utf-8 | BMC Complement Altern Med | 2,005 | 10.1186/1472-6882-5-12 | oa_comm |
==== Front
BMC BioinformaticsBMC Bioinformatics1471-2105BioMed Central London 1471-2105-6-1361592979310.1186/1471-2105-6-136Methodology ArticleAVID: An integrative framework for discovering functional relationships among proteins Jiang Taijiao [email protected] Amy E [email protected] Department of Biology, Massachusetts Institute of Technology, Cambridge, USA2005 1 6 2005 6 136 136 14 12 2004 1 6 2005 Copyright © 2005 Jiang and Keating; 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
Determining the functions of uncharacterized proteins is one of the most pressing problems in the post-genomic era. Large scale protein-protein interaction assays, global mRNA expression analyses and systematic protein localization studies provide experimental information that can be used for this purpose. The data from such experiments contain many false positives and false negatives, but can be processed using computational methods to provide reliable information about protein-protein relationships and protein function. An outstanding and important goal is to predict detailed functional annotation for all uncharacterized proteins that is reliable enough to effectively guide experiments.
Results
We present AVID, a computational method that uses a multi-stage learning framework to integrate experimental results with sequence information, generating networks reflecting functional similarities among proteins. We illustrate use of the networks by making predictions of detailed Gene Ontology (GO) annotations in three categories: molecular function, biological process, and cellular component. Applied to the yeast Saccharomyces cerevisiae, AVID provides 37,451 pair-wise functional linkages between 4,191 proteins. These relationships are ~65–78% accurate, as assessed by cross-validation testing. Assignments of highly detailed functional descriptors to proteins, based on the networks, are estimated to be ~67% accurate for GO categories describing molecular function and cellular component and ~52% accurate for terms describing biological process. The predictions cover 1,490 proteins with no previous annotation in GO and also assign more detailed functions to many proteins annotated only with less descriptive terms. Predictions made by AVID are largely distinct from those made by other methods. Out of 37,451 predicted pair-wise relationships, the greatest number shared in common with another method is 3,413.
Conclusion
AVID provides three networks reflecting functional associations among proteins. We use these networks to generate new, highly detailed functional predictions for roughly half of the yeast proteome that are reliable enough to drive targeted experimental investigations. The predictions suggest many specific, testable hypotheses. All of the data are available as downloadable files as well as through an interactive website at . Thus, AVID will be a valuable resource for experimental biologists.
==== Body
Background
High-throughput technologies, including genome sequencing, expression profiling, and large-scale interaction and localization assays, have provided a wealth of data about proteins and their properties, particularly for the model organism Saccharomyces cerevisiae [1-5]. The process of inferring functional information from these data is not straightforward. The data are not of uniformly high quality and must be weighted in an appropriate way [6-8]. Computational methods such as machine learning hold promise for this task, but they require a clear definition of "protein function". Gene Ontology (GO) has emerged as a unifying framework that makes it possible to carry out computational function annotation [9].
GO uses expert curation to systematically describe the role of proteins in the cell. Three hierarchical ontologies are used: molecular function (MF) describes the specific molecular task performed by a protein, biological process (BP) describes the broader biological activity a protein participates in, and cellular component (CC) describes the subcellular location or complex where a protein is found. The deeper an annotation is in one of the GO hierarchies, the more informative and specific it is. For example, at a low level the yeast protein Mcm2 is annotated with "catalytic activity" as a molecular function, but at the most detailed level of the MF hierarchy its description is "ATP-dependent DNA helicase activity". Similarly, Mcm2's biological process of "cell growth and/or maintenance" is refined to "DNA unwinding", and the general descriptor "nucleus" is refined as "pre-replicative complex" at the most descriptive level of the CC classification. An important goal is to provide the most detailed possible annotations for all proteins. Currently, however, only ~60% of the S. cerevisiae proteome has annotation at the most descriptive level of at least one of the MF, BP or CC classifications.
Many groups have explored computational analysis as a way to expand functional assignments. For example, the genomic context of a gene can reveal functional relationships, especially in prokaryotes, as can patterns of co-evolution [10-12]. Several methods have been proposed for using experimental protein-protein interaction networks to assign functional descriptors to unknown proteins on the basis of their interaction partners [13-17]. Other researchers have developed methods for combining multiple sources of data [11,12,18,19]. Troyanskaya et al. used a Bayesian approach to increase the accuracy of functional predictions of GO BP terms [20], and Jansen et al. predicted new members of protein complexes in this way [21,22]. Tanay et al. (2004) integrated several sources of experimental data to generate statistically significant protein "modules" and used the modules to assign GO BP terms to 874 uncharacterized yeast proteins [23]. Most recently, Lee et al. used a Bayesian framework to build networks reflecting functional relationships between 4,677 yeast proteins [24].
We present a method called AVID (Annotation Via Integration of Data) for predicting functional relationships among proteins. AVID integrates the results of high-throughput experiments, and incorporates sequence data, to build unified, high-confidence networks in which proteins are connected if they are likely to share a common annotation. We illustrate one use of these networks by treating functional annotation as a classification problem and assigning GO terms to individual proteins based on their neighbors in the networks. AVID is distinct from previous computational function prediction methods in several ways that will make it a useful tool for experimental biologists. First, AVID predicts functional annotation in all three GO categories: MF, BP and CC. This provides a more complete view of an uncharacterized protein's possible role in the cell than any single term alone. Second, the functional terms predicted by AVID are very detailed. We adopted only the most specific terms in GO as a list of possible annotations and refer to these as AVID GO terms. AVID GO terms have no functional subcategories. There are 841 AVID GO terms for MF, 602 for BP and 192 for CC that are used for yeast. Such terms are considerably more useful than general ones, and they are also harder to predict. Third, by considering five different types of input data, AVID achieves good coverage; we report here predictions of new GO terms for about half of the yeast genome. Fourth, AVID is reliable. The functional networks generated are 65–78% accurate and annotation of proteins is 52–67% accurate. The trade-off obtained between coverage and accuracy is superior to that obtained using a naïve Bayesian framework. Finally, AVID predicts relationships among proteins that are largely distinct from those that have been suggested by other computational methods [12,13,20,21,24-26].
Here, we describe three stages used in AVID to construct functional correlation networks and a fourth stage that is used to assign specific functions to individual proteins. We report the estimated accuracy of AVID at different stages using known data. Then we describe the results of applying AVID to the entire yeast proteome to generate new GO annotations.
Results and discussion
Description and performance of the four stages of AVID
Figure 1 outlines the multi-stage method and its application to predicting the MF of unannotated protein YOL137W. AVID can be regarded as a filtering process. Initially, all proteins are considered as potentially functionally related. Stages 2 and 3 remove lower confidence associations. The links remaining after stage 3 form networks that contain a wealth of information about functional similarity. These networks are the primary output of AVID, and in stage 4 they are used to assign specific functional terms to individual proteins.
In stage 1, diverse features, such as the presence or absence of sequence similarity, or the observation of a protein-protein interaction, are considered as potential indicators of whether two proteins share an AVID GO term (Table 1 and additional file 1). For each type of evidence, i, and each GO category, j = {MF, BP, CC}, we define a correlation coefficient PijAVID1 to describe how well the evidence predicts functional similarity. For all three ontologies, MF, BP and CC, there is a weak positive correlation between the experimental observation of an interaction (by yeast 2-hybrid or by co-purification) and similarity of annotation. Sequence similarity is correlated with MF and BP, but less so with CC, consistent with the expectation that evolutionarily related proteins frequently have related functions but act in diverse parts of the cell. The cellular localization data of Huh et al. [1] correlate positively but very weakly with CC. This is because GO cellular components, at the most detailed level, are protein complexes that are much more descriptive than this experimental localization data. AVID GO terms in the CC classification should often, in fact, be regarded as descriptions of protein interaction rather than cellular localization. Thus it is not surprising that, e.g., two proteins sharing the location "nucleus" have a low probability of participating in the same complex (and thus sharing an AVID GO CC term). Gene co-expression profiles correlate with all three GO functional types; the higher the Pearson correlation between two expression profiles is, the more likely the two proteins share a GO term of any kind. Most of the correlations between data and function are very weak; none of the correlation coefficients are greater than 35%. Nevertheless, differences between the (+) and (-) data sets (see Table 1) make these sources valuable for inferring functional similarity.
Stage 2 is a filter that combines data sources and removes protein pairs that lack sufficient evidence of functional relatedness. For each pair of proteins, a PjAVID2 value is defined as the product of the normalized conditional probabilities PijAVID1 from all sources of evidence in stage 1. Tests using known proteins show that pairs with a PjAVID2 value greater than 12.8 have more than 66.8%, 45.8% or 69.3% probability of MF, BP or CC relatedness (Figure 2A). Pairs with PjAVID2 < 12.8 are not considered further. Good coverage is preserved using this cutoff: 60.8%, 74.0% or 79.0% of proteins in test sets with existing MF, BP or CC annotation are retained by the filter.
To improve accuracy while still making predictions for a large number of proteins, a machine learning scheme (a decision tree) is employed in stage 3 [27]. The tree takes the stage 1 conditional probabilities as input, and returns a binary decision about whether two proteins are likely to share a function. The entire process used to predict the similarity of three pairs of proteins is illustrated in detail in additional file 2, which includes diagrams of the decision trees for MF, BP and CC.
Stages 1, 2 and 3 result in the construction of a reliable protein correlation network for each of the GO functional types. The three networks relate proteins likely to share a similar function or be part of the same protein complex. Ten-fold cross-validation testing using proteins with existing GO annotation showed that 77% of MF, 65% of BP and 78% of CC pair-wise relations predicted in stage 3 were correct. Furthermore, 1,432 proteins (55.4%), 1,122 proteins (48.6%) and 974 proteins (72.3%) for MF, BP and CC, respectively, were retained in pairs judged to have functional similarity after stage 3. Alternative machine-learning strategies were tested but did not show any improvement in performance. All of the stages employed to construct the networks are important. In particular, a decision tree applied to unfiltered data was not as effective, resulting in 4–21% reduced coverage and 2–3% reduced accuracy for BP and CC, and ~10% reduced accuracy for MF, compared to the full 3-stage process.
In stage 4, a "majority rule" algorithm is used to annotate uncharacterized proteins based on the correlation networks. AVID often predicts several GO terms for a protein. The average number of functions predicted (compared to the number of existing annotations in GO, in parentheses) is 2.19 (1.45) for MF, 1.95 (1.07) for BP, and 1.87 (1.14) for CC. In testing, if any of the AVID predictions were identical to an existing GO annotation the prediction was counted as correct. If one term was correctly predicted, all terms were correctly predicted 85% of the time for MF, 74% of the time for BP and 83% of the time for CC.
Figure 2B shows that the success rates for predicting AVID GO terms in stage 4 depend on the fraction of proteins treated as unknown in the networks during testing. If only 10% of proteins have known function, the success rate for the remaining 90% (the test set) is only 30–50%. However, if functions are known for 90%, the remaining 10% of proteins can be annotated with 65–75% accuracy. Whereas the assessment data in Figure 2B are based on reference sets of annotated proteins, divided into training and testing sets, the actual fraction of yeast proteins currently unannotated at the AVID GO level is 56.4%, 56.5% and 70.3% for MF, BP and CC, respectively. Thus we can use Figure 2B to estimate that MF, BP and CC prediction success, when applied to the entire proteome, will be ~67%, ~52% and ~66%, respectively. As more proteins are annotated, predictive accuracy for unknown proteins will improve further.
As mentioned above, AVID predicts 1.5 to 1.8 times as many AVID GO terms per protein as already exist in GO. This is expected, because current annotation is incomplete. In tests on annotated proteins, the percentage of functional terms predicted by AVID that are already included in GO is 46%, 33% and 47% for MF, BP and CC, respectively. We recover 63% (MF), 44% (BP) and 60% (CC) of previously annotated AVID GO terms.
New networks including unannotated S. cerevisiae proteins
We applied AVID to the entire yeast proteome, generating three new networks that include proteins for which detailed AVID GO annotation is not yet available. These predicted networks have similar connectivity to the networks generated from existing GO data for testing. The average numbers of edges per node in the testing networks were 10.2, 8.2, and 17.2 for MF, BP and CC, respectively, whereas in the prediction networks these averages were 13.4, 8.7 and 12.5. The distribution of connectivities in the different networks are provided in additional file 3. Using the predicted networks, we made two types of functional predictions. All predictions are available in additional files 4 and 5.
Refined predictions
First, we predicted a more detailed function or location for proteins previously characterized only at less descriptive levels in GO, we refer to these as refined predictions. We evaluated whether refined predictions are consistent with existing coarser annotations at GO levels 2, 3 and 4 for MF, BP and CC, respectively. An AVID GO prediction was judged consistent with a coarser one if the less descriptive term is its ancestor in the GO hierarchy. There are 17 functional categories in MF level 2, 24 functional categories in BP level 3 and 40 functional categories in the CC level 4, so this comparison is a non-trivial test. Because the definition of "level" can be ambiguous, the categories used are listed in additional file 6.
Table 2 summarizes the results; refined predictions are 75 – 87% consistent with existing coarser annotations. Note that these coarser annotations were not used by AVID in any way when predicting specific terms. This indicates that new AVID GO terms, when traced back to MF level 2, BP level 3 or CC level 4, are 75–87% accurate. This estimate is higher than the accuracy for predicting AVID GO terms themselves because it is easier to assign more general functions correctly. For cases in which a refined AVID prediction is not a subcategory of existing GO annotation, we identified many cases where the prediction is, nevertheless, biologically relevant. In a trivial example, AVID assigns YOR244W (Esa1) a MF of "chromatin binding". Our formal accounting treated this as incorrect in testing because "chromatin binding" is not a sub-category of the existing GO term for Esa1, "histone acetyltransferase activity". However, Esa1 catalyzes acetylation of chromatin substrates [28], so the disagreement in this test is merely an artefact of the structure of GO. This suggests that refined AVID GO predictions are more consistent with existing biological knowledge than the estimate of 75–87%.
Novel predictions
In a second type of prediction, we assigned one or more AVID GO terms to proteins without any existing annotation in GO at any level, we refer to these as novel predictions. Note that these are novel in the sense that they annotate proteins not previously included in GO. Other sources of evidence regarding function may exist, e.g. in the literature or at the Saccharomyces Genome Database (SGD) [29]. We made novel MF predictions for 950 proteins, novel BP predictions for 504 proteins and novel CC predictions for 907 proteins (Table 2). The refined and novel predictions for MF, BP and CC together cover 51% of the yeast proteome and, when combined with existing annotations, provide AVID GO descriptors for ~80% of yeast proteins. Cross-validation testing indicated that these predictions are ~52–67% accurate (previous section). The accuracy of specific novel predictions is hard to evaluate systematically, but we assessed the plausibility of a subset of both our refined and novel predictions using the experimental data of Hazbun et al. [30] (Figure 3), which were not included in the version of GO used to develop our method. In this work, 100 uncharacterized open reading frames were labelled with a tandem affinity purification tag. Mass spectrometry was used to identify proteins that form a complex with the gene product of interest. Overlap with the high-throughput affinity purification data used as input to AVID was very low (7%). For each novel annotation predicted by AVID for a protein localized to a complex by Hazbun et al., we classified it as GO-consistent if another member of the complex had the same annotation in GO, and as AVID-consistent if another member of the complex had the same annotation predicted by AVID. We found that ~46% of MF, ~16% of BP and ~27% of CC predictions were GO-consistent; ~73% (43 out of 59) of MF, 63% (27 out of 44) of BP and 71% (45 out of 62) of CC predictions were GO- or AVID-consistent. Among the annotations not formally classified as GO- or AVID-consistent some are nevertheless clearly relevant. For example, AVID assigns "mitotic chromosome segregation" to YNL313C. In the Hazbun experiments, YNL313C co-purified with Tub3p, and Tub3p is annotated in GO with the very similar term "homologous chromosome segregation". Thus, AVID predicted the functional similarity of YNL313C and Tub3, and this was supported later by the co-purification of these proteins. Notably, the AVID prediction of similarity of YNL313C and Tub3 did not come directly from experimental evidence; there is no direct edge between these two proteins in any of the AVID networks.
AVID predictions of MF, BP and CC were made using different weighting of the input data, different functional categories and different stage 3 correlation networks. For 85 proteins, AVID provided novel predictions in all three of these categories. In many examples, the three novel predictions are related and consistent (additional file 7). Here we list several examples from these 85 proteins where experimental data, or descriptions in SGD, support our predictions. First, YOR179C (Syc1) was assigned by AVID a BP of "mRNA polyadenylation", a MF of "cleavage/polyadenylation specificity factor activity" and CCs of "mRNA cleavage factor complex" and "cleavage and polyadenylation specificity factor complex". GO annotation added after our predictions were completed assigned YOR179C "mRNA cleavage and polyadenylation specificity factor complex" as a cellular component, based on the experiments of Nedea et al. [31]. In another example, AVID assigned "tRNA-intron endonuclease activity" (MF), "tRNA splicing" (BP), and "tRNA-intron endonuclease complex" (CC), to YLR375W (see additional file 2). SGD lists the description "involved in pre-tRNA splicing and in uptake of branched-chain amino acids" for YLR375W, although this information is not included in GO or supported by literature references at SGD [29]. In a third example, AVID assigned YEL018W (Eaf5), YER092W (Ies5) and YDL002C (Nhp10) to "histone acetylation" (BP), "chromatin binding" (MF), and "nucleosome remodelling complex" (CC); SGD reports for Eaf5 the description "subunit of the NuA4 acetyltransferase complex", and for Ies5 "protein that associates with the INO80 chromatin remodelling complex under low-salt conditions". The involvement of Nhp10 in chromatin remodelling is also supported by Shen et al. [32]. Similar consistency among MF, BP and CC predictions supports the annotation of YBR043C, YIL121W (Qdr2) and YOL137W (Bsc6) as plasma membrane-associated proteins involved in glucose transport and/or galactose metabolism and the assignment of YER084W, YGR017W, YLR154C (Rnh203) and YPL014W as being related to (possibly regulating) protein kinase CK2 activity. Many other suggestive examples are available in additional file 7.
We have constructed an interactive web server that allows users to individually trace the data that contributed to any prediction [33]. For example, to understand the origins of the MF prediction made for YOL137W, shown in Figure 1, a user can look up the identities and functions of all neighbors of this protein in the stage 3 networks. For each neighbor, we provide information about what experimental or sequence data was used to establish the relationship, as well as the stage 1 weight assigned to that data source, and an additional measure of confidence from the decision tree processing (see Methods). We also give its status as "known", "refined" or "novel". This makes it possible to establish, for example, that YOL137W was assigned GO term 0005355 ("glucose transporter activity") because the majority rule vote by neighbors of known function was won by YOL156W and YDL138W, which share GO term 0005355. The association of these proteins with YOL137W was determined primarily from sequence similarity and mRNA co-expression. Users can also see, however, that YOL137W is predicted to share functional similarity with other proteins, e.g. YOL103W (GO:0005365, "myo-inositol transporter activity"), also on the basis of sequence similarity and mRNA co-expression. This demonstrates how examining AVID network relationships can provide a broader picture than the stage 4-assigned terms alone.
Comparison to other methods
A variety of methods have been proposed in the literature for the computational annotation of protein function, but these are difficult to compare given the lack of adequate "gold standard" data sets for universal testing. There are further obstacles to rigorous comparison. First, groups formulate the function prediction problem differently. For example, our use of highly specific GO terms is distinct from the generally broad (and widely varying) functional classifications used by others. Second, methods use different sources of evidence and training sets. Finally, various methods provide different forms of output, ranging from sets of pair-wise relationships between proteins to functional modules to specific functional annotations.
Bayesian frameworks are popular for data integration problems, and have been applied to the prediction of protein functional similarity and protein-protein interactions [20,21,24]. To compare the AVID framework rigorously with a naïve Bayesian net, we implemented the method of Jansen et al. [21] and applied it to our set of reference proteins. As shown in Figure 4, AVID stages 1 and 2 perform comparably to this formalism for MF and BP, although they out-perform it for CC. With the addition of the decision tree in stage 4, however, the trade-off between accuracy and coverage improves significantly for AVID. Whereas both methods are good at making high-confidence predictions at low coverage, the AVID framework maintains much better true positive to false positive ratios than the naïve Bayes net at higher coverage.
Although implementing and comparing a large number of other approaches for the same data is beyond the scope of this work, we can compare overall results obtained by different groups. This turns out to be interesting and surprising. For six published methods that generate pair-wise relationships among yeast proteins, we compare the coverage and overlap of the predicted associations in Table 3. We compare methods at roughly comparable levels of accuracy to that of AVID (~70%), using estimates of the original authors. AVID predicts 37,451 relationships among 4,191 proteins. Lee et al. [24], referred to here and in Table 3 as "MARCOTTE") also obtain very high coverage: 33,919 high-confidence associations among 4,677 proteins. STRING further predicts 23,345 functionally related pairs. However, the largest overlap between any two methods in Table 3 is only 9,873 pair-wise associations predicted by both MARCOTTE and STRING [12,26]. Both of these methods use genomic context as an important predictive element. AVID does not consider genomic context and shares only 3,413 predictions with STRING. These make up 9% of the total AVID predictions, and no other method shows greater overlap. Out of 37,451 high-confidence associations predicted by AVID and 33,919 by MARCOTTE, only 3,020 of these are in common. In light of the fact that all methods show incomplete coverage and imperfect accuracy, the distinct predictions made by different methods are a significant advantage because they provide alternatives that can profitably be considered by experimentalists.
Conclusion
Computational annotation of the proteome has a critical role to play in post-genomic analysis. Although hypotheses about function can often be reached by carefully reading the literature and critically examining high-throughput data, computation can speed and assist this process. Further, computational methods can help discriminate reliable data amidst false positives and negatives. As a tool for this purpose, AVID notably provides functional descriptors at a high level of detail. The strategy of predicting MF, BP and CC terms also provides a more comprehensive description of protein function than many alternative approaches. Finally, AVID performs better than simple naïve Bayesian integration, and the predictions of AVID are largely distinct from those that have been made by other methods.
The stage 3 networks generated by AVID are very accurate (65–78%) and are useful in the absence of stage 4. The specific predictions made in stage 4 are accurate enough to be of practical utility, but they do have limitations. The imperfect majority rule algorithm will sometimes select one function over others that may be equally relevant. Further, because we consider only the most detailed GO categories in training and prediction, some predictions will be incorrect because they are overly specific, even when they correctly reflect the general cellular role of a protein. For these reasons, consideration of the entire stage 3 functional networks is likely to be most useful to experimental biologists. Other algorithms for assigning function based on the AVID networks may give better performance. This is an active area of research [13-16].
AVID can be used to assign new proteins to existing GO functional categories. The catalogue provided by GO is incomplete, however. Accurate descriptors have not yet been defined for all possible functions, processes and compartments. Because of this, proteins with new functions will not be successfully assigned by AVID. Within the limitations imposed by GO, however, performance on novel proteins may be better than estimated by our testing. When assessing the performance of stage 4, we treated known proteins as unknown. This reduced the size of the training set for stage 3 to less than half of that available when making new predictions; this decreases predictive accuracy. Furthermore, most proteins have more than one function, and many are found in more than one cellular compartment or complex. When assessing AVID stage 3, predicted functional similarities among test proteins that are not yet annotated in GO are counted as wrong and thus reduce the estimated success rate, even though many are likely to be correct.
Algorithmic function prediction can be approached from different perspectives, and it will be important for computational biologists to explore various formulations of the problem as well as solutions to it. Ultimately, the value of any approach will be justified through the cumulative success of experiments that it inspires. Our functional networks provide numerous candidate proteins for involvement in important biological processes. Biologists who consult AVID as part of their work are likely to find new predictions of function for their genes of interest that are accurate enough to guide experimental characterization. Thus, AVID is sure to provide a useful resource for the yeast community.
Methods
Data sources
To establish sequence similarity, 6,449 protein sequences from the yeast proteome were downloaded from MIPS [34]. Each sequence in turn was used as a PSI-BLAST query against the entire yeast proteome. Proteins with an E-value less than 0.001 after three iterations of PSI-BLAST were defined as similar to the query sequence [35]. A total of 66,833 similar pairs involving 3,631 proteins were identified in this manner.
A list of high-throughput yeast two-hybrid interactions was obtained from MIPS (file PPI_120803.tab) that included 6,620 pairs among 3,579 proteins (not including 214 self interactions) detected by the high-throughput yeast two-hybrid assays of Uetz and Ito [4,5,36]. Small-scale yeast two-hybrid experiments were not included.
Protein complex file complex052102.tab was downloaded from MIPS. Among these proteins, 67,569 pair-wise relationships were defined between 2,696 proteins reported to occur in the same complex [2,3]. Within a complex, every protein was assigned an interaction with all others.
We used the cellular localization data of Huh et al. [1,37]. A link was assigned to two proteins if they were reported in the same cellular compartment, without considering ~270 proteins with ambiguous localization. This led to the construction of 975,891 pairs among 3,883 proteins. In Table 1 this data set is called "UCSF localization".
The data GDS124 was downloaded from NCBI Gene Expression Omnibus [38]. We used cdc15 block-release time course mRNA expression from the yeast cell cycle. These data consist of 24 time points taken during the course of almost three full cell cycles. We computed the Pearson correlation for each of 19,734,903 pairs among 6,283 proteins.
Yeast protein annotations and hierarchical terms for biological processes, molecular functions and cellular components were downloaded from GO [39]. We only considered annotation categories that do not have any sub-categories. We call these AVID GO terms. A pairing link is assigned to two proteins if they share an AVID GO term.
Data from small-scale experiments were not used. We found that small-scale experiments are already largely captured by existing GO annotation. Furthermore, as implemented, the results we obtain with AVID reflect what is likely to be possible in other organisms, where high-throughput data sets will soon far outnumber small-scale experiments.
Stages of AVID
The following stages were carried out separately for MF, BP and CC.
AVID stage 1 – correlation analysis
We considered five features, fi, that characterize pairs of proteins. Four of these features are either present (fi+) or absent (fi-) for a protein pair: co-localization, two-hybrid interaction, co-occurrence in a complex and sequence similarity. A fifth feature, mRNA expression profile correlation, was described by the Pearson correlation coefficient R. R values were binned into 19 intervals. We considered three GO categories, GOj (one of MF, BP, CC). Conditional probabilities were defined using only the set of proteins, sij, that had records for both fi and GOj. GOij+ is the set of all protein pairs among sij that share an AVID GOj term. For each feature, i, and GO category, j, we computed the conditional probability that a pair of proteins that share feature fi also share some AVID GOj annotation term: PijAVID1= P(common GO term|common data feature) = (pairs in GOij+ with fi)/(pairs with fi). Because protein pairs lacking a relationship (e.g. an interaction) are frequently not reported, the number of such negative pairs was defined as {(possible pairs among p) – (pairs observed to have the corresponding positive feature)}. The correlation coefficients Pij AVID1 were normalized by a factor α = 0.01/ [(pairs in sij and in GOij+)/(pairs among sij)] to account for the different sizes and compositional biases of the sets sij. The value 0.01 is a reference constant used in place of PijAVID1 for protein pairs without records in fi; it is approximately the probability that two randomly chosen proteins will share an AVID GO term. The analysis is insensitive to the value chosen for this constant.
AVID stage 2 – combining data to build a network of correlations
Each pair of proteins in the positive and negative reference sets was described by a set of five PijAVID1 terms for each of MF, BP and CC; the reference value 0.01 was used when feature data was missing (see above). In stage 2 we computed PjAVID2 = Πi100•PijAVID1. Pairs of proteins with PjAVID2 < 12.8 for MF, BP and CC were not considered further. This cutoff was chosen to achieve a good compromise between accuracy and coverage, which can not be simultaneously optimized. Coincidentally, the 12.8 value was reasonable for all three GO categories.
AVID stage 3 – decision tree
In stage 3, the PijAVID1 values from stage 1 for protein pairs that passed the stage 2 filter were used as the input to a decision tree that retuned a binary decision about the presence or absence of functional similarity [27]. Decision trees provide a supervised machine learning scheme for classification and can analyze hierarchical complex relationships. The idea is to recursively subdivide a training set of examples into homogeneous groups, using discriminating attributes. The attribute selection criteria are based on a measure of informational entropy. At each decision point, an attribute is chosen so as to result in the best discrimination of the data into classes. After training, a set of complex rules is represented as a tree structure where non-terminal nodes represent tests on one or more attributes and terminal nodes reflect decision outcomes. We adopted java source code from Weka [40] for the implementation of the decision tree. J48 is an extension to the C4.5 algorithm by Quinlan [41]. It uses a recursive "divide and conquer" strategy to generate a decision tree from the training data. The input training data consist of a list of examples with attribute values and a class label (in our case it is YES or NO to represent correlation or not). Following Zhang et al. [19], we computed a measure of confidence in different paths through the final trees used for making new predictions. This measure is the probability that the decisions made to reach a particular terminal node correctly classified reference data; it is defined for each terminal node and is assigned to each protein pair partitioned to that node. These values are available at the AVID web site.
AVID stage 4 – assigning functions based on correlations
Pairs predicted to be functionally related by the decision tree in stage 3 comprise three correlation networks (one each for MF, BP and CC). In stage 4 these networks are used to classify unknown proteins based on their relationships to categorized neighbors. We assigned functions to unclassified proteins on the basis of the most common function(s) present among their annotated neighbors (the "majority rule" approach) [17]. Wherever possible, functions were assigned based only on GO-annotated proteins. When this was impossible (e.g. no neighbors had known functions), subsequent rounds of majority rule were used to assign functions on the basis of predicted annotations for neighbors. We also tested several iterative methods, such as those discussed by Vazquez et al. [14], but did not find any improvement in performance.
Cross-validation testing
The performance of the AVID framework was evaluated using cross-validation testing by splitting the data into training and test sets prior to the stage 1 correlation analysis. We defined test sets in which increasing percentages (n) of the reference proteins were treated as unknown. The remaining 100-n% of proteins constituted the training set and were used in stages 1, 2 and 3 to generate a predictive model. This model was applied to the entire reference set to generate functional correlation networks. In these networks, the n% of proteins making up the test set remained unannotated. One or more functions were assigned to them using majority rule, and these predictions were compared to original terms assigned by GO. If at least one AVID GO term matched, the annotation of that protein was counted as correct. Accuracy was defined as the number of proteins with ≥ 1 correctly predicted AVID GO term divided by the number of proteins treated as unannotated. For each value of n, ranging from 10 to 90%, 100 random test sets were analyzed (for each of MF, BP and CC). The results reported are the average of all trials, with error bars in Figure 2b showing the standard deviations.
Comparison with other methods
We compared AVID with a naïve Bayesian network approach used by Jansen et al. [21], which is described in detail in the supplementary materials of that reference. In this formalism, the probability of two proteins sharing functional annotation, given evidence sources f1...fn, is proportional to the likelihood ratio L: L(f1...fn) = P(f1...fn|functional similarity)/P(f1...fn|no functional similarity). Data features f1...fn are defined above in the section describing AVID stage 1. L can be computed from contingency tables relating these data features with pairs of reference proteins that are and are not functionally related. These tables are given in additional file 8. Figure 4 compares the performance of AVID stages 1 and 2, AVID stages 1–3 and the naïve Bayes approach by plotting the ratio of true positives to false positives (TP/FP) as a function of sensitivity (defined as TP/P) for reference data. We generated data at various values of TP/FP and TP/P by varying the values of L and the cutoff for PjAVID2.
In Table 3 we compare data from other studies with the content of the AVID functional correlation networks. The following datasets were downloaded from the indicated sources. The cutoff applied (if any) was chosen based on the original reference to generate "high-confidence" protein pairs. Where possible, as detailed below, the cutoff was chosen to result in roughly 70–80% accuracy, to match the estimated accuracy of the AVID pairs. The numbers reported for AVID include edges predicted for all three networks (MF, BP and CC); no protein pair was counted more than once.
1. MARCOTTE. "ConfidentNet", file 1099511s1_5.zip, from the supplementary materials of Lee et al. [24]; the top 34,000 pairs were used. These data are estimated by the authors to be as accurate as small-scale experiments; possibly greater than 70% accurate.
2. STRING. From the STRING database at [42], the file links_high_confidence_v5_1.txt [12,26]. This data set is reported to be ≥ 75% accurate.
3. PIP_600. From the work of Jansen et al. [21], file L_cut_PIP_600.tar from [43]. These predictions are theoretically estimated to be ~50% accurate.
4. LIANG. From the work of Samanta and Liang [13], all pairs with p < 10-8 from [44]. These data are reported as ~70–75% accurate for predicting similarity in broad categories.
5. MAGIC From the data of Troyanskaya et al. [20], predictions.txt from [45]. We used two data sets, one with a probability of being functionally related of ≥ 50%, the other with ≥70%.
6. SCHLITT. File Schlitt-11114_supp_3.txt, with p ≤ 0.01, from the supplementary material of [25]. This p-value cutoff was used in the paper, but what accuracy it corresponds to isn't established.
Authors' contributions
TJ and AEK conceived these studies and analyzed and interpreted the data and their presentation. TJ carried out all of the programming and computation. AK wrote the article, which was read and approved by TJ.
Supplementary Material
Additional File 1
Correlation coefficients and supporting data in detail.
Click here for file
Additional File 2
Detailed description of the AVID process. The first three stages of AVID are illustrated by tracing the prediction of functional relationships between (1) YOL137W and YDL138W, (2) YLR375W and YDR463W, and (3) YIR003W and YIL034C. The structures of the MF, BP and CC decision trees are included.
Click here for file
Additional File 3
Connectivity plots comparing the testing and prediction networks.
Click here for file
Additional File 4
AVID MF, BP and CC predictions – refined and novel.
Click here for file
Additional File 5
Number of proteins with each AVID GO term – original and predicted.
Click here for file
Additional File 6
MF level 2, BP level 3, CC level 4 categories.
Click here for file
Additional File 7
Proteins with novel predictions in all three categories: MF, BP and CC
Click here for file
Additional File 8
Contingency tables for the naïve Bayesian analysis
Click here for file
Acknowledgements
We thank MIT and NSF CAREER award MCB-0347203 to AEK for funding, the CSBi high-performance computing technology platform for resources and support, M. Singh, G. Grigoryan and S. Bell for stimulating discussions, G. Grigoryan for assistance with Figure 1, and S. Altman, S. Sia and members of the Keating lab for comments on the manuscript. The AVID web site was developed by K. Weston.
Figures and Tables
Figure 1 Overview of AVID, using prediction of the molecular function of YOL137W as an example. Initially, YOL137W is treated as potentially functionally related to 6,448 other yeast proteins (not shown). Stage 2 prunes this to 30 putatively similar proteins, and the stage 3 decision tree further reduces this list to eight high-confidence neighbors. See additional file 2 for a description of the decision tree. In stage 4, five of the eight neighbors of YOL137W have existing AVID GO annotation (boxed) and "vote" to assign the GO term GO:0005355. After stage 4, proteins with known annotation in the figure are boxed, those with novel predictions are shown in diamonds and the function of YGR224W, in a hexagon, is a refined prediction. The estimated accuracy of predicted functional similarities (in stages 2 and 3) and annotations (in stage 4) are given in italics (see Methods). MF: molecular function; BP: biological process; CC: cellular component. Rosette figures generated using Graphviz [46].
Figure 2 Evaluation of stages 2 and 4 using known proteins. (A) To evaluate the performance of a simple product model in stage 2, PjAVID2 was calculated for each pair of proteins in an annotated reference set. The plot shows the fraction of protein pairs at different PjAVID2 cutoffs that share an AVID GO term. Pairs with PjAVID2 < 12.8 were not considered in stages 3 or 4. (B) A varying percentage of reference proteins was omitted from the entire training procedure and used as a test set. Functions for these proteins were predicted in stage 4, and the plot shows the success rate for correctly predicting at least one existing GO term at the highest level of annotation. The arrows indicate the expected performance for predicting new functions based on the current status of annotation of the yeast genome. Error bars show the standard deviation from 100 random cross-validation trials. MF (diamonds), BP (squares) and CC (triangles).
Figure 3 GO and AVID annotations of proteins localized to an experimentally identified complex. The complex shown at left was identified using co-purification/mass spectrometry by Hazbun et al. [30]. On the right, proteins with known AVID GO terms are shown as circles; proteins with refined AVID predictions are shown as triangles, proteins with novel AVID predictions are shown as squares. Colors represent AVID GO terms as follows: light blue – small nucleolar ribonucleoprotein complex; grey – processing of 20S pre-rRNA; dark blue – 35S primary transcript processing; dark green – ER to Golgi transport; light green – snoRNA binding; red – ATP dependent RNA helicase activity. This complex is predicted to have a role in RNA transcript processing. The refined and novel functional predictions agree with previous annotations and with each other, increasing confidence that these are meaningful assignments.
Figure 4 Comparison of AVID with a naïve Bayesian network. The performance of a naïve Bayesian net, as described by Jansen et al. [21], is compared to that of AVID, using the same input data and same measures of performance. The plots show the ratio of true positives to false positives (TP/FP) vs. coverage (TP/P). The Bayesian net results are in open triangles, AVID stages 1 and 2 in open circles and AVID stages 1, 2 and 3 in closed circles. The performance of AVID is notably superior at higher coverage. At low coverage, all methods can achieve high accuracy.
Table 1 Correlation between experimental or sequence-based measures of protein relatedness and GO annotation similarity.
Dataa MFb BPb CCb
UCSF localization (+) 0.031 0.029 0.058
UCSF localization (-) 0.007 0.007 0.004
yeast 2-hybrid (+) 0.120 0.167 0.254
yeast 2-hybrid (-) 0.010 0.010 0.010
MIPS complex (+) 0.122 0.116 0.263
MIPS complex (-) 0.007 0.008 0.002
sequence similarity (+) 0.187 0.344 0.078
sequence similarity (-) 0.009 0.007 0.010
Microarrayc
-1 < R ≤ -0.9 0.008 0.003 0.006
0.9 < R ≤ 1.0 0.086 0.118 0.125
aExcept for the microarray data, each data source is divided into two sets: one contains protein pairs observed to share the feature described by the data (+), the other contains protein pairs that lack the feature or are not reported (-). bThe normalized conditional probabilities Pij>AVID1 are defined in the Methods; they are the probability of sharing a common GO term given observation of a particular data feature. A perfect correlation would give a value of 1.0. cR is the Pearson correlation coefficient for pairs of mRNA expression profiles, which were binned into 19 intervals. Only two intervals are shown here. See additional file 1 for further details.
Table 2 Summary of AVID prediction performance.
MF BP CC
total predicted proteins 1852 1458 2304
proteins with novel predictions (no existing GO annotation) 950 540 907
proteins with refined predictions (existing GO annotation is less detailed) 902 918 1397
accuracy of stage 3 pair-wise similarities, from 10-fold cross validation 77% 65% 78%
frequency of unannotated proteins in the correlation networks 56.4% 56.5% 70.3%
estimated success rate for AVID GO term assignment ~67% ~52% ~66%
consistency of refined predictions with existing annotation 74.7% 80.3% 86.9%
Table 3 Overlap of pair-wise predictions of functional (or localization) similarity by different methods.
METHODS AVID MARCOTTE STRING PIP_600a LIANG MAGICb MAGICc SCHLITT
AVID 37451 3020 3413 785 2570 2740 49 9
MARCOTTE 33919 9873 3528 2454 1971 66 26
STRING 23245 3614 1740 1819 32 21
PIP_600a 9897 425 260 8 1
LIANG 7963 1647 41 6
MAGICb 7922 397 7
MAGICc 397 1
SCHLITT 526
a. This study was intended to predict pairs of proteins that co-localize to the same complex.
b. MAGIC pairs with p ≥ 0.5
c. MAGIC pairs with p ≥ 0.7
==== Refs
Huh WK Falvo JV Gerke LC Carroll AS Howson RW Weissman JS O'Shea EK Global analysis of protein localization in budding yeast Nature 2003 425 686 691 14562095 10.1038/nature02026
Gavin AC Bosche M Krause R Grandi P Marzioch M Bauer A Schultz J Rick JM Michon AM Cruciat CM Remor M Hofert C Schelder M Brajenovic M Ruffner H Merino A Klein K Hudak M Dickson D Rudi T Gnau V Bauch A Bastuck S Huhse B Leutwein C Heurtier MA Copley RR Edelmann A Querfurth E Rybin V Drewes G Raida M Bouwmeester T Bork P Seraphin B Kuster B Neubauer G Superti-Furga G Functional organization of the yeast proteome by systematic analysis of protein complexes Nature 2002 415 141 147 11805826 10.1038/415141a
Ho Y Gruhler A Heilbut A Bader GD Moore L Adams SL Millar A Taylor P Bennett K Boutilier K Yang L Wolting C Donaldson I Schandorff S Shewnarane J Vo M Taggart J Goudreault M Muskat B Alfarano C Dewar D Lin Z Michalickova K Willems AR Sassi H Nielsen PA Rasmussen KJ Andersen JR Johansen LE Hansen LH Jespersen H Podtelejnikov A Nielsen E Crawford J Poulsen V Sorensen BD Matthiesen J Hendrickson RC Gleeson F Pawson T Moran MF Durocher D Mann M Hogue CW Figeys D Tyers M Systematic identification of protein complexes in Saccharomyces cerevisiae by mass spectrometry Nature 2002 415 180 183 11805837 10.1038/415180a
Ito T Chiba T Ozawa R Yoshida M Hattori M Sakaki Y A comprehensive two-hybrid analysis to explore the yeast protein interactome Proc Natl Acad Sci U S A 2001 98 4569 4574 11283351 10.1073/pnas.061034498
Uetz P Giot L Cagney G Mansfield TA Judson RS Knight JR Lockshon D Narayan V Srinivasan M Pochart P Qureshi-Emili A Li Y Godwin B Conover D Kalbfleisch T Vijayadamodar G Yang M Johnston M Fields S Rothberg JM A comprehensive analysis of protein-protein interactions in Saccharomyces cerevisiae Nature 2000 403 623 627 10688190 10.1038/35001009
Kemmeren P van Berkum NL Vilo J Bijma T Donders R Brazma A Holstege FC Protein interaction verification and functional annotation by integrated analysis of genome-scale data Mol Cell 2002 9 1133 1143 12049748 10.1016/S1097-2765(02)00531-2
Salwinski L Eisenberg D Computational methods of analysis of protein-protein interactions Curr Opin Struct Biol 2003 13 377 382 12831890 10.1016/S0959-440X(03)00070-8
von Mering C Krause R Snel B Cornell M Oliver SG Fields S Bork P Comparative assessment of large-scale data sets of protein-protein interactions Nature 2002 417 399 403 12000970 10.1038/nature750
Ashburner M Ball CA Blake JA Botstein D Butler H Cherry JM Davis AP Dolinski K Dwight SS Eppig JT Harris MA Hill DP Issel-Tarver L Kasarskis A Lewis S Matese JC Richardson JE Ringwald M Rubin GM Sherlock G Gene ontology: tool for the unification of biology. The Gene Ontology Consortium Nat Genet 2000 25 25 29 10802651 10.1038/75556
Huynen MA Snel B von Mering C Bork P Function prediction and protein networks Curr Opin Cell Biol 2003 15 191 198 12648675 10.1016/S0955-0674(03)00009-7
Marcotte EM Pellegrini M Thompson MJ Yeates TO Eisenberg D A combined algorithm for genome-wide prediction of protein function Nature 1999 402 83 86 10573421 10.1038/47048
von Mering C Huynen M Jaeggi D Schmidt S Bork P Snel B STRING: a database of predicted functional associations between proteins Nucleic Acids Res 2003 31 258 261 12519996 10.1093/nar/gkg034
Samanta MP Liang S Predicting protein functions from redundancies in large-scale protein interaction networks Proc Natl Acad Sci U S A 2003 100 12579 12583 14566057 10.1073/pnas.2132527100
Vazquez A Flammini A Maritan A Vespignani A Global protein function prediction from protein-protein interaction networks Nat Biotechnol 2003 21 697 700 12740586 10.1038/nbt825
Letovsky S Kasif S Predicting protein function from protein/protein interaction data: a probabilistic approach Bioinformatics 2003 19 Suppl 1 I197 I204 12855458 10.1093/bioinformatics/btg1026
Karaoz U Murali TM Letovsky S Zheng Y Ding C Cantor CR Kasif S Whole-genome annotation by using evidence integration in functional-linkage networks Proc Natl Acad Sci U S A 2004 101 2888 2893 14981259 10.1073/pnas.0307326101
Schwikowski B Uetz P Fields S A network of protein-protein interactions in yeast Nat Biotechnol 2000 18 1257 1261 11101803 10.1038/82360
Pavlidis P Weston J Cai J Noble WS Learning gene functional classifications from multiple data types J Comput Biol 2002 9 401 411 12015889 10.1089/10665270252935539
Zhang LV Wong SL King OD Roth FP Predicting co-complexed protein pairs using genomic and proteomic data integration BMC Bioinformatics 2004 5 38 15090078 10.1186/1471-2105-5-38
Troyanskaya OG Dolinski K Owen AB Altman RB Botstein D A Bayesian framework for combining heterogeneous data sources for gene function prediction (in Saccharomyces cerevisiae) Proc Natl Acad Sci U S A 2003 100 8348 8353 12826619 10.1073/pnas.0832373100
Jansen R Yu H Greenbaum D Kluger Y Krogan NJ Chung S Emili A Snyder M Greenblatt JF Gerstein M A Bayesian networks approach for predicting protein-protein interactions from genomic data Science 2003 302 449 453 14564010 10.1126/science.1087361
Jansen R Lan N Qian J Gerstein M Integration of genomic datasets to predict protein complexes in yeast J Struct Funct Genomics 2002 2 71 81 12836664 10.1023/A:1020495201615
Tanay A Sharan R Kupiec M Shamir R Revealing modularity and organization in the yeast molecular network by integrated analysis of highly heterogeneous genomewide data Proc Natl Acad Sci U S A 2004 101 2981 2986 14973197 10.1073/pnas.0308661100
Lee I Date SV Adai AT Marcotte EM A probabilistic functional network of yeast genes Science 2004 306 1555 1558 15567862 10.1126/science.1099511
Schlitt T Palin K Rung J Dietmann S Lappe M Ukkonen E Brazma A From gene networks to gene function Genome Res 2003 13 2568 2576 14656964 10.1101/gr.1111403
Snel B Lehmann G Bork P Huynen MA STRING: a web-server to retrieve and display the repeatedly occurring neighbourhood of a gene Nucleic Acids Res 2000 28 3442 3444 10982861 10.1093/nar/28.18.3442
Russell S Norvig P Artificial Intelligence: A Modern Approach 1995 Englewood Cliffs, NJ, Prentice-Hall
Boudreault AA Cronier D Selleck W Lacoste N Utley RT Allard S Savard J Lane WS Tan S Cote J Yeast enhancer of polycomb defines global Esa1-dependent acetylation of chromatin Genes Dev 2003 17 1415 1428 12782659 10.1101/gad.1056603
Saccharomyces Genome Database
Hazbun TR Malmstrom L Anderson S Graczyk BJ Fox B Riffle M Sundin BA Aranda JD McDonald WH Chiu CH Snydsman BE Bradley P Muller EG Fields S Baker D Yates JR Davis TN Assigning function to yeast proteins by integration of technologies Mol Cell 2003 12 1353 1365 14690591 10.1016/S1097-2765(03)00476-3
Nedea E He X Kim M Pootoolal J Zhong G Canadien V Hughes T Buratowski S Moore CL Greenblatt J Organization and function of APT, a subcomplex of the yeast cleavage and polyadenylation factor involved in the formation of mRNA and small nucleolar RNA 3'-ends J Biol Chem 2003 278 33000 33010 12819204 10.1074/jbc.M304454200
Shen X Ranallo R Choi E Wu C Involvement of actin-related proteins in ATP-dependent chromatin remodeling Mol Cell 2003 12 147 155 12887900 10.1016/S1097-2765(03)00264-8
AVID: Annotation Via Integration of Data
MIPS. The MIPS comprehensive yeast genome database (CYGD).
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 Nucleic Acids Res 1997 25 3389 3402 9254694 10.1093/nar/25.17.3389
Ito T Tashiro K Muta S Ozawa R Chiba T Nishizawa M Yamamoto K Kuhara S Sakaki Y Toward a protein-protein interaction map of the budding yeast: A comprehensive system to examine two-hybrid interactions in all possible combinations between the yeast proteins Proc Natl Acad Sci U S A 2000 97 1143 1147 10655498 10.1073/pnas.97.3.1143
Yeast GFP fusion localization database
NCBI Gene Expression Omnibus
Gene Ontology
weka
Quinlan JR Programs for machine learning, C5 1993 San Francisco, Morgan-Kaufmann
String: functional protein association networks
A Bayesian networks approach for predicting protein-protein interactions from genomic data
Common partners of proteins
Bayesian framework for biological data integration - Download
Ellson J Gansner E Koren Y Koutsofios E Mocenigo J North S Woodhull G Dobkin D Alexiev V Lilly B Scheerder J Richard G. D Graphviz
| 15929793 | PMC1177925 | CC BY | 2021-01-04 16:02:48 | no | BMC Bioinformatics. 2005 Jun 1; 6:136 | utf-8 | BMC Bioinformatics | 2,005 | 10.1186/1471-2105-6-136 | oa_comm |
==== Front
BMC BioinformaticsBMC Bioinformatics1471-2105BioMed Central London 1471-2105-6-1391593875210.1186/1471-2105-6-139SoftwarePentaPlot: A software tool for the illustration of genome mosaicism Hamel Lutz [email protected] Olga [email protected] J Peter [email protected] Department of Computer Science and Statistics, University of Rhode Island, Kingston, RI 02881, USA2 Department of Molecular and Cell Biology, University of Connecticut, Storrs, CT, 06269-3125, USA3 Department of Biochemistry and Molecular Biology, Dalhousie University, 5850 College Street, Halifax, NS B3H 1X5, Canada2005 6 6 2005 6 139 139 4 3 2005 6 6 2005 Copyright © 2005 Hamel et al; licensee BioMed Central Ltd.2005Hamel 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
Dekapentagonal maps depict the phylogenetic relationships of five genomes in a visually appealing diagram and can be viewed as an alternative to a single evolutionary consensus tree. In particular, the generated maps focus attention on those gene families that significantly deviate from the consensus or plurality phylogeny. PentaPlot is a software tool that computes such dekapentagonal maps given an appropriate probability support matrix.
Results
The visualization with dekapentagonal maps critically depends on the optimal layout of unrooted tree topologies representing different evolutionary relationships among five organisms along the vertices of the dekapentagon. This is a difficult optimization problem given the large number of possible layouts. At its core our tool utilizes a genetic algorithm with demes and a local search strategy to search for the optimal layout. The hybrid genetic algorithm performs satisfactorily even in those cases where the chosen genomes are so divergent that little phylogenetic information has survived in the individual gene families.
Conclusion
PentaPlot is being made publicly available as an open source project at .
==== Body
Background
Trees have a long history as models for the evolutionary history of organisms [1,2]. Lineage sorting and reticulate evolution have long been recognized as processes that make it difficult to infer species evolution from gene trees [3,4]. However, the extent of gene transfer between divergent species, particularly in case of microorganisms, has initiated a reassessment of the applicability of a tree-based concept for organismal evolution [5,6]. Individual genes coexisting in a present day genome can have very different evolutionary histories [7,8]. In particular, horizontal gene transfer is recognized as an alternative to within lineage processes like duplication and de-novo evolution of genes for an organism to acquire new properties [9]. Here we present a software tool, which computes dekapentagonal maps to depict the phylogenetic relationships of five genomes in a visually appealing diagram as an alternative to bifurcating trees. Dekapentagonal maps allow for the recognition of a plurality or majority signal and they can serve as a visual tool to pre-screen for putative instances of horizontally transferred genes (e.g., see [10]).
Given five genomes we can characterize all possible phylogenetic relationships between the genomes with fifteen different unrooted tree topologies. One way to depict all fifteen relationships is to use a generalization of barycentric coordinates, so called dekapentagonal maps (see below and [10]). The support value vector for a gene family contains the posterior probabilities for each of the fifteen tree topologies given the aligned sequences, or the frequencies with which the fifteen different tree topologies are recovered from bootstrapped samples generated from the aligned sequences. The dekapentagonal map of five genomes depicts the support value vectors for all gene families that have a representative in each of the five genomes. The successful construction of dekapentagonal maps critically depends on an optimal layout of the fifteen different tree topologies along the fifteen vertices of the dekapentagon. Figure 1 is an example of a particular layout of the tree topologies along the dekapentagon's vertices (see [10] for detailed discussion of these analyses). The points within the diagram denote actual data support for particular tree topologies, with each point representing one family of orthologous genes [11,12]. The individual regions within the map demark areas of support for individual topologies. The region in the center of dekapentagonal map represents an area of no support for any particular topology. The resulting display facilitates recognition of frequently supported tree topologies (topologies #5, #10 and #15 in Figure 1) and their shared features (e.g., Chlorobium tepidum (Ct) grouping with Rhodobacter capsulatus (R)). The placement of a support value vector to the inside of the dekapentagon depends on how the fifteen topologies are laid out along the vertices. A gene family that has equal support for only two of the tree topologies will map to the periphery, if these two topologies occupy neighboring vertices, but it will map into the center, if the two topologies occupy opposing vertices. We define as optimal a layout of tree topologies along the vertices, if it minimizes the distance of the support value vectors from the periphery. This way an analysis of genomes related only through strict vertical inheritance will result in a cluster of points neighboring a single vertex; the horizontal transfer of several genes will result in points close to other not necessarily neighboring vertices (e.g. topology #2 in Figure 1), and tree topologies between which the data frequently cannot decide will be neighboring each other.
Figure 1 Dekapentagonal map for the analyses of five photosynthetic genomes: Synechocystis sp. (S), Chloroflexus aurantiacus (Ca), Chlorobium tepidum (Ct), Rhodobacter capsulatus (R) and Heliobacillus mobilis (H), based on posterior probabilities. Each point plotted within the dekapentagon represents a family of orthologous proteins – there are a total of 188 sets of orthologs common to the five genomes [26]). The dekapentagon is divided into zones of proximity to topologies: points that fall into one of the 15 zones that correspond to the 15 tree topologies favor either that topology most or several neighboring topologies, and points that fall into the single central zone represent unresolved relationships. The tree topology number (1 to 15) is given first, followed by the number of points per zone in parentheses. Abbreviations: Ca, Chloroflexus aurantiacus; Ct, Chlorobium tepidum; H, Heliobacillus mobilis; R, Rhodobacter capsulatus. This figure was previously published in [10].
Computing the optimal layout of the tree topologies along the vertices of the dekapentagon presents a difficult optimization problem given the large number of possible layouts: (15-1)!/2 = 14!/2≈4*1010 (only free circular permutations [13] are counted, and the arrangements that become equivalent by rotation or flipping of the dekapentagon are considered the same arrangements).
This setting seems ideally suited for optimization based on genetic algorithms [14]. At the core of the software tool presented here is the design and implementation of a hybrid genetic algorithm which computes optimal tree topology layouts along the dekapentagon vertices using demes in order to avoid premature convergence and which employs a local search strategy to complement the global search behavior of the crossover and mutation operators.
Implementation
Design overview
PentaPlot is written as a program, which is comprised of multiple processing steps implemented both in Perl and C++. The processing steps are hidden from the user by means of a master Perl script that ties all the processing steps together. In normal usage the user prepares a probability matrix (see software documentation for formatting details), which provides information about the support of particular tree topologies by the families of orthologous genes. This matrix is processed by the program into the visual dekapentagonal map as shown in Figure 1. The individual processing steps are as follows:
• Compute tree topology layout from probability matrix.
• Map the polar coordinates for orthologous gene family into a Cartesian coordinates.
• Summarize the number of genes, which fall into the individual zones that support particular tree topologies.
• Construct the dekapentagonal map.
The generated dekapentagonal map is available as an image, which either can be viewed in an interactive previewer or saved in the post-script and PDF formats. The implementation heavily relies on Wall's genetic algorithm C++ component library [15] and TeX [16].
PentaPlot also provides access to a number of tuning parameters for the construction of the maps, which are accessible via command line arguments:
• Iterations (default 50): Optimizing tree topology layouts with a genetic algorithm is a stochastic approach, therefore, in order to obtain some confidence in the computed solutions the solutions should be recomputed a number of times. This parameter controls how often the computation is to be repeated.
• Populations (default 10): A fundamental concept in genetic algorithms is the notion of a population. Here we apply a genetic algorithm that utilizes multiple populations at the same time in order to prevent premature convergence. This parameter specifies how many populations the algorithm should use.
• Population sizes (default 30): Population sizes are critical in genetic algorithms. If the population size is too small there will not be enough genetic diversity in the population to effectively explore the search space. If the population size is too large then a large amount of computation time might be wasted. This parameter controls the sizes of the populations.
• Maximum number of generations (default 500): The genetic algorithm as implemented in PentaPlot has an automatic stopping criteria built in based on 99% convergence within 50 generations, that is, if the performance of the best layout of the current generation T and the performance of the best layout of the generation T-50 are within 1% from each other then the genetic algorithm terminates. However, in particularly difficult optimization landscapes this convergence might never occur and the genetic algorithm might run forever (or at least for a very long time). To avoid this situation the maximum generation parameter allows the user to limit the number of generations the genetic algorithm is allowed to compute.
Data preparation: The probability matrix
The first step in the phylogenetic analysis of genomes is the detection of sets of orthologous genes from the genomes making up the set of five genomes under consideration, i.e., genes that share common ancestry and are related through speciation and not gene duplication events. We use reciprocal top scoring hits as a criterion to select orthologous genes, i.e., each of the genes picks the other members of the quintet as the top scoring hit in a BLAST search [12,17,18]. While this selection scheme aims at detecting orthologous genes, the resulting sets can only be considerate to be putatively orthologous. While this selection scheme aims at detecting orthologous genes, the resulting sets only can be considered to be putatively orthologous. These gene sets include horizontally transferred genes (xenologs), especially those instances, where the transferred gene replaced its ortholog, synologs that were brought into a single genome through lineage fusion [19,20] and duplicated genes where one or the other paralog was lost in each the analyzed genomes [11]. When we refer to sets of orthologous genes in the rest of the manuscript we mean that those orthologs are putative. Once these putatively orthologous genes are detected, they are aligned and all possible unrooted tree topologies are evaluated (fifteen topologies for five genomes) and either their posterior probabilities or bootstrap support values are calculated (see [12] for details on methodology). Therefore, each family of putatively orthologous genes is associated with a 15-dimensional support value vector. This construction results in probability matrices where each row represents a family of orthologous genes and each column represents a particular unrooted tree topology. A value in a matrix represents the probability of support for a particular tree topology by a particular gene family. It is important to note that all the probabilities in one record have to sum up to one. Any other method that calculates support value vectors can be used to produce valid probability matrices. Please note that the construction of the probability matrix is a preprocessing step and is not included in the PentaPlot program.
Mapping of probabilities into barycentric coordinate systems
Barycentric coordinate systems (coordinate systems based on centers of gravity) are most easily explained using triangles instead of generalized polygons. A simple probability matrix is shown in Table 1.
Table 1 A simple probability matrix (see Figure 1 and text for more details)
Tree #1 (T1) Tree #2 (T2) Tree #3 (T3)
Support value vector for a set of orthologous genes P1 P2 P3
With the tree topologies at the vertices of a triangle we can interpret the probabilities as weights at each corner of the triangle with the restriction that P1 + P2 + P3 = 1. We can then visualize this as shown in Figure 2. It is interesting to note that, given the particular assignment of the tree topologies to the vertices, the point P in the area of the triangle is completely defined by the support at each vertex, that is, the point P represents the center of gravity of this construction. We can now interpret the point P as a visual characterization of the support for each of the tree topologies by the set of orthologous genes in the example dataset above. If our dataset had more than one record (e.g., Table 2) we would see multiple points in the triangle, one for each set of orthologous genes. Figure 3 shows a visualization of a dataset with numerous orthologous genes and their support of three different tree topologies involving four genomes A, B, C, and D, respectively. The areas at the vertices demark the regions of support for the individual tree topologies.
Table 2 A three dimensional probability matrix that is used to calculate support value map depicted in Figure 3 where n is the number of orthologous genes considered. An analogous matrix with 15 columns can be constructed for 15-dimensional support value vectors.
Tree #1 (T1) Tree #2 (T2) Tree #3 (T3)
Support value vector for a set #1 of orthologous genes P11 P12 P13
Support value vector for a set #2 of orthologous genes P21 P22 P23
... ... ... ...
Support value vector for a set #n of orthologous genes Pn1 Pn2 Pn3
Figure 2 A barycentric coordinate system with three coordinates. See text and [12, 25] for more details.
Figure 3 An example (modified from [11, 12]) of visualization of the support of three different unrooted tree topologies by multiple sets of orthologous genes from four genomes [12, 25].
We can generalize this to the case of the dekapentagon as shown in Figure 4. Similarly to the case of the triangle in Figure 2, we only show the construction of the center of gravity of a single set of orthologous genes. Again, the weights at the vertices represent the support of a particular tree topology by the set of orthologous genes. The points Mij denote the centers of gravity due to the weights indicated by the subscripts. When we include all the weights in this construction we will obtain a unique M that represents the center of gravity of all the weights for a particular arrangement of the weights along the vertices of the dekapentagon. Similar to the previous case we can interpret the center of gravity as a visual representation of the support for the tree topologies by this set of orthologous genes. Should there be multiple sets of orthologous genes in the dataset we will obtain multiple centers of gravity, one for each set of orthologous genes.
Figure 4 Schematic presentation of calculating and plotting support value vectors into a dekapentagon. Support values associated with each vertex are represented as weights attached to the vertices. Points M indicate locations of center of gravities of vertices that are mentioned in the index associated with each point M. See Implementation section for details of the calculation of the coordinates. This figure was previously published in [10].
Computing the center of gravity in a dekapentagon
Computing the center of gravity in a triangle is straightforward; however, the same computation in a dekapentagon deserves some remarks. To construct the center of gravity we place the dekapentagon into a Cartesian coordinate system with its origin coinciding with the center of the dekapentagon. Then the Cartesian coordinates (xi, yi) of a vertex i can be computed with the equations:
where R is the distance from origin to the vertex (equal for all the vertices due to the location of the origin of the coordinate system and here we assume that R is equal to 1), and 1 ≤ i ≤ 15. For each pair of vertices i and j the coordinates (xM, yM) of the center of gravity Mij are calculated according to the law of the lever:
where pi and pj represent the support ("weights") at the vertices i and j, respectively. The process is repeated for all pairs of vertices, and then iteratively for all "intermediate" centers of gravities until only one pair of coordinates remains, which represents the center of gravity of the dekapentagon using all the "weights". The resulting coordinates of the dekapentagon's center of gravity do not depend on the order in which the masses are combined but they do depend on the particular arrangement of the support p along the vertices, i.e. on the way the 15 different tree topologies are assigned to the vertices. For example, one could envision a family of orthologous genes that supports only two of the fifteen possible tree topologies. If these two topologies were assigned to two opposing vertices, the support value vector would map to the center of the dekapentagonal map, indicating no particular support for any tree topology, whereas if the two topologies are assigned to neighboring vertices, the support value vector will map onto the periphery between these two vertices, revealing support for these two topologies over the other 13 alternatives. Therefore, it is crucial to compute a layout that moves the centers of gravity as close to the periphery as possible.
As mentioned above, in our case we not only compute a single center of gravity or barycentric point, but instead, given a data set with N orthologous genes we will compute N different barycentric points, one for each record in the data set.
The layout algorithm
There are about 4*1010 possible arrangements of topologies on a dekapentagon's vertices. An arrangement is considered optimal when the topologies are arranged at the polygon vertices in a way that maximizes the sum of the distances of all barycentric points from the center of the polygon. There are too many arrangements of topologies around the dekapentagon to search for the optimal arrangement exhaustively. Therefore, we used a heuristic search based on genetic algorithms [21].
In the genetic algorithm setup, each generation consists of a population of arrangements where each individual in a population encodes a particular mapping of the possible tree topologies identified by numerical identifiers (1 through 15) to vertices of the dekapentagon. The fittest individuals in a population maximize the sum of all distances of the barycentric points from the center of the polygon. As is typical in evolutionary computation, the genetic algorithm applies mutation and crossover operations to each successive generation of arrangements until an optimal solution is obtained [14]. Genetic algorithms today provide many different implementation strategies beyond the basic bit string genetic algorithm first developed by Holland [14]. We chose an array-based, hybrid genetic algorithm that uses demes to avoid premature convergence [15].
Genetic algorithms are good at finding approximate solutions in large search spaces but they are notoriously inefficient when it comes to fine tuning these solutions. By equipping a genetic algorithm with a local search strategy we avoid these problems. This is referred to as hybrid genetic algorithms [21] (sometimes also referred to as memetic algorithms [22]). Our hybrid genetic algorithm is summarized by the following pseudo code:
function evolve
create initial population
do
// perform crossover and mutation
population : = compute-new-population (population)
best : = fittest-individual (population)
optimized : = local-optimization (best)
// if optimized is fitter than best replace best
// with optimized in the population
if (fitness (optimized) > fitness (best))
replace (population,best,optimized)
until (stopping-condition)
return fittest-individual (population)
This algorithm is replicated over the demes giving rise to our hybrid deme genetic algorithm. It is noteworthy that we deviate from the standard notion of hybrid genetic algorithm slightly by only applying the local optimization function to the fittest individual of the population at each generation in each deme due to the computational cost of our local optimization: given the tree topology layout of the fittest individual, our local optimization strategy attempts to find an even better layout by systematically swapping tree topologies in the layout. The pseudo code of the local search heuristic follows:
function local-optimization (layout [1..15])
bestScore : = bestSource : = bestTarget : = -1
for s : = 1 to 14 do
for t : = s+1 to 15 do
swap-topologies (layout [s], layout [t])
score : = scoreArrangement (layout)
if (score > bestScore)
bestScore : = score
bestSource : = s
bestTarget : = t
endif
undo-swap (layout [s], layout [t])
endfor
endfor
swap-topologies (layout [bestSource], layout [bestTarget])
return layout
This procedure steps through all topologies in a layout and evaluates the score obtained by swapping each topology with all other positions. Here 'scoreArrangement' is a procedure that computes the barycentric coordinates for each record in the data set and the score is derived from the sum of all the distances of the barycentric points from the center of the polygon. Most likely we can improve the computational requirements of this local search procedure by sampling possible swaps in the layouts instead of trying them all [23].
Another key notion beyond the local search strategy is that we have to constrain the structure of the individuals in the populations in such a way that each individual can only encode legal sequences of topologies, that is, each individual can only encode layouts of topologies around the perimeter of the polygon that do not have repetitions. This is analogous to the term closure condition that arises in genetic programming where any term constructor combined with any other legal term constructor must give rise to a legal term [24]. Here we opted for an array representation where each position in the array denotes a vertex on the polygon. The contents at each array position denote a tree topology assigned to that vertex. Each individual is initialized in such a way that the tree topologies 1 through 15 are assigned to the vertices in such a way that there are no repetitions. In order to make this work the crossover and mutation operators have to preserve the uniqueness property of the topology layouts. Goldberg's PMX (partially matched crossover) operator [21] and Wall's swap mutation implemented in GALIB [15] fulfill our uniqueness requirement and have been implemented in PentaPlot.
Results and discussion
We tested the design of our algorithm with four experiments of increasing difficulty [10]. Each experiment involved the comparison of five genomes. We applied both posterior probability mapping and bootstrap support value mapping [11,12,25] to two different genome quintets:
1. An inter-domain genome quintet consisting of representatives of all three domains of life: Saccharomyces cerevisiae (Y), Rhodobacter capsulatus (R), Bacillus subtilis (B), Archaeoglobus fulgidus (A), Sulfolobus solfataricus (S).
2. Bacterial genomes representing the five phyla that contain organisms with chlorophyll-based photosynthesis: Chlorobium tepidum (Ct), Chloroflexus aurantiacus (Ca), Heliobacillus mobilis (H), Rhodobacter capsulatus (R), Sulfolobus solfataricus (S).
The two datasets resulting from the first genome quintet each had 53 records, that is, 53 families of orthologous genes with one representative in each of the five genomes. The datasets resulting from the second genome quintet each had 188 records. Our investigation reported in [10] corroborated the layouts produced by our algorithm. The increase in difficulty in these experiments arises from the fact that (a) maximum likelihood mappings tend to produce barycentric points which lay close to the circumference of the polygon making it more difficult to discern an optimal layout and (b) the bacterial genomes contained a large number of orthologous genes, that is, there were a large number of barycentric points that needed to be considered during optimization.
We also compared the performance of our hybrid deme genetic algorithm to other genetic algorithm implementations such as the binary string genetic algorithm, the array based genetic algorithm, and a deme based genetic algorithm in each of these four experiments. The population size of the genetic algorithms in all the experiments for all genetic algorithms was 300 individuals. In the case of the deme configurations this population was distributed over 10 subpopulations. We used a convergence stopping criterion of 99% with a window of 50 generations. Typical runs lasted between 50 and 70 generations. The convergence percentages were computed by averaging the number of times a genetic algorithm computed exactly the same layout over the fifty runs. We postulate that a high degree of reproducibility indicates either a global optimal solution or a very strong local minimum, which can be considered a quasi-optimal solution. Given that the reproducible solutions found by the genetic algorithms were corroborated in bipartition analyses [10] we are confident that the genetic algorithms did converge on a global optimum. The results are summarized in Table 3.
Table 3 Application of the different genetic algorithms to four experiments with increasing difficulty (the percentages indicate reproducability of solutions over fifty runs).
Experiments \ GA Type Binary Array Deme Hybrid Deme
1) Inter-domain Bootstrap 0% 0% 100% 100%
2) Inter-domain ML Mapping 0% 0% 84% 84%
3) Bacterial Bootstrap 0% 0% 62% 72%
4) Bacterial ML Mapping 0% 0% 56% 72%
In the case of the binary string and array genetic algorithms it was interesting to see that we did not achieve reproducible solutions. Introducing demes into the genetic algorithms produced the most dramatic performance jump as can be seen from Table 3. In the case of the first experiment the performance jumped from 0% to 100% and dropped off with increasing difficulty of the experiments. The deme genetic algorithm performed well over the range of the experiments. However, in the fourth experiment it only converged on an optimal solution on every other run. The fourth column shows the performance of our hybrid deme genetic algorithm. We can see that it shares the performance characteristics of the deme genetic algorithm on the easier experiments but the performance of the hybrid deme genetic algorithm did not degenerate as fast with increasing difficulty of the experiments.
Tables 4 and 5 summarize the performance of the hybrid deme genetic algorithm using parameters other than the default parameters. In these experiments we applied our hybrid deme genetic algorithm to the maximum likelihood mapping of the bacterial genomes (our most difficult experiment). Table 4 shows the convergence behavior of the genetic algorithm given different number of populations and different sizes of the populations. Here, convergence is defined as above. What is most intriguing here is that bigger populations are not necessarily better. One possible explanation for this premature convergence might be the fact that the single best individual we are selecting in the local search for improvement does not have as much an impact on the large populations as it does in smaller populations. Thus, larger populations are more prone to early stagnation in their search. Table 5 highlights the fact that the deme idea is indeed very important to this algorithm in order to prevent premature convergence. As can be seen from both tables, our default values present a reasonable tradeoff between convergence behavior and computational complexity implied in larger number of populations or large populations.
Table 4 Investigating the genetic algorithm with different population and size settings.
Pop. Number \ Pop. Size 10 20 30 50 80 100 300 500
1 12% 24% 26% 34% 44% 44% 20% 10%
2 36% 28% 38% 42% 54% 52% 16% 12%
3 24% 34% 46% 70% 60% 56% 24% 12%
Table 5 Assuming population sizes of 30 and 50, investigating the behavior of the genetic algorithm with different numbers of populations.
Pop. Number \ Pop. Size 30 50
2 38% 42%
4 52% 68%
8 60% 78%
10 72% 88%
15 82% 88%
20 82% 88%
Conclusion
Dekapentagonal maps provide an alternative to a single evolutionary tree to visualize phylogenetic relationships between organisms. Here we presented a tool, which computes such dekapentagonal maps given an appropriate probability matrix. The visualization critically depends on the optimal layout of unrooted tree topologies along the vertices of the dekapentagon. Given the large number of possible layouts, this represents a difficult optimization problem well suited for genetic algorithms. At its core our tool utilizes a genetic algorithm with demes and a local search strategy. The chosen optimality criterion moves the individual barycentric points representing orthologous genes as far to the periphery as possible. The resulting arrangement places tree topologies between which individual data sets frequently do not decide next to each other. The developed hybrid genetic algorithm performs satisfactorily even in those cases where the chosen genomes are so divergent that little phylogenetic information has survived in the individual gene families.
Availability and requirements
• Project name: PentaPlot
• Project home page:
• Operating system(s): linux, kernel version 2.4.18 and above
• Programming language: Perl, C++, LaTeX
• Other requirements: Perl 5, BioPerl 1.4, LaTeX 2e, GAlib 2.4.5
• License: GNU GPL
• Any restrictions to use by non-academics: contact authors
Authors' contributions
LH wrote the code for the PentaPlot program. OZ and JPG developed overall methodology for dekapentagonal mapping. OZ prepared probability matrices for analyzed genome quintets. All authors contributed equally to the manuscript preparation.
Acknowledgements
This work was supported through NASA's Exobiology (NAG5-11470) and Applied Information Systems Research (NNG04GP90G) Programs, and in part through the NSF award MCB-0237197 to JPG.
==== Refs
Darwin C On the Origin of Species by Means of Natural Selection, or the Preservation of Favoured Races in the Struggle for Life 1859 London: John Murray, Albemarle Street
Lamarck J-B Histoire naturelle des animaux sans vertebras Paris 1815
Gogarten JP Murphey RD Olendzenski L Horizontal gene transfer: pitfalls and promises Biol Bull 1999 196 359 361 10390834
Li W-H Molecular Evolution 1997 Sunderland, MA: Sinauer Associates
Doolittle WF Phylogenetic classification and the universal tree Science 1999 284 2124 9 10381871 10.1126/science.284.5423.2124
Olendzenski L Zhaxybayeva O Gogarten JP Seckbach J What's in a Tree? Does Horizontal Gene Transfer Determine Microbial Taxonomy? Cellular Origin and Life in Extreme Habitats Vol 4: Symbiosis 2001 Netherlands: Kluwer Academic Publishers
Koonin EV Makarova KS Aravind L Horizontal gene transfer in prokaryotes: quantification and classification Annu Rev Microbiol 2001 55 709 42 11544372 10.1146/annurev.micro.55.1.709
Hilario E Gogarten JP Horizontal transfer of ATPase genes – the tree of life becomes a net of life Biosystems 1993 31 111 9 8155843 10.1016/0303-2647(93)90038-E
Gogarten JP Doolittle WF Lawrence JG Prokaryotic evolution in light of gene transfer Mol Biol Evol 2002 19 2226 38 12446813
Zhaxybayeva O Hamel L Raymond J Gogarten JP Visualization of Phylogenetic content of Five Genomes with Dekapentagonal Maps Genome Biology 2004 5 R20 15003123 10.1186/gb-2004-5-3-r20
Zhaxybayeva O Gogarten JP An Improved Probability Mapping Approach to Assess Genome Mosaicism BMC Genomics 2003 4 37 12974984 10.1186/1471-2164-4-37
Zhaxybayeva O Gogarten JP Bootstrap, Bayesian probability and maximum likelihood mapping: Exploring new tools for comparative genome analyses BMC Genomics 2002 3 4 11918828 10.1186/1471-2164-3-4
Weisstein EW Circular Permutations, In: Eric Weisstein's World of Mathematics
Holland JH Adaptation in Natural and Artificial Systems 1975 University of Michigan Press
Wall M GALIB: A C++ Library of Genetic Algorithm Components. Distributed through 1996
Knuth DE The TeXbook 1986 Reading, Mass.: Addison-Wesley
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 Nucleic Acids Res 1997 25 3389 402 9254694 10.1093/nar/25.17.3389
Montague MG Hutchison CA 3rd Gene content phylogeny of herpesviruses Proc Natl Acad Sci U S A 2000 97 5334 9 10805793 10.1073/pnas.97.10.5334
Fitch WM Homology a personal view on some of the problems Trends Genet 2000 16 227 31 10782117 10.1016/S0168-9525(00)02005-9
Gogarten JP Which is the most conserved group of proteins? Homology-orthology, paralogy, xenology, and the fusion of independent lineages J Mol Evol 1994 39 541 3 7807544 10.1007/BF00173425
Goldberg DE Genetic Algorithms in Search, Optimization and Machine Learning 1989 Addison-Wesley
Moscato P An introduction to population approaches for optimization and hierarchical objective functions: The role of tabu search Annals of Operations Research 1993 41 85 121
Krasnogor N Smith J Emergence of Profitable Search Strategies Based on a Simple Inheritance Mechanism Proceedings of the Genetic and Evolutionary Computation Conference 2001 432 439
Koza JR Genetic programming: on the programming of computers by means of natural selection 1992 Cambridge, Mass.: MIT Press
Strimmer K von Haeseler A Likelihood-mapping: a simple method to visualize phylogenetic content of a sequence alignment Proc Natl Acad Sci, USA 1997 94 6815 6819 9192648 10.1073/pnas.94.13.6815
Raymond J Zhaxybayeva O Gogarten JP Gerdes SY Blankenship RE Whole-genome analysis of photosynthetic prokaryotes Science 2002 298 1616 20 12446909 10.1126/science.1075558
| 15938752 | PMC1177926 | CC BY | 2021-01-04 16:27:46 | no | BMC Bioinformatics. 2005 Jun 6; 6:139 | utf-8 | BMC Bioinformatics | 2,005 | 10.1186/1471-2105-6-139 | oa_comm |
==== Front
BMC BioinformaticsBMC Bioinformatics1471-2105BioMed Central London 1471-2105-6-1501596085410.1186/1471-2105-6-150SoftwareAnovArray: a set of SAS macros for the analysis of variance of gene expression data Hennequet-Antier Christelle [email protected] Hélène [email protected] Karine [email protected] Séverine [email protected] Isabelle [email protected] Jean-Paul [email protected] François [email protected] Stéphane [email protected] Unité Mathématique, Informatique et Génome, INRA, 78352 Jouy-en-Josas, France2 INA-PG / INRA 16 rue Claude Bernard, F-75231 PARIS cedex 05 – France3 Unité Biologie du Développement et de la Reproduction, INRA, 78352 Jouy-en-Josas, France2005 16 6 2005 6 150 150 18 1 2005 16 6 2005 Copyright © 2005 Hennequet-Antier et al; licensee BioMed Central Ltd.2005Hennequet-Antier 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
Analysis of variance is a powerful approach to identify differentially expressed genes in a complex experimental design for microarray and macroarray data. The advantage of the anova model is the possibility to evaluate multiple sources of variation in an experiment.
Results
AnovArray is a package implementing ANOVA for gene expression data using SAS® statistical software. The originality of the package is 1) to quantify the different sources of variation on all genes together, 2) to provide a quality control of the model, 3) to propose two models for a gene's variance estimation and to perform a correction for multiple comparisons.
Conclusion
AnovArray is freely available at and requires only SAS® statistical software.
==== Body
Background
Macroarray and microarray experiments are powerful technologies to simultaneously study thousands of genes. These technologies are efficient to identify differentially expressed genes between two or more samples (conditions). However the complexity of experiments may require an experimental design including different kinds of repeats. Analysis of variance (ANOVA) is a well suited statistical method to analyse gene expression depending on several sources of variation ([1-3]). It permits to decide which effects are important and to estimate them. One great interest of anova is to estimate the variations due to the gene factor taking into account interactions with other experimental factors.
Recently, several tools have been developped to perform anova on gene expression data (library YASMA [4] written in R, MAANOVA [5] written in R and Matlab, GeneANOVA [6] written in JAVA,...). Compared to these tools, we propose three improvements : a control quality procedure, a choice of two models for gene's variance estimation, and an additional correction for multiple comparisons.
Implementation
Language selection
The AnovArray package is a flexible tool to perform analysis of variance. It is developped in SAS® statistical language in order to take advantage of statistical capabilities and powerful data management. Thanks to the SAS® environment, AnovArray makes it possible to quantify gene effect and their interactions with other factors assuming all genes have the same variance. The anova method implemented in R is based on a matrix inversion and could not produce results on a large dataset with thousands of genes. For balanced factorial designs, the anova method implemented in SAS® provides the anova table and some statistics (means, tests of effects,. . .). Our package uses these statistics and gives additional ones in order to quantify and estimate the factors, and also to identify differentially expressed genes. Moreover, AnovArray package benefits of all SAS® possibilities (clustering, supervised classification, singular value decomposition, regression, etc...) for further analyses.
Models of anova
AnovArray is a set of five SAS® macros: global_analysis, adjust, cleandata, differential_analysis and comparison. It is dedicated to analysis of variance in the case of a balanced experimental design. A user-defined analysis of variance model is used for the macros global_analysis and differential_analysis. In the macro global_analysis, the anova model includes factor "Gene" and their interactions with other factors. It is performed on all genes together making the model more powerful to estimate experimental factors than if the model was defined for each gene. If necessary, several models can be considered by the user in order to test the impact of possible experimental factors.
To identify differentially expressed genes, AnovArray offers the choice between two variance models: an homogeneous model (HOM) and an heterogeneous model (HET). All genes have the same variance in the homogeneous model whereas each gene has its own variance in the heterogeneous model. If the hypothesis HOM is acceptable, the power of differentially expressed genes detection is greater, especially when only few repeats are available.
Quality control
AnovArray contains facilities to evaluate the quality of the anova model defined by the user in the macro global_analysis. First, the classical anova table gives an estimation of the variability due to each factor. This table permits the user to classify the most important factors. Second, the model assumptions can be checked graphically (figure 2) by the user. The model assumes that residuals follow a gaussian distribution, are independent and have the same variance. Typically, figure 2a, permits to check that the residuals close to zero and do not show any structure. Figure 2b describes two kinds of graphs to check the residuals distribution: by fitting a gaussian distribution on residuals histogram and by using a Q-Q plot (plot of residuals distribution quantiles versus gaussian distribution quantiles). The hypothesis that residuals have the same variance (and also genes) is checked by fitting a chi-squared distribution (figure 2c). These graphs can also be very useful to describe which experimental factor affects particular genes. If the model is well adapted, the user could carry on the identification of differentially expressed genes. If not, the user has either to identify a subgroup of genes which do not satisfy the assumptions of the model (atypical genes), or to reconsider the factors included in the model. So thanks to AnovArray, it is possible to correct undesirable effets identified in the graphical representation using adjust macro and to remove a group of genes having an atypical behaviour using cleandata macro.
Figure 2 Graphs produced by the macro global_analysis. Residuals versus predicted, Gaussian distribution and Distribution chi2 of variances.
Way of use
The five macros of the package can be used either independently or in a concerted way as indicated in the strategy analysis described in figure 1. The anova model is defined in the macro global_analysis by the user. This macro computes the classical anova table which permits to identify factors which are important to explain observed differences in gene expression. As explained in the previous section, several graphs described in figure 2 are available to check model assumptions: variance homogeneity and gaussian distribution of residuals. These graphs can also be very useful to highlight which experimental factor affects a sub-population of genes. Several models can be tested and the quality control facilities (statistics in the table of anova, graphs) permit to select which one is the more accurate. Depending on the results given by the macro global_analysis, it could be necessary to use macros adjust and cleandata. The macro adjust will then permit to systematically remove undesirable effects (factors) observed in graphs obtained by the macro global_analysis. In the same manner, the macro cleandata makes it possible to remove genes which do not respect the assumptions of the model. We advise to use this iterative process (global_analysis, cleandata and adjust) before using the macro differential_analysis. The aim of this process is to make sure that data are well fitted by the model and that model assumptions are satisfied. This process is very important to get reliable results on differentially expressed genes.
Figure 1 AnovArray Package. AnovArray package: map of the use of the major macros.
As explained in the previous section, the package also permits the differential analysis under two hypotheses: either genes have equal variance (homogeneous model – HOM) or each gene has its own variance (heterogeneous model – HET). The macro differential_analysis produces the list of genes differentially expressed between several experimental conditions using p-values and adjusted p-values statistics. A p-value is defined as the probability of rejecting the null hypothesis {The interaction gene x condition is null.}, if true. P-values are calculated for each gene under the hypothesis that all genes have the same variance and under the hypothesis that each gene has its own variance. By using the correction for multiple comparisons FDR [7](False Discovery Rate), a gene is differentially expressed if its adjusted p-value is lower than a significance level given by the user.
Finally, the macro comparison enables to compare graphically the results obtained by the two models of variance. In a way, the plot of adjusted p-values under hypothesis of homogeneous variance versus adjusted p-values under hypothesis of heterogeneous variance indicates the genes which probably do not satisfy the homogeneity of variance hypothesis.
To summarize AnovArray package permits successive usage of different anova models (principal effects and their interactions) in order to construct an adaptive approach of gene expression data analysis.
Results and discussion
AnovArray has been used for the analysis of a macroarray dataset resulting on the hybridisation of 72 membranes. This dataset contains the level of expression of 1920 bovine embryonic cDNA pieces under three conditions of complex sample preparation into two tissues (ovary, brain). The package turned out to be useful and rapid to identify differentially expressed genes between both tissues and three protocols of complex sample preparation (Degrelle et al., manuscrit in preparation). In particular, the anova model emphasizes the existence of an interaction between gene and sample preparation method, between gene and tissue and between gene and sample and tissue. This analysis highlights that the sample preparation could affect differential analysis results.
The dataset described in the manual is a subset of the previous one. The frame of the experiment is conserved and only hybridisation on six membranes have been retained. Three samples obtained from bovine tissues A and B are hybridised on one macroarray membrane. This dataset containing 1525 cDNA was created to give an usage of the AnovArray package. The aim of this analysis is the detection of genes differentially expressed between tissues. The anova model used is
Ytgi = μ + αt + βg + γtg + εtgi
where Ytgi is the expression (transformed by logarithm in base 2) of the ith observation of the gene g in tissue t, μ is the mean effect, αt is the effect of tissue t with t in {A, B}, βg is the effect of gene g with 1525 levels, γtg the interaction between tissue t and gene g, and εtgi is the residual error. The model assumes that the residuals εtgi are independent and normally distributed with equal variance and mean zero (εtgi ~ (0, σ2)) if variance is homogeneous, or (εtgi ~ (0, )) if variance is heterogeneous. To perform a differential analysis, we test the null hypothesis {the interaction γtg is null}. The Fisher statistic is calculated in homogeneous (resp. heterogeneous) model using the variance σ2 (resp. ). The power of the Fisher test depends on the accuracy of the variance estimation and at least six measures are necessary to estimate . Two genes are found differentially expressed between the two tissues with the hypothesis of homogeneity of variance and none with the hypothesis of heterogeneity of variance. Methods and statistics are described in a user's guide available at .
Conclusion
We have presented a tool for analysis microaray and macroarray based on the analysis of variance. This package contains some useful graphs to describe and analyse microarray and macroarray data. It permits to evaluate the source of bias, the assumptions underlying the model (distribution of residuals, distribution of variances). It gives also the list of differentially expressed genes between more than two conditions using the false discovery rate (FDR).
Our future development will concern an extension to mixed models and an addition of other multiple correction methods.
Availability and requirements
Project name: AnovArray: a set of SAS macros for the analysis of variance of gene expression data.
Project home page:
Operating system(s): Platform independent
Programming language: SAS®
Other Requirements: SAS® release 8.01 with modules BASE SAS, SAS/Stat and SAS/Graph.
Licence:
Any restrictions to use by non-academic user: citation
Authors' contributions
KP developed the software. CHA, HC, FR and SR conceived the study, participated in its design and coordination. SD, IH, JPR provided experiment dataset and tested software.
==== Refs
Kerr M Martin M Churchill G Analysis of Variance for Gene Expression Microarray Data Journal of Computational Biology 2000
Sekowska A Robin S Daudin J Hénaut A Danchin A Extracting biological information from DNA arrays: an unexpected link between arginine and methionine metabolism in Bacillus subtilis Genome Biology 2001
Draghici S Data Analysis Tools for DNA Microarrays, CHAPMAN & HALL/CRC 2003 7 155 187
library YASMA
MAANOVA software
Didier G Brézellec P Remy E Hénaut A GeneANOVA-gene expression analysis of variance Bioinformatics 2002
Benjamini Y Hochberg Y Controlling the false discovery rate: A pratical and powerful approach to multiple testing Journal of The Royal Statistical Society B 1995
| 15960854 | PMC1177927 | CC BY | 2021-01-04 16:27:46 | no | BMC Bioinformatics. 2005 Jun 16; 6:150 | utf-8 | BMC Bioinformatics | 2,005 | 10.1186/1471-2105-6-150 | oa_comm |
==== Front
BMC BiotechnolBMC Biotechnology1472-6750BioMed Central London 1472-6750-5-181596085010.1186/1472-6750-5-18Methodology ArticleCell membrane array fabrication and assay technology Yamazaki Victoria [email protected] Oksana [email protected] Robert J [email protected] Luat [email protected] Thomas [email protected] Lore [email protected] Jay T [email protected] Synamem Corporation, 863 Mitten Road – Suite 101, Burlingame, CA94010, USA2 Division of Medical and Biochemical Microbiology, Research Center Borstel, Center for Medicine and Biosciences, Borstel, Germany3 Department of Chemistry, 109 Lewis Hall, University of California – Berkeley, Berkeley, CA 94720, USA2005 16 6 2005 5 18 18 31 3 2005 16 6 2005 Copyright © 2005 Yamazaki 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
Microarray technology has been used extensively over the past 10 years for assessing gene expression, and has facilitated precise genetic profiling of everything from tumors to small molecule drugs. By contrast, arraying cell membranes in a manner which preserves their ability to mediate biochemical processes has been considerably more difficult.
Results
In this article, we describe a novel technology for generating cell membrane microarrays for performing high throughput biology. Our robotically-arrayed supported membranes are physiologically fluid, a critical property which differentiates this technology from other previous membrane systems and makes it useful for studying cellular processes on an industrialized scale. Membrane array elements consist of a solid substrate, above which resides a fluid supported lipid bilayer containing biologically-active molecules of interest. Incorporation of transmembrane proteins into the arrayed membranes enables the study of ligand/receptor binding, as well as interactions with live intact cells. The fluidity of these molecules in the planar lipid bilayer facilitates dimerization and other higher order interactions necessary for biological signaling events. In order to demonstrate the utility of our fluid membrane array technology to ligand/receptor studies, we investigated the multivalent binding of the cholera toxin B-subunit (CTB) to the membrane ganglioside GM1. We have also displayed a number of bona fide drug targets, including bacterial endotoxin (also referred to as lipopolysaccharide (LPS)) and membrane proteins important in T cell activation.
Conclusion
We have demonstrated the applicability of our fluid cell membrane array technology to both academic research applications and industrial drug discovery. Our technology facilitates the study of ligand/receptor interactions and cell-cell signaling, providing rich qualitative and quantitative information.
==== Body
Background
DNA and protein microarrays have enabled the measurement of gene expression and biomarkers on an unprecedented scale, and have greatly affected fields ranging from oncology to medical diagnostics. However, assaying biological events occurring on cell membranes, including signal transduction and intercellular communication, has proven to be considerably more difficult to industrialize in microarray format. Basic methods for reconstituting membrane proteins into phospholipid vesicles were developed quite a number of years ago [1], facilitating fundamental discoveries in such diverse cellular processes as metabolism, solute transport, protein trafficking, and signal transduction. In subsequent years, it was demonstrated that planar phospholipid bilayers could be generated simply by fusing these vesicles to various solid substrates (e.g. glass [2] and quartz [3]), generating "supported membranes" which have lateral fluidity. The characteristic fluidity, uniformity of lipid/protein orientation, and planarity make these supported membranes ideal for studying most membrane components, including receptors, transporters, channels, and adhesive factors. Alternatively, planar lipid bilayers can be deposited on surfaces derivatized with various polymeric films [4-6]. Regardless of how the planar membrane interacts with the solid support, display of these lipid bilayers represents a powerful way of developing technologies ranging from biosensors [4,7] to platforms for the study of ion channel electrophysiology [8]. Given the increasing complexity of current basic research experiments and industrial drug discovery schemes, the ability to pattern arrays of lipid bilayers has become almost essential for performing multiplexed, high information-content assays. To address this need, there have been a number of methods developed over the past few years to pattern arrays of lipid bilayers on solid supports for purposes of studying the biology of membranes and their component lipids and proteins [9-12]. While these unique approaches all generate patterns of lipid bilayers on solid supports, they vary greatly in a number of important ways, including the manner of lipid bilayer display (e.g. supported above or directly associated with a polymer linked to the solid support) and the means of generating the membrane pattern (e.g. lithographic modification of the solid support to allow selective deposition of membranes [9,10,12,13] or modification of the planar bilayer following deposition [11]).
We have recently developed a technology for industrializing physiologically fluid supported membrane microarrays relying upon the discovery that fused silica, normally a compatible substrate for supported lipid bilayers, can readily be made membrane-incompatible through various lithographic procedures [9,13]. This development led to the realization that fluid cell membranes could be isolated in independently-addressable corrals on a silicon chip through lithographic patterning. In essence, microarrays consisting of spatially-indexed corrals of supported lipid bilayers can be generated at very high density (Figure 1) [14]. Membrane corrals are regions of a solid substrate, in this case fused silica, which are compatible with lipid bilayer deposition and are isolated from each other by membrane-barriers composed of chrome. In this manner, cell membranes are supported on a silicon chip with an aqueous interface between the membrane and solid support, as well as a bulk aqueous phase above the membrane [9,13,14]. The isolated membranes in each corral of the array are physiologically fluid as confirmed by fluorescence recovery after photobleaching (FRAP). In this technique, excitation light shadows a small region of the corral, photobleaching the fluorescent molecules in that area. Following this excitation step, the photobleached region of the corral is examined for a recovery of fluorescence which is dependent upon the ability of fluorescent lipids in the surrounding membrane to diffuse into the photobleached area. The extent of membrane fluidity determines the rate at which these fluorescent molecules diffuse into the photobleached area and diminish the shadow formed by the excitation step. Another technique used to measure the fluidity of supported lipid bilayers is the lateral migration of charged lipids in the presence of an electric field [9,13,14]. Given their aqueous surroundings, both upper and lower bilayers of the membrane exhibit a physiological level of fluidity. The fluidity of our supported lipid bilayer arrays permits interactions among membrane components (e.g. oligomerization of cell surface receptors) on a physiological time scale, a phenomenon thought to underlie numerous signal transduction pathways. This report demonstrates that the individually-corralled lipid bilayers can faithfully mimic the structural and functional characteristics of in vivo cell membranes, making them valuable for the study of biological processes ranging from simple ligand/receptor interactions to complex cell-cell signaling.
As mentioned previously, fluidity is a critical property of biological membranes in vivo. According to biophysical models of membrane fluidity, there is a logarithmic dependence of the diffusion coefficient of a given molecule (e.g. a membrane protein) on its radius, a phenomenon enabling large proteins embedded in the membrane to diffuse nearly as rapidly as lipid molecules [15]. The fluidity dynamics are completely different for membranes interacting directly with solid substrates, where friction greatly attenuates the diffusion of larger molecules (e.g. proteins) in comparison to lipids [15]. The characteristic fluidity of supported membranes depends entirely upon the manner in which the lipid bilayer is prepared and the type of solid substrate supporting the membrane. In order to retain such a high level of fluidity, these membranes cannot be exposed to an air-water interface. Integrated membrane technologies, in which lipid bilayers are deposited on a self-assembled chemical monolayer coating the solid substrate instead of directly on the substrate itself, have sought to eliminate this drawback. However, the increased convenience of coating the substrate with a chemical monolayer is offset by the observation that integrated membranes have considerably reduced fluidity (approximately 40% less) in comparison to supported lipid bilayers deposited directly on bare glass [6]. Consequently, the supported membranes described in this study display a level of fluidity more characteristic of the in vivo environment, a property likely to be critical for numerous cell biological events.
Fluid membrane protein arrays are thus a powerful technology for displaying lipid bilayers which facilitates the study of cell surface processes. Planar synthetic membranes and their molecular components are readily displayed in a fluid and functional state characteristic of the in vivo environment. While the basic principles of fluid bilayer arrays were invented in academia [9,13], we have industrialized the technology to a level suitable for high throughput drug screening. By combining lipid biochemistry with semiconductor microfabrication and robotic handling, membranes are arrayed at high density in discrete corrals on a silicon chip. In order to complement the membrane array, we have also developed methods for efficiently expressing, purifying, and displaying the vast majority of membrane proteins in the genome. Taken together, the membrane array and membrane protein display technologies enable the study of biological processes occurring on cell membranes. In this article, we report the validation of our membrane array technology for studies pertaining to ligand/receptor binding and live cell-based assays. We initially characterized our supported lipid bilayers by displaying glycolipid molecules found in mammalian (ganglioside GM1) and bacterial (LPS) cells. We have chosen to study the interaction between CTB and the membrane ganglioside GM1. Cholera toxin, naturally secreted by the bacterium Vibrio cholerae, is thought to exist as a hexamer consisting of two subunits in an AB5 configuration. The pentameric CTB subunit is thought to bind up to five GM1gangliosides in a cooperative fashion [16,17]. This multivalent interaction likely requires membrane fluidity for proper stoichiometric assembly of the ligand/receptor complex. Additionally, we demonstrate the successful display of various other membrane components, including LPS from Gram-negative bacteria and the mammalian proteins ICAM-1 and I-Ek, molecules which are important drug targets for treating diseases ranging from septic shock in the case of the former to autoimmunity in the case of the latter. Gram-negative bacteria display LPS in the outer membrane of their cell wall. Also referred to as endotoxin, LPS consists of two distinct regions: a hydrophilic polysaccharide moiety and the highly conserved hydrophobic toxin lipid A [18,19]. As LPS is present only in bacterial membranes, it has been a target of continual interest to drug discovery programs aimed at developing more efficacious antibiotics. Our studies examining T cell adhesion and activation on membrane protein arrays displaying immunological proteins demonstrates the utility of our platform for biological assays. In these T cell experiments, we have displayed proteins present on antigen presenting cells which are critical to generating an immunological response: the adhesion protein ICAM-1 and the class II dimeric antigen-presentation molecule I-Ek. The ability to study complex multivalent ligand/receptor interactions and the engagement of live cells with our supported lipid bilayers could aid not only industrial drug screening against membrane targets, but also basic research studies in academia.
Results
Arraying high quality supported membranes
We sought to characterize the quality and precision of our fluid membrane arraying capabilities for industrial production. The CTB/GM1 interaction afforded a system to study our supported lipid bilayers, as its pentavalent structure is likely to require a high degree of membrane fluidity for efficient complex formation. As depicted in Figure 2A, membranes containing unique lipid compositions (red- or green-fluorescent dopant lipids in this case) were arrayed at very high density in a manner which prevents mixing of contents from corral to corral. Moreover, the arrayed membranes were of high quality in terms of their physiological level of fluidity, as determined by FRAP measurements (Figure 2B). In order to demonstrate the precision of arraying membranes of varying compositions into specific corrals of an array, we used a pre-written program (8 × 8 arrow pattern with 8 samples) to deposit 8 different concentrations of ganglioside GM1 on a chip in a manner so as to generate an arrow-shaped pattern (Figure 3A). GM1 was detected using fluorescence microscopy following an incubation with fluorescently-labeled CTB (red), a method we have previously used to investigate this ligand/receptor pair [14].
High information-content binding of CTB to GM1 arrayed on supported membranes
Given the CTB/GM1 binding illustrated in Figure 3A, we next wished to quantify this ligand/receptor interaction. Figure 3B depicts a binding experiment performed using eight 8 × 8 (64 corral) supported membrane arrays displaying varying concentrations of GM1. We used 8 different concentrations of GM1 (0.00, 0.02, 0.05, 0.10, 0.18, 0.25, 0.35 and 0.50 mol percent) and 8 different concentrations of CTB (0, 5, 10, 20, 30, 50, 100, and 300 nM). Hence, we performed a ligand binding experiment using 8 replicates for each concentration of CTB applied so as to obtain highly quantitative and information-rich binding data with very little reagent material required. The fluid membranes were arrayed using a pre-written program consisting of an 8 × 8 standard loop with 8 samples. As in the case of the membranes in Figure 2, fluidity was confirmed using the FRAP technique. Figure 3B is a contour plot of the quantitation of the CTB/GM1 data, indicating the expected specific and saturable binding of the ligand to its receptor.
Incorporation of diverse membrane components into supported fluid lipid bilayers and live cell assays
In order to further validate our system for industrial drug discovery and basic research applications, we have purified and displayed important molecules mediating microbial pathogenesis (LPS) and various immunological disorders (ICAM-1 and I-Ek). Lipid A, the principal endotoxic moiety of LPS, was displayed in lipid bilayers and detected specifically by the anti-lipid A antibody (A6), yielding an approximately 4-fold increase in immunoreactivity compared with membranes devoid of the endotoxin (data not shown). We next sought to extend our technology for assaying membrane biology important to the immune system. The ectodomains of ICAM-1 (an adhesion molecule) and I-Ek (an antigen-presentation molecule) were over-expressed in mammalian cells, purified, and displayed in supported lipid bilayers through glycosylphosphatidylinositol (GPI) tethering (data not shown). Given that these two proteins are particularly important in mediating the adhesion and activation of T cells on antigen-presenting cells, we examined these proteins for their ability to capture a transgenic murine T cell line [20] specific for the pigeon cytochrome c (PCC) antigen. Figure 4A illustrates that surfaces displaying either ICAM-1 or I-Ek-complexed to PCC as GPI-linked proteins mediate specific adhesion of T cells when displayed either alone or in combination, while membranes devoid of these proteins confer almost no adhesion. Moreover, in order to mimic a potential therapeutic intervention, we show that monoclonal antibodies specific for either ICAM-1 or I-Ek nearly abolish T cell adhesion, while having little effect on the low level of non-specific adhesion to membranes devoid of protein (Figure 4A). In addition to measuring adhesion, the physiological level of fluidity of our supported membranes makes them ideal for assaying the full T cell activation process. Following adhesion of T cells to antigen-presenting cells, there is a dramatic relocalization of signaling molecules at the cell-cell interface known as the "immunological synapse," which is thought to be the most proximal marker of antigen-dependent T cell activation [21]. Upon initial encounter of the PCC-specific murine T cells with a membrane displaying fluorescently-labeled ICAM-1 and I-Ek complexed to PCC peptide, T cells adhere and form an immature "synapse" characterized by ICAM-1 localized centrally and I-Ek at the periphery (data not shown). During the next 30 minutes of T cell engagement, however, this "synaptic" pattern changes dynamically, as the peptide-loaded I-Ek migrates centrally and ICAM-1 adopts a more peripheral localization (data not shown). In addition to measuring "immunological synapse" formation qualitatively as has been previously demonstrated [21], we can also obtain quantitative information about this marker of T cell activation by measuring the clustering of cell surface molecules. As presented in Figure 4B, we have quantified the intensity of clustered I-Ek to measure the strength of "synapse" formation and, hence, T cell activation. As a control for quantitative measurement, we have compared the strength of "synapse" formation in the presence of I-Ek complexed to a wild-type PCC peptide to that complexed to either no peptide or a mutated PCC* peptide which cannot elicit T cell expansion. Figure 4B demonstrates not only the capability of obtaining robust quantitative data for T cell activation using our assay system, but also the high degree of antigen-specificity recorded in the experiment.
Discussion
We report here the industrialization of membrane protein microarrays suitable for basic research and drug discovery applications in numerous aspects of membrane biology, from multivalent ligand/receptor interactions to live cell assays. Through robotic handling approaches, it is possible to construct high density arrays of supported lipid bilayers containing unlimited unique combinations of membrane components. The experiments illustrated in Figure 2 and Figure 3 underscore the physiological integrity of our membranes as determined by their in vivo-like fluidity, as well as the ability to array these membranes at high density without causing mixing of contents among corrals. The high quality of the arrayed membranes is extremely reproducible between corrals and among different chips. Further, the precision of the arraying equipment facilitates deposition of unique membrane compositions in specific corrals of a membrane array. The physiological fluidity of each discretely-arrayed bilayer is truly what enables the study of membrane processes, which often require the dynamic lateral motion of proteins and lipids.
In order to validate our system for studying membrane biology, we first examined the interaction between CTB and GM1, a multivalent binding event likely to require fluidity of the bilayer for proper complex formation. As illustrated in Figure 3, a vast amount of high content data can be obtained rapidly (within minutes) for the CTB/GM1 system, or nearly any other ligand/receptor system, given the ease with which concentrations of ligand and receptor can be varied and analyzed. This advantage would be extremely beneficial for investigating the pharmacodynamics of essentially any ligand against membrane components. Indeed, the experiment depicted in Figure 3 was performed rapidly and only required 8 wells of a 96-well microtitre plate. Given the physiological level of fluidity in our supported membrane arrays, it is possible to study ligand/receptor interactions in a manner beyond capturing mere binding data. Actual signaling can be examined using our technology, given that the fluid and dynamic motion of membrane components is preserved. We chose the CTB/GM1 interaction precisely for its multivalency, a property which may require fluidity of the surrounding lipid membrane environment for proper complex formation. Indeed, pentavalent CTB is thought to recruit five GM1molecules in a process that demonstrates positive cooperativity among the components [16,17]. Hence, the characteristic fluidity of our lipid bilayers is what permits the study of CTB/GM1 and likely numerous other membrane receptors on an almost in vivo-like level. While the CTB/GM1 system is instructive for studying multivalent ligand/receptor interactions in the context of fluid membranes, we wanted to validate our technology for studying various other types of membrane components. Our ability to display bacterial LPS could facilitate academic studies in the area of host/pathogen interaction, as well as lead to the identification of novel therapeutics to treat life-threatening septic shock. While the polymyxin family of amphiphilic antibiotics target LPS, there is a continuing push to develop more efficacious anti-endotoxin molecules to treat sepsis [22]. For example, candidate antibody or small molecule therapeutics for septic shock could be identified from studying the interaction of immune system cells with membrane surfaces displaying LPS. In order to validate our technology for live cell-based assays, we have examined the activation of intact T cells on membrane arrays displaying molecules critical for adhesion and antigen-presentation. As in the case of the CTB/GM1 interaction, we chose a cell-cell association (in this case T cells and antigen-presenting cells) which would highlight the power of our fluid lipid bilayer technology to faithfully mimic in vivo membrane environments. The adhesion of T cells to the supported membranes (Figure 4A) models the first mechanistically-distinct stage of T cell activation, which happens to be an important point for therapeutic intervention in a number of immunological disorders. Indeed, Raptiva, the recently-approved antibody therapeutic for psoriasis, directly targets the ICAM-1/LFA-1-mediated T cell adhesion process [23]. In addition to cell adhesion, our ability to measure full T cell activation in both qualitative and quantitative formats underscores our ability to capture actual signal transduction events occurring on our fluid lipid bilayers, which enable the dynamic reorganization of surface receptors so crucial for this process. While T cell adhesion and partial activation has been demonstrated on arrays consisting solely of MHC deposited on a glass surface without membranes or other molecules present, it is possible that this technology fails to capture important nuances of the activation pathway which may be critical in identifying novel membrane components capable of modulating these processes [24]. In addition, drugs targeting the T cell activation process may have subtle effects which can be distinguished only through qualitative observation of the dynamic "immunological synapse." Studying T cell activation in array format could be quite useful in identifying the roles of various T cell or antigen-presenting cell membrane proteins in this process. Specifically, these proteins could be arrayed in the presence of antigen-presentation machinery in order to assess their precise roles in the distinct stages of T cell activation, including T cell adhesion, T cell stopping, and synapse formation. In essence, activation of live T cells on our supported membranes greatly facilitates high information-content immunological assays.
Conclusion
Our membrane protein microarray system extends the capabilities of DNA and protein arrays. While the latter have successfully industrialized the analysis of gene and protein expression, fluid cell membrane arrays represent the industrialization of membrane biology. Given that the vast majority of pharmaceuticals, including most of the recent protein- and antibody-based therapeutics, target membrane components, this system has the capability to advance the drug discovery process. Moreover, the technology is also useful for academic research applications, which will undoubtedly form the basis of many future therapeutic opportunities. The combination of physiological in vivo-like qualities of our membranes with robotic arraying has created a powerful system for studying membrane biology in a high throughput fashion.
Methods
Reagents
Egg phosphatidylcholine (egg-PC) and 1-Palmitoyl-2- [12-[(7-nitro-2-1,3-benzoxadiazol-4-yl)amino]dodecanoyl]-sn-Glycero-3- [Phospho-rac-(1-glycerol)] (ammonium salt) (16:0-12:0 NBD-PG) were obtained from Avanti Polar Lipids. N-Texas Red sulfonyl-1,2-dihexadecanoyl-sn-glycero-3-phosphoethanolamine (triethylammonium salt) (Texas Red DHPE) and CTB (recombinant) Alexa Fluor® 594 conjugate (labeled CTB) were obtained from Molecular Probes (presently Invitrogen). Unlabeled CTB was purchased from Sigma Chemical Co.
Fabrication of supported membrane arrays
Micropatterned substrates were fabricated on glass wafers, micropatterned with chrome barriers following a "piranha clean" (3:1 mixture of sulfuric acid and 30% solution of hydrogen peroxide). Array surfaces were cleaned by soaking for 20 minutes in freshly prepared piranha solution. Subsequently, the surfaces were rinsed under deionized water. The small unilamellar vesicles (SUVs) were prepared by standard extrusion methods. Briefly, lipids were mixed in the desired concentrations in chloroform, followed by evaporation of the chloroform in a round bottom flask using a rotary evaporator (Buchi, Rotovap R200). The lipids were resuspended in distilled water and stored overnight. Following hydration overnight at 4°C, they were passed 21 times through an Avanti extruder containing a membrane with 100-nm pores. The extruded vesicles were stored at 4°C until use within several weeks.
Planar supported bilayers were formed by fusion of SUVs onto the piranha-cleaned surfaces. Prior to supported membrane deposition, spreading solutions were prepared by mixing the SUV suspensions in equal ratios with phosphate-buffered saline (PBS). The fabrication of lipid microarrays was carried out using a microarrayer (Cartesian technologies model MycroSys™ SQ series) equipped with software (AxSys™ software) for programmable aspiration and dispensing. Due to the very small drop size (nanoliters), the microarrayer was enclosed in a humidified chamber (~98% humidity) at room temperature to avoid evaporation. In order to prevent contamination from carryover between different lipid solutions, an automatic wash cycle was incorporated into every arraying program that consisted of consecutive washes in deionized water and dry cleaning before each sample aspiration. Following the arraying step, the arrayed elements were submerged in deionized water. They were then washed 3 times with PBS, followed by incubation with CTB in PBS on a rocker (set at 2.75) for one hour at room temperature. Subsequently, they were washed 6 times with PBS on the rocker (set at 2.75) with 5-minute washes for the last three.
Reconstitution of LPS into SUVs
The pathogenic moiety of LPS (lipid A) was isolated from R595 LPS (LPS from deep rough mutant of Salmonella enterica Minnesota strain R595) by sodium acetate buffer treatment (0.1 M, pH 4.4, 100°C for 1 hour) and was dissolved in a 65:25:4 ratio mixture of chloroform:methanol:deionized water. Multilamellar vesicle formation was performed using standard methods. Briefly, the solvent was evaporated from the lipid mixture using a rotary evaporator (Buchi Rotary Evaporators, C Assembly). The remaining lipid cake was hydrated overnight at 4°C to form multilamellar vesicles. Multilamellar vesicles were sonicated in order to increase the incorporation of LPS into unilamellar vesicles, followed by extrusion to isolate the SUV population. The monoclonal IgG2b anti-lipid A antibody (A6) was prepared as previously described [25]. The goat anti-mouse IgG (H&L) antibody conjugated to Cy2 was purchased from Jackson ImmunoResearch Laboratories. Standard immunofluorescence protocols as recommended by Jackson Immuno were followed.
Generation and display of ICAM-1 and I-Ek
cDNAs encoding the extracellular domains (excluding signal sequences) of human ICAM-1 and murine I-Ek (alpha and beta chains) were amplified from ESTs (Invitrogen) and subcloned into pcDNA3.1 (Invitrogen) modified to encode a fused in-frame signal sequence and His6 sequence in tandem at the N-terminus and a GPI-specifying sequence at the C-terminus (pYBS101) as HindIII-XbaI fragments with enterkinase-encoding cleavage sites (DDDDK) at their N-termini to generate proteins having a C-terminal GPI anchor. CHO cells were transfected with expression plasmids using calcium phosphate transfection (Promega) or SuperFect (Qiagen) following manufacturers' recommendations. 48 hours after transfection, cells were split at 1:5, 1:10 and 1:20 into growth media containing 500 μg/ml of G418 (Invitrogen) for generation of stable lines expressing human ICAM-1 or murine I-Ek. Polyclonal stable cell lines were maintained and passaged in G418-containing media for 4 weeks. Expression of these proteins was confirmed by FACS analysis using PE-conjugated antibodies (Becton Dickinson). Confluent CHO cells from 6–8 10-cm dishes were washed 2X with cold PBS, and were scraped into 600 μL (per dish) of lysis buffer containing 50 mM Tris-Cl pH 8.0, 150 mM NaCl, and 1% Triton X-100, and a protease inhibitor cocktail (Sigma). Harvested cells were solubilized on ice for 30 minutes, followed by microcentrifugation for 15 min at the highest speed. Ni-NTA agarose (Qiagen) was washed 1X with lysis buffer and added to clarified extracts (50 μL of beads per 600 uL of cell extract). Cell extracts were incubated with beads for 2 hours with rotation at 4°C. Beads were washed 4X with buffer containing 50 mM Tris-Cl pH 8.0, 150 mM NaCl, 10 mM imidazole (Sigma), and 1% Octyl-Glucopyranoside (Calbiochem) plus a protease inhibitor cocktail (Calbiochem). Bound proteins were eluted with 600 μL of buffer containing 50 mM Tris-Cl pH 8.0, 150 mM NaCl, 200 mM imidazole, and 1% Octyl-Glucopyranoside plus a protease inhibitor cocktail for 1 hour at 4°C with rotation. Aliquots of protein extracts were added to an egg phosphatidylcholine (PC) lipid mix (1% NBD-PG, 99% egg PC) and dialyzed overnight at room temperature against PBS, with 3X buffer changes. While egg PC was used to generate membranes in this case, these proteins could easily be reconstituted in vesicles with various lipid compositions. To form the supported lipid membranes, 20 μL of undiluted (or diluted 1:2) protein-lipid mix was deposited onto silicon chips or cover slips for 3–5 min at room temperature, followed by 5 washes with buffer. Incorporation of protein into the supported membrane was confirmed by immunofluorescence using antibodies to the respective protein. Membranes were stained with 30 μL of PE-conjugated antibodies (BD Pharmingen) in 5% fetal calf serum (FCS) for 30 min, and then washed extensively with PBS and assessed by fluorescence microscopy. Protein densities are approximately 10,000–20,000 molecules/μm2, as estimated from a proteolipid concentration of 0.6 μg/μL (50% pure) in a 20-μL drop on a chip surface of 7 mm diameter. It was assumed that only 1% of the drop contents is retained on the chip surface for membrane display.
T cell isolation, adhesion, and immunological synapse formation
Murine T cells specific for PCC were derived from spleens harvested from the T cell receptor transgenic mouse strains B10.Cg-Tg(TcrAND)53Hed/J and B6;SJL-Tg(TcrAND)53Hed/JCell (The Jackson Laboratory). T cells were isolated from splenocytes by magnetic bead selection (Dynal) according to the manufacturer's instructions. T cell adhesion assays to supported membranes displaying either ICAM-1 or I-Ek as GPI-linked proteins were performed as previously described with several modifications [26]. Briefly, membranes were blocked with 10% fetal bovine serum (FBS) in PBS for 1 hour at room temperature, followed by incubation for 10 minutes with RPMI 1640 media (Invitrogen) containing 10% FBS at 37°C. Approximately 1 × 105 T cells were added to the well for 15 minutes. Following incubation with cells, corrals were washed for 10–15 minutes in the presence of PBS with 5% serum. Where appropriate, blocking antibodies were used at an approximately 1:100 dilution. The approximate number of cells per membrane corral was determined by counting under a brightfield microscope (usually 150–200 cells/corral). Supported lipid bilayers were subsequently inverted into a dish containing 10% FBS in PBS for 15–20 minutes to wash unbound cells. The final cell number was determined by counting under microscope directly or after imaging. In experiments where monoclonal antibodies against human ICAM-1 or murine I-Ek (Santa Cruz) were used to block T cell adhesion, supported membranes were pre-incubated with antibody followed by 3 washes with PBS prior to incubation with cells. "Immunological synapse" formation experiments were performed as described previously [21]. Imaging was performed on a Nikon Eclipse 300 vertical microscope, and analysis and quantitation were performed using Adobe Photoshop.
Imaging of membrane arrays
Imaging of the membrane microarrays was performed on a Nikon Eclipse 300 vertical microscope using a 4 × objective. Fluorescence images were taken with the same objective in fluorescence mode. The supported membranes were fluorescently labeled with 1 mol % of NBD, Texas Red-tagged lipids or labeled CTB. Images were recorded with a digital color camera driven by CoolSnap software. In cases where it was not possible to image the entire chip in one field (e.g. a 64 corral chip), the image was reconstructed in piecemeal fashion. We used Adobe Photoshop software to measure the intensity of each corral and Microsoft Excel to analyze the data.
Authors' contributions
VY performed the endotoxin experiments and oversaw the completion of all other experiments described in this manuscript. LN contributed the arrays depicted in Figure 2 and Figure 3A, and RS performed the CTB/GM1 experiment shown in the quantitation in Figure 3B. OS performed the immunological experiments depicted in Figure 4. TG provided endotoxin and LB provided the lipid A monoclonal antibodies. JG provided advice throughout the course of this work.
Acknowledgements
This work was financially supported in part by NSF SBIR Grants (DMI-0320515 and DMI-0320471), the Deutsche Forschungsgemeinschaft (SFB 470, project B5), and private funding. We gratefully acknowledge Nick Ulman and Steve Sundberg for design and fabrication of the membrane array substrate, and Sebastien Payen for the beautiful presentation of results in Figure 3.
Figures and Tables
Figure 1 Schematic diagram of cell membrane arrays for drug discovery. The 64 250-μm × 250-μm corrals are separated by 750 μm of chrome which prevents contamination of supported membranes from one corral to another. Each corral contains a supported lipid bilayer membrane which displays different molecules or different concentrations of the same molecule. The array elements are subsequently exposed to ligands or live cells, allowing for observation of differences in binding or cell biological events among the corrals. Each array resides in a single well of a 96-well plate, facilitating multiplexed assays.
Figure 2 Independently addressable cell membrane arrays. (A) An 8 × 8 64-corral supported membrane array containing different molecules (1 mol % NBD-PG or 1 mol % Texas Red DHPE in egg PC) in an alternating pattern was arrayed robotically in a manner which prevents mixing of contents from corral to corral. (B) Photobleaching a region of membrane containing 1 mol % NBD-PG in egg PC is followed by a recovery.
Figure 3 A single-pass high information-content binding curve of the CTB/GM1 interaction. (A) 8 different concentrations of ganglioside GM1 (0.00, 0.02, 0.05, 0.10, 0.18, 0.25, 0.35 and 0.50 mol %) are microarrayed in the pattern of an arrow. The chips were then exposed to a fluorescently-labeled CTB, revealing the variation in GM1 concentration. (B) This experiment was performed on 8 different supported membrane chips. Using the microarrayer, the eight 8 × 8 chips were prepared with the same distribution of ganglioside GM1. They were subsequently exposed to different concentrations of fluorescently-labeled CTB. The arrays were imaged microscopically and fluorescence was quantitated using Adobe Photoshop.
Figure 4 Live cell assays for academic research and industrial drug discovery. (A) Control membranes (devoid of protein) or membranes reconstituted with either ICAM-1 or I-Ek-complexed to PCC peptide alone or together were incubated with murine T cells specific for PCC in the presence or absence of specific antibodies to either protein. Adhered cells were counted under the microscope and averaged from different corrals. (B) Murine T cells specific for PCC were incubated with supported lipid bilayers displaying fluorescently-labeled ICAM-1 and I-Ek – complexed to wild-type PCC, a mutated PCC*, or no peptide, and were subsequently assayed for the intensity of clustered I-Ek at the center of synapses. Values represent means ± SEM from at least 3 independent determinations.
==== Refs
Silvius JR Solubilization and functional reconstitution of biomembrane components Annu Rev Biophys Biomol Struct 1992 21 323 348 1326355 10.1146/annurev.bb.21.060192.001543
McConnell HM Watts TH Weis RM Brian AA Supported planar membranes in studies of cell-cell recognition in the immune system Biochim Biophys Acta 1986 864 95 106 2941079
Pearce KH Hiskey RG Thompson NL Surface binding kinetics of prothrombin fragment 1 on planar membranes measured by total internal reflection fluorescence microscopy Biochemistry 1992 31 5983 5995 1627541 10.1021/bi00141a005
Sackmann E Tanaka M Supported membranes on soft polymer cushions: fabrication, characterization and applications Trends Biotechnol 2000 18 58 64 10652510 10.1016/S0167-7799(99)01412-2
Zhang L Booth CA Stroeve P Phosphatidylserine/Cholesterol Bilayers Supported on a Polycation/Alkylthiol Layer Pair J Colloid Interface Sci 2000 228 82 89 10882496 10.1006/jcis.2000.6910
Fang Y Frutos AG Lahiri J Membrane protein microarrays J Am Chem Soc 2002 124 2394 2395 11890761 10.1021/ja017346+
Bayley H Cremer PS Stochastic sensors inspired by biology Nature 2001 413 226 230 11557992 10.1038/35093038
Mayer M Kriebel JK Tosteson MT Whitesides GM Microfabricated teflon membranes for low-noise recordings of ion channels in planar lipid bilayers Biophys J 2003 85 2684 2695 14507731
Groves JT Ulman N Boxer SG Micropatterning fluid lipid bilayers on solid supports Science 1997 275 651 653 9005848 10.1126/science.275.5300.651
Janshoff A Kunneke S Micropatterned solid-supported membranes formed by micromolding in capillaries Eur Biophys J 2000 29 549 554 11156297 10.1007/s002490000104
Yee CK Amweg ML Parikh AN Direct photochemical patterning and refunctionalization of supported phospholipid bilayers J Am Chem Soc 2004 126 13962 13972 15506757 10.1021/ja047714k
Burridge KA Figa MA Wong JY Patterning adjacent supported lipid bilayers of desired composition to investigate receptor-ligand binding under shear flow Langmuir 2004 20 10252 10259 15518521 10.1021/la0489099
Groves JT Boxer SG Micropattern formation in supported lipid membranes Acc Chem Res 2002 35 149 157 11900518 10.1021/ar950039m
Yamazaki V Nguyen LT Schafer RJ Ulman N Groves JT MembraneChip™: Arrays of natively displayed membrane targets for novel drug discovery Nanotech 2003 3 400 403
Sackmann E Supported membranes: scientific and practical applications Science 1996 271 43 48 8539599
Schon A Freire E Thermodynamics of intersubunit interactions in cholera toxin upon binding to the oligosaccharide portion of its cell surface receptor, ganglioside GM1 Biochemistry 1989 28 5019 5024 2765522 10.1021/bi00438a017
Merritt EA Sarfaty S Van Den Akker F L'Hoir C Martial JA Hol WG Crystal structure of cholera toxin B-pentamer bound to receptor GM1 pentasaccharide Protein Sci 1994 3 166 175 8003954
Heine H Rietschel ET Ulmer AJ The biology of endotoxin Mol Biotechnol 2001 19 279 296 11721624 10.1385/MB:19:3:279
Wiese A Brandenburg K Ulmer AJ Seydel U Muller-Loennies S The dual role of lipopolysaccharide as effector and target molecule Biol Chem 1999 380 767 784 10494826 10.1515/BC.1999.097
Kaye J Hsu ML Sauron ME Jameson SC Gascoigne NR Hedrick SM Selective development of CD4+ T cells in transgenic mice expressing a class II MHC-restricted antigen receptor Nature 1989 341 746 749 2571940 10.1038/341746a0
Grakoui A Bromley SK Sumen C Davis MM Shaw AS Allen PM Dustin ML The immunological synapse: a molecular machine controlling T cell activation Science 1999 285 221 227 10398592 10.1126/science.285.5425.221
David SA Towards a rational development of anti-endotoxin agents: novel approaches to sequestration of bacterial endotoxins with small molecules J Mol Recognit 2001 14 370 387 11757070 10.1002/jmr.549
Kupper TS Immunologic targets in psoriasis N Engl J Med 2003 349 1987 1990 14627782 10.1056/NEJMp038164
Soen Y Chen DS Kraft DL Davis MM Brown PO Detection and characterization of cellular immune responses using peptide-MHC microarrays PloS Biol 2003 1 429 438 10.1371/journal.pbio.0000065
Kuhn HM Brade L Appelmelk BJ Kusumoto S Rietschel ET Brade H Characterization of the epitope specificity of murine monoclonal antibodies directed against lipid A Infect Immun 1992 60 2201 2210 1375194
Chan PY Lawrence MB Dustin ML Ferguson LM Golan DE Springer TA Influence of receptor lateral mobility on adhesion strengthening between membranes containing LFA-3 and CD2 J Cell Biol 1991 115 245 255 1717480 10.1083/jcb.115.1.245
| 15960850 | PMC1177928 | CC BY | 2021-01-04 16:02:57 | no | BMC Biotechnol. 2005 Jun 16; 5:18 | utf-8 | BMC Biotechnol | 2,005 | 10.1186/1472-6750-5-18 | oa_comm |
==== Front
BMC CancerBMC Cancer1471-2407BioMed Central London 1471-2407-5-571593263210.1186/1471-2407-5-57Research ArticlePrevalence of von Hippel-Lindau gene mutations in sporadic renal cell carcinoma: results from the Netherlands cohort study van Houwelingen Kjeld P [email protected] Dijk Boukje AC [email protected] de Kaa Christina A [email protected] Leo J [email protected] Hanneke JM [email protected] Jack A [email protected] den Brandt Piet A [email protected] Egbert [email protected] Department of Urology, University Medical Center Nijmegen, P.O. Box 9101, 6500 HB Nijmegen, the Netherlands2 Department of Epidemiology, NUTRIM, Maastricht University, P.O. Box 616, 6200 MD Maastricht, the Netherlands3 Department of Pathology, University Medical Center Nijmegen, P.O. Box 9101, 6500 HB Nijmegen, the Netherlands2005 2 6 2005 5 57 57 7 1 2005 2 6 2005 Copyright © 2005 van Houwelingen 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
Biallelic von Hippel-Lindau (VHL) gene defects, a rate-limiting event in the carcinogenesis, occur in approximately 75% of sporadic clear-cell Renal Cell Carcinoma (RCC). We studied the VHL mutation status in a large population-based case group.
Methods
Cases were identified within the Netherlands cohort study on diet and cancer, which includes 120,852 men and women. After 11.3 years of follow-up, 337 incident cases with histologically confirmed epithelial cancers were identified. DNA was isolated from paraffin material collected from 51 pathology laboratories and revised by one pathologist, leaving material from 235 cases. VHL mutational status was assessed by SSCP followed by direct sequencing, after testing SSCP as a screening tool in a subsample.
Results
The number of mutations was significantly higher for clear-cell RCC compared to other histological types. We observed 131 mutations in 114 out of 187 patients (61%) with clear-cell RCC. The majority of mutations were truncating mutations (47%). The mean tumor size was 72.7 mm for mutated tumors compared to 65.3 mm for wildtype tumors (p = 0.06). No statistically significant differences were observed for nuclear grade, TNM distribution or stage. In other histological types, we observed 8 mutations in 7 out of 48 patients (15%), 1 mutation in 1 of 6 oncocytoma, 3 mutations in 2 of 7 chromophobe RCC, 2 mutations in 2 of 30 papillary RCC, no mutations in 1 collecting duct carcinoma and 2 mutations in 2 of 4 unclassified RCC.
Conclusion
VHL mutations were detected in 61% of sporadic clear-cell RCC. VHL mutated and wildtype clear-cell RCC did not differ with respect to most parameters.
==== Body
Background
Historically, the classification of Renal Cell Cancer (RCC) was based on morphological features. The majority of RCC are of the clear cell type (~80%); other subtypes are papillary RCC (10%), chromophobe RRC (5%), collecting-duct carcinoma (1%) and unclassified RCC (3–5%). Based on the work of numerous investigators, it became evident that RCC could be divided into genetically distinct classes: this resulted in the so-called Heidelberg classification, which partly overlaps the former pathological classification of RCC based on morphological criteria. The most prominent common genetic aberration for clear-cell (conventional) RCC is loss of 3p. Characteristic for papillary RCC is trisomy of chromosomes 3q, 7,8,12,16,17,20, and loss of the Y chromosome, and chromophobe RCC is characterized by a combination of loss of heterozygosity at chromosomes 1,2,6,10,13,17, and 21 [1].
Von Hippel-Lindau disease (VHL) is a rare inherited disorder associated with, amongst others, an enhanced risk for clear-cell RCC [2]. The VHL gene responsible for this syndrome was identified through linkage analyses and molecular cloning and is located on chromosome 3p25. After its identification it became evident that the VHL gene is also involved in the development of sporadic clear-cell RCC. Together with loss of the homologous chromosome 3p allele (3p LOH), VHL mutations are rate-limiting events in the carcinogenesis of clear-cell RCC [3,4]. Mutations have been observed in the entire gene and usually lead to a truncated inactive protein [5]. The VHL gene is considered a tumor suppressor gene, involved in cell cycle regulation, regulation of hypoxia inducible genes and proper fibronectin assembly in extracellular matrix [6,7]. It was estimated that approximately 75% of all sporadic clear-cell RCC harbor biallelic VHL defects [8]. In approximately 19% of sporadic clear-cell RCC, methylation of the VHL gene promoter appeared to be involved [9]. In approximately 10%–20% of sporadic clear-cell RCC no alteration in the VHL alleles was detected, indicating that other genes are involved in clear-cell RCC carcinogenesis, possibly affecting the same signaling pathway as VHL.
Several risk factors for developing RCC have been identified: tobacco smoking, obesity, drugs, such as phenacetin, hypertension and/or its medication, and occupational exposure to trichloroethylene, gasoline, petroleum products, asbestos, and iron processing fumes. The influence of dietary factors, such as vegetable, fruit vitamin C, carotenoid, meat and milk product consumption, is controversial [10]. Multiple and specific types of VHL mutations in RCC have been associated with exposure to the industrial solvent trichloroethylene [11,12]. Consumption of vegetables and citrus fruit decreased the frequency of VHL mutations among smokers and consumption of citrus fruit decreased VHL mutation frequency for all patients [13]. These findings and investigations in animals [14] suggest that mutational patterns in the VHL gene may serve as an etiological imprint to factors causing renal cancer. Thus, it may be possible to improve our etiological insight in particular risk factors when a more specific endpoint than "RCC" can be defined, e.g. based on histology and mutational status of a gene involved in tumor carcinogenesis.
We decided to determine the mutational status of the VHL gene of RCC cases identified within a population-based cohort of 120,852 men and women aged 55–69 which was recruited in the Netherlands to study associations between dietary habits, lifestyle and cancer occurrence. To validate whether SSCP could serve as a prescreening method, SSCP and direct sequencing was evaluated in a subset of 20 patients. In this article we report on histopathological and clinical parameters and the type and number of mutations in VHL observed in a large population-based sample of cases.
Methods
Study population
The RCC cases selected for this study were incident cases identified in the population of the Netherlands Cohort Study (NLCS) on diet and cancer. The NLCS has been described in detail elsewhere [15]. Briefly, this prospective study was initiated in 1986 and included 58,279 men and 62,573 women, 55 to 69 years of age, who completed a self-administered questionnaire on diet, family history of cancer, and other risk factors for cancer at baseline. The entire cohort is being monitored for cancer occurrence by annual record linkage to the Netherlands Cancer Registry and to PALGA, a nationwide network and registry of histo- and cytopathology [16]. Information on sex and family history of RCC (at baseline) was retrieved from the NLCS database, while information on age at diagnosis and tumor localization was retrieved from the Netherlands Cancer Registry. Tumor stage was classified according to the 1987 revision of the UICC-TNM classification [17], using information from the cancer registry and the pathology report. Tumor size was retrieved from the pathology report.
Tissue sample collection
Paraffin material was collected after approval by the Medical Ethical Committees of Maastricht University, PALGA and the Netherlands cancer registry. From 1986 to 1997 (11.3 years follow up) 355 kidney cancer cases (ICD-O-3: C64.9) were identified. Urothelial cell carcinomas were excluded and only histologically confirmed epithelial cancers were included (ICD-O: M8010-8119, 8140–8570), leaving 337 cases. The PALGA database was also used to identify the location of tumor tissue storage in the Dutch pathology laboratories. For 273 cases, a PALGA record with information on the location of paraffin material was available at the start of the collection of paraffin blocks. The tissue samples were distributed over 51 pathology laboratories throughout the Netherlands.
For 251/273 cases (92%) paraffin blocks were collected. Failure to retrieve material was the result of the refusal of the pathology laboratory to cooperate (3 laboratories with material for 10 cases), the unavailability of suitable material (i.e. only material from a biopsy, cytology or a metastasis was present) (8 cases), not being able to locate the paraffin block at the laboratory (3 cases), and for 1 case the reason was not recorded. Paraffin blocks of parallel normal tissue were collected if available. Collected archival tissue sample blocks were registered and coded using a unique identification number.
Revision
Genomic DNA was extracted from five 20-μm slices from each tumor and normal specimen. A flanking section was haematoxylin and eosin (HE) stained for histological purposes, e.g. grading and estimation of percentage of tumor cells. One experienced pathologist (CAHK) reviewed all HE stained slides. The RCC were classified according to the World Health Organization (WHO) classification of Tumours of 2002 [18]. Nuclear grading was performed according to Fuhrman [19]. Grading was based on the most atypical focus in the paraffin block used for DNA extraction, with a dimension of at least one high power field.
Material of 16 cases was discarded after revision. The collected material was unsuitable for analysis, because it concerned a biopsy (N = 2) or a metastasis (N = 2) or no tumor tissue was present (N = 4) or material contained less than 10% malignant cells (N = 7) (tumor samples had to contain at least 10% malignant cells to decrease the possibility of ignoring mutations). Material from one case was reclassified as urothelial cell carcinoma and subsequently excluded. As a result, tumor DNA from 235 cases was available for further analysis.
DNA isolation
Genomic DNA was prepared as follows: paraffin was removed with xylene and genomic DNA was extracted by salt-precipitation. Briefly, 450 μl of cell lysis solution (10 mM Tris/HCl (pH 7.4), 400 mM NaCl, 2 mM EDTA), 25 μl of 10% SDS and 50 μl of proteinase K solution (20 mg/ml) were added to the tissue samples and incubated over-night at 55°C. Proteins were precipitated using 175 μl of saturated NaCl, followed by centrifugation (2 minutes, 13.200 rpm). DNA was precipitated by the addition of 0.6 volumes of iso-propanol, dissolved in TE (pH 7.4) and stored at -20°C. The DNA concentration and purity was measured at 260 and 280 nm.
VHL mutation analysis
PCR primers used for amplification are described in Table 1. Amplification of DNA for SSCP analysis and sequence analysis was performed as follows: 100 ng of the extracted DNA was subjected to 35 cycles of PCR: 40 sec at 92°C, 40 sec at Tm and 40 sec at 72°C. Exon amplification was performed in 30 μl of buffer: 50 mM KCl, 10 mM Tris-HCl (pH 9.0), 1.5 mM MgCl2, 1% triton X-100, 0.1% (w/v) gelatin, 250 uM dNTP's, 25 pmol of each primer and 1U Taq polymerase (HT Biotechnology Ltd.). For SSCP analysis, 0.2 μl of [α 32P] dATP (10 mCi/ml, ~3000 Ci/mmol, FIRMA) was added.
For SSCP analyses, 5 μl of the radiolabeled PCR product was diluted in 5 μl loading buffer (96% formamide, 20 mM EDTA, 0.05% bromophenol blue and xylene cyanol), boiled for 3 min, and quenched on ice before loading. Three μl of each sample was loaded on a 0.5x MDE gel, containing 0.6x TBE either with or without 10% (v/v) glycerol. Electrophoresis was performed at room temperature at 5W for MDE gels without glycerol, and at 7W for MDE gels with 10% (v/v) glycerol, using 0.6x TBE as electrophoresis buffer. After 20 hours of electrophoresis, the gels were transferred to Whatmann 3 MM paper (FIRMA) and dried on a gel dryer. The separated fragments were visualized by Hyperfilm-MP (FIRMA) exposure.
For sequence analysis, the PCR products were purified using the Wizard™ PCR preps purification system (Promega Corp.). Sequencing was performed at the local central sequence facility using BigDye Terminator and the ABI basecaller (Applied Biosystems). Mutations were identified by visual inspection of sequences provided by the ABI basecaller, and called when unequivocally present on the sense and/or antisense strand, and when the area under the curve of peaks changed by more than 5%, compared to the area under the curve of the normal signal.
Feasibility and comparability pilot
Twenty samples were analyzed by PCR-SSCP and direct sequencing to investigate the suitability of PCR-SSCP as a screening tool preceding direct sequencing. After evaluation, we decided to use PCR-SSCP as a screening instrument before direct sequencing. All remaining samples were subjected to PCR-SSCP, which was only followed by direct sequencing in case of an aberrant band pattern on the PCR-SSCP. Samples were also sequenced when equivocal PCR-SSCP results were obtained.
Statistical analyses
The overall frequency of VHL mutations as well as the type of mutation and affected exon and codon were computed for all 235 cases. Differences between histological subtypes were assessed by the χ2-test. For clear-cell RCC, mutated and wildtype tumors were compared with respect to age at diagnosis, sex, grade, TNM classification [17], stage and family history of RCC. Differences in mean values of age at diagnosis as a continuous variable were evaluated by the student's t-test. Differences in the categorical variables sex, family history of RCC, grade, TNM classification and stage were evaluated for significance by the χ2-test. A p-value of 0.05 or less was considered statistically significant. Statistical analyses were performed with the STATA statistical software package (STATA statistical software, Release 7, STATA corporation, College Station, TX, USA, 2001).
Results
PCR-SSCP as a screening tool preceding direct sequencing
To evaluate whether PCR-SSCP could be used to distinguish between wildtype and mutated VHL, we determined the VHL status of 20 cases by SSCP and direct sequencing (Table 2). SSCP and direct sequencing were mostly in agreement. In all cases where disparate results were obtained, this concerned aberrant signals in PCR-SSCP, followed by a negative result by direct sequencing (Table 2). Most importantly, we never found a mutation in the sequence analysis after a negative PCR-SSCP (0% false-negatives) (Table 2). However, we are aware that the sampling size is limited, and it is known that SSCP as a screening tool is neither 100% sensitive nor specific. Therefore, we estimated the chance of a positive result on direct sequencing after a negative result on the SSCP (i.e., the chance that a mutation will be missed by using SSCP as a screening instrument). The point estimate for the percentage of false-negatives based on SSCP is 0%, with an upper confidence limit between 3.7% and 16.9%, based either on the individual analyses (N = 97), or on the number of cases (N = 20) (see Additional file 1: Calculation of the estimated upper 95% confidence limit). Different types of mutations were identified by SSCP followed by direct sequencing in this pilot.
VHL mutations and clinical parameters
We were able to analyze at least one tumor block for 235 patients with 236 tumors (one patient had two primary tumors). After revision by CAHK, the analyzed samples could be subdivided as follows: clear-cell RCC (188/236, 79.7%), papillary RCC (30/236, 12.7%); chromophobe RCC (7/236, 3.0%); oncocytoma (6/236, 2.5%), collecting duct carcinoma (1/236, 0.4%), and unclassified RCC (4/236, 1.7%). For 198 cases one paraffin block was analyzed, while for 34 cases 2 paraffin blocks were analyzed and for 3 cases 3 paraffin blocks were analyzed. In 10 cases, a different genotype was observed in different tumor blocks obtained from 1 tumor. In 6 cases, 1 sample contained wildtype VHL while the other sample harbored one or more mutations. For the statistical analyses these cases were analyzed as "mutated". In 13 cases, 2 mutations in the same paraffin block were observed. In one patient with two primary tumors, both tumors showed a different mutation. In total we observed 253 outcomes (139 mutations (in 121 patients) and 114 wildtype) for 235 patients with 236 primary tumors. All 139 observed mutations are described. (see Additional file 2: Description of observed mutations (N = 139), histological parameters and personal characteristics for 121 cases).
VHL gene mutations were mostly observed in clear-cell RCC, but also in unclassified RCC (2/4); papillary RCC (2/30); chromophobe RCC (3 mutations in 2/7 patients) and oncocytoma (1/6) (see Additional file 2: Description of observed mutations (N = 139), histological parameters and personal characteristics for 121 cases). The percentage of patients with a mutation was significantly higher for clear-cell tumors compared to tumors of other histological types (χ2: 32.9; p-value<0.001). In the remainder of this paper, we will only consider clear-cell RCC, leaving 204 outcomes (131 mutations (in 114 patients) and 73 wildtype) for 187 patients with 188 primary tumors. Nine outcomes were mutations in intronic sequences, possibly affecting splicing. The majority of observed mutations in coding regions lead to a truncated VHL product (insertion/deletion mutations leading to a frameshift and nonsense mutations, 62/122) or to a deleted/inserted or altered amino acid, 47/122). The percentage of point mutations was much higher in exon 1 (50.7%) than in exon 2 and 3 (29.6 and 23.1, respectively).
Figure 1 shows the type of mutation plotted against the codon number. We observed 62 truncating mutations, which consisted of 4 nonsense mutations and 58 deletions or insertions leading to a shift of the reading frame (Figure 1A), 15 insertions or deletions that did not affect the reading frame (Figure 1B), 32 missense mutations (Figure 1C) and 13 silent mutations (Figure 1D).
Cases with a mutation were older (mean: 67.8 years; sd: 4.7) than cases without a mutation (mean: 67.1; sd: 4.6). The mean age difference at diagnosis was 0.8 years with a 95% CI ranging from -0.6 through 2.2 years. The percentage of men with clear-cell RCC with a VHL mutation was somewhat higher than the percentage of males among patients with a clear-cell RCC without a VHL mutation, 63.2 % vs. 53.4% (χ2: 1.75; p-value: 0.19).
The mean tumor size was 72.7 mm for tumors with a VHL mutation compared to 65.3 mm for tumors without a VHL mutation. The difference was 7.3 mm with a 95% CI from -2.1 mm through 16.8 mm. Other tumor parameters for patients with a primary clear-cell RCC with or without a VHL mutation are shown in Table 3. We did not observe a difference in nuclear grade, T, N, M, or stage between mutated and wildtype tumors.
For 3 patients, a positive family history was reported; 2 of these patients (Sample id's: 1003 and 306/1600) had a clear-cell tumor harboring a mutation. One patient with a papillary tumor reported a positive family history; no mutations in the VHL gene were observed in the corresponding tumor sample. Obviously, these numbers are too small to engage in statistical testing.
Discussion
The von Hippel-Lindau (VHL) gene is a tumor suppressor gene predisposing to both sporadic clear-cell (conventional) RCC and von Hippel-Lindau disease. We analyzed 235 cases of primary sporadic RCC, of which 187 cases presented with the histological subtype clear-cell RCC, identified within the Netherlands cohort study on diet and cancer. VHL mutations were assessed by PCR-SSCP, followed by direct sequencing when aberrant signals were detected in SSCP. Screening by SSCP was considered appropriate, since we never detected a mutation by direct sequencing after a negative result on the SSCP in a pilot study. Furthermore, we estimated the upper 95% confidence limit for cases missed at 3.7% or 16.9%, based on the individual analyses (N = 97) or on the number of cases (N = 20). Therefore, mutated cases may have been overlooked, which placed them in the wildtype group. This would have resulted in a reduction of contrast between wildtype and mutated groups.
We observed 131 mutations in 114 out of 187 patients (61%) in patients with clear-cell tumors. Estimates reported in the literature [2,20-30] range from 36% [25] to 54% [27], and therefore our estimate appears high. However, our estimate is close to the estimate of 75% as discussed in the review by Cohen [8]. This difference in the observed proportion of mutations in the VHL gene may be attributed to the use of different techniques, e.g., SSCP vs. direct sequencing, to contamination with normal tissue components, or to the type of material analyzed (fresh or archival paraffin material). In our study, we used SSCP followed by direct sequencing, as most investigations did [2,20-29]. To avoid ignoring VHL mutations through contamination with normal tissue, samples needed to contain at least 10% tumor cells. The vast majority of samples (180 out of 190) contained at least 50% viable tumor tissue. The sample with the lowest percentage of viable tumor tissue, in which we observed a mutation, contained 30% of viable tumor tissue. For 2 samples the percentage of viable tumor tissue was lower (10% and 20%), thus we cannot formally exclude the possibility that we overlooked an existing mutation. We identified cases in a population-based study after 11.3 years of follow-up; the time span and the number of patients and hospitals involved excluded the possibility to collect fresh material. Three other studies on sporadic RCC used paraffin embedded material and the percentages of mutations detected were 46% [29], 50% [22] and 54% [27].
The occurrence of various types of mutations is comparable to other studies of this size [21,24,26,27], but some differences are also apparent. The presence of silent mutations is not commonly reported; only Ma et al. [27] reported 13% of silent mutations, similar to 11% observed by us. It is possible that others do not report silent mutations because the investigators do not wish to designate these as true mutations since their relevance is unclear. The percentage of frameshift mutations was approximately 50% in most studies [21,26], comparable to the 48% observed in the current study. The percentage of in-frame deletions/insertions was higher in our study (12%) as compared to the percentage in other studies (<5%) [21,26]. Compared to the distribution of mutation data present in the Universal VHL Mutation database with 747 mutations [31], the relative amount of point mutations is much higher in the database (64%), compared to the percentage of point mutations we observed (44%), as well as the frequency of nonsense mutations (11% vs. 3%). However, many of the mutations in the Mutation database concern germline mutations identified in VHL-families. It has been suggested that the mutational spectrum between germline and sporadic VHL mutations may differ [25], which could explain the disparity. We were not able to select records for sporadic tumors in the database, but Gallou et al. [26] described mutations for 145 sporadic cases from this database. Compared to our results, the percentage of nonsense mutations was higher (8% vs. 3%) and the percentage of in-frame deletions or insertions was lower (4% vs. 12%) [26].
Furthermore, we searched the Universal VHL Mutation database [31] for the mutations observed in the cases with a positive family history of RCC. The T>A (silent) mutation at codon 139 (sample id 1003) was not reported before, the deletion of 1 nucleotide at codon 146 (sample id 306) was recorded in the database twice, while the second 1 nucleotide deletion at codon 167 (sample id 1600) had not been reported before.
For 10 RCC cases, we observed different VHL genotypes in different tumor samples obtained from one tumor. Samples were re-analyzed to exclude technical errors. Again, different VHL genotypes, identical to the earlier genotypes, were observed, showing that tumor heterogeneity appears to play a role. This was an unexpected finding in view of the general acceptance that a mutation in the VHL gene is an early event in clear-cell RCC carcinogenesis, because one would then expect monogenetic VHL expression. Nevertheless, it is possible that genetic instability leads to additional mutational events, eventually leading to different VHL genotypes in clear-cell RCC. This was underscored by 13 cases showing 2 mutations within the same paraffin block and other studies have also reported multiple mutations per primary tumor [21,27,29].
The distribution of histological subtypes was comparable to the distribution described before [[32], p2692]. Surprisingly, we also found VHL mutations in some of the non-clear-cell RCC samples that were collected. These results are in contrast with the current opinion that VHL mutations are exclusively restricted to clear-cell RCC. VHL mutations, however, have also been reported in chromophobe RCC, although these were clustered in the 5'UTR/promoter region [22], and Brauch et al. [12] described a renal oncocytoma with a VHL mutation. To our knowledge, others have not yet described VHL mutations in papillary RCC. It is conceivable that genetic instability following tumor-initiating events leads to VHL mutations acquired at a later stage of tumor development. As expected, the relative occurrence of VHL mutations in the non-clear-cell RCC renal tumors was much lower than in clear-cell RCC, emphasizing the crucial role of VHL mutations in clear-cell RCC carcinogenesis.
The high percentage of mutations observed, the observation of mutational heterogeneity (different genotypes in 1 or different samples from the same tumor), and the observation of mutations in other histological types, raises the question whether paraffin material is suitable for analysis. It has been shown, however, that VHL gene mutations can be detected in archival paraffin material [33]. Identical mutations were observed in exon 2 in 3 cases for which archival paraffin material and tumor derived cell lines were assessed by SSCP and sequencing [33]. On the other hand, it has also been reported that artifacts may occur when archival paraffin material is used [34]. Williams et al. reported up to 1 mutation artifact per 500 bases recorded. The chance of such artificial mutations in formalin-fixed material was inversely correlated with the number of cells used in the PCR: the fewer the cells, the more artifacts. No artifacts were observed when 300 cells were used and only 1 artifact was observed in 150 cells (0.03%). The highest frequency of artifacts was observed when using 10 or 20 cells, and equaled 0.2 % [34]. The number of cells in our analyses, however, was large (>500 cells), since 5 20-μm slices of paraffin embedded material were used and biopsies were excluded.
The mean tumor size was larger for mutated tumors than for wildtype tumors, however not statistically significant. Two other studies did not show a difference in tumor size [24,28]. VHL gene mutations leading to reduced or inactive VHL protein (pVHL) could theoretically lead to a larger tumor size. In the absence of pVHL, hypoxia inducible factor becomes stabilized and upregulates a myriad of hypoxia-inducible genes, resulting in hypervascular tumors [35]. This should be a growth advantage. Our results and results from others [24,28] showed no association with nuclear grade (according to Fuhrman) [19]. It was difficult to compare TNM and stage, since different classifications were used in different studies and the TNM classification has changed definitions in recent years. We reviewed all cases and classified them according to the 1987 version of the TNM classification [17]. We did not observe a difference in T, N, M, or stage between mutated and wildtype tumors, as was confirmed by 2 other studies [24,28]. One study, however, reported a higher percentage of mutations in pT3 tumors with a VHL mutation or hypermethylation, compared to tumors without VHL alteration (64% vs. 40%) [21].
Phenotype should also be considered. Mutations may impair pVHL function. However, in order for pVHL to lose its function, biallelic VHL defects have to occur. We did not measure LOH, since it is well documented that LOH commonly occurs in RCC cases, with estimates ranging from 74% [30] to 93% [21]. Kondo et al. showed 97% LOH in mutated or hypermethylated samples compared to 81% LOH in wildtype samples [24]. Furthermore, silent mutations do not lead to an altered protein, but these types of mutations were not common (13/131) in our study. Also, pVHL formation may be inhibited by epigenetic silencing, which has been reported to occur in 19% of clear-cell RCC [9]. This could have resulted in misclassification in comparing pathological parameters, because samples with epigenetic silencing only were classified as wildtype tumors, as we were not able to investigate epigenetic silencing in the current study. Finally, cells in cell culture have been shown to produce two types of proteins: pVHL30 and pVHL19 (a results of internal translation from the second methionine within the VHL open reading frame (Met 54) [36-38]. The pVHL19 product has been shown to inhibit the production of hypoxia-inducible genes, but does not bind to fibronectin [36,37]. Reintroduction of pVHL19 has been shown to inhibit tumor formation in nude mice [37]. We detected mutations 5' of codon 54 in 11 cases; these are unusual and would be predicted to affect pVHL30, but not pVHL19 translation products. However in 5 of these 11 cases, a second mutation was observed 3'of codon 54. So, in 5 of these cases the other mutation may be relevant. For the other 6 cases, these mutations may not lead to a functionally inactive protein, and thus we may have misclassified these. However, as remarked before, the pVHL19 variant does not bind to fibronectin and thus may not contribute to the maintenance of the extracellular matrix.
Conclusion
VHL mutations were observed in 61% of sporadic clear-cell RCC. No differences were observed in nuclear grade, TNM distribution or stage between VHL mutated and wildtype clear-cell RCC. However, the tumor size was larger for VHL mutated clear-cell RCC. This study forms a good basis to study possible associations between potential risk factors and VHL mutations in clear-cell RCC. Ultimately, this should lead to greater insight in the etiology of clear-cell RCC.
List of abbreviations
CI Confidence Interval
NLCS Netherlands Cohort Study on diet and cancer
RCC Renal Cell Carcinoma
RR Rate Ratio
VHL von Hippel-Lindau
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
Paraffin embedded material was collected by BACD. KPH carried out laboratory analyses, aided by HJMG. CAHK revised all HE-stained slides from collected material. BACD carried out statistical analyses. KPH, BACD and EO co-drafted the document. CAHK, LJS, JAS and PAB revised the document critically for important intellectual content. PAB also initiated the Netherlands Cohort Study on diet and cancer, of which this study is a part. All authors read and approved the final manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Supplementary Material
Additional File 1
Calculation of the estimated upper 95% confidence limit.
Click here for file
Additional File 2
Description of observed mutations (N = 139), histological parameters and personal characteristics for 121 cases.
Click here for file
Acknowledgements
This study was financially supported by the Dutch Kidney Foundation (Grant C99.1863) and the Dutch Cancer Society. We wish to thank Dr. R.A. Goldbohm, Dr. E. Dorant, C.A. de Brouwer, Prof. Dr. Kiemeney, Prof. Dr. Geurts van Kessel and Prof. Dr. Ruiter for their preparatory work for this study; Dr. A. Kester for statistical advice; S. van de Crommert, H. Brants, J. Nelissen, C. de Zwart, M. Moll, W. van Dijk, M. Jansen and A. Pisters for assistance; and H. van Montfort, T. van Moergastel, L. van den Bosch and R. Schmeitz for programming assistance. The authors also thank the staffs of the Dutch regional cancer registries and the Netherlands national database for pathology (PALGA) for providing incidence data. Finally, we would like to thank the pathology laboratories from the following hospitals for providing paraffin material: UMC Nijmegen, Academisch Ziekenhuis Groningen, Leids Universitair Medisch Centrum (LUMC), Rijnland ziekenhuis, Antoni van Leeuwenhoek ziekenhuis, Laboratorium Pathologie Oost Nederland, Academisch Ziekenhuis Utrecht, Ziekenhuis Rijnstate, Laboratorium voor de Volksgezondheid in Friesland, Stichting samenwerkende ziekenhuizen Oost-Groningen, Martini ziekenhuis, Stichting Samenwerking Delftse Ziekenhuizen, Ziekenhuis Leyenburg, Vrije Universiteit Medisch Centrum, Academisch Medisch Centrum, Academisch Ziekenhuis Maastricht, Groene Hart Ziekenhuis, Canisius Wilhelmina Ziekenhuis, Slotervaartziekenhuis, Maaslandziekenhuis, Atrium Medisch Centrum Heerlen, Atrium Medisch Centrum Kerkrade, Deventer ziekenhuis, IJsselmeerziekenhuizen Lelystad, Isala klinieken, Elkerliekziekenhuis, Jeroen Bosch Ziekenhuis, Pathologisch laboratorium voor het Gooi en Almere, Regionaal Pathologisch & Cytologisch Laboratorium voor Eemland en Noord/West Veluwe, Diakonessenhuis Utrecht, St. Antonius Ziekenhuis, Onze Lieve Vrouwe Gasthuis, Boven-IJ Ziekenhuis, Stichting Pathologisch Anatomisch Laboratorium Kennemerland, Ziekenhuis de Heel, Diaconessenhuis Leiden, Rode Kruis Ziekenhuis, Bronovo ziekenhuis, Laurentius ziekenhuis Roermond, Pathologisch Laboratorium voor Dordrecht en omstreken, Zuiderziekenhuis, St. Claraziekenhuis, Medisch Centrum Haaglanden, Havenziekenhuis, Stichting Streek Laboratorium Zeeland, Stichting Pathologisch en cytologisch laboratorium West-Brabant, Stichting Ignatius ziekenhuis, St. Elisabeth ziekenhuis, Catharinaziekenhuis, St. Maartensgasthuis and Spaarne ziekenhuis.
Figures and Tables
Figure 1 Type of mutation plotted against the codon number. A. Truncating mutations (Frameshift & Nonsense mutations) B. In-frame insertions/deletions C. Missense mutations D. Silent mutations Codon 1–114 encodes exon 1, codon 114–155 encode exon 2, and codon 155–213 encode exon 3.
Table 1 Primers used to amplify the VHL gene
Name Sequence 5'>3' Exon Tm Fragment length
Sense3a GGT CTG GAT CGC GGA GGG A 1 64°C * 191 bp
Asense4a GCC CGG CCT CCA TCT CCT 1
Sense5a AGT CGG GCG CCG AGG AGT 1 64°C * 184 bp
Asense6a CCG TCG AAG TTG AGC CAT AC 1
Sense7a CCC AGG TCA TCT TCT GCA AT 1 64°C † 159 bp
Asense8a CTG CTG GGT CGG GCC TAA G 1
Sense9 GTG GCT CTT TAA CAA CCT TTG C 2 60°C‡ 194 bp
Asense10 CCT GTA CTT ACC ACA ACA ACC TTA TC 2
Sense11a CAC TGA GGA TTT GGT TTT TGC 3 55°C 162 bp
Asense12a TCC AGG TCT TTC TGC ACA TTT 3
Sense13a GAC ATC GTC AGG TCG CTC TA 3 55°C 150 bp
Asense14a TCA AAA GCT GAG ATG AAA CAG TG 3
* 3% DMSO added for PCR
† 1.5% DMSO and 2.5 mM MgCl2 added for PCR
‡ 4% DMSO and 1.0 mM MgCl2 added for PCR
Table 2 Mutation analysis in 20 samples by SSCP and direct sequencing
Direct sequencing
Positive Negative
Primerset 3/4
SSCP Positive 0 5
Negative 0 15
Primerset 5/6
SSCP Positive 2 0
Negative 0 18
Primerset 7/8
SSCP Positive 4 2
Negative 0 14
Primerset 9/10
SSCP Positive 3 0
Negative 0 17
Primerset 11/12
SSCP Positive 4 1
Negative 0 15
Primerset 13/14
SSCP Positive 2 0
Negative 0 18
Positive on SSCP means: indication of a mutation or not assessable after two attempts.
Positive on direct sequencing means: proof of a mutation.
Table 3 Tumor parameters for patients with a clear-cell tumor, also stratified by VHL mutation status
All tumors (N = 188) VHL mutated tumors (N = 115) Wildtype tumors (N = 73)
N % N % N % χ2 p-value †
T (TNM*)
1 5 2.7 4 3.5 1 1.4
2 93 49.5 55 47.8 38 52.1
3A 35 18.6 20 17.4 15 20.6
3B 50 26.6 34 29.6 16 21.9
4 2 1.1 1 0.9 1 1.4 0.18§** 0.67
X 3 1.6 1 0.9 2 2.7
N (TNM*)
0 126 67.0 78 67.8 48 65.8
1 7 3.7 6 5.2 1 1.4
2 7 3.7 5 4.4 2 2.7 1.51¶** 0.22
X 48 25.5 26 22.6 22 30.1
M (TNM*)
0 131 69.7 77 67.0 54 74.0
1 26 13.8 18 15.7 8 11.0 0.99** 0.32
X 31 16.5 20 17.4 11 15.1
Stage (TNM*)
1 5 2.7 4 3.5 1 1.4
2 83 44.2 46 40.0 37 50.7
3 66 35.1 42 36.5 24 32.9
4 32 17.0 22 19.1 10 13.7 1.41§** 0.24
X 2 1.1 1 0.9 1 1.4
Nuclear grading †‡
I 46 24.5 27 23.5 19 26.0
II 67 35.6 43 37.4 24 32.9
III 48 25.5 29 25.2 19 26.0
IV 27 14.4 16 13.9 11 15.1 0.43 0.94
* Based on pathological TNM unless unknown, then clinical TNM (UICC, 1987) was used [17]
† Based on the review by one experienced pathologist (CAHK)
‡ According to Fuhrman [19]
§1–2 vs. 3–4
¶1–2 vs. 0
** Excludes "X (unknown) " category
==== Refs
Kovacs G Akhtar M Beckwith BJ Bugert P Cooper CS Delahunt B Eble JN Fleming S Ljungberg B Medeiros LJ Moch H Reuter VE Ritz E Roos G Schmidt D Srigley JR Storkel S van den Berg E Zbar B The Heidelberg classification of renal cell tumours [editorial] J Pathol 1997 183 131 133 9390023 10.1002/(SICI)1096-9896(199710)183:2<131::AID-PATH931>3.3.CO;2-7
Gnarra JR Tory K Weng Y Schmidt L Wei MH Li H Latif F Liu S Chen F Duh FM Lubensky I Duan DR C. F Pozzatti R Walther MM Bander NH Grossman HB Brauch H Pomer S Brooks JD Isaacs WB Lerman MI Zbar B Linehan WM Mutations of the VHL tumour suppressor gene in renal carcinoma Nat Genet 1994 7 85 90 7915601 10.1038/ng0594-85
Latif F Tory K Gnarra J Yao M Duh FM Orcutt ML Stackhouse T Kuzmin I Modi W Geil L Schmidt L Zhou F Li H Wei MH Chen F Glenn G Choyke P Walther MM Weng Y Duan DR Dean M Glavac D Richards FM Crossey PA Ferguson-Smith MA Le Paslier D Chumakov I Cohen D Chinault AC Maher ER Linehan WM Zbar B Lerman MI Identification of the von Hippel-Lindau disease tumor suppressor gene Science 1993 260 1317 1320 8493574
Zbar B Brauch H Talmadge C Linehan M Loss of alleles of loci on the short arm of chromosome 3 in renal cell carcinoma Nature 1987 327 721 724 2885753 10.1038/327721a0
Gnarra JR Duan DR Weng Y Humphrey JS Chen DY Lee S Pause A Dudley CF Latif F Kuzmin I Schmidt L Duh FM Stackhouse T Chen F Kishida T Wei MH Lerman MI Zbar B Klausner RD Linehan WM Molecular cloning of the von Hippel-Lindau tumor suppressor gene and its role in renal carcinoma Biochim Biophys Acta 1996 1242 201 210 8603073
Richards FM Molecular pathology of von Hippel-Lindau disease and the VHL tumor suppressor gene Expert Rev Mol Med 2001 2001 1 27 14987375 10.1017/S1462399401002654
Kaelin WG Jr The von hippel-lindau tumor suppressor gene and kidney cancer Clin Cancer Res 2004 10 6290S 6295S 15448019
Cohen HT Advances in the molecular basis of renal neoplasia Curr Opin Nephrol Hypertens 1999 8 325 331 10456263 10.1097/00041552-199905000-00008
Herman JG Latif F Weng Y Lerman MI Zbar B Liu S Samid D Duan DR Gnarra JR Linehan WM Baylin SB Silencing of the VHL tumor-suppressor gene by DNA methylation in renal carcinoma Proc Natl Acad Sci U S A 1994 91 9700 9704 7937876
McLaughlin JK Lipworth L Epidemiologic aspects of renal cell cancer Semin Oncol 2000 27 115 123 10768591
Brauch H Weirich G Hornauer MA Storkel S Wohl T Bruning T Trichloroethylene exposure and specific somatic mutations in patients with renal cell carcinoma J Natl Cancer Inst 1999 91 854 861 10340905 10.1093/jnci/91.10.854
Brauch H Weirich G Klein B Rabstein S Bolt HM Bruning T VHL mutations in renal cell cancer: does occupational exposure to trichloroethylene make a difference? Toxicol Lett 2004 151 301 310 15177666 10.1016/j.toxlet.2003.12.074
Hemminki K Jiang Y Ma X Yang K Egevad L Lindblad P Molecular epidemiology of VHL gene mutations in renal cell carcinoma patients: relation to dietary and other factors Carcinogenesis 2002 23 809 815 12016154 10.1093/carcin/23.5.809
Shiao YH Rice JM Anderson LM Diwan BA Hard GC von Hippel-Lindau gene mutations in N-nitrosodimethylamine-induced rat renal epithelial tumors [see comments] J Natl Cancer Inst 1998 90 1720 1723 9827526 10.1093/jnci/90.22.1720
Van den Brandt PA Goldbohm RA Van 't Veer P Volovics A Hermus RJ Sturmans F A large-scale prospective cohort study on diet and cancer in The Netherlands J Clin Epidemiol 1990 43 285 295 2313318 10.1016/0895-4356(90)90009-E
Van den Brandt PA Schouten LJ Goldbohm RA Dorant E Hunen PM Development of a record linkage protocol for use in the Dutch Cancer Registry for Epidemiological Research Int J Epidemiol 1990 19 553 558 2262247
Hermanek P Sobin L TNM Classification of Malignant Tumours 1987 Fourth Springer-Verlag
Eble J Sauter G Epstein J Sesterhenn I World Health Organization Classification of Tumours. Pathology and Genetics. Tumours of the Urinary System and Male Genital Organs 2004 Lyon: IARC Press
Fuhrman S Lasky L Limas C Prognostic significance of morphologic parameters in renal cell carcinoma Am J Surg Pathol 1982 6 655 663 7180965
Bailly M Bain C Favrot MC Ozturk M Somatic mutations of von Hippel-Lindau (VHL) tumor-suppressor gene in European kidney cancers Int J Cancer 1995 63 660 664 7591282
Brauch H Weirich G Brieger J Glavac D Rodl H Eichinger M Feurer M Weidt E Puranakanitstha C Neuhaus C Pomer S Brenner W Schirmacher P Storkel S Rotter M Masera A Gugeler N Decker HJ VHL alterations in human clear cell renal cell carcinoma: association with advanced tumor stage and a novel hot spot mutation Cancer Res 2000 60 1942 1948 10766184
Barnabas N Amin MB Pindolia K Nanavati R Worsham MJ Mutations in the von Hippel-Lindau (VHL) gene refine differential diagnostic criteria in renal cell carcinoma J Surg Oncol 2002 80 52 60 11967908 10.1002/jso.10086
Foster K Prowse A van den Berg A Fleming S Hulsbeek MMF Crossey PA Richards FM Cairns P Affara NA Ferguson-Smith MA Buys CHCM Maher ER Somatic mutations of the von Hippel-Lindau disease tumour suppressor gene in non-familial clear cell renal carcinoma Hum Mol Genet 1994 3 2169 2173 7881415
Kondo K Yao M Yoshida M Kishida T Shuin T Miura T Moriyama M Kobayashi K Sakai N Kaneko S Kawakami S Baba M Nakaigawa N Nagashima Y Nakatani Y Hosaka M Comprehensive mutational analysis of the VHL gene in sporadic renal cell carcinoma: relationship to clinicopathological parameters Genes Chromosomes Cancer 2002 34 58 68 11921283 10.1002/gcc.10054
Gallou C Joly D Mejean A Staroz F Martin N Tarlet G Orfanelli MT Bouvier R Droz D Chretien Y Marechal JM Richard S Junien C Beroud C Mutations of the VHL gene in sporadic renal cell carcinoma: definition of a risk factor for VHL patients to develop an RCC [In Process Citation] Hum Mutat 1999 13 464 475 10408776 10.1002/(SICI)1098-1004(1999)13:6<464::AID-HUMU6>3.0.CO;2-A
Gallou C Longuemaux S Delomenie C Mejean A Martin N Martinet S Palais G Bouvier R Droz D Krishnamoorthy R Junien C Beroud C Dupret JM Association of GSTT1 non-null and NAT1 slow/rapid genotypes with von Hippel-Lindau tumour suppressor gene transversions in sporadic renal cell carcinoma Pharmacogenetics 2001 11 521 535 11505222 10.1097/00008571-200108000-00007
Ma X Yang K Lindblad P Egevad L Hemminki K VHL gene alterations in renal cell carcinoma patients: novel hotspot or founder mutations and linkage disequilibrium Oncogene 2001 20 5393 5400 11536052 10.1038/sj.onc.1204692
Suzuki H Ueda T Komiya A Okano T Isaka S Shimazaki J Ito H Mutational state of von Hippel-Lindau and adenomatous polyposis coli genes in renal tumors Oncology 1997 54 252 257 9143408
Yang K Lindblad P Egevad L Hemminki K Novel somatic mutations in the VHL gene in Swedish archived sporadic renal cell carcinomas [In Process Citation] Cancer Lett 1999 141 1 8 10454237 10.1016/S0304-3835(99)00031-2
Hamano K Esumi M Igarashi H Chino K Mochida Ji Ishida H Okada K Biallelic inactivation of the von Hippel-Lindau tumor suppressor gene in sporadic renal cell carcinoma J Urol 2002 167 713 717 11792959 10.1097/00005392-200202000-00071
Beroud C Collod-Beroud G Boileau C Soussi T Junien C UMD (Universal mutation database): a generic software to build and analyze locus-specific databases Hum Mutat 2000 15 86 94 10612827 10.1002/(SICI)1098-1004(200001)15:1<86::AID-HUMU16>3.0.CO;2-4
Novick AC Campbell SC Walsh PC, Retik AB, Darracott Vaughan Jr E, Wein AJ, Kavoussi LR, Novick AC, Partin AW, Peters CA 75. Renal tumors Campbell's Urology 2002 4 eighth Philadelphia: Saunders 2672 2719
Zhuang Z Gnarra JR Dudley CF Zbar B Linehan WM Lubensky IA Detection of von Hippel-Lindau disease gene mutations in paraffin-embedded sporadic renal cell carcinoma specimens Mod Pathol 1996 9 838 842 8871925
Williams C Ponten F Moberg C Soderkvist P Uhlen M Ponten J Sitbon G Lundeberg J A high frequency of sequence alterations is due to formalin fixation of archival specimens Am J Pathol 1999 155 1467 1471 10550302
Maynard MA Ohh M Von Hippel-Lindau tumor suppressor protein and hypoxia-inducible factor in kidney cancer Am J Nephrol 2004 24 1 13 14654728 10.1159/000075346
Iliopoulos O Ohh M Kaelin WG Jr pVHL19 is a biologically active product of the von Hippel-Lindau gene arising from internal translation initiation Proc Natl Acad Sci U S A 1998 95 11661 11666 9751722 10.1073/pnas.95.20.11661
Blankenship C Naglich JG Whaley JM Seizinger B Kley N Alternate choice of initiation codon produces a biologically active product of the von Hippel Lindau gene with tumor suppressor activity Oncogene 1999 18 1529 1535 10102622 10.1038/sj.onc.1202473
Schoenfeld A Davidowitz EJ Burk RD A second major native von Hippel-Lindau gene product, initiated from an internal translation start site, functions as a tumor suppressor Proc Natl Acad Sci U S A 1998 95 8817 8822 9671762 10.1073/pnas.95.15.8817
| 15932632 | PMC1177929 | CC BY | 2021-01-04 16:03:05 | no | BMC Cancer. 2005 Jun 2; 5:57 | utf-8 | BMC Cancer | 2,005 | 10.1186/1471-2407-5-57 | oa_comm |
==== Front
BMC CancerBMC Cancer1471-2407BioMed Central London 1471-2407-5-591594903510.1186/1471-2407-5-59Research ArticleOverexpression of extracellular superoxide dismutase reduces acute radiation induced lung toxicity Rabbani Zahid N [email protected] Mitchell S [email protected] Rodney J [email protected] Emerald [email protected] Hong [email protected] Liguang [email protected] Maria L [email protected] Thaddeus S [email protected] Mark W [email protected] Zeljko [email protected] Departments of Radiation Oncology, Duke University Medical Center, Durham, North Carolina USA2 Department of Medicine, Division of Pulmonary, Allergy and Critical Care Medicine, Duke University Medical Center, Durham, North Carolina, USA2005 10 6 2005 5 59 59 8 9 2004 10 6 2005 Copyright © 2005 Rabbani 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
Acute RT-induced damage to the lung is characterized by inflammatory changes, which proceed to the development of fibrotic lesions in the late phase of injury. Ultimately, complete structural ablation will ensue, if the source of inflammatory / fibrogenic mediators and oxidative stress is not removed or attenuated. Therefore, the purpose of this study is to determine whether overexpression of extracellular superoxide dismutase (EC-SOD) in mice ameliorates acute radiation induced injury by inhibiting activation of TGFβ1 and downregulating the Smad 3 arm of its signal transduction pathway.
Methods
Whole thorax radiation (single dose, 15 Gy) was delivered to EC-SOD overexpressing transgenic (XRT-TG) and wild-type (XRT-WT) animals. Mice were sacrificed at 1 day, 1 week, 3, 6, 10 and 14 weeks. Breathing rates, right lung weights, total/differential leukocyte count, activated TGFβ1 and components of its signal transduction pathway (Smad 3 and p-Smad 2/3) were assessed to determine lung injury.
Results
Irradiated wild-type (XRT-WT) animals exhibited time dependent increase in breathing rates and right lung weights, whereas these parameters were significantly less increased (p < 0.05) at 3, 6, 10 and 14 weeks in irradiated transgenic (XRT-TG) mice. An inflammatory response characterized predominantly by macrophage infiltration was pronounced in XRT-WT mice. This acute inflammation was significantly attenuated (p < 0.05) in XRT-TG animals at 1, 3, 6 and 14 weeks. Expression of activated TGFβ1 and components of its signal transduction pathway were significantly reduced (p < 0.05) at later time-points in XRT-TG vs. XRT-WT.
Conclusion
This study shows that overexpression of EC-SOD confers protection against RT-induced acute lung injury. EC-SOD appears to work, in part, via an attenuation of the macrophage response and also decreases TGFβ1 activation with a subsequent downregulation of the profibrotic TGFβ pathway.
==== Body
Background
Acute radiation (RT) induced lung toxicity and subsequently occurring pulmonary fibrosis are considered to be critical, dose-limiting factors for radiation therapy of thoracic malignancies [1]. Subclinical molecular and cellular events commence during RT therapy, but the clinical features and histological findings may not be revealed for months, or even years after the treatment [2].
In RT-induced lung injury, damage to endothelial and epithelial cells are thought to be the initial steps leading to acute lung toxicity [3]. Tissue damage and repair initiated by irradiation is associated with the production of important biological mediators, such as cytokines [4]. These cytokines perpetuate the inflammatory and fibrogenic processes associated with RT injury [5]. The relative role of cytokine dysregulation versus direct tissue injury from irradiation in the pathogenesis of acute and late toxicities is not well defined. Previous observations and recent advancements in understanding the role of these cytokines implicate transforming growth factor-β1 (TGF-β) as a key component in the development of RT-induced normal tissue injury in multiple organs, including lungs [6].
In addition to cytokine dysregulation, an oxidant/antioxidant imbalance in the lower respiratory tract has been proposed as the mechanism of lung injury in a number of inflammatory lung conditions [7,8]. Ionizing RT is associated with increased production of free radicals [9], which is reflected by the accumulation of oxidatively damaged cellular macromolecules. RT may impair lung cells either directly via generation of reactive oxygen species (ROS) [9] or indirectly via the action on parenchymal and inflammatory cells through biological mediators [4,5]. This process may subordinate the cellular antioxidant defenses and lead to the accumulation of toxic levels of ROS.
Extracellular superoxide dismutase (EC-SOD), one of the subtypes of naturally occurring superoxide dismutases, is the dominant antioxidant enzyme found in a variety of extracellular compartments [10]. These enzymes act by catalyzing the dismutation of the superoxide anion radical to oxygen and hydrogen peroxide. EC-SOD is secreted into the extracellular spaces, and in the lung it is expressed primarily by the alveolar type II pneumocytes [11]. EC-SOD is mainly bound to the extracellular matrix but it is also detectable in plasma [12]. The protective role of this enzyme has been studied in RT, bleomycin or hyperoxia induced oxidative stress lung injury [13-15].
Recently, we have reported that overexpression of EC-SOD in transgenic mice appears to protect against RT-induced chronic injury [13]. In the present study, we analyzed the early events at the molecular and cellular levels to clarify the mechanisms by which overexpression of EC-SOD protects against RT-induced lung damage. We hypothesized that continuous over-production of ROS, during and well after radiotherapy has been completed, is responsible for the pathogenesis of RT-induced lung injury and that continuous overexpression of EC-SOD would ameliorate this injury.
Methods
Animals
Transgenic (TG) B6C3 mice that overexpress human EC-SOD (hEc-SOD) in alveolar and airway epithelial cells and wild-type (WT) littermates were utilized for this study. The generation of the TG mice has been described in detail by Folz et. al. [15]. These mice are maintained in a B6C3 background. Heterozygous transgenic positive mice are always bred to an F1 (C57BL/6 × C3H) mouse, and the transgenic positive pups are compared to their transgenic negative littermates. Polymerase chain reaction (PCR) analysis of tail DNA was performed to confirm the genotype status. PCR positive and negative animals were used for all these studies. The animals were maintained in cages at room temperature, with a 12 hr light-dark cycle with access to food and water ad libitum. The protocol was approved by the Duke University Medical Center Institutional Animal Care and Use Committee.
Thoracic irradiation
The animals were randomly distributed into four groups; unirradiated, wild-type (Con -WT), EC-SOD overexpressing transgenic (Con -TG), irradiated EC-SOD overexpressing transgenic (XRT-TG) and wild-type mice (XRT-WT). RT was delivered with a single dorsal-ventral field using 4-MV photons, for a dose of 15 Gy to the whole thorax with 0.5 cm bolus material. Mice were sacrificed at six different time points post irradiation (1 day, 1, 3, 6, 10 and 14 wks), along with matched controls.
Breathing frequency
Functional assessment of lung damage was performed weekly, by measuring breathing frequency, using a whole-body plethysmography chamber (Model RM-80; Columbus Instruments, Columbus, Ohio). Four readings were taken at each time point for every animal, and the mean value was used for analysis.
Broncho-alveolar lavage fluid
Mice were anesthetized by intraperitoneal injection of pentobarbital (50 μg/g). After the onset of adequate anesthesia, the trachea was exposed and a 20-gauge needle was used to puncture the trachea, which was replaced with a 21-gauge blunt needle. A silk ligature was fastened around the trachea to secure the needle. Bronchoalveolar lavage (BAL) was performed with 1 ml PBS using a 1-ml syringe and repeated once. The collected fluid was immediately processed as follows. BAL fluid (300 μL) was cytocentrifuged and stained with a leukostat stain kit (Fisher Scientific Co., Pittsburgh, Pennsylvania, USA). A minimum of 300 cells were subsequently counted to obtain a differential cell count. Cellular BALF (30–50 μL) was stained with trypan blue, and the total cell count was manually obtained using a Neubauer hemocytometer (Reichert, Buffalo, New York, USA). The remaining BALF was centrifuged at 1,800 g for 5 minutes at 4°C, and the supernatant was decanted and centrifuged a second time at 12,000 g for 10 minutes at 4°C. The supernatant was collected and saved at -80°C for further analysis, and the pelleted cells were snap frozen in liquid nitrogen and stored at -80°C.
Wet lung weights
At the time of sacrifice, the right lung was removed, gently blotted, and the wet weight recorded. The right lung was snap frozen in liquid nitrogen and then stored in -80°C freezer.
Histopathology
The left lungs from mice in each group and at each time point were fixed in situ for microscopy by intratracheal instillation of 4% paraformaldehyde in PBS at 20 cm H2O pressure for 10 minutes and subsequently removed from the thorax and immersed in additional fixative for 4–6 hours. These tissues were paraffin-embedded, sectioned, and stained with hematoxylin and eosin. Histopathologic evidence for lung injury was scored on a scale from 0 (normal) to 4 (most severe) as described earlier [16]. Parameters assessed were the degree of vascular congestion, thickening of alveolar wall (edema), hyaline membrane formation in the distal airway, and inflammatory cell accumulation.
Tissue TGF-β1 activity
Frozen lung tissue was thawed at room temperature, then placed in 1X phosphate buffered saline (PBS) with 0.5 mM EDTA according to weight (10 μL of buffer per 1 mg of tissue). The sample was homegenized by sonication, then spun at 15,000 rpm for 30 minutes at 4°C. The supernatant was used for the sandwich enzyme-linked immunosorbent assay (ELISA) in order to quantify the TGF-β1. To measure the active TGF-β1, anti-TGF-β1 monoclonal antibody and biotinylated anti-TGF-β1 antibodies were used as the capture and probe antibodies, respectively. Horseradish peroxidase-conjugated with streptavidin was used to bind the biotin, followed by substrate reagent and stop solution (all products are from R&D Systems Inc, Minneapolis, MN). To quantify the total TGF-β1, latent TGF-β1 was activated by adding 0.1 ml 2.5 N Acid / 10 M Urea into 0.1 ml sample and neutralized by adding 0.1 ml 2.7 N NaOH / 1 M Hepes.
Ninety-six-well microtiter plates were read at OD 450 nm using a kinetic microplate reader (v-max; Molecular Devices, Menlo Park, CA). The level of TGF-β1 (total and active) was determined by comparing the peroxidase activity to known concentrations of purified TGF-β1 (R&D Systems Inc, Minneapolis, MN).
Immunohistochemistry (Smad3 And p-Smad2/3)
Immunohistochemistry was carried out as previously described [17]. Briefly, paraffin embedded tissues were sectioned at 5 μm thickness, and deparaffinized and hydrated using a xylene solution and graded ethanol. An endogenous peroxide blocking solution of 3% hydrogen peroxide was applied for 20 minutes at room temperature then rinsed in deionized water. Sections were then placed in an antigen retrieval solution of citrate buffer from Biogenex (San Ramon, CA). The solution was heated in a microwave for 3–4 minutes, cooled for 2–3 minutes, then heated again in a microwave. After cooling at room temperature for 20 minutes, sections were rinsed in deionized water. Next, the sections were incubated in 10% donkey serum in phosphate-buffered saline (PBS) for 25 minutes at room temperature to reduce non-specific binding of antibody. Sections were then rinsed three times with 1X PBS, covered with primary antibodies against the Smad 3 and p-Smad 2/3 (dilution 1:200, Santa Cruz Biotechnology), incubated in a humidity chamber at 37°C for 60 minutes. After rinsing three times with 1% PBS, secondary and tertiary antibodies were added to the sections per manufacturer instructions (314KLD, Innovex, Richmond, CA), with incubation in the humidity chamber at 37°C for 30 minutes. Sections were again rinsed three times in 1% PBS, then incubated for 2 minutes at room temperature in a peroxidase substrate solution with 3, 3'-diaminobenzidine tetrahydrochloride (Sigma, St. Louis, MO). The slides were washed three times for 10 minutes in deionized water, and counterstained with hematoxylin. After rinsing again with deionized water, sections were dehydrated in graded ethanol and xylene, then mounted. Normal serum was used as controls with each set of stains.
Scoring of positive cells
The lung specimens were subjected to blinded evaluation. After scanning the whole lung section and calculating the extent of damaged lung parenchyma, 4 representative fields (40X objective) were selected and analyzed. After initial qualitative assessment of morphologic changes, cells that showed staining after RT exposure to the Smad-3 and p-Smad-2/3 antibodies were recorded. The sum of the measurement from the 4 fields was totaled, and the arithmetic mean was calculated. Positive cell counts were expressed as the average number of cells per field. Means from lungs of irradiated animals were compared with the corresponding values obtained from control animals.
Statistical methods
The Student's t test and One-way ANOVA was used to test the significance of any differences between groups at each time. A p-value ≤ 0.05 was considered statistically significant. All datas are presented as mean values (± 95% confidence interval).
Results
Breathing frequency
The pulmonary function of the animals was assessed by monitoring the respiratory rate every week for 14 weeks. The XRT-WT mice exhibited a progressive increase in breathing rates after irradiation beginning at 3 weeks in comparison to control animals (Figure 1). XRT-TG animals have a delay in the onset of the functional lung damage (3 weeks), with the significant decrease (p < 0.05) in respiratory rate as compared to XRT-WT mice. During the follow-up period of 14 weeks the mean breathing rates of the two control groups were not changed from baseline values (249 ± 2.5 vs. 251 ± 2).
Right lung wet weights
The amount of acute lung damage elicited by RT was measured by an increase in right lung wet weights, which are an index of pulmonary edema and consolidation. RT treatment resulted in a significant increase in right lung wet weights in the XRT-WT group (Figure 2). The XRT-TG mouse strain showed significantly less (p < 0.05) increase in wet weight from 3 week to 14 weeks, when compared with their XRT-WT littermates.
Broncho-alveolar lavage fluid cell counts
There was no difference in total cell counts between XRT-WT and XRT-TG mice 1 day after irradiation (Figure 3a). However, the total cell count increased significantly by 1 week in the XRT-WT mice when compared with the XRT-TG mice and remained elevated till 14 weeks (p < 0.05).
The increase in BALF cell count was mainly due to an increase in the number of macrophages (Figure 3b). Beginning at 1 week, the XRT-WT mice showed a significantly greater increase in macrophages, which persisted throughout the period of observation (XRT-WT vs. XRT-TG, at 1, 3, 10, and 14 weeks, p < 0.05).
The lymphocyte count was also significantly increased in XRT-WT animals beginning at 3 weeks (Figure 3c) as compared to the XRT-TG animals (XRT-WT vs. XRT-TG, at 3 and 10 weeks, p < 0.05).
Histopathology
At 3 weeks after irradiation, the H & E stained lung sections of XRT-WT mice showed evidence for decreased pulmonary compliance, as indicated by increased edema, thickening of the alveolar walls, vascular and interstitial congestion, and diffuse inflammatory cell infiltration, which worsened with time (Figure. 4C, E). Histological changes were also observed in XRT-TG mice after 3 weeks (Figure. 4D); however they were lesser in degree and sparse in distribution compared to the XRT-WT animals (Figure. 4C). In the later time points, the extent and severity of injury was greater in XRT-WT groups compared to their XRT-TG animals (Figure 4E, F).
TGFβ activity
The results of the quantitative assessment of levels of active and total TGFβ1 protein expression in lung tissue are shown in Figure 5. Unirradiated lung tissue of control animals expressed lower levels of TGFβ1. Following thoracic irradiation, lung tissue levels of active TGFβ1 in XRT-WT were increased within the first week. The active TGFβ1 levels were significantly increased from 6 weeks onward. The XRT-TG animals demonstrated significantly less TGFβ1 activation as compared to XRT-WT mice (XRT-WT vs. XRT-TG, at 6, 10, and 14 weeks, p < 0.05, Figure 5).
Signal transduction
Increased immunoreactivity for Smad 3 in lung parenchyma of XRT-WT animals was noticeable at 1 week, and which steadily increased from 3–14 weeks. The temporal increase with XRT-TG was less steep than XRT-WT (XRT-WT vs. XRT-TG, at 3, 6, 10 and 14 weeks, p < 0.05, Figure 6a) (Representative images of XRT-WT vs. XRT-TG at 3 and 14 weeks after irradiation shown in Figure 7C–F).
In XRT-WT mice, there was significant increase in p-Smad2/3 positivity by 3 weeks post irradiation (XRT-WT vs. XRT-TG, at 3, 6, 10 and 14 weeks, p < 0.05, Figure 6b). With time, the fraction of positively staining cells also increased in XRT-TG animals but this p-Smad2/3 expression was significantly diminished when compared with XRT-WT. Thus, RT not only increased TGFβ expression and activation, but also increased signal transduction down the fibrosis pathway. This effect was significantly attenuated by over-expression of ECSOD. (Representative images of XRT-WT vs. XRT-TG at 3 and 14 weeks after irradiation shown in Figure 7G–J).
Discussion
The current evidence indicates that acute RT-induced damage to the lung is characterized by inflammatory changes, which proceed to the development of fibrotic lesions in the late phase. Ultimately, complete structural ablation will ensue, if the source of inflammatory/fibrogenic mediators and oxidative stress is not removed or attenuated [13]. Our data demonstrate that overexpression of ECSOD in mouse lung appears to confer protection from acute RT damage by reducing the inflammatory response, decreasing in TGFβ1 activation and, consequently, downregulation of TGFβ signal transduction pathway.
Prior work by our laboratory has revealed that EC-SOD is protective against chronic RT-induced lung injury, but we were unable to differentiate between the acute phase inflammatory response (pneumonitis /alveolitis) and the fibrotic response to pulmonary irradiation because only a single late time point was studied [13]. To begin to answer this question, we used TG mice in which alveolar and airway epithelial cells overexpress EC-SOD. Thus, by using this TG mouse model we were able to enhance the antioxidant capacity of the extracellular pulmonary compartment in which antioxidant reserves may play critical roles in preserving pulmonary function against early RT damage.
The lung is among one of the most radiosensitive viscera to ionizing RT. Ionizing RT has long been recognized to induce an inflammatory response within irradiated tissues [18-22]. Strong evidence implicates a key role for inflammation in the development of radiation induced pulmonary injury [23]. Suppression of this acute inflammation may lead to a reduction in subsequent fibrosis.
The mechanism underlying the inflammatory response in RT-induced lung injury is complex. Endothelial damage leads to leukocyte adhesion and transmigration into tissues. Inflammatory cells are a known source of ROS [24,25]. These cells can also release inflammatory and fibrogenic cytokines [26], which further contribute to injury in part through the recruitment of more inflammatory cells to the site of injury [5]. The time course for recruitment of the inflammatory cells in the lower respiratory tract of XRT-WT animals in this study demonstrates the important role of inflammation in early radiation injury. In the presence of a potent anti-inflammatory stimulus, i.e., EC-SOD, XRT-TG animals suffered significantly less RT-induced pulmonary damage.
The anti-inflammatory effects of EC-SOD also appear to reduce the risk of subsequent fibrosis. The elevated levels of EC-SOD in transgenic mice has been shown to be protective against the development of pulmonary fibrosis from a number of stimuli, including RT, hyperoxia and bleomycin induced pulmonary fibrosis [13-15]. EC-SOD TG mice have been shown to better tolerate hyperoxia, in part by reducing the excessive influx of inflammatory cells into the lung [15]. In a mouse model of influenza virus-induced pneumonia, the overexpression of EC-SOD markedly ameliorated the inflammatory and oxidant responses in the lung [27]. This protective effect of EC-SOD was attributed to inhibitory action against a series of inflammatory mediators. Similar effects were also noticed after exogenous administration of SOD's [28,29] in influenza- induced lung injury.
In contrast to EC-SOD overexpressing animals, EC-SOD knockout mice have shown evidence for increased sensitivity to hyperoxia exposure, which was exhibited by earlier onset of inflammatory reactions and interstitial edema, when compared with wild-type mice [30]. In another study, EC-SOD null mice demonstrated increased susceptibility to inflammation and pulmonary fibrosis. From these results the authors proposed that one mechanism by which EC-SOD protects against pulmonary fibrosis is by inhibiting inflammation [31]. This appears to be the case following radiation as well.
One of the means by which EC-SOD appears to reduce pulmonary inflammation is through suppressing activation and recruitment of macrophages. Macrophages increase in number shortly after irradiation and have been shown to produce a broad spectrum of cytokines, which stimulate fibroblast proliferation and the production of an extracellular matrix [32]. These include platelet-derived growth factor (PDGF) [33], interleukin-1 (IL-1) [34], tumor necrosis factor α (TNFα) [35], basic fibroblast growth factor (bFGF) [36], and TGFβ [37]. These cytokines not only regulate fibroblast function and collagen synthesis, but also recruit other inflammatory cells and contribute to the pathogenesis of both acute and late RT damage [38]. The macrophage count of XRT-TG mice in this study at different time points revealed a significant reduction when compared with XRT-WT. This observation supports the notion that the protective effect of EC-SOD is mediated, at least in part, by its ability to attenuate the influx of macrophages.
Several lines of evidence have supported the use of therapeutic strategies targeting TGFβ1 to prevent fibroproliferative disorders of various organs [3]. Due to its implication in lung fibrogenesis, blockade of the TGFβ pathway has been proposed as a molecular strategy to ameliorate pulmonary fibrosis and is therefore of potential clinical significance [39]. Using a soluble version of TGFβ type II receptor, Wang et al [40] found that pulmonary fibrosis, including excess collagen accumulation, was greatly reduced in hamsters after bleomycin administration. Similarly, we found that soluble TGFβ type II receptor gene therapy led to a significant reduction of acute RT-induced lung injury in rats 4 weeks after irradiation [41]. Interference with the TGFβ signal transduction pathway has also been shown to protect against fibrosis formation. Recently, Smad7, an intracellular antagonist of TGFβ signaling, has been shown to attenuate bleomycin-induced lung fibrosis in mice receiving intratracheal injection of recombinant adenovirus overexpressing Smad7 [42]. Smad3 knockout mice have been shown to be protected from RT-induced soft tissue injury [43]. This loss of Smad3 interfered with the chemotactic response of neutrophils, macrophages, keratinocytes, and fibroblasts to TGFβ [44-47]. Upregulation of TGFβ at sites of injury was also Smad3 dependent [43].
Our data further support the notion that TGFβ1 signaling through the Smad 3 pathway is critical to the development of RT-induced injury. Overexpression of EC-SOD leads to downregulation of this signaling pathway as evidenced by both reduced expression and phosphorylation of Smad3. Inhibition of Smad3, or even partial reduction in its activity might afford protection against the toxicities of irradiation. Taken together, these results suggest that abrogation of TGFβ signaling might be sufficient to offer protection against lung fibrogenesis.
In summary, this study shows that overexpression of EC-SOD confers protection against RT induced acute lung injury. EC-SOD appears to work, in part, via an attenuation of the macrophage response. In addition, sustained EC-SOD expression decreases TGFβ activation with a subsequent downregulation of the profibrotic TGF-β pathway. Thus, EC-SOD in the extracellular compartments of the lung is an important modulator of the complex response of normal tissue to acute radiation insult.
Abbreviations
TGFβ1 : Transforming growth factor Beta
EC-SOD : Extracellular Superoxide Dismutase
RT : Radiation
XRT-WT : Irradiated wildtype
XRT-TG : Irradiated transgenic
ROS : Reactive oxygen species
BALF : Broncho-alveolar lavage fluid
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
ZNR performed the study (breathing frequency, wet lung weight, IHC, statistical analysis of the data) and drafted the manuscript. RJF and MG provided animals and assisted in BALF cell count analysis. EA and LC helped in histology/IHC and assisted in conducting the study. TVS performed whole thorax irradiation to the animals. HH performed TGF-β ELISA. MWD helped in editing of manuscript. ZV and MSA conceived of the study, participated in its design and coordination, and assisted in editing of manuscript. All authors read and approved the final manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
Supported by National Cancer Institute Grant CA98452.
Figures and Tables
Figure 1 Comparison of breathing frequency over time. Changes in breathing rate after 15 Gy of single dose of irradiation to whole thorax in XRT-WT and XRT-TG groups. XRT-WT group showed a significant increase in respiratory rate vs. XRT-TG, at 3, 6, 10 and 14 wks, * p < 0.05. Error bars represent 95% confidence intervals.
Figure 2 Comparison of right lung wet weights over time. Lung weights at 1 day, 1,3, 6, 10 and 14 weeks of exposure to radiation. A significant increase in the wet weight of XRT-WT lungs was observed by 3 wks of radiation exposure (XRT-WT vs. XRT-TG, at 3, 6, 10 and14 wks, * p < 0.05). Error bars represent 95% confidence intervals.
Figure 3 Comparison of cell counts in bronchoalveolar lavage fluid (BALF) from XRT-WT and XRT-TG mice exposed to radiation. (a) Total cell counts. and (b) Macrophage cell counts at 1day, 1,3, 6, 10 and 14 weeks of exposure to radiation. Both total cell and macrophage counts significantly increased (* p < 0.05) with time of exposure in XRT-WT animals (c) Lymphocyte cells count was significantly increased by 3 wk, in the BALF of the XRT-WT group as compared to the XRT-TG mice (XRT-WT vs. XRT-TG, at 3, and 10 wks, * p < 0.05). Error bars represent 95% confidence intervals.
Figure 4 Light microscopy of hematoxylin and eosin-stained section of representative lungs from controls (A, B ) and treatment groups (XRT-WT vs. XRT-TG) at 3 wk (C, D ) and 14 wks (E, F ). Representative lung sections from XRT-WT exposed to 3 wk of radiation showing mild to moderate damage at 3 wk after radiation (C ), which gradually increased with time course of injury and become moderate to severe with increased thickening of alveolar wall and enhanced infiltration composed mainly of macrophages and other inflammatory cells at 14 weeks (E ). Whereas the XRT-TG mice showing mild-modest alveolar septal thickness and fewer numbers of alveolar macrophages and other inflammatory cells (D, F ). Magnification × 400.
Figure 5 Comparison of TGFβ1 activation in irradiated lung tissue. Increase active TGFβ1 expression began by 1 week. By 6 weeks, there was significantly more activated TGFβ1 in the lungs of XRT-WT mice (grey squares) vs. XRT-TG animals (grey triangles) (XRT-WT vs. XRT-TG, at 6, 10 and 14 wks, # p < 0.05). There were significantly more total TGFβ1 levels in XRT-WT (black squares) vs. XRT-TG animals (black triangles) at late time points (XRT-WT vs. XRT-TG, at 10 and 14 wks, * p < 0.05). Error bars represent 95% confidence intervals.
Figure 6 a): Smad3 expression after irradiation. Cells expressing Smad3 were counted at 40X magnification in four random fields by two blinded observers. The Smad3 positive cell count for each animal was the mean of all eight readings. Increased expression of Smad3 in XRT-WT lungs was noticeable at 1 wk, and which steadily increased from 3–14 wks. This temporal increase in XRT-TG was significantly reduced than XRT-WT (XRT-WT vs. XRT-TG, at 3, 6, 10 and14 wks, * p < 0.05). b): p-Smad2/3 expression after irradiation. For p-Smad2/3, cell positivity was counted at 40X and mean value for each animal was taken. In XRT-WT lung, there was significant increase in p-Smad2/3 expression by 3 weeks post irradiation. With time, the fraction of positively staining cells also increased in XRT-TG animals but the level of expression in this group was significantly decreased when compared with XRT-WT (XRT-WT vs. XRT-TG, at 3, 6, 10 and 14 wks, * p < 0.05). Error bars represent 95% confidence intervals.
Figure 7 Representative images of Smad3 immunohistochemistry of controls (A, B ) and XRT-WT vs. XRT-TG groups at 3 wk (C, D ) and 14 wks (E, F ). For Smad3, most positively stained cells are present in damaged areas of XRT-WT lungs, which showed significantly increased cell positivity when compared with XRT-TG animals. Representative images of p-Smad2/3 immunohistochemistry of XRT-WT vs. XRT-TG groups at 3 wk (G, H ) and 14 wks (I, J ) after irradiation. XRT-WT animals exhibited significantly increased positivity for p-Smad2/3 when compared with XRT-TG mice. See the results section for the details. Magnification × 400.
==== Refs
Vujaskovic Z Marks LB Anscher MS The physical parameters and molecular events associated with radiation-induced lung toxicity Semin Radiat Oncol 2000 10 296 307 11040330
Stone HB Coleman CN Anscher MS McBride WH Effects of radiation on normal tissue: consequences and mechanisms Lancet Oncol 2003 4 529 536 12965273 10.1016/S1470-2045(03)01191-4
Martin M Lefaix J Delanian S TGF-beta1 and radiation fibrosis: a master switch and a specific therapeutic target? Int J Radiat Oncol Biol Phys 2000 47 277 290 10802350 10.1016/S0360-3016(00)00435-1
Rubin P Johnston CJ Williams JP McDonald S Finkelstein JN A perpetual cascade of cytokines postirradiation leads to pulmonary fibrosis Int J Radiat Oncol Biol Phys 1995 33 99 109 7642437 10.1016/0360-3016(95)00095-G
Rubin P Finkelstein J Shapiro D Molecular biology mechanisms in the radiation induction of pulmonary injury syndromes: interrelationship between the alveolar macrophage and the septal fibroblast Int J Radiat Oncol Biol Phys 1992 24 93 101 1512168
Hill RP Rodemann HP Hendry JH Roberts SA Anscher MS Normal tissue radiobiology: from the laboratory to the clinic Int J Radiat Oncol Biol Phys 2001 49 353 365 11173128 10.1016/S0360-3016(00)01484-X
Crystal RG Oxidants and respiratory tract epithelial injury: pathogenesis and strategies for therapeutic intervention Am J Med 1991 91 39S 44S 1928210 10.1016/0002-9343(91)90282-3
Borok Z Buhl R Grimes GJ Bokser AD Hubbard RC Holroyd KJ Roum JH Czerski DB Cantin AM Crystal RG Effect of glutathione aerosol on oxidant-antioxidant imbalance in idiopathic pulmonary fibrosis Lancet 1991 338 215 216 1676780 10.1016/0140-6736(91)90350-X
Riley PA Free radicals in biology: oxidative stress and the effects of ionizing radiation Int J Radiat Biol 1994 65 27 33 7905906
Marklund SL Extracellular superoxide dismutase and other superoxide dismutase isoenzymes in tissues from nine mammalian species Biochem J 1984 222 649 655 6487268
Folz RJ Guan J Seldin MF Oury TD Enghild JJ Crapo JD Mouse extracellular superoxide dismutase: primary structure, tissue-specific gene expression, chromosomal localization, and lung in situ hybridization Am J Respir Cell Mol Biol 1997 17 393 403 9376114
Oury TD Crapo JD Valnickova Z Enghild JJ Human extracellular superoxide dismutase is a tetramer composed of two disulphide-linked dimers: a simplified, high-yield purification of extracellular superoxide dismutase Biochem J 1996 317 ( Pt 1) 51 57 8694786
Kang SK Rabbani ZN Folz RJ Golson ML Huang H Yu D Samulski TS Dewhirst MW Anscher MS Vujaskovic Z Overexpression of extracellular superoxide dismutase protects mice from radiation-induced lung injury Int J Radiat Oncol Biol Phys 2003 57 1056 1066 14575837 10.1016/S0360-3016(03)01369-5
Bowler RP Nicks M Warnick K Crapo JD Role of extracellular superoxide dismutase in bleomycin-induced pulmonary fibrosis Am J Physiol Lung Cell Mol Physiol 2002 282 L719 26 11880297
Folz RJ Abushamaa AM Suliman HB Extracellular superoxide dismutase in the airways of transgenic mice reduces inflammation and attenuates lung toxicity following hyperoxia J Clin Invest 1999 103 1055 1066 10194479
Arkovitz MS Garcia VF Szabo C McConnell K Bove K Wispe JR Decreased pulmonary compliance is an early indicator of pulmonary oxygen injury J Surg Res 1997 67 193 198 9073567 10.1006/jsre.1996.4980
Hsu SM Raine L Fanger H Use of avidin-biotin-peroxidase complex (ABC) in immunoperoxidase techniques: a comparison between ABC and unlabeled antibody (PAP) procedures J Histochem Cytochem 1981 29 577 580 6166661
Rubin P Casarett GW Clinical radiation pathology as applied to curative radiotherapy Cancer 1968 22 767 778 5212301
Travis EL Down JD Holmes SJ Hobson B Radiation pneumonitis and fibrosis in mouse lung assayed by respiratory frequency and histology Radiat Res 1980 84 133 143 7454976
Penney DP Rubin P Specific early fine structural changes in the lung irradiation Int J Radiat Oncol Biol Phys 1977 2 1123 1132 599063
Penney DP Siemann DW Rubin P Shapiro DL Finkelstein J Cooper RAJ Morphologic changes reflecting early and late effects of irradiation of the distal lung of the mouse: a review Scan Electron Microsc 1982 413 425 7167757
Steinberg F Quabeck K Rehn B Kraus R Mohnke M Costabel U Kreuzfelder E Molls M Bruch J Schaefer UW Lung effects after total body irradiation of mice and bone marrow transplant patients: comparison of experimental and preliminary clinical data Recent Results Cancer Res 1993 130 133 143 8362083
Ward PA Hunninghake GW Lung inflammation and fibrosis Am J Respir Crit Care Med 1998 157 S123 9 9563771
Hagiwara SI Ishii Y Kitamura S Aerosolized administration of N-acetylcysteine attenuates lung fibrosis induced by bleomycin in mice Am J Respir Crit Care Med 2000 162 225 231 10903246
MacNee W Rahman I Oxidants/antioxidants in idiopathic pulmonary fibrosis Thorax 1995 50 Suppl 1 S53 8 7570466
Johnston CJ Williams JP Okunieff P Finkelstein JN Radiation-induced pulmonary fibrosis: examination of chemokine and chemokine receptor families Radiat Res 2002 157 256 265 11839087
Suliman HB Ryan LK Bishop L Folz RJ Prevention of influenza-induced lung injury in mice overexpressing extracellular superoxide dismutase Am J Physiol Lung Cell Mol Physiol 2001 280 L69 78 11133496
Akaike T Noguchi Y Ijiri S Setoguchi K Suga M Zheng YM Dietzschold B Maeda H Pathogenesis of influenza virus-induced pneumonia: involvement of both nitric oxide and oxygen radicals Proc Natl Acad Sci U S A 1996 93 2448 2453 8637894 10.1073/pnas.93.6.2448
Oda T Akaike T Hamamoto T Suzuki F Hirano T Maeda H Oxygen radicals in influenza-induced pathogenesis and treatment with pyran polymer-conjugated SOD Science 1989 244 974 976 2543070
Carlsson LM Jonsson J Edlund T Marklund SL Mice lacking extracellular superoxide dismutase are more sensitive to hyperoxia Proc Natl Acad Sci U S A 1995 92 6264 6268 7603981
Fattman CL Chang LY Termin TA Petersen L Enghild JJ Oury TD Enhanced bleomycin-induced pulmonary damage in mice lacking extracellular superoxide dismutase Free Radic Biol Med 2003 35 763 771 14583340 10.1016/S0891-5849(03)00402-7
Vignaud JM Allam M Martinet N Pech M Plenat F Martinet Y Presence of platelet-derived growth factor in normal and fibrotic lung is specifically associated with interstitial macrophages, while both interstitial macrophages and alveolar epithelial cells express the c-sis proto-oncogene Am J Respir Cell Mol Biol 1991 5 531 538 1958380
Martinet Y Rom WN Grotendorst GR Martin GR Crystal RG Exaggerated spontaneous release of platelet-derived growth factor by alveolar macrophages from patients with idiopathic pulmonary fibrosis N Engl J Med 1987 317 202 209 3600711
Matsushima K Taguchi M Kovacs EJ Young HA Oppenheim JJ Intracellular localization of human monocyte associated interleukin 1 (IL 1) activity and release of biologically active IL 1 from monocytes by trypsin and plasmin J Immunol 1986 136 2883 2891 2420874
Beutler B Cerami A Cachectin: more than a tumor necrosis factor N Engl J Med 1987 316 379 385 3543677
Gospodarowicz D Neufeld G Schweigerer L Fibroblast growth factor Mol Cell Endocrinol 1986 46 187 204 2427371 10.1016/0303-7207(86)90001-8
Assoian RK Fleurdelys BE Stevenson HC Miller PJ Madtes DK Raines EW Ross R Sporn MB Expression and secretion of type beta transforming growth factor by activated human macrophages Proc Natl Acad Sci U S A 1987 84 6020 6024 2888109
Graham MM Evans ML Dahlen DD Mahler PA Rasey JS Pharmacological alteration of the lung vascular response to radiation Int J Radiat Oncol Biol Phys 1990 19 329 339 2168354
Roberts AB Russo A Felici A Flanders KC Smad3: a key player in pathogenetic mechanisms dependent on TGF-beta Ann N Y Acad Sci 2003 995 1 10 12814934
Wang Q Wang Y Hyde DM Gotwals PJ Koteliansky VE Ryan ST Giri SN Reduction of bleomycin induced lung fibrosis by transforming growth factor beta soluble receptor in hamsters Thorax 1999 54 805 812 10456973
Rabbani ZN Anscher MS Zhang X Chen L Samulski TV Li CY Vujaskovic Z Soluble TGFbeta type II receptor gene therapy ameliorates acute radiation-induced pulmonary injury in rats Int J Radiat Oncol Biol Phys 2003 57 563 572 12957270 10.1016/S0360-3016(03)00639-4
Nakao A Fujii M Matsumura R Kumano K Saito Y Miyazono K Iwamoto I Transient gene transfer and expression of Smad7 prevents bleomycin-induced lung fibrosis in mice J Clin Invest 1999 104 5 11 10393693
Flanders KC Sullivan CD Fujii M Sowers A Anzano MA Arabshahi A Major C Deng C Russo A Mitchell JB Roberts AB Mice lacking Smad3 are protected against cutaneous injury induced by ionizing radiation Am J Pathol 2002 160 1057 1068 11891202
Datto MB Frederick JP Pan L Borton AJ Zhuang Y Wang XF Targeted disruption of Smad3 reveals an essential role in transforming growth factor beta-mediated signal transduction Mol Cell Biol 1999 19 2495 2504 10082515
Yang X Letterio JJ Lechleider RJ Chen L Hayman R Gu H Roberts AB Deng C Targeted disruption of SMAD3 results in impaired mucosal immunity and diminished T cell responsiveness to TGF-beta Embo J 1999 18 1280 1291 10064594 10.1093/emboj/18.5.1280
Ashcroft GS Yang X Glick AB Weinstein M Letterio JL Mizel DE Anzano M Greenwell-Wild T Wahl SM Deng C Roberts AB Mice lacking Smad3 show accelerated wound healing and an impaired local inflammatory response Nat Cell Biol 1999 1 260 266 10559937 10.1038/12971
Flanders KC Major CD Arabshahi A Aburime EE Okada MH Fujii M Blalock TD Schultz GS Sowers A Anzano MA Mitchell JB Russo A Roberts AB Interference with transforming growth factor-beta/ Smad3 signaling results in accelerated healing of wounds in previously irradiated skin Am J Pathol 2003 163 2247 2257 14633599
| 15949035 | PMC1177930 | CC BY | 2021-01-04 16:03:04 | no | BMC Cancer. 2005 Jun 10; 5:59 | utf-8 | BMC Cancer | 2,005 | 10.1186/1471-2407-5-59 | oa_comm |
==== Front
BMC CancerBMC Cancer1471-2407BioMed Central London 1471-2407-5-721600017010.1186/1471-2407-5-72Research ArticleLung cancer symptoms and pulse oximetry in the prognostic assessment of patients with lung cancer Martins Sandro J [email protected] Nelson [email protected] Sueli O [email protected] Cecilia M [email protected] Crystina A [email protected] Teresa Y [email protected] Pulmonary Division, University of São Paulo Medical School, Av. Dr. Arnaldo 455, 01246-903 São Paulo, Brazil2 Psychology Section, Hospital das Clínicas, University of São Paulo Medical School, Av. Dr. Arnaldo 455, 01246-903 São Paulo, Brazil3 Oncology Unit, Hospital Santa Izabel, Pca Almeida Couto 500, 40050-410 Salvador, Brazil2005 6 7 2005 5 72 72 9 7 2004 6 7 2005 Copyright © 2005 Martins 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
Medical oncologists continue to use performance status as a proxy for quality of life (QOL) measures, as completion of QOL instruments is perceived as time consuming, may measure aspects of QOL not affected by cancer therapy, and interpretation may be unclear. The pulse oximeter is widely used in clinical practice to predict cardiopulmonary morbidity after lung resection in cancer patients, but little is known on its role outside the surgical setting. We evaluated whether the Lung Cancer Symptom Scale and pulse oximetry may contribute to the evaluation of lung cancer patients who received standard anticancer therapy.
Methods
We enrolled forty-one consecutive, newly diagnosed, patients with locally advanced or metastatic lung cancer in this study. We developed a survival model with the variables gender, age, histology, clinical stage, Karnofsky performance status, wasting, LCSS symptom scores, average symptom burden index, and pulse oximetry (SpO2).
Results
Patient and observer-rated scores were correlated, except for the fatigue subscale. The median SpO2 was 95% (range: 86 to 98), was unrelated to symptom scores, and was weakly correlated with observer cough scores. In a multivariate survival model, SpO2 > 90% and patient scores on the LCSS appetite and fatigue subscales were independent predictors of survival.
Conclusion
LCSS fatigue and appetite rating, and pulse oximetry should be studied further as prognostic factors in lung cancer patients.
==== Body
Background
The symptoms of lung cancer are a burden to patients, and a major detriment to their quality of life (QoL) and ability to function. Several different instruments have been developed to assess QoL of lung cancer patients, with recognized reliability and validity [1,2], yet their use is not widespread. Medical oncologists continue to use performance status as a proxy for QOL measures, as completion of QOL instruments is perceived as time consuming, may measure aspects of QOL not affected by cancer therapy, and interpretation may be unclear [3]. Whether formal QoL evaluation can improve patient care and disease outcome, perhaps through helping medical reasoning, by adding information to other well-known prognostic factors is a matter of active research.
The Lung Cancer Symptom Scale (LCSS) is a disease-and site-specific instrument developed in 1985 by Patricia Hollen and colleagues at the Memorial Sloan-Kettering Cancer Center, primarily to assess quality of life for individuals with lung cancer, both in its physical and functional aspects. It is focused on six clinically important lung cancer symptoms that are likely to affect patients' physical and functional status. Although it lacks detail in many QoL domains, such as social, spiritual, and psychological parameters, the LCSS has demonstrated high inter-rater reliability, feasibility, reliability, and content validity when evaluated against the Sickness Impact Profile (appetite), Profiles of Mood States (fatigue), American Thoracic Society Questionnaire (cough and dyspnea) and the McGill Pain Questionnaire, and their normative data in patients with non-small cell lung cancer (NSCLC) has been published [4-7].
Pulse oximetry is widely used to rapidly monitor arterial oxygen saturation (SpO2) [8]. It has many of the characteristics of an ideal monitoring technique: portability, non-invasiveness, ease of use (calibration is not required) and the capability for continuous on-line monitoring of SpO2. SpO2 measurement may be particularly useful in the care patients with advanced lung cancer, leading to early detection and appropriate management of hypoxemia, a condition likely to occur in this clinical setting [9].
In this study, we evaluate the prognostic value of baseline LCSS scores and pulse oximetry in a cohort of ambulatory patients, with advanced lung cancer, receiving standard medical therapy.
Method
Design
This was an observational cohort study. Consecutive outpatients with advanced or metastatic lung cancer who attended the pulmonology division of the Hospital das Clínicas of the University of São Paulo, Brazil, were invited to participate. All procedures were conducted in accordance with the ethical standards of the World Medical Association Declaration of Helsinki and Institutional board.
New patients admitted for lung cancer therapy were eligible if they presented with histological diagnosis of lung cancer, and locally advanced or metastatic disease extent. Exclusion criteria included past infection (last month) or fever (last week); use of supplementary oxygen; and potential causes for bad signal capture by pulse oximeter: black skin, nail abnormalities or pigmentation, anemia (hemoglobin < 10 g/dL), jaundice (bilirubin < 3 mg/dL), hypotension (systolic bp ≤ 90 mmHg), or tachycardia (> 100 bpm) were not included. All patients gave written signed informed consent before participating.
Lung Cancer Symptom Scale
The LCSS comprises two scales, one for patients and other for health professionals as observer. The patient scale consists of nine items: six subscales related to major lung cancer symptoms (appetite, cough, dyspnea, fatigue, hemoptysis, and pain), all using 100 mm visual analogue measurements, and three summation items (activity status, symptomatic distress, and overall QoL). The observer scale is a five-point scale that measures the intensity of those symptoms (100 = none, 75 = mild, 50 = moderate, 25 = marked, and 0 = severe). The average symptom burden is an ancillary index, obtained as the mean of the six symptom scales scored separately by patients and the observer. At study admission, two psychologists explained all procedures and oriented patients to focus on the past day information about their symptom perception. A medical oncologist recorded symptom scores using the observer scale. In this study we used a Portuguese version of the LCSS developed for Brazil, through a forward and backward translation process [10].
Pulse oximetry
The attending investigator at the initial medical examination recorded the mean of three consecutive finger pulse oximetry measures (Nonim Med Inc; Plymouth, USA), read at least 20 seconds apart with up to 2% variability, as the non-invasive estimate of SpO2.
Medical therapy
Chemotherapy, radiotherapy, and supportive measures were administered on individual basis, and consisted of non-investigational therapies only. The attending physician was responsible for indicating the proper therapeutic modality according to patients' individual characteristics (comorbidities, clinical status, disease sites, and personal preferences). All chemotherapy regimens used were platinum-based.
Data analysis
The following variables were selected for this analysis: gender, age (years), histology, clinical stage, Karnofsky performance status (KPS), weight loss (% usual body weight), symptom subscores (appetite, cough, dyspnea, fatigue, hemoptysis, and pain), average symptom burden index, and SpO2.
Due to non-normality constraints, bivariate association among symptom scores, Karnofsky performance status, and SpO2 were evaluated by nonparametric statistics (Spearman's rho). A Cox model was fit by step-wise procedure, as follows: a global test was performed to show the simple, unadjusted relationship between factors and survival, and seven possible explanatory variables were selected to model building due to its anticipated prognostic value (Wald chi square > 2.7); a stepwise model selection procedure was used and included explanatory variables that were significant at the 0.10 level; the likelihood ratio test was used to assess every improvement in model fitting. We used the SPSS package to perform statistical analysis (version 11.01, SPSS Inc, Chicago, IL).
Results
From Oct 2000 to Apr 2001, forty-one consecutive eligible new patients with lung cancer were studied. Baseline demographic and clinical data are presented in Table 1. Our sample was largely comprised of aged, male, advanced lung cancer patients. Most patients (83%) had NSCLC and presented with poor performance status and weight loss. All patients received initial chemotherapy: in NSCLC, mitomycin C, vinblastine and cisplatin (61%) or carboplatin (22%); in small cell lung cancer, cisplatin and etoposide (17%). Nine patients (22%) received concurrent cisplatin radiotherapy (50.4 Gy) consolidation after four chemotherapy cycles. Median number of LCSS symptoms was 5 (range: 1 to 6); fatigue (95%) was the major presenting symptom, followed by pain (80%), dyspnea (78%), cough (73%), loss of appetite (68%), and hemoptysis (39%). All patients complied with LCSS administration.
LCSS scores
To compare individual and health professional perceptions on symptom distress, the relationship between patient and observer LCSS scores are showed in Table 2. We expected significant correlation in all subscales since they measured the same distress dimension, but we found disagreement on fatigue evaluation.
Observer-derived average symptom burden index was not associated with age, histology, clinical stage, and KPS; except for hemoptysis (p = 0.11) and pain (p = 0.09), it was correlated with patient-derived subscales: appetite (rho = 0.33, p = 0.034), fatigue (rho = 0.34, p = 0.031), cough (rho = 0.54, p = 0.005), and dyspnea (rho = 0.46, p = 0.002). Patient-derived average symptom burden was associated with KPS (p = 0.001) and independent of age, histology, or clinical stage; it was also related to all but one obsever-derived symptom subscale (pain, p = 0.16): appetite (rho = 0.39, p = 0.012), fatigue (rho = 0.42, p = 0.006), cough (rho = 0.38, p = 0.014), dyspnea (rho = 0.52, p = 0.0001), and hemoptysis (rho = 0.39, p = 0.013).
Pulse oximetry and LCSS scores
Median SpO2 was 95% (range: 86 to 98); eight patients (19%) had SpO2 below 90% but were not deemed candidates for domiciliary oxygen therapy. As presented in Table 3, the SpO2 measurement was unrelated to patient-derived symptom scores and exhibited only a weak correlation with observer-derived cough scores.
Survival analysis
As of December 2003, thirty-eight patients (93%) had died and three were lost to follow up. The median survival was 42 weeks and 36% survived the first year. A multivariate survival model developed from seven possible explanatory variables is presented in Table 4. SpO2 > 90%, patient-derived scores on the appetite and fatigue subscales were found to be the only independent prognostic factors in this cohort. Clinical stage retained borderline prognostic value in the final model. Age (p = 0.92), gender (p = 0.79), and Karnofsky's performance status (p = 0.94) did not behave as independent prognostic factors. Global symptom scores failed to show relevant prognostic value on univariate tests and were not selected to the survival modeling.
Discussion
Despite significant advances in cancer medicine, most physicians still deal with patients suffering from relentless, incurable illnesses, and their therapeutic decisions are often made based on subjective judgments about patients' QoL. Performance status scales, such as the KPS introduced in 1949, are widely accepted tools to evaluate physical functioning of cancer patients. Several recent studies suggest that a broad QoL evaluation using current well-structured instruments can provide more detailed prognostic information. This has been demonstrated for the Functional Living Index-Cancer [11,12], the European Organization for Research and Treatment of Cancer QLC43 form [2,13], and the Functional Assessment of Cancer Therapy – Lung questionnaire [14,15]. As shown here, the LCSS and pulse oximetry may be regarded as valuable tools for prognostic assessment in advanced lung cancer. SpO2 was independent of most LCSS data, and the combination of SpO2 and LCSS offered independent prognostic information on overall survival.
A SpO2 below 90% and variation of 10 mm in patient scores on appetite and fatigue were associated with an increased hazard of death. Measurement of pulse oximetry and completion of the LCSS form were fast and were not burdensome for the health care staff. The pulse oximeter is widely used in clinical practice to predict cardiopulmonary morbidity after lung resection in cancer patients [16], but little is known on its role outside the surgical setting.
The presence of anorexia and fatigue at initial evaluation has been associated with shortened overall survival. Loss of appetite is manifested by 75% of patients with advanced lung cancer at some point of their disease course [7], although in cancer centers its assessment eventually lead to therapeutic interventions in only 60% of the time [17]. Besides, its true prognostic value may be hidden or minimized in some studies given the fact that weight loss is usually the variable of interest. Recently, Hoang et al. [18] analyzed data from 1,436 patients with stage IV or IIIB NSCLC included in two randomized phase III Eastern Cooperative Oncology Group trials and found that loss of appetite conferred a hazard ratio for death of 1.62.
Fatigue also is a common symptom in patients with cancer, and is associated with both physical and psychological distress particularly patients with lung [19]. Using validated specific scales (Cancer Fatigue Scale and Fatigue Numerical Scale), fatigue that interferes with any daily activities may be detected in up to 59% of these patients [20]. Its independent negative impact on survival was reported in a retrospective study of 1,154 patients with lung carcinoma [21].
There are shortcomings on the use of SpO2 as a proxy measure of an estimate of SaO2 in patients with lung cancer. The presence of high carboxyhemoglobin values in heavy smokers could disturb the SpO2 reading by the pulse oximeter [22]. Many patients with lung cancer have some degree of chronic obstructive pulmonary disease (COPD), a condition in which SpO2 may not accurately represent arterial oxygen saturation [23]. Furthermore, the small sample size, heterogeneity of therapeutic approaches, and absence of full data on pulmonary comorbidity or carboxyhemoglobin values limit generalization of our model. Our four-factors model was fit from seven variables, from data on 41 patients and 38 events, and thus could not incorporate further prognostic information. In order to achieve reliable estimates of the three major functions (survival, probability density, and hazard) and their standard errors at each time interval by Cox regression the minimum recommended sample size should allow five to ten events per factor in the modeling set of variables [24,25].
In contrast to the patient's perspective, observers do not grade symptom distress only but also take into account the clinical setting related to that problem and discrepancies may rise. In the study presenting the conceptual model of the LCSS [26], fatigue was found to be the greatest significant predictor of symptomatic distress in NSCLC patients throughout therapy. We noticed a lack of correlation between patient and observer fatigue subscales, with a tendency for physicians to underrate this item. Although this may be due to misclassification errors on the categorical differences within the fatigue item by the observer, a bias towards an underestimation of symptom intensity by health professionals could not be ruled out. Discrepancies between observer and patient rated assessments have been reported for other symptom-based instrument, the Edmonton Symptom Assessment System [27,28].
To our knowledge, this is the first report suggesting the existence of a role for pulse oximetry in the assessment of newly diagnosed lung cancer patients receiving palliative anticancer therapy. Several issues still need to be clarified, including the pulse oximetry's performance under influence of smoking status or the coexistence COPD in lung cancer patients. Conceivably, its negative impact on prognosis could rely on the relationship of SpO2 with continued cigarette smoking, a condition associated to therapy failure and poor survival [29], or low SpO2 would reflect severity of underlying COPD. Our results add to the evidence that evaluation of lung cancer symptoms improve assessment of prognosis in the palliative setting, and support additional larger studies to refine and validate these findings so that they can be used in daily practice.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
SJM participated in the design of the study and performed the statistical analysis. NH participated in the patient care and carried out the pulse oximetry measurements. CAY, SOC and CMH participated in the design of the study, and carried out QOL measurements and their analysis. TYT conceived of the study, and participated in its design and coordination. All authors read and approved the final manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
We thank Dr. Augusto Mota that kindly provided us thoughtful commentaries on our manuscript.
Figures and Tables
Table 1 Patient characteristics
Median age (range), in years 63 (34 to 80)
Gender
Male 32 (78%)
Female 9 (22%)
Karnofsky performance status
60 % 3 (7%)
70 % 22 (54%)
80% 11 (27%)
90% 5 (12%)
Clinical Stage
IIIA 7 (17%)
IIIB 12 (29%)
IV 22 (54%)
Histology
Adenocarcinoma 16 (39%)
Squamous cell carcinoma 9 (22%)
Poorly differentiated carcinoma 6 (15%)
Large cell carcinoma 3 (7%)
Small cell carcinoma 7 (17%)
Metastatic sites
Bone 8 (19%)
Addrenal 6 (15%)
Skin 4 (10%)
Lung 3 (7%)
Brain 2 (5%)
Liver 1 (2%)
Wasting 22 (54%)
Table 2 Correlation between patient and observer scores on symptom subscales
Patient scale (mm) * Observer scale (point) * Spearman's rho p-value†
Appetite 31 (19–43) 75 (50–100) -0.52 0.0001
Fatigue 56 (46–65) 50 (0–100) -0.21 N.S.
Cough 29 (19–39) 75 (25–100) -0.71 0.0001
Dyspnea 38 (27–49) 75 (0–100) -0.51 0.001
Pain 44 (32–57) 75 (25–100) -0.36 0.02
Hemoptysis 13 (0–22) 100 (50–100) -0.50 0.001
Global symptom score 35 (28–42) 69 (42–100) -0.57 0.0001
* Median values and range.
† N.S.: Not significant at the 0.05 level (two-tailed)
Table 3 Correlation between pulse oximetry and LCSS subscales
Patient scale Observer scale
Spearman's rho p-value * Spearman's rho p-value
Appetite 0.02 N.S -0.20 N.S.
Fatigue -0.23 N.S -0.19 N.S.
Cough -0.16 N.S -0.33 0.03
Dyspnea 0.07 N.S -0.26 N.S.
Pain -0.13 N.S 0.02 N.S.
Hemoptysis -0.85 N.S -0.20 N.S.
Global symptom score -0.23 N.S. -0.31 0.049
* N.S.: Not significant at the 0.05 level (two-tailed)
Table 4 Survival analysis by Cox regression
β S.E. OR CI 95% p-value
Pulse oximetry (SpO2 < 95%) 1.39 0.63 4.02 1.16–13.85 0.022
Clinical stage (III vs. IV) 0.78 0.43 2.18 0.93–5.06 0.070
Appetite (10 mm) 0.13 0.06 1.14 1.01–1.30 0.041
Fatigue (10 mm) 0.21 0.09 1.23 1.03–1.47 0.019
Overall model: Chi-square = 10.21, p = 0.037
Variables not in the equation: age, gender, and Karnofsky performance status
==== Refs
Gridelli C Perrone F Nelli F Ramponi S De Marinis F Quality of life in lung cancer patients Ann Oncol 2001 12 S21 25 11804379 10.1023/A:1013034729575
Dancey J Zee B Osoba D Whitehead M Lu F Kaizer L Latreille J Pater JL Quality of life scores: an independent prognostic variable in a general population of cancer patients receiving chemotherapy. The National Cancer Institute of Canada Clinical Trials Group Qual Life Res 1997 6 151 158 9161115 10.1023/A:1026442201191
Hollen PJ Gralla RJ Cox C Eberly SW Kris MG A dilemma in analysis: issues in the serial measurement of quality of life in patients with advanced lung cancer Lung Cancer 1997 18 119 136 9316004 10.1016/S0169-5002(97)00059-7
Hollen PJ Gralla RJ Kris MG Potanovich LM Quality of life assessment in individuals with lung cancer: testing the Lung Cancer Symptom Scale (LCSS) Eur J Cancer 1993 29A S51 58 8381294
Hollen PJ Gralla RJ Kris MG Cox C Belani CP Grunberg SM Crawford J Neidhart JA Measurement of quality of life in patients with lung cancer in multicenter trials of new therapies. Psychometric assessment of the Lung Cancer Symptom Scale Cancer 1994 73 2087 2098 8156514
Hollen PJ Gralla RJ Comparison of instruments for measuring quality of life in patients with lung cancer Semin Oncol 1996 23 31 40 8610235
Hollen PJ Gralla RJ Kris MG Eberly SW Cox C Normative data and trends in quality of life from the Lung Cancer Symptom Scale (LCSS) Support Care Cancer 1999 7 140 148 10335932 10.1007/s005200050244
McCarthy K Decker MJ Strohl KP Kacmarek RM, Hess D, Stoller JK Pulse Oximetry Monitoring in respiratory disease 1993 St. Louis: Mosby 309 347
Tremper KK Pulse oximetry Chest 1989 95 713 715 2647421
Hollen PJ Gralla RJ Kris MG Lung Cancer Symptom Scale (LCSS): Administration, scoring, and procedures manual 1995 Boston: Northeastern University
Ganz PA Lee JJ Siau J Quality of life assessment. An independent prognostic variable for survival in lung cancer Cancer 1991 67 3131 3135 1710541
Ruckdeschel JC Piantadosi S Quality of life in lung cancer surgical adjuvant trials Chest 1994 106 324S 328S 7988255
Herndon JE 2ndFleishman S Kornblith AB Kosty M Green MR Holland J Is quality of life predictive of the survival of patients with advanced nonsmall cell lung carcinoma? Cancer 1999 85 333 340 10023700 10.1002/(SICI)1097-0142(19990115)85:2<333::AID-CNCR10>3.0.CO;2-Q
Cella D Eton DT Fairclough DL Bonomi P Heyes AE Silberman C Wolf MK Johnson DH What is a clinically meaningful change on the Functional Assessment of Cancer Therapy-Lung (FACT-L) Questionnaire? Results from Eastern Cooperative Oncology Group (ECOG) Study 5592 J Clin Epidemiol 2002 55 285 295 11864800 10.1016/S0895-4356(01)00477-2
Moinpour CM Lyons B Grevstad PK Lovato LC Crowley J Czaplicki K Buckner ZM Ganz PA Kelly K Gandara DR Quality of life in advanced non-small-cell lung cancer: results of a Southwest Oncology Group randomized trial Qual Life Res 2002 11 115 126 12018735 10.1023/A:1015048908822
Dunn WF Scanlon PD Preoperative pulmonary function testing for patients with lung cancer Mayo Clin Proc 1993 68 371 377 8455398
Brown JK Radke KJ Nutritional assessment, intervention, and evaluation of weight loss in patients with non-small cell lung cancer Oncol Nurs Forum 1998 25 547 553 9568609
Hoang T Xu R Schiller JH Bonomi P Johnson DH Clinical model to predict survival in chemonaive patients with advanced non-small-cell lung cancer treated with third-generation chemotherapy regimens based on eastern cooperative oncology group data J Clin Oncol 2005 23 175 183 15625371 10.1200/JCO.2005.04.177
Cooley ME Short TH Moriarty HJ Symptom prevalence, distress, and change over time in adults receiving treatment for lung cancer Psychooncology 2003 12 694 708 14502594
Okuyama T Tanaka K Akechi T Kugaya A Okamura H Nishiwaki Y Hosaka T Uchitomi Y Fatigue in ambulatory patients with advanced lung cancer: prevalence, correlated factors, and screening J Pain Symptom Manage 2001 22 554 564 11516597 10.1016/S0885-3924(01)00305-0
Tammemagi CM Neslund-Dudas C Simoff M Kvale P Lung carcinoma symptoms – an independent predictor of survival and an important mediator of African-American disparity in survival Cancer 2004 101 1655 1663 15378496 10.1002/cncr.20547
Glass KL Dillard TA Phillips YY Torrington KG Thompson JC Pulse oximetry correction for smoking exposure Mil Med 1996 161 273 276 8855058
Carone M Patessio A Appendini L Purro A Czernicka E Zanaboni S Donner CF Comparison of invasive and noninvasive saturation monitoring in prescribing oxygen during exercise in COPD patients Eur Respir J 1997 10 446 451 9042647 10.1183/09031936.97.10020446
Klein JP Moeschberger ML Survival analysis: techniques for censored and truncated data 1997 New York: Springer-Verlag
Steinberg M Cox Regression SPSS Advanced Models 100 1999 Chicago: SPSS Inc 283 319
Hollen PJ Gralla RJ Kris MG Cox C Quality of life during clinical trials: conceptual model for the Lung Cancer Symptom Scale (LCSS) Support Care Cancer 1994 2 213 222 8087439 10.1007/BF00365725
Pautex S Berger A Chatelain C Herrmann F Zulian GB Symptom assessment in elderly cancer patients receiving palliative care Crit Rev Oncol Hematol 2003 47 281 286 12962902
Stromgren AS Groenvold M Pedersen L Olsen AK Spile M Sjogren P Does the medical record cover the symptoms experienced by cancer patients receiving palliative care? A comparison of the record and patient self-rating J Pain Symptom Manage 2001 21 189 196 11239737 10.1016/S0885-3924(01)00264-0
Tammemagi CM Neslund-Dudas C Simoff M Kvale P Smoking and lung cancer survival: the role of comorbidity and treatment Chest 2004 125 27 37 14718417 10.1378/chest.125.1.27
| 16000170 | PMC1177931 | CC BY | 2021-01-04 16:03:05 | no | BMC Cancer. 2005 Jul 6; 5:72 | utf-8 | BMC Cancer | 2,005 | 10.1186/1471-2407-5-72 | oa_comm |
==== Front
BMC CancerBMC Cancer1471-2407BioMed Central London 1471-2407-5-731600017210.1186/1471-2407-5-73Research ArticleFurther evidence for increased macrophage migration inhibitory factor expression in prostate cancer Meyer-Siegler Katherine L [email protected] Kenneth A [email protected] Pedro L [email protected] Research and Development Service (151), Bay Pines Veterans' Administration Medical Center, Bay Pines, FL 33744, USA2 Department of Surgery, University of South Florida, Tampa, FL 33612, USA3 Department of Pathology, Malcom Randall Veterans' Administration Medical Center, Gainesville, FL 32608, USA4 Department of Pathology, Immunology, and Laboratory Medicine, University of Florida College of Medicine, Gainesville, FL 32610, USA2005 6 7 2005 5 73 73 19 4 2005 6 7 2005 Copyright © 2005 Meyer-Siegler 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
Macrophage migration inhibitory factor (MIF) is a cytokine associated with prostate cancer, based on histologic evidence and circulating (serum) levels.
Recent studies from another laboratory failed to document these results. This study's aims were to extend and confirm our previous data, as well as to define possible mechanisms for the discrepant results. Additional aims were to examine MIF expression, as well as the location of MIF's receptor, CD74, in human prostatic adenocarcinoma compared to matched benign prostate.
Methods
MIF amounts were determined in random serum samples remaining following routine PSA screening by ELISA. Native, denaturing and reducing polyacrylamide gels and Western blot analyses determined the MIF form in serum. Prostate tissue arrays were processed for MIF in situ hybridization and immunohistochemistry for MIF and CD74. MIF released into culture medium from normal epithelial, LNCaP and PC-3 cells was detected by Western blot analysis.
Results
Median serum MIF amounts were significantly elevated in prostate cancer patients (5.87 ± 3.91 ng/ml; ± interquartile range; n = 115) compared with patients with no documented diagnosis of prostate cancer (2.19 ± 2.65 ng/ml; n = 158). ELISA diluent reagents that included bovine serum albumin (BSA) significantly reduced MIF serum detection (p < 0.01). MIF mRNA was localized to prostatic epithelium in all samples, but cancer showed statistically greater MIF expression. MIF and its receptor (CD74) were localized to prostatic epithelium. Increased secreted MIF was detected in culture medium from prostate cancer cell lines (LNCaP and PC-3).
Conclusion
Increased serum MIF was associated with prostate cancer. Diluent reagents that included BSA resulted in MIF serum immunoassay interference. In addition, significant amounts of complexed MIF (180 kDa under denaturing conditions by Western blot) found in the serum do not bind to the MIF capture antibody. Increased MIF mRNA expression was observed in prostatic adenocarcinoma compared to benign tissue from matched samples, supporting our earlier finding of increased MIF gene expression in prostate cancer.
==== Body
Background
Macrophage migration inhibitory factor (MIF) is a cytokine initially isolated based upon its ability to halt the in vitro random migration of macrophages [1,2]. Subsequently, it was defined as a proinflammatory cytokine with a key regulatory role in inflammation and both specific and nonspecific immunity [3]. Currently MIF is recognized as being constitutively expressed in various cell types, able to manifest itself as a cytokine, hormone or enzyme, and as a mediator of many pathologic conditions [4].
As a proinflammatory cytokine, MIF counter-regulates the effects of glucocorticoids and stimulates the secretion of other cytokines such as tumor necrosis factor (TNF)-α and interleukin (IL)-1β [3], thus assuming a role in the pathogenesis of inflammatory, immune diseases and cancer [5-7]. MIF is directly associated with the growth of lymphoma, melanoma, and colon cancer [8]. Indications of MIF's key role in tumor progression derive from studies where treatments with either anti-MIF immunoglobulin therapy and/or MIF antisense oligonucleotide confer anti-tumor activity [9,10]. Although MIF is associated with cancer angiogenesis, progression, and metastasis, the exact mechanism of this cytokine's action is uncertain since a receptor for MIF has only recently been identified as the cell surface form of the invariant chain (CD74) [11].
This laboratory was first in localizing MIF within prostatic epithelium and in establishing that MIF is consistently increased within both prostate tissue and serum in prostate cancer patients [7,12,13]. The expression of CD74 in the human prostate or its association with prostate cancer has not been investigated, but if MIF affects prostate cancer cell growth, then it is predicted that MIF's receptor should be localized within prostate tissue. Our recent paper established that patients with prostate cancer had elevated serum MIF levels [13]. However, a study from Michael et al. failed to confirm these findings [14]. This discrepancy is worth investigating since establishing the exact role of MIF in prostate cancer may prove useful as a diagnostic or prognostic indicator (in addition to the well-established standard, prostate specific antigen – PSA) for this important disease entity.
The objectives of this study are: (1) to confirm and extend our previous serum findings by analyzing serum samples remaining following routine serum PSA analysis from both prostate cancer and non-prostate cancer patients; (2) to identify potential confounds within the ELISA protocol that may account for the contradictory findings reported [14]; (3) to localize, quantify, and compare MIF mRNA and protein amounts in matched cancer and benign biopsy samples using commercially available tissue arrays; (4) to localize the MIF receptor, CD74, within prostate tissue and (5) to compare MIF secretion by normal prostate epithelial cells and prostate cancer cells in vitro.
Methods
Approval for all facets of this study was obtained from the Bay Pines Veterans' Administration Medical Center (VAMC) Institutional Review Board and the study protocols comply with the Helsinki Declaration. The serum samples were obtained with waiver of informed consent under section 46.116(d) of the Department of Health and Human Services human subject regulations at 45 CFR 46.
Comparison of MIF serum amounts in prostate cancer and non-cancer patients
Random serum samples remaining following routine PSA evaluation prior to being discarded were collected (n = 212) and stored at -80°C prior to analysis. Clinical PSA values for the serum samples (determined by the radioimmunoassay laboratory at the Bay Pines VAMC) and documentation of prostate cancer were obtained through analysis of computerized records. Normal patients were defined as those without documented prostate cancer and with serum PSA values at or below 2 ng/ml (n = 158).
Samples were assayed for MIF using the protocol described by the manufacturer (DuoSet ELISA, R&D Systems, Minneapolis, MN) with the following modifications: Milk diluent, (500-92-01, Kirkegaard and Perry Laboratories, Gaithersburg, MD) was used at a 1:20 dilution as the blocking agent and serum diluent. Our previously published data set of MIF serum levels in prostate CaP patients was also included for comparison (n = 61) [13]. Determined values for serum PSA and MIF were not normally distributed, therefore differences in serum PSA and MIF amounts were assessed by Kruskall-Wallis ANOVA followed by post-hoc tests (Dunn's multiple comparison test). Significance was defined as p < 0.05. Data are reported as median ± interquartile range. Statistical analysis was performed using Prism statistical analysis software version 4.02 (GraphPad Software, Inc., San Diego, CA).
Comparison of different ELISA protocols on MIF serum determination
Michael et al. used the MIF DuoSet ELISA system to determine serum MIF [14]. However, there were several deviations from the manufacturer's protocol and from our own protocol, principally the use of bovine serum albumin (BSA) as a blocking agent and serum diluent. Therefore, to test whether differences in ELISA protocols result in discrepant findings, in a separate experiment 10 random serum samples (n = 5 CaP; n = 5 non-CaP) were assayed in parallel using either milk diluent (our established protocol in agreement with the manufacturer's suggestions) or PBS with 1% BSA (as reported by Michael, et al. [14] as the blocking agent and serum diluent). All other steps in the protocol were as described by the manufacturer (R&D Systems). Comparisons between different ELISA protocols were evaluated using two-tailed unpaired nonparametric t-tests (Mann-Whitney). Data are reported as median ± interquartile range.
MIF forms in human serum
The discrepancy between these ELISA data could be attributed to blockage of the monoclonal capture antibody (mAb) epitope due to MIF interaction with other serum proteins. Changes in the dilution buffer composition may alter these interactions resulting in unmasking of capture antibody binding sites. To address this possibility, serum proteins (diluted 1:20 with PBS) were separated by native, denaturing and reducing polyacrylamide gel electrophoresis following the manufacturer's protocols (4–20% Tris-glycine or 4–12 % NuPAGE Bis-Tris gels, Invitrogen, Carlsbad, CA) and transferred to Invitrolon (Invitrogen). Reducing conditions included addition of anti-oxidant (Invitrogen) to running buffer and transfer buffer to prevent oxidation of reduced cysteine, methionine and tryptophan residues. Blots were incubated overnight with a biotinylated MIF specific, affinity purified, antibody (BAF289, R&D Systems) at 1:1000 dilution overnight at 4°C as described previously [15].
Serum MIF was immunoprecipitated using agarose-immobilized ELISA capture mAb (R&D Systems antibody; Profound co-immunoprecipitation system, Pierce, Rockford, IL) to characterize further mAb MIF binding forms. For immunoprecipitation studies, 2 ml of serum was first cleared of contaminating IgG by passage through a Protein G column (Pierce) and 1 ml PBS fractions collected. The eluted fraction with the highest protein content was batch incubated with 200 μl immobilized capture mAb (200 μg MIF mAb per 100 μl of packed resin) overnight with continuous rocking at 4°C. Following overnight incubation the agarose resin was transferred to a spin column, proteins that did not stick were collected, and the column washed with 5, 1 ml PBS washes. The bound proteins were eluted and proteins in all fractions separated under denaturing and reducing conditions on 4–12% NuPAGE Bis-Tris gels, and then transferred to Invitrolon membranes following the manufacturer's protocols (Invitrogen). In some instances, fractions with low total protein content were first concentrated 800 fold (Eppendorf vacuum concentrator, Westbury, NY).
In all instances, MIF Western blots were performed using a polyclonal biotinylated antibody at 1:1000 dilution followed by strepavidin-horseradish peroxidase conjugate used at a 1:500 dilution (R&D Systems). Antibody specific bands were detected using a chemiluminescent substrate (Pierce) and bands quantified as described previously [15]. Data are expressed as net intensity values normalized to the staining intensity of a recombinant MIF standard (7.5 ng).
Determination of MIF and CD74 in matched prostate cancer and benign samples
Tissue array slides were obtained from Ambion (Low density; Austin, TX) and used for both in situ hybridization (MIF only) and immunohistochemistry (IHC; MIF and CD74). The slides contained paired samples of prostatic adenocarcinoma and matched benign specimens from the same patient. The present investigation was restricted to comparison of expression in matched benign versus cancer samples. Samples that could not be matched (e.g. missing section on the slide) were not included in the analysis.
In situ hybridization
RNase inhibitor (Protect-RNA, Sigma, St. Louis, MO) was included in all aqueous solutions. Tissue sections were dewaxed and rehydrated through ethanol, then treated with Proteinase K for 30 min at 37°C and endogenous peroxidase quenched by 10 min incubation in 3% hydrogen peroxide. MIF biotinylated probes were prepared by reverse transcription, followed by amplification of a 254 base pair (bp) MIF fragment [13]. T7 RNA polymerase promoters were ligated to the PCR product (Ambion) and single stranded RNA oligonucleotide probes produced by in vitro T7 transcription (Promega, Madison, WI) with incorporation of biotinylated-dUTP. Hybridized probes were detected with avidin-peroxidase (ISH-B1, Sigma). Negative controls consisted of hybridization of tissue arrays with sense biotinylated probes and hybridization of test prostatic tissue sections without probe addition. A poly-d(T) biotinylated probe was used as a positive control in test prostatic tissue sections. Tissue arrays were not counterstained.
Immunohistochemistry
Paraffin sections (4 μm) of formalin-fixed tissue arrays were dewaxed and processed for antigen retrieval (citrate buffer, pH 6.0; microwave 600 W to boiling, 2 min rest, and repeat cycle 10 times for a total time of 20 min). Sections were treated with 3% H202 for 2 min to quench endogenous peroxidase, preincubated in normal donkey serum (3%; 30 min) and then exposed to MIF biotinylated polyclonal antibody (BAF289, R&D Systems; 1:1000 in PBS, 0.3% Triton X-100) or CD74 antibody (sc-5438, Santa Cruz Biotechnology, Santa Cruz, CA; 1:400 in PBS, 0.3% Triton X-100) overnight at 4°C.
Previous experiments demonstrated the specificity of MIF labeling of prostatic tissues [12]. Specificity of CD74 immunostaining was confirmed on a separate set of prostatic tissue slides by omitting the primary antiserum (data not shown). Bound primary antibody was detected using biotinylated secondary antibody (Vector Laboratories, Burlingame, CA) and diamidinobenzidine (DAB) according to the manufacturer's protocol. The slides were counterstained with hematoxylin, dehydrated, cleared in xylene and coverslipped.
All tissue sections were scored for the location and intensity of staining (0 = no staining detected through 3 = very strong staining). MIF mRNA and MIF/CD74 immunostaining was examined in matched tissue samples. Differences were assessed in epithelial MIF staining intensity in matched versus adenocarcinoma samples using paired, two-tailed t-tests.
Determination of MIF expression in normal and prostate cancer epithelial cells in vitro
Normal prostate epithelial cells (PrEC, Cambrex, Walkersville, MD) and prostate cancer cell lines (LNCaP and PC-3; American Type Culture Collection, Manassas, VA) were cultured until 80% confluent according to previously described protocols [16]. Media were exchanged for serum free PrEGM (Cambrex) and following 48 h incubation, samples of culture medium containing equal total protein concentrations were analyzed by denaturing Western blot as described above. The resultant protein bands from three separate Western blots from each cell line were quantified by digital imaging (Kodak, Rochester, NY). Data are expressed as fold change compared to normal cells and differences assessed by unpaired, two-tailed t-tests.
Results
Comparison of MIF serum amounts in prostate cancer and non-cancer patients
We determined MIF serum levels in normal (non-CaP) and prostate cancer (CaP) patients using milk based buffer as the ELISA blocking agent and sample diluent. Using this protocol, the median MIF serum level in CaP patients (Fig. 1, CaP 1; 4.70 ± 4.32 ng/ml; ± interquartile range) was significantly greater than that observed in non-CaP patients (2.19 ± 2.65 ng/ml; Fig 1; <0.001). Thus, we confirm our previous observation of increased MIF serum levels in CaP patients [13].
Furthermore, we compared our previously analyzed cancer patients with our current group to detect any possible differences. Our previous CaP group had a median MIF serum level of 6.50 ± 3.10 ng/ml (Fig. 1, CaP 2) that was not statistically different from our current CaP group (Fig. 1, CaP 1; p = 0.495) although it was still significantly elevated when compared to the control group (Fig 1; p < 0.001).
Since no significant differences were observed between the two cancer groups, we have pooled the serum MIF levels to result in a combined prostate CaP group. The controls from our first study were from serum samples used for thyroid hormone analysis (n = 14) and were not combined with this normal group since the corresponding PSA values are unknown [13]. Thus, our data show that CaP patients have a significantly higher MIF serum level (5.87 ± 3.91 ng/ml, n = 115) when compared to non-cancer patients (2.19 ± 2.65 ng/ml, n = 158) and this difference is statistically significant (p < 0.0001). All serum samples were obtained from patients at least 15 months post-CaP diagnosis and median PSA serum amounts were significantly increased in the CaP (1.20 ± 4.60 ng/ml, ± interquartile range, n = 115) patients (Fig. 1B, p < 0.001) suggesting that this group of patients had recurrent or metastatic prostate cancer.
Comparison of different protocols on MIF serum determination
We noted several differences between the MIF ELISA protocol of Michael et al. [14] and our serum MIF ELISA protocol. Since MIF's interactions with albumin subfractions have been described [17], we tested the possibility that use of BSA in the MIF ELISA protocol would interfere in the MIF ELISA assay. Thus, we compared a subset of serum samples from prostate cancer patients (n = 5) and from patients with no documented prostate cancer (n = 5), which were assayed in parallel using either milk based buffer (our protocol) or PBS with 1% BSA (Michael et al. [14] protocol) as the serum diluent and blocking agent. Serum MIF levels were significantly increased, even in this small sample, when we used the milk-diluent based MIF ELISA (our protocol), with non-CaP having 4.18 ± 0.99 ng/ml compared to prostate cancer patients that had 13.06 ± 6.15 ng/ml of MIF in the serum (Fig. 2). However, the same samples, when run using the BSA diluent (as reported by Michael et al. [14]) showed no significant difference between the two groups (Fig. 2), in agreement with their findings. Moreover, when the same set of normal sera was compared using milk-diluent versus BSA diluent, it was observed that the BSA protocol resulted in a statistically significant lower MIF serum level (1.84 ± 0.73 for BSA compared to 4.18 ± 0.99 for milk). These data demonstrate that addition of BSA to the MIF serum ELISA protocol results in not only interference with the ability to detect an increase in MIF serum levels, but it also decreases the ability of the assay to detect MIF (as evidenced by the significantly lower MIF serum levels observed in patients without prostate cancer with the BSA protocol).
MIF forms in human serum
The success of antigen capture ELISA reactions is dependent upon accessibility of the epitope. Detecting antigen bound to endogenous proteins in serum is often difficult because of epitope masking or sensitivity problems. MIF's association with other proteins (either as carrier proteins or activators) has not received much attention although it may shed light on the mechanism of action of this complex molecule. While many studies have detected serum MIF, to date there is no information on the MIF form(s) in human serum. MIF Western blots under native conditions showed only a large high molecular weight band in the range of 150 to 500 kDa. (Fig. 3A). No monomeric, dimeric or trimeric MIF was observed. Under denaturing conditions, a single MIF immunoreactive band was detected (180 kDa). Again no monomeric, dimeric or trimeric MIF was observed. Under reducing conditions only a single 12 kDa MIF band was detected, suggesting that this MIF-protein complex is either held by covalent bonds or that the reducing conditions alters the complex conformation allowing the MIF to be released. The same banding pattern was observed using a well characterized MIF monoclonal antibody (III.D.9), suggesting that this banding pattern was not the result of antibody cross reactivity or non-specific binding (data not shown).
The ability of the ELISA mAb to detect all the MIF in serum samples was analyzed by affinity chromatography with the ELISA capture antibody followed by Western blotting (Fig. 3C). Following removal of immunoglobulin by Protein G chromatography, serum MIF was found in a 180 kDa complex (Fig. 3C, Denatured, NS Protein G). The MIF capture monoclonal antibody (Fig. 3C, Denatured, NS MIF mAb) did not recognize this 180 kDa MIF form. Affinity captured MIF was not detectable by Western blot without concentration of the eluted fraction (Fig. 3C, Denatured, Elution, Elution conc). The concentrated (800 fold) mAb captured MIF was separated into four bands of less than 40 kDa (12, 24, 30 and 36) with the 24 kDa band being the predominant form detected. The serum concentration of the180 kDa MIF form was 220 ng/ml, while the concentration of the 24 kDa band was 1.5 ng/ml as determined by normalization of band intensities to the MIF recombinant protein. Therefore, the MIF ELISA appears to detect less than 1% of the total serum MIF detected by Western blot.
The 180 kDa MIF complex was disrupted by reducing agents as only the 12 kDa monomeric MIF form was detected with addition of dithiolthreitol to the sample (Fig. 3C, Reduced). Under reducing conditions the MIF captured by the monoclonal antibody could not be detected without concentration (Fig. 3C, Reduced, Elution, Elution conc).
Determination of MIF and CD74 in matched prostate cancer and benign samples
Only matched samples of benign and adenocarcinoma tissue were included in the final analyses (MIF in situ, n = 20; MIF IHC analysis, n = 20; CD74 IHC, n = 18). The patient ages ranged from 59 to 94 years, with a median of 72.5 years. All tumors were TNM classification T1a or T1b.
MIF mRNA in the prostate
In situ hybridization of 20 matched samples of benign and cancerous prostate tissue revealed minimal MIF mRNA in the epithelium of benign samples, which was restricted to the cytoplasm (Fig. 4A). Increased MIF mRNA was detected in the matched cancer samples (Fig. 4B), which was localized to the cytoplasm with some evidence of nuclear or perinuclear localization of the signal. The mean in situ hybridization intensity of prostate cancer was 1.5, significantly higher than in matched benign tissues whose mean score was 0.5 (Fig. 5A; p < 0.001).
MIF protein location and intensity in the prostate
MIF immunostaining was readily observed in both cancerous and benign prostate in the 21 matched samples (Fig. 4C and 4D). MIF protein was localized predominantly within the epithelium in both benign prostate and cancer (Figs. 4C and 4D). The mean staining intensity within the epithelium in benign samples was 1.4, while that of the epithelium in adenocarcinoma was 1.5 (Fig. 5B). Stromal staining was evident in both benign and adenocarcinoma matched samples with no significant difference in the mean staining score (data not shown).
CD74 location and intensity in the prostate
Weak CD74 immunostaining was detected in both benign and prostate cancer tissues (Fig. 4E and 4F). The staining was patchy and localized to epithelial cells. There was no significant difference noted in staining intensity when comparing matched benign (mean intensity score = 0.6, Fig. 5C) and prostate cancer specimens (mean intensity score = 0.4).
Determination of MIF expression in normal and prostate cancer epithelial cells in vitro
Western blotting of cell culture medium from prostate epithelial cells showed that all cell lines secreted only 12 kDa MIF into the PrEGM serum-free culture media (Fig. 6A). However, prostate cancer cells (LNCaP and PC-3) showed significantly greater MIF secretion when compared to the normal cell line (Fig. 6A). Analysis of the resulting band intensities documents an 8 to 12 fold increase in MIF amounts secreted by prostate cancer cell lines compared with normal prostate cells (Fig. 6B).
Discussion
Recently attention has been focused on the possibility that epithelial cell injury as a result of chronic inflammation plays an important role in prostate carcinogenesis [18]. Thus, we proposed that MIF, a proinflammatory cytokine constitutively expressed by the prostatic epithelium and upregulated in prostate cancer, plays an important role in this disease entity [7]. Our present findings confirm our hypothesis since: 1) increased MIF serum levels were observed in an additional group of prostate cancer patients when compared to patients with no documented prostate cancer; 2) MIF mRNA was increased in prostate cancer tissue suggesting increased MIF synthesis in prostate cancer; 3) CD74, recently described as a receptor for MIF[11], is localized to prostatic epithelia; 4) prostatic epithelial cells in vitro secrete MIF, with much greater secretion observed in prostate cancer cells. Moreover, increased MIF expression in prostate cancer has been independently documented by other laboratories, further supporting our hypothesis [19,20]. The possible utility of MIF as a prognostic marker for CaP is suggested by the finding that serum MIF amounts are statistically increased in patients with CaP as long as 15 months post diagnosis.
Our current findings also illustrate that different protocols, or even seemingly small variations of the same protocol (such as choice of diluent), can have large effects on the ability to detect serum MIF. When we compared the same group of patients using our milk-diluent protocol (as recommended by the manufacturer) or a BSA-diluent protocol (as reported by Michael et al. [14]), we noticed that we were not able to detect differences in MIF serum levels of prostate cancer patients using the BSA-diluent protocol. This observation agrees with the findings of Michael et al. [14]. Furthermore, addition of BSA appears to diminish the ability of the ELISA protocol to detect MIF, since even the values in the same normal group were markedly smaller in the BSA-diluent protocol (reduced by 66%) when compared to our milk-diluent protocol. Thus, it is likely that the difference in the choice of diluent for the ELISA protocol accounts in part for the discrepant findings of Michael et al. [14], and their inability to detect significantly elevated MIF serum levels in their prostate cancer group. Additionally, the post-diagnosis duration was much later for our cancer group, so it is also possible that there may also be intrinsic group differences. However, the rather large discrepancy observed in a parallel test of both protocols (Fig. 2) suggests that differences in the ELISA procedure alone account for a large part of their negative findings. Particularly since MIF is documented to bind to subfractions of human albumin [17] and that albumin is a known interfering protein in serum ELISA [21], which appears to interfere with the determination of serum MIF amounts. In the aforementioned study, BSA at concentrations as high as 2.5 mg/ml was not shown to bind to immobilized recombinant MIF [17]. However, the human serum MIF ELISA as performed by both Michael et al. [14] and as reported here (Fig. 2) utilizes 4-fold higher concentrations of BSA and the MIF-BSA binding in the ELISA would occur in solution. BSA interference is a likely explanation for the discrepant findings as there were no other major differences in the serum diluent/blocking agent used including pH and ionic strength.
Interference in immunoassay is a well-documented phenomenon, especially prevalent when the analyte is detected in serum [21]. Matrix effects (defined as the sum of the effects of all of the components in a system with the exception of the analyte to be measured and including the reagents used in the assay) can increase or decrease the measured result [21]. In fact, cross-platform fluctuations in serum PSA have also been documented [22]. These interassay variations were attributed to operator variability, technical errors and/or inadequate estimation of this heterogeneous molecule and its complexes with other proteins [22,23]. All of these factors probably also affect determination of MIF serum levels and thus, strict adherence to established protocols aimed at standardization should allow for comparison of results between different laboratories.
The physiologically relevant form of MIF serum remains to be determined. However, in the serum it does not appear to be predominantly free uncomplexed MIF (monomeric, dimeric or trimeric) since these MIF forms were not detected by WB under native or denatured conditions (Fig 3A and 3B). The monomeric MIF form (12 kDa) and a small amount of the dimeric form (24 kDa) were detected in serum under reducing conditions (Fig. 3B). Crystallized recombinant bioactive MIF assumes a trimeric structure [24]. However, mixtures of all three forms of recombinant MIF have been found in solution [25]. The active in vivo produced MIF form(s) have not been determined. However, it was recently reported that the physiologically active form of MIF, purified from a human B-cell line was primarily dimeric (24 kDa) [26]. Based upon the results reported here the monomer, dimeric and trimeric MIF forms would account for less than 1% of the total MIF present in the serum as detected by Western blotting. In addition, our data suggest that ELISA assay conditions can alter the amount of MIF captured by the monoclonal antibody.
Interestingly, the present study also documents that majority of MIF in the serum is associated in a high molecular weight complex (180 kDa under denaturing conditions) that cannot be captured by the monoclonal antibody. We recently documented that rat MIF released by the urothelium into the bladder lumen is also part of a high molecular weight complex [27]. The rat MIF binding protein was identified as the acute phase protein α-1-inhibitor 3 [27]. α-1-inhibitor 3 is a rodent member of the α-2-macroglobulin protease inhibitor family that is found in high concentration in serum [28]. In general, the α-2-macroglobulin protease inhibitor family is recognized as carriers of a numerous cytokines and growth factors [29] and it is believed that interaction of cytokines and growth factors with α-2-macroglobulin family proteins modulates cytokine and growth factor biological activity [30]. Thus, identification of the serum complex protein may provide information about physiologically relevant MIF forms. The identities of the protein(s) in the human serum complex are presently under investigation.
The present study is the first report to document MIF mRNA expression in prostate epithelial cells and to reveal altered MIF expression in cancer using the complementary techniques of in situ hybridization and immunostaining within matched samples from the same patient. Low levels of MIF mRNA were restricted to the cytoplasm of prostatic epithelium in benign samples, but the MIF mRNA signal was significantly more intense in the cytoplasm of prostatic epithelial cells in matched cancer samples, suggesting that increased cytoplasmic MIF mRNA is associated with tumorigenesis. Data from our previous study (using a much smaller data set) gave some indication that increased MIF mRNA was associated with more invasive, aggressive tumor [13]. In that study, using laser capture microscopy, we reported that invasive epithelial cells exhibited higher MIF mRNA amounts than benign epithelium within the same prostatic biopsy specimen. In the present study, we demonstrate a significant increase in MIF mRNA associated with prostatic tumors compared with matched benign tissue from the same patients, documenting that upregulation of MIF gene expression is a hallmark of prostatic tumorigenesis (Fig. 1A).
This report demonstrates that MIF protein is localized to both the malignant and matched histologically benign epithelia in prostate cancer patients. This observation confirms our earlier findings [12], as well as those of other investigators [31,32]. del Vecchio et. al determined a decrease in MIF immunostaining in high grade tumors [31]. These authors suggested that this finding may be the consequence of a reduced MIF synthesis or of an enhanced and altered secretion by tumor cells into the surrounding stroma [31]. However, these studies localized and quantified MIF protein only. The alternative explanation for reduced MIF immunostaining is supported by the data presented here, namely, there is increased MIF secretion by aggressive prostate cancer cells. This hypothesis is based upon two separate sets of findings: (1) increased mRNA expression in prostate biopsy samples without a concomitant increase in cytoplasmic protein in prostate cancer epithelium suggests that tumor cells may secrete more MIF and (2) documented increased MIF secretion by prostate cancer cell lines compared with normal prostate epithelial cells in culture. We recognize that there is no correlation between MIF immunohistochemistry and serum levels. In fact, this anomaly is described for serum PSA versus PSA immunohistochemistry [33]. Based upon these findings it is tempting to speculate that the increased MIF secretion by prostate cancer cells is responsible for the increased serum MIF levels observed in prostate cancer patients. Prostate cancer disrupts acinar structure and function resulting in "leakage" of proteins normally released into the acinar ducts (e.g. PSA), into the stroma and then absorption into the circulation [34]. It is possible that a similar mechanism accounts for the increased MIF secretion by prostate cancer epithelial cells and increased MIF serum levels in cancer patients.
CD74 was identified as the receptor for MIF using in vitro models [11]. The primary function of this protein is to regulate peptide loading of exogenously derived peptides onto major histocompatibility class II heterodimers, but a small portion of the total cell CD74 content is expressed on the cell surface [35]. However, neither the exact function of the cell surface CD74, nor the physiological relevance of the interaction of CD74 with MIF is known [11,36]. We recently documented that MIF associates with CD74 and CD44 in an urothelial cancer cell line [10] and in an in vivo model of bladder inflammation [15] suggesting that the effects of MIF on the bladder are mediated by interaction with both CD74 and CD44. If MIF-CD74 interaction is needed for MIF to affect prostate tumor growth then CD74 should be localized within the prostate. To our knowledge, ours is the first report to show that CD74 is present in human prostate tissue. However, these data do not show any significant change in CD74 amount in tumor compared to matched benign prostate. Thus, the physiological relevance of CD74 in prostate cancer remains to be determined.
Conclusion
In summary, we confirmed and extended our observations of increased serum MIF and MIF mRNA are associated with CaP. Serum MIF is predominantly part of a complex with other proteins that undetected by ELISA using a commercially available MIF mAb. MIF mRNA was observed in the cytoplasm of all prostatic epithelia, but was increased in matched cancer samples. In addition, this study documents the location of CD74, the MIF receptor, within prostate epithelial cells. Whether increased serum MIF has predictive value that is independent of known predictive markers such as rising PSA values or Gleason score in prostate cancer awaits a larger trial that compares these variables prospectively with PSA levels and patient outcomes.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
KLMS participated in the design of the study, ELISA, Western blotting, immunoprecipitations, statistical analyses, performed in situ hybridizations and carried out all cell culture studies. KAL photographed all tissue slides, scored all in situ hybridizations and immunohistochemical slides. PLV participated in the design of the study, ELISA, Western blotting, immunohistochemical analysis and statistical analyses. All authors read and approved the final manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
Supported by the VA Merit Review Program and the Bay Pines Foundation (KLM-S, PLV) and VA MREP (KAI). Michael Bellino, Rachel Lyle and Gary A. Smith, Jr. provided technical assistance.
Figures and Tables
Figure 1 Scatter plots of PSA and MIF in random serum samples obtained following routine PSA testing. Median values for each group are represented by horizontal lines. Diagnosis of prostate cancer was determined by chart review. *** = p < 0.001 A) Serum MIF – All data points were mean values of duplicate ELISA analyses. B) Serum PSA – All data were obtained from computerized records of clinical laboratory analysis.
Figure 2 Comparison of Serum MIF amounts determined by different ELISA conditions. A subset of random serum samples were from pulled from our serum bank from patients with no documented diagnosis of prostate cancer (n = 5) and from patients with CaP (n = 5). The serum MIF amounts were determined in parallel using either a milk based serum blocking/diluent solution or phosphate buffered saline pH 7.4 with 1% bovine serum albumin. Data are expressed as median ± interquartile range, ** = p < 0.01, when compared to Milk normal serum samples.
Figure 3 MIF Western blot of serum proteins. A. Native serum MIF – Serum was diluted 1:20 in PBS and proteins were separated under native conditions then analyzed for MIF by Western blot analysis. Under native conditions, a broad band between 150 – 500 kDa was detected. B. Denatured and reduced serum MIF – Serum was diluted as in A and proteins separated under denaturing or reducing conditions. Denaturing conditions resulted in a single prominent band at 180 kDa and some minor bands between 80 and 120 kDa. Under reducing conditions, a single prominent 12-kDa band and a minor band at 24 kDa were observed. C. mAb Affinity purified serum MIF – MIF containing serum proteins were purified by sequential affinity chromatography through Protein G agarose, followed by MIF mAb affinity chromatography. Nonsticking proteins as well as elution fractions were separated by NuPAGE gel electrophoresis under denaturing or reducing conditions, transferred to PVDF and MIF containing bands identified using a polyclonal antibody (R&D Systems, 1:1000 dilution). Lanes left to right: NS Protein G – non-sticking fraction from Protein G column; NS MIF mAb – non-sticking fraction from MIF monoclonal antibody column; Elution – third elution fraction; Elution conc – third elution fraction concentrated 800 fold; MIF standard – 7.5 ng of recombinant MIF protein. Numbers and lines to the left of each Western blot indicate location of molecular weight markers.
Figure 4 Matched prostatic tissue staining. All panels are 400 × total magnification. A) In situ hybridization – MIF in benign tissue B) In situ hybridization – MIF in matched cancer tissue C) Immunohistochemistry – MIF in benign tissue D) Immunohistochemistry – MIF in matched cancer tissue E) Immunohistochemistry – CD74 in benign tissue F) Immunohistochemistry – CD74 in matched cancer tissue
Figure 5 MIF in situ hybridization and IHC staining intensity in prostatic tissue arrays. Mean values for each group are represented by horizontal lines. Values are expressed as mean ± standard error. A) MIF in situ – intensity of epithelial cell staining of matched benign and tumor regions of prostate biopsies from prostate cancer patients (n = 20). B) MIF IHC – intensity of epithelial cell staining of matched benign and tumor regions of prostate biopsies from prostate cancer patients (n = 20). C) CD74 IHC – intensity of epithelial cell staining of matched benign and tumor regions of prostate biopsies from prostate cancer patients (n = 18). (*** = p < 0.001).
Figure 6 MIF secreted by in vitro prostate cells. A) Western blot of 48 h culture medium from 80% confluent normal, LNCaP and PC-3 prostate cells. B) MIF bands were quantified by digital imaging and are expressed as fold increase compared to normal. Data are presented as the mean ± standard error of three separate Western blots. (* = p < 0.05).
==== Refs
Bloom BR Bennett B Mechanism of a reaction in vitro associated with delayed-type hypersensitivity Science 1966 153 80 82 5938421
David JR Delayed hypersensitivity in vitro : its mediation by cell-free substances formed by lymphoid cell-antigen interaction Proc Natl Acad Sci U S A 1966 56 72 77 5229858
Baugh JA Bucala R Macrophage migration inhibitory factor Crit Care Med 2002 30 S27 35 11782558 10.1097/00003246-200201001-00004
Metz CN Bucala R Role of macrophage migration inhibitory factor in the regulation of the immune response MIF Adv Immunol 1997 66 197 123 9328642
Kamimura A Kamachi M Nishihira J Ogura S Isobe H Dosaka-Akita H Intracellular distribution of macrophage migration inhibitory factor predicts the prognosis of patients with adenocarcinoma of the lung Cancer 2000 89 334 341 10918163 10.1002/1097-0142(20000715)89:2<334::AID-CNCR18>3.0.CO;2-N
Bini L Magi B Marzocchi B Arcuri F Tripodi S Cintorino M Protein expression profiles in human breast ductal carcinoma and histologically normal tissue Electrophoresis 1997 18 2832 2841 9504817 10.1002/elps.1150181519
Meyer-Siegler K Hudson PB Enhanced expression of macrophage migration inhibitory factor in prostatic adenocarcinoma metastases Urology 1996 48 448 452 8804500 10.1016/S0090-4295(96)00207-5
Nishihira J Macrophage migration inhibitory factor (MIF): its essential role in the immune system and cell growth J Interferon Cytokine Res 2000 20 751 762 11032394 10.1089/10799900050151012
Chesney J Metz C Bacher M Peng T Meinhardt A Bucala R An essential role for macrophage migration inhibitory factor (MIF) in angiogenesis and the growth of a murine lymphoma Mol Med 1999 5 181 191 10404515
Meyer-Siegler KL Erica Leifheit C Vera PL Inhibition of Macrophage Migration Inhibitory Factor Decreases Proliferation and Cytokine Expression in Bladder Cancer Cells BMC Cancer 2004 4 34 15248897 10.1186/1471-2407-4-34
Leng L Metz CN Fang Y Xu J Donnelly S Baugh J MIF signal transduction initiated by binding to CD74 J Exp Med 2003 197 1467 1476 12782713 10.1084/jem.20030286
Meyer-Siegler K Fattor RA Hudson PB Expression of macrophage migration inhibitory factor in the human prostate Diagn Mol Pathol 1998 7 44 50 9646034 10.1097/00019606-199802000-00008
Meyer-Siegler KL Bellino MA Tannenbaum M Macrophage migration inhibitory factor evaluation compared with prostate specific antigen as a biomarker in patients with prostate carcinoma Cancer 2002 94 1449 1456 11920501 10.1002/cncr.10354
Michael A Stephan C Kristiansen G Burckhardt M Loening SA Schnorr D Diagnostic validity of macrophage migration inhibitory factor in serum of patients with prostate cancer: A re-evaluation Prostate 2005 62 34 39 15389818 10.1002/pros.20104
Meyer-Siegler KL Vera PL Substance P induced changes in CD74 and CD44 in the rat bladder J Urol 2005 173 615 620 15643275 10.1097/01.ju.0000143188.02802.f3
Meyer-Siegler KL Increased stability of macrophage migration inhibitory factor (MIF) in DU-145 prostate cancer cells J Interferon Cytokine Res 2000 20 769 778 11032396 10.1089/10799900050151030
Zeng FY Kratzin H Gabius HJ Migration inhibitory factor-binding sarcolectin from human placenta is indistinguishable from a subfraction of human serum albumin Biol Chem Hoppe Seyler 1994 375 393 399 7980871
De Marzo AM DeWeese TL Platz EA Meeker AK Nakayama M Epstein JI Pathological and molecular mechanisms of prostate carcinogenesis: implications for diagnosis, detection, prevention, and treatment J Cell Biochem 2004 91 459 477 14755677 10.1002/jcb.10747
Karan D Kelly DL Rizzino A Lin MF Batra SK Expression profile of differentially-regulated genes during progression of androgen-independent growth in human prostate cancer cells Carcinogenesis 2002 23 967 975 12082018 10.1093/carcin/23.6.967
Lin JC Chang SY Hsieh DS Lee CF Yu DS The association of Id-1, MIF and GSTpi with acquired drug resistance in hormone independent prostate cancer cells Oncol Rep 2005 13 983 988 15809768
Selby C Interference in immunoassay Ann Clin Biochem 1999 36 704 721 10586307
Link RE Shariat SF Nguyen CV Farr A Weinberg AD Morton RA Variation in prostate specific antigen results from 2 different assay platforms: clinical impact on 2304 patients undergoing prostate cancer screening J Urol 2004 171 2234 2238 15126793 10.1097/01.ju.0000127736.86597.e7
Kioukia-Fougia N Christofidis I Strantzalis N Physicochemical conditions affecting the formation/stability of serum complexes and the determination of prostate-specific antigen (PSA) Anticancer Res 1999 19 3315 3320 10652627
Sun HW Bernhagen J Bucala R Lolis E Crystal structure at 2.6-A resolution of human macrophage migration inhibitory factor Proc Natl Acad Sci U S A 1996 93 5191 5196 8643551 10.1073/pnas.93.11.5191
Philo JS Yang TH LaBarre M Re-examining the oligomerization state of macrophage migration inhibitory factor (MIF) in solution Biophys Chem 2004 108 77 87 15043922 10.1016/j.bpc.2003.10.010
Potolicchio I Santambrogio L Strominger JL Molecular interaction and enzymatic activity of macrophage migration inhibitory factor with immunorelevant peptides J Biol Chem 2003 278 30889 30895 12740374 10.1074/jbc.M302854200
Vera PL Iczkowski KA Leng L Bucala R Meyer-Siegler KL Macrophage migration inhibitory factor is released as a complex with α-1-inhibitor-3 in the intraluminal fluid during bladder inflammation in the rat J Urol 2005 174 338 343 15947686 10.1097/01.ju.0000161606.15696.79
Wassler M Esnard F Fries E Posttranslational folding of alpha 1-inhibitor 3. Evidence for a compaction process J Biol Chem 1995 270 24598 24603 7592680 10.1074/jbc.270.41.24598
Gonias SL Carmichael A Mettenburg JM Roadcap DW Irvin WP Webb DJ Identical or overlapping sequences in the primary structure of human alpha(2)-macroglobulin are responsible for the binding of nerve growth factor-beta, platelet-derived growth factor-BB, and transforming growth factor-beta J Biol Chem 2000 275 5826 5831 10681572 10.1074/jbc.275.8.5826
Liu Q Ling TY Shieh HS Johnson FE Huang JS Huang SS Identification of the high affinity binding site in transforming growth factor-beta involved in complex formation with alpha 2-macroglobulin. Implications regarding the molecular mechanisms of complex formation between alpha 2-macroglobulin and growth factors, cytokines, and hormones J Biol Chem 2001 276 46212 46218 11583997 10.1074/jbc.M105177200
del Vecchio MT Tripodi SA Arcuri F Pergola L Hako L Vatti R Macrophage migration inhibitory factor in prostatic adenocarcinoma: correlation with tumor grading and combination endocrine treatment- related changes Prostate 2000 45 51 57 10960842 10.1002/1097-0045(20000915)45:1<51::AID-PROS6>3.0.CO;2-9
Arcuri F del Vecchio MT de Santi MM Lalinga AV Pallini V Bini L Macrophage migration inhibitory factor in the human prostate: identification and immunocytochemical localization Prostate 1999 39 159 165 10334104 10.1002/(SICI)1097-0045(19990515)39:3<159::AID-PROS3>3.0.CO;2-M
Weir EG Partin AW Epstein JI Correlation of serum prostate specific antigen and quantitative immunohistochemistry J Urol 2000 163 1739 1742 10799172 10.1097/00005392-200006000-00024
Stenman U-H Leinonen J Zhang W-M Finne P Prostate-specific antigen Semin Cancer Biology 1999 9 83 93 10202130 10.1006/scbi.1998.0086
Wraight CJ van Endert P Moller P Lipp J Ling NR MacLennan IC Human major histocompatibility complex class II invariant chain is expressed on the cell surface J Biol Chem 1990 265 5787 5792 1690714
Ong GL Goldenberg DM Hansen HJ Mattes MJ Cell surface expression and metabolism of major histocompatibility complex class II invariant chain (CD74) by diverse cell lines Immunology 1999 98 296 302 10540230 10.1046/j.1365-2567.1999.00868.x
| 16000172 | PMC1177932 | CC BY | 2021-01-04 16:03:04 | no | BMC Cancer. 2005 Jul 6; 5:73 | utf-8 | BMC Cancer | 2,005 | 10.1186/1471-2407-5-73 | oa_comm |
==== Front
BMC CancerBMC Cancer1471-2407BioMed Central London 1471-2407-5-761600197510.1186/1471-2407-5-76Research ArticleThe proteasome inhibitor MG-132 sensitizes PC-3 prostate cancer cells to ionizing radiation by a DNA-PK-independent mechanism Pajonk Frank [email protected] Ophoven Arndt [email protected] Christian [email protected] William H [email protected] Department of Radiation Oncology, David Geffen School of Medicine at UCLA, 10833 Le Conte Avenue, Los Angeles, CA90095-1714, USA2 Department of Urology, University Hospital Münster, Albert-Schweitzer-Straße 33, D-48149 Münster Germany3 Department of Radiation Oncology, University Hospital Freiburg, Robert-Koch-Straße 3, D-79106 Freiburg, Germany2005 7 7 2005 5 76 76 23 3 2005 7 7 2005 Copyright © 2005 Pajonk 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
By modulating the expression levels of specific signal transduction molecules, the 26S proteasome plays a central role in determining cell cycle progression or arrest and cell survival or death in response to stress stimuli, including ionizing radiation. Inhibition of proteasome function by specific drugs results in cell cycle arrest, apoptosis and radiosensitization of many cancer cell lines. This study investigates whether there is also a concomitant increase in cellular radiosensitivity if proteasome inhibition occurs only transiently before radiation. Further, since proteasome inhibition has been shown to activate caspase-3, which is involved in apoptosis, and caspase-3 can cleave DNA-PKcs, which is involved in DNA-double strand repair, the hypothesis was tested that caspase-3 activation was essential for both apoptosis and radiosensitization following proteasome inhibition.
Methods
Prostate carcinoma PC-3 cells were treated with the reversible proteasome inhibitor MG-132. Cell cycle distribution, apoptosis, caspase-3 activity, DNA-PKcs protein levels and DNA-PK activity were monitored. Radiosensitivity was assessed using a clonogenic assay.
Results
Inhibition of proteasome function caused cell cycle arrest and apoptosis but this did not involve early activation of caspase-3. Short-time inhibition of proteasome function also caused radiosensitization but this did not involve a decrease in DNA-PKcs protein levels or DNA-PK activity.
Conclusion
We conclude that caspase-dependent cleavage of DNA-PKcs during apoptosis does not contribute to the radiosensitizing effects of MG-132.
==== Body
Background
One of the most challenging ongoing efforts in radiotherapy is the search for agents that target tumor-specific characteristics to cause radiosensitization without increasing normal tissue complications. The malignant phenotype is normally associated with acquisition of mutations in genes encoding signal transduction molecules that control cell proliferation and/or cell death. Numerous experimental studies have shown, that the expression of mutated oncogenes and tumor suppressor genes in normal cells alters their intrinsic cellular radiosensitivity and clinical studies have, in some cases, indicated that such mutations influence disease free survival after chemotherapy and radiation therapy [1-4]. These observations suggest that the pathways leading to cell cycle arrest and cell death following exposure to ionizing irradiation are linked to DNA damage and repair. Our knowledge of the molecular circuitry that is involved is still however far from complete.
One of the processes that controls expression of short-lived cell cycle and cell death regulators and transcription factors, such as cyclin A, B and E, p21 and p27, p53, cJun, cFos, and nuclear factor κB (NF-κB), is the proteolytic degradation through the 26S proteasome [5-7]. Indeed, the initial response to many stress signals such as radiation-induced DNA damage [8], hypoxia/reperfusion [9], and exposure to heat [10,11] or cytokines [12] appears to involve rapid alterations in 26S proteasome activity resulting in cell cycle arrest and/or apoptosis.
The 26S proteasome is a 2 MDa proteolytic complex that has 3 different cleavage activities [13]. Its function is highly regulated [14-18]. Highly specific and potent inhibitors of the 26S proteasome like the reversible inhibitor MG-132 or lactacystein, which acts irreversibly, have been shown to induce apoptosis in many tumor cell lines [19-27] and in SV-40-transformed but not normal human fibroblasts [27]. Apoptosis is regulated by the ubiquitin/proteasome system at various levels and inhibition of proteasome function may induce [21] or prevent apoptosis [28]. A possible association between apoptosis and radiosensitization exist through caspase activation and subsequent proteolytic destruction of DNA repair enzymes. Caspase activation has been reported to follow proteasome inhibition [24,29,30] and is known to degrade the catalytic subunit of DNA-dependent protein kinase (DNA-PKcs), a key DNA repair enzyme of non-homologous end-joining (NHEJ) of DNA double strand breaks [31,32].
Still, the exact mechanisms of how proteasome inhibition causes apoptosis are not fully understood but the fact, that malignant cells are much more sensitive to the death-promoting aspects of proteasome inhibition than normal cells [33] makes the ubiquitin/26S proteasome system a target for cancer therapy. With Velcade (formerly known as bortezomib, PS-341) a first proteasome inhibitor has become available for clinical use [34]. The excellent results achieved in patients with multiple myeloma suggest the use of Velcade in solid cancers, especially in combination with classical chemotherapies or radiation therapy. So far, several Phase I/II studies have used Velcade as mono-therapy in advanced chemotherapy refractory solid cancers [35-40] but knowledge of in-vivo effects of combined modality approaches using Velcade is limited. It is known that proteasome inhibition can sensitize cells to chemotherapy [41] and ionizing radiation [26,42] and therefore combination studies will be launched in the near future.
However, a recent in-vitro report demonstrated that Velcade treatment of A549 lung cancer might enhance or attenuate cisplatin toxicity, depending on the sequence of application of both drugs [28]. These findings underline the need for a better understanding of the mechanisms of action of these new compounds as they may interfere with standard therapies.
In this study, we investigated a possible link between induction of the apoptotic death program and radiosensitization following proteasome inhibition.
Methods
Cell culture
Cultures of PC-3 prostate carcinoma cells American Type Culture Collection (ATCC, Manassas) cells were grown in 75 cm2 flasks (Falcon) at 37°C in a humidified atmosphere at 5 % CO2. PC-3 cells were cultured in DMEM medium (Cell Concepts, Umkirch Germany) supplemented with 10 % heat-inactivated fetal calf serum (FCS, Sigma, St. Louis, MO) and 1 % penicillin/streptomycin (Sigma).
MG-132 treatment
MG-132 (Calbiochem, San Diego, CA) was dissolved in DMSO (10 mM) and small aliquots (30 μl) were stored at -20°C. Velcade (Janssen-Cilag, Neuss, Germany) was solubilized in water at a concentration of 1 mg/ml and stored in aliquots of 50 μl at -80°C. Three hours before irradiation growth medium was replaced by medium containing MG-132 (50 μM, 0.5% DMSO) or Velcade (100 nM in water). Control cells were subjected to DMSO treatment alone (0.5 %). In clonogenic assays, cells were incubated at 37°C for 3 hours, washed with DMEM, trypsinized, counted, diluted and irradiated at room temperature.
Irradiation
Exponentially growing PC-3 cells were trypsinized, counted, and diluted. The cell suspensions were immediately irradiated at room temperature using a 137Cs-laboratory irradiator (JLShephard, Mark I) at a dose rate of 5.80 Gy/min and a Gammacell 40 137Cs-laboratory irradiator at a dose rate of 0.83 Gy/min respectively.
Clonogenic survival
Colony-forming assays with PC-3 cells were performed immediately after irradiation by plating an appropriate number of cells into culture dishes, in triplicate. Before plating, the viability of the cells was assessed during counting by a dye exclusion test with tryphan blue. After 14 days, colonies were fixed with methanol, stained with crystal violet, and were counted if they consisted of more than 50 cells. The fraction of cells surviving irradiation was normalized to the surviving fraction of the corresponding control and survival values and curves were fitted to the data using a linear-quadratic model.
Flow cytometry
For analysis of cell cycle distribution, 1 × 105 cells were trypsinized, washed in PBS and fixed with ice-cold ethanol (70 %). After RNAse treatment (1 mg/ml), cells were permeabilized with Triton X-100 and stained with propidium iodide (0.1 mg/ml). To determine the cell cycle distribution, DNA content was measured using a flow cytometer (FACScan, Becton Dickinson). TUNEL-Assay was carried out as described previously [43].
DNA-PK activity assay
In order to assess DNA-PK activity, DNA-PK-dependent phosphorylation of a biotinylated p53-derived peptide (Glu-Pro-Pro-Leu-Ser-Gln-Glu-Ala-Phe-Ala-Asp-Leu-Trp-Lys-Lys. Promega, Madison, WI; [44]) was measured in the presence of [32P-γ]-ATP. Drug-treated or control cells were washed in low salt buffer (10 mM HEPES, 25 mM KCl, 10 mM NaCl, 1.1 mM MgCl2, 0.1 mM ethylenediaminetetraacetic acid [EDTA], 0.5 mM phenylmethanesulfonyl fluoride [PMSF], pH 7.2), pelleted and lysed by one freezing/thawing cycle as described in [45]. After centrifugation at 10,000 × g for 5 minutes at 4°C, the supernatant was collected and used as cell extract. Protein content was determined using the Micro-BCA protocol (Pierce) with bovine serum albumin (Sigma) as standard. 10 μg protein were incubated for 30 minutes at 30°C in DNA-PK reaction buffer (50 mM HEPES (KOH, pH 7.5), 100 mM KCl, 10 mM MgCl2, 0.2 mM Ethylene glycol-bis(2-aminoethylether)-N,N,N',N'-tetraacetic acid [EGTA], 0.1 mM EDTA, 1 mM DTT), 0.025 mM ATP, 0.5 μCi [32P-γ]-ATP, bovine serum albumin [BSA] 0.1 mg/ml, and with human p53 oligopeptide as substrate in the presence or absence of activated calf thymus DNA to measure DNA-DSB-dependent phosphorylation of p53. The final volume was 25 μl. The reaction was stopped by addition of 25 μl 30 % acetic acid. 10 μl was spotted on Whatman P81 membranes in duplicates and washed four times with 15 % acetic acid.
Membranes were placed on a phosphor imager screen for 2 hours. The screen was read on a phosphor imager (Storm 860, Molecular Dynamics) and the activity measured using the ImageQuant software package (Molecular Dynamics). Activity in the absence of activated DNA was assumed to be unspecific and thus subtracted from corresponding measurements in presence of activated DNA.
Caspase-3 activity assay
For assessment of caspase-3 activity, cells were plated into culture dishes 24 hours before drug treatment. After drug treatment, cells were dislodged mechanically and washed twice in PBS. Caspase-3 activity was assessed as described by Enari et al. [46] with minor modifications: After five cycles of freezing and thawing in extraction buffer (50 mM PIPES-NaOH, pH 7.0, 50 mM KCl, 5 mM EGTA, 2 mM MgCl2, 1 mM dithiothreitol [DTT], 20 μM cytochalasin B, 1 mM phenylmethylsulfonyl fluoride [PMSF], 1 μg/ml leupeptin, 1 μg/ml pepstatin A, 50 μg/ml antipain, 10 μg/ml chymopain) lysates were centrifuged at 10,000 × g for 12 minutes (4°C). The supernatant was immediately frozen in liquid nitrogen and stored at -80°C. Protein concentrations were determined using the Micro-BCA protocol (Pierce) with bovine serum albumin (Sigma) as standard. 36 μg protein were diluted in ICE standard buffer (100 mM HEPES-KOH, pH 7.5, 10 % sucrose, 0.1 % CHAPS, 10 mM DTT, 0.1 mg/ml ovalbumin) containing the fluorogenic caspase-3 substrate DEVD-7-amido-4-methylcoumarin (DEVD-AMC, 1 μM) and incubated for 30 minutes at 30°C. Fluorescence was measured using a fluorescence plate reader (Tecon, excitation 380 nm, emission 460 nm).
Immunoblotting
Cells were washed with PBS and lysed in TOTEX buffer (20 mM HEPES [pH 7.9], 0.35 mM NaCl, 20 % glycerol, 1 % Nonidet P-40, 0.5 mM EDTA, 0.1 mM EGTA, 0.5 mM DTT, 50 μM PMSF and 90 trypsin inhibitor units [TIU's]/ml aprotinin) for 30 minutes on ice. The lysate was centrifuged at 12,000 × g for 5 minutes and the supernatant transferred to fresh tubes. Protein concentrations were determined using the Micro-BCA protocol (Pierce, Rockford, IL) with bovine serum albumin as standard. 100 μg or 50 μg of protein were subjected to SDS gel electrophoresis (0.1 % Sodium dodecyl sulfate [SDS]/6 % polyacrylamide) and blotted to PVDF membranes. Equity of protein loading was confirmed by Ponceau staining of the PVDF membranes. After blocking with 5 % skim milk in PBS, membranes were incubated with a mouse polyclonal antibody against human DNA-PKcs (Santa Cruz Biotechnologies). A secondary horseradish-peroxidase-conjugated antibody and the ECL Plus System (Amersham) were used for visualization. Fluorescence was measured using a Typhoon 9410 Phosphorimager (Blue fluorescence, 600 V, Amersham).
Results
MG-132 induces apoptosis and sensitizes PC-3 cells to ionizing radiation
To investigate if transient proteasome inhibition by proteasome inhibition affects PC-3 cells induction of apoptosis and long-term clonogenic survival was examined following MG-132 treatment. Cells were exposed to 50 μM of MG-132 for 3 hours, washed twice, trypsinized and plated into tissue culture dishes in triplicates. Clonogenicity, as assessed by colony counts after 2 weeks incubation, was decreased by MG-132 treatment of PC-3 cells from 65 ± 5.1 % in dimethyl sulfoxide (DMSO)-treated versus 35 ± 2.1 in MG-132-treated cells. Decreased clonogenicity was due to apoptosis, as assessed morphologically. 3 hours exposure of PC-3 cells to MG-132 at 25 μM and 50 μM concentrations increased the TUNEL-positive (terminal deoxynucleotidyl transferase-mediated nick end labeling) population after 24 hours of incubation from initially 8 % to 67 % and 73 % respectively (Fig. 1). Short-term proteasome inhibition for 3 hours was therefore effective in causing apoptosis in a proportion of cells, but a considerable number remained clonogenically viable.
Clinical treatment with proteasome inhibitors might also be expected to spare a proportion of clonogenic tumor cells and it would be an advantage if these were sensitized to some other form of therapy. To test whether they were sensitized to radiation, PC-3 cells were exposed to short-term treatment with MG-132. After 3 hours, cells were washed, irradiated and plated in a clonogenic assay, as described previously [43]. Inhibition of proteasome function by MG-132 sensitized the surviving clonogenic PC-3 cells to the effects of ionizing radiation as shown by the left-shift of the survival curve (Fig. 2).
Effect of 26S proteasome inhibition on radiation-induced cell cycle arrest and apoptosis
Having validated PC-3 cells as a model for radiosensitization by proteasome inhibitors, we first excluded that the observed radiosensitizing effect was a result of changes in cell cycle distribution. Inhibition of 26S proteasome function is known to block cellular transition from G1- to S-phase and from late S- to G2/M-phase, as well as S phase transition [47]. This is considered to be due to alterations in expression of molecules such as p53, p21WAF/CIP1, pRB, p27, and cyclins A, B, and E. Similar effects can be seen following exposure of cells to ionizing radiation, which can cause arrest at the G1 and G2/M checkpoints, as well as S-phase delay, and apoptosis in certain cancer cell lines [48]. In cells with mutated p53, like PC-3 prostate cancer cells, the radiation-induced G1 checkpoint delay and the pro-apoptotic response are often abrogated, but the cells will arrest in G2/M (reviewed in [49]).
The effect of the combined treatment of proteasome inhibitors and ionizing irradiation on cell cycle arrest was tested using PC-3 cells treated continuously with MG-132 (50 μM) from 3 hours prior up to 24 hours after irradiation. This treatment blocks proteasome activity almost completely [50]. Flow cytometric analyses indicated that radiation induced G2/M but not G1 arrest, as would be expected in the p53-null PC-3 cells (Fig. 3). MG-132 treatment caused p53-independent cell cycle arrest in all phases and, as a result, the combination of both treatments did not lead to a radiation-induced G2/M arrest (Fig. 3).
The effect of MG-132 treatment on caspase-3 and DNA-PK activities
Since short-term proteasome inhibition caused both apoptosis and radiosensitization in PC-3 cells, this model can be used to investigate the relationship between these phenomena and to explore the pathways that might interconnect them. We considered the hypothesis that both involve caspase-3 activation. Proteasome inhibition has been reported to induce apoptosis that is mediated by caspase activation [20]. Also, the catalytic subunit of the DNA-PK complex, DNA-PKcs, which is required for NHEJ repair pathway of DNA double strand breaks, is a known substrate of caspase-3 [51]. Its destruction could result in radiosensitization.
MG-132 (50 μM) was maintained in the growth medium of PC-3 cells for various time periods. DNA-PKcs protein levels were monitored by western blotting (Fig. 4). Expression of the intact 460 kDa (upper arrow) protein did not change after 3 (lane 1) and 6 hours (lane 2) but slightly decreased after 24 hours of incubation (lane 3) compared to DMSO-treated control cells (lane 7). At this time-point the specific 160 kDa caspase-3-dependent degradation product of DNA-PKcs appeared (lane 3, lower arrow) which coincided with occurrence of morphological signs of apoptosis. Comparable results were obtained using Velcade (100 nM) another specific proteasome inhibitor (lane 4,5 and 6, Fig. 4).
Because the level of DNA-PKcs might not reflect the total kinase activity of the enzyme complex, functional activity was assessed by phosphorylation of a p53 protein fragment in the presence of DNA double strand breaks. This was unchanged over a period of 5 hours after MG-132 treatment, which is the period over which most DNA repair would be expected to take place (Fig 5). In order to exclude a possible trivial explanation that MG-132 might directly interfere with DNA-PK activity, lysates of PC-3 cells were incubated with MG-132 and studied for p53 phosphorylation. The drug did not affect DNA-PK activity directly (Fig. 6). Since caspase-3 activity has been reported to be increased by proteasome inhibition in the MOe7 cell line [20], caspase-3 activity was measured using a fluorogenic substrate in PC-3 cells at various time points after short-term (3 hours) treatment of PC-3 cells with MG-132. We found a substantial drop in caspase-3-like activity as early as 15' minutes after the end of MG-132 incubation if compared to untreated control cells. Caspase-3-like cleavage activity slowly recovered about 3 hours after the end of MG-132 incubation (Fig. 7), but was never increased above baseline. Radiation (30 Gy) with or without MG-132 pre-treatment also failed to activate caspase-3-like activity during the first 3 hours.
Discussion
Proteasome inhibitors like MG-132 sensitize cancer cells to ionizing radiation. In this study, we investigated interference of proteasome inhibition with NHEJ by caspase-3-dependent cleavage of DNA-PKcs as a possible underlying mechanism.
Treatment with MG-132 caused cell cycle arrest in un-irradiated and irradiated cells, induced apoptosis, decreased clonogenicity, and sensitized the surviving cells to ionizing radiation. This was in accordance with previous reports [19-21,47,52-55].
Drug-induced changes in cell cycle redistribution were most unlikely responsible for this effect because even after 24 hours MG-132 treatment mainly increased the number of cells in late S-phase, which has been shown to coincide with greatest radioresistance [56].
The relationship between pro-apoptotic pathways and radiosensitization is controversial, as is the importance of apoptosis in cell death following exposure to clinically relevant doses of ionizing radiation. In general, the majority of cells in solid carcinomas do not enter the apoptotic death pathway [57]. Instead, they undergo several cell divisions until they die or finally survive. Thus, the fate of cells in solid carcinomas after irradiation may be determined more by their ability to repair DNA damage caused by ionizing radiation than by initiation of apoptosis [58]. However, common molecular pathways may link these two phenomena. So far, any link between survival pathways and molecules involved in DNA repair has been elusive but cannot be excluded.
The most important DNA lesions occurring after exposure of cells to ionizing radiation that determines death or survival of a cancer cell are the double strand breaks. A process called non-homologous end joining (NHEJ) repairs these lesions in eukaryote cells. Concerted concession of NHEJ requires the activity of the catalytic subunit of the DNA-dependent protein kinase (DNA-PKcs), which is a known substrate of caspase-3. In this study we therefore focused on the possibility that proteasome inhibition by MG-132 activates caspase-3, as has been reported previously [24,30]. This could mediate apoptosis and cause degradation of DNA-PKcs, the catalytic subunit of DNA-PK [20], resulting in radiosensitization. Reduction of DNA-PK activity following inactivation or mutation of DNA-PKcs is known to enhance radiosensitivity by decreasing repair of DNA-DSB's [59-61]. Although DSB repair is critical for cell survival after exposure to ionizing radiation [62], it is not clear whether this is a rate-limiting step dependent on the level of expression of DNA-PKcs. For example, DNA-PKcs level has been reported not to correlate with radiosensitivity of gliomas [63] and normal fibroblasts [64]. In any event, we were not able to detect meaningful changes in DNA-PK activity following MG-132 drug treatment. Degradation that did occur, appeared late at 24 hours and was probably an effect rather than a cause of the apoptotic process. According to the kinetics of the DSB-repair process described previously [65], any event interfering with the repair of DNA-DSB's has to take place during the initial 6 hours after irradiation. Consistent with the late cleavage of DNA-PKcs, we were not able to detect early activation of caspase-3 following treatment of PC-3 cells with MG-132. In fact, we observed a substantial drop in DEVD-AMC cleavage activity, which might be explained by the observation, that proteasome activity is necessary to activate caspase-3 at least in some cells [66,67]. Our observations are in accordance with data from Hideshima and coworkers who also could not detect changes in DNA-PKcs levels or caspase-3 activation during the initial 6 hours of Valcade-treatment in multiple myeloma cells [68]. Additionally, Wu and coworkers excluded the involvement of caspases in apoptosis of MO7e cells following proteasome inhibition, as caspase inhibitors failed to prevent DNA fragmentation [20] and it is therefore possible that proteasome inhibition induces caspase-independent apoptosis as described for cells, defective in the ubiquitin pathway [69].
Taken together, we conclude that although proteasome inhibition induces apoptosis in most cancer cells, sensitization of PC-3 cells to ionizing radiation occurs through mechanism that does not involve cleavage of DNA-PKcs.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
FP carried out the molecular studies and drafted the manuscript. AO and ChW participated in the design of the study and helped to draft the manuscript. WMB conceived of the study, and participated in its design and coordination and helped to draft the manuscript. All authors read and approved the final manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
FP was supported by grants from the German Research Foundation Pa 723-2-1 and Pa 723/3-1. WM was supported by a grant from the National Cancer Institute, Department of Health and Human Services PHS grant CA-87887.
Figures and Tables
Figure 1 MG-132 induces apoptosis in PC-3 cells. Representative (n = 3) FACS analysis of PC-3 prostate cancer cells incubated with 0, 25 and 50 μM concentrations of MG-132 for 24 h. TUNEL-staining indicated an increase of apoptotic cells from 7.6 % in control cells to 66.5 % and 72.8 % in MG-132 treated cells.
Figure 2 MG-132 sensitizes human cancer cell lines to ionizing radiation. PC-3 prostate cancer cells were incubated with MG-132 (50 μM) for 3 hours, washed twice, irradiated and plated into culture dishes. After 14 days the colonies were fixed with ethanol, stained with crystal violet, and counted. MG-132 treatment sensitized PC-3 cells to ionizing radiation. Radiobiological parameters obtained from a linear-quadratic fit (LQ-fit): PC-3 control α = 0.3, β = 0.043, α/β = 7.6; PC-3 MG-132 α = 0.3, β = 0.07, α/β = 4.3. Data shows means (± standard deviation) from 3 experiments (each plated in triplicates).
Figure 3 The effect of proteasome inhibition by MG-132 on cell cycle progression and radiation-induced cell cycle arrest and apoptosis. Cell cycle distribution of PC-3 prostate cancer cells assessed by flow cytometry after staining with propidium iodide. Cells were pre-incubated with MG-132 (50 μM) for 3 hours, irradiated, and incubated for additional 24 hours in the presence of MG-132. MG-132 prevented cell cycle progression and, consequently, dose-dependent radiation-induced G2/M-arrest. Additionally, MG-132 treatment increased the number of cells in late S-phase.
Figure 4 Degradation of DNA-PKcs by caspase-3 is a late event in apoptosis induced by proteasome inhibition in PC-3 cells. Representative (n = 3) western blot of lysates from PC-3 prostate cancer cells after 3, 6 and 24 hours incubation with MG-132 (50 μM) or Velcade (100 nM) using a specific antibody against human DNA-PKcs. The antibody recognizes the intact 460 kDa protein as well as the specific 160 kDa fragment observed after caspase-3 dependent cleavage of DNA-PKcs. MG-132 caused no significant change of the DNA-PKcs protein levels over a period of 6 hours. Detection of a caspase-3-cleavage specific 160 kDa fragment coincided with apoptosis at 24 hours, but did not occur during the first 6 hours.
Figure 5 MG-132 does not decrease DNA-PK activity. PC-3 cells were incubated for 3 hours with MG-132 (50 μM). Cells were washed and incubated at 37°C. At indicated times, cells were lysed and DNA-PK activity was measured by phosphorylation of a human p53 derived oligopeptide in the presence of DNA DSB's. Proteasome inhibition did not cause any significant decrease of DNA-PK activity (n = 2, data expressed as mean ± standard error mean).
Figure 6 MG-132 does not interfere with DNA-PK directly. Lysates of untreated PC-3 cells were incubated with MG-132. There was no decrease in DNA-PK activity, excluding any direct interaction with the drug (n = 3, data expressed as mean ± standard error mean).
Figure 7 MG-132 treatment reduces caspase-3-like activity in PC-3 cells. PC-3 cells were pre-incubated with MG-132 (50 μM) for 3 hours, washed, irradiated, and incubated at 37°C. 15 minutes, 1, and 3 hours later, cells were lysed and total cellular protein was assayed for caspase-3-like activity, measuring the release of AMC from the fluorogenic caspase-3 substrate DEVD-AMC. Irradiation with 30 Gy initially decreased constitutive caspase-3 activity in PC-3 cells at 15 minutes followed by a slight increase at 1 and 3 hours after irradiation. In contrast, proteasome inhibition with MG-132 completely inhibited caspase-3 activity at 15 minutes. Activity recovered slowly reaching about 30% of baseline levels after 3 hours. This was not altered by combination of MG-132 treatment with ionizing radiation (n = 3, data expressed as mean ± standard error mean).
==== Refs
Caffo O Doglioni C Veronese S Bonzanini M Marchetti A Buttitta F Fina P Leek R Morelli L Palma PD Harris AL Barbareschi M Prognostic value of p21(WAF1) and p53 expression in breast carcinoma: an immunohistochemical study in 261 patients with long-term follow-up Clin Cancer Res 1996 2 1591 1599 9816338
Bonin SR Pajak TF Russell AH Coia LR Paris KJ Flam MS Sauter ER Overexpression of p53 protein and outcome of patients treated with chemoradiation for carcinoma of the anal canal: a report of randomized trial RTOG 87-04. Radiation Therapy Oncology Group Cancer 1999 85 1226 1233 10189126 10.1002/(SICI)1097-0142(19990315)85:6<1226::AID-CNCR3>3.0.CO;2-B
Hayakawa K Mitsuhashi N Hasegawa M Saito Y Sakurai H Ohno T Maebayashi K Ebara T Hayakawa KY Niibe H The prognostic significance of immunohistochemically detected p53 protein expression in non-small cell lung cancer treated with radiation therapy Anticancer Res 1998 18 3685 3688 9854477
Pomp J Davelaar J Blom J van Krimpen C Zwinderman A Quint W Immerzeel J Radiotherapy for oesophagus carcinoma: the impact of p53 on treatment outcome Radiother Oncol 1998 46 179 184 9510045 10.1016/S0167-8140(97)00163-1
Rock KL Gramm C Rothstein L Clark K Stein R Dick L Hwang D Goldberg AL Inhibitors of the proteasome block the degradation of most cell proteins and the generation of peptides presented on MHC class I molecules Cell 1994 78 761 771 8087844 10.1016/S0092-8674(94)90462-6
Rolfe M Chiu MI Pagano M The ubiquitin-mediated proteolytic pathway as a therapeutic area J Mol Med 1997 75 5 17 9020379 10.1007/s001090050081
Weissman AM Regulating protein degradation by ubiquitination Immunol Today 1997 18 189 198 9136456 10.1016/S0167-5699(97)84666-X
Pajonk F McBride WH Ionizing radiation affects 26s proteasome function and associated molecular responses, even at low doses Radiother Oncol 2001 59 203 212 11325450 10.1016/S0167-8140(01)00311-5
Stempien-Otero A Karsan A Cornejo CJ Xiang H Eunson T Morrison RS Kay M Winn R Harlan J Mechanisms of hypoxia-induced endothelial cell death. Role of p53 in apoptosis J Biol Chem 1999 274 8039 8045 10075703 10.1074/jbc.274.12.8039
Kueckelkorn U Knuehl C Boes-Fabian B Drung I Kloetzel PM The effect of heat shock on 20S/26S proteasomes Biol Chem 2000 381 1017 1023 11076035 10.1515/BC.2000.125
Pajonk F van Ophoven A McBride WH Hyperthermia-induced proteasome inhibition and loss of androgen receptor expression in human prostate cancer cells Cancer Res 2005 65 4836 4843 15930304
Glockzin S von Knethen A Scheffner M Brune B Activation of the cell death program by nitric oxide involves inhibition of the proteasome J Biol Chem 1999 274 19581 19586 10391892 10.1074/jbc.274.28.19581
Hasselgren PO Fischer JE The ubiquitin-proteasome pathway: review of a novel intracellular mechanism of muscle protein breakdown during sepsis and other catabolic conditions Ann Surg 1997 225 307 316 9060588 10.1097/00000658-199703000-00011
Tanaka K Molecular biology of proteasomes Mol Biol Rep 1995 21 21 26 7565659 10.1007/BF00990966
Stohwasser R Kloetzel PM Cytokine induced changes in proteasome subunit composition are concentration dependent Biol Chem 1996 377 571 577 9067255
Nandi D Woodward E Ginsburg DB Monaco JJ Intermediates in the formation of mouse 20S proteasomes: implications for the assembly of precursor beta subunits EMBO J 1997 16 5363 5375 9311996 10.1093/emboj/16.17.5363
Schmidt M Kloetzel PM Biogenesis of eukaryotic 20S proteasomes: the complex maturation pathway of a complex enzyme FASEB J 1997 11 1235 1243 9409542
Kloetzel PM Soza A Stohwasser R The role of the proteasome system and the proteasome activator PA28 complex in the cellular immune response Biol Chem 1999 380 293 297 10223331 10.1515/BC.1999.040
Drexler HC Activation of the cell death program by inhibition of proteasome function Proc Natl Acad Sci U S A 1997 94 855 860 9023346 10.1073/pnas.94.3.855
Wu LW Reid S Ritchie A Broxmeyer HE Donner DB The proteasome regulates caspase-dependent and caspase-independent protease cascades during apoptosis of MO7e hematopoietic progenitor cells Blood Cells Mol Dis 1999 25 20 29 10349510 10.1006/bcmd.1999.0223
Herrmann JL Briones FJ Brisbay S Logothetis CJ McDonnell TJ Prostate carcinoma cell death resulting from inhibition of proteasome activity is independent of functional Bcl-2 and p53 Oncogene 1998 17 2889 2899 9879995 10.1038/sj.onc.1202221
McDade TP Perugini RA Vittimberga FJJ Callery MP Ubiquitin-proteasome inhibition enhances apoptosis of human pancreatic cancer cells Surgery 1999 126 371 377 10455908 10.1067/msy.1999.98785
Adams J Palombella VJ Sausville EA Johnson J Destree A Lazarus DD Maas J Pien CS Prakash S Elliott PJ Proteasome inhibitors: a novel class of potent and effective antitumor agents Cancer Res 1999 59 2615 2622 10363983
Zhang XM Lin H Chen C Chen BD Inhibition of ubiquitin-proteasome pathway activates a caspase-3-like protease and induces Bcl-2 cleavage in human M-07e leukaemic cells Biochem J 1999 340 127 133 10229667 10.1042/0264-6021:3400127
Kitagawa H Tani E Ikemoto H Ozaki I Nakano A Omura S Proteasome inhibitors induce mitochondria-independent apoptosis in human glioma cells FEBS Lett 1999 443 181 186 9989601 10.1016/S0014-5793(98)01709-8
Pajonk F Pajonk K McBride WH Apoptosis and radiosensitization of hodgkin cells by proteasome inhibition Int J Radiat Oncol Biol Phys 2000 47 1025 1032 10863075 10.1016/S0360-3016(00)00516-2
An B Goldfarb RH Siman R Dou QP Novel dipeptidyl proteasome inhibitors overcome Bcl-2 protective function and selectively accumulate the cyclin-dependent kinase inhibitor p27 and induce apoptosis in transformed, but not normal, human fibroblasts Cell Death Differ 1998 5 1062 1075 9894613 10.1038/sj.cdd.4400436
Mortenson MM Schlieman MG Virudachalam S Bold RJ Effects of the proteasome inhibitor bortezomib alone and in combination with chemotherapy in the A549 non-small-cell lung cancer cell line Cancer Chemother Pharmacol 2004 54 343 53 15197486 10.1007/s00280-004-0811-4
Pasquini LA Besio Moreno M Adamo AM Pasquini JM Soto EF Lactacystin, a specific inhibitor of the proteasome, induces apoptosis and activates caspase-3 in cultured cerebellar granule cells J Neurosci Res 2000 59 601 611 10686588 10.1002/(SICI)1097-4547(20000301)59:5<601::AID-JNR3>3.0.CO;2-1
Qiu JH Asai A Chi S Saito N Hamada H Kirino T Proteasome inhibitors induce cytochrome c-caspase-3-like protease- mediated apoptosis in cultured cortical neurons J Neurosci 2000 20 259 265 10627603
Han Z Malik N Carter T Reeves WH Wyche JH Hendrickson EA DNA-dependent protein kinase is a target for a CPP32-like apoptotic protease J Biol Chem 1996 271 25035 25040 8798786 10.1074/jbc.271.40.25035
Song Q Burrows SR Smith G Lees-Miller SP Kumar S Chan DW Trapani JA Alnemri E Litwack G Lu H Moss DJ Jackson S Lavin MF Interleukin-1 beta-converting enzyme-like protease cleaves DNA- dependent protein kinase in cytotoxic T cell killing J Exp Med 1996 184 619 626 8760815 10.1084/jem.184.2.619
Masdehors P Omura S Merle-Beral H Mentz F Cosset JM Dumont J Magdelenat H Delic J Increased sensitivity of CLL-derived lymphocytes to apoptotic death activation by the proteasome-specific inhibitor lactacystin Br J Haematol 1999 105 752 757 10354141 10.1046/j.1365-2141.1999.01388.x
Kane RC Bross PF Farrell AT Pazdur R Velcade: U.S. FDA approval for the treatment of multiple myeloma progressing on prior therapy Oncologist 2003 8 508 513 14657528 10.1634/theoncologist.8-6-508
Maki RG Kraft AS Scheu K Yamada J Wadler S Antonescu CR Wright JJ Schwartz GK A multicenter Phase II study of bortezomib in recurrent or metastatic sarcomas Cancer 2005 103 1431 1438 15739208 10.1002/cncr.20968
Price N Dreicer R Phase I/II trial of bortezomib plus docetaxel in patients with advanced androgen-independent prostate cancer Clin Prostate Cancer 2004 3 141 143 15636679
Papandreou CN Daliani DD Nix D Yang H Madden T Wang X Pien CS Millikan RE Tu SM Pagliaro L Kim J Adams J Elliott P Esseltine D Petrusich A Dieringer P Perez C Logothetis CJ Chatta G Phase I trial of the proteasome inhibitor bortezomib in patients with advanced solid tumors with observations in androgen-independent prostate cancer J Clin Oncol 2004 22 2108 21 15169797 10.1200/JCO.2004.02.106
Blaney SM Bernstein M Neville K Ginsberg J Kitchen B Horton T Berg SL Krailo M Adamson PC Phase I study of the proteasome inhibitor bortezomib in pediatric patients with refractory solid tumors: a Children's Oncology Group study (ADVL0015) J Clin Oncol 2004 22 4804 4809 15570082 10.1200/JCO.2004.12.185
Kondagunta GV Drucker B Schwartz L Bacik J Marion S Russo P Mazumdar M Motzer RJ Phase II trial of bortezomib for patients with advanced renal cell carcinoma J Clin Oncol 2004 22 3720 3725 15365068 10.1200/JCO.2004.10.155
Shah MH Young D Kindler HL Webb I Kleiber B Wright J Grever M Phase II study of the proteasome inhibitor bortezomib (PS-341) in patients with metastatic neuroendocrine tumors Clin Cancer Res 2004 10 6111 6118 15447997
Hideshima T Richardson P Chauhan D Palombella VJ Elliott PJ Adams J Anderson KC The proteasome inhibitor PS-341 inhibits growth, induces apoptosis, and overcomes drug resistance in human multiple myeloma cells Cancer Res 2001 61 3071 3076 11306489
Russo SM Tepper JE Baldwin ASJ Liu R Adams J Elliott P Cusack JCJ Enhancement of radiosensitivity by proteasome inhibition: implications for a role of NF-kappaB Int J Radiat Oncol Biol Phys 2001 50 183 193 11316563 10.1016/S0360-3016(01)01446-8
Pajonk F Pajonk K McBride W Apoptosis and radiosensitization of Hodgkin’s cells by proteasome inhibition Int J Radiat Oncol Biol Phys 2000 47 1025 1032 10863075 10.1016/S0360-3016(00)00516-2
Lees-Miller SP Sakaguchi K Ullrich SJ Appella E Anderson CW Human DNA-activated protein kinase phosphorylates serines 15 and 37 in the amino-terminal transactivation domain of human p53 Mol Cell Biol 1992 12 5041 5049 1406679
Ihara M Suwa A Komatsu K Shimasaki T Okaichi K Hendrickson EA Okumura Y Heat sensitivity of double-stranded DNA-dependent protein kinase (DNA- PK) activity Int J Radiat Biol 1999 75 253 258 10072187 10.1080/095530099140717
Enari M Talanian RV Wong WW Nagata S Sequential activation of ICE-like and CPP32-like proteases during Fas- mediated apoptosis Nature 1996 380 723 726 8614469 10.1038/380723a0
Machiels BM Henfling ME Gerards WL Broers JL Bloemendal H Ramaekers FC Schutte B Detailed analysis of cell cycle kinetics upon proteasome inhibition Cytometry 1997 28 243 252 9222110 10.1002/(SICI)1097-0320(19970701)28:3<243::AID-CYTO9>3.0.CO;2-E
Maity A McKenna WG Muschel RJ The molecular basis for cell cycle delays following ionizing radiation: a review Radiother Oncol 1994 31 1 13 8041894 10.1016/0167-8140(94)90408-1
Hwang A Muschel RJ Radiation and the G2 phase of the cell cycle Radiat Res 1998 150 S52 9 9806609
Lee DH Goldberg AL Selective inhibitors of the proteasome-dependent and vacuolar pathways of protein degradation in Saccharomyces cerevisiae J Biol Chem 1996 271 27280 27284 8910302 10.1074/jbc.271.44.27280
Teraoka H Yumoto Y Watanabe F Tsukada K Suwa A Enari M Nagata S CPP32/Yama/apopain cleaves the catalytic component of DNA-dependent protein kinase in the holoenzyme FEBS Lett 1996 393 1 6 8804412 10.1016/0014-5793(96)00842-3
Dietrich C Bartsch T Schanz F Oesch F Wieser RJ p53-dependent cell cycle arrest induced by N-acetyl-L-leucinyl-L- leucinyl-L-norleucinal in platelet-derived growth factor-stimulated human fibroblasts Proc Natl Acad Sci U S A 1996 93 10815 10819 8855263 10.1073/pnas.93.20.10815
Hashemolhosseini S Nagamine Y Morley SJ Desrivieres S Mercep L Ferrari S Rapamycin inhibition of the G1 to S transition is mediated by effects on cyclin D1 mRNA and protein stability J Biol Chem 1998 273 14424 14429 9603954 10.1074/jbc.273.23.14424
Rao S Porter DC Chen X Herliczek T Lowe M Keyomarsi K Lovastatin-mediated G1 arrest is through inhibition of the proteasome, independent of hydroxymethyl glutaryl-CoA reductase Proc Natl Acad Sci U S A 1999 96 7797 7802 10393901 10.1073/pnas.96.14.7797
Delic J Masdehors P Omura S Cosset JM Dumont J Binet JL Magdelenat H The proteasome inhibitor lactacystin induces apoptosis and sensitizes chemo- and radioresistant human chronic lymphocytic leukaemia lymphocytes to TNF-alpha-initiated apoptosis [see comments] Br J Cancer 1998 77 1103 1107 9569046
Sinclair WK Morton RA X-ray sensitivity during the cell generation cycle of cultured Chinese hamster cells Radiat Res 1966 29 450 474 5924188
Meyn RE Stephens LC Milas L Programmed cell death and radioresistance Cancer Metastasis Rev 1996 15 119 131 8842483 10.1007/BF00049491
Weichselbaum RR Little JB The differential response of human tumours to fractionated radiation may be due to a post-irradiation repair process Br J Cancer 1982 46 532 537 7138762
Rosenzweig KE Youmell MB Palayoor ST Price BD Radiosensitization of human tumor cells by the phosphatidylinositol3- kinase inhibitors wortmannin and LY294002 correlates with inhibition of DNA-dependent protein kinase and prolonged G2-M delay Clin Cancer Res 1997 3 1149 1156 9815794
Chernikova SB Wells RL Elkind MM Wortmannin sensitizes mammalian cells to radiation by inhibiting the DNA-dependent protein kinase-mediated rejoining of double-strand breaks Radiat Res 1999 151 159 166 9952300
Hsieh CL Arlett CF Lieber MR V(D)J recombination in ataxia telangiectasia, Bloom's syndrome, and a DNA ligase I-associated immunodeficiency disorder J Biol Chem 1993 268 20105 20109 8397200
Jeggo PA Identification of genes involved in repair of DNA double-strand breaks in mammalian cells Radiat Res 1998 150 S80 91 9806611
Allalunis-Turner MJ Lintott LG Barron GM Day RS Lees-Miller SP Lack of correlation between DNA-dependent protein kinase activity and tumor cell radiosensitivity Cancer Res 1995 55 5200 5202 7585574
Dikomey E Borgmann K Brammer I Kasten-Pisula U Molecular mechanisms of individual radiosensitivity studied in normal diploid human fibroblasts Toxicology 2003 193 125 135 14599772 10.1016/S0300-483X(03)00293-2
Olive PL Banath JP MacPhail HS Lack of a correlation between radiosensitivity and DNA double-strand break induction or rejoining in six human tumor cell lines Cancer Res 1994 54 3939 3946 8033118
Grimm LM Goldberg AL Poirier GG Schwartz LM Osborne BA Proteasomes play an essential role in thymocyte apoptosis Embo J 1996 15 3835 3844 8670888
Sadoul R Fernandez PA Quiquerez AL Martinou I Maki M Schroter M Becherer JD Irmler M Tschopp J Martinou JC Involvement of the proteasome in the programmed cell death of NGF-deprived sympathetic neurons Embo J 1996 15 3845 3852 8670889
Hideshima T Bergsagel PL Kuehl WM Anderson KC Advances in biology of multiple myeloma: clinical applications Blood 2004 104 607 618 15090448 10.1182/blood-2004-01-0037
Monney L Otter I Olivier R Ozer HL Haas AL Omura S Borner C Defects in the ubiquitin pathway induce caspase-independent apoptosis blocked by Bcl-2 J Biol Chem 1998 273 6121 6131 9497330 10.1074/jbc.273.11.6121
| 16001975 | PMC1177933 | CC BY | 2021-01-04 16:03:04 | no | BMC Cancer. 2005 Jul 7; 5:76 | utf-8 | BMC Cancer | 2,005 | 10.1186/1471-2407-5-76 | oa_comm |
==== Front
BMC Cell BiolBMC Cell Biology1471-2121BioMed Central London 1471-2121-6-291600883910.1186/1471-2121-6-29Research ArticleTim50a, a nuclear isoform of the mitochondrial Tim50, interacts with proteins involved in snRNP biogenesis Xu Hongzhi [email protected] Z Brad [email protected] Melvin L [email protected] Michael D [email protected] Department of Biochemistry, The University of Mississippi Medical Center Jackson, MS 39216-4505, USA2005 11 7 2005 6 29 29 26 1 2005 11 7 2005 Copyright © 2005 Xu et al; licensee BioMed Central Ltd.2005Xu 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 Cajal body (CB) is a nuclear suborganelle involved in the biogenesis of small nuclear ribonucleoproteins (snRNPs), which are vital for pre-mRNA splicing. Newly imported Sm-class snRNPs traffic through CBs, where the snRNA component of the snRNP is modified, and then target to other nuclear domains such as speckles and perichromatin fibrils. It is not known how nascent snRNPs localize to the CB and are released from this structure after modification. The marker protein for CBs, coilin, may play a role in snRNP biogenesis given that it can interact with snRNPs and SMN, the protein mutated in Spinal Muscular Atrophy. Loss of coilin function in mice leads to significant viability and fertility problems and altered CB formation.
Results
In this report, we identify a minor isoform of the mitochondrial Tim50, Tim50a, as a coilin interacting protein. The Tim50a transcript can be detected in some cancer cell lines and normal brain tissue. The Tim50a protein differs only from Tim50 in that it contains an additional 103 aa N-terminal to the translation start of Tim50. Importantly, a putative nuclear localization signal is found within these 103 residues. In contrast to Tim50, which localizes to the cytoplasm and mitochondria, Tim50a is strictly nuclear and is enriched in speckles with snRNPs. In addition to coilin, Tim50a interacts with snRNPs and SMN. Competition binding experiments demonstrate that coilin competes with Sm proteins of snRNPs and SMN for binding sites on Tim50a.
Conclusion
Tim50a may play a role in snRNP biogenesis given its cellular localization and protein interaction characteristics. We hypothesize that Tim50a takes part in the release of snRNPs and SMN from the CB.
==== Body
Background
The biogenesis of most spliceosomal small nuclear ribonucleoproteins (snRNPs) is complicated and requires both cytoplasmic and nuclear maturation steps [1-3]. For example, the spliceosomal small nuclear RNAs (snRNAs) of Ul, U2, U4 and U5 snRNPs are synthesized by RNA polymerase II and may traffic through specific subnuclear domains before being exported to the cytoplasm [1,3]. In the cytoplasm, a septet of Sm proteins (B/B', Dl, D2, D3, E, F, G) binds the Sm motif of the snRNA under the control of the Survival of Motor Neurons (SMN) protein complex. Mutations in the SMN protein cause the neurodegenerative disorder Spinal Muscular Atrophy [4,5]. After the Sm core has been assembled onto the snRNA, the snRNA is subject to further processing, followed by import back into the nucleus; again with the help of the SMN complex [1,5-8].
Upon nuclear re-entry, newly assembled Ul, U2, U4 and U5 snRNPs first localize to a subnuclear domain known as the Cajal body [9]. In the Cajal body (CB), the snRNA component of the snRNP is subjected to pseudouridine-base and 2'-0-methyl sugar-modifications that are guided by small CB-specific RNAs (scaRNAs) [10-12]. These modifications are crucial for proper pre-mRNA splicing in vivo [13]. After their modification in the CB, snRNPs localize to speckles, where they are stored, or perichromatin fibrils, where splicing occurs concurrently with transcription [14]. Unlike Ul, U2, U4 and U5 snRNAs, maturation of the RNA polymerase III-transcribed U6 snRNA does not include a cytoplasmic phase and may take place in the nucleolus and the CB [3]. The U7 snRNP, which is required for histone pre-mRNA 3'-end processing [15-17], is assembled in manner similar to that observed for Ul, U2, U4 and U5 snRNPs, including a cytoplasmic phase [18]. However, U7 snRNA has a noncanonical Sm binding site and thus recruits a different type of Sm core compared to that which binds Ul, U2, U4 and U5 snRNA [19,20]. Like the spliceosomal snRNPs, the U7 snRNP is enriched within CBs [19,21]. CBs have been shown to move in an ATP dependent manner [22,23] as well as associate with various gene loci, including snRNA and histone gene clusters [24]. Consequently, CBs may provide a platform upon which a feedback regulatory mechanism for snRNP and histone biogenesis takes place [25].
The mechanisms by which snRNPs are targeted to and released from the CB are unknown. One possibility is that a factor within the CB interacts with nascent snRNPs and facilitates their modification. Another factor may displace snRNPs from the CB, allowing for their subsequent localization in speckles and perichromatin fibrils. The CB marker protein coilin may play a role in the targeting of snRNPs to CBs. Removal of coilin in Xenopus by immunodepletion decreases snRNP levels in the amphibian equivalent of the CB [26]. Characterization of coilin knockout mice has revealed that they are viable on an outbred mouse strain, but have significant viability and fertility defects on inbred strains [27] (Greg Matera, personal communication). Cell lines derived from coilin knockout mice lack canonical CBs in which snRNPs are enriched [27]. However, add-back experiments demonstrate that typical CBs, containing snRNPs, can be reformed upon the addition of coilin [27]. Furthermore, coilin can interact directly with several Sm proteins of snRNPs [28] (our unpublished observations). Taken together, these data indicate that, while not an essential protein, coilin is important for proper CB formation. Functional CBs may allow for the efficient coordination of the nuclear steps of snRNP biogenesis.
Another protein that requires coilin for its localization to CBs is SMN [27-30]. Coilin directly interacts with SMN via several arginine/glycine (RG) dipeptide repeats in coilin [28]. The arginines within this RG box motif are symmetrically dimethylated [29,30], resulting in an increased affinity for SMN [31,32]. In most cell lines and tissues, SMN localizes to the cytoplasm and CBs [33]. However, SMN in some cell lines and fetal tissue localizes to discrete nuclear structures termed "gems" (for Gemini of Cajal bodies) [34,35]. The presence of gems correlates with a decrease in coilin methylation [29,30]. Although the nuclear role of SMN is not well understood, it is possible that SMN escorts nascent snRNPs to the CB [6,8]. Interestingly, coilin can compete with Sm proteins for binding sites on SMN [28], indicating that locally high concentrations of coilin within the CB may serve to displace snRNPs from the SMN complex. The interplay between SMN, snRNPs and coilin may therefore regulate snRNP accumulation within the CB. We set out to determine if we could identify factors that control the departure of nascent snRNPs from the CB.
In this report, we demonstrate that an isoform of the human mitochondrial Tim50, Tim50a, localizes to speckles and interacts with snRNPs, SMN and coilin. Human Tim50, which has phosphatase activity, is a component of the mitochondrial translocator and regulates mitochondrial integrity and cell death [36]. The protein sequence of Tim50 and Tim50a are identical, with the important exception that Tim50a contains an alternative translational start sequence that adds 103 aa to its N-terminus relative to Tim50. Competition binding experiments show that SmB' does not compete with SMN for Tim50a binding sites, but SmB' does reduce coilin interaction with Tim50a. Furthermore, SMN and coilin compete for Tim50a binding sites and Tim50a forms a more efficient complex with snRNPs in vivo compared to Tim50. Based on these results, we propose that Tim50a is involved in the regulation of snRNP biogenesis and possibly the activity of nuclear SMN.
Results and discussion
Isolation of Tim50a
Given the central role of coilin in the formation and composition of CBs, especially in its ability to recruit the SMN complex and snRNPs, we wanted to identify proteins that interact with coilin and possibly regulate SMN and snRNP localization in the CB. Towards this end, we conducted a yeast two-hybrid screen with coilin as bait. A human brain cDNA library was chosen for this screen because we are interested in assessing CB protein dynamics in neuronal tissue, which is the cell type primarily affected in Spinal Muscular Atrophy. The C-terminal 214 aa of coilin, which interacts with SMN and Sm proteins, was the bait for the screen [28] (our unpublished observations). Consequently, other proteins that bind C214 may regulate the interplay between coilin/SMN and coilin/Sm and thus play a role in CB dynamics.
Several preys were recovered from the screen, one of which was termed CAP50. Sequencing of this prey revealed that it contained the partial coding sequence for a protein known as Tim50. Since Tim50 localizes to the mitochondria and cytoplasm [36], the relevance of CAP50 interaction with coilin, a nuclear protein, was not initially clear. However, upon conducting EST (expressed sequence tag) database searches, we found that there are two possible translational starts for the TIMM50 gene product. The majority of ESTs terminate slightly upstream of the published translational start sequence of Tim50 (MAASA, [36]), suggesting that Tim50 is the major product of the TIMM50 gene. However, three ESTs, all from cancer cells lines, demonstrate that an alternative translational start sequence can be utilized which would add 103 aa to the N-terminus of the Tim50 protein. We have named this protein Tim50a (Figure 1A). Importantly, a putative nuclear localization signal (NLS) [37] is found within this region, indicating that Tim50a, unlike Tim50, may reside in the nucleus.
Figure 1 N-terminal amino acid alignment and relative expression of Tim50a and Tim50. (A) Compared to Tim50, Tim50a contains additional sequence that contains a putative nuclear localization sequence (NLS) upstream of the translational start of Tim50. (B) Schematic representation (not to scale) of Tim50a and Tim50 showing the locations of the primers used in the PCR reactions. (C) The relative expression of Tim50 versus Tim50a in four cancer cell lines. (Upper panel) A standard PCR reaction using cDNA from each cell line as template and ComFor+ComRev primers yields a product that could be from Tim50 and Tim50a. Water serves as a negative control while Tim50a plasmid cDNA serves as a positive control. Nested PCR (lower panel) demonstrates that Tim50a is detected in two cell lines but accounts for only a small fraction of the products from the TIMM50 gene. The cDNA from the following cell lines was used: A-549 (non-small cell lung cancer), NIH: OVCAR-3 (ovarian cancer), MDA-MB-231 (breast cancer) and HCT-15 (colon cancer). The lower band in the NIH: OVCAR-3 nested PCR reaction is a non-specific product.
To verify the existence of Tim50a in other cell lines and to obtain a general idea as to the abundance of this transcript relative to the Tim50 isoform, we performed nested PCR with primers specific to Tim50a on cDNAs from four additional cancer cell lines: A-549, NIH: OVCAR-3, MDA-MB-231 and HCT-15. Another primer set was used in a standard (not nested) PCR reaction to amplify a region common to both Tim50a and Tim50 (Figure 1B). As shown in Figure 1C (upper panel), a standard PCR reaction (30 cycles) using primers that can bind both Tim50a and Tim50 (ComFor+ComRev) generate a product from all of the cell lines tested. In contrast, standard PCR reactions using Tim50a-specific primers (50aFor+50aRev) fail to yield a product from cell line cDNAs, but successfully generate a product using a positive control plasmid template (our unpublished observations). However, nested PCR (20 cycles) reveals that two of the four cell lines, A-549 and NIH: OVCAR-3, contain the Tim50a transcript (Figure 1C, lower panel), as monitored by sequencing of the product. The lower band in the NIH: OVCAR-3 reaction is from an unrelated transcript. Therefore, while Tim50a mRNA is not nearly as abundant as the Tim50 message, this species is present in at least four different cell lines (two lines from this study and two from the EST database).
Cellular localization of Tim50a
To test if Tim50a is a nuclear protein, we obtained an EST encoding the additional residues and cloned the entire Tim50a coding sequence into a GFP expression vector. We also made a GFP fusion to the truncated Tim50 protein isolated from the two-hybrid screen (CAP50) as well as GFP fusions to various mutants of Tim50a. The localization of these proteins, and GFP fused to the mitochondrial Tim50 isoform, was monitored after transient transfection into HeLa cells (Figures 2 and 3). In contrast to Tim50, GFP-Tim50a is exclusively nuclear with speckle-like accumulations (Figure 3). To confirm that the nuclear accumulations are indeed speckles, in which snRNPs are stored, we stained the cells with antibodies against Sm proteins (Y12). There is overlap between Tim50a and snRNPs in speckles (arrows), as shown by the yellow-orange signal in the merged image (Figure 4, row 1, right panel). We do not detect Tim50a in Cajal bodies (arrowheads), indicating that the association of Tim50a with CBs or coilin is weak or transient. Alternatively, coilin may associate with Tim50a in the nucleoplasm. Indeed, while coilin is highly enriched in the CB, the majority (70%) of the protein is diffusely localized throughout the nucleoplasm [38]. We observed identical localization patterns for Tim50a upon fusion to a smaller myc-tag (our unpublished results), suggesting that the GFP-tag does not affect localization. The diffuse cytoplasmic and nuclear localization of GFP-CAP50 and GFP-Tim50a(N87) demonstrate that the region containing the putative NLS is necessary for nuclear localization (Figure 4). Tim50a mutants that retain the putative NLS display nuclear accumulation. For example, an internal deletion of Tim50a (Tim50aSac) is found in a speckled nuclear pattern while the first 135 aa (Tim50aN135) is nuclear and nucleolar. Tim50a localization to speckles is therefore contingent upon the NLS and residues found between aa 135 and 256.
Figure 2 Schematic representation and cellular localization of Tim50a constructs used in this study. The Tim50a protein is 456 aa in length. Note that Tim50, reported by [36], lacks an NLS, which is present in Tim50a, Tim50aSac and Tim50aN135. Tim50aSac contains a deletion between aa 256 and 431.
Figure 3 Cellular localization of GFP-tagged Tim50a and Tim50. HeLa cells were transfected with GFP-Tim50a (top row) or GFP-Tim50 (bottom row) and the cells were stained with DAPI to define the nucleus. The right panels show the overlay of the GFP signal (green) with the DAPI signal (blue). Note that GFP-Tim50 has extensive cytoplasmic staining but Tim50a is exclusively nuclear. The nuclear signal for GFP-Tim50 was not reported by [36] and may be caused by overexpression and/or fusion of GFP to the N-terminus of the protein.
Figure 4 Cellular localization of various Tim50a proteins. Hela cells were transfected with GFP-tagged proteins (first column), followed by staining with antibody Y12 (second column) to detect Sm proteins (snRNPs). The merged image in color is shown in the last column (GFP is green, Sm proteins are red). Arrows mark speckles whereas arrowheads demarcate Cajal bodies. Note that cells for GFP-CAP50 and GFP-Tim50aN87 are shown at a lower magnification compared to the other images in order to demonstrate the cytoplasmic localization of these proteins.
Actinomycin D redistribution of Tim50a
The transcription inhibitor actinomycin D can be used to partially characterize the interaction of proteins in the nucleus. This is because actinomycin D inhibits transcription and causes the reorganization of splicing factors and Cajal bodies [39]. In particular, speckles enlarge and CBs are relocalized to the periphery of the nucleolus in the presence of actinomycin D [39]. If Tim50a were an integral part of speckles, then we would suspect that this protein would be found in enlarged speckles when cells are treated with actinomycin D. However, in the presence of actinomycin D, Tim50a no longer localizes with snRNPs in speckles but is instead found in discrete structures (Figure 5, bottom panels, arrows). This is evident in the merged image, which shows separate green (GFP-Tim50a, arrows) and red (snRNPs, arrowheads) signals with little overlap. In contrast, non-treated cells display overlap between Tim50a and snRNPs, resulting in a yellow-orange signal in the merged image (Figure 5, top panels). We tested whether Tim50a in actinomycin D treated cells localized to paraspeckles, another subnuclear domain of unknown function [40]. No colocalization of Tim50a with the paraspeckle marker protein, PSP1, was observed (data not shown). Taken together, the localization data indicate that Tim50a is likely to have a different function compared to Tim50, possibly in the regulation of CB activity or snRNP biogenesis. However Tim50a is not an integral component of speckles given the separate localization pattern of Tim50a and snRNPs in actinomycin D treated cells.
Figure 5 Actinomycin D treatment alters the co-localization of Tim50a with snRNPs. HeLa cells transfected with GFP-tagged Tim50a (green) were untreated or subject to actinomycin D, followed by fixation and staining with the anti-Sm antibody Y12 (red). The overlay of the images is shown in the last column. For untreated cells, arrowheads mark the location of Cajal bodies and arrows show speckles. For actinomycin D treated cells, arrows mark accumulations of Tim50a while arrowheads show enlarged Sm speckles.
Direct interaction of Tim50a with coilin, SmB' and SMN
The two-hybrid screen suggests that Tim50 isoforms may interact with coilin. To prove this, we conducted GST-pulldown assays using bacterially purified T7-tagged coilin and Tim50a fused to GST. Coilin is not recovered over beads fused to GST alone, but is recovered over GST-Tim50a beads (Figure 6A). We next tested if Tim50a would interact with the C-terminal 214 aa of coilin (C214), which is this same fragment of coilin used to isolate CAP50 in the two-hybrid screen. In agreement with the two-hybrid results, soluble Tim50a is recovered over GST-C214 beads, but not by GST alone (Figure 6B). Since Tim50a localizes to speckles, we were interested in assessing if Tim50a interacts with Sm proteins of snRNPs. Using SmB' as a representative of Sm proteins, we conducted GST-pulldown experiments with a variety of GST-Tim50a proteins. Consistent with the localization data, SmB' directly interacts with GST-Tim50a (Figure 6C, lane 4). SmB' also associates with two mutants of Tim50a, CAP50 and Tim50Sac. Based on the binding results, we suspect that the interaction domain for SmB' on Tim50a lies somewhere between residues 121 and 256 of Tim50a, although we cannot exclude the role of the residues at the extreme C-terminus of Tim50a (aa 431–456, Figure 2). Given that Tim50a localizes to speckles and directly interacts with coilin and SmB', we next tested if Tim50a could interact with another protein crucial for snRNP formation, SMN. As shown in Figure 6D, SMN binds to the GST-Tim50a proteins but not to GST alone. We also examined if the isoform known to localize in mitochondria, Tim50, also interacts with the coilin fragment C214 as found for Tim50a. Purified full-length Tim50 does indeed interact specifically with the coilin fragment as shown in Figure 6E. From these in vitro binding studies, we conclude that coilin, SmB' and SMN can interact with both Tim50a and Tim50 isoforms, but the interaction of these proteins in vivo is likely regulated by their subcellular localization.
Figure 6 Tim50a directly interacts with coilin, SMN and SmB'. GST-pulldown assays were conducted on bacterially purified proteins. (A) Tim50a interacts with coilin. Soluble coilin was incubated with beads loaded with GST or GST fused to Tim50a. After incubation and washes, the reactions were subject to Western blotting and probing with antibodies to coilin. The input lane accounts for 26% of the coilin used in the pulldown reactions. The bottom panel is a coomassie stained gel demonstrating the levels of GST-Tim50a (lane 2) and GST (lane 3) used in the pulldown assay. (B) Tim50a interacts with the C-terminus of coilin. Soluble TV-tagged Tim50a was incubated with beads loaded with GST or GST fused to the C-terminal 214 aa of coilin (C214, the same portion of coilin used in the two-hybrid screen). After incubation and washes, the proteins were subject to Western blotting and probing with antibodies to T7. The input lane accounts for 20% of the Tim50a used in the pulldown reactions. (C) SmB' interacts with Tim50a and mutants thereof. T7-tagged SmB' was incubated with GST, GST-CAP50, GST-Tim50a or GST-Tim50aSac, followed by washing and probing with anti-T7 antibodies. The input lane accounts for 20% of the SmB' used in the pulldown reactions. (D) SMN interacts with GST-CAP50, GST-Tim50a and GST-Tim50aSac. Soluble T7-SMN was incubated with various GST fusions, followed by Western blotting and probing with anti-T7 antibodies. The input lane accounts for 10% of the SMN used in the reactions. (E) The mitochondrial Tim50 protein interacts with the C-terminus of coilin in vitro. Soluble T7-tagged Tim50 was incubated with GST or GST-C214 beads, followed by Western blotting and probing with antibodies to T7. The input lane (lane 1) accounts for 20% of the Tim50 used in the pulldown reactions. The membrane was stained with Ponceau S to verify that equivalent amounts of GST proteins were used in the assay.
SmB' and SMN compete with coilin for Tim50a binding sites
The ability of Tim50a to bind coilin, SmB' and SMN raises the interesting possibility that this protein may regulate the dynamics of CB proteins. To test this hypothesis, we conducted competition GST-pulldown experiments using GST-Tim50aSac. This protein was used instead of GST-Tim50a (Figure 2) because it is more easily purified compared to GST-Tim50a and has similar binding capabilities (Figure 6C and 6D). For the first experiment, we tested if SMN and SmB' vie for the same binding sites on Tim50a. This was accomplished by incubating a fixed amount of GST-Tim50aSac with a constant amount of SMN and, in separate reactions, an increasing amount of SmB'. Despite binding a large amount of SmB' (Figure 7A, top panel), the level of SMN recovered by the GST-Tim50aSac beads does not change (Figure 7A, bottom panel, compare the SMN level in lane 6 to that in lane 2). We conclude that Tim50a has separate binding sites for SMN and SmB' and that one protein does not preclude the other from binding.
Figure 7 SmB' and SMN compete with coilin for Tim50a binding sites. Competition GST-pulldown experiments are shown. (A) SMN does not compete with SmB' for Tim50a binding sites. GST-Tim50aSac was incubated in separate reactions with SmB' (lane 2, top) or SMN (lane 2 bottom) followed by probing with anti-T7 (to detect SmB') and anti-SMN (to detect SMN). Lanes 3–6 show separate reactions in which a fixed amount of SMN was incubated with GST-Tim50aSac along with an increasing amount of SmB'. In lane 6, 27 times more SmB' was used in the reaction compared to the level of SMN. The input lanes account for 20% of the amounts used in all the reactions for SMN and the reactions shown in lane 2 and 3 for SmB'. (B) SmB' competes with coilin for Tim50a binding sites. GST-Tim50aSac was incubated with a fixed amount of the C-terminal coilin fragment (C214) and an increasing amount, in separate reactions, of SmB'. After Western blotting, anti-T7 was used to detect SmB' (top) and anti-coilin was used to detect C214 (bottom). Lane 4 shows the result of a reaction in which 3 times more SmB' compared to C214 was used in the binding assay with GST-Tim50aSac. Note that the level of recovered C214 is reduced in lane 4 compared to lane 2. The input lanes account for 20% of the amounts used in all the reactions for C214 and the reactions shown in lane 2 and 3 for SmB'. (C) Coilin competes with SMN for Tim50 binding sites. A competition experiment is shown wherein individual reactions were set-up containing a fixed amount of GST-Tim50aSac and SMN with increasing amounts of coilin fragment (C214). In lane 7, 27 times more coilin fragment was used in the reaction compared to the level of SMN. After SDS-PAGE and Western transfer, SMN and the coilin fragment were detected using antibodies specific to each protein. Lanes 2 and 3 show the binding of the coilin fragment or SMN to GST-Tim50aSac in the absence of competitor, respectively. The coilin fragment was added to the reactions containing GST-Tim5aSac and SMN starting at lane 4. As can been seen in lane 7, as more coilin fragment binds to GST-Tim50aSac, very little SMN is recovered. The input lane accounts for 20% of the amounts used in all the reactions for SMN and the reactions shown in lane 2 and 4 for the coilin fragment.
We then tested if coilin and SmB' compete for binding sites on Tim50a. For this experiment, we incubated a fixed amount of GST-Tim50aSac with a constant amount of a coilin fragment (C214) and, in separate reactions, an increasing amount of SmB'. As shown in Figure 7B, increasing the level of SmB' decreases the recovery of the coilin fragment (C214) by GST-Tim50aSac (bottom panel, compare the amount of C214 recovered in lane 4 to that in lane 2). Thus coilin and SmB' compete for binding sites on Tim50a.
Finally we monitored if coilin and SMN compete for the same binding sites on Tim50a. This experiment was conducted with a fixed amount of GST-Tim50aSac and a constant amount of SMN with increasing levels of the coilin fragment (C214). The binding of the respective proteins to the Tim50a substrate in the absence of competitor (Figure 7C) is shown in lane 2 (upper panel) for the coilin fragment and lane 3 (lower panel) for SMN. In reactions with both soluble proteins, as the level of coilin fragment (C214) is increased, the amount of SMN recovered by the Tim50a beads is reduced (compare the SMN signal in lane 7 to that in lane 3). These data are consistent with a competition for Tim50a binding sites by SMN and coilin. Given its protein interaction potential, we hypothesize that Tim50a may affect CB function by interacting with proteins that localize to this domain. One possible scenario is that Tim50a displaces snRNPs from the CB and facilitates their translocation to speckles. Tim50a may also disrupt the interaction between coilin and SMN, allowing for SMN to leave the CB.
Tim50a interacts with SMN and snRNPs in vivo
To corroborate the direct interaction data, we tested if Tim50a could associate with SMN in cells. For these experiments, we transfected HeLa cells with empty GFP vector, GFP-tagged full-length Tim50a or GFP fused to the N-terminal 87 aa of Tim50a. After incubation, lysates were made and subjected to immunoprecipitation with anti-GFP antibodies, followed by western blotting with anti-SMN antibodies. As shown in Figure 8A (upper panel), no SMN is co-immunoprecipitated with GFP alone but SMN is recovered by GFP-Tim50a. Only a faint SMN signal is observed in the GFP-Tim50a(N87) reaction (lane 3), suggesting that this fragment does not contribute significantly to the ability of SMN to bind Tim50a. Reprobing of this same blot with anti-GFP (bottom panel) demonstrates that while a large amount of GFP and GFP-Tim50a(N87) were immunoprecipitated, little SMN was recovered. In contrast, the amount of immunoprecipitated GFP-Tim50a is less than that obtained for GFP and GFP-Tim50a(N87), yet more SMN was recovered. Thus Tim50a can form a complex with SMN in cells.
Figure 8 Tim50a interacts with SMN and snRNPs in vivo. (A) SMN and Tim50a form a complex in vivo. HeLa cells were transfected with empty GFP vector or GFP-tagged Tim50a and Tim50aN87. Lysate was subjected to immunoprecipitation with polyclonal anti-GFP antibodies, followed by Western blotting and probing with anti-SMN antibodies (top panel). The blot was reprobed with monoclonal anti-GFP (lower panel) to visualize the level of immunoprecipitated GFP, GFP-Tim50a and GFP-Tim50aN87. (B) Tim50a and Tim50 form a complex with snRNPs. HeLa cells were transfected with empty GFP vector, GFP-tagged Tim50a or GFP-Tim50 and lysate was subjected to immunoprecipitation with anti-Sm (Y12) antibodies, followed by Western blotting and probing with anti-GFP antibodies. Immunoglobulin heavy IgG(H) chains are marked. The input lanes represent 2% of the amount of lysate used in the immunoprecipitation reactions and have been overexposed five-fold relative to the immunoprecipitation reactions in order to visualize the GFP-Tim50a signal. (C) Tim50a and Tim50 bind the Sm protein, SmB'. HeLa cells were transfected with myc-tagged Tim50a or Tim50 and lysate was incubated with beads loaded with GST or GST-SmB', followed by Western blotting and probing with anti-myc antibodies. The input lanes represent 2% of the amount of lysate used in the pulldown reactions. Approximately equal amounts of GST and GST-SmB' proteins were used.
Given that Tim50a localizes to speckles, in which snRNPs are stored, and directly interacts with a subunit of the Sm core, it is likely that Tim50a will interact with intact snRNPs. However, the in vitro binding data indicate that the mitochondrial isoform, Tim50, also has the potential to interact with snRNPs, which are known to have a cytoplasmic phase. To test this, we conducted a co-immunoprecipitation assay on HeLa cells transfected with empty GFP vector or GFP fused to Tim50a or Tim50. Lysates were generated and subject to incubation with the anti-Sm antibody Y12. After SDS-PAGE and Western transfer, the blot was probed with anti-GFP antibodies to monitor the association of GFP-Tim50a or GFP-Tim50 with snRNPs. As shown in Figure 8B, a signal corresponding to GFP-Tim50a is present in the Y12 immunoprecipitation reaction (lane 5). In contrast, less Tim50 is co-immunoprecipitated by Y12 (lane 6) compared to Tim50a. A GFP-alone signal is not observed in the Y12 reaction (lane 4), demonstrating the specificity of the reaction. Therefore, in vivo, Tim50a appears to be more efficient in snRNP interaction than Tim50, although we have not excluded the possibility that the GFP-tag may adversely affect Tim50 interactions. To address this point, Tim50a and Tim50 proteins were fused to smaller myc-tags, expressed in HeLa cells and tested for interaction with an immobilized Sm protein, SmB'. Both proteins bind GST-SmB', indicating that the GFP-tag may indeed contribute to the reduced interaction of Tim50 with snRNPs (Figure 8C). Alternatively, the faint Tim50 signal observed in the Y12 co-immunoprecipitation reaction (lane 6) may be an artifact due to the overexpression of GFP-Tim50 and subsequent nuclear localization, as shown in Figure 3. In support of this thought, immunofluorescence analysis of endogenous Tim50 demonstrates that this protein is predominantly cytoplasmic with little nuclear accumulation [36]. Given that Tim50a is a nuclear protein, we speculate that Tim50a interactions may regulate the nuclear activity of SMN and trafficking of snRNPs, possibly by altering the dynamics of the association between SMN, coilin and snRNPs. Additionally, it is plausible that cytoplasmic SMN can interact with Tim50 before mitochondrial import.
Conclusion
The cellular localization and protein interaction characteristics of Tim50a suggest that this isoform of Tim50 may play a role in snRNP biogenesis. Based on the abundance of ESTs, it is likely that Tim50, not Tim50a, is the major product of the TIMM50 gene. However, the presence of Tim50a ESTs from cancer cell lines suggests that this protein may be induced in cellular transformation. It is known that cancer cells contain CBs, while some normal cells and tissues do not [41,35]. For example, normal adult lung tissue does not contain CBs, but lung cancer cells do. Since CBs play a role in snRNP biogenesis, it is likely that the increase of CBs in transformed cells reflects on the heightened demand in splicing requirements for upregulated genes that participate in the establishment and/or maintenance of the altered cellular physiology [42]. The induction of CBs may also facilitate the production of telomerase, the RNA component of which localizes to CBs [43,44]. In addition to cancer cells neuronal cells contain robust CBs [45], possibly because high levels of snRNPs are necessary to process the abundance of messages subject to alternative splicing in neuronal tissue [46]. Therefore, Tim50a expression may be limited to tissue with high demands in snRNP metabolism. We are currently developing an antibody to the unique N-terminus of Tim50a in order to explore this possibility. Additionally, experiments are underway to verify the role of Tim50a in snRNP biogenesis. In particular, we are interesting in determining if Tim50a is involved in the liberation of snRNPs from the CB after they have been modified. The functional consequence of the interaction between Tim50a and SMN is also being investigated. Finally we note that Tim50a contains the phosphatase domain shown to be functional in the Tim50 isoform [36], and thus may alter CB activity by modifying coilin, a known phosphoprotein [45].
Methods
Yeast two-hybrid screen, plasmid construction and mutagenesis
A human brain cDNA library cloned into the prey vector pACT2 and pretransformed into the yeast strain Y187 (BD Biosciences, Palo Alto, CA) was mated with the strain PJ69-2A harboring the bait vector pAS2-l-coilin(362–576, C214) per the manufacturer's instructions. After mating, the yeast were plated onto medium lacking tryptophan, leucine, histidine and adenine (to select for bait, prey and protein interaction). Colonies were picked after several days of incubation and the prey plasmid was isolated and transformed into PJ69-2A containing pAS2-1-C214 or control bait plasmids to confirm the specificity of the interaction. Restriction digests and sequencing revealed that nine different prey cDNAs were recovered, one of which was termed CAP50. CAP50 was sequenced and the results were used to conduct a BLAST search on the NCBI website. Based on this analysis, we concluded that CAP50 contained the partial coding sequence for the protein Tim50 [36]. Several expressed sequence tags (ESTs) corresponding to Tim50 are present in the database. One EST, BE386959, contains upstream sequence information not found in the majority of Tim50 ESTs or reflected in the Tim50 protein sequence. We procured EST BE386959 and used this cDNA as a template in a PCR reaction with primers designed to amplify the Tim50a isoform (BamHI Forward primer: 5'GTCGGGATCCATGGCCTCAGCTTTATCTCT-3' and EcoRl reverse primer: 5'-TCGGAATTCGAGGCCCAGAGTTCAGGGCT-3'). The TIM50a amplified product was cloned in frame into the pEGFP-Cl vector (BD Biosciences, Palo Alto, CA) at EcoRI + BglII restriction enzyme sites. Tim50a was also cloned in frame into pGEX-2T (Amersham Pharmacia, Uppsala, Sweden) and pET28a (Novagen, Madison, WI) vectors at EcoRI + BamHI sites and utilized for bacterial protein expression. Various mutants of TIM50a were produced from specific restriction enzyme sites that were located within the TIM50a coding sequence or in pEGFP-C1. These constructs included TIM50aN87 (SmaI deletion), TIM50aSac (SacI internal deletion) and TIM50aN135 (ApaI deletion). Also, CAP50 obtained from the yeast two-hybrid screen was digested with BamHI + BglII. and cloned in frame into pGEX-2T at a BamHI site and utilized for bacterial protein expression, or digested with BamHI + EcoRI and cloned in frame into pEGFP-Cl at the equivalent sites for transient transfection experiments. The mitochondrial Tim50 was cloned into various expression vectors by standard techniques.
PCR and nested PCR of Tim50 isoforms on cancer cell line cDNA
PCR reactions were performed in the DNA Engine PTC-200 Peltier Thermal Cycler (MJ Research, Watertown, MA). The cDNA from four cancer cell lines were kindly provided by Dr. Laree Hiser (The University of Mississippi, Medical Center): NIH: OVCAR-3 (ovarian cancer), HCT-15 (colon cancer), A-549 (non-small cell lung cancer) and MDA-MB-231 (breast cancer). These cDNA were shown by Dr. Hiser to be free of genomic DNA. Two sets of primers were used in the PCR reactions, the sequence of which can be provided upon request. The first set (ComFor+ComRev) was designed to amplify a region common to both Tim50a and Tim50, thus providing a method to monitor the total expression level of the TIMM50 gene. The second set of primers (50aFor+50aRev or 50anRev) bind specifically to Tim50a and not Tim50, thereby allowing the level of the Tim50a transcript to be assessed. A standard reaction, consisting of 30 cycles of amplification and the ComFor + ComRev primers, was used to amplify the common region of Tim50/Tim50a from cDNA. Plasmid containing the Tim50a cDNA (10 ng) and water served as positive and negative controls, respectively. To observe the Tim50a isoform, nested PCR had to be conducted. For these reactions, 30 cycles of amplification using the 50aFor+50aRev primers was conducted using the cDNA as template in a 25 μl reaction. Water and 10 ng of Tim50a plasmid served as negative and positive controls, respectively. 0.1 μl of each product, including the negative and positive controls, was then used as a template in a 25 μl reaction with a second pair of primers (50aFor+50anRev) for 20 cycles of amplification. The resultant products were cloned and sequenced to verify that they were derived from the Tim50a isoform. Nested PCR using NIH:OVCAR-3 cDNA consistently yielded two products. Sequencing demonstrated that the upper product was from Tim50a while the lower product was from an unrelated message with partial homology to the primers used in the nested PCR reaction.
Cell Culture, Transfection, Immunofluorescence, and Immunoprecipitation
HeLa cells from the American Type Culture Collection (ATCC) were cultured, transfected with Superfect (Qiagen, Valencia, CA) and processed as described [28,29,47]. Where indicated, actinomycin D was used at 5 μg/ml for two hours. For immunofluorescence, cells were grown on chamberslides and fixed with 4% paraformaldehyde followed by the permeabilization with 0.5% Triton X-100. The permeable cells were then blocked with 10% normal goat serum for 30 min and probed with the corresponding primary and secondary antibodies. For immunoprecipitation, whole cell lysate was obtained from HeLa cells transfected with specific constructs. Cells were lysed in 50 mM Tris, pH 7.5, 150 mM NaCl, 1% NP-40,1% sodium deoxycholate, 0.1% SDS, 2 mM EDTA plus 1 tablet of complete protease inhibitor cocktail (Roche, Mannheim, Germany) per 50 ml lysis solution. The lysates were then repeatedly passed through a 25-gauge needle to shear DNA. The cell lysate was incubated at 4°C for one hour with primary antibody, followed by the addition of Sepharose beads (Amersham Pharmacia, Uppsala, Sweden) and an additional one hour of incubation. Both incubation periods were accompanied with gentle rotation. The beads were washed with 1 ml lysis solution 5 times, resuspended in 5X SDS loading buffer, boiled, and subject to SDS-PAGE and Western blotting as described [47]. Antibodies used for this experimentation include anti-SMN (BD Biosciences, Palo Alto, CA, catalog number 610646), anti-GFP (polyclonal, BD Biosciences, Palo Alto, CA, catalog number 632459; monoclonal, Roche, Indianapolis, IN, catalog number 1814460) and anti-Sm (Y12) (Lab Vision, Freemont, CA, catalog number MS-450-P).
In vitro binding assays
GST-tagged and His-tagged constructs, after transformation into E. coil BL21(DE3)pLysS cells, were induced and purified as described [28]. In binding reactions, approximately 1 μg of His-T7-tagged protein was incubated with 1 μg of the GST-fusion protein in 1 ml of lysis buffer plus 2 mM DTT. After incubation for 1 hour at 4°C with gentle inversion, the beads were washed 5 times (1 ml each) with lysis buffer plus DTT, resuspended in 10 μl 5X SDS loading buffer, boiled and subjected to SDS-PAGE. Primary antibodies used included anti-T7 (Novagen, Madison, WI, 1:1000), anti-SMN (described above) and anti-coilin [48], 1:500). Competition experiments were performed as described [28], except that increasing amounts of SmB' or coilin fragment were used.
Authors' contributions
HX conducted binding assays and made DNA constructs. ZBS did cell localization, immunoprecipitation studies and competition binding assays. MRII conducted some competition binding assays. MDH conceived the experiments, coordinated the study and wrote the paper.
Acknowledgements
We thank Kecia Grant for her excellent technical assistance and Dr. Greg Matera (Case Western Reserve University, Cleveland, OH) for the anti-coilin antibodies. We also thank Dr. Laree Hiser (The University of Mississippi, Medical Center) for the cancer cell line cDNAs and Drs. Archa Fox and Angus Lamond (University of Dundee, Scotland) for their kind gift of anti-PSPl antibodies. This work was supported by a grant from the Muscular Dystrophy Association.
==== Refs
Will CL Luhrmann R Spliceosomal UsnRNP biogenesis, structure and function Curr Opin Cell Biol 2001 13 290 301 11343899 10.1016/S0955-0674(00)00211-8
Carmo-Fonseca M New clues to the function of the Cajal body EMBO Rep 2002 3 726 7 12151329 10.1093/embo-reports/kvf154
Gerbi SA Borovjagin AV Lange TS The nucleolus: a site of ribonucleoprotein maturation Curr Opin Cell Biol 2003 15 318 25 12787774 10.1016/S0955-0674(03)00049-8
Meister G Eggert C Fischer U SMN-mediated assembly of RNPs: a complex story Trends Cell Biol 2002 12 472 8 12441251 10.1016/S0962-8924(02)02371-1
Massenet S Pellizzoni L Paushkin S Mattaj IW Dreyfuss G The SMN complex is associated with snRNPs throughout their cytoplasmic assembly pathway Mol Cell Biol 2002 22 6533 41 12192051 10.1128/MCB.22.18.6533-6541.2002
Narayanan U Ospina JK Frey MR Hebert MD Matera AG SMN, the spinal muscular atrophy protein, forms a pre-import snRNP complex with snurportinl and importin beta Hum Mol Genet 2002 11 1785 95 12095920 10.1093/hmg/11.15.1785
Mouaikel J Narayanan U Verheggen C Matera AG Bertrand E Tazi J Bordonne R Interaction between the small-nuclear-RNA cap hypermethylase and the spinal muscular atrophy protein, survival of motor neuron EMBO Rep 2003 4 616 22 12776181 10.1038/sj.embor.embor863
Narayanan U Achsel T Luhrmann R Matera AG Coupled in vitro import of U snRNPs and SMN, the spinal muscular atrophy protein Mol Cell 2004 16 223 34 15494309 10.1016/j.molcel.2004.09.024
Sleeman JE Lamond AI Newly assembled snRNPs associate with coiled bodies before speckles, suggesting a nuclear snRNP maturation pathway Curr Biol 1999 9 1065 74 10531003 10.1016/S0960-9822(99)80475-8
Darzacq X Jady BE Verheggen C Kiss AM Bertrand E Kiss T Cajal body-specific small nuclear RNAs: a novel class of 2'-O-methylation and pseudouridylation guide RNAs EMBO J 2002 21 2746 56 12032087 10.1093/emboj/21.11.2746
Verheggen C Lafontaine DL Samarsky D Mouaikel J Blanchard JM Bordonne R Bertrand E Mammalian and yeast U3 snoRNPs are matured in specific and related nuclear compartments EMBO J 2002 21 2736 45 12032086 10.1093/emboj/21.11.2736
Jady BE Darzacq X Tucker KE Matera AG Bertrand E Kiss T Modification of Sm small nuclear RNAs occurs in the nucleoplasmic Cajal body following import from the cytoplasm EMBO J 2003 22 1878 88 12682020 10.1093/emboj/cdg187
Yu YT Shu MD Steitz JA Modifications of U2 snRNA are required for snRNP assembly and pre-mRNA splicing EMBO J 1998 17 5783 95 9755178 10.1093/emboj/17.19.5783
Lamond AI Sleeman JE Nuclear substructure and dynamics Curr Biol 2003 13 R825 8 14588256 10.1016/j.cub.2003.10.012
Muller B Schumperli D The U7 snRNP and the hairpin binding protein: Key players in histone mRNA metabolism Semin Cell Dev Biol 1997 8 567 76 9642171 10.1006/scdb.1997.0182
Dominski Z Marzluff WF Formation of the 3' end of histone mRNA Gene 1999 239 1 14 10571029 10.1016/S0378-1119(99)00367-4
Schumperli D Pillai RS The special Sm core structure of the U7 snRNP: far-reaching significance of a small nuclear ribonucleoprotein Cell Mol Life Sci 2004 61 2560 70 15526162 10.1007/s00018-004-4190-0
Stefanovic B Hackl W Luhrmann R Schumperli D Assembly, nuclear import and function of U7 snRNPs studied by microinjection of synthetic U7 RNA into Xenopus oocytes Nucleic Acids Res 1995 23 3141 51 7667090
Pillai RS Will CL Luhrmann R Schumperli D Muller B Purified U7 snRNPs lack the Sm proteins D1and D2 but contain Lsm10, a new 14 kDa Sm D1-like protein EMBO J 2001 20 5470 9 11574479 10.1093/emboj/20.19.5470
Pillai RS Grimmler M Meister G Will CL Luhrmann R Fischer U Schumperli D Unique Sm core structure of U7 snRNPs: assembly by a specialized SMN complex and the role of a new component, Lsm11, in histone RNA processing Genes Dev 2003 17 2321 33 12975319 10.1101/gad.274403
Frey MR Matera AG Coiled Bodies Contain U7 Small Nuclear RNA and Associate with Specific DNA Sequences in Interphase Cells Proc Natl Acad Sci USA 1995 92 5915 5919 7597053
Platani M Goldberg I Lamond AI Swedlow JR Cajal body dynamics and association with chromatin are ATP-dependent Nat Cell Biol 2002 4 502 8 12068306 10.1038/ncb809
Platani M Goldberg I Swedlow JR Lamond AI In vivo analysis of Cajal body movement, separation, and joining in live human cells J Cell Biol 2000 151 1561 74 11134083 10.1083/jcb.151.7.1561
Frey MR Matera AG RNA-mediated interaction of Cajal bodies and U2 snRNA genes J Cell Biol 2001 154 499 509 11489914 10.1083/jcb.200105084
Matera AG RNA splicing: More clues from Spinal Muscular Atrophy Curr Biol 1999 9 R140 R142 10074419 10.1016/S0960-9822(99)80083-9
Bauer DW Gall JG Coiled bodies without coilin Mol Biol Cell 1997 8 73 82 9017596
Tucker KE Berciano MT Jacobs EY LePage DF Shpargel KB Rossire JJ Chan EK Lafarga M Conlon RA Matera AG Residual Cajal bodies in coilin knockout mice fail to recruit Sm snRNPs and SMN, the spinal muscular atrophy gene product J Cell Biol 2001 154 293 307 11470819 10.1083/jcb.200104083
Hebert MD Szymczyk PW Shpargel KB Matera AG Coilin forms the bridge between Cajal bodies and SMN, the spinal muscular atrophy protein Genes Dev 2001 15 2720 9 11641277 10.1101/gad.908401
Hebert MD Shpargel KB Ospina JK Tucker KE Matera AG Coilin methylation regulates nuclear body formation Dev Cell 2002 3 329 37 12361597 10.1016/S1534-5807(02)00222-8
Boisvert FM Cote J Boulanger MC Cleroux P Bachand F Autexier C Richard S Symmetrical dimethylarginine methylation is required for the localization of SMN in Cajal bodies and pre-mRNA splicing J Cell Biol 2002 159 957 69 12486110 10.1083/jcb.200207028
Friesen WJ Massenet S Paushkin S Wyce A Dreyfuss G SMN, the product of the spinal muscular atrophy gene, binds preferentially to dimethylarginine-containing protein targets Mol Cell 2001 7 1111 7 11389857 10.1016/S1097-2765(01)00244-1
Brahms H Meheus L de Brabandere V Fischer U Luhrmann R Symmetrical dimethylation of arginine residues in spliceosomal Sm protein B/B' and the Sm-like protein LSm4, and their interaction with the SMN protein RNA 2001 7 1531 42 11720283 10.1017/S135583820101442X
Matera AG Nuclear bodies: multifaceted subdomains of the interchromatin space Trends Cell Biol 1999 9 302 9 10407409 10.1016/S0962-8924(99)01606-2
Liu Q Dreyfuss G A novel nuclear structure containing the survival of motor neurons protein EMBO J 1996 15 3555 3565 8670859
Young PJ Le TT Dunckley M Nguyen TM Burghes AH Morris GE Nuclear gems and Cajal (coiled) bodies in fetal tissues: nucleolar distribution of the spinal muscular atrophy protein, SMN Exp Cell Res 2001 265 252 61 11302690 10.1006/excr.2001.5186
Guo Y Cheong N Zhang Z De Rose R Deng Y Farber SA Fernandes-Alnemri T Alnemri ES Tim50, a component of the mitochondrial translocator, regulates mitochondrial integrity and cell death J Biol Chem 2004 279 24813 25 15044455 10.1074/jbc.M402049200
Dingwall C Laskey R Nuclear targeting sequences – a consensus? Trends Biochem Sci 1991 16 478 481 1664152 10.1016/0968-0004(91)90184-W
Lam YW Lyon CE Lamond AI Large-scale isolation of Cajal bodies from HeLa cells Mol Biol Cell 2002 13 2461 73 12134083 10.1091/mbc.02-03-0034
Carmo-Fonseca M Pepperkok R Carvalho MT Lamond AI Transcription-dependent colocalization of the U1, U2, U4/U6 and U5 snRNPs in coiled bodies J Cell Biol 1992 117 1 14 1532583 10.1083/jcb.117.1.1
Fox AH Lam YW Leung AK Lyon CE Andersen J Mann M Lamond AI Paraspeckles: a novel nuclear domain Curr Biol 2002 12 13 25 11790299 10.1016/S0960-9822(01)00632-7
Spector DL Lark G Huang S Differences in snRNP localization between transformed and nontransformed cells Mol Biol Cell 1992 3 555 569 1535243
Mercatante DR Kole R Control of alternative splicing by antisense oligonucleotides as a potential chemotherapy: effects on gene expression Biochim Biophys Acta 2002 1587 126 32 12084454
Zhu Y Tomlinson RL Lukowiak AA Terns RM Terns MP Telomerase RNA accumulates in Cajal bodies in human cancer cells Mol Biol Cell 2004 15 81 90 14528011 10.1091/mbc.E03-07-0525
Jady BE Bertrand E Kiss T Human telomerase RNA and box H/ACA scaRNAs share a common Cajal body-specific localization signal J Cell Biol 2004 164 647 52 14981093 10.1083/jcb.200310138
Gall JG The centennial of the Cajal body Nat Rev Mol Cell Biol 2003 4 975 80 14685175 10.1038/nrm1262
Grabowski PJ Black DL Alternative RNA splicing in the nervous system Prog Neurobiol 2001 65 289 308 11473790 10.1016/S0301-0082(01)00007-7
Hebert MD Matera AG Self-association of coilin reveals a common theme in nuclear body localization Mol Biol Cell 2000 11 4159 71 11102515
Almeida F Saffrich R Ansorge W Carmo-Fonseca M Microinjection of anti-coilin antibodies affects the structure of coiled bodies J Cell Biol 1998 142 899 912 9722604 10.1083/jcb.142.4.899
| 16008839 | PMC1177934 | CC BY | 2021-01-04 16:31:30 | no | BMC Cell Biol. 2005 Jul 11; 6:29 | utf-8 | BMC Cell Biol | 2,005 | 10.1186/1471-2121-6-29 | oa_comm |
==== Front
BMC Clin PatholBMC Clinical Pathology1472-6890BioMed Central London 1472-6890-5-51594148210.1186/1472-6890-5-5Case ReportUnexpected combination of acute croup and myocarditis: Case report Briassoulis George [email protected] Athina [email protected] Emmanuel [email protected] John [email protected] Pediatric Intensive Care Unit, "P & A Kyriakou" Children's Hospital, Athens, Greece2 Paediatric Intensive Care Unit, University Hospital of Heraklion, Crete, Greece3 Microbiology and Transfusion Departments, NIMTS Hospital, Athens, Greece4 Department of Pathology, School of Medicine, National University of Athens, Greece2005 7 6 2005 5 5 5 13 12 2004 7 6 2005 Copyright © 2005 Briassoulis et al; licensee BioMed Central Ltd.2005Briassoulis 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
Lower vaccination coverage among foreign-born children is of concern because they live in households and communities characterized by more intense exposure to infectious diseases. Because of their higher prevalence rates, there is an increasing occurrence of infectious diseases imported into developed countries. This case report emphasizes the emerging necessity for new clinicians and pathologists of having competence with old infectious disease pathology.
Case presentation
A three and a half year old girl, who presented with croup history of 5 days and has been in severe respiratory distress, was admitted to the Pediatric Intensive Care Unit in shock and acute respiratory failure. The patient was immediately intubated, and a grayish nonadherent membrane extending through the glottis down into the larynx was apparent during the procedure. Echocardiographic findings, which were consistent with acute myocarditis, confirmed poor left ventricular contractility despite escalating high doses of inotropes. Autopsy showed numerous strains of toxigenic corynobacterium diphtheriae, which also grew on the Loeffler cultures of membranes received during the intubation.
Conclusion
It is critical that new generations of clinicians and bio-pathologists not only be trained in the subspecialty of infectious disease pathology, but that they also be willing participants in the diagnosis and investigation of infectious diseases.
==== Body
Background
Diphtheria, caused by corynebacterium diiphtheriae, is acquired by contact with either a carrier or a person with the disease. Before the discovery of antitoxin at the turn of the 20th century, the "straggling angel of children", as diphtheria once was called, was a significant cause of mortality in children and adults. In severe cases, respiratory and circulatory collapse may occur. Complications secondary to elaborated diphtheritic toxin may affect any system, but myocarditis and nervous system involvement are most characteristic. The incidence of diphtheria declined markedly after the extensive use of diphtheria toxoid after World War II.
Inadequate immunization cover is deemed responsible for the continued menace of diphtheria in developing countries [1]. A large number of immigrants have come to Greece and other western countries from diphtheria-endemic countries during the past 10 years. Lower vaccination coverage among foreign-born children is of concern because foreign-born children often live in households and communities characterized by more intense exposure to infectious diseases, and many originate from countries with much higher prevalence rates of diphtheria and other diseases than the developed countries [2]. Additionally, the majority (65%) of internationally adopted children have no written records of overseas immunizations [3].
The increasing occurrence of infectious diseases imported into the United States and other nations, including human immunodeficiency virus-1 group O, dengue fever, tuberculosis, malaria, diphtheria and cholera in immigrants and travelers, and Ebola virus in nonhuman primates, emphasizes the necessity for clinicians and pathologists of having competence with infectious disease pathology [4]. It is critical that new generations of intensivists and bio-pathologists not only be trained in the subspecialty of infectious disease pathology, but that they also be willing participants in the diagnosis and investigation of infectious diseases. We describe a case of acute diphtheritic laryngotracheobronchitis and myocarditis in an immigrant girl, the first diagnosed in Greece during the last 30 years.
Case presentation
A three and a half year old girl, who presented with a croup history of 4 days duration, was transferred to the Pediatric Intensive Care Unit because of worsening respiratory distress. The past history was uncontributory except that the patient was an unvaccinated immigrant, living in close contact with her alcoholic grandfather. The patient was immediately intubated, and a grayish nonadherent membrane extending through the glottis down into the larynx was apparent during the procedure. Material from the membranes was sent for cultures and the patient was started on intravenous penicillin and specific antitoxin. Although blood gases improved initially, respirator settings had soon to be increased significantly to sustain a pO2 value above 60 mm Hg (7.9 kPa). A chest roentgenogram revealed extensive bilateral diffuse pulmonary infiltrations (Fig. 1). Laboratory findings included white blood cells 70,110 cu mm (70*109/L) with segmented neutrophils 46%, band forms 8%, platelets 134,000 cu mm (134*109/L), CRP 177. Serum alanine aminotransferase and aspartate aminotransferase concentrations peaked at 340 U/L and 721 U/L, respectively and serum creatinine increased to 1.4 mg/dL (124 μmol/L). On hospital day 1 she became hypotensive and remained in shock despite escalating doses of inotropic agents. Echocardiographic findings, which were consistent with acute cardiomyopathy, confirmed extremely poor left ventricular contractility (asystole; shortening fraction <5%), despite high dose continuous dobutamine (20 μg.kg-1.min-1) and epinephrine (up to 5 μg.kg-1.min-1) infusions. After having completed 36 hours of hospitalization in the PICU with multiple cardiac arrests, the child was no more responding to resuscitation efforts and died.
Figure 1 Chest roentgenogram upon admission. Frontal chest roentgenogram revealed extensive bilateral diffuse pulmonary infiltrations, left greater than right.
Autopsy findings revealed numerous diphtheritic membrane formations extended down into the larynx and further obstructing the distal tracheobronchial tree (Fig. 2). Myocardium edema, congestion and mononuclear and neutrophil cell infiltration, completed the clinicopathologic picture of diphtheritic cardiomyopathy. Membrane formations and visceral hemorrhagic necrotic lesions were associated with numerous strains of corynobacterium diphtheriae (producing the typical club-shaped appearance and distributed in a typical Chinese-letter appearance on the smear) (Fig. 3). The organism, however, of Gravis type, grew on the Loeffler cultures of portions of membranes, which were received during the intubation. As also expected, the strain was shown to be toxigenic.
Figure 2 Tracheobronchial membranes at autopsy. Autopsy findings revealed numerous diphtheritic membrane formations (arrowheads) extended down into the larynx and further obstructing the distal tracheobronchial tree.
Figure 3 Strains of corynobacterium diphtheriae on a smear. A Giemsa stain (original magnification × 1000) of visceral hemorrhagic necrotic lesions was associated with numerous strains of corynobacterium diphthetiae, producing the typical club-shaped appearance and distributed in a typical Chinese-letter appearance on the smear (arrows).
Conclusion
This manuscript reports on an exceptionally rare disease – almost forgotten today – in a western country. Most of the recent literature concerning clinical manifestations and treatment modalities of diphtheritic croup [5,6] or myocarditis [7,8] almost exclusively comes from endemic areas in developing countries [9,10]. Recently, however, concerns regarding the emergence of this obsolete disease in the developed word are reported with an increasing frequency, mostly as quizzes [11] exceptional [12] or unusual case reports [13,14] – like ours – or as alarming updated review articles of progress in clinical, epidemiological and microbiological aspects of diphtheria in the European region [15]. Researchers concern with the emerging threat of diphtheria being re-introduced from Eastern Europe or Asia and re-established in the West, highlight the need for improved immunisation coverage, surveillance and epidemiological studies to sustain control of diphtheria in European Region [16], and overemphasize the need for keeping a high index of clinical suspicion and initiation treatment without delay [17].
Airway obstruction is the most common cause of death among culture-confirmed cases of diphtheria involving the respiratory tract [18]. Although diphtheritic membranes in the larynx, dyspnea and leukocytosis are all poor prognostic signs, tracheobronchial membrane is not usually identified before death in patients with respiratory tract diphtheria. In a recent study examining clinical features and predictors of diphtheritic cardiomyopathy in Vietnamese children, it was shown that fatal outcome was best predicted by the combination of myocarditis on admission and a pseudomembrane score of >2 [7]. Similarly, in Indian children, the extension of membrane formation to two or more sites was a highly sensitive predictor of mortality, whereas a total leukocyte count > 25,000 cu mm (25*109/L) had a high specificity and positive predictive value while serum alanine aminotransferase levels of > 80 U/L had high sensitivity and a negative predictive value [19], findings which were verified in our case report. Cardiomyopathy and neuropathy occur frequently, complicating severe diphtheria, but in most instances the cardiac manifestations appear between the second and the third week of the disease. Occasionally, myocarditis, as in our case, and neuritis may be noted as early as the first week of the disease [20]. Because diphtheritic antitoxin cannot neutralize penetrating toxin or toxin absorbed to cells, its efficacy decreases as the duration of pharyngeal diphtheria increases. Therefore, the late provision of specific treatment after the 4th day is associated with a 20-fold increase in mortality rate [21]. Nowadays, reports from endemic areas offer new scientific information on current treatment modalities for obsolete diseases. A Russian study showed that patients with severe diphtheritic myocarditis had a high risk of lowering of left ventricular ejection fraction below 35%, akinesia of left ventricular segments (asinergia index > 2) and death [22]. Interestingly, temporary insertion of a cardiac pacemaker in these patients has been shown to improve outcome [23].
Diphtheria caused by the gravis strain is usually more toxigenic and carries a poorer prognosis. Although adequate immunization rate has been suggested to result in a form of herd immunity by dampening the survival of toxigenic strains of corynebacterium diphtheriae, the gene for virulence can be transferred by lysogeny to nontoxigenic strains to confer the toxigenic state among people who are not immune [24].
Diphtheria has been regarded eliminated in Greece like in most Western European countries after the start of mass immunizations 3 decades ago, and no indigenous case was reported since 1970s. Meanwhile, it seems that immunologic status of population has been changed. Most recent diphtheria outbreaks in western countries have increasingly involved alcoholic urban adults and cutaneous infection has been recognized as an important epidemiologic feature of such modern outbreaks [25]. That accorded with the continuing decline of diphtheria antitoxin titter after vaccination with protective levels (>0.1 IU/mL) detected in only 15% of those more than 40 years of age [26]. Thus, the effect of massive vaccination routines is of particular concern now, because, although the incidence of diphtheria in general has been decreased, the mortality from toxic diphtheria remains high, especially between unvaccinated persons [27]. For that reason, regular booster doses with the combined tetanus-diphtheria toxoid should be routinely given at mid-decade ages and whenever tetanus toxoid is indicated. Moreover, since it is more than obvious now that immigration of nonimmunized, impoverished populations has already provided a continuing pool of susceptible persons, prompt massive immunization campaign should be urgently aimed at such unimmunized groups.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
GB conceived of the study, participated in its design and coordination, and drafted the manuscript. JP has made substantial contributions to conception. TA helped to draft the manuscript. EA participated in the technical procedures, collection of tissue samples, immunohistochemistry, special staining and cultures. All authors read and approved the final manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
==== Refs
Nandi R De M Browning S Purkayastha P Bhattacharjee AK Diphtheria: the patch remains J Laryngol Otol 2003 117 807 810 14653924 10.1258/002221503770716250
Strine TW Barker LE Mokdad AH Luman ET Sutter RW Chu SY Vaccination coverage of foreign-born children 19 to 35 months of age: findings from the National Immunization Survey, 1999–2000 Pediatrics 2002 110 e15 12165614 10.1542/peds.110.2.e15
Schulte JM Maloney S Aronson J San Gabriel P Zhou J Saiman L Evaluating acceptability and completeness of overseas immunization records of internationally adopted children Pediatrics 2002 109 E22 11826232 10.1542/peds.109.2.e22
Schwartz DA Emerging and reemerging infections. Progress and challenges in the subspecialty of infectious disease pathology Arch Pathol Lab Med 1997 121 776 784 9278604
Havaldar PV Dexamethasone in laryngeal diphtheritic croup Ann Trop Paediatr 1997 17 21 23 9176573
Grigorev GM Kortev AI Shemiakina EK Shmatova NK [Diagnosis of pharyngeal and laryngeal diphtheria in adults] Vestn Otorinolaringol 1995 2 31 3 [Article in Russian] 7785153
Kneen R Nguyen MD Solomon T Pham NG Parry CM Nguyen TT Ha TL Taylor A Vo TT Nguyen TT Day NP White NJ Clinical features and predictors of diphtheritic cardiomyopathy in Vietnamese children Clin Infect Dis 2004 39 1591 1598 15578357 10.1086/425305
Perles Z Nir A Cohen E Bashary A Engelhard D Atrioventricular block in a toxic child: do not forget diphtheria Pediatr Cardiol 2000 21 282 283 10818196 10.1007/s002460010062
Lodha R Dash NR Kapil A Kabra SK Diphtheria in urban slums in north India Lancet 2000 355 204 10675127 10.1016/S0140-6736(99)04847-3
Kadirova R Kartoglu HU Strebel PM Clinical characteristics and management of 676 hospitalized diphtheria cases, Kyrgyz Republic, 1995 J Infect Dis 2000 181 S110 115 10657201 10.1086/315549
Stewart CE Not just a sore throat.... Emerg Med Serv 2000 29 56, 59 66 quiz 112 11183096
McCluney NA McKerrow WS Should we be concerned about diphtheria in the UK? Surgeon 2004 2 234 235 15570833
Lumio J Olander RM Groundstroem K Suomalainen P Honkanen T Vuopio-Varkila J Epidemiology of three cases of severe diphtheria in Finnish patients with low antitoxin antibody levels Eur J Clin Microbiol Infect Dis 2001 20 705 710 11757971 10.1007/s100960100594
Stastny B Leonhardt-Horti H Hagel KJ Schranz D [Fulminant diphtheria myocarditis in an unvaccinated preschool child] Klin Padiatr 1999 211 462 464 [Article in German] 10592928
De Zoysa A Efstratiou A Eighth International Meeting of the European Laboratory Working Group on Diphtheria and the Diphtheria Surveillance Network – June 2004: Progress is needed to sustain control of diphtheria in European Region Euro Surveill 2004 9 [Epub ahead of print]
Bader HM Egler P [Immunisation coverage in the adult workforce 2003. Utilisation of routine occupational health checks to ascertain vaccination coverage in employees.] Bundesgesundheitsblatt Gesundheitsforschung Gesundheitsschutz 2004 47 1204 1215 [Article in German] 15583892 10.1007/s00103-004-0942-0
Butenandt O Weiss M [Diphtheria. An illness, which many physicians no longer see in personal experience] Fortschr Med 1996 114 281 283 [Article in German] 8974967
Dobie RA Tobey DN Clinical features of diphtheria in the respiratory tract JAMA 1979 229 2197 2201 490806 10.1001/jama.242.20.2197
Havaldar PV Sankpal MN Doddannavar RP Diphtheritic myocarditis: clinical and laboratory parameters of prognosis and fatal outcome Ann Trop Paediatr 2000 20 209 215 11064774
Munford RS Ory HW Brooks GF Feldman RA Diphtheria deaths in the United States, 1959–1970 JAMA 1974 229 1890 1893 4213033 10.1001/jama.229.14.1890
Feigin RD Stechenberg BW Feigin RD, Cherry JD Diphtheria Textbook of Pediatric Infectious Diseases 1987 2 Philadelphia: WB Saunders Co 1134 1137
Ordian AM Iushchuk ND Karetkina GN Golikov AP Gazarian GA [Clinical and prognostic significance of disturbed global and regional contractility of left ventricle in diphtheria myocarditis] Klin Med (Mosk) 2000 78 20 23 [Article in Russian] 10723145
Dung NM Kneen R Kiem N Bethell DB Phu NH Solomon T Chau TT Mai NT Day NP White NJ Treatment of severe diphtheritic myocarditis by temporary insertion of a cardiac pacemaker Clin Infect Dis 2002 35 1425 1429 12439808 10.1086/344176
Editorial Diphtheria outbreaks in immunized populations N Engl J Med 1988 318 41 42 3336382
Harnisch JP Tronca E Nolan CM Diphtheria among alcoholic urban adults Ann Int Med 1989 111 71 82 2472081
Christenson B Bottiger M Serologic immunity to diphtheria in Sweden in 1978 and 1984 Scan J Infect Dis 1986 18 227 233
Karzon D Edwards K Diphtheria outbreaks in immunized populations N Engl J Med 1988 318 41 43 3336382
| 15941482 | PMC1177935 | CC BY | 2021-01-04 16:27:55 | no | BMC Clin Pathol. 2005 Jun 7; 5:5 | utf-8 | BMC Clin Pathol | 2,005 | 10.1186/1472-6890-5-5 | oa_comm |
==== Front
BMC Dev BiolBMC Developmental Biology1471-213XBioMed Central London 1471-213X-5-101592705110.1186/1471-213X-5-10Research ArticleChanges in gravitational force affect gene expression in developing organ systems at different developmental times Shimada Naoko [email protected] Gbolabo [email protected] Stephen J [email protected] Robert Wood Johnson Medical School, Department of Neuroscience and Cell Biology, 675 Hoes Lane, Piscataway, NJ 08854, USA2005 31 5 2005 5 10 10 9 3 2005 31 5 2005 Copyright © 2005 Shimada et al; licensee BioMed Central Ltd.2005Shimada 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
Little is known about the affect of microgravity on gene expression, particularly in vivo during embryonic development. Using transgenic zebrafish that express the gfp gene under the influence of a β-actin promoter, we examined the affect of simulated-microgravity on GFP expression in the heart, notochord, eye, somites, and rohon beard neurons. We exposed transgenic zebrafish to simulated-microgravity for different durations at a variety of developmental times in an attempt to determine periods of susceptibility for the different developing organ systems.
Results
The developing heart had a period of maximum susceptibility between 32 and 56 hours after fertilization when there was an approximately 30% increase in gene expression. The notochord, eye, somites, and rohon beard neurons all showed periods of susceptibility occurring between 24 and 72 hours after fertilization. In addition, the notochord showed a second period of susceptibility between 8 and 32 hours after fertilization. Interestingly, all organs appeared to be recovering by 80 hours after fertilization despite continued exposure to simulated-microgravity.
Conclusion
These results support the idea that exposure to microgravity can cause changes in gene expression in a variety of developing organ systems in live embryos and that there are periods of maximum susceptibility to the effects.
==== Body
Background
Research has shown that the space environment can alter gene expression, cell structure and function, organ systems and the behavior of organisms. While some of these changes may constitute adaptation to the space environment, we do not have a comprehensive understanding of the mechanisms driving these changes, nor are we wholly aware of the changes themselves. This is especially true with respect to development. The zebrafish is a powerful model for studying vertebrate development. We have previously demonstrated that zebrafish embryos can be sustained for at least 5 days in simulated-microgravity using a bioreactor that NASA designed for cells in culture [1,2]. Using this bioreactor, we have begun to examine the effect of simulated-microgravity on gene expression in specific developing organ systems in vivo using transgenic zebrafish that express the green fluorescent protein gene (gfp) under the influence of different promoters [3]. The use of gfp as a 'reporter-gene' has two significant advantages; gene expression can be monitored directly in a live animal, and changes in morphology that might be due to changes in expression of other genes can easily be detected.
We have been judicious in our choice of transgenic fish. β-actin is an ubiquitous cytoskeletal protein and β-actin gene expression is not thought to be affected by microgravity in cells in culture. However, the experiments that were done on cells in culture involved microarray analysis. It is not clear that changes in expression levels that we have documented in zebrafish embryos [3] and others have documented in cultured endothelial cells in simulated-microgravity [4] would have been detected using microarrays because the changes would not meet the 1–2 fold change threshold established for significance using microarrays. It is possible that there were changes in β-actin gene expression in the cultured cells of the same magnitude that we have seen in zebrafish embryos. Those changes would not have been deemed significant.
With the advent of bioinformatics, there has been a natural tendency to try to maximize the amount of data from individual experiments performed in Space using microarray technology. However, gene microarrays and northern blots can give misleading information, especially if they are done using materials from whole embryos, because changes in minor organs might be masked. For instance, a dramatic increase in β-actin gene expression exclusively in the hypochord, might not be detected because it would be averaged out by the lack of any significant change in any other structure. Using fluorescence imaging of whole embryos and a gfp-reporter gene approach, changes in gene expression should be detectable regardless of the size of the organ. In addition, data from microarrays and northern blots must be normalized against the expression level for genes that we know are not influenced by the experimental conditions to yield quantitative measures of expression changes. However, we cannot predict which genes are not affected by changes in gravitational force in vivo. Therefore, the choice of 'housekeeping gene' for normalization of microarray data could bias the results. For example, it is not unusual to use β-actin expression for normalization. Clearly this would not have been a good choice if the present experiments had been done using either microarrays or northern blots.
In our previous study heart and notochord development were examined at a single developmental time point [3]. Heart development was examined because previous studies had clearly demonstrated that the adult cardiovascular system is affected by changes in gravitational force [5-7]. However, the effects of microgravity on a developing heart are still poorly understood. The notochord, in most vertebrates, has adult derivatives but as a structure is present only during development. The notochord has been implicated in patterning the developing vertebral column [8], spinal cord [9], somites [10], and gut [11] among other structures. Despite its importance in patterning developing organ systems, the effects of microgravity on the notochord remain unknown. We demonstrated that simulated microgravity had effects on gene expression in the heart and notochord [3] but because of the limited extent of the study, we were not able to determine whether there were specific developmental periods when the heart or notochord were differentially susceptible to the effects.
We have now extended our previous work to include both a greater range of developmental time frames and an analysis of effects on other organ systems. Interestingly, of all the organs studied, the heart and notochord showed the most dramatic changes in gene expression.
Results
Whole embryo (Table 1 & Figure 1A)
Since body length is a better predictor of developmental stage than age [12], we measured body length in all embryos/larvae. There was no significant difference between the length of embryos/larvae in any of the experimental groups compared to their age matched controls (Data not shown). This suggests that none of the exposure durations used for this study caused any general developmental delay or acceleration.
Table 1 Table of percent changes in fluorescence intensity in different organs/cell types that resulted from exposure to simulated-microgravity for different developmental times.
Whole Embryo Heart Notochord
hpf mean SEM p-value hpf mean SEM p-value hpf mean SEM p-value
8–24 -1.24 2.25 0.37 8–24 8–24 0.58 4.51 0.25
8–32 8.40 3.61 0.07 8–32 8–32 16.07 5.60 0.03
8–56 11.59 3.66 0.003 8–56 9.41 8.12 0.17 8–56 10.11 6.23 0.08
24–48 8.81 2.86 0.01 24–48 12.19 5.83 0.04 24–48 16.17 4.56 0.003
24–72 18.62 4.58 0.0004 24–72 27.75 4.82 0.00003 24–72 31.68 6.80 0.0001
24–80 10.07 5.72 0.11 24–80 11.15 6.44 0.13 24–80 17.77 8.11 0.06
32–56 2.67 5.19 0.29 32–56 35.60 12.44 0.02 32–56 2.25 7.29 0.32
32–80 1.21 4.14 0.43 32–80 0.38 5.97 0.48 32–80 3.91 7.17 0.35
48–72 1.97 2.50 0.28 48–72 8.42 3.96 0.04 48–72 4.13 4.51 0.22
56–80 -2.58 2.82 0.27 56–80 5.01 5.61 0.27 56–80 -7.11 4.35 0.15
Rohon Beard Neurons Eye Lens
hpf mean SEM p-value hpf mean SEM p-value hpf mean SEM p-value
8–24 8–24 2.07 4.80 0.40 8–24
8–32 8–32 3.01 5.23 0.36 8–32
8–56 9.65 5.84 0.08 8–56 14.06 6.73 0.04 8–56 17.73 6.69 0.01
24–48 8.05 4.65 0.09 24–48 12.65 4.57 0.01 24–48 10.75 4.74 0.03
24–72 17.29 5.08 0.01 24–72 15.25 7.18 0.03 24–72 18.55 7.28 0.02
24–80 14.36 7.28 0.09 24–80 11.27 7.78 0.15 24–80 10.41 7.78 0.18
32–56 1.90 6.78 0.40 32–56 8.96 8.58 0.18 32–56 11.73 9.13 0.14
32–80 -1.60 5.72 0.41 32–80 0.70 5.94 0.47 32–80 -4.09 5.57 0.34
48–72 5.53 4.10 0.17 48–72 -3.77 3.73 0.24 48–72 -4.73 4.52 0.22
56–80 -4.76 4.30 0.24 56–80 1.25 5.06 0.44 56–80 2.83 5.91 0.37
Somites
hpf mean SEM p-value
8–24 4.28 3.38 0.23
8–32 4.10 5.71 0.31
8–56 7.07 5.25 0.13
24–48 7.50 4.52 0.09
24–72 22.24 6.30 0.002
24–80 15.39 7.80 0.08
32–56 3.11 7.17 0.32
32–80 2.08 5.89 0.41
48–72 8.24 6.52 0.12
56–80 -6.88 3.50 0.13
hpf = hours post-fertilization that the embryos were kept in the bioreactor (see Figure 2);
SEM = standard error of the mean.
Figure 1 Graphs of percent changes in fluorescence intensity in different organs/cell types that resulted from exposure to simulated-microgravity for different developmental times. For each embryo, fluorescence intensity was measured in the regions illustrated in Figure 3. Each graph represents the changes in a specific organ/cell type. Regions of each graph are shaded to indicate exposure times that began at the same developmental time. All graphs are presented with the same exposure-time sequence on the x-axis and the same scale on the y-axis. Error bars = SEM
A brief exposure to simulated-microgravity during very early development (8–24 hpf) had no significant effect on overall fluorescence intensity of the whole embryo nor did any exposure that began after 32 hpf. However, exposure durations of at least 24 hours beginning prior to 32 hours caused increases in fluorescence with a peak occurring after exposure from 24 to 72 hours after fertilization. After 72 hours, continued exposure did not induce a further increase in expression but rather expression decreased. This suggests that a younger embryo is more susceptible to the effects of simulated-microgravity and that as the embryo matures, it develops mechanisms to adapt to simulated-microgravity.
Heart (Table 1 & Figure 1B)
Gene expression in the heart appeared to follow the same general trend seen in the embryo as a whole except the changes were more dramatic. The heart showed changes that increased in magnitude with a peak after 72 hours followed by recovery despite continued exposure to simulated-microgravity. In addition, the developing heart had a period between 32 and 56 hours when it showed the most dramatic increase in fluorescence intensity, 35.60 +/- 12.44% (mean +/- SEM), while the embryo as a whole showed no change.
Notochord (Table 1 & Figure 1C)
Gene expression in the notochord also appeared to follow the same general trend seen in the embryo, as a whole except, like the heart, the changes were more dramatic. In addition, the notochord, like the heart, showed a variation from the pattern of effects seen in the whole embryo. Unlike the heart, which showed a period of later susceptibility, the notochord had an additional period of earlier susceptibility (8–32 hpf) than the embryo as a whole.
Rohon beard neurons, eye, lens, and somites (Table 1 & Figure 1D–G)
Gene expression in the rohon beard neurons, eye, lens, and somites followed the same general trend seen in the embryo as a whole.
Discussion
The interpretation of the results presented here is based on the assumptions that gfp mRNA transcription, mRNA stability, and translation are similar to that of the native β-actin gene, and that an increase in GFP expression would be correlated with a similar increase in β-actin expression. This has been shown to be true for many other transgenic zebrafish (for instance see: [13-18]. Therefore, we can draw 3 general conclusions from the current work. First, simulated-microgravity can have global effects on gene-expression. Second, different organs/tissues, such as the heart and notochord, are more susceptible to the effects of simulated-microgravity on gene expression. Third, there are organ/tissue-specific periods of greater susceptibility that coincide with specific developmental events.
Simulated-microgravity's global effects on gene expression are suggested by the trends that become apparent when all the panels in Figure 1 are compared. For instance, early exposures of 24 to 48 hour durations tend to cause increases in β-actin gene expression in all the tissues examined. However, prolonging the exposure for durations longer than 48 hours and the embryos and tissues tend to show a recovery from the earlier effects. This is consistent with recent work using a similar bioreactor, showing that cultured endothelial cells up-regulate actin expression after early exposure to simulated-microgravity and recovery after prolonged exposure [4]. In addition, younger embryos tend to be more susceptible to the effects of simulated-microgravity on β-actin gene expression than older embryos (compare exposure times beginning at either 8 or 24 hours with exposure times beginning at 32 hours). In general, younger embryos are more susceptible and the effects are reversible.
Of all the organs/tissues that we have analyzed, the heart and notochord showed the most dramatic changes in β-actin gene expression after exposure to simulated-microgravity. Both of these organs showed periods of greater susceptibility, but the periods were different for the two organs. The heart, unlike the rohon beard neurons, the eye, the lens and the somites is involved in active, physical processes. This supports the idea that it is some aspect of the heart contraction/pumping that causes the more dramatic changes in the heart and could explain the smaller, similar changes in gene expression seen in the other tissues. The heart has a period of peak sensitivity between 32 and 56 hours post fertilization. This is the developmental time frame during which the beating heart tube is changing its shape as folding occurs and the atrio-ventrcular septum begins to form [19]. This might also be the developmental period during which the developing heart is most susceptible to the morphological influences of changes in shear forces associated with blood flow [20,21]. Since fluid dynamics are known to be influenced by changes in gravitational force, it is possible that the changes in β-actin gene expression could be due to changes in shear forces associated with blood flow through the heart in simulated-microgravity. Shear forces are also known to impact gene expression in cultured endothelial cells [21]. If simulated-microgravity related changes in shear force are initiating the change in gene expression in the zebrafish heart, then microgravity might have an impact on gene expression in the heart and blood vessels of astronauts during space flight. This raises the possibility that effects of microgravity on gene expression in endothelial cells in vivo might be a contributing factor in cardiovascular changes in astronauts.
The notochord shows more substantial early changes in β-actin gene expression than the developing heart and also appears to have two different periods of susceptibility (Figure 1C). We proposed above that the effects on the developing heart might be due to changes in shear forces associated with changes in fluid dynamics. The changes seen in the notochord are more difficult to explain. If intracellular micro-convective currents are significantly reduced in simulated-microgravity, actin polymerization-depolymerization could be affected because of a change in the local availability of actin subunits. In this case, the change in β-actin gene expression could reflect an attempt to maintain a stable cytoskeleton. However, the vector averaged nature of the bioreactor simulating microgravity might still allow for micro-convective currents that would play a role in regulating polymerization-depolymerization of actin filaments. If this is the case, then the notochord might be more susceptible to direct effects of microgravity than any other part of the embryo. The notochord is the major support structure of the embryo serving the same purpose as the vertebrae in older zebrafish. The two periods of susceptibility coincide with the 'straightening' of the embryo and the onset of spontaneous muscle contractions and movements of the embryo. These are the two developmental periods that the notochord cells would need to strengthen their cytoskeleton to provide additional stiffness, during the first period, and support for the embryo during the second period. Changes in gravitational force would result in a change in the stiffness and support that would need to be provided by the notochord. The same changes in support requirements result in changes in the skeletal system in astronauts during space flight suggesting that studying the zebrafish notochord in simulated-microgravity might provide insight into possible direct effects of microgravity on the skeletal system in astronauts.
Conclusion
All-in-all, our results support the idea that exposure to microgravity can cause changes in gene expression in a variety of developing organ systems in live embryos and that there are periods of maximum susceptibility to the effects. In addition, the periods of susceptibility differ for different organ systems. Interestingly, organs that show effects in these experiments either develop into or regulate the development of organ systems that are affected in astronauts during space flight [22-24]. This suggests that ground-based research using zebrafish embryos might have predictive value for the affects of long-duration space flight on astronaut health.
Methods
Animals
Adult zebrafish (strain: Tg(actin:egfp)zp1) that expressed gfp gene under the influence of the promoter/enhancer of the zebrafish β-actin gene [3] were maintained at 28°C in communal tanks on a 14/10 hour light/dark cycle. Eggs were collected within 3 hours of when they were laid and fertilized. 8 hours post-fertilization (hpf) the eggs were transferred to room temperature (20°C) and phenyl-thiourea (0.003%: Sigma, St Louis MO) was added to the water to inhibit pigment production. The eggs were maintained at room temperature for the duration of the experiments. All experiments were performed under an approved IACUC protocol (#I02-020-02).
Simulated-microgravity
To simulate many of the aspects of microgravity for zebrafish embryos and larvae, we used a bioreactor that NASA designed for cells in culture [1-3]. We placed the eggs in the bioreactor at the times indicated in Figure 2. 25–30 eggs were placed in the bioreactor at the beginning of each experiment. When eggs were removed from the bioreactor, 9 embryos were chosen at random and manually removed from the chorion. 9 embryos were prepared for imaging in case some of the embryos did not express the transgene. Only 6 embryos/larvae were used for acquiring images of GFP fluorescence. Each experiment was repeated 4 times to give a total of N = 24 in each group.
Figure 2 Developmental time that embryos were in the bioreactor. The bars indicate the developmental time, in hours post-fertilization, that embryos were kept in the bioreactor. N = 24 control & N = 24 Experimental for each exposure time.
Before beginning the experiments we compared gene expression in embryos incubated in an identical bioreactor oriented in an upright position so that it functioned as a slow speed centrifuge with embryos incubated in a Petri dish placed on the support base of the bioreactor. The Petri dish with the embryos (N = 24) was placed on the support base of the bioreactor so that the embryos were exposed to the same temperature, light cycle, and vibration as the embryos in the bioreactor. Embryos (N = 24) were placed in the bioreactor at 24 hpf and maintained in the bioreactor for 48 hrs. There were no differences in morphology, or gene expression for the whole embryo (mean = -1.51; sem = 5.79; p = 0.43), the heart (mean = 1.13; sem = 6.28; p = 0.45), the notochord (mean = -1.49; sem = 6.21; p = 0.49), the eye (mean = -1.21; sem = 5.37; p = 0.49), the lens (mean = 0.41; sem = 6.57; p = 0.45), the somites (mean = 1.49; sem = 6.21; p = 0.49), or the rohon beard neurons (mean = 1.26; sem = 6.34; p = 0.47) between the two groups of embryos. Therefore, all subsequent experiments were performed using control embryos in a Petri dish on the support base of the bioreactor.
GFP-fluorescence imaging
All images were collected using a Leica DMRE microscope (Leica Microsystems Inc., Bannockburn, IL) equipped with a Ludl BioPrecision motorized stage (Ludl Electronic Products Ltd., Hawthorne, NY) and a Hamamatsu Orca-ER camera (Hamamatsu Photonics, Hamamatsu City Japan). The microscope, stage, and camera were controlled using OpenLab software (Improvision, Lexington, MA) running on an Apple Dual-processor G4 computer. Prior to collecting fluorescence images, a bright-field image of each embryo/larvae was acquired using a 5× objective. This image was used to measure rostral-caudal length as an indication of the age of the embryo/larva. After the brightfield image was acquired, a complete Z-series of fluorescence images was acquired. For this series, images were collected at 3 μm intervals using a 10× objective. The entire stack of images was saved to disk. The camera gain, offset, and exposure time were kept constant for all homozygous embryos/larvae. Because heterozygous embryos/larvae had approximately half the fluorescence intensity as homozygous embryos/larvae, a second set of camera settings was used for all of the heterozygous embryos/larvae. Embryos older than 24 hpf were anesthetized using tricaine (0.04% 3-amino benzoic acidethylester: Sigma, St. Louis) during imaging to prevent movements of the embryo.
GFP intensity measurements
Using the OpenLab software, an image where the organ of interest was in focus was selected from the z-series stack. A region of interest was then drawn around the organ (Figure 3) and the software automatically calculated the average intensity within the region. To estimate the average intensity of fluorescence for the entire embryo/larva, an image was selected from the middle of the z-series stack and the average intensity for the entire image was calculated.
Figure 3 Images of the same transgenic zebrafish embryo (strain: Tg(actin:egfp)zp1) 48 hours post fertilization. A. Bright field image of the embryo. B-D. Fluorescence images of the same embryo at different focal planes. B. Image at the focal plane for heart measurement. The heart is outlined in red. C. Image at the focal plane for notochord measurement. A length of notochord adjacent to 4 somites is outlined in red. D. Image at the focal plane for rohon beard neuron measurement. 4 rohon beard neurons are indicated by red arrows. E. Image at the focal plane for eye and lens measurement. The eye and lens are outlined, separately, in red. F. Image at the focal plane for somite measurement. 4 somites are outlined in red. Scale bar = 1 mm (A), 0.5 mm (B – E), 0.05 mm (F).
Data analysis and statistics
The measurements of fluorescence intensity in homozygous and heterozygous control embryos/larvae were used to normalize the measurements in homozygous and heterozygous experimental embryos/larvae respectively. After the individual measurements were normalized, the mean, standard deviation, and standard error of the mean were calculated for each group. The means for the control and experimental groups were compared using a t-test.
Authors' contributions
NS carried out the majority of the experiments and participated in the interpretation of the results and drafting of the manuscript. GS carried out some of the experiments and participated in the design of the study. SJM conceived of the study, and participated in its design and coordination and helped to draft the manuscript. All authors read and approved the final manuscript.
Acknowledgements
We thank Meghal Desai for excellent technical help and fish care and Denise Dehnbostel for editing assistance. This work was funded by NASA (#NAG2-1591).
==== Refs
Moorman SJ Burress C Cordova R Slater J Stimulus dependence of the development of the zebrafish (Danio rerio) vestibular system J Neurobiol 1999 38 247 258 10022570 10.1002/(SICI)1097-4695(19990205)38:2<247::AID-NEU7>3.0.CO;2-3
Moorman SJ Cordova R Davies SA A critical period for functional vestibular development in zebrafish Dev Dyn 2002 223 285 291 11836792 10.1002/dvdy.10052
Gillette-Ferguson I Ferguson DG Poss KD Moorman SJ Changes in gravitational force induce alterations in gene expression that can be monitored in the live, developing zebrafish heart Adv Space Res 2003 32 1641 1646 15002421 10.1016/S0273-1177(03)90405-4
Carlsson SI Bertilaccio MT Ballabio E Maier JA Endothelial stress by gravitational unloading: effects on cell growth and cytoskeletal organization Biochim Biophys Acta 2003 1642 173 179 14572900 10.1016/j.bbamcr.2003.08.003
Perhonen MA Franco F Lane LD Buckey JC Blomqvist CG Zerwekh JE Peshock RM Weatherall PT Levine BD Cardiac atrophy after bed rest and spaceflight J Appl Physiol 2001 91 645 653 11457776
Goldstein MA Cheng J Schroeter JP The effects of increased gravity and microgravity on cardiac morphology Aviat Space Environ Med 1998 69 A12 6 10776447
Prisk GK Fine JM Elliott AR West JB Effect of 6 degrees head-down tilt on cardiopulmonary function: comparison with microgravity Aviat Space Environ Med 2002 73 8 16 11817623
Fleming A Keynes RJ Tannahill D The role of the notochord in vertebral column formation J Anat 2001 199 177 180 11523820 10.1017/S0021878201008044
Altmann CR Brivanlou AH Neural patterning in the vertebrate embryo Int Rev Cytol 2001 203 447 482 11131523
Pourquie O Vertebrate somitogenesis Annu Rev Cell Dev Biol 2001 17 311 350 11687492 10.1146/annurev.cellbio.17.1.311
Cleaver O Krieg PA Notochord patterning of the endoderm Dev Biol 2001 234 1 12 11356015 10.1006/dbio.2001.0214
Bird NC Mabee PM Developmental morphology of the axial skeleton of the zebrafish, Danio rerio (Ostariophysi: Cyprinidae) Dev Dyn 2003 228 337 357 14579374 10.1002/dvdy.10387
Kaneko M Cahill GM Light-Dependent Development of Circadian Gene Expression in Transgenic Zebrafish PLoS Biol 2005 3 e34 15685291 10.1371/journal.pbio.0030034
Her GM Chiang CC Wu JL Zebrafish intestinal fatty acid binding protein (I-FABP) gene promoter drives gut-specific expression in stable transgenic fish Genesis 2004 38 26 31 14755801 10.1002/gene.10248
Ward AC McPhee DO Condron MM Varma S Cody SH Onnebo SM Paw BH Zon LI Lieschke GJ The zebrafish spi1 promoter drives myeloid-specific expression in stable transgenic fish Blood 2003 102 3238 3240 12869502 10.1182/blood-2003-03-0966
Shafizadeh E Huang H Lin S Transgenic zebrafish expressing green fluorescent protein Methods Mol Biol 2002 183 225 233 12136757
Lawson ND Weinstein BM In vivo imaging of embryonic vascular development using transgenic zebrafish Dev Biol 2002 248 307 318 12167406 10.1006/dbio.2002.0711
Wan H He J Ju B Yan T Lam TJ Gong Z Generation of two-color transgenic zebrafish using the green and red fluorescent protein reporter genes gfp and rfp Mar Biotechnol (NY) 2002 4 146 154 14961274 10.1007/s10126-001-0085-3
Glickman NS Yelon D Cardiac development in zebrafish: coordination of form and function Semin Cell Dev Biol 2002 13 507 513 12468254 10.1016/S1084952102001040
Hove JR Koster RW Forouhar AS Acevedo-Bolton G Fraser SE Gharib M Intracardiac fluid forces are an essential epigenetic factor for embryonic cardiogenesis Nature 2003 421 172 177 12520305 10.1038/nature01282
Topper JN Gimbrone MAJ Blood flow and vascular gene expression: fluid shear stress as a modulator of endothelial phenotype Mol Med Today 1999 5 40 46 10088131 10.1016/S1357-4310(98)01372-0
Hawkey A The physical price of a ticket into space J Br Interplanet Soc 2003 56 152 159 14552355
Droppert PM The effects of microgravity on the skeletal system--a review J Br Interplanet Soc 1990 43 19 24 12856692
Williams DR The biomedical challenges of space flight Annu Rev Med 2003 54 245 256 12471177 10.1146/annurev.med.54.101601.152215
| 15927051 | PMC1177936 | CC BY | 2021-01-04 16:40:17 | no | BMC Dev Biol. 2005 May 31; 5:10 | utf-8 | BMC Dev Biol | 2,005 | 10.1186/1471-213X-5-10 | oa_comm |
==== Front
BMC DermatolBMC Dermatology1471-5945BioMed Central London 1471-5945-5-61591889710.1186/1471-5945-5-6Research ArticleTopical immunotherapy of severe alopecia areata with diphenylcyclopropenone (DPCP): experience in an Iranian population Aghaei Shahin [email protected] Department of Dermatology, Jahrom Medical School, Jahrom, Iran2005 26 5 2005 5 6 6 29 1 2005 26 5 2005 Copyright © 2005 Aghaei; licensee BioMed Central Ltd.2005Aghaei; 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
Highly variable results of topical diphenylcyclopropenone (DPCP) in the treatment of alopecia areata have been reported so far. The purposes of the present study were to evaluate the efficacy and tolerability of DPCP treatment in severe alopecia areata.
Methods
Twenty-eight patients (16 female and 12 male, 10–35 years old, mean age 25 years) with extensive alopecia areata were enrolled in an open-label clinical trial. After sensitization with 2% DPCP, progressively higher concentrations beginning at 0.001% were applied weekly for 6 months to one side of the scalp, after which, if terminal hair growth was noted, the entire scalp was then treated under the same weekly protocol. The maximum concentration of DPCP in acetone was 2%.
Results
Twenty-seven of 28 patients completed therapy. The overall response rate was 81.5% (22/27), complete remission (90%-100% terminal hair re-growth) was obtained 22.2% (6/27) and partial remission (10%-90% terminal hair re-growth) in 59.3% (16/27). In all patients an eczematous reaction consisting of erythema, itching, and scaling at the site of application were observed. During therapy, other side effects including, occipital lymphadenopathy 40.7% (11/27), severe eczema/blister formation 40.7% (11/27), hyperpigmentation 18.5% (5/27) were observed, but no hypopigmentation, vitiligo, contact urticaria, and erythema multiforme-like reaction were seen in the patients. Nineteen of 27 (70.4%) patients had at least one side effect, other than eczematous reaction. Notably, partial recurrence was observed in 50.9% (13/22) of these patients after 6 to 12 months of follow-up. During the follow-up time the maintenance DPCP immunotherapy was continued.
Conclusion
Topical DPCP treatment for alopecia areata is an effective therapy with a slightly high relapse rate during bilateral maintenance treatment. According to the author's knowledge this is the first experience with DPCP in Iran.
==== Body
Background
Dermatosis of the scalp such as alopecia areata (AA) are a burden for many patients and often resistant, even to extensive therapy [1-4]. Furthermore, treatment results can be difficult to interpret because of spontaneous remissions and recurrences, as well as the use of several therapies simultaneously, such as tretinoin creams, minoxidil lotions and zinc supplementation [4]. Topical and intra-lesional corticosteroid therapy is frequently tried, but the benefit of such treatment is often questionable or temporary. Systemic corticosteroid treatment may be effective in some cases, but the maintenance dose needed is often high [4,5]. Some success, has been reported with anthralin, but results seem variable [6,7]. Other therapies which have been tried, with variable success, include minoxidil [8,9], cyclosporine [10-12], alpha-interferon [13], and acupuncture [14].
Topical immunotherapy has been used for more than 10 years for the treatment of severe AA and various contact allergens such as dinitrochlorobenzene (DNCB), squaric acid dibutylester (SADBE), and diphenylcyclopropenone (DPCP) have determined re-growth of hair in patients with AA [15-17].
The effectiveness of topical immunotherapy with AA has been demonstrated in several reports, although the response rate varied greatly from 4% to 85% [18]. The present study was performed as an open-label clinical trial to evaluate the efficacy and tolerability of topically administered DPCP in patients with extensive AA. In addition, the clinical response with prognostic factors was correlated and assessed the patients' response after a follow-up of 6 to 12 months, in which bilateral DPCP maintenance treatment was continued.
Methods
During 2 years between April 2001 and May 2003 at the Department of Dermatology, Saadi Hospital, Shiraz University of Medical Sciences, 28 patients (12 men, 16 women) with severe AA (> 40% scalp hair loss) were enrolled in an open-label clinical trial. Their ages ranged from 10 to 35 years (mean, 25 years). The DPCP used (Acros Organics, New Jersey, USA) 98% pure, was dissolved in acetone. Informed consent was obtained. Women of childbearing age were required to use a reliable form of birth control. Individuals with AA were ineligible for DPCP treatment if they presented with less than 40% scalp involvement, the age less than 10 years, significant cardiovascular disease, pregnancy, or serious intercurrent medical illnesses. In brief, sensitization was performed with a 2% solution of DPCP applied to an area of 5× 5 cm on one side of the scalp. Two weeks following sensitization, treatment was started by weekly ipsilateral applications of incremental concentrations of DPCP (between 0.001% and 2%) adjusted to the patient's reactivity to the contact allergen. The aim was to maintain mild contact eczema and itch for about 48 hours after application. Patients were instructed to avoid direct sun exposure of the scalp and not to wash the scalp for 48 hours after each application. The opposite half of the scalp served as a control to rule out possible spontaneous remission. Once hair re-growth occurred on the treated side, the applications were extended over the entire scalp. The overall duration of therapy varied from 6 to 12 months. The patients, who achieved clinically significant re-growth, were under follow-up of 6–12 months, with ongoing maintenance DPCP immunotherapy.
The types of AA before treatment were classified as follows: (1) multilocular AA including ophiatic type with > 40% loss of scalp hair; (2) subtotal or total AA; (3) AA partim universalis (loss of all scalp hair with some body hair left) or universalis (loss of all scalp and body hair). According to the response to topical DPCP immunotherapy, patients were grouped into 3 categories: no hair re-growth (< 10% terminal hair), partial hair re-growth (10–90% terminal hair) and complete hair re-growth (90–100% terminal hair).
Efficacy evaluation was performed with clinical examination. If no re-growth was observed within 6 months of treatment, the patient was considered to be a non-responder and was dropped from the trial.
In all patients, the following laboratory tests were performed at baseline, every 3 months during treatment and after treatment: complete blood cell count; chemistry profile; urinalysis; levels of thyroid hormones (free tri-iodothyronine, free thyroxine, thyroid stimulating hormone); fasting blood sugar; and antinuclear antibody titres (ANA). Other laboratory tests such as anti-gastric parietal cells, anti-thyroglobulin and anti-smooth muscle antibody titres were not done, because of loss of these capabilities in our institute.
The influence of the following factors on the outcome of topical immunotherapy were investigated: sex, age at onset, atopy, disease duration before treatment, type of AA, presence of nail changes, presence of naevus simplex, and family history of AA and other auto-immune disorders such as diabetes mellitus (DM) and thyroid diseases. To assess any correlations between the above mentioned parameters the X2 test was used.
Results
Twenty-seven out of 28 patients completed the treatment. The demographic and clinical data of the patients included in the study are summarized in table 1 and 2.
Table 1 Demographic data
Number of patients 27
Sex M/F 12/15
Age (range, mean) 10–35, 25 years
Disease duration before treatment (range, mean) 0.58–28, 5.82 years
Table 2 Clinical data
Type of AA
(1) Multi-locular including ophiatic type (>40% of scalp hair loss) 11/27 (40.7%)
(2) Subtotal, total 5/27 (18.6%)
(3) Universalis, partim universalis 11/27 (40.7%)
Positive family history of:
AA 10/27 (37%)
Thyroid disease 7/27 (6%)
DM 9/27 (33.3%)
Positive personal history of:
Thyroid disease 3/27 (11.1%)
DM None
Presence of nail changes 16/27 (59.3%)
Presence of naevus simplex on the posterior neck 11/27 (40.7%)
Circulating ANA None
Disease duration before treatment ranged from 0.58 (7 months) to 28 years. The duration of therapy ranged from 6 to 24 months, including long-term patients with repeated DPCP treatment for maintaining hair re-growth.
Overall 81.5% (22 of 27 patients) responded to therapy: 22.2% (6 of 27 patients) achieved complete hair re-growth (90–100% terminal hair), and 59.3% (16 of 27 patients) had partial hair re-growth (10–90% terminal hair). Five patients had no hair re-growth (18.5%). Of the 22 patients with complete and partial remission, 13 (50.9%) suffered a relapse either simultaneously maintenance treatment of follow-up or following termination of therapy. No abnormalities were detected on baseline and follow-up laboratory data in these patients.
Side effects following sensitization were seen in 19 of 27 patients (70.4%): occipital lymphadenopathy in 40.7% (11 of 27 patients), severe eczema/blister formation in 22.2% (6 of 27 patients), and hyperpigmentation in 18.5% (5 of 27 patients). No cases of hypopigmentation, vitiligo, contact urticaria, and erythema multiforme-like reaction were observed. The therapeutically induced mild contact eczema with itching was not considered as adverse effects.
Discussion
Topical immunotherapy, using DPCP, is currently considered the most effective mode of treatment. However, the way in which DPCP operates on hair follicles in AA still remains to be elucidated. Vascular endothelial growth factor (VEGF), essential for angiogenesis and vascular permeability, may be responsible for maintaining proper vasculature around hair follicles, and several studies provide evidence that apoptosis is a central element in the regulation of hair follicle and vascular regression [19]. Moreover, topical immunotherapy considerably alters the peribulbar CD4/CD8 ratio in human and experimental animal studies, restoring a condition close to normal scalp skin [19,20].
In the present study, growth of terminal hair on the entire scalp was achieved totally in 81.5% of patients with AA after 6–12 months of treatment with DPCP. This results are less than those reported by Cotellessa et al [18] who observed a 48% complete success rate in a series of 52 patients, and those of Weise et al [21] and Van der Steen et al [22] who detected 40% and 50.4% complete re-growth in 124 and 139 patients, respectively (complete hair re-growth in the present study was 22.2%). In other reports, however, the percentage of success greatly varied from 4% to 85% [17,21,23-25]. The discrepancy of response rates may be due to the number of patients in clinical trials; the type, duration, and severity of the AA; and different methods of assessing clinical efficacy.
In the present study, concomitant spontaneous re-growth of eyebrows and body could be observed in approximately half of the successfully treated patients, which are similar to those reported by Cotellessa et al [18]. Nine of 27 patients had positive family history of atopy, but no significant correlations between its frequency and treatment response revealed (P = 0.07). Also, positive personal/family history of thyroid disease, atopy, and diabetes mellitus (DM); sex; age of patients; type of AA; and duration of disease before treatment had no significant correlations with treatment response (P > 0.05). In addition, the X2 test revealed significant correlation between the presence of naevus simplex (salmon patch) on the posterior neck and less favorable treatment response (P = 0.03). The same correlation between the presence of nail changes (such as pitting, ridging, leukonychia punctata, brittleness, "red spotted lunula", "sandpaper nails", and onychomadesis) and less favorable response rate was detected (P = 0.02). Remarkably, with regard to patients' follow-up with ongoing maintenance therapy in patients with significant hair re-growth, partial recurrences of AA were observed in about one-half of the patients.
Wiesman et al. referred to the risks of relapse at long-term follow-up. After 35 months of follow-up, 62.6% of the subjects who had been successfully treated experienced a relapse, and this risk was not influenced by the implementation of maintenance therapy [26]. This result is in agreement with the present study. But, other investigators believe that if this result could be reproduced in other studies, the patient could stop treatment when hair re-growth was complete because maintenance treatment would be of no use [27].
Among the investigated "prognostic factors" for the outcome of DPCP treatment, which have formerly been studied with variable results [21,25,28], only the presence of naevus simplex and nail changes were found to be of significance.
The main side effects observed during therapy with relatively high frequency were severe eczema/bullous formation, occipital lymphadenopathy, and hyperpigmentation. In none of the patients hypo-pigmentation or vitiligo were observed.
Conclusion
The study findings are in agreement with former studies showing the efficacy of topical immunotherapy with DPCP in the treatment of severe AA, but with a slightly high relapse rate during bilateral maintenance treatment. According to the author's knowledge this is the first experience with DPCP in Iran.
List of abbreviations used
AA: Alopecia areata
DPCP: Diphenylcyclopropenone
DNCB: Dinitrochlorobenzene
SADBE: Squaric acid dibutylester
DM: Diabetes mellitus
ANA: Antinuclear antibody test
VEGF: Vascular endothelial growth factor
Competing interests
The author declares that he has no competing interests.
Figure 1 A 12-year-old patient with alopecia totalis before treatment.
Figure 2 The same patient 4-weeks after treatment.
Figure 3 After 20-weeks of treatment (bilateral treatment started on 12-weeks, no photo available).
Figure 4 After 24-weeks of treatment (right view).
Figure 5 After 24-weeks of treatment (left view).
Figure 6 After 28-weeks of treatment (front view).
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
The author thanks the patients and/or their families for giving informed consent about the application of the medication.
==== Refs
Fiedler VC Alaiti S Treatment of alopecia areata Dermatol Clin 1996 14 733 8 9238331 10.1016/S0733-8635(05)70399-5
Fiedler VC Alopecia areata: a review of therapy, efficacy, safety and mechanism Arch Dermatol 1992 128 1519 29 1444509 10.1001/archderm.128.11.1519
Schwartz RA Jonninger CK Alopecia areata Cutis 1997 59 238 41 9169260
Shapiro J Alopecia areata: Update on therapy Dermatol Clin 1993 11 35 46 8435916
Olsen EA Carson SC Turney EA Systemic steroids with or without 2% topical minoxidil in the treatment of alopecia areata Arch Dermatol 1992 128 1467 73 1444500 10.1001/archderm.128.11.1467
Schmoeckel C Weissman I Pelwig G Braun-Falco O Treatment of alopecia areata by anthralin-induced dermatitis Arch Dermatol 1979 115 1254 5 159668 10.1001/archderm.115.10.1254
Fiedler-Weiss VC Buys CM Evaluation of anthralin in the treatment of alopecia areata Arch Dermatol 1987 123 1491 3 3314718 10.1001/archderm.123.11.1491
Fiedler VC Wendrow A Szpunar GJ Metzier C De Villez RL Treatment-resistant alopecia areata. Response to combination therapy with minoxidil plus anthralin Arch Dermatol 1990 126 756 9 2140670 10.1001/archderm.126.6.756
Fiedler-Weiss VC Topical minoxidil solution (1% and 5%) in the treatment of alopecia areata J Am Acad Dermatol 1987 16 745 8 3549811
Gupta AK Ellis CN Cooper KD Nickoloff BJ Ho VC Chan LS Hamilton TA Tellner DC Griffiths CE Voorhees JJ Oral cyclosporine for treatment of alopecia areata. A clinical and immunohistochemical analysis J Am Acad Dermatol 1990 22 242 50 2138175
Gilhar A Pillar T Etzioni A Topical cyclosporine A in alopecia areata Acta Derm Venereol (Stockh) 1989 69 252 3 2566234
de Prost Y Teillac D Paquez F Carrugi L Bachelez H Touraine R Placebo-controlled trials of topical cyclosporine in severe alopecia areata Lancet 1986 ii 803 4 2876247 10.1016/S0140-6736(86)90315-6
Magee KL Hsu SM Tucker SB Trial of intralesional interferon-alpha in the treatment of alopecia areata Arch Dermatol 1990 126 760 2 2189361 10.1001/archderm.126.6.760
Ge S Treatment of alopecia areata with acupuncture J Trad Chinese Med 1990 10 199 200
Daman L Rosenberg W Drake L Treatment of alopecia areata with DNCB Arch Dermatol 1978 114 1036 8 308348 10.1001/archderm.114.7.1036
Happle R Kalveran KJ Buchner U Echternacht-Happle K Goggelmann W Summer KH Contact allergy as a therapeutic tool for alopecia areata: application of squaric acid dibutylester Dermatologica 1980 161 289 97 7439477
Happle R Hansen BM Wiesner-Menzel L Diphenylcyclopropenone in the treatment of alopecia areata Acta Derm Venereol (Stockh) 1983 63 49 52 6191489
Cotellessa C Peris K Caracciolo E The use of topical diphenylcyclopropenone for the treatment of extensive alopecia areata J Am Acad Dermatol 2001 44 73 6 11148480 10.1067/mjd.2001.109309
Simonetti O Lucarini G Bernardini ML Simoncini C Biagini G Offidani A Expression of vascular endothelial growth factor, apoptosis inhibitors (survivin and p16) and CCL27 in alopecia areata before and after diphencyprone treatment: an immunohistochemical study Br J Dermatol 2004 150 940 948 15149507 10.1111/j.1365-2133.2004.05881.x
Tang L Lui H Sundberg JP Bissonnette R McLean DI Shapiro J Restoration of hair growth with topical diphencyprone in mouse and rat models of alopecia areata J Am Acad Dermatol 2003 49 1013 9 14639378 10.1016/S0190-9622(03)02141-8
Weise K Kretzschmar L John SM Hamm H Topical immunotherapy in alopecia areata: anamnestic and clinical criteria of prognostic significance Dermatology 1996 92 129 33 8829494
Van der Steen PHM Boezeman JBM Happle R Topical immunotherapy for alopecia areata: Re-evaluation of 139 cases after an additional follow-up of 19 months Dermatology 1992 184 198 201 1392112
Hausen BM Stute J Diphenylcyclopropenone: a strong contact sensitizer Chem Industry 1980 17 699 700
Kietzmann H Hardung H Christophers E Terapie der alopecia areata mit Diphenylcyclopropenon Hautarzt 1985 36 331 5 Abstract 3160683
Gordon PM Aldridge RD McVittie E Hunter JAA Topical diphencyprone for alopecia areata: evaluation of 48 cases after 30 months' follow-up Br J Dermatol 1996 134 869 71 8736327 10.1046/j.1365-2133.1996.119854.x
Wiseman MC Shapiro J McDonald N Lui H Predictive model for immunotherapy of alopecia areata with diphencyprone Arch Dermatol 2001 137 1063 1068 11493099
Happle R Diphencyprone for the treatment of alopecia areata Arch Dermatol 2002 138 112 113 11790175 10.1001/archderm.138.1.112
Van der Steen PHM Van Baar HJM Perret CM Happle R Treatment of alopecia areata with diphenylcyclopropenone J Am Acad Dermatol 1991 24 253 7 2007671
| 15918897 | PMC1177937 | CC BY | 2021-01-04 16:29:46 | no | BMC Dermatol. 2005 May 26; 5:6 | utf-8 | BMC Dermatol | 2,005 | 10.1186/1471-5945-5-6 | oa_comm |
==== Front
BMC DermatolBMC Dermatology1471-5945BioMed Central London 1471-5945-5-81600017110.1186/1471-5945-5-8Research ArticleComputer-aided dermoscopy for diagnosis of melanoma Barzegari Masoomeh [email protected] Haiedeh [email protected] Parisa [email protected] Arash [email protected] Zahra S [email protected] Masood [email protected] Department of Dermatology, Razi Hospital, Tehran University of Medical Sciences, Vahdat-Eslami Street, 11996 Tehran, Iran2 Department of Pathology, Razi Hospital, Tehran University of Medical Sciences, Vahdat-Eslami Street, 11996 Tehran, Iran2005 6 7 2005 5 8 8 22 2 2005 6 7 2005 Copyright © 2005 Barzegari et al; licensee BioMed Central Ltd.2005Barzegari 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
Computer-aided dermoscopy using artificial neural networks has been reported to be an accurate tool for the evaluation of pigmented skin lesions. We set out to determine the sensitivity and specificity of a computer-aided dermoscopy system for diagnosis of melanoma in Iranian patients.
Methods
We studied 122 pigmented skin lesions which were referred for diagnostic evaluation or cosmetic reasons. Each lesion was examined by two clinicians with naked eyes and all of their clinical diagnostic considerations were recorded. The lesions were analyzed using a microDERM® dermoscopy unit. The output value of the software for each lesion was a score between 0 and 10. All of the lesions were excised and examined histologically.
Results
Histopathological examination revealed melanoma in six lesions. Considering only the most likely clinical diagnosis, sensitivity and specificity of clinical examination for diagnosis of melanoma were 83% and 96%, respectively. Considering all clinical diagnostic considerations, the sensitivity and specificity were 100% and 89%. Choosing a cut-off point of 7.88 for dermoscopy score, the sensitivity and specificity of the score for diagnosis of melanoma were 83% and 96%, respectively. Setting the cut-off point at 7.34, the sensitivity and specificity were 100% and 90%.
Conclusion
The diagnostic accuracy of the dermoscopy system was at the level of clinical examination by dermatologists with naked eyes. This system may represent a useful tool for screening of melanoma, particularly at centers not experienced in the field of pigmented skin lesions.
==== Body
Background
The incidence of melanoma is much lower in Asia than in western countries [1]. Clinicians sometimes misdiagnose early melanoma especially in areas with lower incidence of disease [2]. Because advanced cutaneous melanoma is still incurable, early detection by means of accurate screening is an important step towards a reduction in mortality.
Recently, computer-aided dermoscopy using artificial neural networks (ANNs) has been reported to be an accurate tool for the evaluation of pigmented skin lesions (PSLs) [3-5].
To our knowledge, the accuracy of such system for diagnosis of PSL has not been demonstrated in the Middle East, where most of the patients have Fitzpatrick skin type III-IV. We set out to determine the sensitivity and specificity of this system in Iranian patients.
Methods
One hundred and twenty two consecutive PSLs equal or smaller than 15 mm in diameter, with a clinical diagnosis of one of the pigmented melanocytic lesions, which were referred to dermatology clinic of Razi Hospital for diagnostic evaluation or cosmetic reasons were included in the study. Each lesion was examined by two clinicians (an attending dermatologist and a third year dermatology resident) with naked eyes. They consulted with each other and recorded all of their clinical diagnostic considerations in a list. The first diagnostic consideration in the list was the most likely clinical diagnosis. After clinical examination, the lesions were analyzed using a microDERM® dermoscopy unit. The system consists of a special camera, which had ability to take images at ×15, ×20, ×30, and ×50 magnifications and contains a 752 × 582 pixel charge-coupled device. The image analysis software was Visiomed AG (Ver. 3.50) based on an ANN that was trained using images collected in a Europe-wide multicenter study (DANAOS) [6]. The output value of the software is a score ranging from 0 to 10 for each lesion. Informed consent was obtained from each patient. All of the lesions were excised and examined histologically. The final diagnosis was made based on pathological examination.
The study was approved by the Research Ethics Committee of Razi Hospital.
Results
One hundred and twenty two pigmented skin lesions from 91 Iranian patients (30 male, 61 female; mean age 32.3 years, age range 6–94) were included in the study.
Melanoma was in the list of clinical diagnostic considerations in 19 lesions. It was the most likely clinical diagnosis in nine lesions.
Table 1 shows the frequency and dermoscopy score of the PSLs. Histopathological examination revealed melanoma in six lesions.
Table 1 Frequency and dermoscopy score of the lesions according to histopathological diagnosis.
Histopathological diagnosis No. of lesions Dermoscopy Score
N % Mean Range
Melanoma 6 4.9 8.05 (7.34–8.47)
Lentigo Maligna 3 2.5 7.90 (7.34–8.47)
Lentigo maligna melanoma 2 1.6 8.19 (7.92–8.47)
Acral lentiginous melanoma 1 0.8 8.24 -----
Nonmelanoma 116 95.1 3.86 (0.00–8.35)
Melanocytic 111 91 3.74 (0.00–8.35)
Junctional nevus 4 3.2 2.54 (1.92–3.21)
Compound nevus 10 8.1 2.86 (1.84–5.23)
Intradermal nevus 76 62.3 3.68 (0.00–8.35)
Congenital nevus 7 5.7 5.34 (0.54–7.52)
Blue nevus 5 4.1 2.24 (0.35–3.72)
Combined nevus 2 1.6 4.54 (1.39–7.69)
Dysplastic nevus 7 5.7 5.66 (1.70–8.06)
Nonmelanocytic 5 4.1 6.42 (0.64–8.09)
Seborrheic keratosis 2 1.6 8.00 (7.91–8.09)
Dermatofibroma 1 0.8 7.96 -----
Epidermal nevus 1 0.8 0.64 -----
Actinic keratosis 1 0.8 7.53 -----
Total 122 100 4.06 (0.00–8.47)
N, number
Considering only the most likely clinical diagnosis, sensitivity and specificity of clinical examination for diagnosis of melanoma were 83% and 96%, respectively. Considering all clinical diagnostic considerations, the sensitivity and specificity were 100% and 89%.
Figure 1 shows receiver operating characteristic (ROC) curve for the separation of benign PSLs and melanoma using dermoscopy score in our study. In order to compare the sensitivity and specificity of clinical examination with dermoscopy score, we selected two points on the ROC curve that showed sensitivity and specificity near that of clinical examination. Choosing a cut-off point of 7.88 for dermoscopy score, the sensitivity and specificity of the score for diagnosis of melanoma were 83% and 96%, respectively. Setting the cut-off point at 7.34, the sensitivity and specificity were 100% and 90%.
Figure 1 Receiver operating characteristic curve for the separation of benign pigmented skin lesions and melanoma using dermoscopy score.
Cohen's kappa statistic was used for evaluation of agreement between diagnostic tests. Kappa values are shown in Table 2.
Table 2 Cohen's kappa values for evaluation of agreement between diagnostic tests.
Tests Kappa value
PD versus CD1 0.65
PD versus CD2 0.44
PD versus DS1 0.50
PD versus DS2 0.46
CD1 versus DS1 0.64
CD2 versus DS2 0.65
CD1, clinical diagnosis when considering only the most likely clinical diagnosis; CD2, clinical diagnosis when considering all clinical diagnostic considerations; DS1, dermoscopy score when setting the cut-off point at 7.88; DS2, dermoscopy score when setting the cut-off point at 7.34; PD, pathological diagnosis.
Discussion
The ABCD rule is one of the most widely used methods for evaluating PSLs with the naked eye. However, the diagnostic accuracy is not very high [7,8].
Dermoscopy is a noninvasive method that enables clinicians to evaluate numerous morphological features that are not visible to the naked eye. Several studies have shown that this method improves diagnostic accuracy by 20–30% compared with simple clinical observation [9-11]. Recently, computer-aided dermoscopy has been introduced as an additional tool to improve the diagnosis of pigmented skin lesions. In previous reports, the sensitivity of digital dermoscopy analysis for diagnosis of melanoma has ranged between 80% and 100% [12,13]. The specificity has ranged between 46% and 98% [14,15]. The diagnostic accuracy of most of the reported softwares have been at the level of a dermatologist experienced in dermoscopy and higher than inexperienced clinicians [3-5,16]. In one study carried out by Seidenari et al, diagnostic accuracy of the software was higher than clinical assessment by an experienced observer [17]. They reported that the sensitivity and specificity of computer analysis for diagnosis of melanoma were 93% and 95%, respectively. In that study, clinical assessments were performed on 20-fold magnified images. Sensitivity and specificity of clinical diagnosis made by an experienced observer were 81% and 95%. When an untrained dermatologist assessed the images, sensitivity and specificity of clinical diagnosis were 74% and 75%. Another study carried out by Bauer et al revealed that diagnostic accuracy of computer-aided dermoscopy was higher than dermoscopy by a trained dermatologist [15].
Visiomed AG, our image analysis software, has been trained using images collected from European countries. However, it could detect melanoma in Iranian patients with a high level of accuracy. The accuracy of the computerized dermoscopy system in our study is comparable with that of the most accurate reported systems [5,6,16].
Comparing clinical examination with dermoscopy score for diagnosis of melanoma in our study, there is no considerable difference in sensitivity and specificity. However, a larger study with higher power may detect a possible difference.
Conclusion
This system could not help us to reduce unnecessary excisions or improve early melanoma detection. Nevertheless, it may improve the diagnostic accuracy of an inexperienced clinician in the clinical evaluation of PSLs and represent a useful tool for screening of melanoma, particularly at centers not experienced in the field of PSLs. However, the cost benefit ratio of using this system needs to be assessed in developing countries with a low incidence of melanoma. Furthermore, it is of paramount importance to clarify that computer analysis has been developed in order to assist and not to replace physicians in the diagnosis of PSLs.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
MB participated in the design of the study and oversaw the drafting process and gave critical inputs to the manuscript.
HG participated in the design of the study, evaluation of the patients, and dermoscopic analysis of the lesions and oversaw the drafting process and gave critical inputs to the manuscript.
PM participated in the design of the study, evaluation of the patients, dermoscopic analysis of the lesions, histological evaluation of the lesions, statistical analysis, and drafting the manuscript.
AT participated in the design of the study, statistical analysis, and drafting the manuscript.
ZSN participated in histological evaluation of the lesions.
MA participated in histological evaluation of the lesions.
All authors read and approved the final manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
==== Refs
Diepgen TL Mahler V The epidemiology of skin cancer Br J Dermatol 2002 146 1 6 11966724 10.1046/j.1365-2133.146.s61.2.x
Cassileth BR Clark WH JrLusk EJ Frederick BE Thompson CJ Walsh WP How well do physicians recognize melanoma and other problem lesions? J Am Acad Dermatol 1986 14 555 560 3958271
Binder M Steiner A Schwarz M Knollmayer S Wolff K Pehamberger H Application of an artificial neural network in epiluminescence microscopy pattern analysis of pigmented skin lesions: a pilot study Br J Dermatol 1994 130 460 465 8186110
Piccolo D Ferrari A Peris K Diadone R Ruggeri B Chimenti S Dermoscopic diagnosis by a trained clinician vs. a clinician with minimal dermoscopy training vs. computer-aided diagnosis of 341 pigmented skin lesions: a comparative study Br J Dermatol 2002 147 481 486 12207587 10.1046/j.1365-2133.2002.04978.x
Blum A Luedtke H Ellwanger U Schwabe R Rassner G Garbe C Digital image analysis for diagnosis of cutaneous melanoma. Development of a highly effective computer algorithm based on analysis of 837 melanocytic lesions Br J Dermatol 2004 151 1029 1038 15541081 10.1111/j.1365-2133.2004.06210.x
Hoffmann K Gambichler T Rick A Kreutz M Anschuetz M Grunendick T Orlikov A Gehlen S Perotti R Andreassi L Newton Bishop J Cesarini JP Fischer T Frosch PJ Lindskov R Mackie R Nashan D Sommer A Neumann M Ortonne JP Bahadoran P Penas PF Zoras U Altmeyer P Diagnostic and neural analysis of skin cancer (DANAOS). A multicentre study for collection and computer-aided analysis of data from pigmented skin lesions using digital dermoscopy Br J Dermatol 2003 149 801 809 14616373 10.1046/j.1365-2133.2003.05547.x
NIH Consensus conference Diagnosis and treatment of early melanoma JAMA 1992 268 1314 1319 1507379 10.1001/jama.268.10.1314
Barnhill RL Current status of the dysplastic melanocytic nevus J Cutan Pathol 1991 18 147 159 1918502
Steiner A Pehamberger H Wolff K In vivo epiluminescence microscopy of pigmented skin lesions. II. Diagnosis of small pigmented skin lesions and early detection of malignant melanoma J Am Acad Dermatol 1987 17 584 591 3668003
Nachbar F Stolz W Merkle T Cognetta AB Vogt T Landthaler M Bilek P Braun-Falco O Plewig G The ABCD rule of dermatoscopy. High prospective value in the diagnosis of doubtful melanocytic skin lesions J Am Acad Dermatol 1994 30 551 559 8157780
Bafounta ML Beauchet A Aegerter P Saiag P Is dermoscopy (epiluminescence microscopy) useful for the diagnosis of melanoma? Results of a meta-analysis using techniques adapted to the evaluation of diagnostic tests Arch Dermatol 2001 137 1343 1350 11594860
Ercal F Chawla A Stoecker WV Lee HC Moss RH Neural network diagnosis of malignant melanoma from color images IEEE Trans Biomed Eng 1994 41 837 845 7959811 10.1109/10.312091
Seidenari S Pellacani G Giannetti A Digital videomicroscopy and image analysis with automatic classification for detection of thin melanomas Melanoma Res 1999 9 163 171 10380939
Farina B Bartoli C Bono A Colombo A Lualdi M Tragni G Marchesini R Multispectral imaging approach in the diagnosis of cutaneous melanoma: potentiality and limits Phys Med Biol 2000 45 1243 1254 10843103 10.1088/0031-9155/45/5/312
Bauer P Cristofolini P Boi S Burroni M Dell'Eva G Micciolo R Cristofolini M Digital epiluminescence microscopy: usefulness in the differential diagnosis of cutaneous pigmentary lesions. A statistical comparison between visual and computer inspection Melanoma Res 2000 10 345 349 10985668 10.1097/00008390-200008000-00005
Rosado B Menzies S Harbauer A Pehamberger H Wolff K Binder M Kittler H Accuracy of computer diagnosis of melanoma: a quantitative meta-analysis Arch Dermatol 2003 139 361 367 12622631 10.1001/archderm.139.3.361
Seidenari S Pellacani G Pepe P Digital videomicroscopy improves diagnostic accuracy for melanoma J Am Acad Dermatol 1998 39 175 181 9704824
| 16000171 | PMC1177938 | CC BY | 2021-01-04 16:29:46 | no | BMC Dermatol. 2005 Jul 6; 5:8 | utf-8 | BMC Dermatol | 2,005 | 10.1186/1471-5945-5-8 | oa_comm |
==== Front
BMC Evol BiolBMC Evolutionary Biology1471-2148BioMed Central London 1471-2148-5-361593264510.1186/1471-2148-5-36Research ArticleAn emerging phylogenetic core of Archaea: phylogenies of transcription and translation machineries converge following addition of new genome sequences Brochier Céline [email protected] Patrick [email protected] Simonetta [email protected] Laboratoire EGEE (Evolution, Génomique, Environnement) Université Aix-Marseille I, Centre Saint-Charles, Case 36, 3 Place Victor Hugo, 13331 Marseille, Cedex 3, France2 Unite Biologie Moléculaire du Gène chez les Extremophiles, Institut Pasteur, 25 rue du Dr. Roux, 75724 Paris Cedex 15, France3 Atelier de Bioinformatique, Université Paris 6, 12 rue Cuvier, Paris, France2005 2 6 2005 5 36 36 13 1 2005 2 6 2005 Copyright © 2005 Brochier et al; licensee BioMed Central Ltd.2005Brochier 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 concept of a genomic core, defined as the set of genes ubiquitous in all genomes of a monophyletic group, has become crucial in comparative and evolutionary genomics. However, it is still a matter of debate whether lateral gene transfers (LGT) may affect the components of genomic cores, preventing their use to retrace species evolution. We have recently reconstructed the phylogeny of Archaea by using two large concatenated datasets of core proteins involved in translation and transcription, respectively. The resulting trees were largely congruent, showing that informational gene components of the archaeal genomic core belonging to two distinct molecular systems contain a coherent signal for archaeal phylogeny. However, some incongruence remained between the two phylogenies. This may be due either to undetected LGT and/or to a lack of sufficient phylogenetic signal in the datasets.
Results
We present evidence strongly favoring of the latter hypothesis. In fact, we have updated our transcription and translation datasets with five new archaeal genomes for a total of 6384 and 2928 amino acid positions, respectively, and 25 taxa. This increase in taxonomic sampling led to the nearly complete convergence of the transcription-based and translation-based trees on a single phylogenetic pattern for archaeal evolution. In fact, only a single incongruence persisted between the two phylogenies. This concerned Methanopyrus kandleri, whose placement remained strongly biased in the transcription tree due to its above average evolutionary rates, and could not be counterbalanced due to the lack of availability of closely related and/or slower-evolving relatives.
Conclusion
To our knowledge, this is the first report of evidence that the phylogenetic signal harbored by components of the archaeal translation apparatus is confirmed by additional markers belonging to a second molecular system (i.e. transcription). This rules out the risk of circularity when inferring species evolution by small subunit ribosomal RNA and ribosomal protein sequences, since it has been suggested that concerted LGT may affect these markers. Our results strongly support the existence of a core of proteins that has evolved mainly through vertical inheritance in Archaea, and carries a bona fide phylogenetic signal that can be used to retrace the evolutionary history of this domain. The identification and analysis of additional molecular markers not affected by LGT should continue defining the emerging picture of a genuine phylogenetic core for the third domain of life.
==== Body
Background
The discovery that Lateral Gene Transfers (LGT) play a major role in the evolution of prokaryotic organisms has raised concerns about the possibility of reconstructing species phylogenies [1]. Some biologists even believe that LGT have obscured the phylogenetic record to such an extent that the task may be in fact hopeless [1,2]. However, others have argued that the careful selection of a 'core' of genes that have been refractory to transfer may help solving (at least partly) this conundrum [3-6]. The genomic core concept, i.e. the set of homologous genes present in all -or most-genomes of a phylogenetically coherent group, has become crucial in comparative and evolutionary genomics [7]. Indeed, the identification of 'genomic cores' can provide crucial information on the composition of ancestral genomes [8,9], as well as on organisms evolution at various phylogenetic depths [6,10,11]. However, homology-based analyses to define core genes cannot discriminate between vertically transmitted components and horizontally exchanged ones (i.e. "cryptic orthologous replacements" [7]). Thus, it is still a matter of debate whether Lateral Gene Transfers (LGT) may affect the components of genomic cores, preventing their use to retrace species evolution. Nevertheless, the extent and nature of the horizontal component of genomic cores can be identified by molecular phylogeny. The use of conserved gene cores to retrace species evolution has mainly focused on the translation apparatus, since the ribosome appears to be one of the best conserved macromolecular machines in the living world. The concatenation of either bacterial and archaeal ribosomal protein sequences has produced global phylogenies that are roughly similar to those obtained with both small and large ribosomal subunit rRNA genes (16S and 23S rRNA) [4,5,11]. Moreover, careful individual analyses have indicated that ribosomal proteins have been apparently never exchanged between the three Life domains, and rarely between different lineages within domains [4,5,12,13]. However, it may be argued that concerted LGT involving rRNA and ribosomal protein genes, since they belong to the same macromolecular machinery, could escape detection in such analyses. Nevertheless, this hypothesis could be discarded if phylogenies based on additional sets of genomic core proteins belonging to other molecular machineries are congruent with those of the components of the translation apparatus.
Accordingly, we have recently performed an in-depth analysis of proteins involved in transcription and translation from Archaea [11]. Individual phylogenies of these proteins confirmed that the components of these informational molecular machineries are little affected by LGT in the archaeal domain, and permitted the assembly of two large concatenated datasets of likely vertically-transmitted genes to reconstruct the phylogeny of the third domain of life [11]. The trees based on the 'translation' dataset (53 ribosomal proteins, Figure 1A) and the 'transcription' dataset (11 RNA polymerase subunits and 3 transcription factors, Figure 1B) were globally congruent, suggesting that the two informational systems contain a coherent phylogenetic signal for the archaeal phylogeny [11]. However, a number of incongruent nodes remained between the two trees (Figure 1A and 1B). First, the hyperthermophilic methanogen Methanopyrus kandleri was close to other methanogens in the translation tree (Figure 1A), whereas it emerged with a strong statistical confidence at the base of the euryarchaeal phylum in the transcription tree (Bootstrap Value BV = 90%, Figure 1B). A second incongruence concerned the position of the euryarchaeon Archaeoglobus fulgidus, since this archaeon was grouped, albeit with weak support (BV = 41%), with Thermoplasmatales in the translation tree (Figure 1A), whereas in the transcription tree it was strongly placed as sister group to the clade composed of Methanosarcinales and Halobacteriales (BV = 100%) (Figure 1B). Finally, although in both phylogenies Methanobacteriales and Methanococcales were located in-between Thermococcales and a large clade comprising Thermoplasmatales, Archaeoglobus, Methanosarcinales and Halobacteriales, they were paraphyletic in the translation tree (Figure 1A) whereas they were monophyletic in the transcription tree (Figure 1B).
Figure 1 Unrooted Maximum Likelihood (ML) trees based on concatenation of ribosomal proteins (A) and RNA polymerase subunits and transcription factors (B). For clarity, the two datasets from Brochier et al. (2004) were scaled down to the same number of species by removing Methanosarcina mazei and Methanosarcina acetivorans from the transcription dataset. Numbers at nodes are bootstrap values (BV). The scale bars represent the number of changes per position for a unit branch length. Trees were produced by exhaustive searches performed by PROTML. Branch lengths and likelihood values were calculated by TREE-PUZZLE (JJT model including a Γ-correction (8 categories of sites)). Numbers at nodes are bootstrap values computed with PUZZLEBOOT from 1000 replications. Asterisks indicate constrained nodes (supported by BV = 100% in preliminary NJ and heuristic ML analyses). The names of groups showing incongruent positions between the two trees are underlined.
In the case of M. kandleri, we suggested that the discrepancy between the translation and transcription trees was likely due to the very fast evolutionary rate of its RNA polymerase subunits (reflected by the very long branch of M. kandleri in the transcription tree, Figure 1B). Such an accelerated rate of evolution may be due to the lack in this archaeon of the critical transcription factor TFS [11,14]. A Long Branch Attraction (LBA) artefact [15] between the very long branch of M. kandleri and the outgroup (i.e. Crenarchaeota) may thus be responsible for the basal position of this methanogen in the transcription tree. In contrast, the incongruence between the two trees in the position of A. fulgidus, and in those of Methanobacteriales and Methanococcales may be either due to undetected LGT, and/or result from an insufficient phylogenetic signal in the two protein datasets. Only in the latter case should an increased taxonomic sampling help resolving this incongruence, whereas if LGTs are responsible, the addition of more taxa should not increase resolution and will possibly add more confusion. The recent sequencing of several new genomes from Euryarchaeota now permits tackling these two alternatives.
Results and discussion
We have updated our previous datasets of the components of the translation and transcription machineries [11] to include a total of 25 Archaea. In particular, we included the psychrophilic methanogen Methanogenium frigidum [16] and the mesophilic methanogen Methanococcoides burtonii [16]-two lineages belonging to Methanomicrobiales [17] and Methanosarcinales [18], respectively. We also included the halophile Haloferax volcanii [19], the Thermococcale Thermococcus gammatolerans (Yvan Zivanovic and Fabrice Confalonieri, personal communication), and Nanoarchaeum equitans, a highly divergent archaeon that has been suggested as the representative of a new archaeal phylum, the Nanoarchaeota [20-22]. As in our previous studies [5,11] we did not include any eukaryotic outgroup in order to limit biases due to LBA.
As described previously [11], separated phylogenetic analyses were performed on each of these new datasets in order to identify and remove potential lateral gene transfer (LGT) events (data not shown). Despite the fact that most relationships were largely unresolved in several trees due to the small size of most proteins, we checked for any possible strongly supported departure from undisputed nodes in the archaeal phylogeny, such as the clades of Thermoplasmatales, Halobacteriales, Sulfolobales, Thermococcales, Methanosarcinales and Methanococcales. Following the addition of novel taxa, no new clear-cut case of LGT could be observed with respect to these nodes, confirming that transfers are indeed very rare for these markers[5,11]. A few proteins gave an instable placement for Nanoarchaeum equitans. However, since the position of this taxon in the archaeal phylogeny has not yet been firmly tested, we did not judge these proteins as clear-cut cases of LGT. The 53 ribosomal proteins and the 14 proteins involved in transcription were thus concatenated into two large 'translation' and 'transcription' datasets, whose sizes were 6384 and 2928 amino acid positions, respectively.
Exhaustive Maximum Likelihood (ML) searches were performed on the two updated translation and transcription fusion datasets, with a few constraints given to undisputable nodes (i.e. supported by BV = 100% in preliminary Neighbor Joining and ML heuristic analyses (not shown)). The best ML topologies for the translation and transcription datasets are presented in Figure 2A and 2B, respectively. Topologies not significantly less likely than the ones presented in Figure 2 differed by minor rearrangements on nodes that are feebly supported by bootstrap, such as the branching order within halobacteriales in the transcription tree, or the grouping of methanopyrus with methanococcales/methanobacteriales in the translation tree (not shown).
Figure 2 Unrooted ML trees for the updated translation dataset (A) and transcription dataset (B). For details and tree computation see legend to Figure 1. The five new taxa are indicated in bold.
Interestingly, the addition of five new archaeal taxa led to convergence of the transcription and translation trees on a coherent phylogenetic pattern, to the exclusion of the position of M. kandleri-still emerging after Thermococcales in the translation tree (Figure 2A), but recovered as a very long branch at the base of the euryarchaea in the transcription tree (Figure 2B)-(the same trees were obtained when removing M. kandleri from the datasets, data not shown). This only incoherence between the two phylogenies is most likely due to the fact that the LBA artefact affecting the position of M. kandleri in the transcription tree persisted even after increase in taxonomic sampling, due to unavailability of closely related and slower evolving species. Nanoarchaeum equitans emerged as a separate branch distinct from those leading to Crenarchaeota and Euryarchaeota domains, in both translation and transcription trees (Figure 2A and 2B), supported by strong bootstrap values (BV = 100%). This position is congruent with previous results based on ribosomal proteins concatenation [22]. T. gammatolerans branched off of at the base of Thermococcales, that were confirmed as the first emerging euryarchaeal phylum, as in our previous studies [5,11]. Interesting, Methanobacteriales and Methanococcales formed now a monophyletic group in both translation and transcription trees (BV = 55% and 82%, Figure 2A and 2B, respectively). This suggests that the paraphyly of these groups observed in our previous translation tree (Figure 1A) was likely incorrect due to a lack of phylogenetic signal rather than to a LGT bias. The Methanobacteriales/Methanococcales monophyletic group is sister to a large cluster including both methanogenic and non-methanogenic species: A. fulgidus, the three Thermoplasmatales, the three Halobacteriales and the five Methanomicrobia (Methanomicrobiales and Methanosarcinales) (BV = 97% and BV = 77%, Figure 2A and 2B, respectively). This supports the hypothesis of an ancient origin of methanogenesis in Archaea followed by subsequent loss in some lineages (A. fulgidus, Thermoplasmatales and Halobacteriales). Moreover, the position of A. fulgidus, while left uncertain in our previous analyses (i.e. either sister-group of Thermoplasmatales in the translation tree, or of Halobacteriales/Methanomicrobia in the transcription tree, Figure 1A and 1B, respectively) was now robustly indicated as sister to Methanomicrobiales, Methanosarcinales and Halobacteriales in both translation and transcription trees (BP = 85% and 96%, respectively, Figure 1A and 1B). The strong placement of A. fulgidus in our updated translation tree is likely due the stabilisation of the node following addition of new taxa. This result further supports the hypothesis of a late and independent emergence of aerobic respiration in Euryarchaeotes (Halobacteriales), possibly via the recruitment of bacterial genes. Finally, both translation and transcription trees confidently grouped M. burtonii and M. frigidum with the three Methanosarcina (BV = 100% and BV = 100%, Figure 2A and BV = 100% and BV = 79%, Figure 2B, respectively) within the Methanomicrobia group. The very close relationship between M. burtonii and the three Methanosarcinales constitutes a novel phylogenetic argument justifying its inclusion in the order Methanosarcinales, at present based only on 16S rRNA phylogeny [18].
Conclusion
The congruence we obtained between the archaeal phylogenies based on the components of the translation and transcription machineries strongly supports the existence of a core of genes that evolved mainly through vertical inheritance in Archaea, and carry a bona fide phylogenetic signal that can be used to infer the phylogeny of this domain. Our results confirm also that the addition of new taxa strongly improves phylogenetic inference, and support the idea that evolutionary considerations should be included in the choice of new genomes to be sequenced. However, our conclusions should not be considered as the "last word" on the subject. For example, the misplacement of M. kandleri at the base of Euryarchaea in the transcription tree was not cured by the increase in taxonomic sampling. The inclusion of sequences from slower evolving and close relatives, when they will be available, may help resolving this bias. Similarly, the very long branch displayed by N. equitans (Figure 2A and 2B) suggests that its placement as a separate branch distinct from that leading to Euryarchaeota and Crenarchaeota, although congruent between the transcription and translation trees, should be taken with caution due to the risk of an LBA artefact. The analysis of the components of additional molecular systems and the inclusion of more taxa may eventually lead to a confident placement for these two interesting species in the archaeal phylogeny.
Finally, our results make us confident that the construction of a phylogeny that retraces the vertical history of the archaeal domain is a feasible task. The identification and analysis of additional molecular markers not affected by LGT on large phylogenetic scales and their phylogenetic analysis by approaches that minimise reconstruction artefacts should continue defining the emerging picture of a genuine phylogenetic core for the Archaea. The application of a similar strategy to the bacterial and eukaryal domains could also lead to a bona fide reconstruction of their respective evolutionary histories.
Methods
In order to update the datasets of our previous analysis [11], we included the two methanogens Methanogenium frigidum [16] and Methanococcoides burtonii [16], the halophile Haloferax volcanii [19], the Thermococcale Thermococcus gammatolerans (Yvan Zivanovic and Fabrice Confalonieri, personal communication), and Nanoarchaeum equitans. In addition, the sequences of the two Methanosarcinales Methanosarcina mazei and Methanosarcina acetivorans were added to the ribosomal dataset. Sequences were retrieved by TBLASTN at genome sequencing web sites for H. volcanii , M. burtonii and M. frigidum , or using BLASTP in NCBI for N. equitans, M. acetivorans and M. mazei [23]. For each dataset, novel sequences were manually added to previous alignments by using the ED program of the MUST package [24]. Regions were the alignment was ambiguous were removed from the each dataset.
Trees were computed by a number of different approaches. Neighbor-Joining (NJ) trees were calculated by the NEIGHBOR program of the PHYLIP package [25], using Maximum Likelihood (ML) distance matrices (JTT model including a Γ-correction) computed by TREE-PUZZLE 5.1 [26]. Heuristic ML trees were computed using PHYML with the JTT model including a Γ-correction [27]. Exhaustive tree topology searches with limited constraints were performed using PROTML of the MOLPHY package [28]. The likelihoods and branch lengths of ML topologies were performed by TREE-PUZZLE (JTT model including a Γ-correction). For exhaustive ML searches, constraints (asterisks in Figures 1 and 2) were given to undisputable nodes (supported by BV = 100%), based on preliminary NJ and ML heuristic analyses (not shown).
The SEQBOOT program of the PHYLIP package [25] was used for the generation of bootstrapped datasets, and PUZZLEBOOT [29] and CONSENSE in the PHYLIP package [25] were used for bootstrap value calculations on 1000 replications and consensus tree reconstructions, respectively.
Datasets and their corresponding phylogenies are available on request from CB.
Authors' contributions
CB carried out the analyses. CB, PF and SG conceived the study and drafted the manuscript. All authors read and approved the final manuscript.
Acknowledgements
We wish to thank Eric Armanet and Gael Stefan for allowing some of the calculations to be run on their computers. We thank also Yvan Zivanovic, Fabrice Confalonieri for kindly providing sequences from T. gammatolerans and Shiladitya DasSarma and the members of the university of Scranton for the sequences of H. volcanii free available by BLAST on .
==== Refs
Doolittle WF Phylogenetic classification and the universal tree Science 1999 284 2124 2129 10381871 10.1126/science.284.5423.2124
Nesbo CL Boucher Y Doolittle WF Defining the core of nontransferable prokaryotic genes: the euryarchaeal core J Mol Evol 2001 53 340 350 11675594 10.1007/s002390010224
Jain R Rivera MC Lake JA Horizontal gene transfer among genomes: The complexity hypothesis Proc Natl Acad Sci U S A 1999 96 3801 3806 10097118 10.1073/pnas.96.7.3801
Brochier C Bapteste E Moreira D Philippe H Eubacterial phylogeny based on translational apparatus proteins Trends Genet 2002 18 1 5 11750686 10.1016/S0168-9525(01)02522-7
Matte-Tailliez O Brochier C Forterre P Philippe H Archaeal phylogeny based on ribosomal proteins Mol Biol Evol 2002 19 631 639 11961097
Lerat E Daubin V Moran NA From Gene Trees to Organismal Phylogeny in Prokaryotes:The Case of the gamma-Proteobacteria PLoS Biol 2003 1 E19 12975657 10.1371/journal.pbio.0000019
Charlebois RL Doolittle WF Computing prokaryotic gene ubiquity: rescuing the core from extinction Genome Res 2004 14 2469 2477 15574825 10.1101/gr.3024704
Koonin EV Comparative genomics, minimal gene-sets and the last universal common ancestor Nat Rev Microbiol 2003 1 127 136 15035042 10.1038/nrmicro751
Harris JK Kelley ST Spiegelman GB Pace NR The genetic core of the universal ancestor Genome Res 2003 13 407 412 12618371 10.1101/gr.652803
Daubin V Gouy M Perriere G A phylogenomic approach to bacterial phylogeny: evidence of a core of genes sharing a common history Genome Res 2002 12 1080 1090 12097345 10.1101/gr.187002
Brochier C Forterre P Gribaldo S Archaeal phylogeny based on proteins of the transcription and translation machineries: tackling the Methanopyrus kandleri paradox Genome Biol 2004 5 R17 15003120 10.1186/gb-2004-5-3-r17
Makarova KS Ponomarev VA Koonin EV Two C or not two C: recurrent disruption of Zn-ribbons, gene duplication, lineage-specific gene loss, and horizontal gene transfer in evolution of bacterial ribosomal proteins Genome Biol 2001 2 RESEARCH0033.
Coenye T Vandamme P Organisation of the S10, spc and alpha ribosomal protein gene clusters in prokaryotic genomes FEMS Microbiol Lett 2005 242 117 126 15621428 10.1016/j.femsle.2004.10.050
Lange U Hausner W Transcriptional fidelity and proofreading in Archaea and implications for the mechanism of TFS-induced RNA cleavage Mol Microbiol 2004 52 1133 1143 15130130 10.1111/j.1365-2958.2004.04039.x
Felsenstein J Cases in which parsimony or compatibility methods will be positively misleading Syst Zool 1978 27 401 410
Saunders NF Thomas T Curmi PM Mattick JS Kuczek E Slade R Davis J Franzmann PD Boone D Rusterholtz K Feldman R Gates C Bench S Sowers K Kadner K Aerts A Dehal P Detter C Glavina T Lucas S Richardson P Larimer F Hauser L Land M Cavicchioli R Mechanisms of thermal adaptation revealed from the genomes of the Antarctic Archaea Methanogenium frigidum and Methanococcoides burtonii Genome Res 2003 13 1580 1588 12805271 10.1101/gr.1180903
Franzmann PD Liu Y Balkwill DL Aldrich HC Conway de Macario E Boone DR Methanogenium frigidum sp. nov., a psychrophilic, H2-using methanogen from Ace Lake, Antarctica Int J Syst Bacteriol 1997 47 1068 1072 9336907
Franzmann PD Springer N Ludwig W Conway de Macario E Rohde M A methanogenic Archaeon from Ace Lake, Antarctica: Methanococcoides burtonii sp. nov. Syst Appl Microbiol 1992 15 573 581
DasSarma S http://halo.umbi.umd.edu/cgi-bin/blast/blast_hvo.pl
Huber H Hohn MJ Rachel R Fuchs T Wimmer VC Stetter KO A new phylum of Archaea represented by a nanosized hyperthermophilic symbiont Nature 2002 417 63 67 11986665 10.1038/417063a
Huber H Hohn MJ Stetter KO Rachel R The phylum Nanoarchaeota: present knowledge and future perspectives of a unique form of life Res Microbiol 2003 154 165 171 12706504 10.1016/S0923-2508(03)00035-4
Waters E Hohn MJ Ahel I Graham DE Adams MD Barnstead M Beeson KY Bibbs L Bolanos R Keller M Kretz K Lin X Mathur E Ni J Podar M Richardson T Sutton GG Simon M Soll D Stetter KO Short JM Noordewier M The genome of Nanoarchaeum equitans: insights into early archaeal evolution and derived parasitism Proc Natl Acad Sci U S A 2003 100 12984 12988 14566062 10.1073/pnas.1735403100
Altschul SF Gish W Miller W Myers EW Lipman DJ Basic local alignment search tool J Mol Biol 1990 215 403 410 2231712 10.1006/jmbi.1990.9999
Philippe H MUST, a computer package of Management Utilities for Sequences and Trees Nucleic Acids Res 1993 21 5264 5272 8255784
Felsenstein J Phylogeny Inference Package (Version 3.2). Cladistics 1989 5 164 166
Schmidt HA Strimmer K Vingron M von Haeseler A TREE-PUZZLE: maximum likelihood phylogenetic analysis using quartets and parallel computing Bioinformatics 2002 18 502 504 11934758 10.1093/bioinformatics/18.3.502
Guindon S Gascuel O A simple, fast, and accurate algorithm to estimate large phylogenies by maximum likelihood Syst Biol 2003 52 696 704 14530136 10.1080/10635150390235520
Adachi J Hasegawa M MOLPHY version 2.3: programs for molecular phylogenetics based on maximum likelihood Comput Sci Monogr 1996 28 1 150
Holder ME Roger AJ A shell-script program called "puzzleboot" that allows the analysis of multiple data sets with PUZZLE even though PUZZLE lacks the "M" option of many PHYLIP programs 2002
| 15932645 | PMC1177939 | CC BY | 2021-01-04 16:37:17 | no | BMC Evol Biol. 2005 Jun 2; 5:36 | utf-8 | BMC Evol Biol | 2,005 | 10.1186/1471-2148-5-36 | oa_comm |
==== Front
BMC Endocr DisordBMC Endocrine Disorders1472-6823BioMed Central London 1472-6823-5-61596322810.1186/1472-6823-5-6Research ArticleInsulin resistance in adolescents with Down syndrome: a cross-sectional study Fonseca Cristina T [email protected] Daniela M [email protected] Márcia G [email protected] Izabel CR [email protected]ães Marília M [email protected] Post-Graduate Program of Endocrinology, Medicine School, Hospital Universitário Clementino Fraga Filho (HUCFF), Federal University of Rio de Janeiro (UFRJ) - Av. Brigadeiro Trompowski, s/n, HUCFF, Ilha do Fundão, Rio de Janeiro, Brazil2 Genetics Department, Instituto de Puericultura e Pediatria Martagão Gesteira (IPPMG), UFRJ - Av. Brigadeiro Trompowski, s/n, HUCFF, Ilha do Fundão, Rio de Janeiro, Brazil3 Pediatrics Department, IPPMG, UFRJ - Av. Brigadeiro Trompowski, s/n, HUCFF, Ilha do Fundão, Rio de Janeiro, Brazil4 Endocrinology Department - HUCFF, UFRJ - Av. Brigadeiro Trompowski, s/n, HUCFF, Ilha do Fundão, Rio de Janeiro, Brazil2005 17 6 2005 5 6 6 3 2 2005 17 6 2005 Copyright © 2005 da Fonseca et al; licensee BioMed Central Ltd.2005da Fonseca 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 prevalence of diabetes mellitus is higher in individuals with Down syndrome (DS) than in the general population; it may be due to the high prevalence of obesity presented by many of them. The aim of this study was to evaluate the insulin resistance (IR) using the HOMA (Homeostasis Model Assessment) method, in DS adolescents, describing it according to the sex, body mass index (BMI) and pubertal development.
Methods
15 adolescents with DS (8 males and 7 females) were studied, aged 10 to 18 years, without history of disease or use of medication that could change the suggested laboratory evaluation. On physical examination, the pubertal signs, acanthosis nigricans (AN), weight and height were evaluated. Fasting plasma glucose and insulin were analysed by the colorimetric method and RIA-kit LINCO, respectively. IR was calculated using the HOMA method. The patients were grouped into obese, overweight and normal, according to their BMI percentiles. The EPIINFO 2004 software was used to calculate the BMI, its percentile and Z score.
Results
Five patients were adults (Tanner V or presence of menarche), 9 pubertal (Tanner II – IV) and 1 prepubertal (Tanner I). No one had AN. Two were obese, 4 overweight and 9 normal. Considering the total number of patients, HOMA was 1.7 ± 1.0, insulin 9.3 ± 4.8 μU/ml and glucose 74.4 ± 14.8 mg/dl. The HOMA values were 2.0 ± 1.0 in females and 1.5 ± 1.0 in males. Considering the nutritional classification, the values of HOMA and insulin were: HOMA: 3.3 ± 0.6, 2.0 ± 1.1 and 1.3 ± 0.6, and insulin: 18.15 ± 1.6 μU/ml, 10.3 ± 3.5 μU/ml and 6.8 ± 2.8 μU/ml, in the obese, overweight and normal groups respectively. Considering puberty, the values of HOMA and insulin were: HOMA: 2.5 ± 1.3, 1.4 ± 0.6 and 0.8 ± 0.0, and insulin: 13.0 ± 5.8 μU/ml, 7.8 ± 2.9 μU/ml and 4.0 ± 0.0 μU/ml, in the adult, pubertal and prepubertal groups respectively.
Conclusion
The obese and overweight, female and adult patients showed the highest values of HOMA and insulin.
==== Body
Background
Down syndrome (DS) is a common chromosomal disorder, affecting 1 per 700 live births [1]. Life expectancy has increased in the last decades, due to the improvement of the scientific knowledge about the syndrome and its acquired complications, and also, due to the easier access and the more recent diagnostic and therapeutic means for the patients, their families and the multiprofessional health team that accompany them [2,3].
Early hypotonicity and typical dimorphisms are among the clinical characteristics of the syndrome. There is a higher incidence of diseases in various organic systems, including the endocrinologic system, such as diabetes mellitus among others [4].
Diabetes mellitus has a higher prevalence in DS than in the general population [4,5]. It can appear as an autoimmune disorder (type 1 diabetes mellitus) or as a disorder in which the insulin resistance is the predominant physiopathological factor (type 2 diabetes mellitus) [6-8].
Insulin resistance may present without any clinical manifestation, often, many years before the appearance of frank diabetes mellitus, which would occur when the pancreas failed to secrete enough insulin to compensate such resistance and keep the person euglycemic [6]. Therefore, it would be interesting to know the level of insulin resistance in this syndrome, since adolescence, as diabetes mellitus evolve with several complications which would further increase the morbidity and mortality of this population. This way, preventive measures could be applied, such as giving the patient adequate dietetic orientations and leading him to an early consultation with a nutrition service, stimulating and developing special programs of physical activity, treating those who have insulin resistance with medications, which may be a therapeutic option in the future.
The aim of this study was to estimate the insulin resistance (IR) through the HOMA (Homeostasis Model Assessment) method, in adolescents with DS, and to describe it according to the sex, body mass index (BMI) and presence of puberty.
Methods
A descriptive cross-sectional study was designed with 15 out patient adolescents (8 males and 7 females) with Down syndrome, aged 10 to 18 years.
Adolescents with previously known diabetes mellitus or impaired glucose tolerance (IGT) or impaired fasting glucose (IFG) were excluded from this study. Furthermore, the studied patients did not have any comorbidity and did not use any medication that could compromise the suggested laboratory evaluation, such as anti-hyperlipidemia drugs, topic or systemic glucocorticoids, anti-hypertensive drugs, sexual steroids and GnRH analogues.
The local Institutional Ethical Committee approved the study and the informed consent was obtained from all patients' parents or tutors.
The adolescents were measured and weighed; the stature was annotated in meters and the weight in kilograms. They were also submitted to physical examination, including the evaluation of the presence of acanthosis nigricans and the pubertal developmental degree according to Tanner's stage [9]. Considering puberty, they were grouped, as follows:
• Prepubertal – absence of sexual development;
• Pubertal – pubertal development according to Tanner's stage II to IV;
• Adult – pubertal development according to Tanner's stage V or menarche in girls.
Blood samples were collected after a 12-hour overnight fasting period, for the analysis of plasma glucose and insulin. The glucose levels were analysed by the colorimetric method, considering the normal range from 70 to 100 mg/dl, according to the American Diabetes Association, 2003. Insulin levels were analysed by a radioimmunoassay kit (Linco Research Incorporation). This kit is specific for human insulin and does not cross react with pro-insulin (<0.2%). The normal fasting range, for adults, after an 18-hour fasting period, varies from 5 to 15 μU/ml. The method's sensitivity is <2 μU/ml, specificity is 100% and the range value within and between assay variations are 3.1% and 6.0% respectively.
Insulin resistance was calculated according to the HOMA method, through the formula: {[glucose(mg/dl)/18] X insulin(μU/mL)}/ 22.5 [10].
The EPIINFO software version 2004, provided the BMI and its percentile and Z score.
The adolescents were grouped by the nutritional status according to the National Center for Health Statistics 2000 – Center for Disease Control and Prevention(CDC – NCHS) as follows [11]:
• Undernourished – BMI percentile below 5;
• Normal -BMI percentiles 5 to 85;
• Overweight -BMI percentiles 85 to 95;
• Obese – BMI percentile above 95.
Results
The mean age of the patients was 12.9 ± 2.6 years (varying from 10 to 18 years); none had acanthosis nigricans. Considering puberty, 5 were adults,9 pubertal and 1 prepubertal, and regarding the nutritional status, 2 (13.3%) were obese, 4 (26.7%) overweight and 9 (60%) normal (Table 1). The mean BMI was 20.9 ± 4.3 (from 15.5 to 29.6) and the mean BMI Z score was +0.55 ± 0.9.
Table 1 Distribution of patients' data.
CA Sex BMI BMIp BMI Z Puberty Glucose (mg/dl) Insulin (μU/ml) HOMA
10y F 15.7 29.01 -0.55 Pubertal 52 4.4 0.6
10y5m M 17.8 64.79 0.38 Pubertal 71 5.4 0.9
10y7m F 20.7 86.11 1.09 Pubertal 74 10.9 2.0
10y9m F 19.9 80.09 0.84 Pubertal 74 13.7 2.5
10y9m M 15.5 20.5 -0.82 Prepubertal 85 4.0 0.8
11y4m F 26.1 96.81 1.85 Adult 60 19.3 2.9
11y11m M 17.2 39.23 -0.27 Pubertal 72 6.0 1.1
12y4m M 19.4 69.16 0.5 Pubertal 98 7.6 1.8
12y4m F 16 15.17 -1.03 Pubertal 81 6.9 1.4
13y1m M 22.6 88.02 1.18 Pubertal 53 7.4 1.0
13y1m M 23.7 91.93 1.4 Pubertal 69 8.0 1.4
15y6m F 19.8 44.94 -0.13 Adult 82 6.3 1.3
16y4m F 26.3 89.93 1.28 Adult 97 15 3.6
16y10m M 23.7 78 0.77 Adult 57 7.4 1.0
18y M 29.6 95.93 1.74 Adult 91 17 3.8
CA = chronological age; M = male; F = female; BMIp = BMI percentile; BMI Z = BMI Z score.
Considering the total number of patients, the mean values were: HOMA = 1.7 ± 1.0 (from 0.6 to 3.8), insulin = 9.3 ± 4.8 μU/ml (from 4.0 to 19.3 μU/ml) and glucose = 74.4 ± 14.8 mg/dl (from 52 to 98 mg/dl). Considering sex, the mean values of HOMA were 2.0 ± 1.0 in females and 1.5 ± 1.0 in males.
Considering the subgroups, divided by the nutritional status, the mean values were: HOMA: 3.3 ± 0.6, 2.0 ± 1.1 and 1.3 ± 0.6, and insulin: 18.15 ± 1.6 μU/ml, 10.3 ± 3.5 μU/ml and 6.8 ± 2.8 μU/ml, in the obese, overweight and normal, respectively (Figure 1).
Figure 1 Mean values of HOMA and insulin according to the nutritional classification. The highest values of HOMA and insulin were found in the obese, followed by the overweight, and lastly, by the normal-weighed patients.
Considering puberty, the mean values were: HOMA: 2.5 ± 1.3, 1.4 ± 0.6 and 0.8 ± 0.0, and insulin: 13.0 ± 5.8 μU/ml, 7.8 ± 2.9 μU/ml and 4.0 ± 0.0 μU/ml, in the adult, pubertal and prepubertal groups, respectively (Figure 2).
Figure 2 Mean values of HOMA and insulin according to the pubertal classification. The highest values of HOMA and insulin were found in the adult, followed by the pubertal, and lastly, by the prepubertal patients.
Discussion
Insulin resistance may be defined as a diminished response to the biological actions of insulin; it may involve not only the carbohydrate metabolism, but also the lipid metabolism [12]. In pharmacological terms, it represents a state in which normal amounts of insulin produce a subnormal biological response. It is characterized by hyperinsulinemia and can be associated with normoglycemia or hyperglycemia [13].
The hyperinsulinemic euglycemic clamp has been regarded as the standard method to evaluate insulin resistance. Considering that it is a difficult and relatively invasive technique that requires a specialized team and, in general, it is not available to the clinical practice [14], researchers have been studying other more practical methods that would be able to measure insulin resistance. One of these studied methods is the HOMA method, a mathematic model based on measurement of fasting plasma glucose and insulin levels, which is especially useful for DS patients in whom the use of other methods based on the results of an oral glucose tolerance test (OGTT), for example, would be extremely difficult to perform. Although it is not as sensitive or reproducible a measure of IR as the clamp technique, it has been validated both in adults [10,14] and in children and adolescents without DS [15,16] and, for this reason, it has been extensively used, not only in epidemiological studies, but also in clinical practice [10].
In Down syndrome there is a high prevalence of overweight and obesity [17-19] and one study suggests that there may be a different distribution of body fat, more truncal than peripheral, in such children, which may represent only one more trait of the syndrome or be the result of the muscle hypotonicity that is well known to be part of the syndrome's characteristics, and this may reduce the activity level and energy needs of these patients [17]. Therefore, the higher frequency of type 2 diabetes mellitus in DS may be associated not only with the higher propensity for obesity, but also with large abdominal fat stores, reflecting larger amounts of visceral adiposity, which may contribute to insulin resistance [20,21] and consequently, to the development of type 2 diabetes mellitus in a variable period of time. Such correlation between type 2 diabetes mellitus and the higher proportion of body fat mass, mainly concentrated in the abdominal region, has also been demonstrated by some studies involving type 2 diabetic, non-obese [21] or obese individuals, without DS [13].
Android obesity, that is, the obesity in which there is accumulation of fat in the central region of the body, especially in the abdomen, is more associated with insulin resistance than the gynecoid obesity [20]. It is also known that the visceral adipocytes, when compared to the subcutaneous, have a higher basal lipolysis, are poorly sensitive to insulin and that such characteristics are amplified when they show hypertrophy [22]. Consequently, this increase in lipolysis offers more free fatty acids to the liver, which stimulates, this way, the liver production of glucose and the inhibition of glucose uptake and its oxidation in muscle tissue [23]. In an attempt to keep euglycemia, insulin secretion increases and the resultant compensatory hyperinsulinemia, in obese patients, reduces the expression of the membrane insulin receptor (down regulation), which generates more resistance to the insulin action [24].
In the present study, the highest values of HOMA were found in the obese and overweight patients, but we can not affirm if such patients really have insulin resistance, as there is no consensus for ideal HOMA values for children and adolescents and much less for patients with DS. No references about insulin resistance indexes in DS were found in literature. Nevertheless, comparing our results to those showed in some studies involving youth without DS, such as the study of Yeckel et al [25], we could say that none of our patients had IR because such authors found mean values of HOMA of about 7.00 for normal glucose tolerant children and adolescents without DS, based on clamp and OGTT studies, while our highest HOMA value was 3.8. Nevertheless, such study involved exclusively moderate to severe obese patients whose BMI was much higher than that of our obese patients (38.1 × 27.8 respectively), fact that may have contributed to the highest values of HOMA presented by them, confirming the direct relation between obesity and HOMA values, which is in accordance with our results. Moreover, our sample involved just 2 obese patients; most of them had normal weight. On the other hand, Allard et al showed lower mean values of HOMA, ranging from 0.83 to 1.62 in a representative sample of 2244 children and adolescents without DS [26]. A study performed in Rio de Janeiro, Brazil, with normal-weighed healthy students, also without DS, suggested a mean value of HOMA of 2.36 for girls and 2.66 for boys, using the same method of insulin dosage as the one used in our study [27]. Based on such reference values of Brazilian children and adolescents, we found 4 (26.7%) patients (3 females and 1 male) in our study with insulin resistance when assessed through the HOMA method. Among them, 2 had obesity, 1 overweight and the other, normal weight, that is, 100% (2/2) of the obese, 25% (1/4) of the overweight and 11.1% (1/9) of the normal weight groups had insulin resistance. Unfortunately, our small sample did not permit us to confirm these data with statistical analysis as referred in previous reports that correlated obesity and insulin resistance, in obese patients, without DS [28,29].
We also found higher HOMA values in adolescents with complete pubertal development than in the pubertal ones, which was similar to literature data; this fact demonstrates that puberty represents a period marked by higher levels of insulin resistance not only in healthy adolescents, but also in those with DS [13,30]. Nevertheless, healthy adolescents have physiological compensatory mechanisms that avoid the appearance of deleterious effects in the future that could result from this reduction in insulin action, whereas in DS, it may be that mechanisms fail in the adolescent years, resulting in higher insulin resistance levels than in the healthy population of the same age. The detection of higher HOMA values in females than in males has been previously described in literature, but only in adults and in clinically healthy patients [13].
Conclusion
We conclude that the obese and overweight, female and adult patients showed the highest values of HOMA and insulin. Nevertheless, additional studies are necessary, including larger samples and longitudinal follow-up in order to confirm the high prevalence of insulin resistance in Down syndrome, especially in the obese patients, and to verify if the patients with the highest values of HOMA may develop any glucose metabolism disturbance in the future.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
CTF and DMA selected and examined the patients, collected the blood samples, evaluated the analysis results and drafted the manuscript. MGR helped to select the patients and ICRB participated in the design of the study. MMG conceived the study and participated in its design and coordination and helped to draft the manuscript. All authors read and approved the final manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
We especially thank to Sérgio Franco Laboratory, on the person of Dalva Margareth Boelter dos Santos, for the insulin analysis.
==== Refs
Sugayama SMM Kim CA Setian N Anormalidades Cromossômicas Endocrinologia Pediátrica – Aspectos físicos e metabólicos do recém-nascido ao adolescente 2002 Sĕo Paulo, Brazil, Sarvier 638 639
Cooley WC Graham JM Jr Down syndrome – an update and review for the primary pediatrician Clin Pediatr 1991 30 233 242
Baird PA Sadovnick AD Life expectancy in Down syndrome J Pediatr 1987 110 849 854 2953876
Smith DS Health care management of adults with Down syndrome Am Fam Physician 2001 64 1031 1038 11578024
Milunsky A Neurath PW Diabetes mellitus in Down's syndrome Arch Environ Health 1968 17 372 376 Quoted from: Ohyama Y, Utsugi T, Uchiyama T, Hanaoka T, Tomono S, Kurabayashi M: Prevalence of diabetes in adult patients with Down's syndrome living in a residential home. Diabetes Care 2000, 23 (5): 705–706. 4233868
Amiel SA Buchanan CR Brook CGD, Hindmarsh PC Diabetes mellitus Clinical Pediatric Endocrinology 2001 England, Blackwell Science Ltd 411 439
Sperling MA Sperling MA Diabetes mellitus Pediatric Endocrinology 2002 United States of America, Saunders 323 366
Chen H Down syndrome Access on: June 2003
Marshall WA Tanner JM Variations in the pattern of pubertal changes in girls Arch Dis Child 1969 44 291 303 Quoted from: Rosenfield RL: Puberty in the female and its disorders. In: Pediatric Endocrinology. Edited by Sperling MA. United States of America, Saunders 2002, 455-518. 5785179
Matthews DR Hosker JP Rudenski AS Naylor BA Treacher DF Turner RC Homeostasis model assessment: insulin resistance and β-cell function from fasting plasma glucose and insulin concentrations in man Diabetologia 1985 28 412 419 3899825 10.1007/BF00280883
National Center for Health Statistics. National Center for Chronic Disease Prevention and Health Promotion and Center for Disease Control and Prevention Access on: July 2003
Wallace TM Matthews DR The assessment of insulin resistance in man Diabetic Medicine 2002 19 527 534 12099954 10.1046/j.1464-5491.2002.00745.x
Krentz AJ Fortnightly review: insulin resistance BMJ 1996 313 1385 1389 8956710
Bonora E Targher G Alberiche M Bonadonna RC Saggiani F Zenere MB Monauni T Muggeo M Homeostasis model assessment closely mirrors the glucose clamp technique in the assessment of insulin sensitivity Diabetes Care 2000 23 57 63 10857969
Gungor N Saad R Janosky J Arslanian S Validation of surrogate estimates of insulin sensitivity and insulin secretion in children and adolescents J Pediatr 2004 144 47 55 14722518 10.1016/j.jpeds.2003.09.045
Uwaifo GI Fallon EM Chin J Elberg J Parikh SJ Yanovski JA Indices of insulin action, disposal, and secretion derived from fasting samples and clamps in normal glucose-tolerant black and white children Diabetes Care 2002 25 2081 2087 12401760
Baer MT Waldron J Gumm H van Dyke DC Chang H van Dyke DC, Lang DJ, Heide F, van Duyne S, Soucek MJ Nutrition assessment of the child with Down syndrome Clinical perspectives in the management of Down syndrome 1990 United States of America, Springer-Verlag New York Inc 107 125
Cronk CE Chumlea WC Roche AF Assessment of overweight children with trisomy 21 Am J Ment Defic 1985 89 433 436 3156498
Al Husain M Body mass index for Saudi children with Down's syndrome Acta Paediatr 2003 92 1482 1485 14971803 10.1080/08035250310006656
Bjorntorp P Metabolic implications of body fat distribution Diabetes Care 1991 14 1132 1143 1773700
Banerji MA Chaiken RL Gordon D Kral JG Lebovitz HE Does intra-abdominal adipose tissue in black men determine whether NIDDM is insulin-resistant or insulin-sensitive? Diabetes 1995 44 141 146 7859931
McCarty MF Modulation of adipocyte lipoprotein lipase expression as a strategy for preventing or treating visceral obesity Med Hypotheses 2001 57 192 200 11461172 10.1054/mehy.2001.1317
Boden G Chen X Ruiz J White JV Rossetti L Mechanisms of fatty-acid induced inhibition of glucose uptake J Clin Invest 1994 93 2438 2446 8200979
Seely BL Olefsky JM Moller D Potential celular and genetic mechanisms for insulin resistance in the common disorders of diabetes and obesity Insulin resistance 1993 Chichester, Wiley 187 252
Yeckel CW Weiss R Dziura J Taksali SE Dufour S Burgert TS Tamborlane WV Caprio S Validation of insulin sensitivity indices from oral glucose tolerance test parameters in obese children and adolescents J Clin Endocrinol Metab 2004 89 1096 1101 15001593 10.1210/jc.2003-031503
Allard P Delvin EE Paradis G Hanley JA O'Loughlin J Lavallée C Levy E Lambert M Distribution of fasting plasma insulin, free fatty acids, and glucose concentrations and of Homeostasis Model Assessment of insulin resistance in a representative sample of Quebec children and adolescents Clin Chem 2003 49 644 649 12651818 10.1373/49.4.644
Pessoa CHCN Pessoa MBHCN Amaral DM Teixeira AM Cunha EF Guimarães MM Cordeiro JGH Estudo do Homeostasis Model Assessment (HOMA) em grupo de escolares saudáveis Arq Bras Endocrinol Metab 2003 47 101
Guzzaloni G Grugni G Mazzilli G Moro D Morabito F Comparison Between β-cell Function and Insulin Resistance Indexes in Prepubertal and Pubertal Obese Children Metabolism 2002 51 1011 1016 12145774 10.1053/meta.2002.34029
Barja S Arteaga A Acosta AM Hodgson MI Resistencia Insulínica y otras expresiones del síndrome metabólico em niños obesos chilenos Rev Méd Chile 2003 131 259 268
Amiel SA Sherwin RS Simonson DC Luaritano AA Tamborlane WV Impaired insulin action in puberty: a contributing factor to poor glycemic control during adolescence N Engl J Med 1986 315 215 219 3523245
| 15963228 | PMC1177940 | CC BY | 2021-01-04 16:27:55 | no | BMC Endocr Disord. 2005 Jun 17; 5:6 | utf-8 | BMC Endocr Disord | 2,005 | 10.1186/1472-6823-5-6 | oa_comm |
==== Front
BMC Fam PractBMC Family Practice1471-2296BioMed Central London 1471-2296-6-261596084910.1186/1471-2296-6-26Research ArticleThe effect of illustrations on patient comprehension of medication instruction labels Hwang Stephen W [email protected] Carolyn QN [email protected] Nadia [email protected] Centre for Research on Inner City Health, St. Michael's Hospital, Toronto, Ontario, Canada2 University of Toronto, Ontario, Canada2005 16 6 2005 6 26 26 19 10 2004 16 6 2005 Copyright © 2005 Hwang et al; licensee BioMed Central Ltd.2005Hwang 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
Labels with special instructions regarding how a prescription medication should be taken or its possible side effects are often applied to pill bottles. The goal of this study was to determine whether the addition of illustrations to these labels affects patient comprehension.
Methods
Study participants (N = 130) were enrolled by approaching patients at three family practice clinics in Toronto, Canada. Participants were asked to interpret two sets of medication instruction labels, the first with text only and the second with the same text accompanied by illustrations. Two investigators coded participants' responses as incorrect, partially correct, or completely correct. Health literacy levels of participants were measured using a validated instrument, the REALM test.
Results
All participants gave a completely correct interpretation for three out of five instruction labels, regardless of whether illustrations were present or not. For the two most complex labels, only 34–55% of interpretations of the text-only version were completely correct. The addition of illustrations was associated with improved performance in 5–7% of subjects and worsened performance in 7–9% of subjects.
Conclusion
The commonly-used illustrations on the medication labels used in this study were of little or no use in improving patients' comprehension of the accompanying written instructions.
==== Body
Background
Health literacy is the ability to read, comprehend, and act on health-related materials such as consent forms, prescription drug labels, and medical instructions [1]. Approximately 44 million Americans are functionally illiterate and another 50 million have marginal literacy skills [2]. In a study conducted at two urban public hospitals, about one-third of patients had inadequate or marginal health literacy, and 42% were unable to understand written directions for taking medication on an empty stomach [3].
Pharmaceutical labels must provide critical instructional information to people with different experiences and education levels. Patients are faced with the responsibility of converting declarative information into procedural application, resulting in a substantial risk of errors [4]. Pictorial instructions may be provided in an effort to make the procedural information more readily accessible and its comprehension less dependent on the individual's background or prior knowledge [5-7]. Relatively little research has been conducted on the cognitive processes involved in the interpretation of written and pictorial instructions on medication labels [4,7].
Special instructions regarding how a prescription medication should be taken or its possible side effects are often applied to medication bottles in the form of auxiliary labels. These labels typically include small illustrations that are intended to enhance comprehension. We conducted this study to determine whether the addition of illustrations to prescription medication instruction labels affects patients' comprehension of the accompanying written information.
Methods
Setting and subjects
This study was conducted at three family practice clinics affiliated with an urban academic teaching hospital in Toronto, Ontario. These clinics provide primary care to a large patient population living in the central area of the city. From January to September 2001, consecutive patients presenting to the clinic during regular office hours on selected weekdays were approached and asked to participate in a study of comprehension of prescription labels. Days on which patients were enrolled were selected on the basis of availability of a member of the research team. A total of 130 participants were enrolled. Patients were excluded if they were too ill to participate or were unable to communicate in English. Patients who participated in the study gave written informed consent and received a $5 payment. The St. Michael's Hospital Research Ethics Board approved this study.
Data collection
One of the investigators (CQNT) conducted face-to-face interviews with participants that obtained information on demographic characteristics, native language, and education. Participants were then presented with five instruction labels regarding how certain prescription medications should be taken or their potential side effects (Figure 1). Pharmacies usually affix these labels to the bottles or packaging of certain medications when they are dispensed to the patient. We selected five labels from among those in common use by pharmacies in Toronto. Labels were deliberately chosen to provide a wide range in terms of our assessment of the complexity of both the written information and the accompanying illustration. The labels were presented as black-and-white images that were enlarged from the original label size of 4 × 1 cm to a final size of 8 × 2 cm to enhance readability. The five labels were presented on a single sheet of letter-sized paper.
Figure 1 Prescription medication instruction labels.
Participants were first shown five labels with text only (Figure 1, left column), then a separate sheet of paper with the labels with identical text and the addition of visual illustrations (Figure 1, right column). In each case, participants were asked, "If this label were on your pill bottle, how would you take this medication?" Participants were allowed unlimited time to reply. The interviewer wrote down participants' verbatim responses on a survey form; interviews were not audiotaped.
Following presentation of all labels, participants were given the Rapid Estimate of Adult Literacy in Medicine (REALM) test. The REALM test is a previously validated instrument that uses an objective scoring system to assign a grade-range estimate of literacy (grade 0 to 6, grade 7 to 8, or grade 9 and above) [8]. This test is a simple and widely used research instrument that is highly correlated with other measures of health literacy such as the Test of Functional Health Literacy in Adults (TOFHLA) [9].
Data coding
Two researchers independently coded participants' interpretations of medication labels as incorrect, partially correct, or completely correct. Disagreements in coding were resolved by consensus after further discussion. Labels A, C, and D were deemed to each convey a single main informational component, and interpretations of these labels were coded as either incorrect or completely correct. Labels B and E were deemed to each convey three main informational components. For example, for label B the informational components were: (1) possibility of drowsiness, (2) alcohol may worsen effect, and (3) caution if driving or handling machinery. If none of these components could be identified in the participant's verbatim response, their answer was coded as incorrect. If one or two of these components was found, the response was coded as partially correct. If all three components were found, the response was coded as completely correct. This method of coding, although not previously validated, offered useful detail regarding the completeness of the participants' comprehension. At the time of coding, investigators were blinded to participants' literacy level. The effect of the addition of illustrations on participants' performance was classified as improved, worse, or unchanged, based on the categorization of their first and second responses as incorrect, partially correct, or completely correct. When improvement or worsening was noted, responses were examined to characterize the nature of the change.
Statistical analyses
The sign test was used to assess whether there was significant improvement or worsening in the interpretation of the label with the addition of the illustration. Chi-square analyses were used to assess whether the effect of the addition of illustrations on participants' performance on label interpretation (improved, worse, or no change) was significantly associated (p < 0.05) with sex, age (under 25 years, 25 to 39 years, 40 to 64 years, or 65 years and over), native language (English or other language), or grade-range estimate of health literacy (assessed using the REALM). All significance tests were two-sided. Statistical analyses were performed using SPSS 10.0 (SPSS, Inc., Chicago, IL).
The power of this study was dependent on both the number of individuals whose performance on label interpretation changed with the addition of illustrations, and the anticipated magnitude of this effect. Assuming that about 25% of subjects or 30 individuals would demonstrate a change in performance, this study would have had 80% power to detect a ratio of 3:1 or greater in terms of the proportion of individuals with improved performance vs. worse performance.
Results
Characteristics of participants are shown in Table 1. All subjects across all literacy levels correctly interpreted labels with instructions to take medication with water (label A), with food (label C), or not in conjunction with alcohol (label D), regardless of whether they were accompanied by visual illustrations. Participants' interpretations of label B and label E are shown in Table 2. A large number of responses were only partially correct. In the case of label B, participants often failed to note the intensifying effect of alcohol or the need to avoid operating a car or machinery. In the case of label E, participants often failed to note the recommended hours of separation between meals and medication use.
Table 1 Characteristics of Study Participants (N = 130)
Characteristic Number (%)
Sex
Male 57 (44)
Female 73 (56)
Age Group
Under 25 years 25 (19)
25 to 39 years 40 (31)
40 to 64 years 51 (39)
65 years and over 14 (11)
Native Language
English 92 (71)
Other language 38 (29)
Highest Educational Attainment
Less than high school 5 (4)
Some high school 8 (6)
Completed high school 35 (27)
Post-secondary 82 (63)
REALM Score
Grade 0 to 6 6 (5)
Grade 7 to 8 29 (22)
Grade 9 and above 95 (73)
Table 2 Label interpretations without and with illustrations.
Without Illustration With Illustration P-value*
Number (%) Number (%)
Interpretation of Label B
Incorrect 23 (18) 29 (22)
Partially Correct 63 (49) 57 (44)
Completely Correct 44 (34) 44 (34)
Change in Interpretation of Label B
Improved ... 6 (5)
No Change ... 113 (87) 0.33
Worse ... 11 (9)
Interpretation of Label E
Incorrect 13 (10) 14 (11)
Partially Correct 46 (35) 44 (34)
Completely Correct 71 (55) 72 (55)
Change in Interpretation of Label E
Improved ... 9 (7)
No Change ... 112 (86) 1.00
Worse ... 9 (7)
* Using the sign test to assess whether there was significant improvement or worsening in the interpretation of the label with the addition of the illustration.
The addition of visual illustrations in labels B and E resulted in some changes in interpretations (Table 2). With each label, however, the number of subjects whose performance worsened was approximately equal to the number of those whose performance improved. Thus, for both labels B and E, the addition of illustrations did not significantly improve or worsen the overall accuracy of label interpretation (Table 2). Most subjects whose performance worsened did so because they focused exclusively on the meaning of the illustration (for example, "drowsiness" or "empty stomach") and neglected the accompanying written information. This resulted in partially correct answers, as described above. Most subjects whose performance improved appeared to have attended to the written information on the label.
Sex, age group, being a native English speaker, and health literacy level were not significant predictors of improved or worsened comprehension when illustrations were added. It is important to note, however, that this study was not powered to detect such differences among subgroups of participants.
Discussion
We found that the illustrations selected for examination in this study provided little or no benefit in improving patients' comprehension of labels informing them how medications should be taken or their potential side effects. Participants in this study had perfect comprehension of three relatively simple labels when they were presented without visual illustrations, leaving no opportunity for the addition of illustrations to have any positive effect. In the case of two more complex labels that a number of patients had difficulty fully understanding, the addition of illustrations had positive and negative effects on comprehension in approximately equal numbers of patients.
Our finding of mixed effects of illustrations may be explained by a number of factors. The illustrations on the labels we presented may have been ambiguous and failed to clearly convey a specific message, or at worst they may have been misleading. For example, the illustration showing the outline of a stomach could not be understood without some knowledge of human anatomy. Even if it were understood, such an illustration conveyed only one component of a more complex message. Furthermore, the illustrations may have captured the patient's attention, distracting him or her from the full message contained in the accompanying writing. This phenomenon might occur regardless of a patient's level of health literacy.
This study has several limitations. We only examined patients in urban family practice units associated with an academic teaching hospital. Because the vast majority of our study sample had an adequate or high literacy level, our data do not support any conclusions regarding the effect of medication label illustrations in patients with low literacy. In addition, our relatively small sample size and the large proportion of subjects whose performance was unchanged with the addition of illustrations raises the possibility of type 2 error, that is, a failure to detect true differences due to inadequate power. We used a selected set of medication labels to provide a range of complexity in terms of both text and illustrations; because our choice of labels was non-random, our results may not accurately represent all of the illustrations in use or their effect on comprehension. The size and readability of the images we used may have affected our results. The use of enlarged versions of the labels may have enhanced their readability, and participants did not undergo formal screening for visual acuity. Some degree of sequencing bias may have occurred, since the labels were presented in the same order for all subjects. Finally, we did not obtain qualitative data through detailed interviews that might have allowed us to clarify how participants perceived the meaning of the illustrations and how the participants related the illustrations to the text.
Conclusion
Pictorial illustrations can improve comprehension of medication labels if the illustrations and text are well-matched to each other and are appropriate to the educational and cultural background of the user [4,7]. However, two older studies found that the use of symbols did not improve understanding of prescription instructions compared to written directions alone [5] and that comprehension of prescription labels was unaffected by the use of auxiliary labels (different from those used in this study) [10]. The inconsistent results of previous research in this area are not entirely surprising, given significant variation in the quality of illustrations and their ability to act as a complement to written instructions to improve comprehension.
Thus, while illustrations can be effective in improving patients' understanding of how their medications should be taken or their potential side effects, physicians should not assume that this is always the case. At least some of the images currently in use on prescription medication labels do not appear to be helpful, and the effectiveness of such illustrations should be further evaluated. Given their potential to confuse or distract patients, an alternative would be to replace illustrations with a common symbol, such as a large exclamation mark, to reinforce the importance of the accompanying text. Additional research is needed to better understand the cognitive processes that affect comprehension when written and pictorial pharmaceutical instructions are combined, and to develop more effective communication strategies to improve patients' understanding of the proper use of their medications [11].
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
SWH participated in the design of the study, data analysis, and drafting of the manuscript. CQNT participated in the design of the study, data collection, data analysis, and drafting of the manuscript. NK participated in the design of the study, data collection, data analysis, and revision of the manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
Dr. Hwang is the recipient of a New Investigator Award from the Canadian Institutes of Health Research. The Centre for Research on Inner City Health is supported in part by a grant from the Ontario Ministry of Health and Long-Term Care. The results and conclusions are those of the authors, and no official endorsement by the organizations named above is intended or should be inferred.
==== Refs
Ad Hoc Committee on Health Literacy for the Council on Scientific Affairs, American Medical Association Health Literacy: Report of the Council on Scientific Affairs JAMA 1999 281 552 557 10022112 10.1001/jama.281.6.552
Kirsch I Jungeblut A Jenkins L Kolstad A Adult Literacy in America: A First Look at the Results of the National Adult Literacy Survey 1993 Washington, DC: National Center for Educational Statistics, US Dept of Education
Williams MV Parker R Baker DW Parikh NS Pitkin K Coates WC Nurss JR Inadequate functional health literacy among patients at two public hospitals JAMA 1995 274 1677 1682 7474271 10.1001/jama.274.21.1677
Patel VL Branch T Arocha JF Errors in interpreting quantities as procedures: The case of pharmaceutical labels Int J Medical Informatics 2002 65 193 211 10.1016/S1386-5056(02)00045-X
Hardie NR Gagnon JP Eckel FM Feasibility of symbolic directions on prescription labels Drug Intell Clin Pharm 1979 13 588 595 10244547
Rymes-Barley C A secret inability to comply: the price of illiteracy Can Pharm J 1989 122 86 94 10292089
Patel VL Eisemon TO Arocha JF Comprehending instructions for using pharmaceutical products in rural Kenya Instructional Sci 1990 19 71 84 10.1007/BF00377986
Davis TC Michielutte R Askov EN Williams MV Weiss BD Practical assessment of adult literacy in health care Health Educ Behav 1998 25 613 624 9768381
Andrus MR Roth MT Health literacy: a review Pharmacotherapy 2002 22 282 302 11898888 10.1592/phco.22.5.282.33191
McKnight PT Schneider PJ Brier KL Effect of Label Format on Information Recall in Patients Receiving Prescription Medications Contemp Pharm Pract 1981 4 150 4 10252152
Moisan J Gaudet M Grégoire JP Bouchard R Non-Compliance with Drug Treatment and Reading Difficulties with Regard to Prescription Labelling among Seniors Gerontology 2002 48 44 51 11844930 10.1159/000048924
| 15960849 | PMC1177941 | CC BY | 2021-01-04 16:29:13 | no | BMC Fam Pract. 2005 Jun 16; 6:26 | utf-8 | BMC Fam Pract | 2,005 | 10.1186/1471-2296-6-26 | oa_comm |
==== Front
BMC GastroenterolBMC Gastroenterology1471-230XBioMed Central London 1471-230X-5-181593875610.1186/1471-230X-5-18Research ArticleA conscious mouse model of gastric ileus using clinically relevant endpoints Firpo Matthew A [email protected] Michael D [email protected] Aniko [email protected] Justin D [email protected] Jeffrey D [email protected] Yuanlin [email protected] Robert E [email protected] Sean J [email protected] Department of Surgery, University of Utah School of Medicine, 30 N 1900 E, Salt Lake City, UT, 84132 USA2 Biostatistics Resource, Huntsman Cancer Institute, University of Utah, Salt Lake City, UT, USA2005 6 6 2005 5 18 18 8 9 2004 6 6 2005 Copyright © 2005 Firpo 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
Gastric ileus is an unsolved clinical problem and current treatment is limited to supportive measures. Models of ileus using anesthetized animals, muscle strips or isolated smooth muscle cells do not adequately reproduce the clinical situation. Thus, previous studies using these techniques have not led to a clear understanding of the pathophysiology of ileus. The feasibility of using food intake and fecal output as simple, clinically relevant endpoints for monitoring ileus in a conscious mouse model was evaluated by assessing the severity and time course of various insults known to cause ileus.
Methods
Delayed food intake and fecal output associated with ileus was monitored after intraperitoneal injection of endotoxin, laparotomy with bowel manipulation, thermal injury or cerulein induced acute pancreatitis. The correlation of decreased food intake after endotoxin injection with gastric ileus was validated by measuring gastric emptying. The effect of endotoxin on general activity level and feeding behavior was also determined. Small bowel transit was measured using a phenol red marker.
Results
Each insult resulted in a transient and comparable decrease in food intake and fecal output consistent with the clinical picture of ileus. The endpoints were highly sensitive to small changes in low doses of endotoxin, the extent of bowel manipulation, and cerulein dose. The delay in food intake directly correlated with delayed gastric emptying. Changes in general activity and feeding behavior were insufficient to explain decreased food intake. Intestinal transit remained unchanged at the times measured.
Conclusion
Food intake and fecal output are sensitive markers of gastric dysfunction in four experimental models of ileus. In the mouse, delayed gastric emptying appears to be the major cause of the anorexic effect associated with ileus. Gastric dysfunction is more important than small bowel dysfunction in this model. Recovery of stomach function appears to be simultaneous to colonic recovery.
==== Body
Background
Ileus is a common post-surgical occurrence characterized by transient impairment of gastrointestinal function. In addition to abdominal surgery, sepsis, trauma, pancreatitis, anesthetic agents, and opioid analgesics are also associated with ileus. The mechanisms of ileus involve neural inhibitory signals and humoral factors including paracrine agents and gut hormones [1-3]. No single event or factor has been clearly implicated as being responsible; it is more likely that multiple mediators act at various times throughout the course of the condition. These mediators communicate throughout the gastrointestinal tract. For example, surgical manipulation of the distal bowel can induce gastric dysfunction [4-8]. Once established, ileus is thought to resolve at different rates in humans with functional inhibition lasting a few hours within the small intestine, 1–2 days within the stomach and 2–3 days within the colon [9,10].
Many recent studies of ileus have focused attention on small bowel smooth muscle dysfunction in isolated muscle strips [11-14]. The weakness of the muscle strip method, however, is the inability to examine complex interactions including brain-gut interactions or organ specific neural reflexes. Similarly, in vitro preparations require tissue harvest to assay the condition and do not allow a systems approach to the problem of ileus. The availability of complete genomic sequence information in the mouse has made the application of a systems approach feasible using global gene expression analysis. However, phenotypic characterizations of ileus in animal models that provide system wide information are lacking. Development of such a model in the mouse would be useful.
In the clinical setting, resolution of ileus is determined by the resumption of normal eating behavior and the passage of flatus or stool. Despite the fact that food intake and fecal output are simple measurements of gastrointestinal function they have not been rigorously analyzed in animal models of ileus. Here we describe a conscious mouse model that uses food intake and fecal output to monitor the time course of ileus, allowing measurement of both the magnitude and duration of the condition in the same animal. This integrative approach to defining the phenotype of ileus may facilitate investigations of potential clinically relevant interventions or preventive strategies. In addition, this conscious mouse model offers the opportunity to study molecular events associated with ileus.
Methods
Animals
Male C3H mice, 8–12 weeks old, were used throughout the study. Mice were maintained on a 12-hour light/dark cycle and given free access to standard rodent chow and water. For the continuous food intake monitoring study, mice were given 20 mg dustless food pellets (Bio-Serv, Frenchtown, NJ). All procedures were approved and monitored by the University of Utah Institutional Animal Care and Use Committee.
Food intake and fecal output measurement
Mice were separated into individual cages with free access to water and a tared amount of food. Food, fecal and mouse mass were measured every 12 hours at the beginning of each scotophase and photophase (7 AM and 7 PM). The amount of food consumed over the 12-hour period was calculated as the difference between the mass of food at the end of the period and the amount of food at the beginning of the period. Fecal output was determined from the mass of fecal pellets collected and measured at the end of the 12-hour period. The potential effect of fecal pellet dehydration over the 12 hour period was addressed in pilot studies in which ileus was induced using either 0.1 mg/kg LPS or laparotomy followed by 4 minute bowel manipulation as well as the corresponding controls. Fecal pellets were collected every 12 hours and fecal output was determined in three ways: 1) mass determination immediately upon collection (as reported in this study), 2) mass determination after a further 24 hour dehydration period, and 3) by counting fecal pellets. The decrease in fecal output relative to controls after ileus induction was similar and not significantly different among the three methods of measurement (data not shown). We concluded that the potential effect of dehydration was inconsequential. After a three-day acclimation period, the mice were randomized into experimental groups and a sham group. Data collection started 24 hours before the induction of ileus and continued after induction until a normal (baseline) circadian pattern of feeding and fecal output returned. For continuous food intake monitoring, mice were placed into individual metabolic chambers with free access to water. Single 20 mg dustless food pellets were delivered on demand using a Coulbourn Instruments Habitest system (Allentown, PA) equipped with a pellet feeder and food trough. The presence of a food pellet within the trough blocked light emitted from a light emitting diode reaching a photodetector. When the mouse removed the pellet, a new pellet was delivered. This exchange was automatically monitored and recorded by the Graphic State Notation 2 software program (Coulbourn Instruments, Allentown, PA). Mice were acclimated to the cage for 4 days prior to data collection.
Induction of ileus
After acclimation and baseline data collection periods, ileus was induced 30 minutes before scotophase unless otherwise noted. To mimic sepsis, lipopolysaccharide (E. coli O111:B4, List Biological Laboratories, Campbell, CA) was administered by intraperitoneal injection. The dose-dependent relationship of endotoxin on food intake and fecal output was investigated by determining the time courses for each endpoint after administration of 0.005, 0.01, 0.02, 0.04, 0.1, 0.2, and 0.4 mg/kg mouse mass of endotoxin. Control mice received a similar volume of saline carrier. Post-surgical ileus was induced in mice anesthetized with isoflurane either by laparotomy followed by manipulation of the cecum for 1 minute or laparotomy, evisceration onto a saline moistened sponge and manipulation of the small bowel, cecum and colon for a total manipulation time of 4 minutes. Manipulation was carried out using cotton tipped swabs. After closure of the incision with sutures, the mouse was removed from anesthesia and allowed to recover under a warming lamp for 30 minutes before data collection resumed. To assess the effects of thermal injury, a 20% total body surface area (TBSA) scald burn was induced on mice from which truncal hair was removed. The scald burn was induced by immersion of the exposed dorsal skin in 70°C water for 7 seconds. Immediately following burn injury, the mice were administered 1 ml of intraperitoneal (IP) lactated Ringer's solution. Mice were given 0.5 ml of LR IP every 12 hours for 72 hours post-burn injury. Control animals had truncal hair removed and received a similar time course of anesthesia and resuscitation with LR. Acute pancreatitis was induced in mice after an eighteen hour fasting period using 3 or 7 hourly IP injections of cerulein (50 μg/kg/dose). Sham treated control mice were given similar volumes of carrier (0.1% BSA in PBS). The timing of the injections was coordinated such that the last injection occurred at the beginning of scotophase.
Gastric emptying and intestinal transit
After a 2-hour fast to empty stomach content, mice were given 200 μl of a 1.5% (w/v) methylcellulose, 0.5% (w/v) phenol red solution in normal saline by gavage. Thirty minutes after gavage, the mice were sacrificed by cervical dislocation, a laparotomy performed and the stomach isolated by clamping the duodenum near the pylorus and the esophagus at the cardia. The entire procedure from sacrifice to clamping was performed in less than one minute. The GIT was removed, separating the stomach from the intestine. The small bowel was dissected from the cecum/colon and divided into four equal length segments by sequential bisection. The amount of phenol red in the stomach and intestinal segments was determined spectrophotometrically after homogenization as described [15]. Gastric emptying was evaluated as the percentage of dye remaining in the stomach relative to the total amount of dye recovered in a standardization group of mice that were sacrificed immediately after gavage. Intestinal transit was determined by measuring the partitioning of dye within the small bowel segments and colon numbered 1–5, proximal to distal. The geometric center of dye transit was calculated for each animal as (∑(% dye per segment X segment number)/100) as described [16].
Mean arterial blood pressure
Mice were anesthetized with isoflurane and placed on a 37°C heating pad. A polyethylene cannula (PE 10), connected to a pressure transducer, was inserted approximately 5 mm into the femoral artery. Mean arterial pressure and respiratory rate (counted over 1 minute) was recorded every five minutes. After 10 minutes of monitoring to ensure stability of pressure and respiratory rate, LPS (25 μg/ml, in doses of 0.1 and 0.4 mg/kg) or carrier was injected IP. Blood pressure and respiratory rate was recorded every five minutes for a total of 1 hour.
Activity monitoring
Digital video images were recorded using a personal computer based system consisting of a WebCam (PC-Cam 300, Creative Labs, Milpitas, CA) and WebCam Control Center version 5.6 software . The camera was placed above four standard mouse cages with wire tops. In lieu of litter, a single shredded paper towel was placed in each cage for bedding. A darkroom light equipped with a single 15-watt bulb and a Kodak GBX-2 Safelight filter (Eastman Kodak, Rochester, NY) provided illumination. Individual mice were placed in each cage with free access to water and a tared amount of food. In order to maintain an unobstructed view, water was provided in a glass bottle and food was limited to 4 standard rodent chow pellets (approximately 5 g each). Mice were acclimated to the cages for 5 days prior to data collection. Mice were administered saline, 0.1 mg/kg LPS or 2 mg/kg LPS in a blinded manner 30 minutes before scotophase. Images were recorded at 1 image per second for 10 seconds every 10 minutes for 36 hours. Investigators blinded to the treatment scored the recorded images for mouse activity and mouse induced movement of the food pellets. Activity was scored if gross movement of the mouse body was evident within any of the 10-second images. Likewise, food pellet movement was scored if the position of any of the food pellets were different when the 10 images were compared. New food pellets were placed and food mass recorded at 12-hour intervals corresponding to scotophase and photophase.
Statistical analyses
Statistical comparisons were made using factorial ANOVA or repeated measures ANOVA (for time course analyses) and Fisher's protected least significant difference (PLSD) post-hoc tests. Comparisons were considered statistically significant at the P < 0.05 level. Values are expressed as mean ± SEM unless data from individual mice are shown.
In order to determine the median-effect dose of LPS on food intake and fecal output, the 12-hour (first night) data was fitted to a dose-response curve. The model had an inverse proportionality form with offsets:
Food intake = A + (B-A)*Dm/ (Dose + Dm) + error
where Dose is the dose of the drug, A is the minimum food intake at arbitrary large doses (horizontal asymptote), B is the baseline food intake (when Dose = 0) and Dm is the median-effect dose, that is the dose at which the food intake is halfway between the baseline B and minimum A. The error term reflects the between-animal variability of food intake, it is assumed to have normal distribution with mean 0 and variance proportional to the response: error~ N(0,s2 (Food Intake)). The adjustment of the variability depending on the response was necessary as the between-animal variability decreased as the food intake decreased. Since each data point corresponds to a different mouse, the observations are independent. The above model was extended to all time points by modeling the median-effect dose Dm as a function of time:
Dm(T) = Dm(1) cT, where T is time in "nights"
that is for each night the dose required to produce a median effect is increased c-fold. We also added two random effect terms: between-mouse variability of the median-effect dose Dm(1) (on log-scale) and of the baseline food intake B. These mouse-specific terms allow us to accommodate the within-mouse dependence of the observations.
Results
Endotoxin transiently decreased food intake and fecal output in a dose dependent manner
Untreated (baseline) and carrier treated mice demonstrated a circadian pattern of food intake. Mice took a large meal during early scotophase and a smaller meal near the end of scotophase or the beginning of photophase. Although food intake was lower during the light period, the mice apparently anticipated the pending dark phase as food intake increased near the end of photophase. Administration of endotoxin at time 0 resulted in an immediate and conspicuous decrease in food intake (Figure 1). Feeding behavior began to recover during the ensuing photophase and scotophase with a return to a normal pattern by the third scotophase after injection.
Given the diurnal feeding behavior of the mice with the majority of food consumed during the dark phase, subsequent experiments were carried out by analyzing food intake during each 12-hour photoperiod. This procedure allowed for simultaneous recovery of fecal pellets. LPS administration resulted in a statistically significant reduction in nocturnal food intake during the first (17% of control), second (60% of control), and third (83% of control) scotophases (Figure 2A). Food intake recovered to near normal levels by the third scotophase with complete recovery of the circadian pattern by the fourth scotophase. These data were consistent with the initial findings using continuous food monitoring. Although the amount of food consumed during the first photophase was not significantly different between control mice and LPS treated mice, there was a significant increase in food intake during the second photophase indicating that, during recovery from the insult, the wave pattern of circadian food intake remained dampened. The pattern of fecal output was similar to the pattern of food intake with significantly inhibited fecal output in the first (26% of control), second (51% of control), and third (86% of control) scotophases (Figure 2B). Unlike the food intake data, fecal output was significantly increased relative to controls during the fourth day after injection (116% of control).
The dose-dependent relationship of endotoxin on food intake and fecal output was investigated by determining the time courses for each endpoint after administration of various doses of endotoxin. Each dose resulted in a statistically significant decrease in food intake and fecal output during the first scotophase, thus, the threshold dose of endotoxin was less than 0.005 mg/kg IP. The estimated parameters and their 95% confidence intervals are given in Table 1 and the plot of the observed data with the fitted dose-response curves are shown in Figure 3. Both fits result in similar conclusions: the median-effect dose Dm was about 0.01 mg/kg and there appeared to be a non-zero minimum food intake/fecal output even at high doses.
The coefficients of the fitted overall model which examined the data over the first four nights are given in the Table 2. Figure 4A shows the fitted model as dose-response curves for an "average" mouse, whereas Figure 4B plots the time-dependent curves for each mouse separately. The model captures most of the observed phenomena except the "overshoot" on the fourth night (84 hours). A more complicated form would be needed to capture that effect. The resulting c-multiplier indicates that, as the mice recovered, the median-effect dose increased 11-fold each night.
Decreased food intake and fecal output effects were common to other insults
Other clinically relevant insults that result in ileus were examined using the conscious mouse model. A transient reduction in food intake and fecal output was identified after laparotomy/bowel manipulation (Figure 5), thermal injury (Figure 6), and cerulein induced acute pancreatitis (Figure 7). All three insults caused a significant reduction in food intake (52%, 21% and 54% of control respectively) and fecal output (50%, 25% and 67% of control respectively) within the first 12 hours. The effects of cerulein induced acute pancreatitis induction on nocturnal food intake and fecal output did not reach their nadir until the second scotophase (37% of control for food intake, 36% of control for fecal output) after cerulein injection indicating that, compared to the other insults, more time was required to develop the full effect (Figure 7).
The specificity of cerulein and laparotomy/bowel manipulation for causing the transient decrease in food intake and fecal output was assessed by altering the severity of the insults. As with endotoxin treatment, there was a direct correlation between severity and magnitude of the effects (Figure 8). Acute pancreatitis was induced in mice using a series of either three or seven hourly injections of cerulein. Figure 8A shows the effect of each series on nocturnal food intake over the ensuing three nights. Three cerulein injections resulted in decreased food intake for the first two nocturnal meals (P ≤ 0.013 vs. corresponding controls) and returned to normal levels by the third night. A more pronounced response was seen after seven cerulein injections, with decreased food intake evident for all three nights (P < 0.0001 vs. corresponding control). Statistical comparison of the two insults revealed a significant difference (P = 0.0018) at each time point indicating that the magnitude of the response was directly related to the severity of the insult. A similar result was seen when laparotomy/bowel manipulation was used to induce ileus (Figure 8B). Both laparotomy followed by either 1 minute cecum manipulation or 4 minute manipulation of the small bowel, cecum, and colon resulted in a reduction for each of the subsequent three night time meals (P = 0.04 vs. corresponding control). Differences between the two insults was not as pronounced as the effect of the cerulein doses, but did result in a significant difference for the first night (P = 0.0024).
Decreased food intake correlated with delayed gastric emptying
To validate the use of these simple measures as markers of ileus, we examined gastric emptying and intestinal transit in this model. Gastric emptying of a methyl cellulose/phenol red dye meal was measured in mice that had been administered 0.1 mg/kg IP LPS one hour prior to gavage (Figure 9A). A period of rapid emptying was seen in both control mice and mice administered endotoxin over the first 15 minutes after gavage. The amount of dye emptied from stomachs of control mice continued to increase between 15 and 45 minutes after gavage whereas emptying from mice administered endotoxin was static over the same period. Overall gastric emptying one hour after LPS administration was dramatically reduced relative to controls (P < 0.0001 vs. corresponding control for 15, 30 and 45 minute time points). The 30-minute rate of gastric emptying was also significantly reduced relative to controls 12 hours after LPS administration (P = 0.0053), but recovered to control levels by 36 hours after LPS administration (Figure 9B). The recovery of gastric emptying apparently preceded the recovery of food intake as normal levels of food intake were not evident until the third or fourth night after LPS administration (see Figures 1 and 2A). To graphically evaluate the correlation of the two recovery rates, nocturnal food intake was plotted versus gastric emptying measured 12 hours prior to the corresponding nocturnal meal (Figure 9C). In essence, we asked if nocturnal food intake data could be "predicted" by the antecedent gastric emptying measurement. Both data sets were plotted as percent of control measurements. The resulting slope of one indicates a positive correlation between food intake and gastric emptying.
Effect of LPS administration on blood pressure
The decrease in gastric emptying in response to LPS administration was unlikely to be a non-specific hemodynamic effect as mean arterial blood pressure remained unchanged relative to controls for 60 minutes after administration of 0.1 mg/kg LPS (Figure 10) yet a significant reduction in gastric emptying was evident within 1 hour after insult (Figure 9A). Administration of 0.4 mg/kg LPS caused a significant decrease in mean arterial blood pressure at several time points as determined by repeated measures ANOVA. This result suggests that at higher LPS doses, changes in hemodynamics or blood flow may explain, at least in part, functional deficiencies in the gastrointestinal tract. We recognize that measurement of systemic blood pressure may not mirror splanchnic blood pressure or perfusion. In the mouse, however, these latter measurements are technically difficult, highly invasive, and subject to their own set of artifacts related to the procedure.
Effect of LPS administration on intestinal transit
In contrast to the effect on gastric and colonic function, intestinal transit was not significantly affected by IP injection of 0.1 mg/kg LPS when measured 12, 36 or 60 hours after treatment (Table 3).
Changes in activity were insufficient to explain decreased food intake
Mice administered the low doses of LPS described above had relatively normal appearance and activity whereas mice administered higher doses (1–2 mg/kg) showed classic signs of sickness behavior including cachexia, diarrhea, lethargy and decreased grooming. To assess the possibility that the anorexic effect of LPS could be due to changes in activity, including the inability to reach the food pellets, we measured activity during the first scotophase after LPS treatment. Mice administered 0.1 mg/kg or 2 mg/kg LPS showed approximately a 50% and 85% reduction in both general activity and mouse-induced movement of food pellets respectively (Figure 11). Within each dose, activity scores and food pellet movement scores were not significantly different from each other. These data suggest a relationship between feeding behavior and activity level. However, relative to controls, the actual amount of food taken during the first scotophase was reduced to a greater extent than activity (92% food intake decrement for mice administered 0.1 mg/kg LPS and greater than 97% for mice administered 2 mg/kg LPS) indicating that the quantity of food taken during the nocturnal meal was not due to reduced activity.
Discussion
We have evaluated the feasibility of using food intake and fecal output as markers for ileus in a conscious mouse model by comparing the effect of various insults resulting in ileus on these endpoints. Low doses of LPS, laparotomy/bowel manipulation, 20% TBSA scald burn, and acute pancreatitis comparably produced a transient anorexic effect and delayed defecation consistent with a clinical picture of ileus. The data indicate that both magnitude and duration of the effects were dose dependent establishing a causal link between the initiating insults and the changes in food intake and fecal output. Each insult was self-limiting and mice appeared to have normal grooming habits, vocalizations, and interactions with cage mates within minutes of treatment or recovery from anesthetic. LPS administration in mice has been shown to increase watery secretions in the bowel and induce diarrhea [17,18], either of which would potentially confound our fecal output measurements. However, in our study, we saw no evidence of increased watery secretion and diarrhea was not evident at any LPS dose in which fecal output was measured. A recent study showed that a minimum dose of 10 mg/kg was required to produce a significant increase in watery secretion, as measured by intestinal content mass [17], whereas the highest dose of LPS used in our study was more than an order of magnitude lower (0.4 mg/kg). It might be argued that the inhibition of food intake was caused by the pain and discomfort associated with the various insults. For endotoxin treatment, the delay in food intake directly correlated with delayed gastric emptying suggesting gastric dysfunction was responsible, at least in part, for decreased food intake. Peripheral administration of low doses of LPS is a well established model for sepsis and is associated with the inhibition of food intake in rodents [19]. In rats, the anorexic effect of LPS appears to be mediated centrally through the activity of serotonin [20,21], melanocortins [22], cytokines [23-26], and prostaglandins [25,27]. In mice, the anorexigenic effect of peripheral LPS is attenuated by vagotomy [19] suggesting a brain-gut interaction. There is also evidence that LPS-induced inhibition of food intake in mice is correlated with the inhibition of the orexigenic and motility promoting peptide ghrelin [28] which is synthesized mainly in gastric tissue and acts centrally. Previous studies have shown that low doses of LPS inhibited gastric emptying [29-32]. Here we show that the time course of the LPS induced anorexic effect directly correlated with delayed gastric emptying (Figure 9C). The parallel recoveries of gastric emptying and food intake suggest that the anorexic effect of LPS was in large part due to gastric dysfunction. Furthermore, although administration of LPS resulted in a significant reduction in movement during the first scotophase, reduced activity alone was not sufficient to explain the anorexic effect (Figure 11) as the decrease in the amount of food taken was significantly higher than the decrease in activity. Delayed gastric emptying has also been demonstrated in rodent models of post-operative ileus [1,4], thermal injury [33] and cerulein induced pancreatitis [34].
The common effects suggest, at some level, a common mechanism is shared by the multiple insults. One likely candidate is the inflammatory response since endotoxemia [13,14,35], intestinal manipulation [12,36-38], and ischemia/reperfusion injury [39] have all been shown to promote cellular inflammation within the small bowel and colon in rats. Recently it has been shown in mice that post-operative ileus was associated with inflammatory cell infiltration within the manipulated small intestine, but not the untouched stomach or colon [1]. In the same study, gastric emptying, measured by scintigraphic imaging, was delayed for 24 hours after insult, recovering within 48 hours. This delay was prevented by inhibition of leukocyte recruitment indicating that the inflammatory cells were responsible for delayed gastric emptying, apparently via inhibitory neural signals since hexamethonium and guanethidine normalized gastric emptying. Inflammatory cell infiltration within the pancreas during cerulein induced acute pancreatitis has been well established, however, it is unclear if similar infiltration within the small bowel muscularis also occurs. Such an event would be confirmatory for the model mentioned above. Our data suggests that inflammatory cell infiltration would be delayed relative to other insults since the effect of higher doses of cerulein did not reach its maximum until the second scotophase. Establishment of inhibitory neural signals is also likely to be sequelae common to the various insults. The role of corticotropin releasing factor (CRF) in both appetite regulation and gastrocolonic motor function has been well established [40]. CRF receptors are widely expressed in the brain, notably within brain centers that control appetite, the gastrointestinal tract, and on vagal afferent neurons. The known effects of central and peripheral CRF receptors and ligands makes them intriguing investigative targets for systemic control of the different organ systems involved with ileus.
Ileus is most commonly associated with a transient decrease in gastrointestinal motility. In humans, motility resolves differentially within the gastrointestinal tract with the small bowel recovering most rapidly, followed by the stomach and then colon [9,10] although it is possible that complete recovery of the entire GI tract is not required for clinical resolution. Food intake and fecal output are the most common clinical markers for resolution of ileus and are more generalized indicators of GI function than measurement of motility, MMC or contractility. As shown in this study, delayed food intake is likely to be modified by behavioral effects and thus provides a more integrated description of gastrointestinal function. In our mouse model, fecal output directly tracked food intake for all insults and doses examined in this study. The fitted models for the 12-hour dose-response curves yielded identical median effective LPS doses for food intake and fecal output providing mathematical verification of this correlation. Since data was collected every 12 hours, it is possible that decreased fecal output lagged food intake by a few hours. Therefore it is difficult to distinguish if this tracking was a function of the initiating insult on colonic activity or normal loss of colonic function due to the fasting caused by the anorexic effects of the insults. The fact that the two endpoints tracked demonstrates the functional coordination of the two organs during ileus and illustrates the necessity for endpoints that examine system wide function in order to develop a complete understanding of the mechanisms of ileus. In the small intestine, 30 minute transit of dye remained unchanged between mice administered endotoxin and control mice 12, 36 and 60 hours after injection. This result indicated that if small bowel transit was compromised by LPS, the effect resolved within 12 hours.
The use of food intake and fecal output as endpoints for ileus in mice revealed dissimilarities from more traditional measures of ileus. In the conscious mouse model described here, food intake and fecal output directly correlated with the extent of bowel manipulation. This result is consistent with at least one other study which showed that, in rats, small bowel smooth muscle impairment and inflammation was directly proportional to the extent of bowel manipulation [37]. However, several groups have concluded that inhibition of bowel motility was independent of the degree of manipulation or the duration of surgery [9,41-43]. It is probable that the discrepancy in outcomes is due to the different endpoints measured. A large body of research into the mechanisms of ileus focuses on changes in intestinal smooth muscle contractility. In rats, changes in small bowel smooth muscle contractility after endotoxin administration were seen only above the threshold dose of 5 mg/kg [13]. In our study, food intake and fecal output were responsive to small changes in LPS dose with the lowest dose of LPS tested (0.005 mg/kg) resulting in a significant reduction in both endpoints indicating that they are highly sensitive measures of gastrointestinal function. These data, along with the lack of change in intestinal transit, suggest that low doses of endotoxin produce limited intestinal ileus.
The ability to conveniently monitor the full time course of ileus in the same animal was lacking in most previous models of ileus (although see [44]). The use of food intake/fecal output as simple, clinically relevant endpoints in un-anesthetized mice should facilitate examination of potential treatment strategies. Improvements in either magnitude or duration of ileus are desirable and treatment strategies would be considered effective if they improved either or both. The fitted model described in this study provides a method for quantitative assessment of both magnitude and duration. Changes in magnitude due to a treatment are most easily assessed by measuring the 12 hour food intake/fecal output in LPS challenged mice. Improvement in treated mice relative to untreated LPS challenged mice would quantitatively assess effectiveness. The specific challenge is not limited to LPS. The 12 hour food intake/fecal output measurements for other insults can be expressed in terms of LPS dose by substituting the amount into the formula for the fitted curves. Thus, in our study, a 20% TBSA burn was comparable to a 0.04 mg/kg LPS dose. Although a more involved assay, duration can be quantitatively monitored by the c-multiplier which will be altered as duration changes. In this scenario, an intervention is used to treat mice challenged with several doses of LPS. An increased c-multiplier, then, would indicate a decrease in the duration of ileus and, thus, indicate an effective intervention. Finally, the ability to monitor ileus without the necessity of harvesting tissue will facilitate gene expression studies which require the same tissue for RNA isolation.
Conclusion
We have demonstrated in conscious mice that a transient decrease in food intake and fecal output are common to multiple insults that result in ileus. Examination of ileus using these endpoints revealed that restoration of gastric emptying preceded recovery of food intake and that gastric and colonic function recovered nearly simultaneously. The easily monitored, clinically relevant endpoints provide a convenient means for examining ileus in a mouse model and should prove useful for analysis of potential intervention strategies and mechanistic investigations of ileus.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
MF coordinated the study, assisted with data collection, carried out data analysis, participated in the study design and drafted the manuscript. MR carried out the pancreatitis experiments and assisted with the transit experiments. AS participated in study design and performed statistical modeling. JG carried out the activity monitoring experiments and assisted with food intake and fecal output data collection. JJ assisted with the activity monitoring experiments and food intake and fecal output data collection. YS performed the laparotomy/bowel manipulations and burn injury animal models, blood pressure monitoring, tissue harvest, and assisted with food intake and fecal output data collection. RG participated in study design, coordination, and data review. SM conceived of the study, participated in the study design, supervised data analysis and edited the manuscript. All authors read and approved the final manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
The authors acknowledge the enthusiastic contribution of John Bade, MD (deceased) to this study.
Figures and Tables
Figure 1 Circadian pattern of food intake and effect of endotoxin administration. Food intake was monitored continuously. Endotoxin (0.1 mg/kg, solid line) or carrier (dashed line) was administered IP at time 0. Black bars along the abscissa indicate scotophase. Data points represent mean measurements of six mice in 3 hour time bins.
Figure 2 Induction of ileus after endotoxin administration. Measurements were made every 12 hours such that each data point represents the food intake (A) or fecal output (B) corresponding to the preceding scotophase or photophase. Bars along the abscissa indicate scotophase. Endotoxin (0.1 mg/kg, solid line) or carrier (dashed line) was administered by intraperitoneal injection at time 0 (arrow). Data points are presented as mean ± SEM (n = 9 mice per group). * P = 0.02 vs. control at corresponding time point by repeated measures ANOVA and Fisher's PLSD.
Figure 3 Effect of endotoxin dose on food intake and fecal output. Twelve hour food intake and fecal output was monitored over a 4 day period after injection of carrier or various doses of endotoxin. The fitted dose-response curves for food intake (A) and fecal output (B) for the first 12 hours after endotoxin injection are shown (n = 6 mice per dose).
Figure 4 Effect of endotoxin dose over time.A. The fitted dose-response curves for food intake are shown for the four nights after injection illustrating the increase in median effective dose each night. B. Time-dependent food intake curves are shown for the individual mice at each dose. (n= 6 mice per dose)
Figure 5 Induction of post-operative ileus. Twelve hour measurements are plotted as described in legend for Figure 2. Mice were anesthetized, a laparotomy performed, and the cecum manipulated with cotton tipped swabs for 1 minute at time 0 (solid line). Control mice received anesthesia for a similar time (dashed line). N = 10 mice per group. * P = 0.008 vs. control at corresponding time point.
Figure 6 Induction of ileus after thermal injury. Twelve hour measurements were are plotted as described in legend for Figure 2. A 20% total body surface area scald burn was induced at time 0 (solid line). Control mice received anesthesia only (dashed line). N = 3 mice per group. * P = 0.05 vs. control at corresponding time point.
Figure 7 Effect of cerulein induced acute pancreatitis on food intake and fecal output. Twelve hour measurements were plotted as described in legend for Figure 2. Seven hourly injections of cerulein (50 μg/kg per dose) were administered after a 12-hour fasting period ending at time 0 (solid line) and compared to control mice receiving carrier alone (dotted line). N = 6 mice per group. * P = 0.009 vs. control at corresponding time point.
Figure 8 Correlation of insult severity and the magnitude of ileus.Nocturnal food intake for the three nights post-insult was plotted for A. after 3 (blue) or 7 (red) hourly injections of cerulein and B. after laparotomy and either a one minute cecum manipulation (blue) or a 4 minute manipulation of the GIT (red). Corresponding control data is shown for each plot (dashed line). Statistical comparisons are discussed in the text. Values are mean ± SEM.
Figure 9 Effect of endotoxin on gastric emptying. A. Gastric emptying curve generated 1 hour after endotoxin (0.1 mg/kg, solid line) or carrier (dashed line) injection. A methylcellulose/phenol red dye meal was administered to different groups of mice by intragastric gavage. The amount of dye remaining in the stomach after a 1, 15, 30 or 45 minute transit period was measured. The percent of dye emptied was calculated relative to the total amount of dye gavaged. Each data point represents measurements from 3 – 7 mice. B. The recovery of gastric function was monitored by measuring the 30 minute emptying at various times after administration of endotoxin (0.1 mg/kg, solid line) or carrier (dashed line). N = 3–12 mice per group. C. Correlation plot comparing the recovery of food intake and gastric emptying. Data points represent percentages of the corresponding control. The data point numbers represent the first, second and third scotophase for food intake measurements (ordinate) or the gastric emptying measurement made preceding each scotophase (abscissa). Values are mean ± SEM.
Figure 10 Effect of endotoxin on mean arterial blood pressure. Mean arterial pressure was recorded every five minutes. Carrier, 0.1 mg/kg, or 0.4 mg/kg was administered IP at time 0 (n = 6 mice per group). * P < 0.05 vs. control at corresponding time.
Figure 11 Activity and feeding behavior during the first scotophase after LPS injection. Digital video images were scored for activity and mouse induced movement of food pellets during the first scotophase after injection of saline (open bars), 0.1 mg/kg LPS (shaded bars) or 2 mg/kg LPS (black bars) and compared to the total amount of food consumed during the same 12-hour period. Movement data are presented as the percent of the total frames that movement was scored (left axis). Food intake in grams is presented on the right axis. Values are mean ± SEM (n = 5 for saline and 2 mg/kg LPS groups, n = 6 for 0.1 mg/kg group). * P < 0.03 vs. corresponding saline control. ‡ P < 0.05 vs. corresponding 0.1 mg/kg LPS dose.
Table 1 Parameters of the 12-hour endotoxin dose-response curve. Results of the fitted model examining the effect of various doses of endotoxin on food intake and fecal output during the first scotophase are shown.
Parameter Estimate 95% Confidence Interval
Food Intake
A – minimum, grams 0.16 (0.08, 0.24)
B – baseline, grams 3.27 (2.90, 3.64)
Dm – median effect dose, mg/kg 0.011 (0.007, 0.015)
s – standard deviation, grams 0.33 (0.28, 0.41)
Fecal Output
A – minimum, grams 0.11 (0.08, 0.14)
B – baseline, grams 1.00 (0.88, 1.12)
Dm – median effect dose, mg/kg 0.010 (0.006, 0.015)
s – standard deviation, grams 0.13 (0.11, 0.16)
Table 2 Parameters of the endotoxin dose-response curves for food intake over time. Results of the fitted model examining the effect of various doses of endotoxin on nocturnal food intake extended to include the four nights after LPS injection.
Parameter Estimate 95% Confidence Interval
A – minimum, grams 0.13 (0.05, 0.21)
B – baseline, grams 3.50 (3.61, 3.72)
Dm(1) – median effect dose for the first night, mg/kg 0.010 (0.007, 0.012)
c – multiplier of the median effect dose for each additional night 11.00 (8.62, 14.04)
sB – standard deviation of B, g 0.27 (0.18, 0.39)
sD – standard deviation of logDm(1), g 0.47 (0.33, 0.66)
s – residual standard deviation, grams 0.23 (0.20, 0.26)
Table 3 Intestinal transit of methyl cellulose/phenol red dye after endotoxin injection. The 30-minute transit of dye was measured in each of five intestinal segments and the mean geometric center calculated. (n = 6 for each group, P = 0.3 by two-way ANOVA)
Treatment Time Post-LPS Injection
12 Hours 36 Hours 60 Hours
Control 2.28 ± 0.17 2.58 ± 0.16 2.36 ± 0.19
0.1 mg/kg LPS 2.59 ± 0.10 2.71 ± 0.09 2.50 ± 0.12
==== Refs
de Jonge WJ van den Wijngaard RM The FO ter Beek ML Bennink RJ Tytgat GN Buijs RM Reitsma PH van Deventer SJ Boeckxstaens GE Postoperative ileus is maintained by intestinal immune infiltrates that activate inhibitory neural pathways in mice Gastroenterology 2003 125 1137 1147 14517797 10.1016/S0016-5085(03)01197-1
Behm B Stollman N Postoperative ileus: etiologies and interventions Clin Gastroenterol Hepatol 2003 1 71 80 15017498 10.1053/cgh.2003.50012
Luckey A Livingston E Tache Y Mechanisms and treatment of postoperative ileus Arch Surg 2003 138 206 214 12578422 10.1001/archsurg.138.2.206
Barquist E Bonaz B Martinez V Rivier J Zinner MJ Tache Y Neuronal pathways involved in abdominal surgery-induced gastric ileus in rats Am J Physiol 1996 270 R888 94 8967419
Bonaz B Plourde V Tache Y Abdominal surgery induces Fos immunoreactivity in the rat brain J Comp Neurol 1994 349 212 222 7860779 10.1002/cne.903490205
Bonaz B Tache Y Corticotropin-releasing factor and systemic capsaicin-sensitive afferents are involved in abdominal surgery-induced Fos expression in the paraventricular nucleus of the hypothalamus Brain Res 1997 748 12 20 9067440 10.1016/S0006-8993(96)01281-4
Martinez V Rivier J Wang L Tache Y Central injection of a new corticotropin-releasing factor (CRF) antagonist, astressin, blocks CRF- and stress-related alterations of gastric and colonic motor function J Pharmacol Exp Ther 1997 280 754 760 9023288
Martinez V Rivier J Tach Y Peripheral injection of a new corticotropin-releasing factor (CRF) antagonist, astressin, blocks peripheral CRF- and abdominal surgery-induced delayed gastric emptying in rats J Pharmacol Exp Ther 1999 290 629 634 10411571
Condon RE Frantzides CT Cowles VE Mahoney JL Schulte WJ Sarna SK Resolution of postoperative ileus in humans Ann Surg 1986 203 574 581 3707236
Livingston EH Passaro EPJ Postoperative ileus Dig Dis Sci 1990 35 121 132 2403907 10.1007/BF01537233
Hierholzer C Kalff JC Billiar TR Bauer AJ Tweardy DJ Harbrecht BG Induced nitric oxide promotes intestinal inflammation following hemorrhagic shock Am J Physiol Gastrointest Liver Physiol 2004 286 G225 33 14715517 10.1152/ajpgi.00447.2002
Kalff JC Buchholz BM Eskandari MK Hierholzer C Schraut WH Simmons RL Bauer AJ Biphasic response to gut manipulation and temporal correlation of cellular infiltrates and muscle dysfunction in rat Surgery 1999 126 498 509 10486602 10.1067/msy.2099.99575
Eskandari MK Kalff JC Billiar TR Lee KK Bauer AJ Lipopolysaccharide activates the muscularis macrophage network and suppresses circular smooth muscle activity Am J Physiol 1997 273 G727 34 9316478
Lodato RF Khan AR Zembowicz MJ Weisbrodt NW Pressley TA Li YF Lodato JA Zembowicz A Moody FG Roles of IL-1 and TNF in the decreased ileal muscle contractility induced by lipopolysaccharide Am J Physiol 1999 276 G1356 62 10362638
Luckey A Wang L Jamieson PM Basa NR Million M Czimmer J Vale W Tache Y Corticotropin-releasing factor receptor 1-deficient mice do not develop postoperative gastric ileus Gastroenterology 2003 125 654 659 12949710 10.1016/S0016-5085(03)01069-2
Miller MS Galligan JJ Burks TF Accurate measurement of intestinal transit in the rat J Pharmacol Methods 1981 6 211 217 7329070 10.1016/0160-5402(81)90110-8
Liang YC Liu HJ Chen SH Chen CC Chou LS Tsai LH Effect of lipopolysaccharide on diarrhea and gastrointestinal transit in mice: roles of nitric oxide and prostaglandin E2 World J Gastroenterol 2005 11 357 361 15637744
Theodorou V Chovet M Eutamene H Fargeau H Dassaud M Toulouse M Bihoreau C Roman FJ Bueno L Antidiarrhoeal properties of a novel sigma ligand (JO 2871) on toxigenic diarrhoea in mice: mechanisms of action Gut 2002 51 522 528 12235074 10.1136/gut.51.4.522
Kent S Bret-Dibat JL Kelley KW Dantzer R Mechanisms of sickness-induced decreases in food-motivated behavior Neurosci Biobehav Rev 1996 20 171 175 8622824 10.1016/0149-7634(95)00037-F
von Meyenburg C Langhans W Hrupka BJ Evidence that the anorexia induced by lipopolysaccharide is mediated by the 5-HT2C receptor Pharmacol Biochem Behav 2003 74 505 512 12479973 10.1016/S0091-3057(02)01029-8
Hrupka BJ Langhans W A role for serotonin in lipopolysaccharide-induced anorexia in rats Pharmacol Biochem Behav 2001 68 355 362 11267641 10.1016/S0091-3057(00)00463-9
Huang QH Hruby VJ Tatro JB Role of central melanocortins in endotoxin-induced anorexia Am J Physiol 1999 276 R864 71 10070149
Dantzer R Bluthe RM Gheusi G Cremona S Laye S Parnet P Kelley KW Molecular basis of sickness behavior Ann N Y Acad Sci 1998 856 132 138 9917873
Laye S Gheusi G Cremona S Combe C Kelley K Dantzer R Parnet P Endogenous brain IL-1 mediates LPS-induced anorexia and hypothalamic cytokine expression Am J Physiol Regul Integr Comp Physiol 2000 279 R93 8 10896869
Langhans W Hrupka B Interleukins and tumor necrosis factor as inhibitors of food intake Neuropeptides 1999 33 415 424 10657519 10.1054/npep.1999.0048
Konsman JP Dantzer R How the immune and nervous systems interact during disease-associated anorexia Nutrition 2001 17 664 668 11448593 10.1016/S0899-9007(01)00602-5
Lugarini F Hrupka BJ Schwartz GJ Plata-Salaman CR Langhans W A role for cyclooxygenase-2 in lipopolysaccharide-induced anorexia in rats Am J Physiol Regul Integr Comp Physiol 2002 283 R862 8 12228055
Basa NR Wang L Arteaga JR Heber D Livingston EH Tache Y Bacterial lipopolysaccharide shifts fasted plasma ghrelin to postprandial levels in rats Neurosci Lett 2003 343 25 28 12749989 10.1016/S0304-3940(03)00312-4
McCarthy DO Tumor necrosis factor alpha and interleukin-6 have differential effects on food intake and gastric emptying in fasted rats Res Nurs Health 2000 23 222 228 10871537 10.1002/1098-240X(200006)23:3<222::AID-NUR6>3.0.CO;2-3
Calatayud S Barrachina MD Garcia-Zaragoza E Quintana E Esplugues JV Endotoxin inhibits gastric emptying in rats via a capsaicin-sensitive afferent pathway Naunyn Schmiedebergs Arch Pharmacol 2001 363 276 280 11284441 10.1007/s002100000359
Collares EF Effect of bacterial lipopolysaccharide on gastric emptying of liquids in rats Braz J Med Biol Res 1997 30 207 211 9239306 10.1590/S0100-879X1997000200008
Takakura K Hasegawa K Goto Y Muramatsu I Nitric oxide produced by inducible nitric oxide synthase delays gastric emptying in lipopolysaccharide-treated rats Anesthesiology 1997 87 652 657 9316972 10.1097/00000542-199709000-00027
Alican I Coskun T Yegen C Aktan AO Yalin R Yegen BC The effect of thermal injury on gastric emptying in rats Burns 1995 21 171 174 7794496 10.1016/0305-4179(95)80003-7
Robert A Lum JT Lancaster C Olafsson AS Kolbasa KP Nezamis JE Prevention by prostaglandins of caerulein-induced pancreatitis in rats Lab Invest 1989 60 677 691 2469859
Turler A Schwarz NT Turler E Kalff JC Bauer AJ MCP-1 causes leukocyte recruitment and subsequently endotoxemic ileus in rat Am J Physiol Gastrointest Liver Physiol 2002 282 G145 55 11751168
Kalff JC Carlos TM Schraut WH Billiar TR Simmons RL Bauer AJ Surgically induced leukocytic infiltrates within the rat intestinal muscularis mediate postoperative ileus Gastroenterology 1999 117 378 387 10419919
Kalff JC Schraut WH Simmons RL Bauer AJ Surgical manipulation of the gut elicits an intestinal muscularis inflammatory response resulting in postsurgical ileus Ann Surg 1998 228 652 663 9833803 10.1097/00000658-199811000-00004
Turler A Moore BA Pezzone MA Overhaus M Kalff JC Bauer AJ Colonic postoperative inflammatory ileus in the rat Ann Surg 2002 236 56 66 12131086 10.1097/00000658-200207000-00010
Hierholzer C Kalff JC Audolfsson G Billiar TR Tweardy DJ Bauer AJ Molecular and functional contractile sequelae of rat intestinal ischemia/reperfusion injury Transplantation 1999 68 1244 1254 10573059 10.1097/00007890-199911150-00006
Zorrilla EP Tache Y Koob GF Nibbling at CRF receptor control of feeding and gastrocolonic motility Trends Pharmacol Sci 2003 24 421 427 12915052 10.1016/S0165-6147(03)00177-9
Kewenter J Kock NG Studies on intestinal motility during the first few days after partial gastrectomy and after cholecystectomy. II. Response to Prostigmine with and without jejunal feeding Acta Chir Scand 1963 125 248 252 14032217
Graber JN Schulte WJ Condon RE Cowles VE Relationship of duration of postoperative ileus to extent and site of operative dissection Surgery 1982 92 87 92 7089872
Noer T Roentgenological transit time through the small intestine in the immediate postoperative period Acta Chir Scand 1968 134 577 580 5731298
Bennink RJ De Jonge WJ Symonds EL van den Wijngaard RM Spijkerboer AL Benninga MA Boeckxstaens GE Validation of gastric-emptying scintigraphy of solids and liquids in mice using dedicated animal pinhole scintigraphy J Nucl Med 2003 44 1099 1104 12843228
| 15938756 | PMC1177942 | CC BY | 2021-01-04 16:03:27 | no | BMC Gastroenterol. 2005 Jun 6; 5:18 | utf-8 | BMC Gastroenterol | 2,005 | 10.1186/1471-230X-5-18 | oa_comm |
==== Front
BMC GenomicsBMC Genomics1471-2164BioMed Central London 1471-2164-6-831593509410.1186/1471-2164-6-83Research Article(TG/CA)n repeats in human gene families: abundance and selective patterns of distribution according to function and gene length Sharma Vineet K [email protected] Samir K [email protected] Srinivasan [email protected] G.N. Ramachandran Knowledge Centre for Genome Informatics, Institute of Genomics and Integrative Biology, Mall Road, Delhi 110 007, India2005 3 6 2005 6 83 83 22 10 2004 3 6 2005 Copyright © 2005 Sharma et al; licensee BioMed Central Ltd.2005Sharma 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
Creation of human gene families was facilitated significantly by gene duplication and diversification. The (TG/CA)n repeats exhibit length variability, display genome-wide distribution, and are abundant in the human genome. Accumulation of evidences for their multiple functional roles including regulation of transcription and stimulation of recombination and splicing elect them as functional elements. Here, we report analysis of the distribution of (TG/CA)n repeats in human gene families.
Results
The 1,317 human gene families were classified into six functional classes. Distribution of (TG/CA)n repeats were analyzed both from a global perspective and from a stratified perspective based on their biological properties. The number of genes with repeats decreased with increasing repeat length and several genes (53%) had repeats of multiple types in various combinations. Repeats were positively associated with the class of Signaling and communication whereas, they were negatively associated with the classes of Immune and related functions and of Information. The proportion of genes with (TG/CA)n repeats in each class was proportional to the corresponding average gene length. The repeat distribution pattern in large gene families generally mirrored the global distribution pattern but differed particularly for Collagen gene family, which was rich in repeats. The position and flanking sequences of the repeats of Collagen genes showed high conservation in the Chimpanzee genome. However the majority of these repeats displayed length polymorphism.
Conclusion
Positive association of repeats with genes of Signaling and communication points to their role in modulation of transcription. Negative association of repeats in genes of Information relates to the smaller gene length, higher expression and fundamental role in cellular physiology. In genes of Immune and related functions negative association of repeats perhaps relates to the smaller gene length and the directional nature of the recombinogenic processes to generate immune diversity. Thus, multiple factors including gene length, function and directionality of recombinogenic processes steered the observed distribution of (TG/CA)n repeats. Furthermore, the distribution of repeat patterns is consistent with the current model that long repeats tend to contract more than expand whereas, the reverse dynamics operates in short repeats.
==== Body
Background
The evolution of organisms with increasing complexity was significantly facilitated by duplication of genes and genomes followed by diversification [1,2]. Gene duplication per se produces two identical copies. Subsequently, one of the copies may either accumulate beneficial changes to give rise to a functionally diversified gene or accrue deleterious mutations to end up as a pseudogene, while the other copy retains its original function. The former mechanism leads to the creation of 'gene families' capable of carrying out diverse functions [2,3].
The classification of genes into gene families by Human Gene Nomenclature Committee (HGNC) on the basis of sequence similarity of the encoded proteins [4] and the availability of human genome sequence [5] allow us to carry out a comprehensive survey of a class of important functional element, namely the (TG/CA)n repeats. Analysis of the distribution of (TG/CA)nrepeats within genes in 'present day' gene families holds the potential to provide insights into the factors steering their abundance and selective distribution. Although the characteristic property of (TG/CA)n repeats exhibiting length polymorphism has been widely used in genetic mapping [6], a growing body of evidence accumulating over several years point to their multiple functional roles in various biological processes.
The (TG/CA)n repeats have a propensity to undergo structural transitions [7-10] and have been shown to modulate transcription in several genes including rat α-lactalbumin [9], rat prolactin [11], MMP-9 [12], IFN-γ [13], EGFR [14], HSD11B2 [15], tilipia prolactin1 [16] and human housekeeping genes [17]. Furthermore, the (TG)n tracts have been observed to act as stimulator in recombination and in mRNA splicing [18-22].
In the current study, the analysis of distribution of (TG/CA)n repeats in human gene families affords assessment of the distribution of these repeats by examining for positive association or negative association with respect to gene length and function.
Results
Characteristics of human gene families and their functional classification
Each of the 1,317 gene families included members with similar functional roles. The family sizes varied in a wide range between 2 to 223 members (Figure 1). The number of gene families was found to bear an inverse exponential relation to family size. About two-fifths of the gene families were duplex. Only three gene families had more than 100 members per family: Immunoglobulin heavy chain (162 genes), Zinc finger proteins (200 genes) and Solute carrier (223 genes).
Figure 1 Distribution pattern of human gene families with respect to family sizes. X axis: family size (number of genes in each gene family). Y axis: number of gene families corresponding to various family sizes. Note the inverse exponential relationship.
The functional classification of 1,317 gene families comprising 7,928 genes in the six functional classes unveiled that the Signaling and communication is largest with 529 families and 3,072 genes (Figure 2). The Cell cycle is the smallest with 82 families and 470 genes.
Figure 2 Global distribution of gene families, genes and proportion of genes containing (TG/CA)n repeats classified into the six functional classes. The numbers correspond to the height of the vertical bars in each group.
Of the 1,317 gene families, 131 were entirely intrachromosomal. Chromosome 1 had the largest number with 17 families followed by chromosomes 19 and 11 with 13 and 12 families respectively. The remaining chromosomes had less than 10 intrachromosomal gene families per chromosome. The functional classification of these 131 intrachromosomal gene families revealed that the highest number (45) belonged to the class of 'Immune and related functions' closely followed by the class of Signaling and communication with 40 families. The remaining classes had the following distribution of gene families: Metabolism (24), Information (15), Structure and motility (5) and Cell cycle (2). These observations indicate that the creation of intrachromosomal human gene families was driven by large number of duplications followed by divergence in selected functional classes.
Global distribution of (TG/CA)n repeats (n ≥ 6 units) in gene families
Of the 1,317 gene families, 732 families had (TG/CA)n repeats in at least one of their members and 326 families had repeats in all their members. Of the 7,928 genes in 1,317 families, 3,986 genes had intragenic (TG/CA)n repeats of length greater than or equal to 6 units. All 3,986 genes had repeats in their introns. Only 277 genes had (TG/CA)n repeats in exons indicating that these repeats are mainly present in introns.
The distribution of genes with (TG/CA)n repeats in the six functional classes is displayed in Figure 2. It is apparent that the class of Signaling and communication had the highest number of genes with (TG/CA)n repeats. Comparison of the proportion of genes with repeats in each class with the global proportion showed that the class of Signaling and communication had significantly higher than the expected proportion (p < 0.0001, Binomial test). In contrast, the classes of Immune and related functions and Information had significantly lower than the expected proportion of genes with repeats (p < 0.0001 and p < 0.0002 respectively). The proportion of genes with repeats was not significantly different from the global proportion in the Cell cycle, Metabolism and Structure and motility classes. These observations show that the (TG/CA)n repeats exhibit positive association with the genes belonging to Signaling and communication whereas, they are negatively associated with the genes belonging to Immune and related functions and Information.
It has been shown that the human genome has an isochore structure that varies in GC content [5]. This variation raises the possibility that the observed selective distribution of (TG/CA)n repeats might have arisen due to fluctuations in the local %(G+C) content of the genomic region as opposed to function. We examined this by comparing the average %(G+C) content of the genes in the six functional classes with the corresponding proportions of genes with repeats. The average %(G+C) content was observed to be in the narrow range (47–49%) in the six functional classes whereas, the proportion of genes with repeats varies widely in the range 29.6–61%. These observations indicate that the proportion of genes with repeats is significantly determined by function instead of small fluctuations in %(G+C) content.
Correlation between gene length, function and global distribution of (TG/CA)n repeats
Comparison of the proportion of genes containing (TG/CA)n repeats with the average lengths of genes in each of the six functional classes revealed a linear relationship (Figure 3, correlation coefficient R = 0.93, p < 0.007). The signaling and communication class had the longest average gene length (74.07 kb) along with the highest proportion of genes with (TG/CA)n repeats (61.23%). The class of Immune and related functions had the shortest average gene length (21.26 kb) with the lowest proportion of genes with (TG/CA)n repeats (29.65%). These observations show that the proportion of genes with (TG/CA)n repeats bears a linear relationship to the length of genes.
Figure 3 Relationship between proportion of genes with (TG/CA)n repeats in each functional class and the average gene length in the corresponding functional classes. X axis: Proportion of genes with (TG/CA)n repeats (%); Y axis: Average gene length (kb) (CC: Cell cycle; IN: Information; IR: Immune and related functions; MET: Metabolism; SC: Signaling and communication; STM: Structure and motility)
Trinity of (TG/CA)n repeats in gene families
In order to examine the characteristics of distribution of (TG/CA)n repeats with respect to multiple functional roles principally governed by their length, we analysed the repeats stratified into three categories: type I (6 ≤ n < 12), type II (12 ≤ n < 23) and type III (n ≥ 23). The results are displayed in Figure 4. The number of genes containing (TG/CA)n repeats decreases with increasing repeat length. It is also apparent that several genes (53% of the total) have multiple types of repeats in various combinations.
Figure 4 A Venn diagram of the genes with trinity of intragenic (TG/CA)n repeats (type I, II and III). Note that several genes (shaded area) have multiple types of repeats in various combinations.
Classification of the distribution of genes with (TG/CA)n repeats stratified into three categories into six functional classes is shown in Figure 5. It is evident that the proportion of genes containing repeats decreases in the order I > II > III in all classes. The proportion of genes containing (TG/CA)n repeats of Signaling and communication were significantly higher than the expected proportion in all three categories of repeats (p < 0.0001, type I, II and III). On the other hand, the proportion of genes with (TG/CA)n repeats of Immune and related functions and Information were significantly lower than expected proportion in all three categories: Immune and related functions (p < 0.0001, type I, II and III), Information (p < 0.0001, type I and II, p < 0.004, type III). The proportion of genes with type III repeats was marginally lower than the expected proportion in Metabolism class (p < 0.01) and marginally higher than the expected proportion in Structure and motility class (p < 0.02). The proportion of genes with repeats in the three categories was not significantly different from the expected value in the class of Cell cycle. These observations show that repeats of all types are positively associated with the genes of Signaling and communication whereas they are negatively associated with the genes of Immune and related functions and Information.
Figure 5 Distribution of proportion of genes with three types of (TG/CA)n repeats in the six functional classes.
The distribution of average number of (TG/CA)n repeats per gene in the three categories in the six functional classes is displayed in Figure 6. Comparison of the average number of repeats per gene in the three categories with the global distribution pattern revealed that in most cases the observed number was significantly lower than the expected value, except for the genes belonging to Signaling and communication and Structure and motility, which had significantly higher average number of repeats per gene than the expected value (p < 0.0004 in all three categories, both classes). The average number of type III repeats per gene in the class of Cell cycle was not significantly different from the expected value. These observations show that the repeat densities were higher in the genes belonging to Signaling and communication and Structure and motility classes whereas, the genes belonging to other classes had lower repeat densities.
Figure 6 Distribution of the densities of three types of (TG/CA)n repeats in the genes of six functional classes.
Large gene families
As a special case of this study, we examined the distribution of (TG/CA)n repeats in the top 2% large families (27). The family sizes of this category varied widely from 32 to 223 members. Functional classification of these large families revealed the following distribution: Immune and related functions (9), Signaling and communication (8), Information (6), Metabolism (2), Structure and motility (1) and Cell cycle (1).
The proportion of genes with (TG/CA)n repeats in large families is displayed in Table 1. Comparison with the global distribution showed that the proportion of genes with repeats was significantly higher than expected value in the Signaling and communication and Structure and motility classes (p < 0.0001, Binomial test). There was no significant difference between the observed and the expected proportion of genes with repeats in the class of Metabolism. In the remaining classes, the proportion of genes with repeats was significantly lower than the expected value (p < 0.0001, Binomial test). As observed with all gene families, a linear relationship was observed between gene lengths and proportion of genes with (TG/CA)n repeats (correlation coefficient R = 0.79, p < 0.0001).
Table 1 Distribution of (TG/CA)n repeats in large gene families
Functional class and gene families Chromosomal Distributiona Average Gene Length (kb)b Genes in the family Proportion of Genes with (TG/CA)n repeats Number of Genes with (TG/CA)n repeats in three categories Average number of repeats per gene in three categories
Type I Type II Type III Type I Type II Type III
Cell cycle classc(1)
Histone proteins family Dispersed 3.55 76 7.9 4 4 0 1.7 0.7 0
Immune and related functions class (9)
Interleukins Dispersed 14.67 43 32.6 9 6 2 1.3 0.9 0.1
Serine (or cysteine) proteinase inhibitor family Dispersed 17.87 32 50 12 9 2 1.4 0.8 0.2
Tumor necrosis factor (ligand) superfamily Dispersed 23.49 38 63.2 21 14 1 1.9 0.9 0.1
CD antigens Dispersed 26.96 54 46.3 20 11 3 1.9 1.3 0.3
Immunoglobulin heavy chains Intrachromosomal 0.38 162 0.6 1 0 0 1 0 0
Immunoglobulin kappa chains Intrachromosomal 0.55 73 5.5 1 3 0 0.3 3 0
Immunoglobulin lambda chains Intrachromosomal 0.35 88 0 0 0 0 0 0 0
Interleukin receptors family Dispersed 29.77 32 59.4 17 10 0 3.4 0.9 0
T cell receptor beta chains 84 Intrachromosomal, 9 Dispersed 0.42 94 9.6 5 3 1 0.8 0.3 0.1
Information class (6)
Homeo box Dispersed 5.48 40 25 6 6 1 1.2 0.9 0.2
Eukaryotic translation initiation factor Dispersed 36.54 33 45.5 13 10 3 2.1 0.9 0.2
Zinc finger protein family Dispersed 30.33 200 42.5 63 44 6 2.2 1.1 0.1
DEAD/H (Asp-Glu-Ala-Asp/His) box polypeptides Dispersed 43.44 32 62.5 17 8 2 1.5 0.6 0.1
Ribosomal protein genes Dispersed 4.94 96 6.3 6 0 0 1 0 0
Mitochondrial ribosomal protein genes Dispersed 16.92 74 23 14 12 1 1.8 0.9 0.1
Metabolism class (2)
Cytochrome P450 superfamily Dispersed 31.34 45 46.7 15 11 2 1.9 1 0.1
Proteasome subunit genes Dispersed 25.64 40 32.5 11 8 0 1.5 0.8 0
Signaling and Communication class (8)
G protein-coupled receptor family Dispersed 24.71 98 33.7 26 19 6 2.3 1.5 0.2
Tripartite motif-containing family Dispersed 29.29 40 60 19 12 2 1.6 0.8 0.1
Solute carrier family Dispersed 59.19 223 62.8 134 87 22 2.9 1.5 0.2
RAS oncogene family Dispersed 39.92 60 65 38 17 4 1.8 0.7 0.1
ATP-binding cassette transporters gene family Dispersed 73.85 44 68.2 29 24 4 3.6 2.5 0.3
Guanine nucleotide binding protein (G protein) polypeptide genes Dispersed 58.83 32 59.4 18 12 5 3.3 1.9 0.4
Potassium voltage-gated channel genes Dispersed 104.95 38 57.9 17 16 6 8.6 4.4 0.5
Protein phosphatase subunit genes Dispersed 65.62 59 57.6 27 22 7 2.9 1.8 0.3
Structure and motility class (1)
Collagen family Dispersed 132.83 37 86.5 29 23 10 5.7 2.3 0.4
a: Chromosomal distribution of the members of gene families. 'Dispersed' indicates that members are distributed on different chromosomes.
b: average gene length (in kb) for each gene family.
c:Numbers in parentheses show the number of large sized gene families in each functional class.
The large Collagen gene family belonging to the class of Structure and motility had the highest proportion of genes containing repeats (86.5%). In order to analyze this further, we examined the sequence conservation of the region flanking 200 bases upstream and downstream in addition to the repeats by comparing the human sequence with the available genome sequence of Chimpanzee (Pan troglodytes), a nearest ancestor to human [46]. We observed, that of the 268 sequence segments including repeats from human Collagen genes, 244 were conserved with greater than 92% identity in the chimpanzee. Of these 244 repeats in human Collagen genes, 73 repeats were identical in length, 142 repeats displayed length polymorphism in the chimpanzee, 27 repeats had point mutations and in 2 cases there were no repeats in the corresponding segments in the chimpanzee. These observations show that both human and chimpanzee Collagen genes have high repeat content, high conservation of position and flanking sequences of the repeats. However, majority of repeats exhibited length polymorphisms, which is consistent with their characteristic property [6].
Discussion
The inverse relationship between the number of gene families and their corresponding sizes, resulting in a large number of small sized gene families, suggests that several duplicated copies may have been lost during the first round of genome duplication itself, considering the hypothesis of two rounds of genome duplication in vertebrate evolution [1,23,24]. The non-uniform distribution of the number of gene families across the six functional classes suggests that widespread gene duplication across gene families spanning a wide range of functions may have been less productive in attaining higher levels of complexity. An alternate course involving large amount of duplications followed by divergence producing a wide range of functions in selected classes might have been favorable. The support for the latter hypothesis emerges from the fact that large sized gene families, inherently low in number, mainly belong to Immune and related functions (required to tackle a wide range of infections), Signaling and communication (required to respond to diverse environmental stimuli) and Information class (required to implement complex molecular processes through supramolecular assemblies or organelles). A few members of large sized gene families of Metabolism class function in bioenergetics and xenobiotic metabolism and of Cell cycle class function in packaging of nuclear DNA. Similarly the large Collagen gene family of Structure and motility class offers a useful repertoire for the formation of multiple tissues [25].
It is apparent that short repeats are abundant in human genes and long repeats are rare. Our findings are consistent with the observations by Whittaker et al. (2003), who showed that longer repeats are more likely to contract than expand [47]. Accordingly, contraction of long repeats in time would result in accumulation of higher number of short repeats.
Of the six functional classes, the Signaling and communication class was the richest in repeats including the proportion of genes with repeats and repeat densities. Many of the genes belonging to this class function at the interface between the body and its environment that appears to be a distinct feature of eukaryotes [28] to confer species-specific advantages [24,41]. The positive association of (TG/CA)n repeats associated with genes of this class strongly argues for a positive temporal regulatory role that could provide for variations in gene expression to complement the enormous diversity characteristic of this class. Compared to this class, the anciently evolved gene families of Information and Cell cycle are poor in repeats. Considering the fact that these genes are highly conserved [28-30] and are involved in implementing the molecular processes acting at the core of cellular physiology, these observations suggest that repeats are negatively associated with these genes to avoid unpredictable consequences for the normal functioning of the cell.
Another argument in favor of these inferences stems from the linear relationship between the average gene length of gene families belonging to the respective functional classes and the proportion of genes with repeats in these classes. The average length of genes belonging to Information class was short and this factor aids in obtaining high levels of expression of these genes [29]. This requirement, however, generates a space constraint to accommodate additional elements. This situation contrasts to that of genes of Signaling and communication class with higher average gene length offering more space for accommodating other regulatory elements. The analysis of Collagen gene family belonging to large sized families presents itself as an interesting case. Most of the members of this family have (TG/CA)n repeats. Sequence comparisons of repeat containing regions of human Collagen genes with the nearest ancestor to humans, the Chimpanzee, revealed that although there is high conservation in terms of content and position of repeats, majority of repeats were polymorphic, which is consistent with their characteristic property [6]. Among repeats that displayed polymorphism between human and Chimpanzee, nearly equal proportions of human repeats were either contracted or expanded in Chimpanzee. These results are also consistent with the Whittaker's model [47].
Strikingly, the genes of Immune and related functions class are poor in (TG/CA)n repeats in general and in type III repeats in particular. A characteristic trend of this class is to have large sized families with their genes arranged juxtaposed on the same chromosomal locations. This arrangement increases the possibility of these gene families to display more uniform sequence characteristics [31]. Further, these genes have the smallest average gene length indicating a compact arrangement, which is likely to act as a space constraint in the accommodation of (TG/CA)n repeats. In addition, the negative association of type III (TG/CA)n repeats in these genes may have a directional role. The immunoglobulin genes use the 7 bp and 9 bp repeats for generation of variants through VDJ recombination [33]. Accommodation of type III (TG/CA)n repeats (n ≥ 23) might introduce variations in this process and could result in loss of directional recombination essential to generate diversity in immunoglobulins and T cell receptor chains in an ordered manner.
Conclusion
The (TG/CA)n repeat distribution pattern observed in human gene families is consistent with Whittaker's model of repeat expansion and contraction. It appears that multiple factors including gene length, function and directionality of recombination processes steered the observed selective patterns of distribution of (TG/CA)n repeats in human gene families.
Methods
Sequence retrieval and mapping of (TG/CA)n repeats
Sequences of 35,114 human genes (build number 33) were retrieved from LocusLink [43] using a JavaScript program. A sum of 192 genes could not be retrieved because of either inaccessibility to the LocusLink page or absence of the link for retrieving the gene sequence. A gene in this analysis is considered as the nucleotide sequence from the start of first exon to the end of last exon. If alternate splicing was reported, the gene length considered was the start of first exon to the last known exon including all alternatively spliced products for that gene.
Perl scripts, 'SimRep' and 'RepGene' were written for the identification and mapping of perfect intragenic (TG/CA)n repeats of length n ≥ 6 units in genes [17]. Throughout this work we have used n ≥ 6 units as the minimum cut-off to identify (TG/CA)n repeats. All repeats were scored in the intragenic region (exons and introns only).
Categorization of (TG/CA)n repeats
We grouped (TG/CA)n repeats into three categories (types I, II and III), according to their length and biological properties. Type I (TG/CA)n repeats, in the range 6 ≤ n <12 units, are short repeats based on the observation that a repeat length of 8 units (n = 8) is minimum to be likely polymorphic [34,35]. Type II (TG/CA)n repeats comprise of 12 ≤ n < 23 units and is based on the observation that more than 93% of the (CA)n repeats of n ≥ 12 units are polymorphic [6]. Further, repeats of this length have also been shown to have preferential binding to nuclear factors compared to short repeats [36] and can also stimulate mRNA splicing [21,22]. Type III repeats consist of relatively long reiterations of (TG/CA)n (n ≥ 23 units) and have propensity to adopt structures such as Z DNA [8,9,37]. Other studies have shown that (TG/CA)n repeats of length greater than 22.5 units can stimulate recombination [18-20].
Clustering of genes into gene families
Functional roles of a large number of human genes are not well known. Presently, these genes are assigned hypothetical annotations. Genes labeled as 'LOC', 'DFKZP', 'FLJ', 'HSPC', 'HSU', 'HT', 'KIAA', 'ORF', 'hypothetical', 'PRO' and 'pseudogenes' without clear functional details were filtered out. A total of 22,688 genes were removed in this filtering exercise. Out of the remaining 12,426 genes, a total of 8,778 genes (25% of total) were clustered into gene families based on their gene root symbols as defined in the guidelines of Human Gene Nomenclature Committee (2002) [4]. The remaining 3,648 genes could not be clustered into gene families and are solitary.
The HGNC guidelines consider sequence and functional similarity of proteins encoded by genes while grouping them into gene families [4,38,39]. A root symbol signifies a gene family. The family members are designated by Arabic numerals placed immediately after the gene root symbol, for example GPR1, GPR2, GPR3 for genes of the G protein-coupled receptor family. A Perl script namely Clustergene was written to cluster 8,778 human genes into 1,556 gene families. The Perl script called ChromoCluster was written to report gene families located on the same chromosome. Subsequently these gene families were classified into the six functional classes as described below.
Functional Classification of gene families for comparative analysis
The gene families were classified into six functional classes namely, 'Information', 'Cell cycle', 'Metabolism', 'Signaling and communication', 'Immune and related functions' and 'Structure and motility' based on the scheme defined by Adams et al. [40]. We combined the functional classes of replication, transcription, RNA processing and translation into 'Information' class based on Andrade et al. [41].
'Cell cycle' includes cell cycle, apoptosis, chromosomal structure and DNA repair; 'Immune and related functions' includes immunology, homeostasis, carrier proteins/membrane transport and stress response; 'Information' includes protein synthesis, translation factors, ribosomal proteins, post-translational modification/targeting, protein degradation, tRNA synthesis/metabolism, RNA synthesis, transcription factors, RNA polymerase, RNA processing, RNA degradation, DNA synthesis/replication and DNA repair; 'Metabolism' includes amino acids, nucleotides, sugars, lipids, cofactors, protein modification, energy and carrier proteins/membrane transport; 'Signaling and communication' includes receptors, hormone/growth factors, intracellular transducers, effectors/modulators, metabolism, cell adhesion and channels/transport proteins; 'Structure and motility' includes cytoskeletal, microtubule-associated proteins/motors and extracellular matrix.
Assignment of gene families to each of the functional classes was carried out according to their annotations in Gene Ontology [42] and LocusLink [43] databases. Out of the total 1,556 gene families, 1,317 could be classified into any of the six functional classes. The remaining 239 families could not be classified unambiguously due to limited information on gene function. Subsequent analysis, with respect to functional classification and distribution of (TG/CA)n repeats, presented here is from 1,317 gene families comprising of 7,928 genes.
Alignment of human (TG/CA)n repeats and flanking sequences with Chimpanzee genome sequence
The repeats present in human Collagen genes were aligned with Chimpanzee (Pan troglodytes) genome by using 'BLAT' software available at UCSC Genome Bioinformatics Site [49]. Nucleotide segments including the repeats and containing 200 nucleotides upstream of the start and 200 nucleotides downstream from the end of each of the (TG/CA)n repeat were extracted for human Collagen genes [48]. These segments were aligned with the Chimpanzee genome (Build 1, version 1, Nov 2003) using BLAT. Only those segments that showed more than 92% identity were noted as conserved.
Statistical methods
Significance of the differences between the proportions of genes containing repeats and repeats densities in the six functional classes compared with global distribution was tested using Binomial proportions test. The observed proportion in each class was tested against the expected proportion, which was computed assuming no preference with respect to function. Correlation coefficient (R) was computed to examine the relationship between average gene length of gene families belonging to a functional class and the proportion of genes with (TG/CA)n repeats in the corresponding functional classes. The 'Interactive Statistical Calculation Pages' website was used to perform the statistical tests.
Authors' contributions
VKS conceived of the idea, developed algorithms in Perl, carried out the analysis and wrote the manuscript. SKB gave scientific suggestions for improving the quality of the work and participated in manuscript preparation. SR is the group leader, gave scientific suggestions, helped in the statistical analysis, critical examination, presentation, writing and manuscript preparation.
Acknowledgements
VKS is a recipient of Senior Research Fellowship from CSIR. We thank Pankaj Bhatnagar for help in writing programs and the anonymous reviewers for their insightful comments. SKB and SR thank CSIR for funding support in the form of a grant (CMM0017) Task Force on "In Silico Biology for Drug target development".
==== Refs
Meyer A Schart M Gene and genome duplications in vertebrates: the one-to-four (-to-eight in fish) rule and the evolution of novel gene functions Current Opinion in Cell Biology 1999 11 699 704 10600714 10.1016/S0955-0674(99)00039-3
Wagner A Selection and gene duplication: a view from the genome Genome Biol 2002 3 reviews1012 10.1186/gb-2002-3-5-reviews1012
King RC Stansfield WD A Dictionary of Genetics 1990 Oxford University Press
Wain HM Bruford EA Lovering RC Lush MJ Wright MW Povey S Guidelines for human gene nomenclature Genomics 2002 79 464 470 11944974 10.1006/geno.2002.6748
International Human Genome Sequencing Consortium Initial sequencing and analysis of the human genome Nature 2001 409 860 921 11237011 10.1038/35057062
Dib C Faure S Fizames C Samson D Drouot N Vignal A Millasseau P Marc S Hazan J Seboun E Lathrop M Gyapay G Morissette J Weissenbach J A comprehensive genetic map of the human genome based on 5,264 microsatellites Nature 1996 380 152 4 8600387 10.1038/380152a0
Brahmachari SK Meera G Sarkar PS Balagurumoorthy P Tripathi J Raghavan S Shaligram U Pataskar S Simple repetitive sequences in the genome: structure and functional significance Electrophoresis 1995 16 1705 14 8582360 10.1002/elps.11501601283
Nordheim A Rich A The sequence (dC-dA)n X (dG-dT)n forms left-handed Z-DNA in negatively supercoiled plasmids Proc Natl Acad Sci 1983 80 1821 1825 6572943
Meera G Ramesh N Brahmachari SK Zintrons in rat α-lactalbuman gene FEBS Lett 1989 251 245 249 2753162 10.1016/0014-5793(89)81463-2
Tripathi J Brahmachari SK Distribution of simple repetitive (TG/CA)n and (CT/AG)n sequences in human and rodent genomes J Biomol Struct Dyn 1991 9 387 97 1741969
Naylor LH Clark EM d(TG)n.d(CA)n sequences upstream of the rat prolactin gene form Z-DNA and inhibit gene transcription Nucleic Acids Res 1990 18 1595 1601 2158081
Shimajiri S Arima N Tanimoto A Murata Y Hamada T Wang KY Sasaguri Y Shortened microsatellite d(CA)21 sequence down-regulates promoter activity of matrix metalloproteinase 9 gene FEBS Lett 1999 455 70 4 10428474 10.1016/S0014-5793(99)00863-7
Pravica V Asderakis A Perrey C Hajeer A Sinnott PJ Hutchinson IV In vitro production of IFN-gamma correlates with CA repeat polymorphism in the human IFN-gamma gene Eur J Immunogenet 1999 26 1 3 10068907 10.1046/j.1365-2370.1999.00122.x
Gebhardt F Zanker KS Brandt B Modulation of epidermal growth factor receptor gene transcription by a polymorphic dinucleotide repeat in intron 1 J Biol Chem 1999 274 13176 13180 10224073 10.1074/jbc.274.19.13176
Agarwal AK Giacchetti G Lavery G Nikkila H Palermo M Ricketts M McTernan C Bianchi G Manunta P Strazzullo P Mantero F White PC Stewart PM CA-Repeat polymorphism in intron 1 of HSD11B2: effects on gene expression and salt sensitivity Hypertension 2000 36 187 94 10948076
Streelman JT Kocher TD Microsatellite variation associated with prolactin expression and growth of salt-challenged tilapia Physiol Genomics 2002 9 1 4 11948285
Sharma VK B-Rao C Sharma A Brahmachari SK Ramachandran S (TG/CA)n repeats in human housekeeping genes J Biomol Struct Dyn 2003 21 303 10 12956614
Tracy RB Baumohl JK Kowalczykowski SC The preference for GT-rich DNA by the yeast Rad51 protein defines a set of universal pairing sequences Genes Dev 1997 11 3423 31 9407034
Dutreix M (GT)n repetitive tracts affect several stages of RecA-promoted recombination J Mol Biol 1997 273 105 13 9367750 10.1006/jmbi.1997.1293
Majewski J Ott J GT Repeats are associated with recombination on human chromosome 22 Genome Research 2000 10 1108 14 10958629 10.1101/gr.10.8.1108
Gabellini N A polymorphic GT repeat from the human cardiac Na + Ca2+ exchanger intron 2 activates splicing Eur J Biochem 2001 268 1076 83 11179974 10.1046/j.1432-1327.2001.01974.x
Hui J Stangl K Lane WS Bindereif A HnRNP L stimulates splicing of the eNOS gene by binding to variable-length CA repeats Nat Struct Biol 2003 10 33 7 12447348 10.1038/nsb875
Hinegardner R Evolution of cellular DNA content in teleost fishes Am Nat 1968 102 517 523 10.1086/282564
Fyodor A Kondrashov FA Rogozin IB Wolf YI Koonin EV Selection in the evolution of gene duplications Genome Biol 2002 3 RESEARCH0008 10.1186/gb-2002-3-2-research0008
Gelse K Poschl E Aigner T Collagens – structure, function, and biosynthesis Adv Drug Deliv Rev 2003 55 1531 46 14623400 10.1016/j.addr.2003.08.002
Hardison RC Conserved noncoding sequences are reliable guides to regulatory elements Trends Genet 2000 16 369 372 10973062 10.1016/S0168-9525(00)02081-3
Tagle DA Koop BF Goodman M Slightom JL Hess DL Jones RT Embryonic epsilon and gamma globin genes of a prosimian primate (Galago crassicaudatus). Nucleotide and amino acid sequences, developmental regulation and phylogenetic footprints J Mol Biol 1988 203 439 455 3199442 10.1016/0022-2836(88)90011-3
Rivera MC Jain R Moore JE Lake JA Genomic evidence for two functionally distinct gene classes Proc Natl Acad Sci USA 1998 95 6239 6244 9600949 10.1073/pnas.95.11.6239
Castillo-Davis CI Mekhedov SL Hartl DL Koonin EV Kondrashov FA Selection for short introns in highly expressed genes Nat Genet 2002 31 415 418 12134150
Dotan I Ziv E Dafni N Beckman JS McCann RO Glover CV Canaani D Functional Conservation between the Human, Nematode, and Yeast CK2 Cell Cycle Genes Biochem Biophys Res Commun 2001 288 603 609 11676486 10.1006/bbrc.2001.5804
Dover GA Evolution of genetic redundancy for advanced players Curr Opin Genet Dev 1993 3 902 10 8118216 10.1016/0959-437X(93)90012-E
Ramachandran GN Sasisekharan V Structure of collagen Nature 1961 190 1004 5 13739287
Alt FW Oltz EM Young F Gorman J Taccioli G Chen J VDJ recombination Immunol Today 1992 13 306 14 1510813 10.1016/0167-5699(92)90043-7
Fondon JW 3rdMele GM Brezinschek RI Cummings D Pande A Wren J O'Brien KM Kupfer KC Wei MH Lerman M Minna JD Garner HR Computerized polymorphic marker identification: experimental validation and a predicted human polymorphism catalog Proc Natl Acad Sci 1998 95 7514 7519 9636181 10.1073/pnas.95.13.7514
Rockman MV Wray GA Abundant raw material for Cis – regulatory evolution in humans Mol Biol Evol 2002 19 1991 2004 12411608
Epplen C Santos EJ Maueler W van Helden P Epplen JT On simple repetitive DNA sequences and complex diseases Electrophoresis 1997 18 1577 85 9378125 10.1002/elps.1150180916
Yagil G Paranemic structures of DNA and their role in DNA unwinding Crit Rev Biochem Mol Biol 1991 26 475 559 1662125
Nelson DR Koymans L Kamataki T Stegeman JJ Feyereisen R Waxman DJ Waterman MR Gotoh O Coon MJ Estabrook RW Gunsalus IC Nebert DW P450 superfamily: update on new sequences, gene mapping, accession numbers and nomenclature Pharmacogenetics 1996 6 1 42 8845856
Mackenzie PI Owens IS Burchell B Bock KW Bairoch A Belanger A Fournel-Gigleux S Green M Hum DW Iyanagi T Lancet D Louisot P Magdalou J Chowdhury JR Ritter JK Schachter H Tephly TR Tipton KF Nebert DW The UDP glycosyltransferase gene superfamily: recommended nomenclature update based on evolutionary divergence Pharmacogenetics 1997 7 255 69 9295054
Adams MD Kerlavage AR Fleischmann RD Fuldner RA Bult CJ Lee NH Kirkness EF Weinstock KG Gocayne JD White O Sutton G Blake JA Brandon RC Chiu M Clayton RA Cline RT Cotton MD Earle-Hughes J Fine LD FitzGerald LM FitzHugh WM Fritchman JL Geoghagen NSM Glodek A Gnehm CL Hanna MC Hedblom E Hinkle PS JrKelley JM Klimek KM Kelley JC Liu L Marmaros SM Merrick JM Moreno-Palanques RF McDonald LA Nguyen DT Pellegrino SM Phillips CA Ryder SE Scott JL Saudek DM Shirley R Small KV Spriggs TA Utterbach TR Weidman JF Li Y Barthlow R Bednarik DP Cao L Cepeda MA Coleman TA Collins E Dimke D Feng P Ferrie A Fischer C Hastings GA He W Hu J Huddleston KA Greene JM Gruber J Hudson P Kim A Kozak DL Kunsch C Ji H Li H Meissner PS Olsen H Raymond L Wei Y Wing J Xu C Yu G Ruben SM Dillon PJ Fannon MR Rosen CA Haseltine WA Fields C Fraser CM Venter JC Initial assessment of human gene diversity and expression patterns based upon 83 million nucleotides of cDNA sequence Nature 1995 377 3 174 7566098
Andrade MA Ouzounis C Sander C Tamames J Valencia A Functional classes in the three domains of life J Mol Evol 1999 49 551 557 10552036
Harris MA Clark J Ireland A Lomax J Ashburner M Foulger R Eilbeck K Lewis S Marshall B Mungall C Richter J Rubin GM Blake JA Bult C Dolan M Drabkin H Eppig JT Hill DP Ni L Ringwald M Balakrishnan R Cherry JM Christie KR Costanzo MC Dwight SS Engel S Fisk DG Hirschman JE Hong EL Nash RS Sethuraman A Theesfeld CL Botstein D Dolinski K Feierbach B Berardini T Mundodi S Rhee SY Apweiler R Barrell D Camon E Dimmer E Lee V Chisholm R Gaudet P Kibbe W Kishore R Schwarz EM Sternberg P Gwinn M Hannick L Wortman J Berriman M Wood V de la Cruz N Tonellato P Jaiswal P Seigfried T White R Gene Ontology Consortium The Gene Ontology (GO) database and informatics resource Nucleic Acids Res 2004 D258 61 14681407
Wheeler DL Church DM Edgar R Federhen S Helmberg W Madden TL Pontius JU Schuler GD Schriml LM Sequeira E Suzek TO Tatusova TA Wagner L Database resources of the National Center for Biotechnology Information: update Nucleic Acids Res 2004 32 D35 40 14681353 10.1093/nar/gkh073
Gu Z Rifkin SA White KP Li Wen-Hsiung Duplicate genes increase gene expression diversity within and between species Nature Genetics 2004 36 577 579 15122255 10.1038/ng1355
Thornton JW DeSalle R Gene family evolution and homology: Genomics Meets Phylogenetics Annu Rev Genomics Hum Genet 2000 1 41 73 11701624 10.1146/annurev.genom.1.1.41
Check E Geneticists study chimp-human divergence Nature 2004 428 242 15029156
Whittaker JC Harbord RM Boxall N Mackay I Dawson G Sibly RM Likelihood-based estimation of microsatellite mutation rates Genetics 2003 164 781 787 12807796
Sainudiin R Durrett RT Aquadro CF Nielsen R Microsatellite mutation models: insights from a comparison of humans and chimpanzees Genetics 2004 168 383 95 15454551 10.1534/genetics.103.022665
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 University of California Santa Cruz The UCSC Genome Browser Database Nucleic Acids Res 2003 31 51 4 12519945 10.1093/nar/gkg129
| 15935094 | PMC1177943 | CC BY | 2021-01-04 16:32:46 | no | BMC Genomics. 2005 Jun 3; 6:83 | utf-8 | BMC Genomics | 2,005 | 10.1186/1471-2164-6-83 | oa_comm |
==== Front
BMC GenomicsBMC Genomics1471-2164BioMed Central London 1471-2164-6-841593874510.1186/1471-2164-6-84Research ArticleDecoding the nucleoid organisation of Bacillus subtilis and Escherichia coli through gene expression data Carpentier Anne-Sophie [email protected]ésani Bruno [email protected] Alex [email protected]énaut Alain [email protected] Laboratoire Génome et Informatique, CNRS UMR 8116, Tour Evry2, 523 Place des Terrasses, 91034 Evry Cedex, France2 CMI, Université de Provence, 39 rue Joliot-Curie, 13453 Marseille cedex 13, France2005 6 6 2005 6 84 84 4 4 2005 6 6 2005 Copyright © 2005 Carpentier et al; licensee BioMed Central Ltd.2005Carpentier 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
Although the organisation of the bacterial chromosome is an area of active research, little is known yet on that subject. The difficulty lies in the fact that the system is dynamic and difficult to observe directly. The advent of massive hybridisation techniques opens the way to further studies of the chromosomal structure because the genes that are co-expressed, as identified by microarray experiments, probably share some spatial relationship. The use of several independent sets of gene expression data should make it possible to obtain an exhaustive view of the genes co-expression and thus a more accurate image of the structure of the chromosome.
Results
For both Bacillus subtilis and Escherichia coli the co-expression of genes varies as a function of the distance between the genes along the chromosome. The long-range correlations are surprising: the changes in the level of expression of any gene are correlated (positively or negatively) to the changes in the expression level of other genes located at well-defined long-range distances. This property is true for all the genes, regardless of their localisation on the chromosome.
We also found short-range correlations, which suggest that the location of these co-expressed genes corresponds to DNA turns on the nucleoid surface (14–16 genes).
Conclusion
The long-range correlations do not correspond to the domains so far identified in the nucleoid. We explain our results by a model of the nucleoid solenoid structure based on two types of spirals (short and long). The long spirals are uncoiled expressed DNA while the short ones correspond to coiled unexpressed DNA.
==== Body
Background
As Lovett and Segall [1] point out in their meeting report on the recently held "Keystone Symposium on Bacterial Chromosomes", we know a lot about the bacterial DNA replication, recombination, repair and other aspects of cell biology, but still rather little about the organisation of bacterial chromosome. The difficulty lies in the fact that the system varies and is difficult to observe directly. A number of different techniques are being employed to answer the problem. The following is meant to give a brief overview and has no claim to be exhaustive:
• Cytology-based approaches include the use of DNA fluorescence microscopy, optical sectioning and FISH (fluorescence in situ hybridisation). These techniques were applied in order to localise within the cell a set of chromosomal segments [2] or to see the relationship between the shapes of the nucleoid and the underlying arrangements of DNA [3].
• Cunha et al [4] approach the question from a cytometric point of view, in order to study the compaction and the internal dynamics of the nucleoid.
• An example of a classical genetic approach is the work by Valens et al [5] who have used a site-specific recombination system in order to reveal spatial proximities of distant DNA sites.
• Various genomic approaches have been adopted. Some authors, like Audit and Ouzounis [6], have taken a sequence-based point of view, in which they face the issue of gene localisation and orientation using 89 complete microbial chromosomes from eubacteria and archeabacteria. This approach leaves aside any physiology-based consideration.
• Other authors have examined the physiological constraints operating placed upon the cell in order to infer chromosomal structure. The idea is that genes which use the same type of resource (e.g. a particular tRNA pool) or which are involved in a part of metabolism that needs a particular environment (e.g. genes involved in sulphur metabolism which is highly sensitive to free radicals) should be in close proximity in the cell, even if they are far away on the chromosome [7,8].
The approaches mentioned above can be spilt in two groups: (i) large-scale analyses, aiming at deciphering the global chromosome organisation; (ii) small-scale analyses, which take a particular point of view (some genes or markers are chosen). The introduction of microarrays has added yet another way to study the chromosomal structure, allowing simultaneously the analysis on small and large scales [9]. Microarrays allow the measure of relative expression levels of the whole genome and therefore the identification of those genes that are co-expressed. Usually the co-expressions observations are used to elucidate the structure of operons and other regulatory structures, see for example [10,11].
The present work aims at understanding the nucleoid structure with the help of microarray data. As transcriptionally active DNA is located near the nucleoid surface or on DNA loops extending from the nucleoid [12], the co-expressed genes which are identified with microarrays probably share some spatial relationship.
However, microarrays give significant information only for those genes the level of expression of which varies across experiments. Consequently, the experimental conditions should be diversified in order to obtain a list of gene correlations as exhaustive as possible and thus an accurate image of the chromosomal structure. To this end, we gathered a number of currently available microarray data from the literature. The data were then pooled together, and treated as just one large data set. This "pooling of information" has already been carried out successfully from human expression data for a study of gene function [13], and from yeast or bacterial data for regulation studies [11,14].
We applied this method to two distant bacteria: Escherichia coli and Bacillus subtilis. Audit and Ouzounis [6] had the same approach, expecting that if observations made on one organism also hold true for the other, it would be reasonable to assume that the inferred chromosomal organisation is indeed a general characteristic of bacteria with double stranded, circular DNA.
Results
The aim of this work is to delineate how the co-expression intensities (correlations) of pairs of genes vary as a function of the inter-gene distance along the chromosome. The co-expression intensity for each couple of genes was evaluated with a non-parametric correlation: the Kendall tau [15,16] (see methods and figure 1 part 2) which depends only on the sign of the observed variation and not on its magnitude. Is is thus a "weaker" describer of the data than the linear correlation coefficient (also called Pearson coefficient of correlation) or the Spearman rank correlation coefficient. The Kendall tau points specifically to monotonic correlations. A high Kendall tau between two genes indicates that their levels of expression vary in the same way: when the expression level of the first gene increases, the expression level of the other one increases also.
Figure 1 Illustration of the methodology used in this study. Example of the results obtained on a hypothetical bacterial circular chromosome model of 4 genes. The gene expression intensities are measured in three experimental conditions. Part 1 is normalised data (mean equal to 0 variance equal to 1) according to experimental conditions. Part 2 is the matrix of Kendall tau (see methods). Part 3 is the autocorrelation matrix with inter-gene distances. Part 4 is the averaged linear autocorrelation.
Then the variation of the Kendall tau coefficient as a function of the distance between genes was measured with a standard linear autocorrelation function [15,16] (see methods and figure 1 part 3). The linear autocorrelation enables to point to regularities in a gene Kendall tau vector and therfore to regularities of expression correlated with particular inter-gene distances.
Bacillus subtilis regularities of co-expression across the genome
The analysis of the B. subtilis transcription data was performed on a set of 262 experimental conditions gathered from eleven independent experiments measuring expression data over the whole genome. A global view of the regularities of co-expression was obtained by summing up the autocorrelation vectors of all the genes (see figure 1 part 4 and results in figure 2 -blue curve).
Figure 2 Long-range averaged autocorrelations in B. subtilis. To identify the regularities which are common to most of the genome, regardless of the genes localisation, the autocorrelation vectors of all the genes were summed (blue curve). This global signal shows the averaged autocorrelation regularities as a function of inter-gene distance. The green curve shows the averaged autocorrelation when the genes positions on the genome were randomly assigned. The red curve represents the resultant of four oscillations of periods 600 ± 55, 240 ± 21, 113 ± 21 and 60 ± 6 genes, which were estimated from the averaged autocorrelation deconvolution. The horizontal scale represents the distance between two genes (the difference of their ranks on the chromosome). The green, blue and red curves have the same vertical scale. The red curve was shifted for readability. Whereas the green signal shows no regularity, long-range correlations can be seen in the blue signal (maxima at ca. 200, 650, 850, 1300, 1500 and 2050 inter gene distance and minima at ca. 550, 900 and 1750–1950).
The averaged linear autocorrelation of changes in gene expression varies as a function of the inter-gene distance. The green curve in figure 2 corresponds to the averaged autocorrelation evaluated after random permutation of the gene positions on the chromosome. Here the variations are small and independent of the inter-gene distances. Those points where the autocorrelation (blue curve) departs from the random signal (green curve) correspond to couples of genes, for which changes in expression levels are statistically correlated (when the blue curve is above the green one) or anti-correlated (when the blue curve is below the green one).
The autocorrelation function shows regular oscillations at large scale, with maxima at a distance of 200, 650, 850, 1300, 1500 and 2050 genes and minima at a distance of 550, 900 and from 1750 to 1950 genes. Note that the inter-gene distance 2050 corresponds to diametrically opposite genes on the B. subtilis chromosome. The autocorrelation function can be seen as the resultant of four oscillations of periods 600 ± 55, 240 ± 21, 113 ± 21 and 60 ± 6 genes. This representation explains 85% of the autocorrelation oscillations (figure 2 – red curve).
The averaged autocorrelation was analysed on a smaller scale with an inter-gene distance comprised between 1 and 150 genes (figure 3 – blue curve). Closely spaced genes on the chromosome show changes in expression levels that are highly correlated. The averaged autocorrelation of two contiguous genes is 0.4. The low-scale autocorrelation can be decomposed into two regimes: (i) inter-gene distances between 1 and 5 (or 6) genes are characterised by a high and rapidly decaying autocorrelation; (ii) beyond a 6 inter-gene distance the autocorrelation shows a regular and slower decay with periodic oscillations of 14 to 15 genes (figure 3 – red curve). The autocorrelation merges with the noise background around an inter-gene distance of 100 genes (corresponding roughly to 100 kb).
Figure 3 Short-range co-expression regularities in B. subtilis. To identify the regularities which are common to most of the genome, regardless of the genes localisation, the autocorrelation vectors of all the genes were summed (blue curve). This global signal shows the averaged autocorrelation regularities as a function of inter-gene distance. The green zone shows the averaged autocorrelation when the genes positions on the genome were randomly assigned (mean of the random signal ± the root mean square deviation). The horizontal scale represents the distance between two genes (the difference of their ranks on the chromosome). Neighbouring genes on the chromosome show highly correlated variations of expression levels. The averaged autocorrelation of two contiguous genes is 0.4. The signal can be decomposed into two parts: (i) inter-gene distances between 1 and 5–6 genes are characterised by a high autocorrelation, which drops steeply; (ii) beyond 6 genes the autocorrelation shows a regular and slower decrease. The autocorrelation merges with the background noise at an inter-gene distance of about 100 genes (similar to 100 kb). The autocorrelation decrease may be seen as the resultant of a linear decrease and 14.5 ± 1 genes period oscillations (red curve).
The oscillations of the averaged autocorrelations of the 4108 B. subtilis genes shown in figure 2 may result (i) either from regularities specific to some genes or some regions; (ii) or from an overall property that would be shared by all the genes regardless of their positions on the chromosome. In order to ascertain which hypothesis is the correct one, the sums of the autocorrelations of continuous groups of 10, 100 and 500 genes were calculated. All the curves obtained are highly similar (data shown for groups of 500 genes, figure 4). The peaks obtained with these groups of genes are identical to those found in the global signal. Hence they do not depend on any particular position on the genome: in other words, the results show that any gene A has its changes in expression level correlated with the changes in expression levels of those genes that are 200, 650, 850, 1300, 1500 and 2050 genes apart and anti-correlated with those that are 550, 900 and 1750–1950 genes apart. This property is independent of the position of gene A.
Figure 4 Partial sums of the autocorrelations in B. subtilis. To analyse if the discovered regularities depend on gene position, the autocorrelation vectors of groups of 500 genes were summed up (9 coloured curves). The horizontal scale represents the distance between two genes (the difference of their ranks on the chromosome). All the curves were vertically shifted for readability. The signals show the co-expression regularities according to inter-gene distance. Long-range periodicities are shared by all the signals regardless of the gene groups.
Escherichia coli regularities of co-expression across the genome
The same work was performed on E. coli with a data set of 106 experimental conditions. This data set is therefore smaller than that used for B. subtilis. In addition there are more missing data for E. coli than for B. subtilis.
Figure 5 represents the variations of the averaged autocorrelation of all the genes as calculated with the actual gene positions (blue curve) and with random gene positions (green curve). The points where the autocorrelation (blue curve) departs from the random signal (green curve) correspond to couples of genes, the change in expression levels of which are correlated (when the blue curve is above the green one) or anti-correlated (when the blue curve is below the green one).
Figure 5 Long-range averaged autocorrelation in E. coli. To identify the regularities which are common to most of the genome, regardless of the genes localisation, the autocorrelation vectors of all the genes were summed (blue curve). This global signal shows the averaged autocorrelation regularities as a function of inter-gene distance. The green curve shows the averaged autocorrelation when the genes positions on the genome were randomly assigned. The red curve represents the resultant of two oscillations of periods 557 ± 30 and 100 ± 18 genes, which were estimated from the averaged autocorrelation deconvolution. The horizontal scale represents the distance between two genes (the difference of their ranks on the chromosome). The green and blue curves have the same vertical scale. The red one is on a scale, which is moved down for readability. Whereas the green signal shows no regularity, long-range periodicities can be seen in the blue signal (maxima at ca. 200, 650, 1100, 1400 and 1700 and minima at ca. 850, 1380 and 2180).
The main characteristics of figures 2 and 5 are similar. Both bacteria share the steep decay of the averaged autocorrelation curve for inter-gene distances lower than 100 genes and two maxima at a distance of 200 and 650 genes. However there are some differences between B.subtilis and E.coli for long-range peaks since some of them are shifted: maxima at 1300 and 1500 in B.subtilis correspond to peaks at 1100 and 1400 in E.coli, respectively. The minimum at 900 in B.subtilis is shifted to 850 in E.coli. Some peaks and troughs, however, are specific to one specie such as those located at 1380, 1700 and 2180 in E.coli and at 550, 850, 1750–1900 and 2050 in B.subtilis. Probably due to the greater number of missing data the autocorrelation function is noisier for E.coli than for B.subtilis.
Discussion
Comparison of our results to already published observations
What has already been observed
The present study of gene expression data from B.subtilis and E.coli has allowed us to confirm and extend some previously published observations:
• We show for both bacteria that closely spaced genes exhibit highly correlated expression levels. This correlation decreases rapidly with oscillations having a period of 14.5 ± 1 genes corresponding to 14.5 ± 1 kb. Short-range correlations are obvious in the study by Sabatti et al [11] of gene expression data from E.coli. Jeong et al [9] have also observed short-range correlations up to 16 kb in their analysis of expression changes during replication in various E. coli strains.
• In this work the averaged autocorrelation function for E. coli may be seen as the resultant of two main oscillations (with periods of 557 ± 30 kb and 100 ± 18 kb). In B. subtilis we observe four oscillations (with periods of 600 ± 55 kb, 240 ± 21 kb, 113 ± 21 kb and 60 ± 6 kb). Rocha et al [17] analysed the distribution of the genes involved in sulphur metabolism in the genome of E.coli and found a number of them to be clustered into statistically significant islands located 650 kb apart. In their study of transcriptional activities in E.coli, Jeong et al [9] have observed significant correlations for genes located 690 kb or 523 kb apart (depending on physiological conditions) together with a clump of periods around 115 kb.
New results
• We show here that the long-range and short-range correlations are similar in E. coli and B. subtilis. That the observed regularities should be shared by two widely distant bacteria immediately suggests that it could be a property common to other bacteria as well.
• In addition, our results are indicative of an unexpected property that may well modify the current model of the nucleoid organisation: the changes in the level of expression of any gene are correlated (positively or negatively) to the changes in the expression level of other genes, located at well-defined long-range distances and regardless of their localisation on the chromosome in both organisms.
• The long-range periods of the autocorrelation function do not correspond to the 100 kb domain organisation, which may result from the control of topological constraints on the rotation of the double helix [12] and was observed in a study of the positions of genes that are controlled by a sequence-specific transcriptional regulator and the genes encoding this regulator [18]. They do not correspond either to the macro-domain of 1 or 2 Mb proposed by Niki et al [2] and by Valens et al [5]. As all the genes exhibit the same long-range correlations, the phenomenon cannot be explained by some process involving regulators. Conversely, the observations made by Jeong et al [9] may be the result of the general phenomenon observed in this study.
Our interpretation
Gene transcription can occur only on the nucleoid surface. Thus the expression correlations that we observed imply that the involved pairs of genes lies on this surface. However all the genes cannot be on the nucleoid surface at the same time. Therefore depending on the external conditions and/or physiological requirements of the cell, different groups of co-expressed genes should be accessible to the transcriptional machinery. Such constraint seems hardly compatible with an unstructured spatial organisation of the chromosome. Similarly a disordered or random packing is very unlikely to result in the significant periodicities described above. Rather, our observations suggest that the nucleoid must be packed in a fairly structured way.
Knowledge about the nucleoid and ribosomes sizes
The genome sizes of E. coli and B. subtilis are respectively 4.6 Mb (4425 genes encoding proteins) and 4.2 Mb (4108 genes encoding proteins). Half of the genes belong to an operon. The operons have an average size of three genes [12,19]. The nucleoid (the chromosome) shows up as a cylinder of approximate size of 0.5 × 0.7 μm [12,20]. Its circumference of 1.5 μm corresponds approximately to 16 kb of uncoiled DNA, or 16 genes. The diameter of a ribosome is 0.025 μm [21], hence 25 to 30 ribosomes can be juxtaposed along the cylinder length of 0.7 μm.
The possible chromosome configuration
We assume that the nucleoid structure consists of a solenoid with two types of spirals (figure 6):
Figure 6 The possible chromosome configuration. We assume that the nucleoid structure consists of a solenoid with two types of spirals: • Large spirals of uncoiled DNA, containing the genes that are transcribed, that lie on the surface of the nucleoid and define its diameter (0.5 μm). • Small spirals of coiled untranscribed DNA that lie inside the nucleoid.
• Large spirals of uncoiled DNA, containing the genes that are transcribed, that lie on the surface of the nucleoid and define its diameter.
• Small spirals of coiled untranscribed DNA that lie inside the nucleoid.
Cellular elements, in particular the ribosomes on the surface of the nucleoid, impose limits to the number of large expressed spirals. The distance between two large spirals cannot be shorter than the diameter of the ribosome; hence a maximum of 25 to 30 uncoiled DNA large spirals may stand on the nucleoid surface (see knowledge about the nucleoid and ribosomes sizes).
Short-range correlations show that contiguous co-expressed genes do not span more than 100 kb, hence no more than 6 large spirals. We can therefore assume that the average length of contiguous uncoiled DNA is 3 large spirals (see figure 6). This will make 8 to 10 groups of three consecutive large DNA spirals distributed along the chromosome.
Explanation of our results by this nucleoid representation
• The short-range correlations may be seen as resulting from two phenomena:
- The co-ordinated expression of the genes within operons. This explains the correlations in the expression of pairs of genes that are less than 5–6 genes apart.
- The presence of one or more consecutive DNA large spirals of approximately 16 genes on the nucleoid surface. The 14.5 ± 1genes period observed in the variations of the autocorrelation function points to those genes that belong to successive spirals and lie on a generatrix of the nucleoid cylinder.
• For long-range correlations we find 10 maxima in E. coli and 11 maxima in B. subtilis. These maxima probably result from groups of large DNA spirals on the nucleoid surface.
However, such a static representation of the nucleoid does explain neither the alternating pattern of maxima and minima nor their positions.
The dynamic of the nucleoid: a phenomenon, which is not fully explained
The dynamic of the nucleoid structure corresponds to the shift between small spirals of unexpressed coiled DNA to large spirals of expressed uncoiled DNA, and vice-versa. The large spirals are present only when there is effective transcription [22]. The transcription process can explain some of our observations:
- Long-range anticorrelations can result from coil-coiled DNA in small spirals next to large expressed spirals. It has been shown indeed that the opening of the double-stranded DNA during transcription leads to waves of compression of those regions of the chromosome that are close to the transcribed DNA [23]. It can therefore be speculated that the expression of the genes in large spirals leads to the super-coiling of the neighbouring small spirals, hence to the impossibility of opening its DNA and to its transcription.
The pattern of maxima is more difficult to explain since it does not correspond to multiples of a single inter-gene distance. In the case of B. subtilis for example, the maxima are at inter-gene distances that are multiples of 650 and multiples of 650 plus 200 (200; 650, 850; 1300, 1500). We speculate that this pattern is a consequence of the dynamic of the nucleoid structure but we currently have no explanation for it. Current work is in progress to try to explain the maxima and minima of the correlation function, which is reminiscent of a beat phenomenon between two stable waves that could be generated by the transcription process.
Conclusion
The analysis of gene expression data compendium provided information on the nucleoid organisation in circular double stranded DNA bacteria. Our results confirm and complete other observations like those obtained by microscopy. Co-expression variations of neighbouring genes on the chromosome suggest that large DNA spires of 14 to 16 genes length stay on the nucleoid surface. This estimation of a large spire length corresponds to the estimation by microscopy of the nucleoid circumference. The contiguous DNA on the nucleoid surface does not exceed around 100 genes (which is equivalent to 100 kb). This segment is organised in several large spirals of 14 to 16 genes.
The long-range correlation pattern is more surprising: the changes in level of expression for any gene are correlated (positively or negatively) to the changes in expression level of genes, located at well-defined long-range distances independently of their location on the chromosome. This original observation is based on the analysis of several independent sets of gene expression data, which put together a great variety of physiological conditions. However the long-range correlations do not correspond to the domains identified so far in the nucleoid. We are currently exploring a model where the long-range correlations could result from a beat phenomenon between compression and decompression waves generated by the transcription process.
Methods
Data used and normalisation
The microarray data sets have been downloaded from the following websites.
Bacillus subtilis
- Helmann et al [24] at the Stanford microarray database
- Yoshida et al[25,26], Ogura et al [27,28], Kobayashi et al [29], Asai et al [30], Doan et al [31], Molle et al [32], and Watanabe et al [33] at KEGG expression database
- Jarmer et al [34] at the Center for Biological Sequence analysis site
Escherichia coli
- Mori et al [35] at KEGG expression database
- Newton et al [36] at
The data were normalised (mean equal to 0 and variance equal to 1) according to the experimental conditions (figure 1 part 1). They were concatenated for each organism leading to a file of gene expression levels made of 262 experimental conditions for B. subtilis and 106 experimental conditions for E. coli.
Estimation of the correlations and the regularities (figure 1)
The aim of this article is to observe how gene co-expressions vary as a function of the inter-gene distance.
1. For each organism the co-expression among each pair of genes is evaluated with a non-parametric correlation: the Kendall tau [15,16] (figure 1 part 2).
To define the Kendall tau τ, we start with the N data points (xi, yi), the expression levels of the genes x and y in the experimental condition i, respectively. Considering all the 1/2N(N - 1) pairs of data points (xi, yi) (xj, yj), we call a pair "concordant" if the differences (xi-xj) and (yi-yj) have the same sign and "discordant" if the differences have opposite signs. If (xi-xj) is equal to zero, we call the pair an "extra y pair." If (yi-yj) is equal to zero, we call the pair an "extra x pair." If both (xi-xj) and (yi-yj) are equal to zero the pair is ignored. Kendall's tau τ is the following simple combination of these various counts:
2. For each gene, we evaluate its distances from those other genes, the expression levels of which vary simultaneously. The variations of co-expression according to inter-gene distance (figure 1 part 3) are evaluated with the linear autocorrelation [16] on the gene's Kendall tau vector.
The autocorrelation for an inter-gene distance of j is calculated as followed:
with y the Kendall tau vector of a gene and the mean of y, N the number of genes
Note that the bacterial chromosome is circular, so there is no boundary problem. Note that the distance between two genes used in this article is the difference their ranks on the chromosome (approximately equivalent to the number of kb).
Signal deconvolution and estimation of the periodicities
The variation of co-expression according to the inter-gene distance is a superimposition of several periodicities (from small to large scale). To identify these periodicities the averaged autocorrelation signal was deconvoluated with Peakfit 4.06 (Jandel Scientific, San Rafael, CA). The percentage of the autocorrelation that this representation explains is calculated as follow:
with y the autocorrelation vector and x the signal generated by the sum of the deconvolution periodicities and N the number of genes.
Authors' contributions
ASC collected the data, performed the statistical analyses and drafted the manuscript. AG and BT participated in the statistical analysis. AH conceived the study, participated in its analysis and coordination. All authors participated to the elaboration of the model, read and approved the final manuscript.
Acknowledgements
The authors wish to thank A. Riva and J.-L. Risler for critical reading of the manuscript. The authors are indebted to Infobiogen for the disk space and calculation time provided on their servers. A.-S. Carpentier was supported by a FCPR of the Ministère de l'Agriculture.
==== Refs
Lovett ST Segall AM New views of the bacterial chromosome EMBO Rep 2004 5 860 864 15319779 10.1038/sj.embor.7400232
Niki H Yamaichi Y Hiraga S Dynamic organization of chromosomal DNA in Escherichia coli Genes Dev 2000 14 212 223 10652275
Zimmerman SB Underlying regularity in the shapes of nucleoids of Escherichia coli: implications for nucleoid organization and partition J Struct Biol 2003 142 256 265 12713953 10.1016/S1047-8477(02)00637-8
Cunha S Woldringh CL Odijk T Polymer-mediated compaction and internal dynamics of isolated Escherichia coli nucleoids J Struct Biol 2001 136 53 66 11858707 10.1006/jsbi.2001.4420
Valens M Penaud S Rossignol M Cornet F Boccard F Macrodomain organization of the Escherichia coli chromosome Embo J 2004 23 4330 4341 15470498 10.1038/sj.emboj.7600434
Audit B Ouzounis CA From genes to genomes: universal scale-invariant properties of microbial chromosome organisation J Mol Biol 2003 332 617 633 12963371 10.1016/S0022-2836(03)00811-8
Nitschke P Guerdoux-Jamet P Chiapello H Faroux G Henaut C Henaut A Danchin A Indigo: a World-Wide-Web review of genomes and gene functions FEMS Microbiol Rev 1998 22 207 227 9862121 10.1016/S0168-6445(98)00014-X
Danchin A Guerdoux-Jamet P Moszer I Nitschke P Mapping the bacterial cell architecture into the chromosome Philos Trans R Soc Lond B Biol Sci 2000 355 179 190 10724454 10.1098/rstb.2000.0557
Jeong KS Ahn J Khodursky AB Spatial patterns of transcriptional activity in the chromosome of Escherichia coli Genome Biol 2004 5 R86 15535862 10.1186/gb-2004-5-11-r86
Steinhauser D Junker BH Luedemann A Selbig J Kopka J Hypothesis-driven approach to predict transcriptional units from gene expression data Bioinformatics 2004 20 1928 1939 15044239 10.1093/bioinformatics/bth182
Sabatti C Rohlin L Oh MK Liao JC Co-expression pattern from DNA microarray experiments as a tool for operon prediction Nucleic Acids Res 2002 30 2886 2893 12087173 10.1093/nar/gkf388
Pettijohn DE F.C. Neidhardt The Nucleoid Escherichia coli and Salmonella Cellular and Molecular Biology 1999 Second Edition Washington, DC, ASM Press
Lee HK Hsu AK Sajdak J Qin J Pavlidis P Coexpression analysis of human genes across many microarray data sets Genome Res 2004 14 1085 1094 15173114 10.1101/gr.1910904
Allocco DJ Kohane IS Butte AJ Quantifying the relationship between co-expression, co-regulation and gene function BMC Bioinformatics 2004 5 18 15053845 10.1186/1471-2105-5-18
Press WH Kendall's Tau Numerical recipes in C : the art of scientific computing 1992 2nd Cambridge, Cambridge University Press, 643 645
Kendall MG Ord JK Stuart A Stationay time-series The advanced theory of statistics 3 Design and analysis, and time series 1983 3 4th London, Griffin
Rocha EP Sekowska A Danchin A Sulphur islands in the Escherichia coli genome: markers of the cell's architecture? FEBS Lett 2000 476 8 11 10878240 10.1016/S0014-5793(00)01660-4
Kepes F Periodic transcriptional organization of the E.coli genome J Mol Biol 2004 340 957 964 15236959 10.1016/j.jmb.2004.05.039
Willenbrock H Ussery DW Chromatin architecture and gene expression in Escherichia coli Genome Biol 2004 5 252 15575978 10.1186/gb-2004-5-12-252
Donachie W Begg KJ Cell Length, Nucleoid Separation, and Cell Division of Rod-Shaped and Spherical Cells of Escherichia coli J Bacteriology 1989 171 4633 4639
Noller HF Nomua M F.C. Neidhardt Ribosomes Escherichia coli and Salmonella Cellular and Molecular Biology 1999 Second Edition Washington, DC, ASM Press
Murphy LD Zimmerman SB Hypothesis: the RNase-sensitive restraint to unfolding of spermidine nucleoids from Escherichia coli is composed of cotranslational insertion linkages Biophys Chem 2002 101-102 321 331 12488011 10.1016/S0301-4622(02)00163-1
Krasilnikov AS Podtelezhnikov A Vologodskii A Mirkin SM Large-scale effects of transcriptional DNA supercoiling in vivo J Mol Biol 1999 292 1149 1160 10512709 10.1006/jmbi.1999.3117
Helmann JD Wu MF Gaballa A Kobel PA Morshedi MM Fawcett P Paddon C The global transcriptional response of Bacillus subtilis to peroxide stress is coordinated by three transcription factors J Bacteriol 2003 185 243 253 12486061 10.1128/JB.185.1.243-253.2003
Yoshida K Kobayashi K Miwa Y Kang CM Matsunaga M Yamaguchi H Tojo S Yamamoto M Nishi R Ogasawara N Nakayama T Fujita Y Combined transcriptome and proteome analysis as a powerful approach to study genes under glucose repression in Bacillus subtilis Nucleic Acids Res 2001 29 683 692 11160890 10.1093/nar/29.3.683
Yoshida K Yamaguchi H Kinehara M Ohki YH Nakaura Y Fujita Y Identification of additional TnrA-regulated genes of Bacillus subtilis associated with a TnrA box Mol Microbiol 2003 49 157 165 12823818 10.1046/j.1365-2958.2003.03567.x
Ogura M Yamaguchi H Yoshida K Fujita Y Tanaka T DNA microarray analysis of Bacillus subtilis DegU, ComA and PhoP regulons: an approach to comprehensive analysis of B.subtilis two- component regulatory systems Nucleic Acids Res 2001 29 3804 3813 11557812 10.1093/nar/29.18.3804
Ogura M Yamaguchi H Kobayashi K Ogasawara N Fujita Y Tanaka T Whole-genome analysis of genes regulated by the Bacillus subtilis competence transcription factor ComK J Bacteriol 2002 184 2344 2351 11948146 10.1128/JB.184.9.2344-2351.2002
Kobayashi K Ogura M Yamaguchi H Yoshida K Ogasawara N Tanaka T Fujita Y Comprehensive DNA microarray analysis of Bacillus subtilis two- component regulatory systems J Bacteriol 2001 183 7365 7370 11717295 10.1128/JB.183.24.7365-7370.2001
Asai K Yamaguchi H Kang CM Yoshida K Fujita Y Sadaie Y DNA microarray analysis of Bacillus subtilis sigma factors of extracytoplasmic function family FEMS Microbiol Lett 2003 220 155 160 12644242 10.1016/S0378-1097(03)00093-4
Doan T Servant P Tojo S Yamaguchi H Lerondel G Yoshida K Fujita Y Aymerich S The Bacillus subtilis ywkA gene encodes a malic enzyme and its transcription is activated by the YufL/YufM two-component system in response to malate Microbiology 2003 149 2331 2343 12949160 10.1099/mic.0.26256-0
Molle V Nakaura Y Shivers RP Yamaguchi H Losick R Fujita Y Sonenshein AL Additional targets of the Bacillus subtilis global regulator CodY identified by chromatin immunoprecipitation and genome-wide transcript analysis J Bacteriol 2003 185 1911 1922 12618455 10.1128/JB.185.6.1911-1922.2003
Watanabe S Hamano M Kakeshita H Bunai K Tojo S Yamaguchi H Fujita Y Wong SL Yamane K Mannitol-1-phosphate dehydrogenase (MtlD) is required for mannitol and glucitol assimilation in Bacillus subtilis: possible cooperation of mtl and gut operons J Bacteriol 2003 185 4816 4824 12897001 10.1128/JB.185.16.4816-4824.2003
Jarmer H Berka R Knudsen S Saxild HH Transcriptome analysis documents induced competence of Bacillus subtilis during nitrogen limiting conditions FEMS Microbiol Lett 2002 206 197 200 11814663 10.1016/S0378-1097(01)00525-0
Mori H Horiuchi T Isono K Wada C Kanaya S Kitagawa M Ara T Ohshima H [Post sequence genome analysis of Escherichia coli] Tanpakushitsu Kakusan Koso 2001 46 1977 1985 11593750
Newton MA Kendziorski CM Richmond CS Blattner FR Tsui KW On differential variability of expression ratios: improving statistical inference about gene expression changes from microarray data J Comput Biol 2001 8 37 52 11339905 10.1089/106652701300099074
| 15938745 | PMC1177944 | CC BY | 2021-01-04 16:39:54 | no | BMC Genomics. 2005 Jun 6; 6:84 | utf-8 | BMC Genomics | 2,005 | 10.1186/1471-2164-6-84 | oa_comm |
==== Front
BMC Health Serv ResBMC Health Services Research1472-6963BioMed Central London 1472-6963-5-441594385910.1186/1472-6963-5-44Research ArticleUnhappy doctors? A longitudinal study of life and job satisfaction among Norwegian doctors 1994 – 2002 Nylenna Magne [email protected] Pål [email protected]ørde Reidun [email protected] Olaf G [email protected] Department of Public Health and General Practice, Norwegian University of Science and Technology, N-7489 Trondheim, Norway2 Directorate for Health and Social Affairs, PO Box 7000, St Olavs plass, N-0130 Oslo, Norway3 Medical Faculty Division, Akershus University Hospital, University of Oslo, Sykehusveien 27, N-1434 Nordbyhagen, Norway4 The Research Institute of the Norwegian Medical Association, PO Box 1152 Sentrum, N-0107 Oslo, Norway5 Section for Medical Ethics, Department of General Practice and Community Medicine, University of Oslo, P.O.Box 1130 Blindern, N-0318 Oslo, Norway6 Department of Health Management and Health Economics, University of Oslo, PO Box 1089, Blindern, N-0317 Oslo, Norway2005 8 6 2005 5 44 44 6 1 2005 8 6 2005 Copyright © 2005 Nylenna et al; licensee BioMed Central Ltd.2005Nylenna 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
General opinion is that doctors are increasingly dissatisfied with their job, but few longitudinal studies exist. This study has been conducted to investigate a possible decline in professional and personal satisfaction among doctors by the turn of the century.
Methods
We have done a survey among a representative sample of 1 174 Norwegian doctors in 2002 (response rate 73 %) and compared the findings with answers to the same questions by (most of) the same doctors in 1994 and 2000. The main outcome measures were self reported levels of life satisfaction and job satisfaction according to the Job Satisfaction Scale (JSS).
Results
Most Norwegian doctors are happy. They reported an average life satisfaction of 5.21 in 1994 and 5.32 in 2002 on a scale from 1 (extremely dissatisfied) to 7 (extremely satisfied). Half of the respondents reported a very high level of general life satisfaction (a score of 6 or 7) while only one third said they would have reported this high level of satisfaction five years ago. The doctors thought that they had a higher level of job satisfaction than other comparable professional groups. The job satisfaction scale among the same doctors showed a significant increase from 1994 to 2002. Anaesthesiologists and internists reported a lower and psychiatrists and primary care doctors reported a higher level of job satisfaction than the average.
Conclusion
Norwegian doctors seem to have enjoyed an increasing level of life and job satisfaction rather than a decline over the last decade. This challenges the general impression of unhappy doctors as a general and worldwide phenomenon.
==== Body
Background
"Doctors are unhappy.... The unhappiness has been illustrated in a plethora of surveys...", Richard Smith, the then editor of BMJ, concluded a few years ago [1]. "Anecdotes and an expanding body of empirical data suggest a widespread professional malaise", Abigail Zuger wrote in The New England Journal of Medicine last year [2]. Doctor discontent is a current topic for discussion in medical journals, and there seems to be a general agreement – at least within the profession itself – that job satisfaction among physicians is declining. But how well founded on empirical investigations is this impression?
A literature search through PubMed [3] in January 2005 on the combination "discontent + physician" revealed 41 English language papers published during the last 25 years. Most of them were editorials and commentaries. Only five of the publications were based on original data on doctors' satisfaction [4-8]. The only evidence of a negative trend over time among the 41 publications was given by Murray et al. [8]. They found a declining satisfaction with most areas of practice among general internists and family practitioners in Massachusetts, USA, based on data from two cross-sectional studies in 1986 and 1997 respectively [8]. Even though there must be other reports than found by this search, there is reason to believe that the current debate is based more on opinion than on evidence.
In Norway, doctors' job satisfaction and subjective well-being have been studied since 1994 as part of an extensive research programme on physicians' health, sickness, and working conditions [9]. A representative reference panel has been surveyed on a regular basis [10]. The latest survey was undertaken in 2002 and includes Norwegian doctors' self-reported happiness and how they consider their own job satisfaction as compared with their belief about the job satisfaction of other professionals. A comparison with data from similar surveys in 1994 and 2000 might show whether there had been a shift over time.
Methods
The Research Institute of The Norwegian Medical Association recruited in 1993 a Reference Panel by inviting a random sample of 2 000 active doctors to participate. 1 272 doctors agreed, and they have since received questionnaires more or less annually. 21 doctors have dropped out, due to death or voluntary withdrawal. In January 2000 another 795 randomly selected doctors who had received their license after 1993 were invited to join the panel, of which 365 agreed. Hence the number of panellists by primo 2000 was 1 616. Between 2000 and 2002 ten doctors died or withdrew from the panel. The remaining 1 606 Norwegian doctors were sent an 11-page questionnaire in June 2002 with one reminder in August, including several questions on subjective well-being:
- On a three point scale (less satisfied, as satisfied, more satisfied) they were asked to compare their own job satisfaction with the one they believe applies to nine other professional groups.
- They were asked to rate on a scale from 1 (extremely dissatisfied) to 7 (extremely satisfied) the question: "When you think about your life at the moment, would you say that by and large you are satisfied with life or are you mostly dissatisfied?" Many of the same doctors had also answered this question in 1994. On a similar scale they were asked to indicate how they would have answered this question five years ago and how they expected to rate it five years ahead. The distribution of responses to these questions is sufficiently close to normal to use parametric tests.
- Ten questions on different aspects of working conditions (responsibility, variation, collaborators, physical working conditions, possibility of using own skills, choice between work methods, positive feedback for good work, pay and work hours) make up the job satisfaction scale (JSS) based on the work by Warr et. al. [11]. These questions are scored from 1 (extremely dissatisfied) to 7 (extremely satisfied) and add up to a scale with a theoretical range from 10 (dissatisfied with job) to 70 (satisfied with job). The distribution is close to normal. JSS results were also available for 1994 and 2000 from most of the respondents.
The 43 medical specialties were reduced to nine categories: no speciality, specialists in family medicine, laboratory medicine (including pathology and radiology), internal medicine (including paediatrics and neurology), surgery, anaesthesiology, gynaecology and obstetrics, psychiatry and public health. Comparisons of the scales over time were made with GLM (general linear modelling) repeated measures, with age as covariate (to control for possible age effects) and between groups with error bar plots. Differences in categorical variables were given as proportions (percentages) with 95% confidence intervals. The analyses were performed with SPSS 12.
Results
A total of 1 174 doctors completed the questionnaire (369 women (31 %), 805 men (69 %)) which gave a response rate of 73 %. For 874 of the doctors we also had life satisfaction data from 1994. For 886 of the doctors we had job satisfaction data from 1994 and 2000, for 981 we had such data from 2000 and 2002, and for 841 from 1994 and 2002. 722 doctors were registered at all three points in time.
With the exception of aviation pilots, who were believed to be on the same level of satisfaction, the respondents thought themselves to be more satisfied with their job than six other professional groups mentioned (lawyers, clergymen, psychologists, craftsmen, farmers, policemen/-women). Three out of four and two out of three, respectively, believed themselves to be more satisfied than teachers and nurses. 67 % would have chosen a medical career over again if they were young today.
The reported level of subjective well-being was high, with 51.3 % (95 % CI 48.4 to 54.3) of the doctors reporting very high life satisfaction (6 or 7 on a scale from 1 to 7). When asked what they would have answered five years ago only 35.3 % (32.5 to 38.1) reported the same high level of satisfaction. The expected life satisfaction five years ahead did not differ from the present situation (51.5% expected a very high life satisfaction (48.4 – 54.5)). The doctors (N = 873) who had answered the question on subjective well-being eight years earlier had an estimated average score of 5.26 (5.19 to 5.33) in1994 and 5.34 (5.26 to 5.42) in 2002.
The 2002 survey showed an average of 52.0 on the job satisfaction scale (range 10 to 70). There was a moderate positive correlation with age (job satisfaction increased with increasing age) (r = .17, p < .0001) but no significant gender difference. Means of the different specialist groups are shown in figure 1. Anaesthesiologists and specialists in internal medicine scored significantly lower than the overall mean, while psychiatrists and specialists in family medicine and public health scored significantly higher.
Figure 1 Job satisfaction scale among 1 117 Norwegian doctors in different specialty groups in 2002 (numbers (N) in parentheses). Means and 95% confidence intervals. The vertical line indicates the overall mean. Higher values mean higher level of satisfaction.
A GLM analysis of the doctors who answered the same questions at all three points in time showed an estimated mean difference between 1994 and 2000 of 1.37 (.71 to 2.03), and between 1994 and 2002 of 1.09 (.42 to 1.76), i.e. positive and statistically significant changes.
Figure 2 shows changes in estimated JSS-means over time for the same 722 doctors in nine different specialty groups. Anaesthesiologists show a steady decrease, whereas psychiatrists and specialists in family medicine and public health show a steady increase in job satisfaction.
Figure 2 Job satisfaction scale among 722 Norwegian doctors in different specialty groups in 1994, 2000 and 2002. GLM-estimated means, controlled for age. Higher values mean higher level of satisfaction.
Discussion
Our findings do not confirm a declining degree of happiness among Norwegian doctors over the last eight years. Data from 2002 show a high level of general satisfaction. Over half of the doctors regarded themselves as very satisfied with life and they believed that they are more satisfied with life now than five years ago. Answers to the identical question on present life satisfaction among the same doctors show no decline over an eight-year period. Most of the doctors believed that they had a higher level of job satisfaction than other professional groups. A comparison of reported job satisfaction among the same doctors between 1994 and 2002 showed an increasing rather than a declining level over the years, but there were interesting differences among specialties. Anaesthesiologists and other hospital-based doctors did not show the trend of increasing job satisfaction seen among psychiatrists and doctors who work outside hospitals. Our next surveys may show whether this difference is increasing.
Although this is one of the few longitudinal studies on doctor satisfaction it has several limitations, the most important being the validity and reliability of questionnaire data on subjective well-being. Reliability tests suggest that some respondents indicate quite different levels of satisfaction over a very short period of time. On the other hand, our samples and scales should be large enough to account for this, larger than those of Murray et al. [8], and as opposed to them we are actually following the same doctors over time. Another limitation lies of course in the response rate and to what extent our respondents are representative of doctors in general. However, our response rates are higher than in most comparable studies, and we have no reason to believe that our panel is not representative of Norwegian doctors, as has been discussed elsewhere [12].
In spite of the impression created by the medical community most physicians continue to report overall career satisfaction [13]. A recent study of doctors who graduated from UK medical schools in 1974 showed a high level of job satisfaction, with a median score of 18 – 19 on a scale from 5 to 25 [14]. Even in the USA, where discontent has become an accepted characterization of doctors' attitude [2,15], satisfaction levels have not changed dramatically [16], though there are reports of a declining career satisfaction related to managed care and decreasing professional autonomy [17,18].
The BMJ has, through a widely cited editorial [1] and a highly selective and non-scientific Internet survey [19], more or less established as an indisputable fact that "unhappy doctors are a worldwide phenomenon" [20]. Our findings challenge the evidence for this statement. No doubt many doctors are discontented with present developments in health care, but the picture is more complex, and at an individual level most Norwegian doctors are still satisfied both with their private and professional life. This accords well with a general paradox: most attention in the press and by the public is given to misery and displeasure, though conversely people across the world report a high degree of happiness whenever they are questioned directly [21].
Conclusion
Most doctors are happy. No decline in professional and personal satisfaction could be found over the last decade. Norwegian doctors had a higher level of job satisfaction in 2002 than in 1994. There were important differences among specialties; general practitioners and psychiatrists were significantly more happy than other doctors. Anaesthesiologists and other hospital-based doctors did not show the trend of increasing job satisfaction found among psychiatrists and primary care doctors.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
MN, PG, RF and OGAa jointly conceived the idea for the study. OGAa analysed the data and all authors interpreted the results. MN drafted the manuscript. All authors revised the paper and approved the final version. MN and OGAa are guarantors.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
The study was funded by The Sickness Compensation Fund of the Norwegian Medical Association
==== Refs
Smith R Why are doctors so unhappy? BMJ 2001 322 1073 4 11337419 10.1136/bmj.322.7294.1073
Zuger A Dissatisfaction with medical practice N Engl J Med 2004 350 69 75 14702431 10.1056/NEJMsr031703
Magee M Hojat M Impact of health care system on physicians' discontent J Community Health 2001 26 357 65 11554499 10.1023/A:1010415301511
Goldacre MJ Lambert TW Parkhouse J Views of doctors in the United Kingdom about their own professional position and the National Health Service reforms J Public Health Med 1998 20 86 92 9602455
Forsberg E Axelsson R Arnetz B Performance-based reimbursement in health care. Consequences for physicians' cost awareness and work environment Eur J Public Health 2002 12 44 50 11968520 10.1093/eurpub/12.1.44
von Vultee PJ Arnetz B The impact of management programs on physicians' work environment and health. A prospective, controlled study comparing different interventions J Health Organ Manag 2004 18 25 37 15133882
Murray A Montgomery JE Chang H Rogers WH Inui T Safran DG Doctor discontent. A comparison of physician satisfaction in different delivery system settings, 1986 and 1997 J Gen Intern Med 2001 16 452 9 11520382 10.1046/j.1525-1497.2001.016007452.x
Aasland OG Olff M Falkum E Schweder T Ursin H Health complaints and job stress in Norwegian physicians. The use of an overlapping questionnaire design Soc Sci Med 1997 45 1615 29 9428082 10.1016/S0277-9536(97)00093-2
Halvorsen PA Kristiansen IS Aasland OG Førde OH Medical doctors'perception of the "number needed to treat" (NNT) Scand J Prim Health Care 2003 21 162 6 14531508 10.1080/02813430310001158
Warr P Cook J Wall T Scales for the measurement of some work attitudes and aspects of psychological well-being J Occup Psychol 1979 52 129 48
Førde R Aasland OG Steen PA Medical end-of-life decisions in Norway Resuscitation 2002 55 235 40 12458059 10.1016/S0300-9572(02)00270-8
Mechanic D Physician discontent JAMA 2003 290 941 6 12928472 10.1001/jama.290.7.941
Davidson JM Lambert TW Goldacre MJ Parkhouse J UK senior doctors' career destinations, job satisfaction, and future intentions: questionnaire survey BMJ 2002 325 685 6 12351361 10.1136/bmj.325.7366.685
Kassirer J Doctor discontent N Engl J Med 1998 339 1543 5 9819454 10.1056/NEJM199811193392109
Landon BE Reschovsky J Blumenthal D Changes in career satisfaction among primary care and specialist physicians, 1997–2001 JAMA 2003 289 442 9 12533123 10.1001/jama.289.4.442
Shearer S Toedt M Family physicians' observations of their practice, well being, and health care in the United States J Fam Pract 2001 50 751 6 11674906
Stoddard JJ Hargraves JL Reed M Vratil A Managed care, professional autonomy, and income J Gen Intern Med 2001 16 675 84 11679035 10.1111/j.1525-1497.2001.01206.x
Edwards N Kornacki MJ Silversin J Unhappy doctors: what are the causes and what can be done? BMJ 2002 324 835 8 11934779 10.1136/bmj.324.7341.835
Myers DG The funds, friends, and faith of happy people Am Psychol 2000 55 56 67 11392866 10.1037//0003-066X.55.1.56
| 15943859 | PMC1177945 | CC BY | 2021-01-04 16:31:50 | no | BMC Health Serv Res. 2005 Jun 8; 5:44 | utf-8 | BMC Health Serv Res | 2,005 | 10.1186/1472-6963-5-44 | oa_comm |
==== Front
BMC Health Serv ResBMC Health Services Research1472-6963BioMed Central London 1472-6963-5-441594385910.1186/1472-6963-5-44Research ArticleUnhappy doctors? A longitudinal study of life and job satisfaction among Norwegian doctors 1994 – 2002 Nylenna Magne [email protected] Pål [email protected]ørde Reidun [email protected] Olaf G [email protected] Department of Public Health and General Practice, Norwegian University of Science and Technology, N-7489 Trondheim, Norway2 Directorate for Health and Social Affairs, PO Box 7000, St Olavs plass, N-0130 Oslo, Norway3 Medical Faculty Division, Akershus University Hospital, University of Oslo, Sykehusveien 27, N-1434 Nordbyhagen, Norway4 The Research Institute of the Norwegian Medical Association, PO Box 1152 Sentrum, N-0107 Oslo, Norway5 Section for Medical Ethics, Department of General Practice and Community Medicine, University of Oslo, P.O.Box 1130 Blindern, N-0318 Oslo, Norway6 Department of Health Management and Health Economics, University of Oslo, PO Box 1089, Blindern, N-0317 Oslo, Norway2005 8 6 2005 5 44 44 6 1 2005 8 6 2005 Copyright © 2005 Nylenna et al; licensee BioMed Central Ltd.2005Nylenna 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
General opinion is that doctors are increasingly dissatisfied with their job, but few longitudinal studies exist. This study has been conducted to investigate a possible decline in professional and personal satisfaction among doctors by the turn of the century.
Methods
We have done a survey among a representative sample of 1 174 Norwegian doctors in 2002 (response rate 73 %) and compared the findings with answers to the same questions by (most of) the same doctors in 1994 and 2000. The main outcome measures were self reported levels of life satisfaction and job satisfaction according to the Job Satisfaction Scale (JSS).
Results
Most Norwegian doctors are happy. They reported an average life satisfaction of 5.21 in 1994 and 5.32 in 2002 on a scale from 1 (extremely dissatisfied) to 7 (extremely satisfied). Half of the respondents reported a very high level of general life satisfaction (a score of 6 or 7) while only one third said they would have reported this high level of satisfaction five years ago. The doctors thought that they had a higher level of job satisfaction than other comparable professional groups. The job satisfaction scale among the same doctors showed a significant increase from 1994 to 2002. Anaesthesiologists and internists reported a lower and psychiatrists and primary care doctors reported a higher level of job satisfaction than the average.
Conclusion
Norwegian doctors seem to have enjoyed an increasing level of life and job satisfaction rather than a decline over the last decade. This challenges the general impression of unhappy doctors as a general and worldwide phenomenon.
==== Body
Background
"Doctors are unhappy.... The unhappiness has been illustrated in a plethora of surveys...", Richard Smith, the then editor of BMJ, concluded a few years ago [1]. "Anecdotes and an expanding body of empirical data suggest a widespread professional malaise", Abigail Zuger wrote in The New England Journal of Medicine last year [2]. Doctor discontent is a current topic for discussion in medical journals, and there seems to be a general agreement – at least within the profession itself – that job satisfaction among physicians is declining. But how well founded on empirical investigations is this impression?
A literature search through PubMed [3] in January 2005 on the combination "discontent + physician" revealed 41 English language papers published during the last 25 years. Most of them were editorials and commentaries. Only five of the publications were based on original data on doctors' satisfaction [4-8]. The only evidence of a negative trend over time among the 41 publications was given by Murray et al. [8]. They found a declining satisfaction with most areas of practice among general internists and family practitioners in Massachusetts, USA, based on data from two cross-sectional studies in 1986 and 1997 respectively [8]. Even though there must be other reports than found by this search, there is reason to believe that the current debate is based more on opinion than on evidence.
In Norway, doctors' job satisfaction and subjective well-being have been studied since 1994 as part of an extensive research programme on physicians' health, sickness, and working conditions [9]. A representative reference panel has been surveyed on a regular basis [10]. The latest survey was undertaken in 2002 and includes Norwegian doctors' self-reported happiness and how they consider their own job satisfaction as compared with their belief about the job satisfaction of other professionals. A comparison with data from similar surveys in 1994 and 2000 might show whether there had been a shift over time.
Methods
The Research Institute of The Norwegian Medical Association recruited in 1993 a Reference Panel by inviting a random sample of 2 000 active doctors to participate. 1 272 doctors agreed, and they have since received questionnaires more or less annually. 21 doctors have dropped out, due to death or voluntary withdrawal. In January 2000 another 795 randomly selected doctors who had received their license after 1993 were invited to join the panel, of which 365 agreed. Hence the number of panellists by primo 2000 was 1 616. Between 2000 and 2002 ten doctors died or withdrew from the panel. The remaining 1 606 Norwegian doctors were sent an 11-page questionnaire in June 2002 with one reminder in August, including several questions on subjective well-being:
- On a three point scale (less satisfied, as satisfied, more satisfied) they were asked to compare their own job satisfaction with the one they believe applies to nine other professional groups.
- They were asked to rate on a scale from 1 (extremely dissatisfied) to 7 (extremely satisfied) the question: "When you think about your life at the moment, would you say that by and large you are satisfied with life or are you mostly dissatisfied?" Many of the same doctors had also answered this question in 1994. On a similar scale they were asked to indicate how they would have answered this question five years ago and how they expected to rate it five years ahead. The distribution of responses to these questions is sufficiently close to normal to use parametric tests.
- Ten questions on different aspects of working conditions (responsibility, variation, collaborators, physical working conditions, possibility of using own skills, choice between work methods, positive feedback for good work, pay and work hours) make up the job satisfaction scale (JSS) based on the work by Warr et. al. [11]. These questions are scored from 1 (extremely dissatisfied) to 7 (extremely satisfied) and add up to a scale with a theoretical range from 10 (dissatisfied with job) to 70 (satisfied with job). The distribution is close to normal. JSS results were also available for 1994 and 2000 from most of the respondents.
The 43 medical specialties were reduced to nine categories: no speciality, specialists in family medicine, laboratory medicine (including pathology and radiology), internal medicine (including paediatrics and neurology), surgery, anaesthesiology, gynaecology and obstetrics, psychiatry and public health. Comparisons of the scales over time were made with GLM (general linear modelling) repeated measures, with age as covariate (to control for possible age effects) and between groups with error bar plots. Differences in categorical variables were given as proportions (percentages) with 95% confidence intervals. The analyses were performed with SPSS 12.
Results
A total of 1 174 doctors completed the questionnaire (369 women (31 %), 805 men (69 %)) which gave a response rate of 73 %. For 874 of the doctors we also had life satisfaction data from 1994. For 886 of the doctors we had job satisfaction data from 1994 and 2000, for 981 we had such data from 2000 and 2002, and for 841 from 1994 and 2002. 722 doctors were registered at all three points in time.
With the exception of aviation pilots, who were believed to be on the same level of satisfaction, the respondents thought themselves to be more satisfied with their job than six other professional groups mentioned (lawyers, clergymen, psychologists, craftsmen, farmers, policemen/-women). Three out of four and two out of three, respectively, believed themselves to be more satisfied than teachers and nurses. 67 % would have chosen a medical career over again if they were young today.
The reported level of subjective well-being was high, with 51.3 % (95 % CI 48.4 to 54.3) of the doctors reporting very high life satisfaction (6 or 7 on a scale from 1 to 7). When asked what they would have answered five years ago only 35.3 % (32.5 to 38.1) reported the same high level of satisfaction. The expected life satisfaction five years ahead did not differ from the present situation (51.5% expected a very high life satisfaction (48.4 – 54.5)). The doctors (N = 873) who had answered the question on subjective well-being eight years earlier had an estimated average score of 5.26 (5.19 to 5.33) in1994 and 5.34 (5.26 to 5.42) in 2002.
The 2002 survey showed an average of 52.0 on the job satisfaction scale (range 10 to 70). There was a moderate positive correlation with age (job satisfaction increased with increasing age) (r = .17, p < .0001) but no significant gender difference. Means of the different specialist groups are shown in figure 1. Anaesthesiologists and specialists in internal medicine scored significantly lower than the overall mean, while psychiatrists and specialists in family medicine and public health scored significantly higher.
Figure 1 Job satisfaction scale among 1 117 Norwegian doctors in different specialty groups in 2002 (numbers (N) in parentheses). Means and 95% confidence intervals. The vertical line indicates the overall mean. Higher values mean higher level of satisfaction.
A GLM analysis of the doctors who answered the same questions at all three points in time showed an estimated mean difference between 1994 and 2000 of 1.37 (.71 to 2.03), and between 1994 and 2002 of 1.09 (.42 to 1.76), i.e. positive and statistically significant changes.
Figure 2 shows changes in estimated JSS-means over time for the same 722 doctors in nine different specialty groups. Anaesthesiologists show a steady decrease, whereas psychiatrists and specialists in family medicine and public health show a steady increase in job satisfaction.
Figure 2 Job satisfaction scale among 722 Norwegian doctors in different specialty groups in 1994, 2000 and 2002. GLM-estimated means, controlled for age. Higher values mean higher level of satisfaction.
Discussion
Our findings do not confirm a declining degree of happiness among Norwegian doctors over the last eight years. Data from 2002 show a high level of general satisfaction. Over half of the doctors regarded themselves as very satisfied with life and they believed that they are more satisfied with life now than five years ago. Answers to the identical question on present life satisfaction among the same doctors show no decline over an eight-year period. Most of the doctors believed that they had a higher level of job satisfaction than other professional groups. A comparison of reported job satisfaction among the same doctors between 1994 and 2002 showed an increasing rather than a declining level over the years, but there were interesting differences among specialties. Anaesthesiologists and other hospital-based doctors did not show the trend of increasing job satisfaction seen among psychiatrists and doctors who work outside hospitals. Our next surveys may show whether this difference is increasing.
Although this is one of the few longitudinal studies on doctor satisfaction it has several limitations, the most important being the validity and reliability of questionnaire data on subjective well-being. Reliability tests suggest that some respondents indicate quite different levels of satisfaction over a very short period of time. On the other hand, our samples and scales should be large enough to account for this, larger than those of Murray et al. [8], and as opposed to them we are actually following the same doctors over time. Another limitation lies of course in the response rate and to what extent our respondents are representative of doctors in general. However, our response rates are higher than in most comparable studies, and we have no reason to believe that our panel is not representative of Norwegian doctors, as has been discussed elsewhere [12].
In spite of the impression created by the medical community most physicians continue to report overall career satisfaction [13]. A recent study of doctors who graduated from UK medical schools in 1974 showed a high level of job satisfaction, with a median score of 18 – 19 on a scale from 5 to 25 [14]. Even in the USA, where discontent has become an accepted characterization of doctors' attitude [2,15], satisfaction levels have not changed dramatically [16], though there are reports of a declining career satisfaction related to managed care and decreasing professional autonomy [17,18].
The BMJ has, through a widely cited editorial [1] and a highly selective and non-scientific Internet survey [19], more or less established as an indisputable fact that "unhappy doctors are a worldwide phenomenon" [20]. Our findings challenge the evidence for this statement. No doubt many doctors are discontented with present developments in health care, but the picture is more complex, and at an individual level most Norwegian doctors are still satisfied both with their private and professional life. This accords well with a general paradox: most attention in the press and by the public is given to misery and displeasure, though conversely people across the world report a high degree of happiness whenever they are questioned directly [21].
Conclusion
Most doctors are happy. No decline in professional and personal satisfaction could be found over the last decade. Norwegian doctors had a higher level of job satisfaction in 2002 than in 1994. There were important differences among specialties; general practitioners and psychiatrists were significantly more happy than other doctors. Anaesthesiologists and other hospital-based doctors did not show the trend of increasing job satisfaction found among psychiatrists and primary care doctors.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
MN, PG, RF and OGAa jointly conceived the idea for the study. OGAa analysed the data and all authors interpreted the results. MN drafted the manuscript. All authors revised the paper and approved the final version. MN and OGAa are guarantors.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
The study was funded by The Sickness Compensation Fund of the Norwegian Medical Association
==== Refs
Smith R Why are doctors so unhappy? BMJ 2001 322 1073 4 11337419 10.1136/bmj.322.7294.1073
Zuger A Dissatisfaction with medical practice N Engl J Med 2004 350 69 75 14702431 10.1056/NEJMsr031703
Magee M Hojat M Impact of health care system on physicians' discontent J Community Health 2001 26 357 65 11554499 10.1023/A:1010415301511
Goldacre MJ Lambert TW Parkhouse J Views of doctors in the United Kingdom about their own professional position and the National Health Service reforms J Public Health Med 1998 20 86 92 9602455
Forsberg E Axelsson R Arnetz B Performance-based reimbursement in health care. Consequences for physicians' cost awareness and work environment Eur J Public Health 2002 12 44 50 11968520 10.1093/eurpub/12.1.44
von Vultee PJ Arnetz B The impact of management programs on physicians' work environment and health. A prospective, controlled study comparing different interventions J Health Organ Manag 2004 18 25 37 15133882
Murray A Montgomery JE Chang H Rogers WH Inui T Safran DG Doctor discontent. A comparison of physician satisfaction in different delivery system settings, 1986 and 1997 J Gen Intern Med 2001 16 452 9 11520382 10.1046/j.1525-1497.2001.016007452.x
Aasland OG Olff M Falkum E Schweder T Ursin H Health complaints and job stress in Norwegian physicians. The use of an overlapping questionnaire design Soc Sci Med 1997 45 1615 29 9428082 10.1016/S0277-9536(97)00093-2
Halvorsen PA Kristiansen IS Aasland OG Førde OH Medical doctors'perception of the "number needed to treat" (NNT) Scand J Prim Health Care 2003 21 162 6 14531508 10.1080/02813430310001158
Warr P Cook J Wall T Scales for the measurement of some work attitudes and aspects of psychological well-being J Occup Psychol 1979 52 129 48
Førde R Aasland OG Steen PA Medical end-of-life decisions in Norway Resuscitation 2002 55 235 40 12458059 10.1016/S0300-9572(02)00270-8
Mechanic D Physician discontent JAMA 2003 290 941 6 12928472 10.1001/jama.290.7.941
Davidson JM Lambert TW Goldacre MJ Parkhouse J UK senior doctors' career destinations, job satisfaction, and future intentions: questionnaire survey BMJ 2002 325 685 6 12351361 10.1136/bmj.325.7366.685
Kassirer J Doctor discontent N Engl J Med 1998 339 1543 5 9819454 10.1056/NEJM199811193392109
Landon BE Reschovsky J Blumenthal D Changes in career satisfaction among primary care and specialist physicians, 1997–2001 JAMA 2003 289 442 9 12533123 10.1001/jama.289.4.442
Shearer S Toedt M Family physicians' observations of their practice, well being, and health care in the United States J Fam Pract 2001 50 751 6 11674906
Stoddard JJ Hargraves JL Reed M Vratil A Managed care, professional autonomy, and income J Gen Intern Med 2001 16 675 84 11679035 10.1111/j.1525-1497.2001.01206.x
Edwards N Kornacki MJ Silversin J Unhappy doctors: what are the causes and what can be done? BMJ 2002 324 835 8 11934779 10.1136/bmj.324.7341.835
Myers DG The funds, friends, and faith of happy people Am Psychol 2000 55 56 67 11392866 10.1037//0003-066X.55.1.56
| 15949037 | PMC1177946 | CC BY | 2021-01-04 20:47:05 | no | BMC Health Serv Res. 2005 Jun 10; 5:46 | latin-1 | BMC Health Serv Res | 2,005 | 10.1186/1472-6963-5-46 | oa_comm |
==== Front
BMC Infect DisBMC Infectious Diseases1471-2334BioMed Central London 1471-2334-5-401591891310.1186/1471-2334-5-40Research ArticleClinical presentation of pertussis in fully immunized children in Lithuania Narkeviciute Irena [email protected] Ema [email protected] Genovaite [email protected] Rimantas [email protected] Center of Pediatrics, Clinic of Children's Diseases, Vilnius University, Santariskiu 4, 08406 Vilnius, Lithuania2 Laboratory of Microbiology, Vilnius University Children's Hospital, Santariskiu 4, 08406 Vilnius, Lithuania3 Department of Mathematics and Informatics, Vilnius University, Naugarduko 24, 03225 Vilnius, Lithuania2005 27 5 2005 5 40 40 15 9 2004 27 5 2005 Copyright © 2005 Narkeviciute et al; licensee BioMed Central Ltd.2005Narkeviciute 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 Lithuania, the vaccination coverage against pertussis is high. Nevertheless, there is a significant increase in pertussis cases in fully immunized children. The aim of our study was to determine the frequency of classical symptoms of laboratory confirmed pertussis and describe its epidemiology in children fully vaccinated against pertussis.
Methods
From May to December 2001, 70 children aged 1 month to 15 years, suffering from prolonged cough were investigated in the Centre of Paediatrics, Vilnius University Children's Hospital. The collected information included personal data, vaccination history, clinical symptoms of the current illness, and treatment before hospitalization. At the admission to the hospital blood samples were taken from all studied children for Bordetella pertussis IgM and IgA.
Results
A total of 53 (75.7%) of the 70 recruited patients with prolonged cough showed laboratory evidence of pertussis. 32 of them were fully vaccinated with whole cell pertussis vaccine (DTP). The age of fully vaccinated patients varied from 4 to 15 years (average 10.9 ± 3.1; median 11). The time period between the last vaccination dose (fourth) and the clinical manifestation of pertussis was 2.6–13 years (average 8.9 ± 3.0; median 9). More than half of the children before the beginning of pertussis were in contact with persons suffering from long lasting cough illness in the family, school or day-care center. The mean duration from onset of pertussis symptoms until hospitalization was 61.4 ± 68.3 days (range, 7 to 270 days; median 30). For 11 patients who had had two episodes (waves) of coughing, the median duration of cough was 90 days, and for 21 with one episode 30 days (p < 0.0002). Most of the children (84.4%) had paroxysmal cough, 31.3% had post-tussive vomiting, 28.1% typical whoop, and 3.1% apnea. Only 15.6% children had atypical symptoms of pertussis.
Conclusion
Fully vaccinated children fell ill with pertussis at the median of 11 years old, 9 years following pertussis vaccination. More than half of the children could catch pertussis at home, at school or day-care center. Clinical picture of pertussis in previously immunized children is usually characterized by such classical symptoms as prolonged and paroxysmal cough, rarely by whopping and post-tussive vomiting, and very rarely by apnea.
==== Body
Background
Pertussis is a highly communicable, vaccine-preventable respiratory disease. The incidence of pertussis has been greatly reduced by massive vaccination. Nevertheless, there is a significant increase in pertussis cases in older children, adolescents and adult people [1-4]. Improved diagnosis, awareness of pertussis, genetic Bordetella pertussis changes and waning of vaccine-induced immunity are the possible reasons for increased incidence of pertussis [1-5]. In the USA the incidence of vaccine-preventable diseases such as measles, rubella, mumps, diphtheria, tetanus has been greatly reduced in the last 15 years. However, the incidence of pertussis cases increased more than twice: 8296 reported cases in 2002 versus 3450 in 1988 [6]. The age distribution of patients with pertussis in the USA in 1994–1996 and 1997–2000 has changed. During the last period, the incidence of pertussis among infants increased by 11%, in children aged 1–4 years decreased to 8%, remained stable for children aged 5–9 years and among adolescents and adults increased by about 60% [7].
In Lithuania immunization of infants and children against pertussis has been introduced since 1956 and massive vaccination started in 1961. According to our standard vaccination schedule, pertussis whole-cell vaccine incorporated in diphtheria-tetanus-pertussis (DTP) vaccine is offered at 3, 4.5 and 6 months of age with a booster dose only at 18 months of age. In 1991, the vaccine coverage among children aged 1 year was 73.2%, whereas this percentage has been increasing and since 1996 reached above 90% (93.6% in 2000, 94.6% in 2001). 35% of all pertussis cases were diagnosed in vaccinated children (at least three DTP vaccine doses) during the period from 1991 to 1995, 33.4% of the cases from 1996 to 2000 and 43.2% in 2001.
Clinical presentation of pertussis in unvaccinated children had been extensively described by several authors [8,9]. The disease in these patients is usually typical and often severe. Data of the clinical course of pertussis in fully immunized children is usually atypical and generally mild [10].
The aim of our study was to determine the frequency of classical symptoms of laboratory confirmed pertussis and describe its epidemiology in fully vaccinated children.
Methods
From May to December 2001, 70 children aged 1 month to 15 years with prolonged cough (duration ≥ 14 days) and siblings with shorter duration cough (but not less than 7 days) were hospitalized and investigated at Vilnius University Children's Hospital, Centre of Paediatrics. The patients were referred to the hospital by general practitioners or pediatricians, because detailed investigation of the children with prolonged cough of unknown etiology was only available in the hospital. The data regarding to the patient's age, vaccination history, clinical symptoms and signs of the current illness, previous treatment was collected on to computer database. Single blood samples were taken from all the children upon admission and sent for detection of specific immunoglobulin (Ig) IgM and IgA antibodies to B. pertussis by an enzyme-linked immunosorbent assay (ELISA). Serological tests were performed and estimated according to the manufacturer's instructions (Labsystems, Finland). Specimens for B. pertussis and B. parapertussis culture were not obtained because of the late illness stage and received previous antimicrobial therapy before hospitalization. Classical symptoms of pertussis were defined as a prolonged cough lasting two weeks or more, paroxysmal cough, inspiratory whoop, post-tussive vomiting and apnea. Confirmed case of pertussis was defined as an episode of cough lasting 7 days or more and positive anti-B. pertussis IgM or IgA or both levels. Two groups of our studied patients were compared. One group of children with prolonged cough who had two episodes (waves) of successive coughing (when first episode was not ended, it means that cough become more intensified and coughing paroxysm renewed) was compared to the second group who had only one episode of prolonged cough.
The descriptive values were expressed as means, medians and standard deviations (SD) or percentages. Two-sided Wilcoxon test for independent samples was used to compare the groups. A p value less than 0.05 were considered statistically significant.
Results
Of the 70 studied children, 53 (75.7%) showed a serological evidence of pertussis. Out of 53 patients, 21 (39.6%) were partially vaccinated or unvaccinated at all and 32 (60.4%) received complete vaccination. Out of 32 fully vaccinated patients, positive B. pertussis IgM antibodies were found in 6, IgA in 3 and both IgM and IgA in 23 patients.
We present epidemiological and clinical analysis of 32 children with confirmed pertussis who received 4 doses whole-cell of pertussis vaccine. Only eight (25%) of 32 fully vaccinated children were referred to the hospital with suspicion of pertussis, other referral diagnosis were bronchitis for 15 children, tracheobronchitis – 4, bronchopneumonia – 4, asthma exacerbation – 1. The characteristics of the children with pertussis are listed in Table 1. There was a difference in gender distribution: male 34.4% and female 65.6%. The age of the patients was from 4 to 15 years (average 10.9 ± 3.1; median 11). Most of the children (75.0%) were older than 9 years. The time period between the last (fourth) vaccination dose and the clinical manifestation of pertussis was 2.6–13 years (average 8.9 ± 3.0; median 9). None of the children had a history of pertussis. Seventeen (53.1%) children before they got ill had been in contact with person suffering from long-lasting cough, 12 of them at home (all of them with siblings and three with one of the parents) and 5 at school or day-care centre. In our study, 9 cases of pertussis were observed in four families: in three families 2 cases in each and in one family 3 cases. Before hospitalization 28 (87.5%) patients had received antibiotic therapy effective against B. pertussis (amoxicillin, ampicillin, macrolide).
Table 1 Characteristics of fully immunized children with pertussis
Characteristics n = 32 %
Age (years):
4–5 2 6.2
6–8 6 18.8
9–11 9 28.1
12–15 15 46.9
Gender:
Male 11 34.4
Female 21 65.6
Antibiotic treatment
before investigation 28 87.5
The rate of pertussis symptoms among fully immunized children is shown in Table 2. The mean duration from the onset of pertussis symptoms until hospitalization was 61.4 ± 68.3 days (range, 7 to 270 days; median 30). Most of the children (22; 68.8%) had cough that lasted more than three weeks. For 11 (34.4%) children who had had two episodes (waves) of successive coughing, the mean duration of cough was 127.3 ± 81.7 days (range, 40 to 270 days; median 90), and for 21 (65.6%) with one episode 26.9 ± 16.2 days (range, 7 to 70 days; median 30). The difference was statistically significant (Wilcoxon test, p < 0.0002).
Table 2 Clinical symptoms of pertussis in fully immunized children
Symptoms n %
Cough 32 100
Duration of coughing (days):
7–14 7 21.9
15–21 3 9.3
22≥ 22 68.8
Paroxysmal cough 27 84.4
Whooping 9 28.1
Post-tussive vomiting 10 31.3
Apnea 1 3.1
The prevalence of other classical symptoms of pertussis was following: paroxysmal cough 27 (84.4%), post-tussive vomiting 10 (31.3%), 9 (28.1%) showed a typical "whoop" and 1 (3.1%) apnea. Five (15.6%) children presented without classical symptoms of pertussis. The duration of their cough was 14 days and more, however, they did not have any coughing paroxysms, whooping, post-tussive vomiting or apnea.
Discussion
Lithuanian Centre of Infectious Disease Control and Prophylaxis reported 162 cases of pertussis in 2001, 43.2% of them were vaccinated against pertussis. Almost half of all patients (47.5%) with pertussis were aged 7–15 years. Our study analyzed the epidemiological data and clinical presentation of 32 fully immunized children with laboratory-confirmed pertussis hospitalized in the Centre of Paediatrics, Vilnius University Children's Hospital from May to December 2001. Detailed analysis of our study data showed that fully vaccinated children got ill with pertussis at the median of 11 years old (range, 4 to 15), after 9 years (range, 2.6 to 13) following four doses vaccination with DTP vaccine. According to the data reported in 1998 in Finland where pertussis vaccination has been in practice for more than 40 years and a four-dose schedule is completed before a child is 2 years old, over the last decade the reported cases of pertussis increased twice and most of the patients were schoolchildren [4]. It is known that vaccine-induced immunity wanes and after 5–10 years makes the vaccinated host vulnerable to infection [11,12]. Reported study from Poland [13] revealed protective immunity against pertussis in 70% of six years old children, in 68% of seven and only in 45% of eight years old children. Thus, waning of immunity, leads to a growing population of pertussis-susceptible older children, adolescents and an increasing proportion of disease cases in these age groups. According to the literature data vaccination with whole-cell vaccines induces protective immunity lasting 6–8 years [11,14], whereas protective immunity induced by vaccination with acellular pertussis vaccine lasts 4–6 years and longer [15-17]. Torvaldsen and McIntyre [18] showed that the fifth dose of pertussis vaccine given at 4–5 years of age reduced the incidence of pertussis in older children.
Many studies have reported the importance of household and school contacts for pertussis infection. The results of the study performed in France during 1993 and 1994 showed that the source of pertussis was a close contact with adult in 46% of identifiable cases, with sibling in 42% of cases and with other family members or friends in 11% of the cases [19]. More recently, in a household contact study in the USA, 53% of people who were identified as the primary source of pertussis infection were aged 13 years or older and 26% were aged 30 years or older [20]. Our study data showed that more than half of the patients had been in contact with persons who had prolonged cough: mostly in families and rarely at schools or day-care centers.
Clinical presentation of pertussis in fully immunized children become talking-point in the clinical practice. Various factors such as patient's age, gender, antibiotic treatment and especially vaccination status may influence clinical presentation of pertussis. Vaccination significantly changes the clinical presentation of pertussis [8-10,21,22]. Prolonged cough is one of the classical symptoms of pertussis. Before investigation the median duration of cough in all our patients was one month. Tozzi et al. [8] in a study of 788 laboratory confirmed cases of pertussis had demonstrated that the duration of cough in vaccinated children was about one month, while in unvaccinated was two times longer. Two episodes of cough and significantly prolonged cough (median 90 days) documented in our study may be accounted for co-infection of B. pertussis with atypical pathogens. Hallander et al. from Sweden [23] demonstrated that the duration of cough increased when more than one agent was detected. A median cough period was 51 days for B. pertussis and 60 days for co-infections with M. pneumoniae or C. pneumoniae.
Unvaccinated children more frequently presented the full spectrum of classical pertussis symptoms than vaccinated children. German study [9] has shown that 90.2% of unvaccinated patients (mean age 4.3 years) had paroxysmal cough, 78.9% whooping and 53.3% post-tussive vomiting. The frequency of paroxysmal cough in fully vaccinated children (median age 11 years) in our study was similar (84.4%), but post-tussive vomiting and whooping was more rare (accordingly 31.3% and 28.2%). Only one 11 years old patient had apnea. The results of our study coincide to those reported from Canada [21]. Of the 103 immunized children (< 5 years of age), 68% developed paroxysmal cough within the first week of their illness and 88% had persistent paroxysmal cough for more than three weeks. Cough generally lasted 16–91 days (median 48). A recent study conducted in Israel [10] has documented the clinical manifestation of pertussis in previously immunized children and young adults (median age 9 years). 21% of the patients had paroxysmal cough, 13% showed post-tussive vomiting, 7% apnea and 6% had classic whoop. The classical symptoms of our studied patients therefore were more frequent compared with Israeli patients. However, we should also take into account other fully vaccinated children who had mild pertussis, but were not referred to the hospital and were not recruited in our study.
Some authors have reported that antibiotic treatment has reduced the severity of pertussis and the duration of cough, especially when they have been started at the beginning of the disease [21]. In opposite, the other studies have demonstrated that antibiotic treatment does not influence on the course of pertussis and is not effective [24,25]. Tozzi et al. [8] has showed that children treated with antibiotics had cough which lasted 6 to 11 days longer and spasmodic cough 4 to 13 days longer than untreated patients. Authors concluded that antibiotic therapy may be a marker of severe disease. We had no possibility to evaluate the role of antibiotics on to the duration of the cough in our patients, because most of our children had received antibiotics before examination.
The results of our study showed that prolonged cough in fully vaccinated children are frequently associated with paroxysmal cough and rarely with other symptoms such as whooping, post-tussive vomiting, apnea. Only 15.6% of the patients with a long duration of cough had no classical symptoms of pertussis. Thus, it is very difficult to suspect pertussis in this group of children in the absence of epidemiological data. Our data suggest that pertussis should be considered in the differential diagnosis of prolonged cough in fully vaccinated children.
Conclusion
Pertussis is a common disease in Lithuania as well as in many other countries with high vaccination coverage. About half of children with pertussis in Lithuania in 2001 were aged 7 to14 years. Fully vaccinated children fell ill with pertussis at the median of 11 years old, 9 years following DTP vaccination. More than half of the children could catch pertussis at home, at school or at day-care center. Clinical picture of pertussis in previously immunized children is usually characterized by such classical symptoms as prolonged and paroxysmal cough, rarely by whopping and post-tussive vomiting, and very rarely by apnea. This study provides useful information for the clinicians about the epidemiology and clinical data of pertussis in fully vaccinated children.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
IN conceived of the study, supervised the design and execution of the study, performed the preliminary and final data analysis, drafted and wrote the manuscript.
EK participated in the design of the study, data assessment, preliminary and final data analysis and writing the manuscript.
GB performed the serologic tests, participated in writing the manuscript.
RE performed the statistical analysis, participated in writing the manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
==== Refs
Christie CD Marx ML Marchant CD Reising SF The 1993 epidemic of pertussis in Cincinnati. Resurgence of disease in a highly immunized population of children N Engl J Med 1994 331 16 21 8202096 10.1056/NEJM199407073310104
de Melker HE Schellekens JF Neppelenbroek SE Mooi FR Rumke HC Conyn-van Spaendonck MA Reemergence of pertussis in the highly vaccinated population of the Netherlands: observations on surveillance data Emerg Infect Dis 2000 6 348 57 10905967
Khetsuriani N Bisgard K Prevots DR Brennan M Wharton M Pandya S Poppe A Flora K Dameron G Quinlisk P Pertussis outbreak in an elementary school with high vaccination coverage Pediatr Infect Dis J 2001 20 1108 12 11740314 10.1097/00006454-200112000-00003
He Q Viljanen MK Arvilommi H Aittanen B Mertsola J Whooping cough caused by Bordetella pertussis and Bordetella parapertussis in an immunized population JAMA 1998 280 635 7 9718056 10.1001/jama.280.7.635
Cherry JD Geffen D The science and fiction of the "resurgence" of pertussis Pediatrics 2003 112 405 6 12897292 10.1542/peds.112.2.405
Rosenthal M Many successes, many challenges with today's immunization program Infectious Diseases in Children 2003
Centers for Disease Control and Prevention Pertussis – United States, 1997–2000 MMWR 2002 51 73 6 11837909
Tozzi AE Rava L Ciofi degli Atti ML Salmaso S the Progetto Pertosse Working Group Clinical presentation of pertussis in unvaccinated and vaccinated children in the first six years of life Pediatrics 2003 112 1069 75 14595048 10.1542/peds.112.5.1069
Heininger U Klich K Stehr K Cherry JD Clinical findings in Bordetella pertussis infections: results of a prospective multicenter surveillance study Pediatrics 1997 100 1 7 9200353 10.1542/peds.100.6.e10
Yaari E Yafe-Zimerman Y Schwartz SB Slater PE Shvartzman P Andoren N Branski D Kerem E Clinical manifestations of Bordetella pertussis infection in immunized children and young adults Chest 1999 115 1254 8 10334136 10.1378/chest.115.5.1254
Jenkinson D Duration of effectiveness of pertussis vaccine: evidence from a 10-year community study BMJ 1988 296 612 4 3126927
Aoyama T Harashima M Nishimura K Saito Y Outbreak of pertussis in highly immunized adolescents and its secondary spread to their families Acta Paediatr Jpn 1995 37 321 4 7645380
Slusarczyk J Dulny G Nowak K Krszyna J Wysokinska T Fordymacka A Gzyl A Janaszek W Gniadek G Immunity of children aged 6–8 against pertussis, tetanus and diphtheria Przegl Epidemiol 2002 56 39 48 12150066
Grimprel E Begue P Anjak I Njamkepo E Francois P Guiso N Long-term human serum antibody responses after immunizations with whole cell pertussis vaccine in France Clin Diagn Lab Immunol 1996 3 93 7 8770511
Salmaso S Mastrantonio P Tozzi AE Stefanelli P Anemona A Ciofi degli Atti ML Giammanco A Stage III Working Group Sustained efficacy during the first six years of life of 3-component acellular pertussis vaccines administered in infancy: the Italian experience Pediatrics 2001 108 E81 11694665 10.1542/peds.108.5.e81
Guiso N Begue P Cohen R Comparison of pertussis antibody levels in children up to 5 years of age primed at 2, 3, 4 months and boostered in a second year of life with either DTPa or DTPw based combination vaccines in France. [Abstract 62] 40th ICAAC, 17–20 September 2000 Toronto, Canada
Lugauer S Heininger U Cherry JD Stehr K Long-term clinical effectiveness of an acellular pertussis component vaccines and a whole cell pertussis component vaccine Eur J Pediatr 2002 161 142 6 11998910 10.1007/s00431-001-0893-5
Torvaldsen S McIntyre PB Effect of the preschool pertussis booster on national notifications of disease in Australia Pediatr Infect Dis J 2003 22 956 9 14614366
Baron S Njamkepo E Grimprel E Begue P Desenclos JC Drucker J Guiso N Epidemiology of pertussis in French hospitals in 1993 and 1994: thirty years after a routine use of vaccination Pediatr Infect Dis J 1998 17 412 8 9613656 10.1097/00006454-199805000-00013
Deen JL Mink CA Cherry JD Christenson PD Pineda EF Lewis K Blumberg DA Ross LA Household contact study of Bordetella pertussis infections Clin Infect Dis 1995 21 1211 9 8589145
Bortolussi R Miller B Ledwith M Halperin S Clinical course of pertussis in immunized children Pediatr Infect Dis J 1995 14 870 4 8584314
Preziosi MP Halloran ME Effects of pertussis vaccination on disease: vaccine efficacy in reducing clinical severity Clin Infect Dis 2003 37 772 9 12955637 10.1086/377270
Hallander HO Gnarpe J Gnarpe H Olin P Bordetella pertussis, Bordetella parapertussis, Mycoplasma pneumoniae, Chlamydia pneumoniae and persistent cough in children Scand J Infect Dis 1999 31 281 6 10482058 10.1080/00365549950163581
World Health Organization Recommended surveillance standarts for pertussis 1999
Mertsola J Ruuskanen O Eerola E Viljanen MK Intrafamilial spread of pertussis J Pediatr 1983 103 359 63 6886900
| 15918913 | PMC1177947 | CC BY | 2021-01-04 16:28:14 | no | BMC Infect Dis. 2005 May 27; 5:40 | utf-8 | BMC Infect Dis | 2,005 | 10.1186/1471-2334-5-40 | oa_comm |
==== Front
BMC Infect DisBMC Infectious Diseases1471-2334BioMed Central London 1471-2334-5-431592706010.1186/1471-2334-5-43Research ArticleEfficacy and safety of telithromycin 800 mg once daily for 7 days in community-acquired pneumonia: an open-label, multicenter study Fogarty Charles M [email protected] Tushar C [email protected] Lala M [email protected] Bruno P [email protected] Spartanburg Medical Research, Spartanburg, South Carolina, USA2 Fall River Walk-in Clinic, Fall River, Massachusetts, USA3 Department of Medicine/Emergency Medicine, Louisiana State University Medical Center, New Orleans, Louisiana, USA4 Aventis Pharmaceuticals, Bridgewater, New Jersey, USA2005 31 5 2005 5 43 43 18 8 2004 31 5 2005 Copyright © 2005 Fogarty et al; licensee BioMed Central Ltd.2005Fogarty 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
Community-acquired pneumonia (CAP) remains a major cause of morbidity and mortality throughout the world. Telithromycin (a new ketolide) has shown good in vitro activity against the key causative pathogens of CAP, including S pneumoniae resistant to penicillin and/or macrolides.
Methods
The efficacy and safety of telithromycin 800 mg orally once daily for 7 days in the treatment of CAP were assessed in an open-label, multicenter study of 442 adults.
Results
Of 149 microbiologically evaluable patients, 57 (9 bacteremic) had Streptococcus pneumoniae. Of the 57 S pneumoniae pathogens isolated in these patients, 9 (2 bacteremic) were penicillin- or erythromycin-resistant; all 57 were susceptible to telithromycin and were eradicated. Other pathogens and their eradication rates were: Haemophilus influenzae (96%), Moraxella catarrhalis (100%), Staphylococcus aureus (80%), and Legionella spp. (100%). The overall bacteriologic eradication rate was 91.9%. Of the 357 clinically evaluable patients, clinical cure was achieved in 332 (93%). In the 430 patients evaluable for safety, the most common drug-related adverse events were diarrhea (8.1%) and nausea (5.8%).
Conclusion
Telithromycin 800 mg once daily for 7 days is an effective and well-tolerated oral monotherapy and offers a new treatment option for CAP patients, including those with resistant S pneumoniae.
==== Body
Background
Community-acquired pneumonia (CAP) remains a major cause of morbidity and mortality throughout the world. In the USA alone, an estimated 3–4 million cases of CAP account for approximately 10 million physician visits, 500,000 hospitalizations, and 45,000 deaths each year [1,2]. Common bacterial pathogens in CAP include Streptococcus pneumoniae, Haemophilus influenzae, Moraxella catarrhalis, Mycoplasma pneumoniae, Chlamydia pneumoniae, and Legionella pneumophila. Ideally, antibiotic therapy for CAP should be based on etiologic diagnosis. The Infectious Diseases Society of America (IDSA) recommends microbiologic testing, including sputum and blood cultures, in hospitalized CAP patients, and encourages sputum culture and Gram staining in CAP outpatients [2]. Practically, however, microbiologic test results are not usually immediately available and a clinician needs to initiate treatment that will cover the most likely pathogens within a few hours of a patient's initial presentation. Complicating the selection of effective empiric therapy is the possibility that a patient may ultimately prove to have a resistant pathogen, an atypical/intracellular pathogen, or both. The increasing prevalence of drug resistance among the typical respiratory tract pathogens is now an important consideration in selecting an antibacterial agent for CAP [2]. High resistance to penicillin (a previously effective therapy for pneumococcal pneumonia) is now > 15% in some areas, and an increasing proportion of S pneumoniae strains are now macrolide-resistant [3,4].
Recognizing the increasing resistance of typical pathogens and increasing incidence of atypical/intracellular pathogens, the American Thoracic Society (ATS) has listed a new class of agents – the ketolides, which belong to the macrolide-lincosamide-streptograminB (MLSB) family – as a potential option for oral therapy in CAP [5]. The ketolide telithromycin has shown good in vitro activity against S pneumoniae resistant to penicillin and/or macrolides [6], and potent activity against both typical and atypical/intracellular respiratory tract pathogens [6-17]. In vitro studies have shown that telithromycin has a low potential to select for resistant strains and does not induce MLSB resistance [18-20]. A once-daily dose of telithromycin 800 mg yields drug concentrations in plasma and bronchopulmonary tissues and fluids in excess of the minimum inhibitory concentration (MIC) of most common respiratory pathogens for 12–24 hours after dosing [21,22]. Concentrations of telithromycin in alveolar macrophages, epithelial lining fluid, and bronchial mucosa 8 hours after a single 800 mg dose exceed those in plasma by factors of 180, 6.5, and 2.2, respectively [21].
The principal objective of the present study was to evaluate the bacteriologic and clinical efficacy and safety of a 7-day regimen of telithromycin 800 mg once daily in patients with CAP, with a focus on infections due to penicillin- and/or erythromycin-resistant S pneumoniae.
Methods
Patient sample
Both in-patients and outpatients ≥ 13 years of age were considered for inclusion in the study. The diagnosis of CAP was based on a new chest x-ray infiltrate and the presence of ≥ 2 of the following: cough; production of purulent sputum; auscultatory findings on pulmonary examination; dyspnea; fever; elevated white blood cell count; or a positive Gram stain in a sputum sample. The decision to admit was based on the clinical judgment of the patient's physician. Women of childbearing potential were required to have a negative pregnancy test and use an accepted method of contraception throughout the study.
Exclusion criteria included: need for parenteral antibacterial therapy; antibacterial therapy for > 24 hours in the previous 7 days; use of azithromycin, ceftriaxone, or dirithromycin within 7 days prior to enrollment; empyema or lung abscess; pulmonary disease requiring a specific treatment that would make the interpretation of the results difficult; suspected nonbacterial pathogen; pathogen resistant to the study medication; known long QTc syndrome; sick sinus syndrome; severe hypokalemia; hypersensitivity to macrolides; terminal illness or immunocompromised status; concomitant medications known to have a potential interaction with telithromycin; and impaired hepatic or renal function, or clinically relevant cardiovascular, neurologic, or endocrine disease.
Study design
Patients were enrolled between January and September 2000 at 37 centers in the USA, 10 centers in South America, 6 centers in South Africa, and 5 centers in Canada. The trial was conducted in accordance with good clinical practice, and the study protocol and other appropriate study-related documents were reviewed and approved by the respective independent ethics committees/institutional review boards in the participating countries. All patients provided written informed consent.
The study groups evaluated comprised:
• Safety population: All patients who received ≥ 1 dose of telithromycin and had ≥ 1 safety assessment after entry into the study
• Modified intent to treat (mITT) population: All patients with a confirmed diagnosis of CAP who received ≥ 1 telithromycin tablet
• Per-protocol clinically evaluable (PPc) population: All mITT patients who completed the study with no major protocol violations
• Per-protocol bacteriologically evaluable (PPb) population: All clinically evaluable patients who also had a bacterial pathogen identified as a causative agent
• Bacteriologic modified intent to treat (bmITT) population: All mITT patients with a bacteriologic sample obtained at pretherapy/entry containing ≥ 1 causative pathogen.
All patients meeting the study inclusion criteria were assigned to 7 days of open-label treatment with telithromycin 800 mg, administered once daily in the morning as two 400 mg tablets. Compliance with treatment was measured by tablet count at the on-therapy visit (Days 3–5) and at or before the post-therapy/test of cure (TOC) visit.
Study endpoints
Bacteriologic outcome in the PPb patients at the post-therapy/TOC visit was the primary efficacy variable. The secondary efficacy variables in the study were bacteriologic outcome in the mITT patients who had ≥ 1 causative pathogen identified at entry (bmITT), and clinical outcome at the post-therapy/TOC visit in the PPc and mITT patients. Special attention was given to infections due to S pneumoniae resistant to penicillin and/or erythromycin.
Bacteriologic evaluation
Sputum samples were obtained for Gram stain, culture, and susceptibility analyses at the 4 study visits: pretherapy/entry (Day 1), on-therapy (Days 3–5), end-of-therapy (Days 8–10), and post-therapy/TOC (Days 17–24), or at early withdrawal if applicable. Blood samples for culture were obtained at pretherapy/entry and at subsequent assessments if cultures were previously positive, a febrile state persisted, or no clinical improvement had occurred after 48 hours of treatment. Susceptibility of pathogens to telithromycin was tested using disk diffusion and dilution methods. Susceptibility to macrolides (i.e. erythromycin) was tested using disk diffusion to establish a minimal resistance profile to these agents for the pathogens isolated. Oxacillin disks were used to screen for penicillin resistance among S pneumoniae isolates, and the nitrocephin test was used to detect β-lactamase production in H influenzae and M catarrhalis.
Tests for detection of atypical/intracellular pathogens included serologic testing of blood samples and/or polymerase chain reaction (PCR) analysis of oropharyngeal swab samples or sputum samples for C pneumoniae and M pneumoniae, and urine and blood samples for L pneumophila. Diagnostic criteria for atypical/intracellular pathogens included negative cultures for common pathogens and the following: for L pneumophila, a 4-fold increase in paired serum immunoglobulin (Ig) G or IgM titers, or a positive urinary antigen for L pneumophila serogroup I; for C pneumoniae, a single IgM titer ≥ 1:32 in combination with a positive PCR, or a 4-fold increase in paired serum IgG/IgM titers; for M pneumoniae, a 4-fold increase in paired serum IgG or a single IgM titer ≥ 1:16 with a positive PCR test. The tests used to confirm the presence of C pneumoniae were based upon approved/agreed FDA procedures, including DNA gene probe of the sputum, acute and convalescent serology for antibodies from the blood, and microimmunofluorescent antigen testing of sputum. M pneumoniae tests were based upon acute and convalescent serology and DNA gene probe. Convalescent serology was not performed on patients initially positive for common pathogens.
Bacteriologic outcome was based on eradication, presumed eradication, or persistence of any pathogen isolated in microbiologic cultures at the pretherapy/entry visit and considered by the investigator to be causative of CAP. Assessments also included analyses for the presence of any new pathogen isolated during or after treatment that was not detected at the pretherapy/entry visit. Bacteriologic outcome was categorized as satisfactory if the causative pathogen was absent in on-therapy or post-therapy cultures (eradication), or if no follow-up culture was available due to clinical improvement (presumed eradication). Bacteriologic outcome was considered unsatisfactory if the causative pathogen was still present (persistence), additional antibacterial therapy was indicated (presumed persistence), a new pathogen emerged during therapy or within 3 days after treatment was completed (superinfection), eradication of the causative organism was followed by replacement with a new species or serotype of the same organism (reinfection), or the causative organism reappeared following eradication (recurrence).
Indeterminate outcomes included cases in which the patient was withdrawn from the study before follow-up cultures were obtained, the microbiologic data were incomplete, concurrent antibacterial treatment was provided for reasons not associated with CAP or respiratory tract infections (RTIs), or death occurred that was not due to CAP or related complications. If ≥ 1 causative pathogen was isolated from the pretreatment culture and the bacteriologic outcome was not the same for all pathogens, the bacteriologic outcome was classified as unsatisfactory.
Clinical evaluation
Patients were evaluated clinically at the four study visits. At entry (Day 1), patients were screened and evaluated for infection-related signs and symptoms, with assessment of vital signs, physical examination, chest x-ray, blood for hematology and chemistry, and a 12-lead electrocardiogram. Similarly, at the on-therapy, end-of-therapy, and post-therapy/TOC visits, patients were assessed for infection-related signs and symptoms, with measurement of vital signs, physical examination and evaluation of overall clinical status, and blood for hematology and chemistry. An electrocardiogram recording was repeated at all visits if the pretherapy/entry QTc was ≥ 500 ms or if the subject was taking a concomitant medication that affected cardiac conduction. A repeat chest x-ray was obtained at the post-therapy/TOC visit.
Evaluation of clinical outcome was based on the investigator's assessment of pneumonia-related signs and symptoms and a chest x-ray at the post-therapy/TOC visit in comparison with the pretherapy/entry visit. Possible clinical outcomes were "cure" (returned to preinfection state or postinfectious stigmata indicative of a normal course of clearance of the infectious process and not requiring further antibiotic treatment), "failure" (residual symptoms or complications of CAP requiring further antibiotic treatment), and "indeterminate" (lost to follow-up or discontinuation not related to study drug).
Safety evaluation
Adverse events (AEs) were spontaneously reported by the patients and/or observed by the investigator from the time of entry to 14 days after the final dose of telithromycin. These events included any sign, symptom, syndrome, or illness that appeared or worsened and that might impair patient well-being. Investigators assessed causality of AEs as being either possibly related or not related to the study drug. Laboratory safety data analyses (i.e. hematology, blood chemistry, urinalysis) were performed according to standard laboratory procedures by Covance Central Laboratory Services, Indianapolis, Indiana.
Statistical analysis
Baseline and outcome data for continuous variables were calculated using summary statistics, and for categoric variables using frequency tabulation and calculated percentages. Two-sided 95% confidence intervals (CIs) were calculated for both bacteriologic outcomes and clinical cure rates. AEs were tabulated, and vital signs and QTc intervals were calculated with summary statistics. Post-therapy/TOC clinical laboratory values were compared with baseline using descriptive statistics, and potentially important laboratory values were compared with predefined levels outside the extended normal range for the respective parameter.
Results
A total of 432 in-patients and outpatients ≥ 13 years of age were enrolled at entry and received open-label treatment with telithromycin; 14 were excluded from the mITT population due to chest x-ray findings that were negative for pneumonia, resulting in an mITT population of 418 patients. Demographic and clinical characteristics of the mITT population are summarized in Table 1. Bacteria other than S pneumoniae identified in bacteremic patients included Streptococcus viridans and Staphylococcus haemolyticus. During the 7-day treatment period, 395(94.5%) patients in the mITT population were 100% compliant.
Table 1 Demographic and clinical characteristics for the modified intent to treat population.
Characteristic n (%)
Total treated 418
Gender 241 (57.7)
Male 177 (42.3)
Female
Age, years 45.0 [13–92]
Median range
< 65 362 (86.6)
≥ 65 56 (13.4)
BMI 417
Mean ± SD 26.9 ± 6.9
Smoking status
Smoker 160 (38.3)
Ex-smoker 79 (18.9)
Nonsmoker 179 (42.8)
Chest x-ray findings
Unilateral 304 (73.6)
Bilateral 109 (26.4)
Single lobe* 351 (85.0)
Multiple lobe 54 (13.1)
Pleural effusion 15 (3.6)
Bacteremia 15 (3.6)
S pneumoniae bacteremia 9 (2.2)
Investigator assessment of current episode
Mild 93 (22.2)
Moderate 279 (66.7)
Severe 46 (11.0)
PSI
Class I 222 (53.1)
Class II 144 (34.4)
Class III 39 (9.3)
Class IV 12 (2.9)
Class V 1 (0.2)
BMI = body mass index; PSI = pneumonia severity index; SD = standard deviation.
*"Single lobe" could apply to either unilateral or bilateral lung involvement (i.e. 1 lobe per right and/or left lung).
Causative pathogens identified at pretherapy/entry in the bmITT (n = 255) and PPb populations (n = 149) are shown in Table 2. Of these, 187/337 (55.5%) isolates in the bmITT population and 119/204 (58.3%) isolates in the bacteriologically evaluable group represented three of the bacteria most commonly implicated in CAP: S pneumoniae, H influenzae, and M catarrhalis. Others included Klebsiella pneumoniae and Staphylococcus aureus. A typical/intracellular pathogens were detected in a small number of patients: C pneumoniae, M pneumoniae, and L pneumophila in 3, 2, and 4 patients, respectively.
Table 2 Causative pathogens identified at pretherapy/entry in the per-protocol bacteriologically evaluable (PPb) and bacteriologic modified intent to treat (bmITT) populations.
Pathogen Number of Isolates in PPb Patients Number of Isolates in bmITT Patients
Total 204 337
S pneumoniae 57 76
H influenzae 49 88
H parainfluenzae 46 81
M catarrhalis 13 23
K pneumoniae 3 9
S aureus 15 23
Other* 21 37
*Includes: Haemophilus haemolyticus (7); Haemophilus parahaemolyticus (7); Enterobacter cloacae (4); Pseudomonas aeruginosa (4); Acinetobacter baumannii (1 bmITT only); Citrobacter freundii (1 bmITT only); Enterobacter aerogenes (1); Enterobacter agglomerans (1 bmITT only); Enterobacter amnigenus (1 bmITT only); Enterobacter not otherwise specified (NOS) (1 bmITT only); Neisseria meningitides (1); Proteus mirabilis (1); Proteus NOS (1 bmITT only); Streptococcus, viridansgroup(5); Staphylococcus haemolyticus (1).
Bacteriologic outcome
Satisfactory bacteriologic outcome was achieved in 137/149 (91.9%, 95% CI: 87.5–96.3) patients in the bacteriologically evaluable subgroup and in 215/255 (84.3%, 95% CI: 79.8–88.8) patients in the bmITT subgroup (Table 3). Results in the bmITT population (a secondary efficacy variable) thus support the positive outcome achieved in the bacteriologically evaluable patients. Of the 12 subjects with an unsatisfactory bacteriologic outcome, 6 had pathogens considered persistent (H influenzae, S aureus, and Pseudomonas aeruginosa) and 6 had pathogens classified as presumed persistent (S aureus, Enterobacter cloacae, and Enterobacter aerogenes). Clinical cure was noted in 2 patients with persistent H influenzae and 2 with persistent P aeruginosa. Eight patients with persistent and presumed persistent bacteria were considered clinical failures.
Table 3 Bacteriologic outcomes at the post-therapy/test of cure evaluation for the bacteriologically evaluable and bacteriologic modified intent to treat (bmITT) populations.
Bacteriologically Evaluable Population bmITT Population
Assessment n (%) 95% CI* n (%) 95% CI*
n 149 255
Satisfactory† 137 (91.9) 87.5; 96.3 215 (84.3) 79.8; 88.8
Unsatisfactory‡ 12 (8.1) 40 (15.7)
Indeterminate - 16 (6.3)
*Two-sided 95% confidence interval (CI).
†Includes eradication, presumed eradication, and colonization.
‡Includes reinfection, superinfection, recurrence, presumed persistence, and persistence.
Eradication rates, including documented and presumed eradication, at post-therapy/TOC for the bacteriologically evaluable and bmITT populations are depicted in Figure 1. Of 204 pathogens identified in bacteriologically evaluable patients at entry or at subsequent study visits, 189 (92.6%) were eradicated. The eradication rate in the bmITT population was 284/315(90.2%). It is notable that all 57 (100%) of the S pneumoniae isolates in the bacteriologically evaluable patients and 70/71 (98.6%) in the bmITT population were eradicated by telithromycin. Eradication rates for H influenzae and M catarrhalis were also high (95.9% and 100%, respectively) in telithromycin bacteriologically evaluable patients.
Figure 1 Bacteriologic eradication rates by causative pathogen in the bacteriologically evaluable (PPb) and the bacteriologic modified intent to treat (bmITT) populations at the post-therapy/test of cure (TOC) visit.
In vitro susceptibility testing of causative pathogens revealed that 70/71 S pneumoniae isolates were susceptible to telithromycin at study entry in the bmITT population, including those exhibiting resistance to other selected antibiotics. One isolate had intermediate susceptibility to telithromycin (MIC 2.00 μg/mL). All M catarrhalis and S aureus strains isolated in the bacteriologically evaluable and bmITT populations, and 91.8% and 90.9% of H influenzae strains isolated in the bacteriologically evaluable and bmITT populations, respectively, were also susceptible to telithromycin (MIC ≤ 1 μg/mL). It is notable that a discrepancy was observed between bacteriologic eradication rates and susceptibility to telithromycin for S aureus and H influenzae: eradication rates were 80% and 96%, respectively, and telithromycin susceptibility was 100% and 90%, respectively, for the two pathogens.
Nine subjects in the per-protocol bacteriologically evaluable population exhibited S pneumoniae strains that were resistant to penicillin or erythromycin. All of these patients (including 2 with bacteremia) had bacteriologic outcomes of eradication or presumed eradication and clinical outcomes of cure (Table 4).
Table 4 Bacteriologic and clinical outcome in per-protocol bacteriologically evaluable patients with S pneumoniae isolates resistant to penicillin G (Pen G) and/or erythromycin A (Ery A).
MIC Susceptibility* (μg/mL)
Subject No. Isolate (Source) TEL Pen G Ery A Bacteriologic Outcome† Clinical Outcome
Subjects with single-pathogen S pneumoniae isolate
1 S pneumoniae (sputum) 0.030 (S) 0.030 (S) b>8.000 (R) Eradication Cure
2 S pneumoniae (sputum) 0.030 (S) 2.000 (R) 0.250 (S) Presumed eradication Cure
3 S pneumoniae (sputum) 0.500 (S) 2.000 (R) 8.000 (R) Presumed eradication Cure
Subjects with multiple-pathogen S pneumoniae isolate
4 S pneumoniae (sputum) 0.060 (S) 2.000 (R) 4.000 (R) Presumed eradication Cure
M catarrhalis 0.120 (S) ND 0.120 Presumed eradication Cure
5 S pneumoniae (sputum) 0.250 (S) 2.000 (R) 8.000 (R) Presumed eradication Cure
6 S pneumoniae (sputum) 0.030 (S) 2.000 (R) 0.060 (S) Presumed eradication Cure
H influenzae 0.002 (S) ND 1.000 Presumed eradication Cure
M catarrhalis 0.120 (S) ND 0.250 Presumed eradication Cure
7 S pneumoniae (sputum) 0.500 (S) 2.000 (R) 8.000 (R) Presumed eradication Cure
H influenzae 2.000 (S) ND 4.000 Presumed eradication Cure
8 S pneumoniae (blood) 1.000 (S) 0.250 (I) 8.000 (R) Presumed eradication Cure
H influenzae 1.000 (S) ND 4.000 Presumed eradication Cure
9 S pneumoniae (blood) 1.000 (S) 0.500 (I) 8.000 (R) Presumed eradication Cure
H influenzae 8.000 (R) ND 4.000 Presumed eradication Cure
S aureus 0.120 (S) ND 0.500 Presumed eradication Cure
MIC = minimum inhibitory concentration; ND = no data; TEL = telithromycin.
*Susceptibility: S = sensitive; I = intermediate; R = resistant.
†Bacteriologic outcome.
In the bmITT population subgroup, penicillin resistance (MIC ≥ 2.0 μg/mL), penicillin intermediate resistance (MIC = 0.12–1.0 μg/mL), and erythromycin (macrolide) resistance (MIC ≥ 1.0 μg/mL) were noted in 8/71 (11.3%), 4/71 (5.6%), and 9/71 (12.7%) S pneumoniae strains, respectively. Of those strains showing resistance to penicillin, 5/8 (62.5%) were also resistant to the macrolides and all 8 (100%) were resistant to trimethoprim/sulfamethoxazole and cefuroxime axetil. Seven of the macrolide-resistant strains (77.8%) were also resistant to trimethoprim/sulfamethoxazole, and 5 (55.6%) were resistant to cefuroxime axetil based on approved National Committee for Clinical Laboratory Standards susceptibility testing breakpoints.
Clinical outcome
Clinical cure with telithromycin in the PPc population at the post-therapy/TOC visit was achieved in 332/357 (93.0%) patients overall (Table 5). Clinical cure in the mITT population at the post-therapy/TOC visit was 357/418 (85.4%), reinforcing the clinical outcome achieved by telithromycin in the clinically evaluable patients.
Table 5 Clinical outcomes for patients according to demographic characteristics of interest (clinically evaluable population at the post-therapy/test of cure visit)
Clinical Outcome
Subgroup n Clinical cure, n (%)
All patients 357 332 (93.0)
Bacteremia 14 14 (100)
S pneumoniae bacteremia 9 9 (100)
Age ≥ 65 years 47 45 (95.7)
Chest x-ray findings
Unilateral 256 236 (92.2)
Bilateral 97 92 (94.8)
Single lobe* 303 282 (93.1)
Multiple lobes 42 38 (90.5)
Fever >38°C (oral or equivalent) 123 117 (95.1)
PSI
Class I 187 176 (94.1)
Class II 126 115 (91.3)
Class III 35 32 (91.4)
Class IV 9 9 (100)
PSI = pneumonia severity index.
*"Single lobe" could apply to either unilateral or bilateral lung involvement (i.e. 1 lobe per right and/or left lung).
No patient with a satisfactory bacteriologic outcome had a clinical failure. Clinical cure was achieved for 100% of patients infected with S pneumoniae, H influenzae, and M catarrhalis. Telithromycin also had excellent activity in clinically evaluable patients infected with atypical/intracellular pathogens. Clinical cure was achieved in 100% of patients infected with C pneumoniae (n = 3), M pneumoniae (n = 2), and L pneumophila (n = 4).
Telithromycin was highly effective in patients with risk factors for increased morbidity. In clinically evaluable patients with bacteremia (n = 14), in addition to those with S pneumoniae bacteremia (n = 9), telithromycin treatment yielded a clinical cure rate of 100%. In clinically evaluable patients ≥ 65 years of age (n = 47), the clinical outcome was 95.7% (45/47 cured). Patients with multiple lobe involvement had a clinical cure rate of 90.5% (38/42 cured) and those with pneumonia severity index (PSI) scores of III (n = 35) and IV (n = 9) had clinical cure rates of 91.4% and 100%, respectively.
Safety evaluation
Of 430 patients who received ≥ 1 dose of study medication and who were evaluated for safety, 154 (35.8%) experienced ≥ 1 treatment-emergent adverse event (TEAE) and 87 (20.2%) experienced AEs considered by the investigator to be treatment-related. (All AEs classified as treatment-emergent and possibly treatment-emergent were considered to be treatment-emergent.) Most TEAEs, regardless of causality, were considered by the investigators to be mild or moderate in severity. The two most commonly reported TEAEs were diarrhea in 35 patients (8.1%) and nausea in 25 patients (5.8%).
One or more serious AEs occurred in 12 patients (2.8%) and were considered unrelated to the study drug by the investigators. Clinically noteworthy abnormal laboratory values (CNALVs) occurred in 102 of the 430 patients in the safety evaluation. The most common CNALVs were decreased creatinine clearance reported in 63 patients (14.7%), increased aspartate aminotransferase (AST) concentration in 16 patients (3.7%), increased alanine aminotransferase (ALT) concentration in 15 patients (3.5%), increased alkaline phosphatase concentration in 12 patients (2.8%), and increased potassium concentration in 12 patients (2.8%). Most of these CNALVs were present at study entry (61.9%). Furthermore, 55% of all CNALVs did not deteriorate further during the treatment period, or laboratory values returned to within the normal range at the final post-therapy laboratory assessment. One case of decreased creatinine clearance resulting in renal insufficiency noted on Day 1 was considered moderate in intensity by the investigator and not related to the study drug. There were no cases of clinical hepatitis reported during the study.
AEs resulting in discontinuation of telithromycin occurred in 11 patients (2.6%). Six patients who discontinued treatment experienced allergic reaction, abdominal pain, diarrhea, vomiting, vertigo, or increased ALT concentration, which were considered possibly related to telithromycin. The remaining events leading to discontinuation (considered unrelated to telithromycin) were confusion, apnea, lung disorder, respiratory disorder, and myocardial infarction. Two deaths occurred during the study (1 resulting from myocardial infarction and 1 due to renal insufficiency and acute aspiration). Neither death was considered related to telithromycin. There were no clinically significant changes in vital signs from baseline mean or median values. Two patients experienced increases of ≥ 60 ms in QTc intervals, but these were not associated with any AEs and were not considered clinically significant.
Discussion
Results of this study demonstrate that 7-day treatment with telithromycin 800 mg once-daily telithromycin, the first ketolide antibacterial to undergo clinical development, is safe and effective in the treatment of CAP. Telithromycin was well tolerated and effective in > 90% of clinically and bacteriologically evaluable patients treated according to the study protocol. The favorable results for telithromycin in the present study are consistent with those reported previously in comparative trials. The safety and efficacy of telithromycin 800 mg once daily for varying treatment durations have been compared with those of clarithromycin, amoxicillin, and trovafloxacin. Respective clinical cure rates for 5-day and 7-day telithromycin versus 10-day clarithromycin were 89.3% and 88.8% versus 91.8%; for 10-day telithromycin versus 10-day amoxicillin, 94.6% versus 90.1%; and for 7- to 10-day telithromycin versus 7- to 10-day trovafloxacin, 90.0% versus 94.2% [23-25]. While the optimal treatment duration for CAP has not yet been established, 7-day telithromycin treatment appears to be effective. This shorter, once-daily regimen may encourage patients to complete their prescribed course of medication.
An important consideration in antibacterial treatment of patients with CAP is efficacy in those at increased risk of morbidity. Of particular concern is the possibility of bacteremic infection when it may not be suspected (a situation in which treatment with standard antibiotics could fail). Musher et al [26] have demonstrated that similar proportions of bacteremic cases occur in patients with pneumococcal CAP regardless of disease severity (i.e. among those with PSI scores ranging from I to IV). Telithromycin was highly effective in clinically evaluable patients with bacteremia, including S pneumoniae bacteremia. In addition, the clinical outcome in this study in patients ≥ 65 years of age was 95.7% (45/47 cured); in those with PSI scores of III and IV, telithromycin resulted in clinical cure rates of 91.4% and 100%, respectively. In other comparative studies, telithromycin showed comparable or slightly superior efficacy in high-risk patients versus clarithromycin and amoxicillin [23-25].
Telithromycin demonstrated bacteriologic eradication rates of 100% for S pneumoniae (57/57 isolates) in the bacteriologically evaluable patient population, including all 9 S pneumoniae strains demonstrating resistance to penicillin and/or erythromycin. Telithromycin was also 100% effective against M catarrhalis (13/13 isolates eradicated) and 96% effective against H influenzae (47/49 isolates eradicated). Susceptibility testing performed in this study revealed that penicillin- and macrolide (erythromycin)-resistant S pneumoniae strains showed resistance to clarithromycin, trimethoprim/sulfamethoxazole, and cefuroxime axetil: all 8 strains of penicillin-resistant S pneumoniae were resistant to both trimethoprim/sulfamethoxazole and cefuroxime axetil, and 5 (62.5%) were resistant to clarithromycin. Seven of the erythromycin-resistant strains (77.8%) were also resistant to trimethoprim/sulfamethoxazole, and 5 (55.6%) were also resistant to cefuroxime axetil.
The number of atypical/intracellular pathogens isolated from CAP patients in this study was small – in part due to the study design, which did not obtain convalescent serum in patients with initial cultures positive for typical pathogens and excluded patients with cultures positive for typical pathogens from the final analysis of atypical/intracellular pathogens. All cases (3 patients with C pneumoniae, 2 with M pneumoniae, and 4 with L pneumophila) were clinically cured.
The safety profile for telithromycin in this study was similar to that reported in a review of previous clinical trials of this ketolide [7]. Telithromycin was very well tolerated, with the most commonly reported AEs (generally involving the gastrointestinal system [diarrhea and nausea]) being mild to moderate. Study discontinuations occurred in only 6/430 (1.4%) patients due to AEs considered by the investigators to be possibly related to the study drug.
In summary, 7 days of once-daily treatment with telithromycin 800 mg orally was effective and well tolerated in adolescent and adult patients with CAP, including those infected with resistant strains of S pneumoniae. Telithromycin may be considered a convenient, well tolerated, and effective treatment option for CAP in this era of increasing antibacterial resistance.
Competing interests
BPL is an employee of sanofi-aventis and holds stocks in sanofi-aventis. CMF has received research funding for clinical trials from Aventis Pharmaceuticals, a member of the sanofi-aventis Group. All other authors declare that they have no competing interests.
Authors' contributions
CMF, TCP, and LMD were investigators on the study. BPL participated in the design of the study. All authors read and approved the final manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
This study was sponsored by Aventis Pharmaceuticals, a member of the sanofi-aventis Group
==== Refs
Bartlett JG Mundy LM Community-acquired pneumonia N Engl J Med 1995 333 1618 1624 7477199 10.1056/NEJM199512143332408
Bartlett JG Dowell SF Mandell LA Guidelines from the Infectious Diseases Society of America: practice guidelines for the management of community-acquired pneumonia in adults Clin Infect Dis 2000 31 347 382 10987697 10.1086/313954
Chenoweth CE Saint S Martinez F Lynch JP IIIFendrick AM Antibacterial resistance in Streptococcus pneumoniae : implications for patients with community-acquired pneumonia Mayo Clin Proc 2000 75 1161 1168 11075746
Felmingham D Grüneberg RN the Alexander Project Group The Alexander Project 1996–1997: latest susceptibility data from this international study of bacterial pathogens from community-acquired lower respiratory tract infections J Antimicrob Chemother 2000 45 191 203 10660501 10.1093/jac/45.2.191
American Thoracic Society Guidelines for the management of adults with community-acquired pneumonia Am J Respir Crit Care Med 2001 163 1730 1754 11401897
CDC fact sheet Quinolones and the clinical laboratory (sic) Centers for Disease Control and Prevention website (accessed December 12, 2001)
Balfour JAB Figgitt DP Telithromycin Drugs 2001 61 815 829 11398913
Roblin PM Hammerschlag MR In vitro activity of a new ketolide antibiotic, HMR 3647, against Chlamydia pneumoniae Antimicrob Agents Chemother 1998 42 1515 1516 9624507
Baltch AL Smith RP Ritz W Franke MA Michelsen PB Antibacterial effect of telithromycin (HMR3647) and comparative antibiotics against Legionella pneumophila J Antimicrob Chemother 2000 46 51 55 10882688 10.1093/jac/46.1.51
Edelstein PH Edelstein MA In vitro activity of the ketolide HMR 3647 (RU 6647) for Legionella spp., its pharmacokinetics in guinea pigs, and use of the drug to treat guinea pigs with Legionella pneumophila pneumonia Antimicrob Agents Chemother 1999 43 90 95 9869571
Bébéar CM Renaudin H Bryskier A Bébéar C Comparative activities of telithromycin (HMR 3647), levofloxacin, and other antimicrobial agents against human mycoplasmas Antimicrob Agents Chemother 2000 44 1980 1982 10858366 10.1128/AAC.44.7.1980-1982.2000
Gustafsson I Hjelm E Cars O In vitro pharmacodynamics of the new ketolides HMR 3004 and HMR 3647 (telithromycin) against Chlamydia pneumoniae Antimicrob Agents Chemother 2000 44 1846 1849 10858341 10.1128/AAC.44.7.1846-1849.2000
Yamaguchi T Hirakata Y Izumikawa K In vitro activity of telithromycin (HMR 3647), a new ketolide, against clinical isolates of Mycoplasma pneumoniae in Japan Antimicrob Agents Chemother 2000 44 1381 1382 10770785 10.1128/AAC.44.5.1381-1382.2000
Okamoto H Miyazaki S Tateda K Ishii Y Yamaguchi K Comparative in vitro activity of telithromycin (HMR 3647), three macrolides, amoxicillin, cefdinir and levofloxacin against Gram-positive clinical isolates in Japan J Antimicrob Chemother 2000 46 797 802 11062201 10.1093/jac/46.5.797
Schülin T Wennersten CB Ferraro MJ Moellering RC JrEliopoulos GM Susceptibilities of Legionella spp. to newer antimicrobials in vitro Antimicrob Agents Chemother 1998 42 1520 1523 9624509
Jalava J Kataja J Seppälä H Huovinen P In vitro activities of the novel ketolide telithromycin (HMR 3647) against erythromycin-resistant Streptococcus species Antimicrob Agents Chemother 2001 45 789 793 11181362 10.1128/AAC.45.3.789-793.2001
Pankuch GA Hoellman DB Lin G Bajaksouzian S Jacobs MR Appelbaum PC Activity of HMR 3647 compared to those of five agents against Haemophilus influenzae and Moraxella catarrhalis by MIC determination and time-kill assay Antimicrob Agents Chemother 1998 42 3032 3034 [Abstract] 9797250
Davies TA Dewasse BE Jacobs MR Appelbaum PC In vitro development of resistance to telithromycin (HMR 3647), four macrolides, clindamycin, and pristinamycin in Streptococcus pneumoniae Antimicrob Agents Chemother 2000 44 414 417 10639373 10.1128/AAC.44.2.414-417.2000
Bonnefoy A Girard AM Agouridas C Chantot JF Ketolides lack inducibility properties of MLS(B) resistance phenotype J Antimicrob Chemother 1997 40 85 90 9249208 10.1093/jac/40.1.85
Andrews JM Weller TMA Ashby JP Walker RM Wise R The in vitro activity of ABT773, a new ketolide antimicrobial agent J Antimicrob Chemother 2000 46 1017 1022 11102425 10.1093/jac/46.6.1017
Muller-Serieys C Soler P Cantalloube C Lemaitre F Gia HP Brunner F Andremont A Bronchopulmonary disposition of the ketolide telithromycin (HMR 3647) Antimicrob Agents Chemother 2001 45 3104 3108 11600363 10.1128/AAC.45.11.3104-3108.2001
Khair OA Andrews JM Honeybourne D Jevons G Vacheron F Wise R Lung concentrations of telithromycin after oral dosing J Antimicrob Chemother 2001 47 837 840 11389116 10.1093/jac/47.6.837
Pullman J Champlin J Vrooman PS Efficacy and tolerability of once-daily oral therapy with telithromycin compared with trovafloxacin for the treatment of community-acquired pneumonia in adults Int J Clin Pract 2003 57 377 384 12846341
Hagberg L Torres A Van Rensburg DJ Leroy B Rangaraju M Ruuth E Efficacy and tolerability of once-daily telithromycin compared with high-dose amoxicillin for treatment of community-acquired pneumonia Infection 2002 30 378 386 12478329 10.1007/s15010-002-2096-z
Tellier G Niederman MS Nusrat R Clinical and bacteriological efficacy and safety of 5 and 7 day regimens of telithromycin once daily compared with a 10 day regimen of clarithromycin twice daily in patients with mild to moderate community-acquired pneumonia J Antimicrob Chemother 2004 54 515 523 15269191 10.1093/jac/dkh356
Musher DM Alexandraki I Graviss EA Bacteremic and nonbacteremic pneumococcal pneumonia: A prospective study Medicine 2000 79 210 221 10941350 10.1097/00005792-200007000-00002
| 15927060 | PMC1177948 | CC BY | 2021-01-04 16:28:16 | no | BMC Infect Dis. 2005 May 31; 5:43 | utf-8 | BMC Infect Dis | 2,005 | 10.1186/1471-2334-5-43 | oa_comm |
==== Front
BMC Infect DisBMC Infectious Diseases1471-2334BioMed Central London 1471-2334-5-461595525010.1186/1471-2334-5-46Technical AdvanceEvaluation of amplified rDNA restriction analysis (ARDRA) for the identification of Mycoplasma species Stakenborg Tim [email protected] Jo [email protected] Patrick [email protected] Dominiek [email protected] Baere Thierry [email protected] Rita [email protected] Johan [email protected] Kruif Aart [email protected] Freddy [email protected] Mario [email protected] Veterinary and Agrochemical Research Centre, Groeselenberg 99, 1180 Brussels, Belgium2 Faculty of Veterinary Medicine, Ghent University, Salisburylaan 133, 9820 Merelbeke, Belgium3 Department of Clinical Chemistry, Microbiology & Immunology, Ghent University Hospital, De Pintelaan 185, 9000 Ghent, Belgium2005 14 6 2005 5 46 46 23 2 2005 14 6 2005 Copyright © 2005 Stakenborg et al; licensee BioMed Central Ltd.2005Stakenborg 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
Mycoplasmas are present worldwide in a large number of animal hosts. Due to their small genome and parasitic lifestyle, Mycoplasma spp. require complex isolation media. Nevertheless, already over 100 different species have been identified and characterized and their number increases as more hosts are sampled. We studied the applicability of amplified rDNA restriction analysis (ARDRA) for the identification of all 116 acknowledged Mycoplasma species and subspecies.
Methods
Based upon available 16S rDNA sequences, we calculated and compared theoretical ARDRA profiles. To check the validity of these theoretically calculated profiles, we performed ARDRA on 60 strains of 27 different species and subspecies of the genus Mycoplasma.
Results
In silico digestion with the restriction endonuclease AluI (AG^CT) was found to be most discriminative and generated from 3 to 13 fragments depending on the Mycoplasma species. Although 73 Mycoplasma species could be differentiated using AluI, other species gave undistinguishable patterns. For these, an additional restriction digestion, typically with BfaI (C^TAG) or HpyF10VI (GCNNNNN^NNGC), was needed for a final identification. All in vitro obtained restriction profiles were in accordance with the calculated fragments based on only one 16S rDNA sequence, except for two isolates of M. columbinum and two isolates of the M. mycoides cluster, for which correct ARDRA profiles were only obtained if the sequences of both rrn operons were taken into account.
Conclusion
Theoretically, restriction digestion of the amplified rDNA was found to enable differentiation of all described Mycoplasma species and this could be confirmed by application of ARDRA on a total of 27 species and subspecies.
==== Body
Background
Mycoplasmas are phylogenetically related to gram-positive bacteria with low GC-content and belong to the class of the Mollicutes. They form a unique group of bacteria that lack a cell-wall and that contain sterols in their cytoplasmatic membrane. They are of great importance, since several species are pathogenic to animals or humans, whereas species of other mollicute genera also infect plants and insects [1]. In addition, a series of mycoplasmas cause trouble in the laboratory, because they infect cell cultures. Already over 100 species have been described, and their number, as well as the number of different hosts is still increasing.
A correct identification of mycoplasmas, mostly performed after a fastidious initial isolation, may be achieved by various methods. Original tools to identify mycoplasmas were mainly based on biochemical and serological differentiation, varying from simple precipitation tests [2], to ELISA [3,4], immunofluorescence [5], or Western blot analysis [6]. These techniques are being replaced by faster DNA-based tools [7]. Many of these methods are based on the 16S rDNA sequence for various reasons. First, the 16S rDNA has been sequenced for all recognized Mycoplasma spp. and is required when describing a new species [8]. Secondly, the 16S rDNA sequences have lower intraspecific variability than most protein encoding genes, hence their use in the construction of phylogenetic topologies [9]. Recently, denaturing gradient gel electrophoresis of amplified 16S rDNA was shown to be useful to differentiate most Mycoplasma spp. [10]. In another approach, correct identification of related Mycoplasma spp. was based on differences of the 16S-23S intergenic spacer (ITS) region. Both size variation [11] as sequence differences [11,12] of the ITS were successfully used to differentiate related species. Compared to the 16S rDNA sequence, ITS sequences may vary more between strains of the same species due to a lower selection pressure [13], although reports of very highly conserved ITS regions are known as well [14].
Amplified rDNA restriction analysis (ARDRA) has already been used for the identification of some avian species [15-17] as well as for pathogenic mycoplasmas in cats [18]. Restriction analysis with PstI of an amplified 16S rDNA fragment was also shown useful to differentiate M. capricolum subsp. capripneumoniae from the other species belonging to the mycoides-cluster [19]. The potential and power of ARDRA to identify members of the Mollicutes was already put forward [20], but was never worked out in detail for a large number of species. In this study, we investigated the value of ARDRA to identify all (to date) recognized Mycoplasma spp.
Methods
Isolates
A total of 60 strains, belonging to 27 different Mycoplasma species and subspecies, were used during this study (Table 1). The Mycoplasma spp. belonging to the mycoides-cluster and the M. hyosynoviae strains, were kindly provided as purified genomic DNA samples by Dr. L. Manso-Silivan (CIRAD, France) and Dr. B. Kokotovic (DFVF, Denmark), respectively. All other Mycoplasma spp. were cultivated using F-medium [21], modified Hayflick medium [22], SP-4-medium [22], SP-4-medium supplemented with L-arginine, HS-medium [23], or Friis'-medium with ampicillin instead of methicillin [24].
Table 1 List of strains used in this study
Mycoplasma species Number of strains Strain designations
M. agalactiae 2 NCTC 10123 (PG2); 5725
M. arginini 1 884/200
M. bovigenitalium 1 MN120
M. bovirhinis 3 ATCC 27748; O475; CODA 8L
M. bovis 4 83/61; 295VD; Widanka309; O422
M. capricolum subsp. capricolum 1 ATCC 27343 (California Kid)
M. capricolum subsp. capripneumoniae 1 NCTC 10192 (F38)
M. columbinasale 1 397
M. columbinum 4 423VD; 446; 447; 448
M. columborale 1 Pul46
M. dispar 2 ATCC 27140; MdispA
M. flocculare 4 ATCC 27399 (Ms42); MP102; MflocF6A; MflocF316
M. gallinarum 3 MgalnA; D63P; MgalnB
M. gallisepticum 3 ATCC 19610; A5969; 2000Myc58
M. glycophilum 2 412VD; MglyF1A
M. hyopneumoniae 4 ATCC 25934 (J); MhF56C; MhF612D; MhF72C
M. hyorhinis 4 MhyorF6A; MhyorF9A; MhyorF7A; MhyorF1A
M. hyosynoviae 4 ATCC 25591 (S16); Mp6; Mp96; Mp178
M. lipofaciens 1 R171
M. mycoides subsp. capri 1 Pg3
M. mycoides subsp. mycoides LC 1 YG
M. mycoides subsp. mycoides SC 1 Pg1
M. neurolyticum 2 MneuF1A; WVU1853
M. orale 1 ATCC 23714
M. pneumoniae 3 0696A, 1285A, 1284A
M. putrefaciens 4 Put85; B387; B731; 7578.95
Mycoplasma sp. bovine group 7 1 Pg50
All isolates were previously identified using biochemical tests and growth precipitation tests with absorbed rabbit antisera [2]. Whenever discrepancies existed between the obtained ARDRA-profiles and the serological results, the 16S rDNA was sequenced for an exact identification [25].
DNA extraction
DNA of growing cultures was extracted using a phenol-chloroform extraction described previously [26] or using alkaline lysis. For alkaline lysis, the cultures were centrifuged (2', 10000 g) and resuspended in 50 μl lysis buffer (0.25% SDS in 0.05 N NaOH). After 5' at 95°C, 300 μl water was added and the bacterial debris was centrifuged (2', 10000 g). One μl of the supernatant was used as template for amplification of the 16S rDNA.
16S PCR amplification
The universal primers pA (5'AGAGTTTGATCCTGGCTCAG) and pH (5'AAGGAGGTGATCCAGCCGCA) were used to amplify the 16S rRNA genes [25], yielding an amplification product of approximately 1500 bp. Thirty cycles (20" 94°C; 15" 57°C; and 30' 72°C) were run on a GeneAmp 9600 Thermal Cycler (Perkin Elmer, USA) using 3 U recombinant Taq DNA polymerase (Invitrogen, UK), 1 × PCR buffer (20 mM Tris-HCl, 1.5 mM MgCl2, and 50 mM KCl; pH 8.4), 10 pmol of each primer and 1 μl of the genomic DNA (~30 ng) as template. Reaction volumes were 50 μl.
Restriction digestion
For all 60 strains, 10 μl of the 16S rDNA PCR product was digested with 5 U of restriction enzyme AluI (Fermentas, Lithuania; sequence: AG^CT) and the associated Y+/Tango restriction buffer (Fermentas) in a total volume of 20 μl for 2 hours at 37°C. For a final identification, the amplified 16S rDNA of some strains were digested in addition with BfaI (New England Biolabs, USA; sequence: C^TAG) or HpyF10VI (Fermentas; sequence: GCNNNNN^NNGC). The restriction fragments were separated on a 3% Nusieve 3:1 agar (Tebu-Bio, France) for 2 hours at 130 V and visualized using a GeneGenius gel documentation system (Westburg, The Netherlands). A 50-bp ladder was used as a DNA marker (Fermentas).
Sequences &in silico ARDRA-profiles
ARDRA-profiles were calculated for all Mycoplasma spp. as acknowledged by the International Committee on Systematics of Prokaryotes (ICPS) to date. The 16S rDNA sequences were downloaded from Genbank (accession numbers are indicated in Figure 1). A consensus sequence was constructed and used for species for which more than one sequence was available. The M. orale 16S rDNA sequence was determined and submitted [Genbank:AY796060], since the only available sequence contained numerous ambiguities. For the members of the M. mycoides-cluster – for which differences between rrnA and rrnB have been published [27] – both sequences were used. For some Mycoplasma spp. only a partial sequence of the 16S rDNA was available. For these sequences, nucleotides were added to the 5' and/or 3' ends to generate fragments of expected length. These lengths and the choice of the nucleotides added were based on a 16S rDNA consensus sequence obtained by alignment of the complete Mycoplasma 16S rDNA sequences available in Genbank using Clustal W. The restriction sites and the exact size of the ARDRA fragments were calculated using Vector NTI Advance V9.0 (Invitrogen) and BioNumerics V3.5 (Applied-Maths, Belgium).
Figure 1 Theoretical ARDRA patterns after in silico digestion with AluI for all currently recognized Mycoplasma spp. Patterns are clustered using UPGMA (Bionumerics V3.5) by way of illustration. The Genbank-accession numbers used are listed together with species name.
By way of illustration, a dendrogram, based on ARDRA patterns, was constructed using the Unweighted Pair Group Method with Arithmetic Means (UPGMA) using 1% tolerance (i.e. bands that differ about 7 nucleotides or less are considered identical) and taking only fragments from 80 to 800 nucleotides into account.
Results
For all Mycoplasma spp., the theoretical AluI,BfaI and HpyF10VI restriction patterns were calculated [see Additional file 1] and are represented in Figure 1, 2, 3. For a number of species, ARDRA was carried out in the laboratory to confirm the in silico obtained results and to check the validity of the technique for identification. ARDRA profiles obtained with AluI and BfaI are shown in Figure 4 and Figure 5, respectively. For a further verification of the technique and for the remaining 9 species that could not be identified with AluI or BfaI alone, ARDRA was also performed with HpyF10VI (Figure 6, 7).
Figure 2 Calculated ARDRA profiles of Mycoplasma spp. that can be differentiated using BfaI, but had undistinguishable AluI restriction profiles.
Figure 3 Calculated ARDRA profiles of Mycoplasma spp. that can be differentiated using HpyF10VI, but had undistinguishable AluI restriction profiles. The restriction pattern of M. capricolum subsp. capricolum represents the not included members of the M. mycoides-cluster as well.
Figure 4 ARDRA profiles after restriction with AluI of 18 different Mycoplasma species. Since all samples of the same species gave identical restriction patterns, the number of strains tested for each species is indicated in parenthesis. A Generuler 50-bp ladder (Fermentas) was used as size-marker.
Figure 5 ARDRA profiles after restriction with BfaI of 18 different Mycoplasma species. Since all samples of the same species gave identical restriction patterns, the number of strains tested for each species is indicated in parenthesis. A Generuler 50-bp ladder (Fermentas) was used as size-marker.
Figure 6 ARDRA profiles after restriction with AluI (left) or HpyF10VI (right) of M. bovigenitalium and of M. columbinum. A Generuler 50-bp ladder (Fermentas) was used as size-marker. The number of strains tested for each species is indicated in parenthesis.
Figure 7 ARDRA profiles of M. putrefaciens and the M. mycoides cluster after restriction with AluI (left) and HpyF10VI (right). The expected band sizes for both rrn operons are indicated in Additional file 1. An O'RangeRuler 50-bp ladder (Fermentas) was used as size-marker. The number of strains tested for each species is indicated in parenthesis.
Two of the four M. columbinum strains showed an unpredicted ARDRA pattern after restriction with AluI. Since the sum of all bands was higher than the length of the 16S sequence, a difference between the 2 rrn operons was expected. This was verified by sequence analysis, which revealed an ambiguity at position 997 (i.e. position 1007 in the E. coli numbering), pointing to the presence of AGCT in one and AGTT in the other operon. As such, a restriction site for AluI in one operon will lack in the other operon and will lead to a mixture of ARDRA profiles. Also for the strains of the M. mycoides-cluster the published sequences of both rrn operons were taken into account [27]. By superimposition of the restriction profiles of both rrnA and rrnB, the correct, expected profiles were obtained. However, a faint band of approximately 370 nucleotides was observed in the HpyF10VI restriction profile of M. capricolum subsp. capripneumoniae, indicating a partial restriction at position 1082 of the rrnA gene (Figure 7). For all other samples, profiles were identical to the calculated restriction profiles using only one consensus sequence of the Genbank entries.
A few species could not be differentiated with the three suggested enzymes and for these, other enzymes were selected. M. cricetuli and M. collis, which have 16S rRNA operons that are 99.8% identical, can be differentiated using Hpy188III. This enzyme cuts the 16S rDNA of M. collis 7 times, while restriction takes place only 6 times in the 16S rRNA gene of M. cricetuli. Also the restriction enzyme EarI can be used, since it only restricts the 16S rRNA gene of M. cricetuli. The very related M. imitans and M. gallisepticum could be differentiated using MseI or HindII. The restriction enzyme BstUI could be used to differentiate the otherwise indistinguishable M. haemocanis (2 restriction sites) and M. haemofelis (3 restriction sites). The determined 16S rDNA sequence of M. orale was almost identical to the 16S rDNA of M. indiense and specific restriction enzymes, like BsaJI or EcoHI, were necessary to differentiate these species. In case of the very related members of the mycoides-cluster, the differentiation is more complicated and a whole series of restrictions are needed. Based on the occurrence of different restriction sites, it is however theoretically possible to correctly identify these species as well, using only commercially available restriction endonucleases (Table 2).
Table 2 Number of restriction sites for the members of the M. mycoides-cluster
Mycoplasma species
Restriction endonuclease Mycoplasma sp. bovine group 7 M. mycoides ssp. mycoides LC M. mycoides ssp. capri M. mycoides ssp. mycoides SC M. capricolum ssp. capripneumoniae M. capricolum ssp. capricolum
BbvI 4 4 4 4 4/2 4
HpyCH4III 3 4 4 3 3 3
HpyF10VI 5 5 5 5 5/4 5
MaeIII 5 5 5 4 5 5
MboII 3/5a 3 3 3 3 3/4
Tsp509I 4 4 4 4/5 4 4
a Two values indicate differences between rrnA and rrnB, based on the Genbank accession numbers indicated in 1.
Discussion
Identification of mycoplasmas still largely relies on serological tests, but owing to the limited availability of quality-controlled sera, the high number of species, the serological cross-reaction between related species and the great variability in the surface antigens of different strains [28], newer techniques are needed. Sequence analysis of the 16S rRNA genes proved a useful tool to identify species, but the need for expensive equipment makes the technique less favorable for routine diagnosis. In this study, we showed that theoretically all Mycoplasma spp. are distinguishable using ARDRA. The in silico determined discriminative power was confirmed in the laboratory and even closely related Mycoplasma spp. could be identified correctly, as exemplified by the restriction with AluI and BfaI of M. agalactiae and M. bovis.
We used universal primers to amplify the entire 16S rDNA to obtain a maximum discriminatory power. Working with universal primers implies that interference from other bacteria is to be expected when starting from clinical samples [29], especially when mycoplasmas are not abundantly present. The use of mycoplasma-specific primers binding to internal regions of the 16S rRNA genes may be helpful and result in a higher specificity as was already proposed by others [20,30]. However, care must be taken since the discriminatory power will decrease if primers are chosen in such a way that less restriction sites are present in the amplification products. Alternatively, McAuliffe et al. [31] proposed a selective enrichment step for 24 hours in Eaton's-medium before amplification of 16S sequences to identify Mycoplasma spp. Also Kiss et al. [16] used ARDRA to identify three avian Mycoplasma species after 48 hours of incubation in Frey media. These suggested approaches may solve most problems, but may still be insufficient for mixed Mycoplasma cultures. The presence of more than one Mycoplasma species in clinical samples will lead to complex patterns, which are not easily resolved.
Differences between rrn operons have been reported in several bacterial classes, but the level of sequence heterogeneity was recently shown to be lower than expected [32]. It is therefore reasonable to assume that rrn operons tend to evolve in concert [33]. For some bacterial species a high level of 16S rDNA sequence heterogeneity has been described [34,35], while for Mycoplasma species, which possess no more than 2 rrn operons, only some micro-heterogeneity (i.e. scattered sequence variation between highly related rRNA genes) has been reported [27,36,37]. Besides, most differences between the two operons will not lead to altered restriction sites and will not influence the ARDRA patterns. In case a mutation is located within one of both restriction recognition sites, as was shown in particular for M. columbinum, restriction will most likely yield an unknown ARDRA profile, rather than lead to a false identification. Moreover, this aberrant pattern can be included in the identification scheme. The significance of the C1007T transition (E. coli numbering) present in two of the four M. columbinum strains is still unknown, but was shown in some strains of E. coli as well [33]. Also, in agreement with an earlier report [27], many differences between the rrnA and rrnB sequences were observed for members of the M. mycoides cluster. Nevertheless, the combined restriction profiles of both rrn sequences resulted in expected patterns with exception of a faint band seen for M. capricolum subsp. capripneumoniae after restriction with HpyF10VI. The reason for this partial restriction is unknown since purifying the PCR product, increasing the enzyme concentration, or lengthening the incubation period made no difference (data not shown). In any case, identification based on ARDRA was shown complex for these very related species and other techniques – like serological tests independent of the 16S rDNA sequences [8] – may be more suitable. However, the extra band visible for M. mycoides subsp. mycoides SC after restriction with AluI was shown sufficiently stable to be used for identification [38] and the value of ARDRA using PstI was also reported for M. capricolum subsp. capripneumoniae [36]. Although the 16S rDNA sequences of these species may be almost identical, ARDRA is able to emphasize the few differences present without the need of extensive 16S rDNA sequence analysis or other tests [19,38-40]. Also for other species with nearly identical 16S rDNA sequences (99.5% identity for M. haemocanis and M. haemofelis; 99.7% for M. gallisepticum and M. imitans; 98.9% for M. orale and M. indiense, and 99.8% for M. criteculi and M. collis), it was calculated that restriction analysis with a single additional enzyme would result in different restriction patterns and therefore to a correct identification.
Conclusion
Restriction digestion with AluI of the amplified 16S rDNA can be used to differentiate between 73 of the 116 described Mycoplasma species and subspecies. An additional restriction with BfaI or HpyF10VI enables the identification of another 31 species and subspecies. Also the remaining 12 species can be differentiated, with the use of additonal enzymes, although other techniques may be preferred for some members of the M. mycoides-cluster.
The simplicity and the general applicability of ARDRA make it possible to implement this technique in most laboratories with basic molecular biology equipment.
List of abbreviations
ARDRA amplified rDNA restriction analysis
ITS intergenic spacer(s)
CIRAD Agricultural Research Centre for International Development (Montpellier, France)
DFVF Danish Institute for Food and Veterinary Research (Copenhagen, Denmark)
UPGMA Unweighted Pair Group Method with Arithmatic Means
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
TS collected and analyzed most of the data and was principal writer of the manuscript. TDB participated in the initial in silico data analysis, while RV helped in the correct identification of the species. JV, PB, DM, JP, and AdK co-drafted the manuscript. FH participated in the discussion of the data, participated in proofreading and management. MV conceived the study and revised the manuscript critically. All authors made contributions, read and approved the final manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Supplementary Material
Additional File 1
Overview of the restriction fragments (and corresponding restriction sites) after ARDRA with AluI, BfaI and HpyF10VI for all current 116 Mycoplasma species and subspecies. The restriction enzymes needed to obtain a correct identification are marked in bold. The fragments are listed according to their size. Based upon available 16S rDNA sequences, ARDRA profiles were calculated for AluI, BfaI and HpyF10VI for all currently acknowledged Mycoplasma species. For these restriction enzymes, the file gives a detailed overview of the in silico determined restriction sites as well as the size of the restriction fragments.
Click here for file
Acknowledgements
This work was supported by a grant of the Federal Service of Public Health, Food Chain Safety and Environment (Grant number S-6136).
The authors thank Sara Tistaert for skilful technical assistance.
==== Refs
Saglio PHM Whitcomb RF Barile MF, Razin S, Whitcomb RF and Tully JG Diversity of wall-less prokaryotes in plant vascular tissue, fungi, and invertebrate animals The mycoplasmas: plant and insects mycoplasmas 1979 III London, Academic Press, Inc. 1 31
Erno H Peterslund K Razin S and Tully JG Growth precipitation test Methods in Mycoplasmology: Mycoplasma Characterization 1983 I , Academic Press 489 492
May JD Branton SL Identification of Mycoplasma isolates by ELISA Avian Dis 1997 41 93 96 9087324
Djordjevic SP Eamens GJ Romalis LF Saunders MM An improved enzyme linked immunosorbent assay (ELISA) for the detection of porcine serum antibodies against Mycoplasma hyopneumoniae Vet Microbiol 1994 39 261 273 8042274 10.1016/0378-1135(94)90163-5
Bencina D Bradbury JM Combination of immunofluorescence and immunoperoxidase techniques for serotyping mixtures of Mycoplasma species J Clin Microbiol 1992 30 407 410 1537910
Thomas CB Sharp P Detection of antigenic variation among strains of Mycoplasma gallisepticum by enzyme-linked immunosorbent inhibition assay (ELISIA) and Western blot analysis Avian Dis 1988 32 748 756 3060088
Rawadi G Dussurget O Genotypic methods for diagnosis of mycoplasmal infections in humans, animals, plants and cell cultures Biotechnol Genet Eng Rev 1998 15 51 78 9573605
Bradbury JM International committee on systematic bacteriology, subcommittee on the taxonomy of Mollicutes Int J Syst Evol Microbiol 2001 51 2227 2230
Woese CR Maniloff J Zablen LB Phylogenetic analysis of the mycoplasmas Proc Natl Acad Sci U S A 1980 77 494 498 6928642
McAuliffe L Ellis RJ Ayling RD Nicholas RA Differentiation of Mycoplasma species by 16S ribosomal DNA PCR and denaturing gradient gel electrophoresis fingerprinting J Clin Microbiol 2003 41 4844 4847 14532239 10.1128/JCM.41.10.4844-4847.2003
Harasawa R Mizusawa H Nozawa K Nakagawa T Asada K Kato I Detection and tentative identification of dominant Mycoplasma species in cell cultures by restriction analysis of the 16S-23S rRNA intergenic spacer regions Res Microbiol 1993 144 489 493 7910696 10.1016/0923-2508(93)90057-9
Harasawa R Kanamoto Y Differentiation of two biovars of Ureaplasma urealyticum based on the 16S-23S rRNA intergenic spacer region J Clin Microbiol 1999 37 4135 4138 10565945
Daffonchio D Cherif A Brusetti L Rizzi A Mora D Boudabous A Borin S Nature of polymorphisms in 16S-23S rRNA gene intergenic transcribed spacer fingerprinting of Bacillus and related genera Appl Environ Microbiol 2003 69 5128 5137 12957895 10.1128/AEM.69.9.5128-5137.2003
Chalker VJ Owen WM Paterson CJ Brownlie J Development of a polymerase chain reaction for the detection of Mycoplasma felis in domestic cats Vet Microbiol 2004 100 77 82 15135515 10.1016/j.vetmic.2004.01.014
Garcia M Jackwood MW Levisohn S Kleven SH Detection of Mycoplasma gallisepticum, M. synoviae, and M. iowae by multi-species polymerase chain reaction and restriction fragment length polymorphism Avian Dis 1995 39 606 616 8561747
Kiss I Matiz K Kaszanyitzky E Chavez Y Johansson KE Detection and identification of avian mycoplasmas by polymerase chain reaction and restriction fragment length polymorphism assay Vet Microbiol 1997 58 23 30 9451458 10.1016/S0378-1135(97)81568-4
Fan HH Kleven SH Jackwood MW Johansson KE Pettersson B Levisohn S Species identification of avian mycoplasmas by polymerase chain reaction and restriction fragment length polymorphism analysis Avian Dis 1995 39 398 407 7677664
Criado-Fornelio A Martinez-Marcos A Buling-Sarana A Barba-Carretero JC Presence of Mycoplasma haemofelis, Mycoplasma haemominutum and piroplasmids in cats from southern Europe: a molecular study Vet Microbiol 2003 93 307 317 12713893 10.1016/S0378-1135(03)00044-0
Bolske G Mattsson JG Bascunana CR Bergstrom K Wesonga H Johansson KE Diagnosis of contagious caprine pleuropneumonia by detection and identification of Mycoplasma capricolum subsp. capripneumoniae by PCR and restriction enzyme analysis J Clin Microbiol 1996 34 785 791 8815084
Deng S Hiruki C Robertson JA Stemke GW Detection by PCR and differentiation by restriction fragment length polymorphism of Acholeplasma, Spiroplasma, Mycoplasma, and Ureaplasma, based upon 16S rRNA genes PCR Methods Appl 1992 1 202 204 1282069
Bolske G Survey of Mycoplasma infections in cell cultures and a comparison of detection methods Zentralbl Bakteriol Mikrobiol Hyg [A] 1988 269 331 340 3146169
Razin S Tully JG Methods in Mycoplasmology Mycoplasma Characterization 1983 I , Academic Press 504
Friis NF Ahrens P Larsen H Mycoplasma hyosynoviae isolation from the upper respiratory tract and tonsils of pigs Acta Vet Scand 1991 32 425 429 1818502
Kobisch M Friis NF Swine mycoplasmoses Rev Sci Tech 1996 15 1569 1605 9190026
Edwards U Rogall T Blocker H Emde M Bottger EC Isolation and direct complete nucleotide determination of entire genes. Characterization of a gene coding for 16S ribosomal RNA Nucleic Acids Res 1989 17 7843 7853 2798131
Miles R Nicholas R Walker JM Mycoplasma protocols Methods in Molecular Biology 1998 104 Totowa, New Jersy, Humana Press 330
Pettersson B Leitner T Ronaghi M Bolske G Uhlen M Johansson KE Phylogeny of the Mycoplasma mycoides cluster as determined by sequence analysis of the 16S rRNA genes from the two rRNA operons J Bacteriol 1996 178 4131 4142 8763941
Rosengarten R Yogev D Variant colony surface antigenic phenotypes within mycoplasma strain populations: implications for species identification and strain standardization J Clin Microbiol 1996 34 149 158 8748292
Conrads G Flemmig TF Seyfarth I Lampert F Lutticken R Simultaneous detection of Bacteroides forsythus and Prevotella intermedia by 16S rRNA gene-directed multiplex PCR J Clin Microbiol 1999 37 1621 1624 10203541
Blanchard A Gautier M Mayau V Detection and identification of mycoplasmas by amplification of rDNA FEMS Microbiol Lett 1991 65 37 42 1874400 10.1016/0378-1097(91)90467-O
McAuliffe L Hatchell FM Ayling RD King AI Nicholas RA Detection of Mycoplasma ovipneumoniae in Pasteurella-vaccinated sheep flocks with respiratory disease in England Vet Rec 2003 153 687 688 14682543
Acinas SG Marcelino LA Klepac-Ceraj V Polz MF Divergence and redundancy of 16S rRNA sequences in genomes with multiple rrn operons J Bacteriol 2004 186 2629 2635 15090503 10.1128/JB.186.9.2629-2635.2004
Martinez-Murcia AJ Anton AI Rodriguez-Valera F Patterns of sequence variation in two regions of the 16S rRNA multigene family of Escherichia coli Int J Syst Bacteriol 1999 49 Pt 2 601 610 10319482
Marchandin H Teyssier C Simeon De Buochberg M Jean-Pierre H Carriere C Jumas-Bilak E Intra-chromosomal heterogeneity between the four 16S rRNA gene copies in the genus Veillonella: implications for phylogeny and taxonomy Microbiology 2003 149 1493 1501 12777489 10.1099/mic.0.26132-0
Mevarech M Hirsch-Twizer S Goldman S Yakobson E Eisenberg H Dennis PP Isolation and characterization of the rRNA gene clusters of Halobacterium marismortui J Bacteriol 1989 171 3479 3485 2722756
Bascunana CR Mattsson JG Bolske G Johansson KE Characterization of the 16S rRNA genes from Mycoplasma sp. strain F38 and development of an identification system based on PCR J Bacteriol 1994 176 2577 2586 8169205
Heldtander M Pettersson B Tully JG Johansson KE Sequences of the 16S rRNA genes and phylogeny of the goat mycoplasmas Mycoplasma adleri, Mycoplasma auris, Mycoplasma cottewii and Mycoplasma yeatsii Int J Syst Bacteriol 1998 48 Pt 1 263 268 9542096
Persson A Pettersson B Bolske G Johansson KE Diagnosis of contagious bovine pleuropneumonia by PCR-laser- induced fluorescence and PCR-restriction endonuclease analysis based on the 16S rRNA genes of Mycoplasma mycoides subsp. mycoides SC J Clin Microbiol 1999 37 3815 3821 10565890
Vilei EM Frey J Differential clustering of Mycoplasma mycoides subsp. mycoides SC strains by PCR-REA of the bgl locus Vet Microbiol 2004 100 283 288 15145506 10.1016/j.vetmic.2004.02.005
Rodriguez JL Ermel RW Kenny TP Brooks DL DaMassa AJ Polymerase chain reaction and restriction endonuclease digestion for selected members of the "Mycoplasma mycoides cluster" and Mycoplasma putrefaciens J Vet Diagn Invest 1997 9 186 190 9211239
| 15955250 | PMC1177949 | CC BY | 2021-01-04 16:28:15 | no | BMC Infect Dis. 2005 Jun 14; 5:46 | utf-8 | BMC Infect Dis | 2,005 | 10.1186/1471-2334-5-46 | oa_comm |
==== Front
BMC Med GenetBMC Medical Genetics1471-2350BioMed Central London 1471-2350-6-271598969410.1186/1471-2350-6-27Research ArticleSpinocerebellar ataxia type 17: Report of a family with reduced penetrance of an unstable Gln49 TBP allele, haplotype analysis supporting a founder effect for unstable alleles and comparative analysis of SCA17 genotypes Zühlke Christine [email protected] Andreas [email protected] Eberhard [email protected] Ulrich [email protected] Institute of Human Genetics, University of Lübeck, Ratzeburger Allee 160, 23538 Lübeck, Germany2 Institute of Human Genetics, Universitätsklinikum Eppendorf, Butenfeld 42, 22529 Hamburg, Germany2005 1 7 2005 6 27 27 10 2 2005 1 7 2005 Copyright © 2005 Zühlke 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
Spinocerebellar ataxia type 17 (SCA17), a neurodegenerative disorder in man, is caused by an expanded polymorphic polyglutamine-encoding trinucleotide repeat in the gene for TATA-box binding protein (TBP), a main transcription factor. Observed pathogenic expansions ranged from 43 – 63 glutamine (Gln) codons (Gln43–63). Reduced penetrance is known for Gln43–48 alleles. In the vast majority of families with SCA17 an expanded CAG repeat interrupted by a CAA CAG CAA element is inherited stably.
Results
Here, we report the first pedigree with a Gln49 allele that is a) not interrupted, b) unstable upon transmission, and c) associated with reduced penetrance or very late age of onset. The 76-year-old father of two SCA17 patients carries the Gln49 TBP allele but presents without obvious neurological symptoms. His children with Gln53 and Gln52 developed ataxia at the age of 41 and 50. Haplotype analysis of this and a second family both with uninterrupted expanded and unstable pathological SCA17 alleles revealed a common core genotype not present in the interrupted expansion of an unrelated SCA17 patient. Review of the literature did not present instability in SCA17 families with expanded alleles interrupted by the CAA CAG CAA element.
Conclusion
The presence of a Gln49 SCA17 allele in an asymptomatic 76-year-old male reams the discussion of reduced penetrance and genotypes producing very late disease onset. In SCA17, uninterrupted expanded alleles of TBP are associated with repeat instability and a common founder haplotype. This suggests for uninterrupted expanded alleles a mutation mechanism and some clinical genetic features distinct from those alleles interrupted by a CAA CAG CAA element.
==== Body
Background
The spinocerebellar ataxias (SCAs), a group of autosomal dominantly inherited human disorders with mainly adult age of onset, are caused by progressive neurodegeneration and significant cerebellar dysfunction, but also involve other regions of the central or peripheral nervous system. Clinically and even histopathologically the differentiation between SCA subtypes may be rather difficult. Genetically, for 25 types causative mutations in various genes are known or defined by linkage to distinct chromosomal regions. For seven SCAs expanded CAG trinucleotide repeats have been discovered [1]. This kind of mutation seems to be specific for the human species.
SCA17 [OMIM: 607136], a rare type of SCA with a variety of clinical features, is caused by an expanded CAG repeat in TBP, the gene of the TATA box-binding protein [TBP; OMIM: 600075]. TBP forms the DNA-binding subunit of the universally essential RNA polymerase II transcription factor D. TBP has been mapped close to the telomeric region at chromosome 6q27 [2]. An imperfect repetitive and polymorphic polyglutamine encoding CAG triplet sequence is part of the coding region of TBP [3]. Initially, 20 different alleles coding for 29 to 42 glutamine residues (Gln) have been identified [4] with the most common alleles containing 32 to 39 Gln codons. The polyglutamine-encoding DNA sequence of TBP wildtype alleles can be subdivided into several regions including two polymorphic (CAG)n stretches: (CAG)3 (CAA)3 (CAG)n CAA CAG CAA (CAG)n CAA CAG. In patients with SCA17, the polymorphic CAG sequence encoding the polyglutamine stretch of TBP is expanded heterozygous. Alleles with 43 – 48 Gln codons represent a zone of incomplete penetrance or very late age of onset [5,6].
Disease causing alleles with up to 63 [7] Gln codons were identified. Clinically, SCA17 may mimic a broad spectrum of neuropsychiatric diseases, including Parkinson and Huntington disease, major psychosis, and multitudinous cerebellar syndromes [8-10]. Neuropathological findings include cerebellar, cortical, and subcortical atrophy, Purkinje cell loss, gliosis [11], and neuronal intranuclear inclusions immunopositive for TBP and polyglutamine.
The clinical relevance of Gln43–48 encoding TBP alleles is not obvious. A Gln43 allele was considered responsible for clinical symptoms in a 64-year-old patient with ataxia and progressive mental deterioration [6]. Gln44 repeats were found in a patient with gait ataxia and behavioral changes and a disease onset at 29 [12] and another patient with early onset cerebellar ataxia (EOCA) and suspected Friedreich ataxia [5]. Regarding the EOCA phenotype in the latter patient, Gln44 was considered a large normal allele. On the other hand, alleles for Gln46 [12] and Gln48 [13] showed reduced penetrance in two families. Therefore, Gln43–48 encoding alleles could present intermediate alleles with incomplete penetrance.
In contrast to the majority of other polyglutamine diseases, the SCA17 repeat expansion shows meiotic stability upon transmission. To date, only two pedigrees are known in which instability of the repetitive sequence has been observed, irrespectively of the gender of the transmitting parent. In one case, the repeat increased by one triplet after maternal transmission [5], while in the other case an increase of 13 triplets and marked anticipation was associated with paternal inheritance [14].
Here, we describe a SCA17 family from northern Germany with a) repeat instability upon paternal transmission and b) reduced penetrance or very late onset in association with an uninterrupted CAG41 sequence (Gln49 allele). Meiotic instability is not common in SCA17 pedigrees. Therefore, we performed haplotype analysis in two unrelated families looking for a founder allele associated with repeat instability at the SCA17 locus. In addition, we included a patient homozygous for a SCA17 allele with a CAG CAA CAG interrupting element to reveal independent mutation events for the different repeat compositions.
Results
Phenotype and genotype
In family A, two of four offspring of unaffected parents (74- and 76-year-old) developed an adult onset ataxia. The 43-year-old index patient reported onset of gait disturbances accompanied by a slowly progressive dysarthria at the age of 41. Brain magnetic resonance image analysis performed approximately half a year after clinical onset revealed cortical cerebellar atrophy and a mildly reduced brain volume. His 56-year-old sister noticed symptoms of gait ataxia, dysarthria, and disturbed handwriting at the age of 50. The 76-year-old father of the siblings reported that a neurological examination had revealed no abnormality. Both parents presented without obvious neurological signs. In order to clarify the repeat status of TBP and the inheritance of the disease in the family, both agreed in molecular genetic analysis for SCA17 but rejected the offer of a detailed clinical neurological examination.
Molecular genetic analyses revealed pathogenic alleles for the TBP gene: 53 repeats for the male patient with age of onset at 41 and 52 repeats for his sister with symptoms starting at 50 years of age. Their clinically non-affected 76-year-old father carries an elongated allele of 49 repeats. Therefore, this allele displays reduced penetrance or very late onset, as well as instability upon transmission. Sequence analysis revealed loss of the CAA CAG CAA interruption in the expanded alleles.
Haplotype analysis
Recently, we described a family (B) with repeat instability and missing CAA triplets at the expanded alleles [5,10]. Genealogic data for both families (A,B), as far as available, revealed no relationship between the families. But Northern German ancestry of both families suggested the possibility of a common founder. This was assessed by haplotype analysis (figure 1) including the unrelated patient H carrying homozygous an expansion in the TBP gene [15]. In both families A and B, the same haplotype consisting of allele 5 of D6S446 and allele 2 of D6S1590 (5_2) was located on the chromosome bearing the expanded TBP allele co-segregating with SCA17, respectively. Patient H did not share this haplotype, she is homozygous for haplotype 2_4. Among the markers analyzed, D6S446 and D6S1590 were the ones most closely located to TBP. Therefore, the data suggest the possibility of a common SCA17 founder allele of the variably expanded TBP allele in families A and B with further instability upon transmission.
Allele frequency
To estimate the probability of a founder effect, the allele frequencies of D6S446 and D6S1590 were determined in 96 anonymous control subjects of the geographic region of families A and B. For D6S446 six alleles were found, for D6S1590 seven. Allele 5 of D6S446 is present in 43 out of 184 chromosomes (23%), allele 2 of D6S1590 in 92 (50%). The computed maximum likelihood frequencies (using Arlequin software version 2.001) of the three most common two-locus haplotypes of D6S446 and D6S1590 with frequencies > 0.1 are 0.219 (6_2), 0.163 (6_1), and 0.153 (5_2). Thus, the three most common haplotypes, including the haplotype 5_2 co-segregating with SCA17 in families A and B account for 53.5% of all haplotypes. The 17 remaining haplotypes predicted had computed frequencies from 0.3–6.8%. Maximum likelihood frequency of haplotype 4_2 (patient H) is 0.042. The likelihood ratio test suggested a highly significant linkage disequilibrium between D6S446 and D6S1590 (p < 5 ×10-6). Observed and maximum likelihood counts of genotypes and haplotypes did not deviate significantly from Hardy Weinberg equilibrium. Therefore, the likelihood of two unrelated subjects sharing haplotype 5_2 is estimated to be ~8 %. The combined likelihood that two unrelated subjects affected by SCA17 share haplotype 5_2 and that this haplotype co-segregates with SCA17 in both families is ~2%. This strongly supports the view of a possible founder haplotype of the SCA17 causing alleles in families A and B.
Discussion
In comparison to other spinocerebellar ataxias, some genetic features in SCA17 are remarkable: The majority of pathological alleles contains an interspersed CAA CAG CAA element separating the (CAG)n sequence into two parts (table 1). In three SCA17 pedigrees known to date the pathogenic allele lacks the CAA CAG CAA interruption and is associated with instability and variable repeat expansion upon transmission. Two of the three families are of northern German origin and were investigated in the study presented. In both families, SCA17 co-segregates with an expanded, uninterrupted, and instable (CAG)n in TBP located on chromosomes sharing a haplotype in the close centromeric neighborhood of TBP, respectively (figure 1). Although the unstable repetitive sequence is linked with one of the three more commonly prevalent haplotypes, statistics point to a nominally significant likelihood that our observation reflects a founder effect rather than coincidence by chance.
In some cases, the expanded glutamine stretch may be the consequence of an intragenic duplication. So, the first published case with SCA17 is a Japanese girl with a de novo duplication event in her paternal allele [7]. In addition, doubling of a stretch of 19 Gln codons was found in a three-generation SCA17 pedigree [16]. Expanded polyglutamine stretches arising from duplications represent rare events in SCAs. Similarly, a pathogenic elongation of the polyalanine part within the polyadenylate binding protein nuclear 1 resulting from duplication has been described for oculopharyngeal muscular dystrophy [17]. The identification of these mutations gives support to the hypothesis of unequal crossing-over as one possible molecular mechanism for repeat expansions. In SCA17, such rare unequal crossing-over events may have been the causative mechanism underlying both the loss of the interspersed CAA CAG CAA element and expansion of (CAG)n. As discussed for SCA2 [18] the presence of CAA interruptions in SCA17 alleles breaks the repetitive sequence into shorter homogenous triplet tracts and may thus protect it from instability by reducing the slippage between the complementary strands.
In SCA17, there is a broad range of intermediate TBP alleles Gln43–48 associated with reduced penetrance. In other SCA types, such intermediate alleles are rarely found [19,20] or even unknown. However, intermediate alleles represent a well-known finding in patients with Huntington disease [21]. We found (CAG/CAA)49 but no visible symptoms at the age of 76 in the father of two SCA17 patients with expanded TBP alleles (CAG/CAA)52–53. This large allele, (CAG/CAA)49, has not been found in control samples. It may be of reduced penetrance, associated with very late age of onset and/or low expression of clinical signs. In addition, we performed haplotype analysis for a sporadic patient with homozygosity of the intermediate allele (CAG/CAA)47 and a rather progressive course [15]. Here, we cannot exclude the possibility of an additive effect of two intermediate alleles with respect to pathogenicity. In this patient, (CAG/CAA)47 is interrupted by CAA CAG CAA and linked with D6S446/D6S1590 haplotype 2_4. Thus, both the type of TBP expansion and the SCA17 linked haplotypes differ between the sporadic homozygous and the familial cases presented and do not point to a single founder. An extended analysis of intragenic markers in larger numbers of SCA17 families with different types of expansions could reveal "mutation prone" founder alleles differing with respect to the type of mutation.
Conclusion
The extraordinary high variability of the clinical expression of pathological TBP repeat expansions in SCA17 is further complicated by the occurrence of genetically instable repeats in association with the lack of an interspersing CAA CAG CAA element. The degree of reduced penetrance and very low expression of symptoms refer to strong modifying factors including potent influence of the genetic background. Unless SCA17 is a rare type of dominantly inherited ataxia, the repeat expansion within the TBP gene arose at unrelated genotypes. This has to be taken into account both in molecular genetic diagnostics and in genetic counseling.
Methods
Molecular analysis and polymorphic markers
Genomic DNA was prepared from peripheral blood leukocytes using standard protocols. SCA17 alleles were amplified by PCR as described [7], separated on 6% denaturing polyacrylamide gels, and visualized by silver staining. For sequencing, PCR products were separated under non-denaturing conditions using the dHPLC-HT-system (WAVE Transgenomic). The eluted fragments were sequenced using dye terminators and the automated capillary sequencer Avant 3100 (Applied Biosystems). For haplotype analysis, the highly polymorphic microsatellite markers D6S1719, D6S264, D6S281, D6S446, and D6S1590 were used as described [15]. Maximum likelihood haplotype frequencies of two-locus haplotypes based on markers D6S446 and D6S1590 genotyped in 96 unrelated anonymous control subjects and linkage disequilibrium were analyzed using Arlequin software version 2.001 [22].
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
CZ conceived of the study and developed the concept, interpreted the data and contributed substantially to the manuscript (corresponding author). AD carried out the molecular genetic studies. ES participated in the design of the study and helped to draft the manuscript. UF participated in the design and performed the 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
We would like to thank Ulrike Gehlken for excellent technical assistance and the German Heredo-Ataxia Society (DHAG), whose cooperation is essential in our work.
Figures and Tables
Figure 1 Pedigrees and haplotype analysis. Pedigrees of families A and B, including six-locus haplotypes based on five 6q-telomeric microsatellite markers (D6S1719, D6S264, D6S281, D6S446, D6S1590) and closely downstream located polymorphic poly-Gln encoding trinucleotide stretch within TBP. Alleles of the microsatellites are denoted by numbers 1 – 8, TBP alleles are denoted by the number of Gln-encoding trinucleotides. SCA17-linked alleles are boxed as well as the disease haplotype D6S446 and D6S1590.
Table 1 Review of Gln Repeats and SCA17 Alleles of 16 Unrelated Cases. Duplicated elements are boxed, instability by maternal (m) or paternal (p) inheritance.
Gln Case Repeat Composition Comments Ref.
49–53 familial (CAG)3 (CAA)3 (CAG)41–45 CAA CAG instability (p) here
63 sporadic (CAG)3 (CAA)3 (CAG)9 CAA CAG CAA (CAG)9 (CAA)3 (CAG)9 CAA CAG CAA (CAG)19 CAA CAG duplication [7]
47 familial (CAG)3 (CAA)3 (CAG)8 CAA CAG CAA (CAG)28 CAA CAG [16]
47 sporadic (CAG)3 (CAA)3 (CAG)6 CAA CAG CAA (CAG)30 CAA CAG
48 familial (CAG)3 (CAA)3 (CAG)6 CAA CAG CAA (CAG)31 CAA CAG
55 familial (CAG)3 (CAA)3 (CAG)9 CAA CAG CAA (CAG)16 CAA CAG CAA (CAG)16 CAA CAG duplication
51 familial (CAG)3 (CAA)3 (CAG)9 CAA CAG CAA (CAG)31 CAA CAG [5]
53 – 55 familial (CAG)3 (CAA)3 (CAG)45–47 CAA CAG instability (m)
48 familial (CAG)3 (CAA)3 (CAG)9 CAA CAG CAA (CAG)28 CAA CAG reduced penetrance [13]
47 sporadic (CAG)3 (CAA)3 (CAG)9 CAA CAG CAA (CAG)27 CAA CAG homozygous mutation [15]
46 familial (CAG)3 (CAA)3 (CAG)9 CAA CAG CAA (CAG)26 CAA CAG [21]
43 (CAG)3 (CAA)3 (CAG)9 CAA CAG CAA (CAG)26 CAA CAG [6]
53 – 66 familial (CAG)3 (CAA)4 (CAG)44/57 CAA CAG instability (p) [14]
44 sporadic (CAG)3 (CAA)3 (CAG)9 CAA CAG CAA (CAG)24 CAA CAG [12]
46 sporadic (CAG)3 (CAA)3 (CAG)11 CAA CAG CAA (CAG)24 CAA CAG reduced penetrance
48 familial (CAG)3 (CAA)3 (CAG)6 CAA CAG CAA (CAG)31 CAA CAG homozygous mutation [23]
==== Refs
Margolis RL The spinocerebellar ataxias: order emerges from chaos Curr Neurol Neurosci Rep 2002 2 447 456 12169226
Imbert G Trottier Y Beckmann J Mandel JL The gene for the TATA binding protein (TBP) that contains a highly polymorphic protein coding CAG repeat maps to 6q27 Genomics 1994 21 667 668 7959752 10.1006/geno.1994.1335
Polymeropoulos MH Rath DS Xiao H Merril CR Trinucleotide repeat polymorphism at the human transcription factor IID gene Nucleic Acids Res 1991 19 4307 1870994
Gostout B Liu Q Sommer SS 'Cryptic' repeating triplets of purines and pyrimidines (cRRY(i)) are frequent and polymorphic: analysis of coding cRRY(i) in the proopiomelanocortin (POMC) and TATA-binding protein (TBP) genes Am J Hum Genet 1993 52 1182 1190 8503450
Zühlke C Hellenbroich Y Dalski A Kononowa N Hagenah J Vieregge P Riess O Klein C Schwinger E Different types of repeat expansion in the TATA-binding protein gene are associated with a new form of inherited ataxia Eur J Hum Genet 2001 9 160 164 11313753 10.1038/sj.ejhg.5200617
Silveira I Miranda C Guimaraes L Moreira MC Alonso I Mendonca P Ferro A Pinto-Basto J Coelho J Ferreirinha F Poirier J Parreira E Vale J Januario C Barbot C Tuna A Barros J Koide R Tsuji S Holmes SE Margolis RL Jardim L Pandolfo M Coutinho P Sequeiros J Trinucleotide repeats in 202 families with ataxia: a small expanded (CAG)n allele at the SCA17 locus Arch Neurol 2002 59 623 629 11939898 10.1001/archneur.59.4.623
Koide R Kobayashi S Shimohata T Ikeuchi T Maruyama M Saito M Yamada M Takahashi H Tsuji S A neurological disease caused by an expanded CAG trinucleotide repeat in the TATA-binding protein gene: a new polyglutamine disease? Hum Molec Genet 1999 8 2047 2053 10484774 10.1093/hmg/8.11.2047
De Michele G Maltecca F Carella M Volpe G Orio M De Falco A Gombia S Servadio A Casari G Filla A Bruni A Dementia, ataxia, extrapyramidal features, and epilepsy: phenotype spectrum in two Italian families with spinocerebellar ataxia type 17 Neurol Sci 2003 24 166 167 14598069 10.1007/s10072-003-0112-4
Bauer P Laccone F Rolfs A Wullner U Bosch S Peters H Liebscher S Scheible M Epplen JT Weber BH Holinski-Feder E Weirich-Schwaiger H Morris-Rosendahl DJ Andrich J Riess O Trinucleotide repeat expansion in SCA17/TBP in white patients with Huntington's disease-like phenotype J Med Genet 2004 41 230 232 14985389 10.1136/jmg.2003.015602
Hagenah JM Zühlke Ch Hellenbroich Y Heide W Klein Ch Focal dystonia as presenting sign of SCA17 Mov Dis 2004 19 217 220 10.1002/mds.10600
Rolfs A Koeppen AH Bauer I Bauer P Buhlmann S Topka H Schols L Riess O Clinical features and neuropathology of autosomal dominant spinocerebellar ataxia (SCA17) Ann Neurol 2003 54 367 375 12953269 10.1002/ana.10676
Stevanin G Fujigasaki H Lebre AS Camuzat A Jeannequin C Dode C Takahashi J San C Bellance R Brice A Durr A Huntington's disease-like phenotype due to trinucleotide repeat expansions in the TBP and JPH3 genes Brain 2003 126 1599 1603 12805114 10.1093/brain/awg155
Zühlke C Gehlken U Hellenbroich Y Schwinger E Bürk K Phenotypical variability of expanded alleles in the TATA-binding protein gene. Reduced penetrance in SCA17? J Neurol 2003 250 161 163 12574945 10.1007/s00415-003-0958-7
Maltecca F Filla A Castaldo I Coppola G Fragassi NA Carella M Bruni A Cocozza S Casari G Servadio A De Michele G Intergenerational instability and marked anticipation in SCA-17 Neurology 2003 61 1441 1443 14638975
Zühlke C Spranger M Spranger S Voigt R Lanz M Gehlken U Hinrichs F Schwinger E SCA17 caused by homozygous repeat expansion in TBP due to partial isodisomy 6 Eur J Hum Genet 2003 11 629 632 12891385 10.1038/sj.ejhg.5201018
Nakamura K Jeong SY Uchihara T Anno M Nagashima K Nagashima T Ikeda S Tsuji S Kanazawa I SCA17, a novel autosomal dominant cerebellar ataxia caused by an expanded polyglutamine in TATA-binding protein Hum Molec Genet 2001 10 1441 1448 11448935 10.1093/hmg/10.14.1441
van der Sluijs BM van Engelen BG Hoefsloot LH Oculopharyngeal muscular dystrophy (OPMD) due to a small duplication in the PABPN1 gene Hum Mutat 2003 21 553 12673802 10.1002/humu.9138
Choudhry S Mukerji M Srivastava AK Jain S Brahmachari SK CAG repeat instability at SCA2 locus : anchoring CAA interruptions and linked single nucleotide polymorphisms Hum Mol Genet 2001 10 2437 2446 11689490 10.1093/hmg/10.21.2437
Zühlke C Dalski A Hellenbroich Y Bubel S Schwinger E Bürk K Spinocerebellar ataxia type 1 (SCA1): phenotype-genotype correlation studies in intermediate alleles Eur J Hum Genet 2002 10 204 209 11973625 10.1038/sj.ejhg.5200788
Andrew SE Goldberg YP Hayden MR Rethinking genotype and phenotype correlations in polyglutamine expansion disorders Hum Mol Genet 1997 6 2005 2010 9328463 10.1093/hmg/6.12.2005
McNeil SM Novelletto A Srinidhi J Barnes G Kornbluth I Altherr MR Wasmuth JJ Gusella JF MacDonald ME Myers RH Reduced penetrance of the Huntington's disease mutation Hum Mol Genet 1997 6 775 779 9158152 10.1093/hmg/6.5.775
Schneider S Roessli D Excofier L Arlequin: A software for population genetics data analysis. Ver 2.000 2000 Genetics and Biometry Lab, Dept. of Anthropology, University of Geneva
Toyoshima Y Yamada M Onodera O Shimohata M Inenaga C Fujita N Morita M Tsuji S Takahashi H SCA17 homozygote showing Huntington's disease-like phenotype Ann Neurol 2004 55 281 286 14755733 10.1002/ana.10824
| 15989694 | PMC1177950 | CC BY | 2021-01-04 16:03:32 | no | BMC Med Genet. 2005 Jul 1; 6:27 | utf-8 | BMC Med Genet | 2,005 | 10.1186/1471-2350-6-27 | oa_comm |
==== Front
BMC MedBMC Medicine1741-7015BioMed Central London 1741-7015-3-121600461510.1186/1741-7015-3-12Research ArticlePublic beliefs about treatment and outcome of mental disorders: a comparison of Australia and Japan Jorm Anthony F [email protected] Yoshibumi [email protected] Helen [email protected] Kumiko [email protected] Kathleen M [email protected] Yuji [email protected] ORYGEN Research Centre, Department of Psychiatry, University of Melbourne, Locked Bag 10, Parkville, Victoria 3052, Australia2 Centre for Mental Health Research, Australian National University, Canberra, ACT 0200, Australia3 Department of Social Work, The Faculty of Human Sociology, Nagasaki International University, 2825-7 Huis Ten Bosch-cho, Sasebo-shi, Nagasaki, 859-3298, Japan4 Department of Human Studies, Bunkyo Gakuin University,1196 Kamekubo, Oi-machi, Iruma-gun, Saitama 356-8533, Japan2005 9 7 2005 3 12 12 20 1 2005 9 7 2005 Copyright © 2005 Jorm et al; licensee BioMed Central Ltd.2005Jorm 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
Surveys of the public in a number of countries have shown poor recognition of mental disorders and beliefs about treatment that often diverge from those of health professionals. This lack of mental health literacy can limit the optimal use of treatment services. Australia and Japan are countries with very different mental health care systems, with Japan emphasising hospital care and Australia more oriented to community care. Japan is also more collectivist and Australia more individualist in values. These differences might influence recognition of disorders and beliefs about treatment in the two countries.
Methods
Surveys of the public were carried out in each country using as similar a methodology as feasible. In both countries, household interviews were carried out concerning beliefs in relation to one of four case vignettes, describing either depression, depression with suicidal thoughts, early schizophrenia or chronic schizophrenia. In Australia, the survey involved a national sample of 3998 adults aged 18 years or over. In Japan, the survey involved 2000 adults aged between 20 and 69 from 25 regional sites spread across the country.
Results
The Japanese public were found to be more reluctant to use psychiatric labels, particularly for the depression cases. The Japanese were also more reluctant to discuss mental disorders with others outside the family. They had a strong belief in counsellors, but not in GPs. They generally believe in the benefits of treatment, but are not optimistic about full recovery. By contrast, Australians used psychiatric labels more often, particularly "depression". They were also more positive about the benefits of seeking professional help, but had a strong preference for lifestyle interventions and tended to be negative about some psychiatric medications. Australians were positive about both counsellors and GPs. Psychiatric hospitalization and ECT were seen negatively in both countries.
Conclusion
There are some major differences between Australia and Japan in recognition of disorders and beliefs about treatment. Some of these may relate to the different health care systems, but the increasing openness about mental health in Australia is also likely to be an explanatory factor.
==== Body
Background
While there is now a range of effective methods for the treatment and management of various mental disorders, many people still receive no professional help or do not receive optimal help [1]. There are many factors that affect this unmet need for treatment. One of these is a lack of mental health literacy on the part of the public, specifically a lack of knowledge of how to recognise mental disorders and beliefs about treatment that are at variance with those of health professionals [2].
Surveys in several countries have found that many members of the public do not correctly recognise disorders in a case vignette [3-5] and that they misunderstand terms such as "schizophrenia" and "mania" [6-8]. Failure to use correct psychiatric labels may cause problems of communication with health practitioners. For example, it is known that GPs are more likely to detect a mental disorder if the patient presents the symptoms in psychological rather than somatic terms [9,10], and if the patient explicitly raises the possibility of a mental disorder with the GP [11,12].
Beliefs about various types of professional help are also important. For example, if a person with a mental disorder believes that consulting a psychiatrist or psychologist is unlikely to be helpful, this will reduce their chance of getting appropriate help. Surveys in Australia and Germany have found that psychiatrists and psychologists are rated less highly than GPs for depression, but are more likely to be seen as helpful for schizophrenia [8,13].
Beliefs about types of treatment also play a role. Surveys in several countries have found predominantly negative attitudes towards psychotropic medication [7,13-18], both because of concern about side effects and the belief that medications only deal with the symptoms rather than the causes [15-18]. Such beliefs may affect adherence to prescribed medication. By contrast, psychological therapies are seen more positively [7,13,16,18-20], as are complementary therapies such as vitamins and herbs [13,20].
While surveys of public beliefs have been carried out in a number of countries, little is known about cross-cultural differences in mental health literacy. In the present paper we report data from surveys in Australia and Japan that were carried out at the same time using as similar a methodology as possible. The contrast between these two countries is interesting because of their very different systems of mental health care. While both countries have a high level of economic development and a high standard of health care, Australia places more emphasis on community care of people with mental disorders and more attention is paid to the high prevalence disorders such as depression. By contrast, in Japan there is more emphasis on hospital care, with much longer in-patient stays than in Australia, and the mental health system is largely concerned with the lower prevalence psychotic disorders. The factors leading to the emphasis on hospital care in Japan include: financial incentives for private hospital in-patient treatment under the national insurance scheme, a lack of community support programs, and the strong stigma against people with mental disorders [21]. Australia also has a system reliant on general practitioners as the first point of call for any health problem and as the gatekeepers to specialist care. By contrast, in Japan there is no specific training in primary care. Family practitioners with offices in the community provide primary care, but they are trained in other specialities [21]. Another difference between the countries, supported by informal observations, is that psychological treatments are more prominent in Australia than in Japan. There may also be cultural differences in the acceptability of expressing negative emotions and displaying behavior that departs from social norms, with Australian society being more individualist and Japanese society more collectivist. Given these differences, we expected some major differences in knowledge and beliefs about mental disorders and their treatment between the two countries.
Methods
Survey interview
Interview questionnaires were developed for each country to have a common core of questions that would allow comparisons between countries, and a country-specific component to allow investigation of issues particular to each country. The common core was based on the interview used in an earlier Australian survey [2], but with additional questions. The interview was based on a vignette of a person with a mental disorder. On a random basis, respondents were shown one of four vignettes: a person with major depression, one with major depression together with suicidal thoughts, a person with early schizophrenia, and one with chronic schizophrenia. All vignettes were written to satisfy the diagnostic criteria for either major depression or schizophrenia according to DSM-IV and ICD-10. The vignette with depression and the one with early schizophrenia were written to satisfy at a minimal level these diagnostic criteria, so that we could ascertain the public's reaction to cases of developing disorder that had reached the point where intervention was needed. The vignette of the person with depression together with suicidal thoughts was identical to the depression vignette in all respects except the suicidal thoughts and was designed to assess how this symptom affected the public's response. The chronic schizophrenia vignette was designed to assess the response to someone with a severe long-standing disorder, where acceptance seemed less likely. Respondents were also randomly assigned to receive either male ("John") or female ("Mary") versions of the vignette.
The depression vignette (John version) was:
John is 30 years old. He has been feeling unusually sad and miserable for the last few weeks. Even though he is tired all the time, he has trouble sleeping nearly every night. John doesn't feel like eating and has lost weight. He can't keep his mind on his work and puts off making decisions. Even day-to-day tasks seem too much for him. This has come to the attention of his boss, who is concerned about John's lowered productivity.
The depression with suicidal thoughts vignette was:
John is 30 years old. He has been feeling unusually sad and miserable for the last few weeks. Even though he is tired all the time, he has trouble sleeping nearly every night. John doesn't feel like eating and has lost weight. He can't keep his mind on his work and puts off making any decisions. Even day-to-day tasks seem too much for him. This has come to the attention of John's boss who is concerned about his lowered productivity. John feels he will never be happy again and believes his family would be better off without him. John has been so desperate, he has been thinking of ways to end his life.
The early schizophrenia vignette was:
John is 24 and lives at home with his parents. He has had a few temporary jobs since finishing school but is now unemployed. Over the last six months he has stopped seeing his friends and has begun locking himself in his bedroom and refusing to eat with the family or to have a bath. His parents also hear him walking about his bedroom at night while they are in bed. Even though they know he is alone, they have heard him shouting and arguing as if someone else is there. When they try to encourage him to do more things, he whispers that he won't leave home because he is being spied upon by the neighbour. They realize he is not taking drugs because he never sees anyone or goes anywhere.
The chronic schizophrenia vignette was:
John is 44 years old. He is living in a boarding house in an industrial area. He has not worked for years. He wears the same clothes in all weathers and has left his hair to grow long and untidy. He is always on his own and is often seen sitting in the park talking to himself. At times he stands and moves his hands as if to communicate to someone in nearby trees. He rarely drinks alcohol. He speaks carefully using uncommon and sometimes made-up words. He is polite but avoids talking with other people. At times he accuses shopkeepers of giving information about him to other people. He has asked his landlord to put extra locks on his door and to remove the television set from his room. He says spies are trying to keep him under observation because he has secret information about international computer systems which control people through television transmitters. His landlord complains that he will not let him clean the room which is increasingly dirty and filled with glass objects. John says he is using these "to receive messages from space".
After being presented with the vignette, respondents were asked two open-ended questions: "What would you say, if anything, is wrong with John/Mary?" and "How do you think John/Mary could best be helped?" Then followed a series of questions asking the respondent to rate the likely helpfulness of various interventions (rated as likely to be helpful, harmful or neither for the person in the vignette). The interventions were: a typical GP or family doctor; a typical chemist (pharmacist); a counselor; a social worker; a telephone counseling service, such as Lifeline; a psychiatrist; a psychologist; help from close family; help from close friends; a naturopath or a herbalist; the clergy, a minister or priest; John/Mary tried to deal with his/her problems on his/her own; vitamins and mineral, tonics or herbal medicines; pain relievers, such as aspirin, codeine or panadol; antidepressants; antibiotics; sleeping pills; anti-psychotics; tranquillizers such as valium; becoming physically more active, such as playing more sport, or doing a lot more walking or gardening; reading about people with similar problems and how they have dealt with them; getting out and about more; attending courses or relaxation, stress management, meditation or yoga; cutting out alcohol altogether; psychotherapy; hypnosis; being admitted to a psychiatric ward of a hospital; undergoing electro-convulsive therapy (ECT); having an occasional alcoholic drink to relax; going on a special diet or avoiding certain foods. Next were questions asking about the likely result for the person in the vignette with and without "the sort of professional help you think is most appropriate" The response options were: Full recovery with no further problems; Full recovery, but problems would probably re-occur; Partial recovery; Partial recovery, but problems would probably re-occur; No improvement; Get worse.
The rest of the common core interview is not relevant to the analyses reported here; it involved questions on knowledge of causes and risk factors, beliefs associated with stigma and discrimination, contact with people like those in the vignette, and the health of the respondent.
The Australian survey
A household survey was carried out of Australian adults aged 18 or over by the company AC Nielsen. Households were sampled from 250 census districts covering all states and territories and metropolitan and rural areas. Up to 5 call backs were made to metropolitan selections and 3 to non-metropolitan selections. Interviews were sought with the person in the household who had the most recent birthday. To achieve a target sample of 4,000 interviews with adults aged 18 years or over, visits were made to 28,947 households. The outcome of these visits was: no contact after repeated visits 14,630; vacant house or lot 306; refused 7,815; person sampled within household temporarily unavailable 1,132; no suitable respondent in household 287; did not speak English 383; incapable of responding 213; and unavailable for the duration of the survey 181. The achieved sample was 3998 persons, with 1001 receiving the depression vignette, 999 the depression with suicidal thoughts vignette, 997 the early schizophrenia vignette, and 1001 the chronic schizophrenia vignette.
In addition to the common core component, the Australian survey interview had questions about awareness of depression in the media and about Australia's national depression initiative.
Ethics approval was given by the Human Research Ethics Committee of the Australian National University.
The Japanese survey
A survey manual supplied from Australia was translated into Japanese and entrusted to Yamate Information Processing Center Ltd. for use with the target population aged 20–69 years, as a rule using the same procedures as Australia. The survey questionnaire, which was developed by the Australian researchers (AFJ, HC, KMG), was tentatively translated into Japanese. Then a native English translator, who had not seen the original English text, translated the Japanese version back into English. By comparing the two English versions, it was possible to confirm the accuracy of the original translation. There were no significant differences between the original text and the reverse translation. Finally, a Japanese version of the questionnaire was produced, which involved formatting the text into Japanese style and making slight wording adjustments. The names of the characters in the vignettes were translated into the Japanese style, viz. "A-o"(putting an o sound at the end is often used for a man's name) or "B-ko" (putting ko at the end is often used for a woman's name), instead of "John" or "Mary" which were used in the English text.
As well as the questions taken from the Australian survey, the Japanese survey asked questions concerning such issues as psychiatric health and welfare policy, the bodies implementing related services, the existence of action groups, and the change in the Japanese name for schizophrenia by the Japanese Society of Psychiatry and Neurology (i.e. from "split personality disorder" to "schizophrenia"). These additions were made to clarify the current Japanese situation and issues in related fields. Further, an original Japanese manual was also created and adopted for use concerning points of interest in the implementation of home visits.
The survey method used was home visit interviews. It was not feasible to do a national survey of randomly selected households in Japan because of constraints of human resources, funding and time. It was therefore decided to sample a range of areas that differed in whether they were large or small cities, whether the area had many psychiatric patients or not, and whether the area had a high suicide rate or not. Using this approach, Japan was divided into 5 areas and 5 research sites were selected in each of these areas, giving a total of 25 geographic sites. As the survey was conducted during the winter, and because it was difficult to ensure that there would be enough survey interviewers, implementation in Hokkaido and Shikoku prefectures proved troublesome. Additional reasons for selection of the 25 regional sites were that they were places of comparatively high population within the relevant regions, the survey interviewers could use public transport, and the urban areas involved no particular inconveniences for the researchers to visit within a certain range using public transportation. 80 households were selected from each site, giving a total of 2000. At each site there were 4 interviewers who took responsibility for 20 households each. The survey was conducted over the period from 19 November to 12 December 2003. Each of the four vignettes was received by 250 people. Half received a male version of a vignette and half the female version.
At the start of the survey, an explanatory meeting was held for the survey interviewers in each region. As many members of the research team as possible attended these explanatory meetings. Eighty-five survey interviewers were recruited for this research with an average age of 50 and an average of 17 years' experience of interviewing in various types of surveys. The areas for the survey interviewers to canvass were allocated on the basis of where they lived. The question of where the individual survey interviewers should go was determined mutually among the survey interviewers themselves, and by the head survey interviewer (supervisor). As a rule, one survey interviewer conducted 20 interviews, but this was considerably flexible, given the number of years of individual experience and what the individual survey interviewer could handle. The interviews were conducted according to the following procedure: visit the target's home and present the written greetings and request (a draft had been prepared by certain survey bodies, which was put into final form after checks by the research team members), then explain the details of the survey using the documents, ask the target for their participation in the research, start the interview and follow through to completion, check that nothing had been omitted from the survey responses, and hand over the remuneration (1000 yen cash voucher). Data were not collected on the refusal rate for this survey.
Statistical analysis
Data were pooled across male and female versions of each vignette and percent frequencies calculated. For the Australian survey, percentages were calculated applying survey weights to give better population estimates. Ninety-five percent CIs were estimated using the Complex Samples procedure in SPSS 12.0. This procedure takes account of sampling weights and geographic clustering in the sample. For the Japanese survey, percentage frequencies and 95% CIs were calculated using unweighted data with SPSS 12.0.
Because of the very different cultures of Australia and Japan, it is possible that any differences in question endorsement rates might be due to subtleties of language or to the social rules applying to the interview situation, as well as to genuine differences in beliefs about treatment and outcome. For this reason, we have not relied on statistical significance of absolute percentages between countries, but rather on the broad patterns of responses, particularly where percent endorsement was ordered very differently across questions.
Results
Recognition of disorders
Table 1 shows the results from both countries. In Australia, "depression" was the term used most often to describe both the depression vignette and the depression with suicidal thoughts vignette. "Schizophrenia/psychosis" was the term used most often to describe both of the schizophrenia vignettes, while the generic term "mental illness" was also commonly used for these vignettes.
Table 1 Percentage (and 95% CI) of respondents mentioning each category to describe the problem shown in the vignette
Category mentioned Country Depression Vignette Depression/Suicidal Vignette Early Schizophrenia Vignette Chronic Schizophrenia Vignette
Depression Australia 65.3 (60.5–69.8) 77.3 (72.7–81.3) 34.8 (30.5–39.4) 9.6 (7.0–13.0)
Japan 22.6 (18.9–26.3) 35.0 (30.8–39.2) 13.6 (10.6–16.6) 9.6 (7.0–12.2)
Schizophrenia/Psychosis Australia 0.0 (0.0–0.0) 0.5 (0.1–1.6) 41.2 (36.5–46.0) 36.1 (31.5–40.9)
Japan 2.2 (0.9–3.5) 1.2 (0.2–2.2) 17.2 (13.9–20.5) 33.4 (29.3–37.5)
Nervous breakdown Australia 0.7 (0.3–2.1) 1.6 (0.8–3.3) 1.7 (0.9–3.2) 1.0 (0.3–3.4)
Japan 2.0 (0.8–3.2) 2.6 (1.2–4.0) 2.6 (1.2–4.0) 2.4 (1.1–3.7)
Psychological/Mental/Emotional problems Australia 4.5 (2.9–6.8) 6.0 (4.2–8.7) 12.9 (10.1–16.3) 14.3 (10.9–18.5)
Japan 29.4 (25.4–33.4) 24.8 (21.0–28.6) 28.4 (24.4–32.4) 27.2 (23.3–31.1)
Mental illness Australia 3.0 (1.7–5.1) 5.5 (3.7–8.2) 23.0 (19.4–27.0) 35.8 (31.4–40.4)
Japan 9.2 (6.7–11.7) 10.2 (7.5–12.9) 21.6 (18.0–25.2) 12.8 (9.9–15.7)
Stress Australia 16.6 (13.1–20.8) 10.9 (8.3–14.3) 3.1 (1.8–5.3) 2.8 (1.4–5.5)
Japan 25.0 (21.2–28.8) 19.8 (16.3–23.3) 5.0 (3.1–6.9) 3.8 (2.1–5.5)
In Japan, no single term predominated for describing the depression vignettes, with "depression", "stress" and "psychological/mental/emotional problems" being the most common. For the early schizophrenia vignette, the generic categories of "mental illness" and "psychological/mental/emotional problems" were used most frequently, while for the chronic schizophrenia vignette, "schizophrenia" and "psychological/mental/emotional problems" were most commonly used.
Best method of help
Table 2 shows the frequency of various responses to the open-ended question about how the person in the vignette could best be helped. In Australia, half the respondents mentioned seeing a GP for the depression vignettes. Other common responses to the depression vignettes were seeing a counselor or talking with friends or family. For the schizophrenia vignettes, seeing a psychiatrist was commonly mentioned, in addition to seeing a GP or counselor or talking with friends or family.
Table 2 Percentage (and 95% CI) of respondents mentioning each category in response to the open-ended question about how the person in the vignette could best be helped
Type of help mentioned Country Depression Vignette Depression/Suicidal Vignette Early Schizophrenia Vignette Chronic Schizophrenia Vignette
See a doctor/GP Australia 56.3 (52.5–60.0) 49.3 (44.4–54.1) 32.1 (27.6–37.0) 21.0 (17.2–25.4)
Japan 20.8 (17.2–24.4) 20.0 (16.5–23.5) 13.4 (10.4–16.4) 15.8 (12.6–19.0)
See a psychiatrist Australia 13.0 (10.7–15.6) 18.2 (14.9–22.0) 32.0 (27.8–36.4) 27.9 (23.6–32.6)
Japan 43.8 (39.4–48.2) 49.0 (44.6–53.4) 58.6 (54.3–62.9) 60.0 (55.7–64.3)
Take medication Australia 6.1 (4.7–7.8) 8.7 (6.5–11.5) 8.3 (6.0–11.4) 11.7 (9.0–15.1)
Japan 3.4 (1.8–5.0) 8.0 (5.6–10.4) 7.8 (5.4–10.2) 5.8 (3.7–7.9)
See a counselor or have counseling Australia 27.7 (24.7–30.9) 37.4 (33.2–41.8) 28.9 (24.7–33.4) 20.8 (17.3–24.7)
Japan 62.0 (57.7–66.3) 74.8 (71.0–78.6) 76.4 (72.7–80.1) 72.2 (68.3–76.1)
Talk over with friends/family Australia 22.9 (19.8–26.4) 24.0 (20.0–28.6) 21.9 (18.0–26.3) 14.4 (11.2–18.4)
Japan 71.8 (67.8–75.8) 70.4 (66.4–74.4) 43.8 (39.4–48.2) 45.2 (40.8–49.6)
Person must first recognize problem Australia 5.4 (3.8–7.6) 6.6 (4.3–9.9) 5.3 (3.3–8.4) 6.0 (4.1–8.7)
Japan 23.4 (19.7–27.1) 24.4 (20.6–28.2) 23.4 (19.7–27.1) 21.8 (18.2–25.4)
Other Australia 37.9 (34.0–41.8) 36.0 (31.1–41.1) 40.1 (34.9–45.4) 49.8 (44.6–55.0)
Japan 8.4 (6.0–10.8) 2.6 (1.2–4.0) 4.4 (2.6–6.2) 7.2 (4.9–9.5)
Don't know Australia 1.8 (1.1–3.0) 2.5 (1.4–4.6) 2.0 (1.0–3.7) 4.8 (3.1–7.5)
Japan 0.4 (0.0–1.0) 1.0 (0.1–1.9) 1.6 (0.5–2.7) 1.2 (0.2–2.2)
Note: because multiple responses were possible, these percentages do not add up to 100%
In Japan, the most commonly mentioned help for the depression vignettes was counseling and family or friends. For the schizophrenia vignettes, seeing a counselor or a psychiatrist were commonly mentioned, but talking it over with family or friends was less commonly mentioned than for the depression vignettes. Seeing a GP was seldom mentioned for any vignette.
Beliefs about specific interventions
Tables 3, 4, 5 show the data on the ratings of likely helpfulness of interventions. The Australian public gave similar ratings for the two depression vignettes. The interventions most commonly endorsed as likely to be helpful were seeing a GP, counselor, close friends, physical activity, reading about the problem, getting out more, learning relaxation, or getting information from a health educator. The responses were similar for the schizophrenia vignettes, except that GPs were rated somewhat lower and psychiatrists somewhat higher. Seeing a social worker was also commonly endorsed for the chronic schizophrenia vignette.
Table 3 Percentage (95% CI) of respondents rating each type of person as "helpful" for the person described in the vignette
Person Country Depression Vignette Depression/Suicidal Vignette Early Schizophrenia Vignette Chronic Schizophrenia Vignette
GP Australia 87.3 (84.0–90.0) 84.1 (80.8–87.0) 76.7 (72.8–80.2) 76.3 (72.1–80.0)
Japan 30.4 (26.4–34.4) 26.0 (22.1–29.9) 19.0 (15.5–22.5) 22.8 (19.1–26.5)
Pharmacist Australia 35.4 (31.3–39.6) 33.2 (29.1–37.6) 23.6 (20.0–27.6) 28.1 (24.1–32.5)
Japan 6.8 (4.6–9.0) 6.6 (4.4–8.8) 4.2 (2.4–6.0) 4.2 (2.4–6.0)
Counselor Australia 82.2 (78.6–85.4) 85.5 (82.1–88.3) 85.0 (81.8–87.8) 83.1 (79.5–86.1)
Japan 85.8 (82.7–88.9) 87.6 (84.7–90.5) 87.0 (84.0–90.0) 88.6 (85.8–91.4)
Social worker Australia 62.8 (58.3–67.0) 67.2 (62.8–71.4) 68.4 (64.0–72.5) 79.1 (74.8–82.8)
Japan 73.4 (69.5–77.3) 70.2 (66.2–74.2) 68.4 (64.3–72.5) 75.2 (71.4–79.0)
Phone counseling Australia 63.5 (59.0–67.9) 66.2 (61.8–70.4) 56.6 (52.4–60.7) 47.5 (43.0–52.1)
Japan 42.4 (38.1–46.7) 49.8 (45.4–54.2) 35.6 (31.4–39.8) 29.6 (25.6–33.6)
Psychiatrist Australia 65.0 (60.8–69.0) 71.3 (67.1–75.1) 80.5 (76.5–84.0) 80.2 (76.4–83.5)
Japan 69.4 (65.3–73.5) 72.4 (68.5–76.3) 73.0 (69.1–76.9) 79.0 (75.4–82.6)
Psychologist Australia 66.9 (62.5–71.1) 69.7 (65.2–73.8) 73.6 (69.4–77.4) 74.9 (70.8–78.6)
Japan 56.6 (52.2–61.0) 51.2 (46.8–55.6) 56.2 (51.8–60.6) 65.2 (61.0–69.4)
Close family Australia 67.9 (63.1–72.3) 64.8 (60.1–69.2) 62.7 (58.4–66.8) 61.4 (56.2–66.3)
Japan 85.0 (81.9–88.1) 84.2 (81.0–87.4) 76.8 (73.1–80.5) 80.4 (76.9–83.9)
Close friends Australia 78.2 (74.1–81.7) 77.1 (73.2–80.5) 73.0 (68.9–76.7) 72.0 (67.5–76.1)
Japan 84.8 (81.6–88.0) 83.2 (79.9–86.5) 70.4 (66.4–74.4) 70.2 (66.2–74.2)
Naturopath/herbalist Australia 34.9 (30.8–39.3) 31.8 (27.5–36.5) 23.7 (20.2–27.7) 19.4 (16.3–22.9)
Japan 11.2 (8.4–14.0) 14.8 (11.7–17.9) 8.4 (6.0–10.8) 9.0 (6.5–11.5)
Clergy Australia 45.3 (41.0–49.7) 51.7 (47.3–56.0) 37.2 (33.1–41.4) 42.9 (38.3–47.7)
Japan 13.6 (10.6–16.6) 20.0 (16.5–23.5) 11.6 (8.8–14.4) 16.2 (13.0–19.4)
Deal with it alone Australia 13.1 (10.1–16.8) 9.7 (7.0–13.2) 11.4 (8.4–15.3) 11.8 (8.9–15.6)
Japan 24.4 (20.6–28.2) 20.4 (16.9–23.9) 22.4 (18.7–26.1) 21.4 (17.8–25.0)
Table 4 Percentage (95% CI) of respondents rating each type of medication as "helpful" for the person described in the vignette
Medication Country Depression Vignette Depression/Suicidal Vignette Early Schizophrenia Vignette Chronic Schizophrenia Vignette
Vitamins, minerals Australia 50.2 (45.4–55.1) 43.7 (39.2–48.3) 31.3 (27.4–35.5) 33.2 (28.9–37.8)
Japan 20.2 (16.7–23.7) 16.4 (13.1–19.7) 10.6 (7.9–13.3) 12.4 (9.5–15.3)
Pain relievers Australia 14.8 (11.7–18.5) 12.3 (9.6–15.7) 7.3 (5.3–9.9) 10.2 (7.7–13.4)
Japan 4.4 (2.6–6.2) 3.6 (2.0–5.2) 4.2 (2.4–6.0) 4.6 (2.8–6.4)
Antidepressants Australia 46.7 (42.4–51.1) 52.5 (48.1–56.7) 49.9 (45.7–54.2) 42.6 (37.9–47.5)
Japan 34.8 (30.6–39.0) 36.0 (31.8–40.2) 38.6 (34.3–42.9) 39.8 (35.5–44.1)
Antibiotics Australia 10.4 (7.9–13.7) 7.9 (5.7–10.8) 4.0 (2.5–6.2) 6.4 (4.5–9.1)
Japan 6.2 (4.1–8.3) 6.0 (3.9–8.1) 4.8 (2.9–6.7) 8.4 (6.0–10.8)
Sleeping pills Australia 23.9 (20.1–28.1) 21.9 (18.6–25.6) 18.1 (14.7–22.1) 11.6 (8.8–15.1)
Japan 31.6 (27.5–35.7) 26.2 (22.3–30.1) 21.4 (17.8–25.0) 24.8 (21.0–28.6)
Antipsychotics Australia 11.2 (8.4–14.8) 16.5 (13.3–20.3) 33.1 (29.0–37.5) 38.2 (34.0–42.6)
Japan 22.6 (18.9–26.3) 21.8 (18.2–25.4) 30.2 (26.2–34.2) 41.2 (36.9–45.5)
Tranquillizers Australia 13.8 (11.0–17.1) 13.8 (11.0–17.1) 17.2 (14.1–20.8) 15.3 (12.6–18.4)
Japan 38.4 (34.1–42.7) 37.0 (32.8–41.2) 38.4 (34.1–42.7) 45.4 (41.0–49.8)
Table 5 Percentage (95% CI) of respondents rating each type of intervention as "helpful" for the person described in the vignette
Intervention Country Depression Vignette Depression/Suicidal Vignette Early Schizophrenia Vignette Chronic Schizophrenia Vignette
Physical activity Australia 92.0 (89.3–94.1) 92.5 (89.9–94.4) 87.4 (84.2–90.1) 79.6 (75.6–83.1)
Japan 69.4 (65.3–73.5) 73.4 (69.5–77.3) 73.4 (69.5–77.3) 70.6 (66.6–74.6)
Read about problem Australia 79.3 (75.3–82.8) 79.8 (75.8–83.3) 79.6 (75.5–83.1) 74.7 (70.6–78.4)
Japan 60.0 (55.7–64.3) 59.4 (55.1–63.7) 57.6 (53.3–61.9) 46.8 (42.4–51.2)
Get out more Australia 87.0 (83.8–89.7) 90.3 (87.4–92.6) 87.1 (83.8–89.8) 76.5 (72.3–80.2)
Japan 67.0 (62.9–71.1) 72.0 (68.1–75.9) 67.2 (63.1–71.3) 61.6 (57.3–65.9)
Learn relaxation Australia 83.6 (80.1–86.7) 85.3 (81.8–88.2) 77.1 (73.3–80.5) 68.7 (64.4–72.8)
Japan 38.2 (33.9–42.5) 41.2 (36.9–45.5) 26.2 (22.3–30.1) 29.4 (25.4–33.4)
Cut out alcohol Australia 56.0 (51.7–60.3) 59.8 (55.3–64.1) 66.1 (61.7–70.2) 53.4 (48.7–58.0)
Japan 10.0 (7.4–12.6) 14.2 (11.1–17.3) 18.6 (15.2–22.0) 17.2 (13.9–20.5)
Psychotherapy Australia 44.1 (39.7–48.5) 50.4 (46.0–54.8) 59.1 (54.5–63.6) 62.3 (57.4–66.8)
Japan 49.0 (44.6–53.4) 48.2 (43.8–52.6) 53.8 (49.4–58.2) 67.0 (62.9–71.1)
Hypnosis Australia 22.4 (18.8–26.5) 23.9 (20.2–28.1) 29.9 (25.9–34.3) 30.9 (26.8–35.2)
Japan 28.0 (24.1–31.9) 28.8 (24.8–32.8) 22.4 (18.7–26.1) 33.2 (29.1–37.3)
Psychiatric ward Australia 16.4 (13.2–20.2) 20.2 (16.7–24.3) 31.9 (27.8–36.4) 37.8 (33.4–42.6)
Japan 13.6 (10.6–16.6) 12.0 (9.1–14.9) 22.0 (18.4–25.6) 30.0 (26.0–34.0)
ECT Australia 5.9 (4.0–8.6) 7.2 (4.9–10.4) 6.4 (4.5–9.1) 6.5 (4.4–9.4)
Japan 2.2 (0.9–3.5) 1.4 (0.4–2.4) 1.4 (0.4–2.4) 1.4 (0.4–2.4)
Occasional drink Australia 44.4 (40.1–48.9) 41.8 (37.1–46.5) 31.1 (26.9–34.7) 27.3 (23.1–31.9)
Japan 31.4 (27.3–35.5) 25.0 (21.2–28.8) 15.2 (12.0–18.4) 20.0 (16.5–23.5)
Special diet Australia 48.3 (43.5–53.1) 45.6 (41.0–50.3) 42.1 (37.9–46.3) 39.3 (35.0–43.9)
Japan 5.6 (3.6–7.6) 6.0 (3.9–8.1) 4.4 (2.6–6.2) 4.4 (2.6–6.2)
Web site Australia 57.9 (53.8–61.9) 55.1 (50.4–59.7) 57.5 (53.0–61.8) 44.1 (39.4–49.0)
Japan 45.6 (41.2–50.0) 45.8 (41.4–50.2) 48.4 (44.0–52.8) 47.0 (42.6–51.4)
Expert via email Australia 53.8 (49.6–58.0) 49.6 (44.9–54.3) 55.4 (51.2–59.5) 44.7 (40.1–49.5)
Japan 54.0 (49.6–58.4) 53.6 (49.2–58.0) 56.8 (52.4–61.2) 56.6 (52.2–61.0)
Book Australia 69.1 (65.3–72.6) 64.7 (60.0–69.1) 70.5 (66.5–74.2) 59.2 (54.3–63.9)
Japan 54.0 (49.6–58.4) 49.8 (45.4–54.2) 57.4 (53.1–61.7) 53.6 (49.2–58.0)
Health educator Australia 86.7 (83.6–89.3) 85.9 (82.3–88.8) 86.2 (83.1–88.8) 83.8 (79.7–87.2)
Japan 55.2 (50.8–59.6) 51.2 (46.8–55.6) 46.6 (42.2–51.0) 50.6 (46.2–55.0)
In Japan, the most commonly endorsed interventions for the depression vignettes were seeing a counselor, or help from friends and family. For the schizophrenia vignettes, seeing a counselor rated highly again, as did close family. Psychiatrists and social workers were highly endorsed for the chronic schizophrenia vignette.
In neither country was there a high level of endorsement for some standard psychiatric interventions: antidepressants for the depression vignettes, antipsychotics for the schizophrenia vignettes, admission to a psychiatric ward for the schizophrenia vignettes, or psychotherapy for the depression vignettes.
Tables 6, 7, 8 show the data on whether interventions were rated as likely to be harmful. In Australia, "harmful" ratings were most common for dealing with the problem alone, sleeping pills, tranquillizers, ECT and admission to a psychiatric ward. In Japan, such ratings were most common for dealing with the problem alone, pain relievers, admission to a psychiatric ward, ECT and going on a special diet.
Table 6 Percentage (95% CI) of respondents rating each type of person as "harmful" for the person described in the vignette
Person Country Depression Vignette Depression/Suicidal Vignette Early Schizophrenia Vignette Chronic Schizophrenia Vignette
GP Australia 0.5 (0.2–1.6) 1.1 (0.5–2.5) 2.5 (1.5–4.3) 2.7 (1.5–4.7)
Japan 9.4 (6.8–12.0) 9.0 (6.5–11.5) 12.6 (9.7–15.5) 12.8 (9.9–15.7)
Pharmacist Australia 8.7 (6.3–12.1) 8.1 (5.7–11.5) 8.6 (6.4–11.4) 8.2 (5.8–11.5)
Japan 23.6 (19.9–27.3) 22.0 (18.4–25.6) 22.4 (18.7–26.1) 23.0 (19.3–26.7)
Counselor Australia 3.1 (1.8–5.3) 2.3 (1.3–3.8) 3.0 (1.8–5.0) 2.4 (1.3–4.4)
Japan 1.0 (0.1–1.9) 1.0 (0.1–1.9) 1.4 (0.4–2.4) 1.6 (0.5–2.7)
Social worker Australia 4.5 (3.0–6.7) 5.5 (3.6–8.2) 4.4 (3.0–6.5) 3.0 (1.8–5.0)
Japan 1.4 (0.4–2.4) 3.0 (1.5–4.5) 4.8 (2.9–6.7) 3.0 (1.5–4.5)
Phone counseling Australia 5.9 (4.1–8.6) 6.3 (4.2–9.2) 7.6 (5.3–10.9) 11.1 (8.4–14.6)
Japan 8.6 (6.1–11.1) 6.6 (4.4–8.8) 11.0 (8.2–13.8) 12.0 (9.1–14.9)
Psychiatrist Australia 7.1 (5.0–10.1) 8.1 (5.9–11.0) 5.2 (3.5–7.7) 4.6 (3.1–6.8)
Japan 5.4 (3.4–7.4) 4.8 (2.9–6.7) 6.0 (3.9–8.1) 2.8 (1.3–4.3)
Psychologist Australia 5.1 (3.3–7.9) 5.2 (3.5–7.7) 3.2 (2.0–5.0) 3.6 (2.3–5.6)
Japan 6.0 (3.9–8.1) 8.0 (5.6–10.4) 6.0 (3.9–8.1) 5.0 (3.1–6.9)
Close family Australia 4.9 (3.3–7.1) 4.1 (2.7–6.1) 5.6 (3.9–8.0) 5.3 (3.7–7.7)
Japan 1.6 (0.5–2.7) 1.6 (0.5–2.7) 4.6 (2.8–6.4) 4.4 (2.6–6.2)
Close friends Australia 2.1 (1.1–3.7) 2.6 (1.5–4.5) 3.0 (1.8–4.8) 3.3 (2.0–5.3)
Japan 1.8 (0.6–3.0) 1.4 (0.4–2.4) 4.0 (2.3–5.7) 4.2 (2.4–6.0)
Naturopath/herbalist Australia 11.1 (8.5–14.5) 13.3 (10.5–16.7) 15.1 (12.1–18.7) 15.0 (11.8–18.9)
Japan 18.8 (15.4–22.2) 17.2 (13.9–20.5) 18.2 (14.8–21.6) 21.4 (17.8–25.0)
Clergy Australia 8.1 (5.8–11.1) 9.3 (7.2–11.9) 11.6 (8.9–15.0) 10.3 (7.7–13.6)
Japan 24.2 (20.4–28.0) 14.6 (11.5–17.7) 26.0 (22.1–29.9) 24.4 (20.6–28.2)
Deal with it alone Australia 64.0 (59.6–68.3) 74.8 (70.4–78.7) 70.4 (65.9–74.5) 67.7 (62.9–72.2)
Japan 41.4 (37.1–45.7) 42.6 (38.3–46.9) 38.8 (34.5–43.1) 40.8 (36.5–45.1)
Table 7 Percentage (95% CI) of respondents rating each type of medication as "harmful" for the person described in the vignette
Medication Country Depression Vignette Depression/Suicidal Vignette Early Schizophrenia Vignette Chronic Schizophrenia Vignette
Vitamins, minerals Australia 4.4 (2.9–6.7) 5.4 (3.7–7.9) 5.8 (4.1–8.4) 6.4 (4.5–9.0)
Japan 14.6 (11.5–17.7) 13.8 (10.8–16.8) 14.6 (11.5–17.7) 14.8 (11.7–17.9)
Pain relievers Australia 37.7 (33.0–42.6) 37.3 (33.1–41.7) 38.9 (34.6–43.4) 34.5 (30.0–39.3)
Japan 43.4 (39.0–47.8) 42.6 (38.3–46.9) 36.6 (32.4–40.8) 35.6 (31.4–39.8)
Antidepressants Australia 27.5 (23.5–31.8) 23.4 (19.8–27.4) 22.6 (19.2–26.5) 29.3 (25.2–33.8)
Japan 18.2 (14.8–21.6) 21.2 (17.6–24.8) 15.2 (12.0–18.4) 10.6 (7.9–13.3)
Antibiotics Australia 38.3 (33.4–43.4) 37.8 (33.0–42.8) 35.9 (31.5–40.4) 36.9 (32.4–41.7)
Japan 29.8 (25.8–33.8) 37.6 (33.3–41.9) 29.0 (25.0–33.0) 22.8 (19.1–26.5)
Sleeping pills Australia 49.6 (44.9–54.3) 50.3 (45.9–54.7) 53.1 (48.2–57.9) 58.8 (54.3–63.1)
Japan 27.0 (23.1–30.9) 27.8 (23.9–31.7) 30.0 (26.0–34.0) 27.0 (23.1–30.9)
Antipsychotics Australia 48.3 (43.5–53.1) 40.4 (35.9–45.0) 24.5 (20.6–28.8) 24.5 (20.9–28.5)
Japan 19.0 (15.5–22.5) 23.8 (20.1–27.5) 17.4 (14.1–20.7) 8.4 (6.0–10.8)
Tranquillizers Australia 60.4 (55.8–64.8) 60.1 (55.5–64.5) 47.5 (42.9–52.2) 55.7 (50.9–60.4)
Japan 15.8 (12.6–19.0) 17.6 (14.3–20.9) 13.4 (10.4–16.4) 9.4 (6.8–12.0)
Table 8 Percentage (95% CI) of respondents rating each type of intervention as "harmful" for the person described in the vignette
Intervention Country Depression Vignette Depression/Suicidal Vignette Early Schizophrenia Vignette Chronic Schizophrenia Vignette
Physical activity Australia 0.8 (0.2–2.3) 0.3 (0.1–2.2) 0.4 (0.1–1.3) 0.6 (0.2–1.6)
Japan 3.6 (2.0–5.2) 4.0 (2.3–5.7) 3.8 (2.1–5.5) 3.6 (2.0–5.2)
Read about problem Australia 4.1 (2.7–6.2) 5.2 (3.6–7.5) 4.6 (3.1–6.8) 3.6 (2.4–5.4)
Japan 7.6 (5.3–9.9) 7.8 (5.4–10.2) 8.0 (5.6–10.4) 10.4 (7.7–13.1)
Get out more Australia 0.4 (0.1–1.7) 0.3 (0.1–1.5) 1.7 (0.9–3.2) 2.2 (1.2–4.1)
Japan 3.0 (1.5–4.5) 4.8 (2.9–6.7) 7.4 (5.1–9.7) 4.6 (2.8–6.4)
Learn relaxation Australia 1.5 (0.7–3.2) 0.8 (0.3–2.0) 1.0 (0.5–2.3) 3.6 (2.0–6.1)
Japan 7.6 (5.3–9.9) 8.2 (5.8–10.6) 16.4 (13.1–19.7) 13.6 (10.6–16.6)
Cut out alcohol Australia 4.7 (3.1–7.0) 5.3 (3.4–8.0) 3.1 (1.9–5.1) 2.7 (1.6–4.7)
Japan 17.2 (13.9–20.5) 15.0 (11.9–18.1) 11.4 (8.6–14.2) 12.2 (9.3–15.1)
Psychotherapy Australia 10.0 (7.5–13.1) 10.6 (8.0–13.9) 5.7 (3.9–8.2) 7.2 (5.2–10.0)
Japan 7.4 (5.1–9.7) 4.2 (2.4–6.0) 5.2 (3.2–7.2) 2.6 (1.2–4.0)
Hypnosis Australia 17.0 (13.9–20.7) 20.4 (16.5–24.9) 12.8 (10.1–16.2) 16.7 (13.4–20.5)
Japan 14.2 (11.1–17.3) 14.0 (10.9–17.1) 17.4 (14.1–20.7) 10.4 (7.7–13.1)
Psychiatric ward Australia 53.3 (48.4–58.1) 49.2 (44.4–54.1) 38.9 (34.4–43.5) 33.2 (28.9–37.7)
Japan 43.0 (38.6–47.4) 43.6 (39.2–48.0) 38.0 (33.7–42.3) 24.6 (20.8–28.4)
ECT Australia 69.4 (64.4–74.0) 65.9 (61.1–70.5) 63.4 (58.5–68.0) 65.4 (60.5–69.9)
Japan 50.2 (45.8–54.6) 54.4 (50.0–58.8) 50.6 (46.2–55.0) 44.0 (39.6–48.4)
Occasional drink Australia 15.4 (12.6–18.7) 19.1 (15.6–23.2) 29.8 (25.9–34.1) 25.2 (21.3–29.7)
Japan 17.4 (14.1–20.7) 20.2 (16.7–23.7) 31.4 (27.3–35.5) 26.8 (22.9–30.7)
Special diet Australia 7.7 (5.7–10.3) 9.2 (6.9–12.2) 7.7 (5.6–10.6) 7.1 (5.0–9.9)
Japan 55.2 (50.8–59.6) 55.6 (51.2–60.0) 53.2 (48.8–57.6) 50.4 (46.0–54.8)
Web site Australia 14.8 (12.0–18.0) 15.3 (12.3–18.8) 12.7 (10.3–15.6) 19.3 (16.2–22.8)
Japan 8.0 (5.6–10.4) 6.2 (4.1–8.3) 7.2 (4.9–9.5) 9.6 (7.0–12.2)
Expert via email Australia 14.3 (11.6–17.4) 16.4 (13.1–20.3) 13.8 (11.3–16.8) 17.3 (14.2–20.8)
Japan 5.0 (3.1–6.9) 5.6 (3.6–7.6) 5.2 (3.2–7.2) 5.8 (3.7–7.9)
Book Australia 7.7 (5.8–10.1) 9.0 (6.9–11.6) 7.1 (5.1–9.7) 9.4 (7.0–12.4)
Japan 3.0 (1.5–4.5) 4.4 (2.6–6.2) 4.2 (2.4–6.0) 5.0 (3.1–6.9)
Health educator Australia 1.4 (0.7–2.8) 2.0 (1.1–3.5) 1.4 (0.7–2.8) 1.6 (0.8–3.2)
Japan 4.4 (2.6–6.2) 3.8 (2.1–5.5) 5.2 (3.2–7.2) 4.6 (2.8–6.4)
Beliefs about outcomes
Table 9 gives the data on beliefs about outcomes after receiving professional help and outcomes without professional help. In Australia, the most common belief is that a person receiving professional help would have either full recovery or full recovery with later relapse. In Japan, the public most commonly believed in either full recovery with relapse or partial recovery with relapse.
Table 9 Percentage (95% CI) of respondents giving each outcome as likely for the person described in the vignette
Likely outcome Country Depression Vignette Depression/Suicidal Vignette Early Schizophrenia Vignette Chronic Schizophrenia Vignette
With professional help
Full recovery Australia 37.3 (32.8–42.1) 29.6 (25.5–34.1) 24.8 (21.1–29.0) 15.8 (12.4–19.9)
Japan 7.4 (5.1–9.7) 5.8 (3.7–7.9) 4.4 (2.6–6.2) 2.8 (1.3–4.3)
Full recovery with relapse Australia 43.6 (39.2–48.1) 48.2 (43.6–52.8) 47.3 (43.0–51.5) 38.9 (34.3–43.7)
Japan 37.2 (32.9–41.5) 33.8 (29.6–38.0) 34.6 (30.4–38.8) 27.8 (23.9–31.7)
Partial recovery Australia 9.9 (7.6–12.8) 9.0 (6.9–11.7) 12.8 (10.0–16.2) 19.1 (15.9–22.7)
Japan 14.8 (11.7–17.9) 15.4 (12.2–18.6) 13.2 (10.2–16.2) 13.6 (10.6–16.6)
Partial recovery with relapse Australia 5.8 (4.0–8.2) 9.5 (7.1–12.6) 11.9 (9.5–14.9) 21.4 (17.9–25.5)
Japan 37.4 (33.1–41.7) 40.6 (36.3–44.9) 42.2 (37.9–46.5) 52.8 (48.4–57.2)
No improvement Australia 0.1 (0.0–0.8) 0.3 (0.1–1.6) 0.3 (0.0–2.3) 0.8 (0.3–2.0)
Japan 2.4 (1.1–3.7) 1.2 (0.2–2.2) 2.4 (1.1–3.7) 1.6 (0.5–2.7)
Get worse Australia 0.5 (0.1–1.9) 0.2 (0.0–1.3) 0.2 (0.0–2.4) 0.3 (0.1–1.3)
Japan 0.0 (0.0–0.0) 0.2 (0.0–0.6) 0.2 (0.0–0.6) 0.0 (0.0–0.0)
Don't know Australia 2.8 (1.7–4.7) 3.1 (1.9–4.9) 2.8 (1.5–4.9) 3.6 (2.3–5.7)
Japan 0.8 (0.0–1.6) 3.0 (1.5–4.5) 3.0 (1.5–4.5) 1.4 (0.4–2.4)
Without professional help
Full recovery Australia 0.6 (0.2–1.9) 0.4 (0.1–1.7) 0.6 (0.2–1.6) 0.1 (0.0–1.2)
Japan 0.6 (0.0–1.3) 0.8 (0.0–1.6) 0.0 (0.0–0.0) 0.4 (0.0–1.0)
Full recovery with relapse Australia 2.2 (1.2–4.2) 1.5 (0.8–2.8) 0.7 (0.2–2.3) 0.5 (0.1–1.7)
Japan 4.2 (2.4–6.0) 2.6 (1.2–4.0) 2.6 (1.2–4.0) 1.2 (0.2–2.2)
Partial recovery Australia 2.8 (1.6–4.8) 2.5 (1.4–4.6) 0.9 (0.4–2.1) 1.2 (0.5–2.7)
Japan 3.8 (2.1–5.5) 3.8 (2.1–5.5) 2.6 (1.2–4.0) 1.2 (0.2–2.2)
Partial recovery with relapse Australia 9.9 (7.4–13.0) 6.7 (4.7–9.3) 3.7 (2.5–5.6) 1.2 (0.5–3.0)
Japan 12.4 (9.5–15.3) 11.2 (8.4–14.0) 8.6 (6.1–11.1) 4.4 (2.6–6.2)
No improvement Australia 19.3 (16.2–22.9) 14.2 (11.4–17.5) 14.7 (11.7–18.3) 19.4 (15.9–23.4)
Japan 29.8 (25.8–33.8) 26.4 (22.5–30.3) 33.6 (29.4–37.8) 39.4 (35.1–43.7)
Get worse Australia 63.9 (59.4–68.1) 72.0 (67.9–75.9) 78.0 (74.2–81.4) 76.8 (72.2–80.8)
Japan 47.6 (43.2–52.0) 50.8 (46.4–55.2) 49.0 (44.6–53.4) 53.2 (48.8–57.6)
Don't know Australia 1.3 (0.6–2.7) 2.7 (1.4–5.0) 1.3 (0.6–2.9) 0.9 (0.3–2.5)
Japan 1.6 (0.5–2.7) 4.4 (2.6–6.2) 3.6 (2.0–5.2) 0.2 (0.0–0.6)
Where professional help was not received, Australians were most likely to believe the person would get worse. This was also the most common response in Japan, although it was less frequently endorsed than in Australia.
Discussion
Below we discuss the results from each country separately and then compare the results from the two countries.
Public beliefs in Australia
The Australian public showed a relatively high level of recognition of depression in the vignettes and this rate was much improved on a similar Australian survey carried out in 1995 [13]. Recognition of the schizophrenia vignettes was not as good, but has also improved since the earlier survey. There was a generally low use of generic lay terms such as "stress", "psychological/mental/emotional problems" and "nervous breakdown". An exception is the generic term "mental illness" which was used by around a quarter of respondents for the early schizophrenia vignette and by around a third for the chronic schizophrenia vignette.
When asked about people who could help, the Australian public showed a high endorsement of GPs and counselors. Psychiatrists were also highly endorsed for the schizophrenia vignettes, more so than in the earlier Australian survey [13]. For medications, only around half endorsed antidepressants for depression, and around a third endorsed antipsychotics for schizophrenia. There were similarly low rates of endorsement of psychotherapy for depression (around half the population) and admission to a psychiatric ward for schizophrenia (around a third). While these endorsement rates are higher than in the 1995 survey [13], they are still low given that these are standard treatments endorsed by most Australian mental health professionals [22,23]. This gap between public and professional beliefs on medication may limit willingness to accept some recommended interventions.
The Australian public sees a range of lifestyle interventions as likely to be helpful, such as increased physical activity, reading about the problem, getting out and about more, and relaxation training. Some of these interventions have supporting evidence for the treatment of depression [24], but not for schizophrenia. In general, the beliefs of the Australian public about treatment are more positive towards lifestyle interventions than towards medical or psychological interventions.
Dealing with the problem alone was seen as likely to be harmful by most Australians, more so than in the earlier survey [13]. A change in such beliefs may help improve the comparatively low rate of help-seeking observed in a 1997 survey of the Australian public [25]. There was also a general belief that seeking professional help would produce a much better outcome for all disorders portrayed in the vignettes, with better outcomes expected for depression than for schizophrenia.
Public beliefs in Japan
When asked what was wrong with the people portrayed in the vignettes, the Japanese public recognized that there was a mental health problem, but tended to use non-psychiatric labels. Fewer than a fifth used the term "schizophrenia" for the early schizophrenia vignette, but this increased to a third for the chronic schizophrenia vignette. Previous research with Japanese teachers has also shown a low rate of using the term "schizophrenia" in relation to a vignette [26]. This term has very negative connotations in Japan [27], leading psychiatrists to be reluctant to give a diagnosis of schizophrenia to their patients [28]. There have also been proposals to replace the term with a more socially acceptable one [27].
When asked about methods of help, the Japanese public most frequently endorsed counselors, close family and friends. The belief in the helpfulness of counselors has been reported previously in a study of Japanese teachers [26]. Psychiatrists and social workers also received a high level of endorsement for the chronic schizophrenia vignette. Dealing with the problem alone was seen to be harmful by more people than helpful, but still around a fifth of the population saw it as helpful. Previous research has shown that there is considerable stigma on seeking help in Japan and a strong desire for confidentiality, leading some people to seek services far away from their place of residence and to pay cash rather than use health insurance (which would lead to their possible identification) [29]. More people saw psychiatric drugs such as antidepressants and antipsychotics as helpful than harmful, but endorsement of these treatments was not high, consistent with results in other developed countries. Admission to a psychiatric ward was in general not viewed favorably, but was more accepted for chronic schizophrenia. This finding is interesting given the high rate of psychiatric hospitalization in Japan compared to other countries. Dietary changes were also seen very negatively; the reason for this is not clear.
When asked about outcomes, the Japanese public was most likely to believe in partial recovery if the person received professional help, but that the person would get worse if there was no help. There is therefore a general belief that professional help would be beneficial, with better outcomes expected for depression than for schizophrenia.
Comparison of Australia and Japan
When asked what was wrong with the people portrayed in the vignettes, the Australian public was generally more likely than the Japanese public to use the term "depression" and less likely to use non-psychiatric terms. However, for the schizophrenia vignettes, Australians used the term "schizophrenia" more often for the early schizophrenia vignette than for the chronic vignette, whereas the Japanese applied it more to the chronic vignette. The Japanese public may be reserving psychiatric labels only for the more severe cases of mental disorder.
When asked about sources of help, the most striking difference between the two countries was in attitudes to GPs, who were seen by the Australians as likely to be helpful much more often than by the Japanese. This difference may be related to the nature of the health systems in the two countries. In Australia, GPs are seen as the first point of contact for any health problem and the gateway to other services. There have been efforts in Australia to improve the training of GPs in mental health and to encourage the public to seek help from GPs for mental disorders. In Japan, the role of the GP is an extremely important issue as well, but their interest in psychiatric treatment at the moment is not necessarily great, and it is difficult to say that their ability to diagnose psychiatric patients correctly is sufficient. Currently, calls are growing regarding the importance of re-educating GPs in psychiatric medicine, and in future, if this is realized, GPs should be able to play a suitable role.
Compared to Australians, the Japanese more often endorse the helpfulness of close family and dealing with the problem on one's own, perhaps reflecting a cultural difference in the extent to which mental health issues should be a private matter or perhaps a lack of knowledge. Similarly, the Japanese were much less positive about receiving information from a health educator, but had similar beliefs to Australians about the helpfulness of more private information sources, such as books and the internet.
Australians were much more positive than the Japanese about lifestyle interventions such as diet, physical activity, getting out more, relaxation, and cutting out alcohol. The Australians were also more likely than the Japanese to see psychiatric medications as harmful, particularly tranquillizers and sleeping pills.
The Japanese had a lower rate of endorsing clergy as likely to be helpful, although this was not a highly endorsed form of help in Australia either. In Australia, it is an accepted role of the clergy to support members of their church at times of crisis. In Japan, it has been extremely rare for a conventional clergyman to play any sort of active role in the medical field. However, traditional local shamans, and groups that have been viewed as the so-called "new religions", often offer incantations and prayers for patients, and in many cases mental illness has been described as some sort of "curse", which can be "swept away" by performing suitable rites.
There were also some similarities between the two countries. Both gave a high rate of endorsement of the helpfulness of counselors. In Australia, counselors are not a registered profession and vary greatly in their training. They are seen as providing a supportive role – someone who will listen to problems and give advice. In Japan, counselors are seen to be associated with "mind care", but are not a common profession. The term "counseling" is also used broadly to cover supportive talking relationships that might be provided by people who are not counselors.
Another similarity between the two countries was the predominantly negative view of psychiatric wards and ECT. The negative view of psychiatric wards in both countries is interesting given the much greater use of this intervention in Japan than Australia.
When asked about long-term outcomes, the Japanese were more likely to believe in partial recovery following treatment, while the Australians were more optimistic about full recovery. On the other hand, the Australians were more negative about outcomes without treatment. In both countries, outcomes for depression were seen more positively than outcomes for schizophrenia. A factor that might produce these differences between the two countries is exposure to people with mental disorders. The more hospital-based system in Japan might mean that the public have less contact with people in various stages of recovery.
Limitations
Both the Australian and Japanese surveys had some methodological limitations. In the Australian survey, the sample was a national one, but there was a considerable amount of non-contact and refusal. In the Japanese survey, the sample was not truly national, but nevertheless covered the country broadly. The representativeness of the sample for the country as a whole is unknown, but is likely to be adequate for investigation of major cross-national differences. Information on refusal was not collected.
Another limitation concerns the problems of making cross-national comparisons between two very different cultures. Because the interview was designed to suit the Australian public, it may not have been optimal for the Japanese. Although we tried to make the survey interviews as close as possible, there will inevitably be subtleties of meaning and cultural factors operating within a structured household survey that could affect the results in unknown ways. For example, there could be differences in the willingness to use various response categories, the acceptability of expressing certain views to an interviewer, or the comparability of interventions or services that may be translated as equivalent. To avoid this limitation as far as possible, we have focused on the broad pattern of responses between countries, rather than small statistically significant differences in percent frequencies.
Some of the interventions listed in the interview are not widely available and respondents would not have had either direct or indirect experience on which to base their beliefs. For example, in Australia, receiving information from a health educator or consulting an expert via email would be rare interventions. Similarly, in Japan, help from priests and naturopaths is rare, while telephone counseling is uncommon but increasing.
The conclusions reached here are limited to the quantitative data collected in community surveys. Future work on cross-cultural comparisons of mental health literacy would benefit from associated qualitative research to document the cultural differences that underpin any quantitative differences found.
Conclusion
Comparing the two countries, some broad themes are apparent. The Japanese public could be described as more reluctant to use psychiatric labels, particularly with milder disorders, and to be less likely to discuss mental disorders with others outside the family. They generally believe in the benefits of treatment, but are not optimistic about full recovery. By contrast, the Australian public has adopted psychiatric labels, particularly "depression", more than the Japanese. They are more positive about the benefits of seeking professional help, but show a strong preference for lifestyle interventions and tend to be negative about some psychiatric medications. Belief in psychological interventions such as counseling and psychotherapy is similar in the two countries.
In subsequent reports from these surveys we intend to examine differences between the countries in beliefs about causes of mental disorders and in stigmatizing attitudes. These data will allow us to see whether the greater reluctance in Japan to label mental disorders and to expose them outside the family is associated with more negative attitudes or with stigmatizing causal explanations.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
AFJ co-designed the Australian survey, analyzed the Australian data, and co-wrote the manuscript.
YN provided overall supervision of the research and provided comments on the manuscript.
HC co-designed the Australian survey and provided comments on the manuscript.
YK provided specific guidance on the Japanese survey, including participation in survey interviewer training, and co-wrote the manuscript.
KMG co-designed the Australian survey and provided comments on the manuscript.
YW was involved in coordination between the Japanese and Australian surveys and co-wrote the manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
This study is part of the Australia-Japan Partnership, which is an agreement between the governments of the two countries for joint projects in areas of health. The Australian research team wishes to thank the Australian Department of Health and Aging, a National Health and Medical Research Council Program Grant, and "beyondblue: the national depression initiative" for support of the Australian survey and Kelly Blewitt for research assistance. The Japanese research team wishes to thank the Ministry of Health, Labor, and Welfare for the Health and Labor Science Research Grants (Research on Psychiatric and Neurological Diseases and Mental Health), which allowed us to conduct our research.
==== Refs
Andrews G Andrews G, Henderson S Meeting the unmet need with disease management Unmet Need in Psychiatry 2000 Cambridge: Cambridge University Press 11 36
Jorm AF Mental health literacy: public knowledge and beliefs about mental disorders Br J Psychiatry 2000 177 396 401 11059991 10.1192/bjp.177.5.396
Jorm AF Angermeyer M Katschnig H Andrews G, Henderson S Public knowledge of and attitudes to mental disorders: a limiting factor in the optimal use of treatment services Unmet Need in Psychiatry 2000 Cambridge: Cambridge University Press 399 413
Lauber C Nordt C Falcato L Rössler W Do people recognise mental illness? Factors influencing mental health literacy Eur Arch Psychiatry Clin Neurosci 2003 253 248 251 14504994 10.1007/s00406-003-0439-0
Magliano L Fiorillo A De Rosa C Malangone C Maj M Beliefs about schizophrenia in Italy: a comparative nationwide survey of the general public, mental health professionals, and patients' relatives Can J Psychiatry 2004 49 322 30 15198469
Brandli H Guimon J, Fischer W, Sartorius N The image of mental illness in Switzerland The Image of Madness: The Public Facing Mental Illness and Psychiatric Treatment 1999 Basel: Karger 29 37 15133598
Hillert A Sandmann J Ehmig SC Weisbecker H Kepplinger HM Benkert O Guimon J, Fischer W, Sartorius N The general public's cognitive and emotional perception of mental illness: an alternative to attitude-research The Image of Madness: The Public Facing Mental Illness and Psychiatric Treatment 1999 Basel: Karger 56 71
Angermeyer MC Matschinger H Guimon J, Fischer W, Sartorius N Social representations of mental illness among the public The Image of Madness: The Public Facing Mental Illness and Psychiatric Treatment 1999 Basel: Karger 20 28
Herran A Vazquez-Barquero JL Dunn G Patients' attributional style is important factor BMJ 1999 318 1558 10356032
Kessler D Lloyd K Lewis G Gray DP Cross sectional study of symptom attribution and recognition of depression and anxiety in primary care BMJ 1999 318 436 439 9974461
Bowers J Jorm AF Henderson S Harris P General practitioners' detection of depression and dementia in elderly patients Med J Aust 1990 153 192 196 2388603
Jacob KS Bhugra D Lloyd KR Mann AH Common mental disorders, explanatory models and consultation behaviour among Indian women living in the UK J R Soc Med 1998 91 66 71 9602740
Jorm AF Korten AE Jacomb PA Christensen H Rodgers B Pollitt P "Mental health literacy": a survey of the public's ability to recognise mental disorders and their beliefs about the effectiveness of treatment Med J Aust 1997 166 182 186 9066546
Croghan TW Tomlin M Pescosolido BA Schnittker J Martin J Lubell K Swindle R American attitudes toward and willingness to use psychiatric medications J Nerv Ment Dis 2003 191 166 174 12637843 10.1097/00005053-200303000-00005
Angermeyer MC Däumer R Matschinger H Benefits and risks of psychotropic medication in the eyes of the general public: results of a survey in the Federal Republic of Germany Pharmacopsychiatry 1993 26 114 120 8234450
Priest RG Vize C Roberts A Roberts M Tylee A Lay people's attitudes to treatment for depression: results of opinion poll for Defeat Depression Campaign just before its launch BMJ 1996 313 858 859 8870574
Fischer W Goerg D Zbinden E Guimon J Guimon J, Fischer W, Sartorius N Determining factors and the effects of attitudes towards psychotropic medication The Image of Madness: The Public Facing Mental Illness and Psychiatric Treatment 1999 Basel: Karger 162 186
Hugo CJ Boshoff DEL Traut A Zungu-Dirwayi N Stein DJ Community attitude toward and knowledge of mental illness in South Africa Soc Psychiatry Psychiatr Epidemiol 2003 38 715 719 14689176 10.1007/s00127-003-0695-3
McKeon P Carrick S Public attitudes to depression: a national survey Ir J Psychol Med 1991 8 116 121
Angermeyer MC Matschinger H Public attitude towards psychiatric treatment Acta Psychiatr Scand 1996 94 326 336 9124079
Mizuno M Murakami M Lefley HP, Johnson DL Differences in strategies for implementing community-based psychiatry in Japan Family Interventions in Mental Illness 2001 Westport Connecticut: Praeger 185 192
Jorm AF Korten AE Jacomb PA Rodgers B Pollitt P Christensen H Henderson S Helpfulness of interventions for mental disorders: beliefs of health professionals compared to the general public Br J Psychiatry 1997 171 233 237 9337975
Caldwell TM Jorm AF Mental health nurses' beliefs about interventions for schizophrenia and depression: a comparison with psychiatrists and the public Aust N Z J Psychiatry 2000 34 602 611 10954391 10.1046/j.1440-1614.2000.00750.x
Jorm AF Christensen H Griffiths KM Rodgers B Effectiveness of complementary and self-help treatments for depression Med J Aust 2002 176 S84 S96 12065003
Andrews G Henderson S Hall W Prevalence, comorbidity, disability and service utilisation. Overview of the Australian National Mental Health Survey Br J Psychiatry 2001 178 145 153 11157427 10.1192/bjp.178.2.145
Kurumatani T Ukawa K Kawaguchi Y Miyata S Suzuki M Ide H Seki W Chikamori E Hwu HG Liao SC Edwards GD Shinfuku N Uemoto M Teachers' knowledge, beliefs and attitudes concerning schizophrenia: a cross-cultural approach in Japan and Taiwan Soc Psychiatry Psychiatr Epidemiology 2004 39 402 409 10.1007/s00127-004-0758-0
Kim Y Berrios GE Impact of the term schizophrenia on the culture of ideograph: the Japanese experience Schizophr Bull 2001 27 181 185 11354585
McDonald-Scott P Machizawa S Satoh H Diagnostic disclosure: a tale of two cultures Psychol Med 1992 22 147 157 1574550
Sugiura T Sakamoto S Kijima N Kitamura F Kitamura T Stigmatizing perception of mental illness by Japanese students: comparison of different psychiatric disorders J Nerv Ment Dis 2000 188 239 242 10790003 10.1097/00005053-200004000-00009
| 16004615 | PMC1177951 | CC BY | 2021-01-04 16:27:56 | no | BMC Med. 2005 Jul 9; 3:12 | utf-8 | BMC Med | 2,005 | 10.1186/1741-7015-3-12 | oa_comm |
==== Front
BMC MedBMC Medicine1741-7015BioMed Central London 1741-7015-3-91593263810.1186/1741-7015-3-9Research ArticleThe impact of glucose-insulin-potassium infusion in acute myocardial infarction on infarct size and left ventricular ejection fraction [ISRCTN56720616] van der Horst Iwan CC [email protected] Jan Paul [email protected] 't Hof Arnoud WJ [email protected] Stoffer [email protected] Kor [email protected] Jan CA [email protected] Jan-Henk E [email protected] AT Marcel [email protected] Maarten WN [email protected] Harry [email protected] Boer Menko-Jan [email protected] Felix [email protected] Department of Cardiology, Thoraxcenter, University Medical Center Groningen, Hanzeplein 1, 9700 RB Groningen, the Netherlands2 Department of Cardiology, Isala Klinieken, locatie Weezenlanden, Groot Wezenland 20, 8011 JW Zwolle, the Netherlands3 Department of Nuclear Medicine, Isala Klinieken, locatie Weezenlanden, Groot Wezenland 20, 8011 JW Zwolle, the Netherlands4 Department of Clinical Chemistry, Isala Klinieken, locatie Weezenlanden, Groot Wezenland 20, 8011 JW Zwolle, the Netherlands5 Department of Surgery, University Medical Center Groningen, Hanzeplein 1, 9700 RB, Groningen, the Netherlands2005 2 6 2005 3 9 9 15 10 2004 2 6 2005 Copyright © 2005 van der Horst et al; licensee BioMed Central Ltd.2005van der Horst 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
Favorable clinical outcomes have been observed with glucose-insulin-potassium infusion (GIK) in acute myocardial infarction (MI). The mechanisms of this beneficial effect have not been delineated clearly. GIK has metabolic, anti-inflammatory and profibrinolytic effects and it may preserve the ischemic myocardium. We sought to assess the effect of GIK infusion on infarct size and left ventricular function, as part of a randomized controlled trial.
Methods
Patients (n = 940) treated for acute MI by primary percutaneous coronary intervention (PCI) were randomized to GIK infusion or no infusion. Endpoints were the creatinine kinase MB-fraction (CK-MB) and left ventricular ejection fraction (LVEF). CK-MB levels were determined 0, 2, 4, 6, 24, 48, 72 and 96 hours after admission and the LVEF was measured before discharge.
Results
There were no differences between the two groups in the time course or magnitude of CK-MB release: the peak CK-MB level was 249 ± 228 U/L in the GIK group and 240 ± 200 U/L in the control group (NS). The mean LVEF was 43.7 ± 11.0 % in the GIK group and 42.4 ± 11.7% in the control group (P = 0.12). A LVEF ≤ 30% was observed in 18% in the controls and in 12% of the GIK group (P = 0.01).
Conclusion
Treatment with GIK has no effect on myocardial function as determined by LVEF and by the pattern or magnitude of enzyme release. However, left ventricular function was preserved in GIK treated patients.
==== Body
Background
It has been suggested that glucose-insulin-potassium (GIK) infusion in acute myocardial infarction (MI) has clinical benefit [1-4]. Both animal studies and early studies in patients to investigate the influence of GIK on infarct size have shown conflicting results [5-14]. Experimental studies on the effect of GIK on preservation of left ventricular function, as determined by hemodynamic parameters, demonstrated a beneficial effect [15,16]. Furthermore, in a small study in patients with acute MI treated by thrombolysis, there was a significant improvement in left ventricular function over a 10-day period [17]. Recently, a small randomized trial of 37 patients has suggested that the addition of GIK to primary percutaneous coronary intervention (PCI) has a beneficial effect [18]. Large studies of GIK in the era of reperfusion yielded some information about effects on myocardial function [1,17,19]. The Polish Glucose-Insulin-Potassium (Pol-GIK) trial with 954 patients, using low-dose GIK, found no difference between median creatinine kinase (CK) levels (920 IU/L in GIK patients versus 925 IU/L in controls) [19]. Recently, in the REeValuation of Intensified Venous metabolic support for Acute infarct size Limitation (REVIVAL) trial with 312 MI patients, the combination of GIK and primary PCI had no effect on myocardial salvage [20]. In the Glucose-Insulin-Potassium Study (GIPS) with GIK infusion as adjunctive therapy to primary PCI in acute MI, the 30-day mortality was not significantly reduced in the overall population [4]. In a predefined subgroup of patients without heart-failure at admission, mortality was 1.2% in GIK patients versus 4.2% in controls (P = 0.01). In the present study we were able to determine the effect of GIK on cumulative enzyme release and left ventricular ejection fraction.
Methods
Study population
An outline of the study and 30-day clinical follow-up has been reported [4]. All consecutive patients with symptoms consistent with acute MI of >30 minutes duration, presenting within 24 hours after the onset of symptoms and with ST-segment elevation more than 1 mm (0.1 mV) in two or more contiguous leads on the electrocardiogram, were evaluated for inclusion in this study. Patients were excluded if they had been pre-treated by thrombolysis or when an illness associated with markedly restricted life expectancy was present. Before randomization, age, gender, history of coronary artery bypass grafting (CABG), previous PCI, stroke and MI, existence of diabetes mellitus, smoking status, heart rate, arterial pressure, body mass index, Killip class, electrocardiographic site of infarction, time of onset of symptoms, and time of hospital admission in both the referring hospital and our hospital were recorded. The research protocol was reviewed and approved by the medical ethics committee of our hospital, and patients were included after informed consent.
Treatment protocol
Patients were randomly assigned to either GIK infusion or no infusion. Assignments to the treatment groups were made with the use of a computerized randomization program. In the GIK group, a continuous infusion of 80 mmol potassium chloride in 500 mL 20% glucose was given at a rate of 3 mL/kg/hour through a peripheral venous line. A continuous infusion of short-acting insulin (50 units Actrapid HM-Novo Nordisk, Copenhagen, Denmark) in 50 ml 0.9% sodium chloride was also initiated using a pump (Perfusor-FM, B. Braun, Melsungen, Germany). The baseline insulin-infusion dose and hourly adjustments of the dose to obtain blood-glucose levels between 7.0 and 11.0 mmol/l were determined using an algorithm based on measurements of whole-blood glucose concentration. The infusion was started as soon as possible after randomization and continued for at least 8 hours with a maximum of 12 hours.
All patients went to the catheterization laboratory as soon as possible after admission and randomization. In the laboratory, both coronary arteries were visualized. PCI was performed using standard techniques if the coronary anatomy was suitable for angioplasty. Additional treatment consisted of unfractionated intravenous heparin, nitroglycerin and aspirin. After sheath removal, low-molecular-weight heparin was given for 1 to 3 days. Before discharge, additional pharmacological treatment consisting of aspirin, clopidogrel, coumarins, beta-blockers, angiotensin-converting enzyme inhibitors and statins was initiated according to the guidelines of the American College of Cardiology/American Heart Association. Enrollment of patients began in April 1998 and ended in September 2001.
Measurements
Enzymatic infarct size was estimated by serial measurements of the CK-MB fraction. The first measurement was taken as soon as possible after admission. Thereafter frequent CK-determinations were performed according to a schedule that called for 4 to 8 measurements in the first 96 hours. To accommodate the problem that exact predefined times for blood sampling were not always followed in practice, the actual times of sampling were recorded, expressed as minutes after the moment of randomization. To allow optimal comparison of the time courses of CK-MB levels and to best approximate the area under the CK-MB curves, we developed an algorithm. This algorithm estimated the time-course of CK-MB for each patient before calculating the patient group means. According to the known kinetics of CK-MB – a rapid rise and a much slower decrease – the following characteristic time points with accompanying time intervals were defined: 0 h (interval from 60 min before to 45 min after randomisation); 2 h (45 min to 3 h); 4 h (3 h to 5 h); 6 h (5 h to 12 h); 24 h (12 h to 36 h); 48 h (36 h to 60 h); 72 h (60 h to 84 h) and 96 h (84 h to 108 h). The algorithm interpolated the CK-MB level for each patient based on the precise times at which the measurements were performed. Next, it determined for each of the predefined intervals whether an actual measurement had been performed in that interval. If no measurement was available for a particular interval, a "not available" value was generated. This avoided inappropriately interpolated values. If one or more measurements were available for an interval, the CK-MB value for the time point with the corresponding interval was calculated from the area under the interpolated curve divided by the interval duration. CK-MB was determined enzymatically using a Hitachi 717 automatic analyzer at 30 degrees Celsius according to the International Federation of Clinical Chemistry (IFCC) recommendation. High enzyme release was defined as a peak CK-MB above the 75% percentile (i.e. the highest quartile). Time to peak CK-MB was determined.
Left ventricular ejection fraction (LVEF) was measured before discharge by radionuclide ventriculography or by echocardiography. Radionuclide ventriculography was performed by the multiple gated equilibrium method following the labeling of the patient's red blood cells with technetium 99mpertechnate. A General Electric 300 gamma camera with a low-energy all-purpose parallel-hole collimator was used. The global ejection fraction was calculated with a Star View computer (General Electric, Wisconsin, USA) using the fully automatic PAGE program. LVEF as assessed by two-dimensional transthoracicechocardiography was reported as a descriptive grade of function, using subjective visual assessment by two independent observers who were unaware of the randomisation code. This approach is less time-consuming than other methods, such as Simpson's rule or wall motion index score, but it can be at least as accurate as other methods [21]. Echocardiography was performed by observers who were unaware of the randomization outcome and clinical data. A LVEF ≤ 30% was defined as a poor left ventricular function before the analysis.
Statistical analysis
All analyses were by intention to treat. Predefined endpoints were pattern and magnitude of CK-MB release, peak CK-MB, high enzyme release (defined as a peak enzyme release in the highest quartile); mean LVEF, and a LVEF lower than or equal to 30%. To allow comparison of the patterns of CK-MB release, the CK-MB values 0, 2, 4, 6, 24, 48, 72 and 96 hours after admission were calculated for each individual patient with linear interpolation. Differences between group means were assessed with the two-tailed Student's t-test. Chi-square analysis was used to test differences between proportions. The Cox proportional-hazards regression model was used to calculate relative risks adjusted for differences in baseline characteristics. To predict the independent association between treatment with GIK and a peak CK-MB release above the highest quartile or LVEF ≤ 30%, a multivariate logistic regression analysis was performed. A two-tailed P value <0.05 was considered statistically significant. The Statistical Package for the Social Sciences (SPSS Inc., Chicago, IL, USA) version 11.5 was used for all statistical analyses.
Results
Study population
Of the 940 patients who underwent randomization, 476 were assigned to receive GIK infusion. There were no differences in the baseline characteristics of the two treatment groups table 1. The mean time between onset of symptoms and admission was 150 min, and 75% of the patients were admitted within 225 min. The interquartile range (IQR) in the GIK group was 100 to 215 min and in the control group 105 to 234 min. After coronary angiography, 860 patients (90.5%) underwent PCI, 38 patients (4%) were referred for CABG within 7 days, and 42 patients (4.5%) were treated conservatively. Thirty-day mortality was 4.8% in the GIK patients and 5.8% in the control patients [4].
Table 1 Baseline clinical and demographic characteristics and initial therapies
Characteristics GIK group Control group
Age, years ± SD 60 ± 12 61 ± 12
Men 351 (73.7) 368 (79.3)
Referred patients 201 (42.3) 180 (38.8)
Body mass index 26.7 ± 3.8 27.0 ± 4.0
Previous MI 52 (10.9) 53 (11.4)
Previous PCI 22 (4.7) 24 (5.2)
Previous CABG 14 (2.9) 12 (2.6)
History of stroke 17 (3.6) 15 (3.2)
Diabetes mellitus 50 (10.5) 49 (10.6)
Hypertension 134 (28.2) 130 (28.0)
Dyslipidaemia 94 (19.7) 95 (20.5)
Currently smoker 225 (47.3) 237 (51.1)
Positive family history 195 (41.0) 179 (38.6)
Time to admission, min (IQR)* 150 (100–215) 150 (105–234)
Door to balloon time, min (IQR) 45 (31–64) 48 (34–69)
Killip class 1 426 (89.5) 430 (92.7)
Killip class 2 24 (5.0) 14 (3.0)
Killip class 3 14 (2.9) 14 (3.0)
Killip class 4 12 (2.5) 6 (1.3)
Anterior MI 250 (52.9) 224 (49.1)
Multi-vessel disease 249 (52.3) 242 (52.2)
PCI 436 (91.6) 424 (91.4)
Stent ** 255 (58.5) 236 (55.7)
GP IIb/IIIa receptor blocker ** 96 (22.1) 109 (25.7)
In-hospital CABG 19 (4.0) 19 (4.1)
TIMI 3 flow 459 (96.4) 435 (93.8)
Successful reperfusion‡ 394 (82.8) 384 (82.8)
Intra-aortic balloon pump 45 (9.5) 42 (9.1)
Mechanical ventilation 12 (2.5) 4 (0.9)
Data are numbers (%) unless otherwise indicated. MI = myocardial infarction. IQR = interquartile range. PCI = percutaneous coronary intervention. CABG = coronary artery bypass grafting. * Time to admission denotes time between onset of symptoms and admission. † In-hospital CABG = CABG defined as patients treated initially conservatively, followed by elective CABG within 7 days. ‡ Successful reperfusion denotes TIMI 3 flow and blush grade 2 or 3. ** Percentage defined as the percentage of the patients who underwent PCI.
Enzyme release
Data on the effect of GIK on enzymatic infarct size (CK-MB) were available in 923 patients (98%), 470 patients treated with GIK and 462 control patients. A mean of 6.9 CK-MB determinations per patient were available. As indicated in figure 1, there was no difference in the pattern or magnitude of CK-MB release between the two groups. Mean peak CK-MB ± SD was 249 ± 228 U/L in the GIK group and 240 ± 200 U/L in the control group (NS). In GIK patients the median time to peak CK-MB was 450 min (IQR 300–733) compared to 468 min (IQR 285–738) in the control patients (NS). High enzyme release (i.e. peak CK-MB >349 U/L) was found in 111 patients (24%) in the GIK group versus 121 patients (25%) in the control group (P = 0.6). Differences between patients with peak CK-MB release below and above 349 U/L are indicated in table 2. Characteristics related to high enzyme release were Killip class ≥ 2, anterior MI and TIMI 0 flow before PCI.
Figure 1 CK-MB release after admission in the GIK group and control group. CK-MB release after admission in the GIK group (unbroken line) and control group (dashed line). Means are indicated for 0, 2, 4, 6, 24, 48, 72 and 96 hours. No significant difference was observed at any time point.
Table 2 Comparison of patients with a peak CK-MB release beneath the highest quartile and patients with a CK-MB release above the highest quartile
Characteristic High enzyme release (N = 232) Low enzyme release (N = 708) P-value
Age, years ± SD 59 ± 12 61 ± 12 0.17
Men 179 (77) 540 (76) 0.86
Hypertension 64 (28) 200 (28) 0.87
Previous MI 24 (10) 81 (11) 0.72
Diabetes mellitus 28 (12) 71 (10) 0.39
Killip class 1 192 (83) 664 (94) <0.001
Anterior MI 161 (69) 313 (44) <0.001
TIMI 0 before PCI 189 (82) 377 (53) <0.001
TIMI 3 after PCI 219 (94) 675 (95) 0.6
GIK treatment 121 (52) 355 (50) 0.6
Data are numbers (%) unless otherwise indicated. MI = myocardial infarction. TIMI = Thrombolysis in Myocardial Infarction. PCI = percutaneous coronary intervention.
Left ventricular ejection fraction
LVEF measurements were available in 419 patients (88%) treated with GIK and in 404 (87%) of the controls. Mean time between admission and LVEF measurement was 2.65 days in the GIK group and 2.5 days in the control group (IQR 2 days in both groups). The mean LVEF was 43.7 ± 11.0 % in the GIK group and 42.4 ± 11.7% in the control group (P = 0.12). Of the 823 patients with a measured LVEF before discharge, a total of 125 (15%) had an ejection fraction lower than or equal to 30%. Differences between patients with LVEF ≤ 30% and those with a higher LVEF are summarized in table 3. Treatment with GIK was associated with a lower frequency of a LVEF ≤ 30% (51 patients (12%) compared to 74 patients (18%), P = 0.01). Other important characteristics associated with LVEF ≤ 30% were age, diabetes mellitus, Killip class, infarct location and TIMI flow after PCI. To identify independent predictors of a low left ventricular ejection fraction (LVEF ≤ 30%) all variables associated with LVEF ≤ 30% in univariate analysis were used in multivariate analysis; see table 3. After adjustment, patients treated with GIK had a RR of 0.53 (95% confidence interval 0.35–0.81, P = 0.01) for LVEF ≤ 30% compared to controls. In patients with Killip class 1 at admission the mean LVEF was 44.1 ± 10.9 % in the GIK group versus 42.8 ± 11.5% in the control group (P = 0.12). In these Killip class 1 patients randomized to GIK, a LVEF ≤ 30% was less often observed (45 patients (12%) versus 65 patients (17%), P = 0.03).
Table 3 Comparison of patients with a low Left Ventricular Ejection Fraction (LVEF ≤ 30%) with those with preserved Left Ventricular Ejection Fraction (LVEF >30%)
Characteristic Low LVEF (N = 125) Preserved LVEF (N = 698) P-value
Age, years ± SD 63 ± 12 59 ± 12 0.008
Men 98 (78) 537 (77) 0.7
Hypertension 43 (34) 182 (26) 0.05
Previous MI 18 (14) 63 (9) 0.06
Diabetes mellitus 29 (23) 57 (8) 0.008
Killip class 1 109 (87) 660 (95) 0.002
Anterior MI 115 (92) 291 (42) <0.001
TIMI 0 before PCI 40 (32) 291 (42) 0.04
TIMI 3 after PCI 114 (91) 677 (97) 0.002
GIK treatment 51 (41) 368 (53) 0.01
Data are numbers (%) unless otherwise indicated. MI = myocardial infarction. TIMI = Thrombolysis in Myocardial Infarction. PCI = percutaneous coronary intervention.
In the overall population the mean peak CK-MB ± SD was 463 ± 318 U/L in patients with LVEF ≤ 30% versus 215 ± 166 U/L in patients with LVEF >30% (P < 0.001).
Discussion
In this study, additional treatment with GIK in patients treated with primary PCI for acute myocardial infarction showed no decrease in enzymatic infarct size and no difference in mean LVEF between GIK and control patients was observed. However, preservation of left ventricular function as determined by a LVEF ≤ 30% was more frequently seen in GIK treated patients. One explanation for the discrepancy between the effects on poor and mean left ventricular function is that the number of patients in our analysis was not sufficient to detect an effect on mean left ventricular ejection fraction. Mean LVEF was 1.3% higher in GIK patients, so and several patients may have been prevented from LVEF ≤ 30%.
Since these results are based on an intention-to-treat analysis in a randomized controlled trial of the effects of GIK, we may have underestimated the effect on preservation of myocardial function. In the control group more patients died, and these patients had a high enzyme release and probably a severely impaired LVEF as well. Nevertheless, our data make it unlikely that GIK has a clinically relevant impact of GIK on enzyme release, but at the same time show an important reduction in the number of patients with a LVEF ≤ 30%. A LVEF ≤ 30% exposes these patients to a high risk of developing heart failure during follow-up, even after successful PCI [22]. We also found a relation between high enzyme release and poor left ventricular function. Possibly the relation between GIK and poor LVEF is related to a selection bias in this sub-analysis. On the other hand it can be hypothesized that GIK mediates effects on LVEF, as measured after 2 to 4 days after admission, that are unrelated to enzymatic infarct size.
In our previous study on the effect of GIK on 30-day mortality, we found a beneficial effect of GIK on outcome in patients who had no signs of heart-failure at admission. Patients with Killip class 1 did not have smaller enzymatic infarct sizes in this analysis. However, the CK-MB curve over the first 96 hours after admission in GIK patients lies below that in controls. Mean LVEF is higher in patients treated with GIK compared to control patients, albeit not significantly (P = 0.12). Even as in the overall population in patients with Killip class 1 treatment with GIK was related with a lower rate of LVEF ≤ 30%.
Previous clinical studies with GIK
Several studies have reported enzymatic infarct size in MI patients treated with GIK, mainly in patients without reperfusion therapy [12,14]. Clinical studies in the era of reperfusion provided sparse information [1,19]. In line with our results, the Pol-GIK trial showed that neither the median and maximal CK nor maximal glutamic-oxaloacetic transaminase differed significantly between GIK and control patients [19]. However, the Pol-GIK trial used low-dose GIK and included patients with a lower risk, i.e. no patients with signs of heart failure were included. One study of 32 MI patients treated by thrombolysis found that infusion of GIK resulted in a shorter time to peak CK and a smaller total enzyme release [6]. In line with our results on enzyme release, the REVIVAL trial found no effect of GIK treatment on the myocardial salvage index measured with technetium-99m sestamibi scintigraphy 6 to 8 hours after admission and after 7 to 14 days (0.50 versus 0.48, P = 0.96). In this trial, GIK consisted of 1000 ml 20% glucose with 40 IU of insulin and 64 mmol potassium chloride at a rate of 1.8 ml/hour for 24 hours. However, in the predefined subgroup of patients with diabetes mellitus, a significant difference in salvage index was found (mean difference 0.19, 95% confidence interval 0.01–0.37). In a trial of 37 patients treated with primary PCI, a beneficial effect of GIK on myocardial function was observed [18]. The increase of the LVEF from baseline to 3 months was 12.5% in the GIK group (P = 0.002) versus 6.1% in the placebo group (NS). However, when GIK and placebo were compared directly, the difference between increments in LVEF was not significant. This may have been the result of a combination of small numbers of patients and profound changes in the LVEF during the 3–6 months after PCI for acute MI [23]. In patients with diabetes mellitus and without myocardial ischemia, infusion of insulin and glucose increased the contractile force, i.e. rest LVEF and LVEF during dynamic exercise measured by radionuclide ventriculography [24]. In 32 anterior MI patients, the left ventricular wall motion score index as measured by echocardiography at baseline was 1.87 and it decreased significantly during GIK infusion to 1.76 (P < 0.001) [25].
In January 2005, the 30-day results of the Clinical Trial of Reviparin and Metabolic Modulation in Acute Myocardial Infarction Treatment Evaluation-Estudios Cardiologicos Latinoamerica (CREATE-ECLA) trial (N = 20201) were published [26]. Infusion of high-dose GIK had no beneficial effect on the overall population, with a mortality rate of 10.0% in the GIK group compared to 9.7% in the control group (P = 0.45). The lack of benefit was also observed in several large subgroups that may benefit in particular from GIK, such as patients treated within 4 hours after symptom onset (N = 8361). Only in patients treated by PCI (N = 1838) was a reduction of 30-day mortality observed, from 6.3% in control patients to 4.8% in GIK patients; but this was not significant. Therefore, in patients treated with a combination of GIK and reperfusion, as in our study, the role of metabolic modulation in mortality and myocardial function needs further investigation.
Animal studies
Several studies on animals have assessed the effects of GIK on infarct size in myocardial infarction, but show conflicting results [5,9,11]. A study of pigs showed that hearts treated with GIK had significantly less tissue acidosis, higher wall motion scores and less tissue necrosis [16]. Similar results were obtained in a study of rats, in which the administration of GIK not only before induction of ischemia but also during acute ischemia seemed beneficial for myocardial function [10]. A recent study of diabetic sheep found that GIK improved left ventricular contractility as determined by stroke work efficiency [27].
Mechanisms of action
We found a relationship between higher enzymatic infarct size and low LVEF. Since we observed an effect of GIK only on LVEF, the preservation of left ventricular function could not be declared a reduction of myocardial necrosis. The potential mechanisms by which GIK might be beneficial include metabolic effects, direct hemodynamic effects, improvement of coronary flow and catecholamine-mediated effects. The main effect of GIK infusion is supposed to be improved delivery of glucose to the ischemic myocardium resulting in a suppression of free fatty acids [28-31]. Treatment with glucose and insulin inhibits lipolysis, thereby decreasing circulating free fatty acid levels, and suppresses fatty acid oxidation. Moreover, in the postischemic myocardium, contractile dysfunction may also be caused by impairment of glucose oxidation and cytosolic proton accumulation, so agents that enhance glucose oxidation in the postischemic heart may also improve contractile function [32]. After prolonged periods of low-flow ischemia, functional recovery is mainly determined by the extent of irreversible ischemic damage. After brief episodes of ischemic damage, oxidative metabolism rapidly returns, well before contractile activity is restored. Therefore GIK treatment is likely to be beneficial when started soon after the onset of ischemia. Protection of ischemic cardiac cell membranes may also improve reflow after reperfusion and protect against the no-reflow phenomenon by reducing cell swelling and microvascular compression [33]. These microvascular protective effects on the coronary circulation are probably similar in patients with and without diabetes mellitus [34,35]. GIK infusion improved regional myocardial perfusion and function in segments adjacent to the infarct area in patients recovering from an acute MI [36]. Another beneficial mechanism may be the protective effect of GIK on the cell membrane [37-39]. Insulin promotes tolerance against ischemic cell death via the activation of innate cell-survival pathways in the heart [40]. However, in a small study of patients with MI treated with PCI, administration of GIK did not improve the enzymatic antioxidant reserve [41].
Study limitations
Left ventricular function was measured in 88% of patients. For some patients transferred back to the referring hospitals within 48 hours the measurement was missing. The same goes for patients who died within 48 hours. Therefore the exact relation between LVEF and 30-day mortality cannot be determined. It is unknown whether these missing data have influenced the results. We determined the left ventricular function before discharge and it has been suggested that measurements after long-term follow-up are more predictive[18,23].
Conclusion
High-dose GIK infusion in patients with acute MI treated with primary PCI had no effect on general left ventricular function or the pattern and magnitude of enzyme release. A lower rate of poor left ventricular function, defined as a left ventricular ejection fraction ≤ 30%, was observed in GIK patients.
Abbreviations
GIK glucose-insulin-potassium
MI myocardial infarction
PCI primary coronary intervention
CABG coronary artery bypass grafting
CK creatine kinase
LVEF left ventricular ejection fraction
TIMI thrombolysis in myocardial infarction
IQR interquartile range
RR relative risk
Pol-GIK Polish Glucose-Insulin-Potassium trial
REVIVAL REeValuation of Intensified Venous metabolic support for Acute infarct size Limitation
GIPS Glucose-Insulin-Potassium Study
CREATE-ECLA Clinical Trial of Reviparin and Metabolic Modulation in Acute Myocardial Infarction Treatment Evaluation-Estudios Cardiologicos Latinoamerica
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
IH, JO, and FZ participated in the design of the study. AH, JH, AG, JD, HS, MB and FZ were responsible for patient inclusions. KM was responsible for the laboratory measurements. SR was responsible for the nuclear measurements. IH, JO, and MN were responsible for the statistical analysis. All authors have been involved in drafting the article. IH, MN, and FZ have been involved in revising the article critically. All authors have given final approval of the version to be published.
Figure 2 CK-MB release after admission in the GIK group and control group of Killip class 1 patients. CK-MB release after admission in the GIK group (unbroken line) and control group (dashed line). Means are indicated for 0, 2, 4, 6, 24, 48, 72 and 96 hours. No significant difference was observed at any time point.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
The research of ICC van der Horst was supported by a generous grant from the Netherlands Heart Foundation (99.028).
==== Refs
Diaz R Paolasso EA Piegas LS Tajer CD Moreno MG Corvalan R Isea JE Romero G Metabolic modulation of acute myocardial infarction. The ECLA (Estudios Cardiologicos Latinoamerica) Collaborative Group Circulation 1998 98 2227 2234 9867443
Fath-Ordoubadi F Beatt KJ Glucose-insulin-potassium therapy for treatment of acute myocardial infarction: an overview of randomized placebo-controlled trials Circulation 1997 96 1152 1156 9286943
Malmberg K Ryden L Efendic S Herlitz J Nicol P Waldenstrom A Wedel H Welin L Randomized trial of insulin-glucose infusion followed by subcutaneous insulin treatment in diabetic patients with acute myocardial infarction (DIGAMI study): effects on mortality at 1 year J Am Coll Cardiol 1995 26 57 65 7797776
van der Horst IC Zijlstra F van't Hof AW Doggen CJ de Boer MJ Suryapranata H Hoorntje JC Dambrink JH Gans RO Bilo HJ Glucose-insulin-potassium infusion in patients treated with primary angioplasty for acute myocardial infarction: the glucose-insulin-potassium study: a randomized trial J Am Coll Cardiol 2003 42 784 791 12957421
Bellows SD Kloner RA Glucose-Insulin-Potassium Does Not Reduce Myocardial Infarct Size in an Ischemic/Reperfusion Rabbit Model J Thromb Thrombolysis 1998 5 25 27 10608046
Chaudhuri A Janicke D Wilson MF Tripathy D Garg R Bandyopadhyay A Calieri J Hoffmeyer D Syed T Ghanim H Aljada A Dandona P Anti-inflammatory and profibrinolytic effect of insulin in acute ST-segment-elevation myocardial infarction Circulation 2004 109 849 854 14757687
Dalby AJ Bricknell OL Opie LH Effect of glucose-insulin-potassium infusions on epicardial ECG changes and on myocardial metabolic changes after coronary artery ligation in dogs Cardiovasc Res 1981 15 588 598 7032699
Heng MK Norris RM Singh BN Barratt-Boyes C Effects of glucose and glucose-insulin-potassium on haemodynamics and enzyme release after acute myocardial infarction Br Heart J 1977 39 748 757 328028
Heng MK Norris RM Peter T Nisbet HD Singh BN The effect of glucose-insulin-potassium on experimental myocardial infarction in the dog Cardiovasc Res 1978 12 429 435 719656
Jonassen AK Aasum E Riemersma RA Mjos OD Larsen TS Glucose-insulin-potassium reduces infarct size when administered during reperfusion Cardiovasc Drugs Ther 2000 14 615 623 11300362
Maroko PR Libby P Sobel BE Bloor CM Sybers HD Shell WE Covell JW Braunwald E Effect of glucose-insulin-potassium infusion on myocardial infarction following experimental coronary artery occlusion Circulation 1972 45 1160 1175 5032816
Reduto L Gulotta S Morisson M Effects of glucose-insulin-potassium infusion on ischemic myocardium Clinical Research 1974 22 685A
Rogers WJ Segall PH McDaniel HG Mantle JA Russell RO JrRackley CE Prospective randomized trial of glucose-insulin-potassium in acute myocardial infarction. Effects on myocardial hemodynamics, substrates and rhythm Am J Cardiol 1979 43 801 809 425917
Sundstedt CD Sylven C Mogensen L Glucose-insulin-potassium-albumin infusion in the early phase of acute myocardial infarction – a controlled study Acta Med Scand 1981 210 67 71 7027747
Whitlow PL Rogers WJ Smith LR McDaniel HG Papapietro SE Mantle JA Logic JR Russell RO JrRackley CE Enhancement of left ventricular function by glucose-insulin-potassium infusion in acute myocardial infarction Am J Cardiol 1982 49 811 820 7039286
Zhu P Lu L Xu Y Greyson C Schwartz GG Glucose-insulin-potassium preserves systolic and diastolic function in ischemia and reperfusion in pigs Am J Physiol Heart Circ Physiol 2000 278 H595 H603 10666092
Satler LF Green CE Kent KM Pallas RS Pearle DL Rackley CE Metabolic support during coronary reperfusion Am Heart J 1987 114 54 58 3300232
Castro PF Larrain G Baeza R Corbalan R Nazzal C Greig DP Miranda FP Perez O Acevedo M Marchant E Olea E Gonzalez R Effects of glucose-insulin-potassium solution on myocardial salvage and left ventricular function after primary angioplasty Crit Care Med 2003 31 2152 2155 12973173
Ceremuzynski L Budaj A Czepiel A Burzykowski T Achremczyk P Smielak-Korombel W Maciejewicz J Dziubinska J Nartowicz E Kawka-Urbanek T Piotrowski W Hanzlik J Cieslinski A Kawecka-Jaszcz K Gessek J Wrabec K Low-dose glucose-insulin-potassium is ineffective in acute myocardial infarction: results of a randomized multicenter Pol-GIK trial Cardiovasc Drugs Ther 1999 13 191 200 10439881
Pache J Kastrati A Mehilli J Bollwein H Ndrepepa G Schuhlen H Martinoff S Seyfarth M Nekolla S Dirschinger J Schwaiger M Schomig A A randomized evaluation of the effects of glucose-insulin-potassium infusion on myocardial salvage in patients with acute myocardial infarction treated with reperfusion therapy Am Heart J 2004 148 105
McGowan JH Cleland JG Reliability of reporting left ventricular systolic function by echocardiography: a systematic review of 3 methods Am Heart J 2003 146 388 397 12947354
Henriques JP Zijlstra F van 't Hof AW de Boer MJ Dambrink JH Gosselink M Hoorntje JC Suryapranata H Angiographic assessment of reperfusion in acute myocardial infarction by myocardial blush grade Circulation 2003 107 2115 2119 12695301
Ottervanger JP van 't Hof AW Reiffers S Hoorntje JC Suryapranata H de Boer MJ Zijlstra F Long-term recovery of left ventricular function after primary angioplasty for acute myocardial infarction Eur Heart J 2001 22 785 790 11350111
Sasso FC Carbonara O Cozzolino D Rambaldi P Mansi L Torella D Gentile S Turco S Torella R Salvatore T Effects of insulin-glucose infusion on left ventricular function at rest and during dynamic exercise in healthy subjects and noninsulin dependent diabetic patients: a radionuclide ventriculographic study J Am Coll Cardiol 2000 36 219 226 10898438
Yetkin E Senen K Ileri M Atak R Tandogan I Yetkin O Kosar F Turhan H Cehreli S Comparison of low-dose dobutamine stress echocardiography and echocardiography during glucose-insulin-potassium infusion for detection of myocardial viability after anterior myocardial infarction Coron Artery Dis 2002 13 145 149 12131017
Mehta SR Yusuf S Diaz R Zhu J Pais P Xavier D Paolasso E Ahmed R Xie C Kazmi K Tai J Orlandini A Pogue J Liu L Effect of glucose-insulin-potassium infusion on mortality in patients with acute ST-segment elevation myocardial infarction: the CREATE-ECLA randomized controlled trial JAMA 2005 293 437 446 15671428
Ramanathan T Shirota K Morita S Nishimura T Huang Y Hunyor SN Glucose-insulin-potassium solution improves left ventricular mechanics in diabetes Ann Thorac Surg 2002 73 582 587 11845878
Apstein CS Gravino FN Haudenschild CC Determinants of a protective effect of glucose and insulin on the ischemic myocardium. Effects on contractile function, diastolic compliance, metabolism, and ultrastructure during ischemia and reperfusion Circ Res 1983 52 515 526 6342832
Depre C Vanoverschelde JL Taegtmeyer H Glucose for the heart Circulation 1999 99 578 588 9927407
Opie LH The glucose hypothesis: its relation to acute myocardial ischemia J Mol Cell Cardiol 1970 1 107 115
Rackley CE Russell RO JrRogers WJ Mantle JA McDaniel HG Papapietro SE Clinical experience with glucose-insulin-potassium therapy in acute myocardial infarction Am Heart J 1981 102 1038 1049 7032266
Lopaschuk GD Wambolt RB Barr RL An imbalance between glycolysis and glucose oxidation is a possible explanation for the detrimental effects of high levels of fatty acids during aerobic reperfusion of ischemic hearts J Pharmacol Exp Ther 1993 264 135 144 8380856
Eberli FR Weinberg EO Grice WN Horowitz GL Apstein CS Protective effect of increased glycolytic substrate against systolic and diastolic dysfunction and increased coronary resistance from prolonged global underperfusion and reperfusion in isolated rabbit hearts perfused with erythrocyte suspensions Circ Res 1991 68 466 481 1991351
McNulty PH Pfau S Deckelbaum LI Effect of plasma insulin level on myocardial blood flow and its mechanism of action Am J Cardiol 2000 85 161 165 10955370
Sundell J Laine H Nuutila P Ronnemaa T Luotolahti M Raitakari O Knuuti J The effects of insulin and short-term hyperglycaemia on myocardial blood flow in young men with uncomplicated Type I diabetes Diabetologia 2002 45 775 782 12107720
Marano L Bestetti A Lomuscio A Tagliabue L Castini D Tarricone D Dario P Tarolo GL Fiorentini C Effects of infusion of glucose-insulin-potassium on myocardial function after a recent myocardial infarction Acta Cardiol 2000 55 9 15 10707753
Aikawa R Nawano M Gu Y Katagiri H Asano T Zhu W Nagai R Komuro I Insulin prevents cardiomyocytes from oxidative stress-induced apoptosis through activation of PI3 kinase/Akt Circulation 2000 102 2873 2879 11104747
Jonassen AK Sack MN Mjos OD Yellon DM Myocardial protection by insulin at reperfusion requires early administration and is mediated via Akt and p70s6 kinase cell-survival signaling Circ Res 2001 89 1191 1198 11739285
Matsui T Tao J del Monte F Lee KH Li L Picard M Force TL Franke TF Hajjar RJ Rosenzweig A Akt activation preserves cardiac function and prevents injury after transient cardiac ischemia in vivo Circulation 2001 104 330 335 11457753
Sack MN Yellon DM Insulin therapy as an adjunct to reperfusion after acute coronary ischemia: a proposed direct myocardial cell survival effect independent of metabolic modulation J Am Coll Cardiol 2003 41 1404 1407 12706939
Diaz-Araya G Nettle D Castro P Miranda F Greig D Campos X Chiong M Nazzal C Corbalan R Lavandero S Oxidative stress after reperfusion with primary coronary angioplasty: lack of effect of glucose-insulin-potassium infusion Crit Care Med 2002 30 417 421 11889322
| 15932638 | PMC1177952 | CC BY | 2021-01-04 16:27:56 | no | BMC Med. 2005 Jun 2; 3:9 | utf-8 | BMC Med | 2,005 | 10.1186/1741-7015-3-9 | oa_comm |
==== Front
BMC MicrobiolBMC Microbiology1471-2180BioMed Central London 1471-2180-5-401599240710.1186/1471-2180-5-40Methodology ArticleA new short-term toxicity assay using Aspergillus awamori with recombinant aequorin gene Kozlova Olga [email protected] Mark [email protected] Nick [email protected] Institute of Cell & Molecular Biology, The University of Edinburgh, King's Buildings, Edinburgh, EH9 3JL, UK2 Pollution Research Unit, Napier University, Merchiston Campus, Edinburgh, EH10 5DT, UK; Presently, Surfactant Technologies Ltd., C/o Avecia Fine Chemicals,, Grangemouth, FK3 8XG, UK3 Pollution Research Unit, Napier University, Merchiston Campus, Edinburgh, EH10 5DT, UK4 LUTESS Ltd., Orchard Brae House,, Edinburgh EH4 2HG, UK2005 2 7 2005 5 40 40 12 1 2005 2 7 2005 Copyright © 2005 Kozlova 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
Most currently available short-term toxicity assays are based on bacterial cells. Therefore there is a need for novel eukaryotic microbial bioassays that will be relevant to higher eukaryotes such as animals and plants. Ca2+ is a universal intracellular signalling molecule found in all organisms from prokaryotes to highly specialized animal cells. In fungi calcium has been demonstrated to be involved in control of many important processes. The recombinant aequorin gene from the jellyfish Aequorea victoria responsible for the expression of the Ca2+-sensitive aequorin photoprotein has been cloned in the filamentous fungus Aspergillus awamori. This has allowed real life monitoring of [Ca2+]c changes in living fungal cells. When subjected to different physico-chemical stimuli fungal cells respond by transiently changing the concentration of free Ca2+ in the cytosol ([Ca2+]c) and the pattern of these changes (Ca2+ signature) is specific to each particular stimulus. Therefore it was interesting to investigate whether different environmental toxicants would be able to affect the pattern of [Ca2+]c changes in a reproducible and dose dependant manner.
Results
Toxicity bioassay has been developed to monitor changes [Ca2+]c of the recombinant fungus in the presence of toxicants representing heavy metals – Cr6+ and Zn2+ and a phenolic polar narcotic -3,5-DCP. The fungus responds to toxicants by a decrease in the amplitude of [Ca2+]c response to 5 mM external CaCl2 and an increase in Ca2+ final resting levels and recovery time.
Conclusion
A novel toxicity bioassay utilizing eukaryotic cells has been developed based on filamentous fungi transformed with the recombinant aequorin gene. A range of parameters characterising changes in [Ca2+]c has been identified, e.g. Amplitude, Length of Transient, Final Resting Level and Recovery Time. These parameters can be used to determine the toxicity of a range of chemicals to eukaryotic cells in a 96-well microtitre plate method.
==== Body
Background
The widespread use and release of natural and synthetic chemicals into the environment, singly or as complex domestic and industrial effluents, has necessitated the development of rapid and cost effective toxicity tests to protect humans and other biota [1,2]. Short-term and long-term bioassays exist utilising microorganisms, invertebrates and higher plants and animals. The short-term microbial toxicity tests involve bacteria, algae, protozoa and fungi (yeasts). One of the most widely used is the Microtox® proprietary test utilising a natural light emitting marine bacterium, Vibrio fischeri. V. fischeri is a gram negative free-living luminescent bacterium [3,4] also found in some species of fish and squid within the light organs [5].
Bioluminescence involves the emission of visible light mediated by the luciferin-luciferase enzyme system [6]. Microbial bioluminescence is a branch of the electron transport chain [7] and as electron transport is involved in cell metabolism, any disruption to this system e.g. by the presence of toxins, will have an effect on light output. Luminescence is controlled by the lux operon [8] within V. fischeri, which produces light at 490 nm [7,9].
Calcium is a well-known second messenger involved in the transduction of different external stimuli and hormonal signals in eukaryotic cells. A great number of studies describing changes in [Ca2+]c have been done describing the role of calcium in animal and plant cells subjected to different treatments. However, very little work in this area has been done on filamentous fungi due to the lack of both routine and reliable methods for monitoring intracellular free Ca2+ in living fungal cells.
Recently recombinant aequorin gene has been expressed in filamentous fungi [10]. Aequorin is a Ca2+-sensitive photoprotein of the jellyfish Aequorea victoria (Mr = 21,400) [11]. The protein consists of a single polypeptide chain, apoaequorin, a hydrophobic luminophore, coelenterazine and bound oxygen [12]. Once Ca2+ ions are bound to the three Ca2+-binding sites in aequorin, the protein is converted into an oxygenase. The oxygenase catalyses the oxidation of the substrate coelenterazine by the bound oxygen and this results in blue light emission. The amount of luminescence emitted by aequorin is dependent upon free Ca2+ concentration and thus aequorin can be used to report cytosolic Ca2+ ([Ca2+]c). Using aequorin transformed fungi it has been shown that different physico-chemical stimuli result in a transient [Ca2+]c increases with a unique Ca2+ signature [10].
The general purpose of toxicity testing is to establish the potential impact of chemicals on biota in the environment. The information gained can then be used to manage the treatment or release of chemicals. Whether a substance is toxic or not depends on physico-chemical factors such as pH, temperature and salinity, but most importantly on the test organism used, the concentration of the chemical and the conditions imposed for the test (see e.g. [13]. No single toxicity test can determine the effect of toxicants on all biota because of differences in response by organisms at different trophic levels. Eukaryotic organisms can behave differently to prokaryotes although with the V. fischeri test a high correlation has been shown with other bioassays using more complex organisms [14].
In this study we have tested the effect of toxic substances on the dynamics of [Ca2+]c in the fungus Aspergillus awamori, containing aequorin gene that constitutively expresses the Ca2+-binding photoprotein aequorin. Data are compared with those obtained using the V. fischeri bioluminescence assay. This represents the first study in using recombinant aequorin genes in a filamentous fungus to assess toxicity of aqueous samples.
Results
All three toxicants tested showed a response (Table 1). Parameters assessed were rise time (characterises the time from application of stimulus to maximum amplitude of the response), amplitude (A), Length of Transient (LT50 at the point where A = 1/2Amax), final resting level (level of [Ca2+]c at the end of the experiment) and recovery time (luminescence integrated from the point when amplitude is maximal to the point when [Ca2+]c reaches its final resting level at the end of the experiment.
At low concentration, 3,5-DCP (5 min preincubation) caused a small increase in the amplitude of the [Ca2+]c response. At high concentrations (11.2 and 112 mgl-1) the phenolic inhibited the amplitude of the [Ca2+]c response at both 5 and 30 min preincubation (Figure 1a &1b; Table 1). The highest concentration of 3,5-DCP (112 mg l-1) also increased LT50 and rise time (Table 1). At all three concentrations of 3,5-DCP there were elevated final resting levels of Ca2+ and increased recovery times. The percentage increase in the final resting levels and recovery time [15] were used as a quantifying parameter for the analysis of the effect of the phenolic toxicant on [Ca2+]c (Table 1).
Chromium did not affect (P < 0.5) the final resting levels and recovery time of Ca2+ but had a strong dose dependant inhibition of the amplitude of the [Ca2+]c response (Figure 1c &1d). Zinc had a combination of these effects on [Ca2+]c (Figure 1e &1f). It caused a decrease in amplitude of the [Ca2+]c response as well as an increase in the final resting levels and recovery times at high concentrations (≥ 350 mg l-1) (Table 1). Both inhibition of the amplitude and the increase in the final resting levels were dose dependent with the exception of the 1300 mg l-1 Zn2+ at 30 min preincubation. The increase of the above mentioned parameters observed at this concentration was lower than when fungi were treated with 700 mg l-1 of Zn2+. This was probably due to aequorin quenching which is known to be caused by the very high concentrations of some heavy metals. At low concentration (180 mg l-1) Zn2+ caused an increase in amplitude of the [Ca2+]c response as well as an increase in the final Ca2+ resting level and recovery time.
Recovery of treated A. awamori to basal Ca2+ levels and light dissipation was fairly rapid in systems challenged with Cr6+ for 5 and 30 min. This was not the case for A. awamori treated with Zn2+ and 3,5-DCP. In these cases there was protracted recovery in Ca2+ content at higher toxicant challenges with both 5 and 30 min incubation.
IC50 is one standard parameter calculated in toxicity studies and was chosen in this study to determine differences in values obtained by the fungal and bacterial luminescence bioassays. In order to calculate IC50 values for all toxicants it was decided to utilise the amplitude of [Ca2+]c changes in the fungal cell after challenge with different concentrations of toxicant. Examples of dose response curves are presented in Figures 2, 3. Calculated IC50 values for the 5 and 30 min toxicant pre-incubations with A. awamori are presented in Table 2. The 30 min IC50 for DCP, Cr6+ and Zn2+ are 36.7, 167.8 and 549.7 mg l-1 respectively. These can be compared to values for the V. fischeri test of results of 3.6, 13.95 and 0.44 mg l-1 for the same toxicants tested over a 30 min incubation period.
Discussion
In this paper data have been presented on the response of A. awamori to different toxicant concentrations by examining changes in Relative Light Units (RLU) and [Ca2+]c in the presence of an external CaCl2 concentration of 5 mM. This study differs from that of Torrecilla et al. [16] which examined changes in intracellular Ca2+ of the cyanobacterium Anabaena able to express apoaequorin constitutively when subjected to heat and cold shock. It is also the first study, as far as we know, examining responses of aequorin to toxicants in a filamentous fungus. The concentrations of Zn2+, Cr6+ and 3,5-DCP used in the experiments to assess toxicity were based on experience with V. fischeri bioassays, through range finding and interest in observing effects at low and high concentrations. Initially, the effect of each of these three chemicals on [Ca2+]c was examined. The [Ca2+]c response was not significantly different from that obtained with the control solution (data not shown). A second approach studied the effect of preincubation of the fungus with toxicants on [Ca2+]c response to external CaCl2.
There is no doubt that aequorin transformed A. awamori, using the protocol developed, responds to toxic chemical challenge in a reproducible way. The aequorin system has more parameters which can be assessed than other bioassays i.e. final [Ca2+]c resting level, recovery time and amplitude of [Ca2+]c/RLU following chemical challenge for 5 or 30 min.
A. awamori has been shown to respond to organic and inorganic compounds by a decrease in the amplitude of [Ca2+]c response to external CaCl2 with increasing toxicant concentration. Zn2+ and DCP also affected the final Ca2+ resting levels and recovery times. In the case of 30 min preincubation with either 112 mg l-1 DCP or 700 mg l-1 Zn2+, the final [Ca2+]c resting levels remained significantly higher than after 5 min preincubation with these toxicants indicating a greater toxicity to the fungus on longer contact with the toxicant. This is not an unusual response even in the V. fischeri bioassay where it is generally observed that toxicity increases (IC50 values decrease) with longer incubation times. IC50 values calculated using amplitude changes of [Ca2+]c showed decreased toxicity to Zn2+ at longer preincubation times. This needs further analytical consideration to determine toxicity values of relevance.
The proprietary Microtox bioassay and other bacterial bioluminescence methods, including that used in the present study, which do not involve genetically modified bacteria, utilise Vibrio (e.g. V. fischeri) and related Photobacterium species. Acute toxicity tests utilising such luminescent bacteria can underestimate the toxicity of chemicals and Backhaus et al. [17] showed that more reliable toxicity estimates can be obtained through the use of long-term toxicity testing with the same organisms. It may be prudent to test the response of A. awamori using longer preincubation times with toxicants e.g. > 24 h, prior to monitoring Ca2+ homeostasis. This may increase the sensitivity of the bioassay. It must be remembered however that for one of the toxicants (Zn2+), a decrease in sensitivity was evident at 30 min compared to the result at 5 min. There was also evidence of toxicity recovery through adaptation with increased incubation. This may be due to acquired resistance by the organism through synthesis of metal-binding proteins (metallothioneins) or constituents in the growth medium removing/immobilising the metal (e.g. EDTA, phosphate precipitation). This needs to be examined in the fungal bioassay.
It is interesting to compare toxicity data obtained with A. awamori with those obtained with the V. fischeri bioassay. The toxicity values for the fungus are higher than those of the bacterial test indicating lower bioassay sensitivity with the parameter used to calculate the IC50 values. For example, the IC50 values for the 30 min Zn2+ incubations were approximately 3 orders of magnitude lower for V. fischeri. Such insensitivity to Zn2+ has been observed with an ATP luminescence assay [18]. The IC50 results with Cr6+ (30 min preincubation) were 400 mg l-1, 12 times higher than in the bacterial biosensor. This value is the same order of magnitude as the 15 min Microtox assay (339.6 mg l-1) carried out by Codina et al. [19]. Indeed, the use of V. fischeri in various proprietary tests (LumisTox, Microtox, ToxAlert etc) have been shown to exhibit differences in sensitivity to different toxicants [20]. It is, therefore, difficult to categorically state that V. fischeri is more sensitive to Cr6+ than A. awamori particularly when different incubation periods are used in bioassays affecting any ultimate IC50 value. It should be also noted that since the aequorin test is based on changes in [Ca2+]c, toxic chemicals often exert an increase in [Ca2+]c thus affecting IC50 calculations. The calculation of IC50 for this fungal bioassay may not be appropriate. Other parameters such as LT50, rise time, final Ca2+ resting levels and recovery time would provide ideal candidate parameters, but more tests and comparisons with other bioassays are needed. The aequorin bioassay can also generate up to 15 parameters to assess the effect of toxicants due to the complex pattern of [Ca2+]c changes [21]. These parameters can be further used to create a profile for toxicants with specific modes of action [15]. With different parameters available for analysis in the aequorin test it could be useful to assess the toxicity by calculating the NOEC (no observable effect concentration) or a nominal IC10 value [22].
Codina et al. [19] also used a yeast bioassay to test metal toxicity. The eukaryotic yeast was found to be less sensitive to metals than prokaryotic organisms including Vibrio and Pseudomonas species. An IC50 value of 549.1 mg l-1 was obtained for Zn2+ using the yeast assay [19] which is comparable to some of the values obtained with A. awamori in our study. The yeast was slightly more sensitive to Cr6+ (30.9 mg l-1) than the filamentous A awamori but this is not unusual among members of the same trophic level (see [19]. Recent chronic toxicity studies have been using wild-type and genetically modified mutants of the yeast Saccharomyces cerevisiae [23]. EC50 values of >1000 mg l-1 were determined for Zn2+ and 1.7-4.79 mg l-1 for Cr6+ indicating variability in assays that rely on conditions imposed including exposure time, species tested and criteria used in the final assessment.
An interesting observation is the enhanced stimulation of light output in some systems with a low concentration of toxicant. This is seen in the 5 min treated systems using 0.11 mg l-1 DCP (Figure 1a), 180 mg l-1 Zn2+ (Figure 1e) and also in the 30 min treated system with 180 mg l-1 Zn2+ (Figure 1f). This stimulation, referred to as hormesis, is a common occurrence in toxicity bioassays [13] and is often observed in our laboratory using the V. fischeri bioassay.
It is not clear based on these preliminary observations what causes Ca2+ and light retention. In the case of DCP it may be that this polar narcotic is affecting membrane permeability and the transport of Ca2+ out of the cytosol (Ca2+ ATPase and efflux of Ca2+ via e.g. Ca2+/H+ antiporters) and hence, a delay in RLU dissipation. The metals may be competing with the transport mechanisms in membranes. The metals thus act by interacting with physiological ions affecting transport and its concomitant effect on light output.
Most of the organic chemicals discharged to the environment exert a narcotic effect on biota. This is either Type I (non-polar narcosis) or Type II (polar narcosis). It would be interesting to utilise A. awamori to test whether non-polar and polar narcoses operate in the fungus and whether these can be predicted by QSAR (Quantitative Structure Activity Relationships). Cronin and Schulz [24] showed, using QSAR, that non-polar narcosis occurs in V. fischeri but they could not validate polar narcosis in the organism.
There is no doubt that a comprehensive testing of metals and organic compounds needs to be carried out to assess the value of A. awamori as a toxicity sensing tool both in systems using pure chemicals and those involving real samples (e.g. complex effluents and soil water matrices). Toxicant preincubation time and effect on fungus response would be an important factor to test fully.
Conclusion
The current paper presents the proof-of-concept study of the novel toxicity bioassay based on using filamentous fungi transformed with the recombinant aequorin gene. The research conducted has shown that the recombinant aequorin method can be used as a novel eukaryotic toxicant biosensor. It offers more parameters which can be readily analysed than the traditionally used bacterial biosensors. The fungal aequorin biosensor responded to toxic substances (3,5-DCP, Zn2+ and Cr6+) by a decrease in the amplitude of the [Ca2+]c response to 5 mM CaCl2 and an increase in the [Ca2+]c resting levels (Zn2+ and 3,5-DCP). Preliminary IC50 values obtained with the fungal aequorin system, and based on the inhibition of the [Ca2+]c amplitude in response to external CaCl2 treatment, indicate that it is less sensitive for detecting 3,5-DCP, Zn2+ and Cr6+ than the V. fischeri luciferase system. However, it is obvious that the changes in amplitude is not the most suitable parameter for calculating IC50 and more research are needed to determine the best possible set of parameters to be used together with the aequorin bioassay.
Methods
Chemicals
Unless otherwise stated all the chemicals used were from Calbiochem-Novabiochem (UK) Ltd. (Nottingham, UK) or Sigma (Dorset, UK).
Organisms and media
Experiments were carried out with the strain of A. awamori 66A, which has been transformed with the recombinant aequorin gene [10]. All media and salt solutions were made up in distilled H2O and sterilised by autoclaving at 121°C at 15 psi for 20 min prior to use. A. awamori cultures were grown in liquid Vogel's medium with 1% sucrose (VS medium [25]).
A method of growing A. awamori cultures for in vivo luminometry was employed utilising static liquid culture in 96-well plates. Twelve ml of sterile VS medium was inoculated with 1 × 105 spores per ml. Coelenterazine (Biosynth AG, Staad, Switzerland) was dissolved in methanol (MeOH) to a final concentration of 2.5 μM. The final MeOH concentration was not more than 0.1%, which has been shown not to affect spore germination or hyphal growth. Using a 12-channel pipette (Anachem, Luton, UK), 100 μl of the inoculated medium was added to each well, and the plate covered with a microplate lid (Labsystems, Helsinki, Finland). Cultures were incubated in a humidity chamber in the presence of free water at 30°C for 24 h.
Luminometry was performed using an EG & G Berthold (Bad Wildbad, Germany) LB96P Microlumat luminometer. The luminometer was designed to measure luminescence using the flat-bottomed 96-well microtitre plates from EG&G Berthold or Dynex Technologies Inc. (Chantily, UK). Two 100 μl built-in injectors allowed stimulation of samples.
The luminometer measures light emission in relative lights units (RLU). To convert RLU into [Ca2+]c concentrations the following empirically derived equation was used:
pCa = 0.332588 (-log k)+5.5593,
where k = luminescence counts s-1/ total luminescence counts [26].
Total luminescence is measured as an integral of all luminescence up to complete aequorin discharge. Aequorin was discharged by adding 2M CaCl2 in 20% ethanol to the wells. Ethanol was used as a permeabilising agent. Conversion of RLU into [Ca2+]c concentration was made using a Macro developed in Excel 7.0 by Mark Knight (University of Oxford).
Three toxicants were tested: 3,5-DCP, ZnSO4 and K2Cr2O7. All toxicants were added 5 min and 30 min before treatment with 5 mM external CaCl2 in a total volume of 25 μl VS medium. 4-6 replicates were performed for each treatment. All the experiments were performed in triplicate. IC50 values were calculated only for the amplitude of the response.
Vibrio fischeri bioassays utilised cultures of the organism grown in Photobacterium Broth (PB) at 37°C until optimum light output. A 100 μl volume of V. fischeri suspension was added to each of the 96 wells in a microtitre plate. During the experiment a 100 μl volume of NaCl solution (giving a final concentration of 2% NaCl in the test) containing the test substance in the appropriate concentration was added to the wells. For each measurement a control consisting of 6 wells with V. fischeri to which the NaCl solution was added. Chemicals were tested in five different concentrations and each concentration was tested in three replicates. Plates were measured in an Anthos Lucy 1 luminometer. Measurements were made over a period of 30 min. IC50 values were calculated for 5 and 30 min incubations. Recording of data started immediately after addition of the test solutions to the wells containing the culture. The standard proprietary Microtox bioassay was also carried out [27].
Authors' contributions
OK conceived the study, designed and developed aequorin bioassay, carried out the tests and has been involved in the writing of the article.
MZ has designed the microbial part of the testing and participated in carrying out the fungal tests.
NC helped design the study and participated in writing of the article.
All authors read and approved the final manuscript.
Acknowledgements
We are grateful to Dr Mark Knight, Department of Plant Science, University of Oxford for the Excel 7.0 spreadsheet used to convert RLU to [Ca2+]c concentrations, University of Edinburgh for providing aequorin transformed fungi and to Darwin Trust for providing funding for Olga Kozlova.
Figures and Tables
Figure 1 Effect of 3,5-DCP, Cr6+ and Zn2+ on [Ca2+]c response to the addition of external CaCl2 (5 mM). A, C and E show 5 min incubation with the toxicant, B, D F show 30 min incubation with the toxicant.
Figure 2 Effect of 3,5-DCP, Cr6+ and Zn2+ on amplitude of [Ca2+]c in A. awamori 5 minute incubation.
Figure 3 Effect of 3,5-DCP, Cr6+ and Zn2+ on final resting level of [Ca2+]c in A. awamori 5 minute incubation.
Table 1 Detailed analysis of the effects of preincubating cultures of A. awamori with 3,5-DCP, Cr6+ and Zn2+ on [Ca2+]c response to 5 mM external CaCl2
Toxicant Features of [Ca2+]C Signatures
A LT50 FRL (%) 5 min pre-incubation RT (%) 5 min pre-incubation FRL (%) 30 min pre-incubation RT (%) 30 min pre-incubation
3,5-DCP 0.112 ↑ - 111 ± 8.4 (↑) 111 ± 8.4 (↑) 106 ± 7.6 (-) 98 ± 1.3 (-)
(mgl-1) 11.2 ↓ - 120 ± 4.8 (↑) 125 ± 4.6 (↑) 127 ± 29.7 (-) 134 ± 31.9 (↑)
112 ↓ ↑ 289 ± 17.6 (↑) 292 ± 6.6 (↑) 416 ± 23.2 (↑) 416 ± 16.6 (↑)
Cr6+ 15 ↓ - - - - -
(mgl-1) 120 ↓ - - - - -
260 ↓ - - - - -
Zn2+ 180 ↑ - 106 ± 9.2(-) 127 ± 4.6 (↑) 97 ± 6.9(-) 92 ± 1.9(↓)
(mgl-1) 350 ↓ - 153 ± 13.6(↑) 167 ± 14.4(↑) 149 ± 11.4(↑) 143 ± 10.2(↑)
700 ↓ ↑ 164 ± 30.7(↑) 179 ± 26.9(↑) 245 ± 16.5(↑) 261 ± 9.6(↑)
1300 ↓ ↑ 169 ± 14.8(↑) 182 ± 10.4(↑) 169± 12.6(↑) 176 ± 16.5(↑)
Note: – no effect compared with untreated control), ↑ increase in parameter, ↓ decrease in parameter, A – amplitude, LT50 – Length of transient, FRL – Final Ca2+ resting level, RT – recovery time. A and LT50 are shown only for 5 min incubation. Results represent mean ± SD. The % values represent the absolute value of the parameter not the relative increase compared with control. N = 6
Table 2 IC50 values (mgl-1) for 3,5-DCP, Cr6+ and Zn2+ using V. fischeri and A. awamori, (based on amplitude of [Ca2+]c)
Treatment V.fischeri A. awamori
Test duration or preincubation time IC50 (5 min) IC50 (30 min) IC50 (6 min) IC50 (30 min)
3,5-DCP 3.62 3.13 46.7 36.7
Cr6+ 29.9 13.95 400.1 167.9
Zn2+ 95.56 0.44 237.2 549.7
Note: N = 6
==== Refs
Dalzell DJB Alte S Aspichueta E de la Sota A Etxebarria J Gutierrez M Hoffmann CC Sales D Obst U Christofi N A comparison of five rapid direct toxicity assessment methods to determine toxicity of pollutants to activated sludge Chemosphere 2002 47 535 545 11996129 10.1016/S0045-6535(01)00331-9
Sherry J Scott B Dutka B Use of various acute, sublethal and early life-stage tests to evaluate the toxicity of refinery effluents Environ Toxicol Chem 1997 16 2249 2257 10.1897/1551-5028(1997)016<2249:UOVASA>2.3.CO;2
Gellert G Stomme A Trujillano A Development of an optimal bacterial medium based on the growth inhibition assay with V. fischeri Chemosphere 1999 39 467 476 10399848 10.1016/S0045-6535(99)00002-8
Hoffmann CC Sales D Christofi N Combination ecotoxicity and testing of common chemical discharges to sewer using the Vibrio fischeri luminescence bioassay Int Microbiol 2003 6 41 47 12730711
Ruby E Lee K The Vibrio fischeri-Euprymna scolopes light organ association: Current ecological paradigms Appl Environ Microbiol 1998 64 805 813 16349524
Steinberg S Poziomek E Engelmann W Rogers K A review of environmental applications of bioluminescent measurement Chemosphere 1995 30 2155 2197 10.1016/0045-6535(95)00087-O
Wilson T Hastings JW Bioluminescence Ann Rev Cell Dev Biol 1998 14 197 230 9891783 10.1146/annurev.cellbio.14.1.197
Meighen E Molecular-biology of bacterial bioluminescence Microbiol Rev 1991 55 123 142 2030669
Hastings J Potrikus C Guptas S Kurfust M Makemson J Biochemistry and physiology of bioluminescent bacteria Adv Microb Physiol 1985 26 235 291 3913293
Nelson G Kozlova-Zwinderman O Collis AJ Knight MR Fincham JRS Stanger CR Renwick A Hessing JGM Punt PJ van den Hondel CAMJJ Read ND Calcium measurement in living filamentous fungi expressing codon-optimized aequorin Molecular Microbiology 2004 52 1437 1450 15165245 10.1111/j.1365-2958.2004.04066.x
Nakajimashimada J Iida H Tsuji FI Anraku Y Galactose-dependent expression of the recombinant ca-2+-binding photoprotein aequorin in yeast Biochem Biophys Res Comm 1991 174 115 122 1989594 10.1016/0006-291X(91)90493-Q
Knight M Campbell AK Smith SM Trewavas A Recombinant aequorin as a probe for cytosolic free Ca2+ in Escherichia coli FEBS Lett 1991 282 405 405 2037058 10.1016/0014-5793(91)80524-7
Christofi N Hoffmann C Tosh L Hormesis responses in free and immobilised light-emitting bacteria Ecotoxicol Environ Saf 2002 52 227 231 12297084 10.1006/eesa.2002.2203
Kaiser KLE Richardson ML Qualitative and quantitative relationships of Microtox data with toxicity data for other aquatic species Ecotoxicology Monitoring 1993 Weinheim: VCH Publishers 197 211
Kozlova-Zwinderman O Eukaryotic biosensor making use of a calcium regulated light emitting enzyme PCT Patent WO2004050905 2004
Torrecilla I Leganés F Bonilla I Fernández-Pinas F Use of recombinant aequorin to study calcium homeostasis and monitor calcium transients in response to heat and cold shock in cyanobacteria Plant Physiol 2000 123 161 175 10806234 10.1104/pp.123.1.161
Backhaus T Froehner K Altenburger R Grimme LH Toxicity testing with V. fischeri : A comparison between long term (24 h) and the short term (30 min) bioassay Chemosphere 1997 35 2925 2938 10.1016/S0045-6535(97)00340-8
Dalzell DJB Christofi N An ATP luminescence method for direct toxicity assessment of pollutants impacting on the activated sewage sludge process Water Research 2002 36 1493 1502 11996339 10.1016/S0043-1354(01)00346-3
Codina JC Pérez-Garcia A Romero P de Vicente A A comparison of microbial bioassays for the detection of metal toxicity Arch Environ Contam Toxicol 1993 25 250 254 8368866 10.1007/BF00212137
Christofi N Dalzell D Hoffmann C Sales D Morton J Arretxe M Heap M Obst U Alte S Etxebarria J de Las Fuentes M Gutierrez M de la Sota A Aspichueta E DTOX- Direct toxicity assessment of complex industrial effluents discharged to sewer Final Report to the European Commission Contract SMT4-CT96-2084 Brussels: Project funded under the Standards, Measurements and Testing Programme 2000
Kozlova-Zwinderman O Calcium measurements in filamentous fungi using recombinant aequorin PhD Thesis 2002 University of Edinburgh, Institute of Cell and Molecular Biology
Hoekstra JA van Ewijk PH The bounded effect concentration as an alternative to the NOEC Sci Total Environ 1993 Part 1 705 711
Schmitt M Gellert G Ludwig J Lichtenberg-Frate H Phenotypic yeast growth analysis for chronic toxicity testing Ecotoxicol Environ Saf 2004 59 142 150 15327869 10.1016/j.ecoenv.2004.06.002
Cronin MTD Schulz TW Validation of Vibrio fischeri acute toxicity data: mechanism of action-based QSARs for non-polar narcotics and polar narcotic phenols Sci Total Environ 1997 204 75 88 9299768 10.1016/S0048-9697(97)00179-4
Vogel HJ A convenient growth medium for Neurospora (Medium N) Microbial Genetics Bulletin 1956 51 107 124
Fricker MD Plieth C Knight H Blancaflor E Knight MR White NS Gilroy S Mason WT Fluorescence and Luminescence Techniques to Probe Ion Activities in Living Plant Cells Fluorescent and Luminescent Probes 1999 London: Academic Press 569 596
AZUR Environmental Microtox Manual 1997 Carlsbad
| 15992407 | PMC1177953 | CC BY | 2021-01-04 16:03:39 | no | BMC Microbiol. 2005 Jul 2; 5:40 | utf-8 | BMC Microbiol | 2,005 | 10.1186/1471-2180-5-40 | oa_comm |
==== Front
BMC Med Inform Decis MakBMC Medical Informatics and Decision Making1472-6947BioMed Central London 1472-6947-5-141595338810.1186/1472-6947-5-14DebateA draft framework for measuring progress towards the development of a national health information infrastructure Sittig Dean F [email protected] Richard N [email protected] Kevin [email protected] Charles [email protected] Barbara [email protected] George [email protected] Laura L [email protected] Lawrence C [email protected] Rainu [email protected] Department of Medical Informatics, Northwest Permanente, P.C., Portland, OR USA2 Center for Medical Informatics, Yale University School of Medicine, New Haven, CT USA3 Department of Health Policy, Management & Evaluation, University of Toronto, Toronto, Ontario Canada4 National Library of Medicine, Rockville, MD USA5 The Leapfrog Group, Washington, DC USA6 Department of Biomedical Informatics, Columbia University, New York, NY USA7 Rhode Island Quality Institute, Providence, RI USA8 Center for Research Evaluation and Planning, Nemours Foundation, Newark, DE USA9 Division of General Internal Medicine, Brigham and Women's Hospital, Boston, MA USA10 Department of Public Health, Cornell Medical School, NY, NY USA2005 13 6 2005 5 14 14 1 2 2005 13 6 2005 Copyright © 2005 Sittig et al; licensee BioMed Central Ltd.2005Sittig 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
American public policy makers recently established the goal of providing the majority of Americans with electronic health records by 2014. This will require a National Health Information Infrastructure (NHII) that is far more complete than the one that is currently in its formative stage of development. We describe a conceptual framework to help measure progress toward that goal.
Discussion
The NHII comprises a set of clusters, such as Regional Health Information Organizations (RHIOs), which, in turn, are composed of smaller clusters and nodes such as private physician practices, individual hospitals, and large academic medical centers. We assess progress in terms of the availability and use of information and communications technology and the resulting effectiveness of these implementations. These three attributes can be studied in a phased approach because the system must be available before it can be used, and it must be used to have an effect. As the NHII expands, it can become a tool for evaluating itself.
Summary
The NHII has the potential to transform health care in America – improving health care quality, reducing health care costs, preventing medical errors, improving administrative efficiencies, reducing paperwork, and increasing access to affordable health care. While the President has set an ambitious goal of assuring that most Americans have electronic health records within the next 10 years, a significant question remains "How will we know if we are making progress toward that goal?" Using the definitions for "nodes" and "clusters" developed in this article along with the resulting measurement framework, we believe that we can begin a discussion that will enable us to define and then begin making the kinds of measurements necessary to answer this important question.
==== Body
Background
In the United States of America, public policy makers recently established the goal of having electronic health records (EHRs) for the majority of Americans by 2014 [1]. This article presents recommendations regarding specific aspects of a conceptual and measurement framework that will help us to measure progress toward that goal. It represents a starting point for what will hopefully be a wide-ranging discussion of exactly how we should measure progress toward the achievement of a functional National Health Information Infrastructure (NHII) [Note: Other terms have also been used to describe this rather nebulous concept including National Health Information Network (NHIN), National Health Information Infrastructure (NHII), Regional Health Information Organization (RHIO), and Local Health Information Infrastructure (LHII), we will use NHII to refer to this concept. Such an NHII would allow all patients, healthcare providers, and those interested in population health to have access to comprehensive electronic health records. This article resulted from a discussion at "The Secretarial Summit on Health Information Technology Launching the National Health Information Infrastructure 2004: Cornerstones for Electronic Healthcare" held in Washington, D.C., July 20–23, 2004
What is the NHII?
The National Health Information Infrastructure (NHII) is:
• An initiative set forth to improve the effectiveness, efficiency and overall quality of health and health care in the United States
• A comprehensive knowledge-based network of interoperable systems of clinical, public health, and personal health information that would improve decision-making by making health information available when and where it is needed.
• The set of technologies, standards, applications, systems, values, and laws that support all facets of individual health, health care, and public health [2].
Measuring the progress in creation, deployment and adoption of health information management and communications technology in support of the healthcare delivery process across the nation will be difficult [3]. As we move from the individual patient's health record to those contained in the entire practice of that patient's primary care physician and the hospital at which that physician practices, to the entire health system in which that hospital participates, to the entire region in which that health system exists, to the entire nation, we will likely be forced to accept less precision in our measurements.
Following an overview of a conceptual model for the NHII, we present a draft measurement framework that would allow us to begin measuring progress towards the successful creation of a fully functional NHII. We will then briefly describe how we might also try to develop a qualitative estimate of the current state of the art regarding various information exchange standards, current and impending legislation, and the "values" of potential users of these systems.
Discussion
A conceptual model of the NHII
"Human endeavor is caught in an eternal tension between the effectiveness of small groups acting independently and the need to mesh with the wider community" [4]. The NHII can be thought of as a collection of healthcare delivery providers that share patient-level information electronically. More specifically, we conceptualize the NHII as a cluster of nodes. We define a node as a physical healthcare environment with the requisite health information management technology to collect, store, display and transmit patient-identifiable, structured, clinical data in an electronic format. Therefore, a sole practitioner in private practice using a simple, electronic health record (EHR) system who has access to the Internet could function as a node. On the other hand, we would also consider a large, academic medical center's inpatient facility as a single node, as well.
To create a functional network infrastructure, individual nodes must be connected in a way that permits sharing of information. Connections rely on the application of agreed upon conventions, or standards for describing clinical and administrative information (i.e., controlled vocabularies, standard identifiers) and for transmitting that information electronically (i.e., message exchange standards, e.g., HL-7 and X.12). The 1996 Health Insurance Portability and Accountability Act and subsequent regulations defined several standardized approaches [5]. Recent research on the growth and behavior of networks suggests we should anticipate significant increases in the capability of these networks as the number of connections grows [6].
We define a cluster as two or more nodes that have a) an existing written data sharing agreement and b) sent (or received) patient-identifiable information to (or from) any other node in the cluster – either directly or through an intermediary which in this case serves as a hub through which others share information. [Note: The process is more than a single claims submission transaction, users must also be able to retrieve, or at least view information from others.] A node may belong to one or more clusters (see Figure 1 for an illustration showing how nodes and clusters can be related). Those aspects of a cluster that contribute to their persistence also help to define clusters. For example, a cluster may be created and maintained by one or more of the following attributes: statutory, or legal, agreements, geographic proximity, or financial ownership. Using this definition, several existing Local Health Information Infrastructures (LHIIs) or Regional Health Information Organizations (RHIOs) would be considered clusters (e.g., the Indianapolis Network for Patient Care (INPC), the Santa Barbara County Care Data Exchange (CDE) or MA-SHARE (Massachusetts Simplifying Healthcare Among Regional Entities [7]).
Figure 1 A conceptual model of the National Health Information Infrastructure. A conceptual model of the National Health Information Infrastructure that illustrates different types of NHII clusters (i.e., one with peer to peer connections the other with a central repository). Once these clusters begin linking up that is the beginning of the NHII.
As we go forward, we anticipate that groups of clusters will form; therefore we add the additional proviso that a cluster can consist of a cluster of clusters. Such a model encapsulates the U.S. Federal government's current articulated plan for achieving a National Health Information Infrastructure (NHII) through the creation of Local or Regional Health Information Organizations (RHIOs) [8].
Key users (stakeholders) of the NHII
The National Committee on Vital and Health Statistics (NCVHS) defined key dimensions of the NHII functionality "by what they encompass, whom they serve, how they are used, and who has primary responsibility for content and control" [9]. These dimensions helped them identify three major groups of users of patient-identifiable health information: patients or consumers, healthcare providers (both individual clinicians and organizations) and communities (or population/public health). Therefore, we believe that we must make measurements with respect to each of these three groups of users.
Using the conceptual model to create a measurement framework
Now that we have a conceptual model for the NHII, we can begin developing a measurement framework that will help us evaluate the nation's progress toward achieving a functional NHII. Borrowing several concepts from conventional quality measurement efforts, we must be able to measure aspects of the structure, process, and outcomes that make up and result from the NHII [10]. These concepts translate into measurements of health information management technology availability, use, and effectiveness at both the nodal and cluster level. In addition, all of these measurements need to be made from the viewpoints of the key users of the NHII, namely, patients, clinicians, and those involved in population health activities (e.g., public health departments). Figure 2 helps illustrate this concept. As in any large-scale measurement and evaluation effort, designing and validating the measures will be one of the most important and difficult challenges to overcome.
Figure 2 A measurement framework for nodes and clusters with the NHII. An illustration of a measurement framework for nodes and clusters within the NHII showing the 3 axes along which measurements should be made: health information management technology availability, use, and effectiveness; NHII level, for example node and cluster; key users of the NHII, namely, patients, clinicians, and those involved in population health activities (e.g., public health departments).
Broadening the conceptual framework to help us better understand the field
The application of information technology to health care is still in its youth. Although much has been learned over the years, there are still many lessons to be learned about how and when a computer-based intervention is most likely to be successful. The scope of harms and benefits to be anticipated when information technology is implemented has not been well catalogued. The development of such understanding represents a key aspect of the formative evaluation of the move towards NHII. One further aspect of our measurement framework borrows from both the case study and the quality improvement frameworks. The accumulation of data from RHIOs and specific initiatives ought to enhance our understanding both of how and when to implement a specific type of intervention in a particular environment, and to improve the nature of the technological innovations themselves.
Phased approach to making measurements
In addition to the conceptual model of the system and identification of the key system users, we believe that we should use an iterative, phased approach, that will allow us to begin making measurements of the NHII, while we continue learning "how best to make these measurements". This iterative approach will also allow us to move forward at varying rates in different regions of the country. This is based on our firm belief that before one can expect to demonstrate improvements in any of the outcome measures associated with the NHII, we must first demonstrate that the key system users are actually using the system. Similarly, we believe that before we can expect to be able to measure any system use, we must be able to demonstrate that the requisite systems are in place and available to our key users. Therefore, we propose a three-phase iterative approach to beginning the measurements: Phase I will consist of the measurements required to demonstrate "Availability" of the systems; Phase II will consist of the measurements required to demonstrate "Use" of the systems; and Phase III will consist of the measurements required to demonstrate the effect of these systems on various outcome measures that are often associated with health information technology (HIT) use.
Phase I – Systems availability
HIT availability can be defined as the existence of, and access to, the requisite technology to collect, store, display and transmit patient-identifiable, structured, clinical data in electronic formats. Therefore, we must be able to identify whether healthcare institutions and their providers have access to various health information technology components. Potential measurements that we could make in this phase include:
• What is the coverage, or percentage, of patients in a region who have their health data available in an electronic format to qualified personnel? As our measurement techniques become more sophisticated, we also hope to be able to measure, or at least estimate, the "completeness" of each patient's health record, although at the present time the definition of a "complete" electronic medical record is still not precisely defined. [Note: On September 1, 2004, the American Health Information Management Association, Healthcare Information and Management Systems Society, and The National Alliance for Health Information Technology announced the formation of a Certification Commission for Healthcare Information Technology. Their charge is to create an efficient, impartial and trusted mechanism for certifying ambulatory electronic health records and other healthcare information technology (IT) products. It is possible that an EHR "completeness" measure could formulated based on their recommendations.]
• Use U.S. census data for a geographic region covered to estimate denominator.
• Use number of unique patient ID's accessible in the system(s) as the numerator.
• Goal: identify 3–5 levels of coverage.
• Number or percentage of clinicians with an RHIO login?
• Use number of unique clinicians with a log-in as numerator.
• Use state licensure records as an estimate of total clinicians in region eligible for logins.
• Goal to identify 3–5 levels.
• Number or percentage of health care organizations in a geographic region with a signed data exchange agreement with the RHIO in place.
• Use total number of healthcare organizations in community as denominator.
• Count the numbers of these LHIIs or RHIOs nationally – perhaps we could even go back a few years and make estimates for 2001–2003
Phase II – Systems use
HIT use can be defined as actual hands-on use of these HIT systems by patients, providers, and those involved in population health. At the nodal level this equates to actual use of various HIT applications such as clinical results review or provider order entry. At the cluster level, HIT use can be measured by the number of clinicians who routinely use the system to enter and review patient-level data. Example measurements we might be able to make here include:
• Patients in a region whose data was accessed by someone other than the originator of the data.
• Clinicians who actually logged-in to the system
• Healthcare institutions that submitted data to the RHIO
Phase III – Effect measurement
The effects of health-related information technology on health and health care represent a vital metric for the NHII. The value of the infrastructure ultimately must be evaluated perhaps using the six quality attributes defined in [11] (i.e., Safety, Timeliness, Efficiency, Effectiveness, Equitability, Patient-centeredness) as measurement axes. Although benefits and costs of HIT have been measured in limited settings, measurements of effect on the scale envisioned for a national infrastructure have never been made. We believe, however, that measurements of the impact of NHII on health outcomes are beyond the scope of our current charge and may distract us from the critical measurements of systems availability and use that must be performed first.
Paying for the RHIOs and the NHII
Clearly, HIT requires significant financial resources to create and maintain it. Therefore, we must be able to at least estimate how much each node or cluster has spent to create and maintain their systems and services and their source of financing. Using these financial estimates, we can then begin to compare different RHIO models based on their return on investment.
When the NHII is up and running
Once we have considerable (e.g., > 25%) penetration of the NHII, then we can begin using electronic, randomly determined, sampling methods of various aspects of the NHII systems to generate objective measures of IT availability, use and effectiveness. For example, we could send queries for a statistically significant number of patients' data (at least one patient in this group should have data from each hospital selected) to randomly selected hospitals and measure both the number and quality of responses received. The number of responses would tell us how many hospitals were able to at least respond to queries of this type, which is essential. The quality of the responses, that is the sensitivity and specificity of the patient matches and the amount and nature of the data returned would tell us how effectively, these institutions had implemented the functionality required to implement such a system.
• Randomly select a statistically significant number of patients and send electronic requests for information to a randomly selected set of healthcare institutions, pharmacies, or labs and count the number of replies. This would provide an estimate of the number of institutions that were capable of working in this system.
• Use the National Provider Identification (NPID) database to estimate number of duplicates as a measure of how well this database is being managed.
Additional measurement features
In addition to the measurements associated with elements of the conceptual model described earlier, we also believe that our measurements of NHII progress should include qualitative reviews of the current state of the art with regard to the legislation that is in place, or impending. Likewise, we believe that similar qualitative studies should be conducted on the state of clinical and administrative information exchange standards and on the "values" of potential users of these systems. While these qualitative estimates of progress will not be as easy to interpret, they provide at least a glimpse of the progress that the nation is making in these critical arenas.
Examples of the types of topics these qualitative reviews might address include:
• Qualitative assessment of the legal climate in each state to support NHII
• Patient privacy protections
• Legal restrictions on sending/receiving various data types
• Electronic signatures
• Prescription transmission to pharmacies
• Legal restrictions on sending laboratory results to patients
• Requirements to submit data in electronic format to local, state, federal payers
• Availability of unique provider ID at federal level
Likewise in assessing the values of key system users one might delve into:
• Qualitative assessment of the perceived value of using HIT for patient care
• Incentives to adoption
• Number of insurance companies reimbursing physicians for use of e-visits
How will we define these measurements?
We recommend that the United States department of Health and Human Services (HHS) convene and co-sponsor an impartial, public-private partnership group, such as an Institute of Medicine (IOM) committee, to create clear and consistent definitions of the components of the NHII (e.g., RHIO, EHR, CPOE) as a basis for the further refinement of specific metrics. This group should be lead by a recognized leader in the field of clinical information systems measurement and evaluation. The remainder of the group should include experts who have expertise in, and can represent and advocate for various measurement perspectives. Key stakeholders from various governmental agencies, healthcare delivery systems, and patient advocacy groups should advise this group.
The group should be charged with defining a set of metrics and developing a methodology to test their reliability and validity. They should release a base set of metrics as soon as they are defined and agreed upon. They should also work to define additional test metrics that are released, but not required to be made for the initial baseline estimates.
Who will make these measurements?
A public-private partnership should be charged with further development of these measurement systems, making the measurements, and reporting the results of these measures on a yearly basis.
For example, a nascent group referred to as the Improve-IT Institute has been formed by two of the authors (DFS and KL). Briefly, ImproveIT is an international coalition of institutions and individuals focused on measuring the progress in adoption and utilization of clinical information technology.
How can or should these measurements be made?
Making measurements of such a multi-faceted, multi-functional set of disparate systems and services will be difficult. Until we have considerable penetration in all aspects of these systems (i.e., inpatient, outpatient, data interchange standards, and unique patient ID mechanisms) measurements will need to be estimated from survey or site visit data.
What should be done first?
We should begin creating a multi-level inventory of NHII components including functionality and interoperability. This inventory should also include estimates of the population covered. We should also create a website documenting existing RHIOs and a means for sharing best practices.
Summary
The NHII has the potential to transform health care in America – improving health care quality, reducing health care costs, preventing medical errors, improving administrative efficiencies, reducing paperwork, and increasing access to affordable health care. While the President has set an ambitious goal of assuring that most Americans have electronic health records within the next 10 years, a significant question remains "How will we know if we are making progress toward that goal?" Using the definitions for "nodes" and "clusters" developed in this article along with the resulting measurement framework, we believe that we can begin a discussion that will enable us to define and then begin making the kinds of measurements necessary to answer this important question.
List of abbreviations
NHII – National Health Information Infrastructure
RHIOs – Regional Health Information Organizations
EHRs – Electronic Health Records
NHIN – National Health Information Network
LHII – Local Health Information Infrastructure
INPC – Indianapolis Network for Patient Care
CDE – Care Data Exchange
MA-SHARE – Massachusetts Simplifying Healthcare Among Regional Entities
NCVHS – National Committee on Vital and Health Statistics
HIT – Health Information Technology
IT – Information Technology
NPID – National Provider Identification
HHS – United States department of Health and Human Services
IOM – Institute of Medicine
CPOE – Computer-based Provider Order Entry
Improve-IT – Indices to Measure Performance Relating Outcomes, Value and Expenditure generated from Information Technology
Competing interests
The Improve-IT Institute has been formed, and is owned, by two of the authors (DFS and KL). Their goal is to further develop the ideas outlined in this manuscript and begin making and reporting the results of these measurements on their website. All of the other authors declare that they have no competing interests.
Authors' contributions
DFS, RNS, KL, CF, BR, GH, LLA, and LCK participated in several conference calls during which the concept for the paper and main ideas were conceived. DFS drafted the manuscript. DFS, RNS, KL, CF, BR, GH, and RK participated in the review and discussion of the "Metrics" topics described in this article at The Secretarial Summit on Health Information Technology Launching the National Health Information Infrastructure 2004: Cornerstones for Electronic Healthcare" held in Washington, D.C., July 20–23, 2004. DFS, RNS, and RK developed the final recommendations. All authors read and approved the final manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
==== Refs
Bush GW Transforming Health Care: The President's Health Information Technology Plan April 26, 2004.
National Health Information Infrastructure (NHII) Frequently Asked Questions (FAQ) Sheet Accessed 5/17/05
Ukoumunne OC Gulliford MC Chinn S Sterne JAC Burney PGJ Methods for evaluating area-wide and organisation-based interventions in health and health care: a systematic review Health Technology Assessment 1999 3 1 98
Berners-Lee T James Hendler J Lassila O The Semantic Web: A new form of Web content that is meaningful to computers will unleash a revolution of new possibilities Scientific American 2001
Health Insurance Portability And Accountability Act of 1996 PUBLIC LAW 104-191. AUG. 21, 1996
Barabási AL Linked: The New Science of Networks 2002 PerseusPublishing, Cambridge, MA
National Health Information Infrastructure (NHII) State and Local Activities Accessed 5/17/05
Yasnoff WA Humphreys BL Overhage JM Detmer DE Brennan PF Morris RW Middleton B Bates DW Fanning JP A consensus action agenda for achieving the national health information infrastructure J Am Med Inform Assoc 2004 11 332 8 15187075 10.1197/jamia.M1616
National Committee on Vital and Health Statistics A Strategy for Building the National Health Information Infrastructure Washington, DC November 15, 2001 (Accessed 6/7/04)
Donabedian A The quality of care. How can it be assessed? JAMA 260 1743 8 1988, Sep 23–30 3045356 10.1001/jama.260.12.1743
Institute of Medicine IOM), Committee on Quality of Health Care in America Crossing the quality chasm: a new health system for the 21st century 2001 Washington (DC): National Academy Press
| 15953388 | PMC1177954 | CC BY | 2021-01-04 23:52:11 | no | BMC Med Inform Decis Mak. 2005 Jun 13; 5:14 | utf-8 | BMC Med Inform Decis Mak | 2,005 | 10.1186/1472-6947-5-14 | oa_comm |
==== Front
BMC Musculoskelet DisordBMC Musculoskeletal Disorders1471-2474BioMed Central London 1471-2474-6-301595815710.1186/1471-2474-6-30Research ArticleDynamic in vitro measurement of patellar movement after total knee arthroplasty: an in vitro study Ostermeier Sven [email protected] Olaf [email protected] Christof [email protected] Christina [email protected] Department of Orthopaedics Hannover Medical School (MHH) Hannover, Germany2005 15 6 2005 6 30 30 28 2 2005 15 6 2005 Copyright © 2005 Ostermeier et al; licensee BioMed Central Ltd.2005Ostermeier 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
Changing the kinematic behaviour of patellar movement could be one of the reasons for anterior knee pain after implantation of a total knee arthroplasty (TKA). The aim of the current study was to measure the potential influence on patellar kinematics of patellar resurfacing during TKA.
Methods
Patellar movement before and after TKA with and without patellar resurfacing was measured under dynamic conditions in an in vitro cadaver simulation. Physiologic Musculus quadriceps forces were applied to five physiologic human knee specimens undergoing simulated isokinetic extension motions, patellar movement was measured using an ultrasonic measurement system. Thereafter, the Interax® I.S.A.-prosthesis system was implanted without and with resurfacing the patella, and patellar movement was again measured.
Results
The physiologic patella center moved on a semilunar path up to 6.4 mm (SD 6.4 mm) medially during extension. After TKA, the unresurfaced patella showed significantly less medial translation (p = 0.04) than the resurfaced patella. Subsequent resurfacing of the patella then resulted in a return to mediolateral positioning of the patella similar to the physiological case, whereas the resurfaced patella tilted up to twice as much as physiologic.
Conclusion
The results of this study suggest that resurfacing of the patella during TKA can result in a restoration of the physiologic mediolateral shift of the patellofemoral joint while angulation of the patella remains unphysiologic.
==== Body
Background
Patellofemoral or anterior knee pain represents one of the most common problems during rehabilitation after total knee arthroplasty (TKA), with postoperative problems reported in 0,5–12% of patients [1-6]. Although the components of the prosthesis may have been implanted correctly, with postoperative radiographs showing no malalignment, patients were often unable to flex the knee or bear weight on the treated joint [5,6]. While many authors observed a theoretical advantage of primary resurfacing of the backside of the patella during TKA in avoiding these problems [1-4,7], others disagreed, and do not recommend routine resurfacing of the patella [8-10]. Furthermore, some authors suggest that only through a correctly implanted patella inlay it is possible to avoid changing patellar tracking in the femoral groove and with it the associated potential malalignment of the degenerated patella [1,11-13].
Several types of retropatellar resurfacing devices have been developed. These included symmetrical, dome shaped, and non anatomic all polyethylene components [3]. Other devices had a metal-backed baseplate with a rotating, asymmetrical high-conforming polyethylene prosthesis [14]. In-vitro studies utilizing pressure-sensitive films showed significantly higher stresses in these designs which exceeded the yield strength on the polyethylene (>12 MPa) [9,14-17]. Furthermore, other studies showed substantial changes in movement of the patella after TKA: An unresurfaced patella moved along a different path in the groove of the femoral component compared to the physiological motion [1]. The patella shifted less medially than the physiologic patella in the physiologic femoral groove, although the angulation of the patella was not changed [1]. Thus, while demonstrating the same amount of flexion around the horizontal axis and rotation around the sagittal axis, the resurfaced patella moved on the same path as a physiologic patella, while at the same time exhibiting a significantly greater lateral tilt around the frontal axis [11-13].
Nonetheless, none of these studies compared the kinematics of the physiologic patella to those of an (un)resurfaced patella after TKA under physiologic loads and dynamic conditions. This study was therefore conducted to measure the path and rotational movement of a physiologic patella, and to compare that motion to the unresurfaced and resurfaced patella after TKA.
Methods
Five (5) fresh frozen left knee specimens (mean age 62, range 52–75 years) were mounted into a specially designed knee simulator in which isokinetic flexion-extension motions were simulated as described by Stukenborg-Colsman et al. before [18]. As this test setup performed an in vitro test no ethic approval was necessary. Each knee specimen was transected approximately 30 cm distal and proximal to the knee joint line. The skin and subcutaneous tissues were removed, preserving the muscles, articular capsule, ligaments, and tendons. Movement of the patella in relation to the femur was measured using an ultrasound based three-dimensional motion analysis system (Zebris CMS-100, Isny, Germany). Marker triangles with a leg length of 50 mm were attached to the femur and the patella by means of uni-cortical screws (Fig. 1). Each marker triangle consisted of three ultrasound emitters placed on the edges of each triangle, and marker position was dynamically measured at a sampling frequency of 10 Hz by an H-shaped array of microphones placed at a distance of 1500 mm medial to the specimen. With this setup, the ultrasound positioning system is reported by the manufacturer to produce a theoretical accuracy of 0.1 mm for translation and 0.1 degrees for rotation of the triangles. The motions of each triangle were geometrically transformed to the geometrical center of the bony patella and the center of the femoral epicondylar axis. Thus, the relative movement of the center of the patella with respect to the center of the femur was measured with directions defined as follows: mediolateral [x], proximodistal [y] and anteroposterior [z] translation. An internal rotation along the long axis of the patella was defined as tilt, a forward rotation along the horizontal axis as flexion and an external rotation along the sagittal axis as rotation (Fig. 2).
Figure 1 Implanted components with attached marker triangles
Figure 2 Directions of translation and rotation of the patellar center
The knee specimens were oriented with the femur fixed horizontally and the patella facing downwards. The tibia was attached to the simulator at mid-length by means of a linear-rotational bearing which permits axial sliding and turning as well as rotation transverse to the axis of the tibia. The bearing in turn was attached to a swing arm which allows varus-valgus rotation. The swing arm was equipped with a strain-gage based load measuring device which allows the extending moment applied to the tibia to be monitored continuously as described by Stukenborg-Colsman et al. [18].
Movement of the tibia was generated by the coordinated activation of two hydraulic cylinders, one to simulate quadriceps muscle force, and the other to apply external flexion moment. The quadriceps force was transmitted through a special clamp which was attached on the quadriceps tendon. An isokinetic extension cycle is simulated from 120 degrees flexion to full extension of the knee with an extension moment of 31 Nm [18]. During this extension cycle the cylinder which simulated the quadriceps force, was required to generated forces of up to 2000 N.
After measuring the movement of the physiologic patella a total knee arthroplasy was performed on the specimens (Interax® I.S.A. size 500, Stryker/Howmedica, Limerick, Ireland). The femoral component, which has straight femoral groove and multiple radii in sagittal plane, was implanted with an external rotation of 3 degrees relative to the posterior femoral condyles, the tibial component was implanted at zero degrees to the mechanical axis with no tibial slope of the tibial baseplate. The mobile bearing inlay used is capable of sliding anterior-posterior as well as rotating on the tibial baseplate. A prosthesis was fitted to each specimen in conjunction with a 8 mm PE-inlay, whereby the implantation was performed according to the manufacturers guidelines for opening the knee joint using a mediopatellar incision. The patella was left unresurfaced and relative movement of the center of the patella was measured a second time.
After this test cycle, the patella was resurfaced with the optional included patella inlay. This inlay is a all-polyethylene inlay with a symmetrical shape in the sagittal plane and a 1.5 times wider lateral than medial facet (Fig. 3). The patella surface was resected at a height of 9 mm to allow a size 300 patella-inlay to be fitted and to restore the original thickness of the unresurfaced patella. The inlay was fixed with bone cement with a medial offset of 5 mm relative to the bony center of the patella. Again the relative movement of the center of the patella was measured.
Figure 3 PE patellar inlay of the Interax®-Prosthesis system set in the femoral component groove, showing the wide lateral flange
After calculating the path and the relative rotational motion of the center of the patella, means were compared using the nonparametric Wilcoxon test for repeated measures at a significance level of α = 0.05.
Results
The physiologic patella (PHY) was observed to move on a semilunar path in the coronal plane (Fig. 4). Relative motions to the femoral centre in the physiologic femoral groove of up to 6.4 mm (SD 6.4 mm) medially were recorded. The most medial point relative to the starting point at 120 degrees of knee flexion was reached at 14.0 degrees flexion after which the patella again moved laterally with further extension (Table 1, Fig. 5). The unresurfaced physiologic patella after TKA (BI) had significantly less (p = 0.04) maximum medial movement of 2.4 mm (2.4 mm, SD 5.9 mm, p = 0.04), with its most medial position being achieved at full extension. The path of this patella was less curved and proceeded almost linearly towards full extension. With a polyethylene resurfacing (TRI) the patella moved on a similar path to the physiologic patella with a maximum medial shift of 5.4 mm (SD 7.1 mm, p = 0.50) being attained. Movements along the Y-axis did not significantly differ with dependence on TKA or patella resurfacing (Fig. 6). Movement along the Z-Axis, which represented the proximodistal direction decreased after TKA. The physiologic patella moved up to 48.9 mm (SD 17.0 mm) proximal while after implantation the unresurfaced patella moved only up to 24.8 mm (SD 20.4 mm, p = 0,33) and the resurfaced patella up to 30.8 mm (SD 19.3 mm, p = 0,47) (Fig. 7). The physiologic patella (PHY) continuously internally tilted up to 10.0° (SD 12.4°) during extension of the knee (Fig. 8). After TKA the patella showed similar (p = 0.34) continuous internal tilting of up to 7.5° (SD 8.7°) until full extension. The resurfaced patella (TRI) showed an internal tilting up to 19.3° (SD 21.7°) during extension (p = 0.50). No significant differences in rotations about the X- and Y-axis of the patella were observed (Fig. 9 and 10).
Figure 4 Patella path in coronal plane relative to the starting position (PHY = physiologic patella, BI = unresurfaced patella after TKA, TRI = resurfaced patella after TKA). •(dot) indicates 120° knee flexion, end maximum extension.
Table 1 Maximum movements and rotations of the physiologic (PHY), unresurfaced (BI) and resurfaced (TRI) patella after TKA.
X-axis Y-axis Z-axis Tilt Flexion Rotation
MAX med. SD MAX ant. SD MAX prox. SD MAX int. SD MAX SD MAX int. SD
[mm] [mm] [mm] [mm] [mm] [mm] [degr.] [degr.] [degr.] [degr.] [degr.] [degr.]
PHY 6.4 6.4 48.9 11.7 48.9 17.0 10.0 12.4 77.0 4.9 15.9 21.2
BI 2.4 5.9 39.9 14.2 24.8 20.4 7.5 8.7 73.5 5.9 19.5 18.2
TRI 5.4 7.1 53.0 6.3 30.8 19.3 19.3 21.7 67.9 6.5 16.8 17.8
Figure 5 Mediolateral shift of the physiologic (PHY), unresurfaced (BI) and resurfaced patella (TRI) after TKA.
Figure 6 Anteroposterior shift of the physiologic (PHY), unresurfaced (BI) and resurfaced patella (TRI) after TKA.
Figure 7 Proximodistal shift of the physiologic (PHY), unresurfaced (BI) and resurfaced patella (TRI) after TKA.
Figure 8 Tilting about the Z axis of the physiologic (PHY), unresurfaced (BI) and resurfaced patella (TRI) after TKA.
Figure 9 Rotation about the Y axis of the physiologic (PHY), unresurfaced (BI) and resurfaced patella (TRI) after TKA.
Figure 10 Flexion about the X axis of the physiologic (PHY), unresurfaced (BI) and resurfaced patella (TRI) after TKA.
Discussion
In this study, the path and rotation of the centre of the physiologic patella as well as that of the unresurfaced and resurfaced patella after TKA were measured under dynamic conditions in an in vitro knee-extension simulation with physiological quadriceps force. The physiologic patella moved in a semilunar bow shaped path from its starting position to a more medial position, before finally moving back to its original medial-lateral position in the course of the from flexion to full extension knee motion. These results correlated with the findings of Patel et al. who showed a similar path of the patella in vivo with a maximum medial shift of 3.2 mm at 30 degrees of flexion [19]. During extension, the patella was observed to tilt medially up to 4.2 degrees.
A limitation of this study of patellar movement is that the simulated extension cycle did not include a weight bearing component, and that the co-contraction of the hamstrings which could also have an additional effect on the patellar path were not considered. Furthermore, the individual contribution of the M. vastus medialis and lateralis were not considered and the resulting quadriceps force vector was directed along the axis of the femur. Nonetheless, unlike other in vitro simulations, physiological muscle forces were applied (up to 1500 N), and the kinematics of knee motion attained using this simulator have been shown to be similar to physiological on physiological specimens [20]. In addition, measurement of patellar movement revealed only motions relative to the individual starting points of each patella at 120 degrees of flexion. But in this study as well, motion of the physiologic patella was comparable to the patellar path reported for physiologic patellae of in vivo investigations [13]. In the present study, we observed significant changes in medial shift of the patella in the femoral groove after TKA. The unresurfaced patella shifted significantly less medially during extension accompanied by less of a bow shaped path in the femoral groove than compared to the physiologic patella and the resurfaced patella. These differences may be explained by the fact that the femoral groove of the Interax®-prosthesis used in this study is symmetrical, in contrast to the physiologic joint. This symmetrical femoral groove functions as a new guideline for the patellar path and represents a simplification of the anatomic shape of physiologic femoral groove which has a higher lateral flange. In the physiologic knee, the patella is guided by this higher lateral flange which pushes the patella in a medial direction with extension. These results contrast to the findings of Omori et al., who showed no significant changes in movement of an unresurfaced patella after TKA with the Genesis®-Prosthesis system (Smith&Nephew Richards, Memphis, USA) [17]. The Genesis®-Prosthesis system has a more anatomical femoral groove with a higher lateral flange, which may explain why the unresurfaced patella moved more medially (physiologically) in that system.
The results of this study further showed that resurfacing of the patella resulted in a similar path with a medial shift similar to the physiologic knee. As the patella inlay of the Interax®-Prosthesis system has a wider lateral facet and an optimized surface fit to the femoral component groove, the patella shifted more medially in a bow shaped path of motion in a similar manner to the physiologic patella. With a resurfaced patella the path of motion observed correlated with the findings of other kinematic studies of patellar path after resurfacing the patella [1,6,12,13,15,17,21,22]. Nonetheless, while translation was similar to physiological, the resurfaced patella tilted two times more internally than the physiologic patella during extension, which was also in contradiction to findings of Omori et al. who again found no different tilting after resurfacing [17]. Again, as with the explanation of the different medial shift observed, the asymmetrically shaped inlay in the horizontal plane of the Interax®-Prosthesis investigated in this study may explain the larger tilt angles observed relative to those reported for the Genesis®-Prosthesis system which has a symmetrical inlay. Because of the asymmetrical shaped patella with a wider lateral flange of this system, the lateral facet could be oversized. As the patella component has a better congruity with the femoral groove than the unresurfaced patella, the patella is better guided, as evidenced by the restoration of mediolateral shift. Concomitantly, larger patellar tilt was seen as the relatively oversized lateral facet was leveled.
Furthermore, a decreased proximal shift of the patella was observed after TKA in the present study, which could potentially lead to a change of the tibiofemoral joint line and produce a patella baja. This may occur as a result of implanting a higher inlay than present in the knee joint before implantation in order to provide stable knee joint conditions, and could lead to increased contact forces on the patellar surface [23-26].
This study demonstrated the ability of a patella resurfacing to restore the physiological mediolateral shift of the patella after TKA. It is hypothesized that the restoration of the physiologic kinematics will result in less patellofemoral complications caused by maltracking of the patella [4]. To date, it is unclear if the higher internal tilt has a negative influence on patellar and knee kinematics, whereas it can be hypothesized that because of lateral oversizing, this internal tilt leads to higher mechanical stresses on the lateral facet of the patellar component [9,12,13]. In addition, the internal tilt could reduce the contact area of the patellar component, which could lead to additional higher contact stresses [17]. Further investigations are planned in order to verify whether this patellar resurfacing design does in fact result in improved in vivo kinematics and the clinically anticipated reduction in patellofemoral complications.
Conclusion
The results of this study suggest that resurfacing of the patella during TKA can result in a restoration of the physiologic mediolateral shift of the patellofemoral joint while angulation of the patella was left in an unphysiological state.
List of abbreviations
BI unresurfaced patella after total knee arthroplasty
MPa Megapascale
PE Polyethylene
PHY physiologic patella
SD standard deviation
TKA total knee arthroplasty
TRI resurfaced patella after total knee arthroplasty
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
Dr. Sven Ostermeier performed the test setup, observed the test cycles and made the biomechanic and statistic analysis. He also wrote the primary manuscript.
Olaf Buhrmester performed the test cycles and developed several mechanical devices for the positioning system.
Dr. Christof Hurschler helped with the biomechanical and statistical analysis. He performed the correction of the final manuscript.
Dr. Christina Stukenborg-Colsman performed the test setup and helped in terms of clinical background.
==== Refs
Anouchi YS Whiteside LA Kaiser AD Milliano MT The effects of axial rotational alignment of the femoral component on knee stability and patellar tracking in total knee arthroplasty demonstrated on autopsy specimens Clin Orthop 1993 170 177 8448937
Boyd AD Ewald FC Thomas WH Poss R Sledge CB Long-term complications after total knee arthroplasty with or without resurfacing of the patella J Bone Joint Surg Am 1993 75-A 674 681 8501082
Harwin SF Patellofemoral complications in symmetrical total knee arthroplasty J Arthroplasty 1998 13 753 762 9802660 10.1016/S0883-5403(98)90026-2
Ranawat C The patellofemoral joint in total condylar knee arthroplasty. Pros and Cons based on five- to tenyear follow-up observation. Clin Orthop 1986 205 93 99 3698397
Rand JA Current Concepts Review: The Patellofemoral Joint in Total Knee Arhtroplasty JBJS 1994 76-A 612 620
Yoshii I Whiteside LA White SE Milliano MT Influence of prosthetic joint line position on knee kinematics and patellar position J Arthroplasty 1991 6 169 177 1875209
Kajino A Yoshino S Kameyama S Kohda M Nagashima S Comparison of the results of bilateral total knee arthroplasty with and without patellar replacement for rheumatoid arthritis. A follow-up note. JBJS[Am] 1997 79 570 574
Abraham W Keefer J Greer RB Daubert H Buchanan JR Should the patellabe resurfaced in total knee arthroplasty Clin Orthop 1988 236 134
Benjamin JB Szivek JA Hammond BS Kubchandhani Z Matthews AI Anderson P Contact areas and pressures between native patellas and prosthetic femoral components J Arthroplasty 1998 13 693 697 9741448
Enis JE Gardener R Robledo MA Latta L Smith R Comparison of patellar resurfacing versus nonresurfacing in bilateral total knee arthroplasty Clin Orthop 1990 260 38 42 2225640
Chew JT Stewart NJ Hanssen AD Luo ZP Rand JA An KN Differences in patellar tracking and knee kinematics among three different total knee designs Clin Orthop 1997 87 98 9418625 10.1097/00003086-199712000-00013
Hsu RW Luo ZP Rand JA An KN Influence of Patellar Thickness on Patellar Tracking and Patellofemoral Contact Characteristics After Total Knee Arthroplasty J Arthroplasty 1996 11 69 79 8676121 10.1016/S0883-5403(96)80163-X
Rhoads DD Noble PC Reuben JD Tullos HS The effect of femoral component position on the kinematics of total knee arthroplasty Clin Orthop 1993 286 122 129 8425333
Buechel FF Pappas MJ Makris G Evaluation of contact stress in metal-backed patellar replacements. A predictor of survivorship Clin Orthop 1991 190 197 1959271
Buff HU Jones LC Hungerford DS Experimental determination of forces transmitted through the patello-femoral joint J Biomech 1988 21 17 23 3339023 10.1016/0021-9290(88)90187-X
Huberti HH W.C. H Patellofemoral contact pressures J Bone Joint Surg [Am] 1984 66 715 6725318
Omori G Koga Y Bechthold JE Gustilo RB Nakabe N Sasagawa K Hara T Takahashi HE Contact pressure and three-dimensional tracking of unresurfaced patella in total knee arthroplasty The Knee 1997 4 15 21 10.1016/S0968-0160(96)00230-X
Stukenborg-Colsman C Ostermeier S Hurschler C Wirth CJ Tibiofemoral contact stress after total knee arthroplasty: comparison of fixed and mobile-bearing inlay designs Acta Orthop Scand 2002 73 638 646 12553510 10.1080/000164702321039598
Patel VV Hall K Ries M Lindsey C Ozhinsky E Lu Y Majumdar S Magnetic resonance imaging of patellofemoral kinematics with weight-bearing J Bone Joint Surg Am 2003 85-A 2419 2424 14668513
Ostermeier S Hurschler C Stukenborg-Colsman C Quadriceps function after TKA--an in vitro study in a knee kinematic simulator Clin Biomech (Bristol , Avon ) 2004 19 270 276 15003342 10.1016/j.clinbiomech.2003.11.006
Glaser FE Gorab RS Lee TQ Edge loading of patellar components after total knee arthroplasty J Arthroplasty 1999 14 493 499 10428232 10.1016/S0883-5403(99)90107-9
Stiehl JB Komistek RD Dennis DA Keblish PA Kinematics of the patellofemoral joint in total knee arthroplasty J Arthroplasty 2001 16 706 714 11547368 10.1054/arth.2001.24443
Miller MC Zhang AX Petrella AJ Berger RA Rubash HE The effect of component placement on knee kinetics after arthroplasty with an unconstrained prosthesis J Orthop Res 2001 19 614 620 11518270 10.1016/S0736-0266(00)00043-7
Nagamine R Whiteside LA White SE McCarthy DS Patellar tracking after total knee arthroplasty. The effect of tibial tray malrotation and articular surface configuration Clin Orthop 1994 262 271 8020227
Petersilge WJ Oishi CS Kaufman KR Irby SE Colwell CWJ The effect of trochlear design on patellofemoral shear and compressive forces in total knee arthroplasty Clin Orthop 1994 124 130 7994950
Singerman R Davy DT Goldberg VM Effects of patella alta and patella infera on patellofemoral contact forces J Biomech 1994 27 1059 1065 8089160 10.1016/0021-9290(94)90222-4
| 15958157 | PMC1177955 | CC BY | 2021-01-04 16:32:04 | no | BMC Musculoskelet Disord. 2005 Jun 15; 6:30 | utf-8 | BMC Musculoskelet Disord | 2,005 | 10.1186/1471-2474-6-30 | oa_comm |
==== Front
BMC Musculoskelet DisordBMC Musculoskeletal Disorders1471-2474BioMed Central London 1471-2474-6-321596323210.1186/1471-2474-6-32Research ArticleA description of physical therapists' knowledge in managing musculoskeletal conditions Childs John D [email protected] Julie M [email protected] Phillip S [email protected] Maria L [email protected] Timothy W [email protected] Anthony [email protected] US Army-Baylor University Doctoral Program in Physical Therapy, Fort Sam Houston, San Antonio, TX, USA2 Department of Physical Therapy, Regis University, Denver, CO, USA3 Department of Physical Therapy, Texas Tech University, Lubbock, TX, USA4 Department of Physical Therapy, Los Angeles Air Force Base, Los Angeles, CA, USA5 Department of Physical Therapy, University of Pittsburgh, Pittsburgh, PA, USA2005 17 6 2005 6 32 32 22 11 2004 17 6 2005 Copyright © 2005 Childs et al; licensee BioMed Central Ltd.2005Childs 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
Physical therapists increasingly provide direct access services to patients with musculoskeletal conditions, and growing evidence supports the cost-effectiveness of this mode of healthcare delivery. However, further evidence is needed to determine if physical therapists have the requisite knowledge necessary to manage musculoskeletal conditions. Therefore, the purpose of this study was to describe physical therapists' knowledge in managing musculoskeletal conditions.
Methods
This study utilized a cross-sectional design in which 174 physical therapist students from randomly selected educational programs and 182 experienced physical therapists completed a standardized examination assessing knowledge in managing musculoskeletal conditions. This same examination has been previously been used to assess knowledge in musculoskeletal medicine among medical students, physician interns and residents, and across a variety of physician specialties.
Results
Experienced physical therapists had higher levels of knowledge in managing musculoskeletal conditions than medical students, physician interns and residents, and all physician specialists except for orthopaedists. Physical therapist students enrolled in doctoral degree educational programs achieved significantly higher scores than their peers enrolled in master's degree programs. Furthermore, experienced physical therapists who were board-certified in orthopaedic or sports physical therapy achieved significantly higher scores and passing rates than their non board-certified colleagues.
Conclusion
The results of this study may have implications for health and public policy decisions regarding the suitability of utilizing physical therapists to provide direct access care for patients with musculoskeletal conditions.
==== Body
Background
Musculoskeletal conditions account for roughly 25% of patient complaints in the primary care setting[1]. However, physicians have been shown to lack confidence in their evaluation and treatment skills of these patients [2-6]. Although its de-emphasis in medical school curricula has been repeatedly implicated,[1,5,7-9] almost half of American medical schools still do not require any formal training in musculoskeletal medicine[10]. This lack of confidence is reflected by poor performance on formal assessments of knowledge in musculoskeletal medicine[7] and less than optimal practice patterns for patients with musculoskeletal conditions[11]. Freedman and Bernstein[7] assessed knowledge in musculoskeletal medicine among 85 physicians during their first week of their internship following graduation from medical school using a standardized examination. The mean score was just under 60%, with only 18% of physicians scoring above a level determined orthopaedic program directors as the minimum threshold necessary to establish competency in musculoskeletal medicine in the primary care setting[7]. Matzkin et al[12] recently demonstrated similar suboptimal levels of knowledge in musculoskeletal medicine among medical students and residents. Except for orthopaedists, they also found that experienced physicians across a variety of specialties demonstrated less than adequate knowledge related to musculoskeletal medicine. The authors concluded that training in both medical school and non-orthopaedic residency training programs was inadequate, a sentiment that has been echoed elsewhere[13].
Considerable evidence supports the benefits of early access to physical therapy care [14-18]. In particular, physical therapists increasingly provide their services without physician referral (ie, direct access). Seventy percent of the public reports they would seek care from a physical therapist without physician referral for musculoskeletal conditions, [19] with 39 states having passed legislation supporting this mode of healthcare delivery [20]. Multiple studies have demonstrated that physical therapists can provide safe and cost-effective care for patients with musculoskeletal conditions in direct access practice settings, supporting the expansion of direct access physical therapy services [21-26]. For example, physician referral episodes of care reportedly increased physical therapy claims by 67%, office visits by 60%, and costs by 123% than when patients directly accessed physical therapy without physician referral[24].
Despite the curricular emphasis placed on the management of musculoskeletal conditions in physical therapy programs, to date few studies have described physical therapists' knowledge of the skills necessary to manage these patients in a direct access setting. A musculoskeletal written examination has been developed and validated for this purpose[7,27]. The examination has been administered to physician interns,[7] medical students and residents,[12] and a variety of physician specialists,[12] making it a pragmatic reference standard for the initial assessment of knowledge in managing musculoskeletal conditions among physical therapist students and licensed physical therapists. Therefore, the purpose of this study was to describe physical therapists' knowledge in managing musculoskeletal conditions using this examination. These data combined with clinical studies demonstrating the benefits of direct access physical therapy [21-26] may further clarify the role of physical therapists in direct access environments.
Methods
We used a cross-sectional design to describe knowledge in managing musculoskeletal conditions among physical therapist students and licensed physical therapists in the uniformed services. The study was approved by the Institutional Review Boards at Wilford Hall Medical Center in San Antonio, TX and at Texas Tech University in Lubbock, TX before subject recruitment and data collection began. All subjects provided informed consent prior to participation.
Based on an a priori sample size estimation, a total of 26 first-professional physical therapy programs accredited by the Commission for Accreditation of Physical Therapy Education were randomly selected for participation. Educational programs are rapidly transitioning to doctoral programs, with approximately 80% of programs having completed the transition to the Doctor of Physical Therapy degree or in the transitioning process at the time of the study [28]. Therefore, randomization was blocked by the degree to be conferred upon graduation: master's (n = 13) vs. doctoral (n = 13) to permit comparisons based on degree status. Program directors were contacted initially by email to inform them of their program's selection, describe the study procedures, and invite the program's participation. Students in these programs were in the terminal phase of their program's curriculum, defined as having completed substantial portions of the didactic curriculum and clinical affiliations. All licensed physical therapists in the four uniformed health services (U.S. Air Force, U.S. Army, U.S. Navy, and U.S. Public Health Service) with at least one year of clinical experience were also invited to participate.
Participants completed the identical examination originally developed by Freedman and Bernstein to assess knowledge in musculoskeletal medicine among physician interns, [7] and more recently administered to medical students, residents, and a variety of physician specialists [12]. The examination consists of 25 open-ended questions that were selected based on commonly encountered musculoskeletal diagnoses encountered in the primary care setting (ie, fractures and dislocations, low back pain, sciatica, and arthritis) and consideration of orthopaedic emergencies that warrant immediate referral to an orthopaedic surgeon or the emergency department (ie, compartment syndrome, hip dislocation, etc.) [7]. Additional details related to the development and validation of the examination are reported elsewhere [7,27].
The examination was administered in a web-based format using Web Surveyor, version 3.6 (Web Surveyor Corporation, Herndon, VA). No time limit was imposed to be consistent with previous methodology [7,12]. Subject confidentiality was strictly maintained through assignment of a unique computer-generated code. Administration of the examination was preceded by a brief demographic survey that queried patients as to their educational background, board-certification status (Orthopaedic and/or Sports Clinical Specialist designation via the American Board of Physical Therapy Specialties), experience in different practice settings, and familiarity with the studies by Freedman and Bernstein [7,27]. Data from any therapists familiar with the studies by Freedman and Bernstein [7,27] were excluded from analysis because the examination questions and answer key were published verbatim in these manuscripts.
Educational programs were requested to have participants complete the study in a group setting with a proctor present (eg, a computer lab) when possible. This would insure that participants did not use any outside resources (ie, textbooks, information available on the internet, personal communication, etc.) to assist them in answering the questions. To maximize participation, however, programs were alternatively given the option to have participants complete the study on their own if a computer lab or similar arrangement was unavailable, or if a participant was not available at the designated time. Licensed physical therapists were also asked to complete the study in a proctored setting. All participants were queried at the end of the study as to whether they used any outside resources to assist them in the completion of the examination. The results of the demographic survey and content of the examination were stored in a secure, password-protected centralized database for subsequent analysis.
Data analysis
A total of 6 judges, blinded to the demographic survey results and whether the participant was a physical therapist student or licensed physical therapist scored blocks of 4–6 questions, resulting in each question being scored by two raters. Judges were physical therapist faculty with considerable experience in providing direct access care for patients with musculoskeletal conditions. Each rater was also trained in the scoring procedures by one of the investigators. An overall score and passing rate were determined using identical procedures as those described by Freedman and Bernstein,[7] however a brief review is provided here. Each question was assigned a maximum possible of 1 point. Partial credit was assigned based on the criteria for partial credit outlined in the answer key [7]. Scores were not penalized for incorrect spelling. Sums of individual scores represented the overall score, which was multiplied by 4 to obtain a percentage score. Inter-rater reliability of the overall score was examined using the intraclass correlation coefficient (ICC), equation 3,1 [29]. The ICC and associated 95% confidence interval was 0.91 (0.89, 0.92). Given a sufficiently high reliability coefficient, only data from the first rater were used in the analysis. Using the results from a single rater is also consistent with the procedures utilized by Freedman and Bernstein [7]. Participants were judged to have passed the examination if their score exceeded the previously established threshold of 73.1% [7].
Descriptive statistics, including frequency counts for categorical variables and measures of central tendency and dispersion for continuous variables were calculated to summarize the data using SPSS for Windows 11.0.1 (Chicago, IL). Independent sample t-tests were used to directly compare differences in knowledge between educational programs conferring the doctoral versus master's degree and between licensed physical therapists who were board-certified and those who were not. Differences in the passing rates among the physical therapist subgroups were examined using the Pearson chi-square statistic. The alpha-level was established a priori to be 0.05 utilizing a two-tailed test.
Results
174 physical therapist students across 12 out of the 26 (46%) randomly selected programs volunteered and completed the study, representing 52% of students within these programs. The mean age of physical therapist student participants was 26.7 (3.3) (range = 22–40). Ninety-two percent of physical therapist student participants (n = 160) completed the study in a proctored setting. 63.8% of physical therapist students (n = 111) were enrolled in doctoral degree programs. 182 licensed physical therapists in the uniformed services completed the study, representing 44% of uniformed physical therapists. The mean age of the licensed physical therapist participants was 37.7 (6.7) (range = 25–55). The average years of experience was 8.7 (6.3) (range = 1–30) and 28.6% (n = 52) of licensed physical therapists were board-certified.
No participant reported having received assistance to complete the examination. One licensed physical therapist reported being familiar with the Freedman and Bernstein studies, [7,27] thus these data were removed from the analysis. No differences in performance on the examination were observed between participants who completed the examination in a proctored versus an un-proctored setting (p = 0.465). Therefore, all responses were included in the analyses.
Overall scores among the physical therapist students and licensed physical therapists are reported in Figure 1. We did not directly compare the results between physical therapists and physicians using inferential statistics because the data from these groups were derived from unrelated studies. However, the identical examination and similar procedures were used in these studies, thus it is reasonable to discuss our findings in relation to the previous data among physicians. To facilitate this discussion, we superimpose the overall scores among the different physician subgroups with those of the different physical therapist subgroups. This provides a frame of reference for visualizing possible differences in knowledge related to managing musculoskeletal conditions between physical therapists and physicians (Figure 1).
Figure 1 Overall scores on the musculoskeletal knowledge examination among physical therapist students, licensed physical therapists, and previous data using the same examination among physicians. All physician-related data was derived from Matzkin et al,[12] except data for the subgroup of physician interns, which was derived from Freedman and Bernstein[7]. PT = physical therapist, Phys = physician, OCS = Orthopaedic Clinical Specialist, SCS = Sports Clinical Specialist, DPT = doctoral physical therapy program, MPT = master's physical therapy program, Ortho = orthopaedics, Other = anesthesia, emergency medicine. ophthalmology, radiology, and transitional, FP = family practice, GS = general surgery, Res = Resident, Peds = Pediatrics, Med = internal medicine, Med stu = medical student, OB = obstetrics-gynecology, and Psy = psychiatry
Licensed physical therapists (n = 182) achieved an average score of 75.9%, with an overall passing rate of 67%. Licensed physical therapists who were board-certified achieved significantly higher scores and passing rates than their non board-certified colleagues (Table 2). Physical therapist students (n = 174) achieved an average score of 66.2%, with an overall passing rate of 24%, versus a 19% passing rate among physician interns [7]. Physical therapist students enrolled in programs conferring the doctoral degree achieved significantly higher scores than students enrolled in programs conferring the master's degree, although passing rates were statistically similar (Table 1).
Table 1 Performance on the musculoskeletal knowledge examination between physical therapists enrolled in a program that confers a master's vs. a doctoral degree. (Participants were judged to have passed if their score exceeded 73.1%[7].)
Degree Status (n = 174) Master's (n = 63) (95% CI) Doctoral (n = 111) (95% CI) p-value
Overall score 63.6 (60.6, 66.6) 67.6 (65.6, 69.6) .022
Passing rate (Overall score >.731) .21 (.11, .31) .26 (.18, .34) .416
Table 2 Performance on the musculoskeletal knowledge examination based on board-certification status. (Participants were judged to have passed if their score exceeded 73.1%[7].)
Board-certification (OCS and/or SCS) (n = 182) Yes (n = 52) (95% CI) No (n = 130) (95% CI) p-value
Overall score 81.3 (79.2, 83.4) 73.7 (71.9, 75.5) <.001
Passing rate .88 (.80, .97) .58 (.50, .67) <.001
Discussion
Physicians assessed in the study by Freedman and Bernstein had just begun their internship year, [7] and Matzkin et al [12] reported data from medical students, residents, and a variety of physician specialists. Given the spectrum of physician experience levels and specialties represented in previous studies, [7,12] these data offer a compelling reference standard for at least a preliminary discussion related to the preparation of physical therapists versus physicians with respect to managing musculoskeletal conditions. It also seems reasonable to make preliminary general observations about possible differences between physical therapists and physicians since we used the identical examination and administered the examination using similar procedures as those used in the previous studies [7,12].
Figure 1 reveals that both physical therapist students and licensed physical therapists tended to have higher levels of knowledge in managing musculoskeletal conditions than medical students, physician interns and residents, and all physician specialists except for orthopaedists. This trend may seem somewhat intuitive since topics related to managing musculoskeletal conditions are emphasized in physical therapy curricula. However, data were previously lacking to support this contention. It is important to consider that the physician data were derived from unrelated studies, [7,12] thus we discuss our results in general terms in relation to the previous studies among physicians [7,12]. The implication that physical therapists have higher levels of knowledge in managing musculoskeletal conditions than physicians provide impetus for further prospective research in this area.
It could be argued that performance among physical therapist students remains suboptimal, supported by the fact that physical therapist students overall achieved an average score of 66.2%. However, the average score among medical students and interns (the most comparable physician group) was 49% [12] and 60%, [7] respectively. One of the primary curricular areas more heavily emphasized in doctoral physical therapy educational programs is the differential diagnosis of these conditions, a proficiency necessary for competence in more autonomous practice settings such as primary care [30]. Although passing rates were statistically similar, overall scores among physical therapists enrolled in doctoral programs was significantly higher than for master's programs (Table 1). These data provide preliminary evidence that an increased focus on the diagnosis of commonly encountered musculoskeletal conditions and orthopaedic emergencies is occurring in the curricula of doctoral physical therapy programs. However, a threshold of 73.1% was established by orthopaedic program directors as a minimum level of knowledge necessary for competency in musculoskeletal medicine. Given similar passing rates, and in light of increasing availability of direct access care for patients with musculoskeletal conditions, orthopaedic curricula among doctoral physical therapy programs should continue to be enhanced.
Both physicians and physical therapists are at a relative early juncture in their clinical education upon graduation from medical school or physical therapy school, thus they might be expected to have scores below the level established by residency program directors. However, the licensed physical therapists in this study demonstrated higher levels of knowledge in managing musculoskeletal conditions than physical therapist students and all physician subgroups, except for orthopaedists (Figure 1). Licensed physical therapists achieved an average score of 75.9% and an overall passing rate of 67%. This seems to be markedly improved compared to the passing rate amongst all physician subgroups except orthopaedists [12]. Furthermore, most physicians, and with increasing frequency physical therapists, receive graduate medical education in the form of clinical residencies which lead to board certification. In fact, board certification in orthopaedic physical therapy represents the largest area of specialization by physical therapists [31]. One of the key findings from this study was that performance among licensed physical therapists who were board-certified was significantly better when compared to their non board-certified colleagues, lending further credibility to the physical therapist board-certification process, which was not initiated until the 1980s.
Several limitations should be considered. Similar to medical education, physical therapy educational programs do not utilize standardized curricula, thus exposure to didactic and clinical education experiences related to the management of musculoskeletal conditions differs. Physical therapists with a stronger background in this area may have achieved higher scores than with less exposure to an orthopaedic curricula. Content of the examination was also primarily focused on the differential diagnosis of commonly encountered musculoskeletal diagnoses in a primary care setting (ie, fractures and dislocations, low back pain, sciatica, and arthritis) and orthopaedic emergencies that warrant immediate referral to an orthopaedic surgeon or the emergency department (ie, compartment syndrome, hip dislocation, etc.) [7]. Therefore, these data may not be generalizable to other physical therapy practice settings. We invited volunteer physical therapist students and licensed physical therapists to participate, thus the potential for selection bias cannot be excluded. However, physician participants in the study by Matzkin et al [12] were also volunteers, posing a similar limitation that likely mitigates any potential bias in discussing our results in relation to this study. Furthermore, although the examination in the Freedman and Bernstein study [7] was apparently completed by all physicians in the intern class, the examination was only administered to one class [7]. The fact that physical therapist students from a wide variety of programs and licensed physical therapists in geographical locations throughout the country participated in this study increases the generalizability of the findings. Future research could be performed to determine if the results demonstrated among licensed physical therapists in the uniformed services who participated in this study would be similar to the results among a group of civilian physical therapists.
Conclusion
The results of this study corroborate existing clinical studies demonstrating that physical therapists can provide safe and effective care for patients with musculoskeletal conditions in a direct access setting [21-26]. In comparison to previous studies among physicians, [7,12] physical therapists demonstrated higher levels of knowledge in managing musculoskeletal conditions than medical students, physician interns and residents, and most physician specialists except for orthopaedists. Physical therapist students enrolled in educational programs conferring the doctoral degree achieved higher scores than their peers enrolled in programs conferring the master's degree. Furthermore, licensed physical therapists who were board-certified achieved higher scores and passing rates than their colleagues who were not board-certified. Nevertheless, despite the benefits of early access to physical therapy [14-17] and favorable legislation in most states, [20] the primary barrier to patients receiving physical therapy services without physician referral is that claims are infrequently reimbursed by third party payers. Combined with existing evidence demonstrating that physical therapists are capable of providing safe and effective care for patients with musculoskeletal conditions in a direct access setting at a reduced cost to the healthcare system and employers, the results of this study may have implications for health and public policy decisions regarding the care of patients with musculoskeletal conditions.
Competing interests
None of the authors of this manuscript have any relevant conflict of interest, financial or otherwise. This study was supported by a grant from the Sports Physical Therapy Section of the American Physical Therapy Association, Inc. The funding organization had no role in the design and conduct of the study, to include data collection; management, analysis, or interpretation of the data. The funding organization was also not involved in the preparation of this manuscript, nor has it been asked to review and/or approve this submission.
Authors' contributions
JC designed and coordinated the study, performed the statistical analysis, and drafted the manuscript. JW assisted with the study design and drafting of the manuscript. PS developed the web survey instrument and provided oversight for the technical aspects of the survey administration. MP coordinated with the first-professional programs and assisted in the data analysis. TF conceived the idea and assisted with study design and analysis. AD assisted with the study design and acted as a liaison to the program directors. All authors read and approved the final manuscript.
Disclaimer
The opinions or assertions contained herein are the private views of the authors and are not to be construed as official or as reflecting the views of the U.S. Air Force or Department of Defense.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
None
==== Refs
Pinney SJ Regan WD Educating medical students about musculoskeletal problems. Are community needs reflected in the curricula of Canadian medical schools? J Bone Joint Surg Am 2001 83-A 1317 1320 11568192
Camp BW Gitterman B Headley R Ball V Pediatric residency as preparation for primary care practice Arch Pediatr Adolesc Med 1997 151 78 83 9006533
Glazier RH Dalby DM Badley EM Hawker GA Bell MJ Buchbinder R Determinants of physician confidence in the primary care management of musculoskeletal disorders J Rheumatol 1996 23 351 356 8882045
Matheny JM Brinker MR Elliott MN Blake R Rowane MP Confidence of graduating family practice residents in their management of musculoskeletal conditions American Journal of Orthopedics (Chatham, Nj) 2000 29 945 952
Saywell RMJ O'Hara BS Zollinger TW Wooldridge JS Burba JL McKeag DB Medical students' experience with musculoskeletal diagnoses in a family medicine clerkship Med Teach 2002 24 186 192 12098439 10.1080/01421590220125303
Clawson DK Jackson DW Ostergaard DJ It's past time to reform the musculoskeletal curriculum Acad Med 2001 76 709 710 11448825
Freedman KB Bernstein J The adequacy of medical school education in musculoskeletal medicine J Bone Joint Surg Am 1998 80 1421 1427 9801210
Laskowski ER Moutvic M Smith J Newcomer-Aney K Showalter CJ Integration of physical medicine and rehabilitation into a medical school curriculum: musculoskeletal evaluation and rehabilitation Am J Phys Med Rehabil 2000 79 551 557 11083306 10.1097/00002060-200011000-00014
Bernstein J Alonso DR DiCaprio M Friedlaender GE Heckman JD Ludmerer KM Curricular reform in musculoskeletal medicine: needs, opportunities, and solutions Clin Orthop 2003 302 308 14612660
DiCaprio MR Covey A Bernstein J Curricular requirements for musculoskeletal medicine in American medical schools J Bone Joint Surg Am 2003 85-A 565 567 12637447
Ahern MJ Soden M Schultz D Clark M The musculo-skeletal examination: A neglected clinical skill Aust N Z J Med 1991 21 303 306 1953507
Matzkin E Smith ME Freccero CD Richardson AB Adequacy of education in musculoskeletal medicine J Bone Joint Surg Am 2005 87-A 310 314 15687152 10.2106/JBJS.D.01779
Akesson K Dreinhofer KE Woolf AD Improved education in musculoskeletal conditions is necessary for all doctors Bull World Health Organ 2003 81 677 683 14710510
Ehrmann-Feldman D Rossignol M Abenhaim L Gobeille D Physician referral to physical therapy in a cohort of workers compensated for low back pain Phys Ther 1996 76 150 156 8592718
Flynn T Fritz J Whitman J Wainner R Magel J Butler B Rendeiro D Garber M Allison S A clinical prediction rule for classifying patients with low back pain who demonstrate short term improvement with spinal manipulation Spine 2002 27 2835 2843 12486357 10.1097/00007632-200212150-00021
Linton SJ Hellsing AL Andersson D A controlled study of the effects of an early intervention on acute musculoskeletal pain problems Pain 1993 54 353 359 8233552 10.1016/0304-3959(93)90037-P
Zigenfus GC Yin J Giang GM Fogarty WT Effectiveness of early physical therapy in the treatment of acute low back musculoskeletal disorders J Occup Environ Med 2000 42 35 39 10652686
Childs JD Fritz JM Flynn TW Irrgang JJ Johnson KK Majkowski GR Delitto A A clinical prediction rule to identify patients with low back pain who will benefit from spinal manipulation: A validation study Ann Intern Med 2004 141 920 928 15611489
Snow BL Shamus E Hill C Physical therapy as primary health care: public perceptions J Allied Health 2001 30 35 38 11265271
Association APT Direct access to physical therapist services: Is yours a direct access state? American Physical Therapy Association website 2004
Overman SS Larson JW Dickstein DA Rockey PH Physical therapy care for low back pain. Monitored program of first-contact nonphysician care Phys Ther 1988 68 199 207 2963349
Byles SE Ling RS Orthopaedic Out-patients: A fresh approach Physiotherapy 1989 75 435 437
Weale AE Bannister GC Who should see orthopaedic outpatients--physiotherapists or surgeons? Ann R Coll Surg Engl 1995 77 71 73 7574300
Mitchell JM de Lissovoy G A comparison of resource use and cost in direct access versus physician referral episodes of physical therapy Phys Ther 1997 77 10 18 8996459
Daker-White G Carr AJ Harvey I Woolhead G Bannister G Nelson I Kammerling M A randomised controlled trial. Shifting boundaries of doctors and physiotherapists in orthopaedic outpatient departments J Epidemiol Community Health 1999 53 643 650 10616677
Hattam P Smeatham A Evaluation of an orthopaedic screening service in primary care Clin Perform Qual Health Care 1999 7 121 124 10848384
Freedman KB Bernstein J Educational deficiencies in musculoskeletal medicine J Bone Joint Surg Am 2002 84-A 604 608 11940622
Association APT The DPT and You: The Future Is Now - Chat Transcript American Physical Therapy Association website 2004
Shrout PE Fleiss JL Intraclass correlations: Uses in assessing rater reliability Psychol Bull 1979 86 420 428 10.1037//0033-2909.86.2.420
Association APT Doctor of Physical Therapy (DPT) Degree Frequently Asked Questions American Physical Therapy Association website 2004
American Board of Physical Therapy Specialties (ABPTS) Specialist Certification (ABPTS Web Site) http://www apta org/Education/specialist 2004
| 15963232 | PMC1177956 | CC BY | 2021-01-04 16:32:04 | no | BMC Musculoskelet Disord. 2005 Jun 17; 6:32 | utf-8 | BMC Musculoskelet Disord | 2,005 | 10.1186/1471-2474-6-32 | oa_comm |
==== Front
BMC NeurolBMC Neurology1471-2377BioMed Central London 1471-2377-5-131601180210.1186/1471-2377-5-13Research ArticleA new approach to estimation of the number of central synapse(s) included in the H-reflex Ghavanini Mohammadreza Alavian [email protected] Alireza [email protected] Shahram [email protected] Mohammadreza [email protected] Department of physical medicine and rehabilitation, Shiraz medical school, Zand avenue, Shiraz, Iran2 Pain research group, Academic centre for education, culture and research, Iran medical science branch. No 31, Karimkhan Zand avenue, Shahid Hosseini alley, Multidiscipnilary pain clinic. Tehran, Iran2005 12 7 2005 5 13 13 8 2 2005 12 7 2005 Copyright © 2005 Ghavanini et al; licensee BioMed Central Ltd.2005Ghavanini 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
Among the main clinical applications of the H-reflex are the evaluation of the S1 nerve root conductivity such as radiculopathy and measurement of the excitability of the spinal motoneurons in neurological conditions. An attempt has been made to reduce the pathway over which H-reflex can be obtained in a hope to localize a lesion to the S1 nerve root, so the S1 central loop has been suggested. The main goal of this study is the estimation of the H-reflex number of synapse(s) for better understanding of the physiology of this practical reflex.
Methods
Forty healthy adult volunteers (22 males, 18 females) with the mean age of (37.7 ± 10.2) years participated in this study. They were positioned comfortably in the prone position, with their feet off the edge of the plinth. Recording electrodes were positioned at the mid point of a line connecting the mid popliteal crease to the proximal flare of the medial malleolus. Stimulation was applied at the tibial nerve in the popliteal fossa and H, F and M waves were recorded. Without any change in the location of the recording electrodes, a monopolar needle was inserted as cathode at a point 1 cm medial to the posterior superior iliac spine, perpendicular to the frontal plane. The anode electrode was placed over the anterior superior iliac spine, and then M and H waves of the central loop were recorded. After processing the data, sacral cord conduction delay was determined by this formula:
* Sacral cord conduction delay = central loop of H-reflex – (delays of the proximal motor and sensory fibers in the central loop).
Results
The central loop of H-reflex was (6.77 ± 0.28) msec and the sacral cord conduction delay was (1.09 ± 0.06) msec.
Conclusion
The sacral cord conduction time was estimated to be about 1.09 msec in this study and because at least 1 msec is required to transmit the signal across the synapse between the sensory ending and the motor cell, so this estimated time was sufficient for only one central synapse in this reflex.
==== Body
Background
The H-reflex was first described by Hoffmann in 1918 [1]. The major clinical application is evaluating the status of the peripheral nervous system with respect to proximal peripheral nerve conduction and potential entrapment of the S1 nerve root. The traditionally performed H-reflex has a very long pathway, reducing its ability to localize a lesion to the S1 nerve root. To overcome this obstacle, H-reflex studies were devised by stimulating the S1 nerve near the first sacral intervertebral foramen by Pease and coworkers [2,3] which was further investigated by others [4-6].
Despite the agreement on its usefulness, there is controversy regarding the synapses involved in its reflex arc. Many authors believe that this is a monosynaptic reflex [7,8]. However, some investigators have hypothesized that the reflex is mainly oligosynaptic [9-11].
In this study, we have tried to estimate the number of the involved synapse(s) in this reflex by calculating the conduction time across the sacral cord, using F-wave and the peripheral and central components of the H-reflex.
Methods
Forty five volunteers were selected to enrol in this study but due to asymmetric ankle reflex in two of them and inability to record H-reflex in three of them, five subjects were excluded from the study, so the final study group consisted of 22 men (55%) and 18 women (45%).
They had no low back pain and no previous history of neurologic problems, intervertebral disc problem, rheumatic diseases, diabetes and renal or metabolic diseases. It was assured that they had normal symmetrical plantar and achilles tendon reflexes, normal muscle strength, were able to tiptoe and walk on their heels and had no sensory deficit and negative straight leg raising (SLR) test.
After obtaining informed consent, the subjects were examined while relaxed in prone position. Examination was performed in room temperature with their skin warmed to reach normal temperature, if cold.
Dantec 2000C equipment was used. The recording electrodes were surface electrodes with 0.5 cm diameter. The active electrode was placed at the middle of the line connecting the popliteal crease to the medial malleolus. The reference electrode was placed 2 cm distal to it. The ground electrode was placed near the active pick-up electrode over the calves. The stimulator electrodes for stimulation at the popliteal fossa were surface electrodes with 0.5 cm in diameter and cathode-anode distance of 2 cm. The cathode was placed proximal to anode over the tibial nerve. Direct rectangular current pulses were used with a duration of stimulation of 1m sec for H-reflex and 0.2 msec for F and M waves. The stimulation frequency was 0.5 Hz for H-reflex and 1 Hz for F-wave. The amplifier had a filter frequency of 2 Hz to 10 KHz, sweep speed of 5 msec/ division and voltage sensitivity of (0.1–2) mv/division.
The soleus H-reflex was obtained with submaximal stimulation of the tibial nerve at the popliteal fossa. Then peak latency and base to peak amplitude of the H-wave were recorded.
With supramaximal stimulation, F and M waves were recorded. Moreover an averaged F-wave latency, onset latency and base to peak amplitude of M-wave were determined. For determination of averaged F-wave latency, at least 10 F-waves were recorded.
To obtain central H-reflex, stimulation was done using monopolar needle electrode for cathode and disc surface electrode with 0.5 cm diameter for anode. The cathode electrode was inserted at a point 1 cm medial to the posterior superior iliac spine, perpendicular to the frontal plane. After the needle touched the sacrum, it was slightly retracted. The anode electrode was placed over the anterior superior iliac spine. The pickup electrodes were not changed; then, stimulation was applied and increased until the largest H-wave could be seen and peak latencies and base to peak amplitudes of M and H waves were recorded (Figure 1).
Figure 1 Recording of M and H waves with the S1 root stimulation.
Finally, the distances were measured:
-Popliteal crease to the needle.
-Needle to the T12 spinous process.
-Popliteal crease to the T12 spinous
Process (via greater trochanter).
Then processing of the data was done as outlined below:
First, the conduction velocity of the proximal motor and sensory fibers were measured by these formulas:
F* : mean of F onset latencies (ms)
M**: onset latency of M-wave (ms)
Subsequently, conduction time in the afferent and efferent fibers in the central loop of H-reflex was calculated:
We considered the above distance equal to the length of S1 nerve root since the 12th thoracic spine is opposite the first sacral segment [12].
The sacral cord conduction time was calculated with this formula as:
Sacral cord conduction time = central loop of H-reflex – (delays of the proximal motor & sensory fibers in the central loop).
The analyses were performed using SPSS 10.0 software. Independent t-test was applied for statistical analysis of the data. P < 0.05 was considered significant.
Results
The results of stimulation of the tibial nerve in the popliteal fossa are displayed in Table 1. The results of direct stimulation of the S1 nerve root and recording from the soleus are shown in Table 2. The mean central loop of the H-reflex was obtained as (6.77 ± 0.28) msec. The results of distances between landmarks are summarized in Table 3. Finally, after calculation (as discussed previously), sacral cord conduction time in H-reflex was obtained as (1.09 ± 0.06) msec (Table 4).
Table 1 M, H&F waves in the tibial nerve stimulation at the popliteal fossa.
Wave Mean ± SD
M Onset latency (msec) 4.3 ± 0.5
Amplitude (mv) 4 ± 1.6
H Peak latency (msec) 34.4 ± 2.5
Amplitude (mv) 1.3 ± 0.7
F Onset latency(msec) 30.7 ± 1.5
Table 2 The M and H waves of the central loop H-reflex.
Parameter Mean ± SD
M-wave Peak latency (msec) 19.1 ± 1.7
Amplitude(mv) 1.4 ± 0.8
H-wave Peak latency (msec) 25.9 ± 1.8
Amplitude (mv) 0.9 ± 0.6
Central loop of H-reflex(msec) 6.77 ± 0.22
Intensity of stimulation (mamp) 41.9 ± 8.5
Table 3 Distances between landmarks.
Distance Mean ± SD
Monopolar needle to T12 spinous process (mm) 174 ± 8*
Monopolar needle to popliteal crease (mm) 567 ± 26*
Popliteal crease to T12 spinous process (mm) 742 ± 32*
* Significant difference between men & women (p < 0.0001).
Table 4 Conduction times and velocities in central loop of H-reflex.
Parameter Mean ± SD
Motor fiber Velocity (m/s) 58.3 ± 2.8
Conduction time(ms) 2.99 ± 0.15*
Sensory fiber Velocity (m/s) 64.9 ± 2.3
Conduction time(ms) 2.69 ± 0.13*
Sacral cord conduction delay (ms) 1.09 ± 0.06
* Significant difference between men & women (P < 0.05).
Discussion
The H-reflex is perhaps the most extensively studied reflex in the literature on human and mammalian neurophysiology. The relative ease with which the reflex can be elicited in muscles throughout the body, involving both spinal and cranial nerves, has also made the H-reflex an attractive clinical and research tool.
There is controversy regarding the synapses involved in the H-reflex. Many authors believe that this is a monosynaptic reflex arc [7,8]. For instance, Ertekin and coworkers stimulated the tibial nerve at the popliteal fossa and recorded from different lumbar epidural intervertebral levels. The time interval between the negative peaks of the ventral and dorsal root potentials was used to calculate the approximate sacral cord conduction time, which was found to be 1.3 msec. Thus they suggested that the reflex be exclusively monosynaptic [8]. In another study, Ertekin and coworker stimulated the tibial nerve and epidurally recorded the potentials. They proposed that the central conduction time of the soleus H-reflex could be about 1.1 msec[13]. The values of 1.3 and 1.1 msec that were obtained are close to the our value of 1.09 msec.
However, some investigators have hypothesized that the reflex is mainly oligosynaptic [9-11]. For example, Burke, et al. proposed that the rising phase of the increase in excitability of the soleus motoneuron pool produced by electrical stimulation of the tibial nerve lasts more than a few milliseconds and the increase in excitability takes several milliseconds to reach the threshold for motoneuron discharge. Therefore the H-reflex is unlikely to be exclusively monosynaptic[10]. Also, in another study they estimated the duration of the rise times of the excitatory post-synaptic potentials (EPSP) produced in soleus motoneurones by electrical stimulation to be 1.9 msec, so they recommended that H-reflex is not a purely monosynaptic reflex[9].
In contrast with some investigators that have hypothesized that there could be some oligosynaptic contributions to the H-reflex [9-11], it can be concluded that the soleus H-reflex is for the most part a monosynaptic reflex, composed of a single synapse between the group Ia fibers and the soleus motor neurons (Figure 2). In this study, the mean sacral cord conduction time was estimated to be about 1.09 msec, so it was sufficient for only one synapse because at least 1m sec is required to transmit the signal across the synapse between the sensory ending and the motor cell [5]. .
Figure 2 Schematic diagram of conduction time values along the dorsal & ventral roots and in the S1 sacral segment of the soleus H-reflex.
As stated earlier, studying the central loop of H-reflex is an important tool to differentiate central from peripheral lesions of the S1 spinal nerve, the normal value of which was (6.77 ± 0.28) msec in this study. This result is in line with the results of previous studies by Pease and coworkers, showing a normal value of (7 ± 0.3) msec[3], by Ghavanini and coworkers with normal value of (6.9 ± 0.4) msec [4], by Zhu and coworkers with normal value of (6.88 ± 0.33) msec [5] and by Sadeghi and coworkers with normal value of (6.78 ± 0.3) msec [6].
Conduction time in the afferent fibers of the central loop H-reflex,(2.69 ± 0.13) msec, and in efferent fibers, (2.99 ± 0.15) msec, were similar to estimations provided by Zhu and coworkers that showed 2.6 msec and 3.2 msec for the afferent and the efferent fibers, respectively [5]. The estimated length of S1 nerve root was (174 ± 8) mm. The resulting value is similar to estimation of Zhu and coworkers in 15 cadavers, that was (175 ± 3) mm [5]. Mean conduction velocity of the afferent (Ia) and the efferent fibers of the H-reflex was (64.9) m/s and (58.3) m/s, respectively. There was no significant difference between men and women in this regard. Its cause seems to be a) longer lower extremity in men and b) more time for proximal sensory and motor fibers conduction in men. (see Tables 3 and 4). Our results are similar to estimations provided by other methods[5,14].
The stimulus intensity to obtain the central H-reflex that had higher amplitude than the central M-wave, was significantly higher than necessary stimulation to obtain the peripheral H-reflex. This more intense stimulation may be due to the fact that the S1 spinal nerve emerges from the sacrum at the anterior aspect. Thus, the needle cannot approach it well from the posterior aspect. Furthermore, the large distance between cathode and anode causes current diffusion, and thereby increasing the best necessary stimulation.
Conclusion
This study, as many previous investigations, showed that the H-reflex is a monosynaptic reflex arc.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
-M.R.A.G : Supervision of the project and helping with calculation.
-A.A. : Data acquisition and testing, writing the paper.
-S.S. : Data acquisition and testing.
-M.R. E : Data acquisition and testing.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
The authors would like to thank Dr. N. SHOKRPOUR, Ph.D., for her great help in the revision of the manuscript; Miss HAGHIGHI for typing the article.
==== Refs
Hoffmann P Über die Beziehungen der Schenreflexe zur willkürlichen Bewegun zum Tonus Zoo Biol 1918 68 351 70
Pease WS Lagattuta FP Johnson EW Spinal nerve stimulation in S1 radiculopathy Am J Phys Med Rehabil 1990 69 77 80 2331343
Pease WS Kozakiewics R Johnson EW Central loop of the H-reflex. Normal Value and use in S1 radiculopathy Am J Phys Med Rehabil 1997 76 182 184 9207699 10.1097/00002060-199705000-00002
Ghavanini MR Ghadi RS Ghavanini AA The central loop of H-reflex in the S1 spinal nerve: Normal values and constitutional influencing factors Electromyogr Clin Neurophysiol 2001 41 259 262 11572185
Zhu Y Starr A Haldeman S Chu JK. Sugerman RA Soleus H-reflex to S1 nerve root stimulation Electroencephalogr Clin Neurophysiol 1998 109 10 14 11003059 10.1016/S0924-980X(97)00058-1
Sadeghi S Ghavanini MR Ashraf AR Jafari P Effects of age and leg length upon central loop of the gastrocnemius – soleus H-reflex latency BMC eurology 2004 4 11 15347428 10.1186/1471-2377-4-11
Magladery JM McDougal DB Electrophysiological studies of nerve and reflex activity in normal man: I. identification of certain reflexes in the electromyogram and the conduction velocity of peripheral nerve fibers Bull Johns Hopkins Hosp 1950 86 265 290 15414383
Ertekin C Mungan B Uludag B Sacral cord conduction time of the soleus H-reflex J Clin Neurophysiol 1996 13 77 83 8988288 10.1097/00004691-199601000-00008
Burke D Ganderia SC Mckeon B The afferent volleys responsible for spinal proprioceptive reflexes in man J Physiol 1983 339 535 552 6887033
Burke D Ganderia Sc Mckeon B Monosynaptic and oligosynaptic contributions to human ankle jerk and H-reflex J Neurophysiol 1984 52 435 448 6090608
Ertekin C Mungan B Ertas M Human root and cord potentials evoked by achilles tendon tap Electromyogr Clin Neurophysiol 1995 35 259 271 7498070
Warwick R Williams PL Gray's Anatomy 1989 37 Edinburg: Churchill Livingstone 947 948
Ertekin C Mungan B Sacral spinal cord and root potentials evoked by the stimulation of the dorsal nerve of penis and cord conduction delay for the bulbocavernosus reflex Neurourol Urodyn 1993 12 9 22 8481731
Zhu Y Starr A Su H Woodward KG Haldeman S The H-reflex to magnetic stimulation of lower limb nerves Arch neurol 1992 49 66 71 1309419
| 16011802 | PMC1177957 | CC BY | 2021-01-04 16:28:53 | no | BMC Neurol. 2005 Jul 12; 5:13 | utf-8 | BMC Neurol | 2,005 | 10.1186/1471-2377-5-13 | oa_comm |
==== Front
BMC NeurosciBMC Neuroscience1471-2202BioMed Central London 1471-2202-6-371592152910.1186/1471-2202-6-37Research ArticleEffects of lesions of the nucleus accumbens core on choice between small certain rewards and large uncertain rewards in rats Cardinal Rudolf N [email protected] Nathan J [email protected] Department of Experimental Psychology, University of Cambridge, Downing Street, Cambridge CB2 3EB, UK2005 28 5 2005 6 37 37 18 4 2005 28 5 2005 Copyright © 2005 Cardinal and Howes; 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
Animals must frequently make choices between alternative courses of action, seeking to maximize the benefit obtained. They must therefore evaluate the magnitude and the likelihood of the available outcomes. Little is known of the neural basis of this process, or what might predispose individuals to be overly conservative or to take risks excessively (avoiding or preferring uncertainty, respectively). The nucleus accumbens core (AcbC) is known to contribute to rats' ability to choose large, delayed rewards over small, immediate rewards; AcbC lesions cause impulsive choice and an impairment in learning with delayed reinforcement. However, it is not known how the AcbC contributes to choice involving probabilistic reinforcement, such as between a large, uncertain reward and a small, certain reward. We examined the effects of excitotoxic lesions of the AcbC on probabilistic choice in rats.
Results
Rats chose between a single food pellet delivered with certainty (p = 1) and four food pellets delivered with varying degrees of uncertainty (p = 1, 0.5, 0.25, 0.125, and 0.0625) in a discrete-trial task, with the large-reinforcer probability decreasing or increasing across the session. Subjects were trained on this task and then received excitotoxic or sham lesions of the AcbC before being retested. After a transient period during which AcbC-lesioned rats exhibited relative indifference between the two alternatives compared to controls, AcbC-lesioned rats came to exhibit risk-averse choice, choosing the large reinforcer less often than controls when it was uncertain, to the extent that they obtained less food as a result. Rats behaved as if indifferent between a single certain pellet and four pellets at p = 0.32 (sham-operated) or at p = 0.70 (AcbC-lesioned) by the end of testing. When the probabilities did not vary across the session, AcbC-lesioned rats and controls strongly preferred the large reinforcer when it was certain, and strongly preferred the small reinforcer when the large reinforcer was very unlikely (p = 0.0625), with no differences between AcbC-lesioned and sham-operated groups.
Conclusion
These results support the view that the AcbC contributes to action selection by promoting the choice of uncertain, as well as delayed, reinforcement.
==== Body
Background
Animals often need to choose between different courses of action on the basis of the eventual rewarding or reinforcing outcomes of those actions. However, the relationship between an action and an outcome is frequently uncertain: animals do not always obtain that for which they work. Therefore, animals must incorporate information on the probability of obtaining different rewards when making decisions about what to do. Little is known of the neural basis of this process. Furthermore, when making decisions under conditions of uncertainty, individuals vary as to how much uncertainty or risk they are willing to tolerate. Formally, individuals differ in how much they 'discount' the value of reinforcers as the uncertainty of the reinforcer increases (i.e. as the probability of the reinforcer declines, or the odds against obtaining the reinforcer increase) [1]. Risk taking is one aspect of the personality trait of impulsivity [2-4] and is a feature of a number of psychiatric disorders, including pathological gambling and certain personality disorders [5-8]. The term 'risk' implies exposure to the possibility of an aversive consequence [9], which may include the possibility of not obtaining an anticipated reward. In the appetitive domain, risk taking is exemplified by the tendency to choose large rewards that are very uncertain, in preference to smaller, certain rewards. Abnormal risk taking may reflect dysfunction of reinforcement learning systems that mediate the effects of uncertain reward or punishment.
The nucleus accumbens (Acb) is one candidate structure that may influence choice involving uncertainty. The Acb responds to anticipated rewards in humans, other primates, and rats [10-17], and is innervated by dopamine (DA) neurons that respond to errors in reward prediction in a manner appropriate for a teaching signal [18-21]. There is clear evidence that the Acb is involved in the processing of delayed reinforcement and its influence upon choice. Damage to the nucleus accumbens core (AcbC) produces impulsive choice in rats [22,23], reducing their ability to choose large, delayed rewards in preference to small, immediate rewards, yet these and other similar lesions do not appear to impair rats' ability to discriminate reward size [23-31]. Furthermore, AcbC lesions impair rats' ability to learn instrumental actions when the outcomes of those actions are delayed [24]. The Acb may also be involved in the processing of uncertain or probabilistic reinforcement. DA neurons that innervate the Acb may fire in a manner related to reward probability [32-34] and the midbrain, the site of the cell bodies of these neurons, responds to stimulus uncertainty in humans [35]. A greater blood flow response is observed in the human Acb during the selection of high-reward/high-risk options, compared to low-reward/low-risk outcomes, in a task where the risk is of not winning [36], with similar activation to high-reward/high-risk option selection in a task where the risk is of losing [37]; this latter activation was correlated with personality measures of harm avoidance. However, these studies are correlative, and it is not known whether the AcbC is causally involved in regulating choice involving uncertain reinforcement.
In the present study, we sought to examine the contribution of the AcbC to choice involving probabilistic reinforcement in rats. We trained rats on a task in which they could choose regularly between a certain, small reward and an uncertain, large reward in discrete trials (Figure 1) and made excitotoxic AcbC lesions before retesting the rats postoperatively. Preoperatively, the proportion of choice trials in which the large reinforcer was chosen was approximately a linear function of the large-reinforcer probability. Postoperatively, after a transient period in which AcbC-lesioned rats were relatively indifferent between the two reinforcers, compared to sham-operated controls, a stable state emerged in which AcbC-lesioned rats chose the large, uncertain reinforcer less often than shams did. This pattern persisted regardless of whether the large-reinforcer probability increased or decreased across the session. AcbC-lesioned rats and controls continued to exhibit a strong preference for the large reinforcer when it was consistently certain, and a strong preference for the small, certain reinforcer when the large reinforcer was very unlikely; the lesioned and sham-operated groups did not differ from each other in either of these conditions. These results suggest that the AcbC is necessary for the normal impact of unlikely (as well as delayed) reinforcers upon choice.
Results
Histology
There were four postoperative deaths. Histological analysis revealed that the lesions were incomplete or encroached significantly on neighbouring structures in two subjects. These subjects were excluded; final group numbers were therefore 6 (AcbC) and 12 (sham). Lesions of the AcbC encompassed most of the core subregion; neuronal loss and associated gliosis extended in an anteroposterior direction from approximately 2.7 mm to 0.2 mm anterior to bregma, and did not extend ventrally or caudally into the ventral pallidum or olfactory tubercle. Damage to the ventromedial caudate-putamen was occasionally seen; damage to the nucleus accumbens shell (AcbSh) was restricted to the lateral edge of the dorsal shell. Schematics of the lesions are shown in Figure 2. Photomicrographs of lesions with identical parameters have been presented before [24,38,39].
Preoperative choice
The groups remained matched for preoperative choice behaviour following later histological selection (Figure 3a). Choice ratios (percentage choice of the large reinforcer, for each trial block) calculated across sessions 10–12 (see Table 1) were analysed using the model lesion intent2 × (large-reinforcer probability5 × S). There was a robust effect of probability (F3.3,52.9 = 70.6, = .826, p < .001) but no effect of lesion intent and no lesion intent × probability interaction (Fs < 1, NS).
Early postoperative choice
In the initial postoperative period, AcbC-lesioned rats exhibited relative indifference between the two alternatives, choosing the large reinforcer close to 50% of the time at all large-reinforcer probabilities; as a result, AcbC-lesioned rats were more likely than shams to choose the large reinforcer when it was most uncertain (Figure 3b). An analysis of choice ratios calculated across sessions 13–15 was performed using the ANOVA model lesion2 × (probability5 × S). This revealed a lesion × probability interaction (F3.3,53.5 = 5.22, = .836, p = .002). Comparison of the two groups at individual large-reinforcer probabilities demonstrated that AcbC-lesioned rats chose the large/uncertain reinforcer more than shams at preinforcer = 0.0625 (pstatistical = .02), and at preinforcer = 0.125 (pstatistical = .009), but did not differ from shams at reinforcer probabilities of 0.25–1 (pstatistical ≥ .158). Nevertheless, simple effects of probability persisted both in shams (F2.8,30.9 = 32.3, = .702, p < .001) and in AcbC-lesioned rats (F4,20 = 5.37, p = .004). Choice at each preinforcer was compared to 50% (indifference) using post hoc two-tailed one-sample t tests, correcting pstatistical values using the Šidák correction for 5 comparisons. For shams, choice differed significantly from 50% at large-reinforcer probabilities of 0.0625 (when choice of the large reinforcer was less than 50%), 0.125 (less than 50%), and 1 (greater than 50%) (corrected pstatistical ≤ 0.007), but for AcbC-lesioned rats, choice did not differ significantly from 50% at any large-reinforcer probability (corrected pstatistical ≥ 0.81).
Final postoperative choice
By the final three sessions of the basic task (sessions 22–24; see Table 1), the pattern of choice in AcbC-lesioned rats had changed (Figure 3c). Once more, an analysis of choice ratios using the model lesion2 × (probability5 × S) revealed a lesion × probability interaction (F2.9,46.4 = 5.78, = .726, p = .002). By now, however, AcbC-lesioned rats did not differ from shams with reinforcer probabilities of 0.0625-0.25 (pstatistical ≥ .386) but chose the large reinforcer less than shams when its probability was 0.5 (pstatistical = .037) and 1 (pstatistical = .015). As before, effects of probability persisted both in shams (F2.3,24.9 = 49.5, = .565, p < .001) and in AcbC-lesioned rats (F4,20 = 9.45, p < .001).
Choice when both reinforcers were certain, or both uncertain
When the large and small reinforcers were both delivered with certainty, AcbC-lesioned and sham-operated rats strongly preferred the large reinforcer; when the small reinforcer was certain and the large reinforcer was consistently unlikely (preinforcer = 0.0625), all rats strongly preferred the small reinforcer (Figure 3d). There were no group differences in either case. This indicates that both AcbC-lesioned and sham-operated rats successfully discriminated the large reinforcer from the small reinforcer, and discriminated the certain large reinforcer from the uncertain large reinforcer. Choice ratios from the final sessions of training in these two conditions (sessions 34 and 52; see Table 1) were analysed using the model lesion2 × (trial block5 × S). In the 'certain' condition (session 34), there was no effect of lesion (F1,15 = 2.54, p = .132), no lesion × block interaction (F = 1.42, NS), and no effect of trial block (F1.5,21.9 = 2.12, = .365, p = .154). Similarly, in the 'uncertain' condition (session 52), there was no effect of lesion (F = 1.35, NS), no lesion × block interaction (F = 1.31, NS), and no effect of trial block (F < 1, NS).
Choice with ascending probabilities
After rats had been trained with the large-reinforcer probability increasing across the session, choice behaviour was similar to that with the decreasing-probability version of the task used initially, with AcbC-lesioned rats choosing the large/uncertain reinforcer less often than shams (Figure 3e; compare Figure 3c). Choice ratios from sessions 44–46 (see Table 1) were analysed using the model lesion2 × (probability5 × S). As before, there was a lesion × probability interaction (F4,64 = 9.29, p < .001), in addition to main effects of lesion (F1,16 = 19.5, p < .001) and probability (F4,64 = 95.6, p < .001), and there were strong effects of probability for both AcbC-lesioned rats (F1,20 = 20.7, p < .001) and shams (F3.1,34.4 = 119.6, = .781, p < .001). AcbC-lesioned rats differed from shams at reinforcer probabilities of 0.125 (pstatistical = .033), 0.25 (pstatistical < .001), 0.5 (pstatistical < .001), and 1 (pstatistical = .013), but not at preinforcer = 0.0625 (pstatistical = .881).
Postoperative choice: analysis by experienced probability
Since the task was genuinely probabilistic, and not pseudorandom, it is possible that the probabilities experienced by subjects differed from the programmed probabilities (although experienced probabilities inevitably tend towards programmed probabilities as the number of trials increases). For example, one subject choosing an uncertain reinforcer at preinforcer = 0.5 for 10 trials might experience 3 rewarded and 7 unrewarded trials (an experienced probability of 0.3), while another might experience 6 rewarded and 4 unrewarded (experienced preinforcer = 0.6). To establish whether such effects accounted to any degree for the pattern of choice observed in AcbC-lesioned and sham-operated rats, choice was re-analysed for four sets of sessions (preoperative sessions 10–12, early postoperative baseline sessions 13–15, late postoperative baseline sessions 22–24, and sessions 44–46 at the end of training on the increasing-probability version of the task; see Figure 4a–d, compared to the corresponding programmed-probability versions in Figure 3a–c,e). In each case, choice ratios were analysed using the model lesion2 × (experienced probabilitycov × S), with the factor × covariate term included in the model. Experienced probabilities were calculated for all trial types (forced and choice trials), across the sessions concerned.
These analyses confirmed the pattern of results obtained on the basis of programmed probabilities. For the preoperative sessions, as expected, there was a main effect of experienced probability (F1,54 = 319.1, p < .001) but no significant terms involving lesion intent (Fs < 1, NS). For the baseline (decreasing-probability) task, both early (sessions 13–15) and late (sessions 22–24) in the postoperative testing, there was a lesion × experienced probability interaction (early: F1,54 = 25.7, p < .001; late: F1,54 = 20.8, p < .001). For the increasing-probability schedule (sessions 44–46), there was no lesion × experienced probability interaction (F1,54 = 1.80, p = .185) but there was a main effect of lesion (F1,16.0 = 9.36, p = .007).
Indifference probabilities
Choice ratios from sham-operated rats on sessions 22–24 (the final 3 postoperative sessions on the basic task; see Table 1) were analysed using four different linear predictors, based either on the probability of delivery of a large reinforcer (given choice of the Large lever), or of the odds against delivery of a large reinforcer, calculated as odds against = (1 - p)/p. This established that choice patterns were predicted best, in linear fashion, by experienced probabilities (within-subject predictor allowing different slopes for each subject, r2 = 0.85) and programmed probabilities (r2 = 0.84), rather than by experienced odds (r2 = 0.61) or programmed odds (r2 = 0.67). Since optimal behaviour would give choice that was a step function of probability (i.e. it is optimal to choose the small/certain lever whenever the 4-pellet reinforcer is delivered with preinforcer < 0.25 and to choose the large/uncertain lever whenever preinforcer > 0.25), a single-parameter continuous function approximating a step function was also used to predict subjects' choice [the logistic function y = 100/e-(x-m)/b with y as the percentage choice of the large reinforcer, x as the programmed probability, b = 0.01 as an approximation to b = 0 and m as the free parameter], but this gave a poor fit (r2 calculated as SSmodel/SStotal for a nonlinear fit: mean r2 = 0.26; note that individual values of r2 can fall outside the range [0,1] when calculated this way for nonlinear models) [40]. Consequently, since choice was best described as a linear function of probability, indifference probabilities were calculated for sham-operated and AcbC-lesioned rats, namely the probability at which rats were equally likely to choose the small/certain and large/uncertain reinforcers. These were calculated via a linear regression of probability on choice (i.e. a regression in which probability was predicted from choice). This method has the potential to produce nonsensical probabilities for individual rats (if, for example, an individual's curve does not go both above and below the 50% choice point in a given set of sessions) but is nonetheless useful for group comparison. Experienced large-reinforcer probabilities (across all types of trials) were used, rather than programmed probabilities, though the pattern of results presented below was not altered by the use of programmed probabilities instead.
The main finding was that by the end of testing, AcbC-lesioned rats had higher indifference probabilities (0.70) than sham-operated rats (0.32) (Figure 5) – that is, while sham-operated rats behaved as if indifferent between a 1-pellet certain reinforcer and a 4-pellet reinforcer delivered with probability 0.32 (mathematically, an expected number of pellets of 0.32 × 4 = 1.28), AcbC-lesioned rats behaved as if indifferent between a 1-pellet certain reinforcer and a 4-pellet reinforcer delivered with probability 0.70 (an expected number of pellets of 2.8). That is, AcbC-lesioned rats appeared to exhibit risk aversion by the end of testing. The full analysis was as follows. Preoperatively (sessions 10–12), indifference probabilities were 0.43 ± 0.08 (AcbC) and 0.54 ± 0.09 (sham); these did not differ (F < 1, NS). In the initial postoperative period (sessions 13–15), indifference probabilities were numerically lower in the lesioned group, being 0.25 ± 0.28 (AcbC) and 0.59 ± 0.12 (sham), but indifference probabilities were highly variable in both groups and these did not differ (F1,16 = 1.76, p = .204). In the later postoperative period (sessions 22–24), indifference probabilities were higher in the lesioned group, being 0.75 ± 0.22 (AcbC) and 0.39 ± 0.15 (sham), though again these did not differ significantly (F1,16 = 1.90, p = .187). In the increasing-probability version of the task (sessions 44–46), indifference probabilities were again higher in the lesioned group, being 0.70 ± 0.15 (AcbC) and 0.32 ± 0.02 (sham). By this stage the difference was highly significant (F1,16 = 12.6, pstatistical = .003), even if corrected for four comparisons (pstatistical = .012) using the Šidák correction.
Omissions and latencies
Omissions were infrequent and not influenced by reinforcer probability or the lesion. Omission data from the final postoperative baseline sessions (sessions 22–24) were analysed. Overall, omissions (either failures to initiate a trial or to respond to an initiated trial) across all trial types occurred at a rate of 2.9 ± 0.9 % (sham) and 5.5 ± 1.9 % (AcbC). Omissions on choice trials for the same sessions were analysed using the model lesion2 × (probability5 × S). There were no effects of lesion (F1,16 = 1.95, NS) or probability (F1.6,25.2 = 2.56, = .394, p = .107), and no interaction (F = 1.04, NS). Almost all omissions were failures to initiate a trial (shams 0.9% of choice trials, AcbC 4.4%) rather than failures to respond once a trial had been initiated (shams 0.06% of choice trials, AcbC 0%).
Initiation latencies on choice trials for sessions 22–24 were analysed in the same manner. They were not affected by the lesion (F < 1, NS), nor by the large-reinforcer probability (F4,64 = 1.41, NS), and there was no lesion × probability interaction (F < 1, NS).
Response latencies were not affected by the lesion, but were affected both by the time in the session, with responding tending to get slower as the session progressed, and by the likelihood of obtaining a large reinforcer, with responding tending to get faster as large-reinforcer delivery became more likely. Response latencies on choice trials for sessions 22–24 were analysed using the model lesion2 × (trial block5 × choice2 × S). Response latencies varied across trial blocks: response latencies were initially 0.82 s (in the first trial block, when the large-reinforcer probability was 1) and slowed to 1.1 s (in the last trial block, when the large-reinforcer probability was 0.0625) (F3.1,25.1 = 2.97, = .785, p = .049). Latencies were not affected by the lesion, or the lever being chosen, and there were no interactions (maximum F was for response: F1,8 = 2.96, p = .124). To establish whether these effects were due to the large-reinforcer probability, or to progressive satiation or the passage of time, data from sessions 44–46 were also analysed, because in these sessions the large-reinforcer probability increased within the session. This time, there was a response × trial block interaction (F4,28 = 6.44, p = .001), with no other terms significant (Fs < 1, NS). Responding on the small/certain lever initially took 0.71 s in the first trial block and slowed to 0.95 s in the last trial block (F2.3,24.8 = 3.58, = .564, p = .038), but responding on the large/uncertain lever initially took 0.97 s (in the first trial block, when the large-reinforcer probability was 0.0625) and speeded up to 0.79 s (in the last trial block, when the large-reinforcer probability was 1) (F3.5,38.8 = 3.222, = .883, p = .027).
The lesion did not affect the latency to collect reward. Food collection latencies on rewarded trials were analysed across sessions 22–24, this time including both forced and choice trials to enable an analysis by response and probability. The model lesion2 × (probability5 × response2 × S) was used; this revealed main effects of response (F1,13 = 13.8, p = .003) and probability (F2.5,32.8 = 3.53, = .631, p = .031), but no other significant terms (maximum F was for lesion × response, F1,13 = 3.94, p = .069). Collection was faster following delivery of the large reinforcer than the small (4.1 versus 5.3 s, respectively), and got slightly slower across the session (4.4 s in the first trial block and 4.9 s in the last).
Amount of food obtained
AcbC-lesioned rats obtained less food as a result of their choices (Figure 6a,b). An analysis of the average number of pellets obtained on choice trials in sessions 13–15 using the model lesion2 × (probability5 × S) revealed a main effect of lesion (F1,16 = 8.69, p = .009), as well as an effect of probability (F1.9,30.1 = 97.5, = .470, p < .001), but no interaction (F < 1, NS). A similar analysis of the final baseline postoperative sessions 22–24 revealed a lesion × probability interaction (F2.1,33.8 = 3.29, = .529, p = .047) in addition to main effects of lesion (F1,16 = 14.2, p = .002) and probability (F2.1,33.8 = 122.2, = .529, p < .001). However, the only probability at which groups significantly differed was p = 1 (statistical p = .014); when the large reinforcer probability was 0.0625-0.5, the two groups did not differ in the amount of food obtained (pstatistical ≥ .129).
Effects of hunger and satiety on choice
Alternating between hunger and satiety had no substantial effects on choice (Figure 6b). Choice ratios for sessions 25–28 were analysed using the model lesion2 × (hunger2 × probability5 × S). As before, a main effect of probability (F2.0,32.6 = 38.4, = .510, p < .001) and a lesion × probability interaction (F2.0,32.6 = 4.29, = .510, p = .022) were present; in addition, there was a marginally significant lesion × hunger × probability interaction (F4,64 = 2.51, p = .05). However, an effect of hunger was not detectable in either group alone, either for shams (hunger: F1,11 = 2.45, NS; hunger × probability: F4,44 = 2.18, NS) or for AcbC-lesioned rats (hunger: F < 1, NS; hunger × probability: F4,20 = 1.79, NS). Similarly, the differences between groups persisted both in the hungry (lesion × probability: F2.6,41.7 = 3.66, = .652, p = .024) and the sated (lesion × probability: F2.5,39.3 = 4.24, = .615, p = .016) conditions.
Locomotor activity and body mass
AcbC-lesioned rats were hyperactive and slower to habituate to a novel environment (Figure 7). AcbC-lesioned rats also gained less mass postoperatively. At the time of surgery, the groups did not differ in mass (shams, 357 ± 4 g; AcbC, 362 ± 6 g; F < 1, NS), but at the end of the experiment AcbC-lesioned rats weighed less than shams (shams, 421 ± 7 g; AcbC, 358 ± 10 g; lesion × time, F1,16 = 80.1, p < .001; simple effect of lesion at final time point: F1,16 = 24.5, p < .001). Both effects are consistent with previous results: AcbC-lesioned rats are known to exhibit locomotor hyperactivity [22,24,38,41] and to weigh less than sham-operated controls [22,24,41,42]. They also eat the food used as the maintenance diet in the present study more slowly than sham-operated controls, and eat less of it in a given time, but do not differ in consumption of the sucrose pellets used as reinforcers in the present study [22,39]. It is not known whether there are metabolic differences in AcbC-lesioned rats above and beyond the tendency to eat somewhat less and to be hyperactive (though see [43]). However, differences in mass between AcbC-lesioned and sham-operated rats are also apparent when they have been fed ad libitum ever since the lesion was made, with AcbC-lesioned rats weighing ~88% as much as sham-operated controls in this situation [39], much as in the present study (85%). This suggests that the food deprivation regimen maintained the proportional relationship between actual and free-feeding mass similarly in sham-operated and AcbC-lesioned rats.
Discussion
These results suggest that the AcbC contributes to the selection of uncertain rewards. AcbC-lesioned rats exhibited risk-averse choice: they chose large, uncertain rewards less than sham-operated controls when offered a smaller, certain alternative, even though they showed a strong and unaltered preference for large rewards over small rewards, and for certain rewards over uncertain rewards. By the end of testing, the control group behaved as if indifferent between a single certain food pellet and four pellets delivered with p = 0.32 (close to the probability of 0.25 that would represent rational indifference), while the AcbC-lesioned group behaved as if indifferent between a single certain pellet and four pellets delivered with p = 0.70.
Though these results establish that the lesions used in this study caused this pattern of behaviour, the precise mechanism by which this occurs is unknown: for example, it is possible that the damage caused to structures adjacent to the AcbC, though limited, played a role in this pattern of choice, or that adaptations in other structures consequent upon the lesion were important in the behavioural effects (particularly given that risk aversion was not apparent immediately but emerged with further time and postoperative experience with the task).
Choice in normal subjects
The dominant model of uncertainty or probability discounting [1,44-46] suggests that subjects calculate a value for each reinforcer, according to its size and other parameters, and discount this by multiplying it by 1/(1+Hθ), where θ represents the odds against obtaining the reinforcer, θ = (1 - p)/p, and H represents an odds discounting parameter that is specific to the individual subject but stable over time for that subject. In this model, value is a hyperbolic function of the odds θ; such a hyperbolic function is supported by empirical research, at least in humans [44,45,47-50]. The present task is not well suited to evaluating such a quantitative model, since in discrete-trial schedules it is often the case that animals maximize, or allocate most of their choices to whichever option is the more favourable [51]. However, the behaviour of normal subjects here can be evaluated as to its optimality. In the present task, neither risk aversion nor risk taking is optimal if carried to extremes. Optimal behaviour, to maximize the expected amount of food, is to choose the small/certain lever when the large (4-pellet) reinforcer probability is less than 0.25, to choose the large/uncertain lever when the probability exceeds 0.25, and to be indifferent at p = 0.25 (i.e. to exhibit a step function in choice). Shams' choice of the large reinforcer behaviour was better described by a linear function of the large-reinforcer probability than by such a step function. Nevertheless, shams' behaviour was reasonably close to the optimal in the most obvious way to measure optimality, namely the amount of food obtained (Figure 6b).
Effects of AcbC lesions in terms of conditioning processes
AcbC-lesioned rats chose the large, uncertain reinforcer less often than shams did, but only when a smaller certain reinforcer was available as an alternative; that is, they exhibited risk-averse choice. A number of simple explanations of the present results may be ruled out. For example, it is unlikely that the pattern of choice exhibited by AcbC-lesioned rats can be explained in terms of perseveration, within a session, on the initially-optimal lever. It might be that animals that perseverated on the lever delivering the small, certain reinforcer, because that lever was initially optimal, would appear to exhibit risk-averse choice in sessions in which the large-reinforcer probability increased across the session (Figure 3e), but this could not explain the same pattern of choice in sessions in which the same lever was initially suboptimal, i.e. when the large-reinforcer probability decreased across the session (Figure 3c). Furthermore, although AcbC lesions are known to affect processes through which Pavlovian conditioned stimuli (CSs) affect behaviour, including Pavlovian-instrumental transfer (PIT), autoshaping, and conditioned reinforcement [38,52-57], there was no Pavlovian CS that was differentially associated with uncertain as opposed to certain reinforcement in this task, so these effects cannot explain the present results. It might be that the AcbC lesion impaired subjects' knowledge of the instrumental action-outcome contingency specifically for the uncertain outcome. There is some debate about the role of the AcbC in instrumental conditioning (see [43,58,59]) and goal-directed action, a subset of instrumental conditioning [58,60,61]. Manipulation of the AcbC can certainly affect instrumental learning [62-65]. However, the AcbC is not required for simple instrumental conditioning: rats with AcbC lesions acquire lever-press responses on fixed-ratio-1 schedules at supernormal levels [24], and rats with Acb or AcbC lesions are fully sensitive to changes in the action-outcome contingency [25,53,66]. However, when acquiring a sequence of random ratio schedules, AcbC-lesioned rats respond somewhat less than sham-operated controls [66], while lesions of the whole Acb made rats respond slightly, though not significantly, less on a similar sequence of random ratio schedules [53]. Random ratio schedules clearly involve probabilistic reinforcement, so these results are consistent with the possibility that the present impairment shown by AcbC-lesioned rats in choosing large, unlikely rewards is due to impaired instrumental conditioning when the outcome is uncertain – and, conversely, that the impairment in simple instrumental learning seen previously [66] was specifically a result of the reward uncertainty inherent in a random ratio schedule, given that AcbC-lesioned rats learn instrumental responses normally or supernormally with certain immediate reinforcement [24]. It is also possible that AcbC-lesioned rats represent the instrumental contingency normally with uncertain reward, but simply value the uncertain outcome less and respond less for it accordingly, as discussed next.
Effects of AcbC lesions in terms of probability discounting and reinforcer magnitude sensitivity
Since the present study required rats to choose between small, certain and large, uncertain rewards, an effect of the lesion to alter the perception of relative reward magnitude might affect choice, just as an alteration in the perception of reward probability might. For example, altering the absolute magnitudes of the reinforcers can affect choice involving probabilistic reinforcement [67,68], as would be predicted if reinforcer 'value' is not simply a linear function of physical magnitude [1]. Specifically, the present results (a tendency for AcbC-lesioned rats to choose the small, certain reinforcer more than shams) could be explained by 'risk aversion' (increased or steeper uncertainty/odds/probability discounting), or if the difference between 1 and 4 pellets was perceived to be smaller by AcbC-lesioned subjects than by shams (due to reduced discrimination between the two reinforcer magnitudes, or perhaps with a normal ability to tell the two apart but with an altered perception of relative value). For example, if a normal subject assigned values of 1 and 4 to the reinforcers, and a lesioned subject assigned values of 1 and 3 to the same reinforcers, then the lesioned subject would be less likely than the sham to choose the large reinforcer when it was made uncertain, even without any primary abnormality in the processing of probability. At first glance, this interpretation would appear to be supported by the observation that AcbC-lesioned rats chose the large reinforcer somewhat less often than shams when it was certain, as well as when it was uncertain. However, several lines of evidence suggest this explanation is not the correct one. When the large and the small reinforcers were both made consistently certain, there were no differences between AcbC-lesioned rats and controls (Figure 3d). Furthermore, other evidence indicates that AcbC lesions do not impair reinforcer magnitude discrimination or the perception of relative reinforcer value. Excitotoxic lesions of the whole Acb do not prevent rats from detecting changes in reward value (induced either by altering the concentration of a sucrose reward or by changing the deprivational state of the subject) [25]. Such lesions also do not impair rats' ability to respond faster when environmental cues predict the availability of larger rewards [26], and nor does inactivation of the Acb with local anaesthetic or blockade of AMPA glutamate receptors in the Acb [27,69]; the effects of intra-Acb NMDA receptor antagonists have varied [69-71]. AcbC-lesioned rats can still discriminate large from small rewards [23,28]. Similarly, DA depletion of the Acb does not affect the ability to discriminate large from small reinforcers [29-31], and systemic DA antagonists do not affect the perceived quantity of food as assessed in a psychophysical procedure [72]. Furthermore, a recent study found evidence that AcbC-lesioned rats may even show somewhat enhanced reinforcer magnitude discrimination (or an exaggerated perception of relative value) [24]. Given that reinforcer magnitude discrimination appears to be unimpaired, at worst, by AcbC lesions, the observation in the present study that AcbC-lesioned rats chose the large reinforcer somewhat less often than controls in the task in which large-reinforcer probabilities changed throughout the session is more likely to be explained by within-session generalization [23,73,74] – i.e. that avoidance of the large reinforcer during trial blocks when it was uncertain generalized to trial blocks when it was certain. Together, these findings suggest that the present results are best explained as an effect of AcbC lesions to increase the rate of uncertainty/odds/probability discounting – effectively, a tendency to behave as if an uncertain outcome were less likely than it really is.
Probability versus delay discounting
It is known that AcbC lesions affect choice and learning involving delayed reinforcement [22-24]. It has been suggested that delay (or temporal) discounting, the process by which delayed reinforcers lose value, and probability (or odds) discounting, the process by which uncertain reinforcers lose value, reflect the same underlying process [44,45,75-81]. For example, in the present task, choosing the uncertain large reinforcer five times but only obtaining it on the fifth response might be seen as equivalent to a very long delay, on average, between choice of the large reinforcer and its eventual delivery. Alternatively, delays may be seen as entailing the ecological risk of losing the reward during the delay. The failure of AcbC-lesioned rats to choose an uncertain reinforcer (risk aversion, as seen in the stable phase of the present results) and their failure to choose a delayed reinforcer may therefore be explained in the same way. However, there is evidence that time and probability discounting are different and dissociable processes [1,46,82]. Most simply, it is not surprising that currency inflation affects human decisions involving delayed but not probabilistic financial reward [83]. Moreover, the absolute magnitude of rewards can have different effects on delayed and probabilistic discounting [46,84,85]. A study looking at human choices in a gambling task found that individuals' propensity to choose rapidly (one, perhaps motoric, measure of delay aversion) and their propensity to bet large amounts of money on uncertain outcomes (a measure of risk taking) represented independent factors [86]. Some studies have found abnormal delay discounting, but not uncertainty discounting, in drug addicts [82,87-89], while gamblers have been observed to discount probabilistic rewards less steeply than controls (i.e. to take risks) without showing differences in delay discounting [8].
Implications for AcbC function and impulsivity
Impulsivity is multifaceted, reflecting – at the least – individual differences in distinct and dissociable processes involving information gathering, the selection of outcomes, and the inhibition of motor actions [90]. Furthermore, as discussed above, delay discounting and probability discounting may also reflect separate processes. Damage to the AcbC can produce impulsive choice in the sense of an impaired ability to choose delayed rewards [22], in addition to hyperactivity [22,24,38,41], though without impairments in attentional function [91] and without motoric impulsivity as assessed by the stop-signal task [92]. In the context of choice involving uncertain appetitive reinforcement, 'impulsivity' would equate to risk taking (less steep uncertainty discounting or greater willingness to choose unlikely rewards). AcbC lesions, however, produced a risk-averse or conservative pattern of choice in the present study. Clearly, then, AcbC-lesioned rats cannot be characterized as impulsive in all senses. A more appropriate unifying concept would seem to be that the AcbC promotes the selection, and perhaps the salience, of uncertain and delayed rewards – perhaps, in general, of rewards that are not certain, imminent, or present [58]. The AcbC promotes choice of [22] and learning with [24] delayed rewards. It appears to promote the selection of uncertain reinforcers (present results), and this is compatible with human imaging studies showing increased Acb blood flow during the selection of high-risk options [36,37]. The Acb is required for PIT, the process by which Pavlovian CSs signalling reward enhance instrumental responding for those rewards [52,53]. It is also required for autoshaping, or locomotor approach to appetitive Pavlovian CSs [38,54-57], and it influences conditioned reinforcement, the process of working for CSs previously paired with reinforcement [38,93-95]. Acb DA also contributes to subjects' motivation to work hard [96-100].
It is not known whether AcbC lesions would produce similar effects on choice involving uncertain aversive events. It would be expected that increased odds/uncertainty/probability discounting – effectively, a tendency to behave as if an uncertain outcome were less likely than it really is – would produce risk aversion for appetitive outcomes (reduced willingness to choose large, unlikely rewards) but risk proneness for aversive outcomes (increased willingness to choose large, uncertain punishments over small, certain punishments) [1]. In humans, at least, the delay and probability discounting processes appear similar for rewards and losses [46,101].
Relationship to structures and neuromodulator systems innervating the AcbC
The prefrontal cortex (PFC), which projects heavily to the AcbC [102], is also involved in decision-making under conditions of uncertainty. Humans with orbitofrontal cortex (OFC) or ventromedial PFC damage are impaired on the Iowa gambling task [103-105], in which subjects must learn to differentiate between low-reward, low-risk card decks that yield a net positive outcome and high-reward, high-risk decks that yield a net negative outcome, though the precise locus and nature of the deficit seen on this task is debated [106-108]. Choice between small, likely rewards and large, unlikely rewards increases cerebral blood flow in orbital and inferior PFC [109,110], and OFC damage also impairs performance of a task requiring human subjects to choose between two possible outcomes and to bet on their choice, with lesioned subjects deciding slowly and failing to choose the optimal, most likely outcome [111]. Excitotoxic lesions of the OFC make rats less likely than sham-operated controls to choose a large, uncertain reward over a small, certain reward [112]; OFC-lesioned rats had lower indifference odds (higher indifference probabilities; steeper uncertainty discounting) and exhibited risk-averse choice, just like the AcbC-lesioned subjects in the present study. There is direct evidence that OFC lesions do alter sensitivity to the relative magnitudes of the two rewards [113], as does OFC DA depletion [114], but the effects on uncertainty discounting are present in addition to those on reinforcer magnitude sensitivity [115].
The Acb is also innervated by the a number of neuromodulator systems, including the serotonin (5-hydroxytryptamine; 5-HT) system [116]. Although manipulations of 5-HT influence choice involving delayed reinforcement, there is less evidence that they influence choice involving uncertainty and risk. Correlational studies have indicated that low cerebrospinal fluid (CSF) levels of the 5-HT metabolite 5-hydroxyindoleacetic acid (5-HIAA) are associated with risk taking in monkeys [117] and impulsive aggression, violence, and suicide in humans [118-122]. Forebrain 5-HT depletion tends to steepen temporal (delay) discounting (reviewed briefly by [28]); however, it does not appear to influence choice involving probabilistic reinforcement. Dietary tryptophan depletion [123-125] decreases levels of 5-HT metabolites in CSF, an indirect indicator of brain 5-HT levels, but has not been shown to affect probability discounting in humans [126,127]; similarly, forebrain 5-HT depletion in rats does not affect choice between small, certain rewards and large, uncertain rewards [128]. The AcbC also receives a substantial DA innervation, and DA neurons respond to reward prediction errors [18-21]. Although systemic D2-type DA receptor antagonists can induce impulsive choice involving delayed reinforcement [129], this effect may not occur in the Acb [130], the response of DA neurons specifically to reward uncertainty is debated [32-34], and little is known of the role of DA in choice involving uncertain rewards. Systemic noradrenergic (NA) blockade has also been shown to affect decision-making under uncertainty in humans, by reducing the discrimination between magnitudes of different losses when the probability of losing was high [131], though NA reuptake inhibition has not been shown to affect the Iowa gambling task [132]. However, the Acb does not receive a substantial NA innervation [133].
Conclusion
We have shown that excitotoxic lesions of the AcbC induce risk-averse choice in rats. AcbC lesions did not prevent rats from discriminating a large reward from a small reward, or a certain reward from an uncertain reward. However, when offered the choice between a small/certain reward and a large/uncertain reward, AcbC-lesioned rats showed a reduced preference for the large/uncertain reward (compared to sham-operated controls) in their final pattern of postoperative choice. AcbC-lesioned rats exhibited a tendency to behave as if an uncertain outcome were less likely than was really the case. Together with previous studies, these results suggest that the AcbC contributes to reinforcement and choice particularly when the reinforcer is temporally distant or uncertain.
Methods
Subjects and housing conditions
The subjects were 24 male Lister hooded rats (Harlan-Olac UK Ltd) housed in a temperature-controlled room (minimum 22°C) under a 12:12 h reversed light-dark cycle (lights off 07:30 to 19:30). Subjects were approximately 15 weeks old on arrival at the laboratory and were given a minimum of a week to acclimatize, with free access to food, before experiments began. Preoperatively, subjects were housed in pairs; postoperatively, they were housed individually. Experiments took place between 09:00 and 21:00, with individual subjects being tested at a consistent time of day. Subjects had free access to water. During behavioural testing, subjects were fed ~15–16 g/day, an amount that maintains ~85–90% of free-feeding mass in normal male Lister hooded rats (the free-feeding mass being a steadily-increasing quantity at this age). Feeding occurred in the home cages at the end of the experimental day. As it was possible for subjects to earn substantial amounts of food in the behavioural tasks, the amount of food actually earned was taken into account when feeding with the maintenance diet in the home cages. All procedures were subject to UK Home Office approval (Project Licence 80/1767) under the Animals (Scientific Procedures) Act 1986.
Behavioural apparatus
Behavioural testing was conducted in one of two types of operant chamber of identical configuration (from Med Associates Inc., Georgia, Vermont, USA, or Paul Fray Ltd, Cambridge, UK). Each chamber was fitted with a 2.8 W overhead house light and two retractable levers on either side of an alcove fitted with an infrared photodiode to detect head entry and a 2.8 W lightbulb ('traylight'). Sucrose pellets (45 mg, Rodent Diet Formula P, Noyes, Lancaster, New Hampshire, USA) could be delivered into the alcove. The chambers were enclosed within sound-attenuating boxes fitted with fans to provide air circulation. The apparatus was controlled by software written by RNC in C++ [134] using the Whisker control system [135]. Equal numbers of subjects were trained in the two brands of operant chamber (12 in each type). Individual subjects were always tested in the same operant chamber.
Initial training
Rats were first trained to press the left lever for single pellets on a fixed-ratio-1 schedule, in 30-min sessions, until they had obtained a total of 100 pellets. This procedure was repeated for the right lever. They were then trained to nosepoke to initiate presentation of a lever in discrete trials. Each session began with the levers retracted and the operant chamber in darkness. Every 40 s, a trial began with illumination of the houselight and the traylight. The subject was required to make a nosepoke response within 10 s, or the current trial was aborted and the chamber returned to darkness. If the subject nosepoked within this time limit, the traylight was extinguished and a single lever presented. If the rat failed to respond on the lever within 10 s, the lever was retracted and the chamber darkened, but if it responded, the houselight was switched off, a single pellet was delivered immediately and the traylight was illuminated until the rat collected the pellet (or a 10-s collection time limit elapsed, whereupon the chamber was darkened). In every pair of trials, the left lever was presented once and the right lever once, though the order within the pair of trials was random. Rats were trained to a criterion of 60 successful trials in one hour (the maximum possible with a 40-s period being 90). They then proceeded to the full task.
Probabilistic choice task
The task was based on delayed reinforcement choice tasks that have been described before [73,74]. The session began in darkness with the levers retracted; this was designated the intertrial state. Trials began at 40-s intervals; the format of a single trial is shown in Figure 2. Each trial began with the illumination of the houselight and the traylight. The rat was required to make a nosepoke response, ensuring that it was centrally located at the start of the trial (latency to poke was designated the initiation latency). If the rat did not respond within 10 s of the start of the trial, the operant chamber was reset to the intertrial state until the next trial began and the trial was scored as an omission. If the rat was already nosepoking when the trial began, the next stage followed immediately. Upon a successful nosepoke, the traylight was extinguished and one or both levers were extended. One lever was designated the Large/Uncertain lever, the other the Small/Certain lever (counterbalanced left/right). The latency to choose a lever was recorded. (If the rat did not respond within 10 s of lever presentation, the chamber was reset to the intertrial state until the next trial and the trial was scored as an omission.) When a lever was chosen, both levers were retracted and the houselight was switched off. Choice of the Small lever caused the certain delivery of one pellet; choice of the Large lever caused the delivery of 4 pellets with a particular probability (see below). When reinforcement was delivered, the traylight was switched on. Multiple pellets were delivered 0.5 s apart. If the rat collected the pellets before the next trial began, then the traylight was switched off and the time from delivery of the first pellet until a nosepoke occurred was recorded as the collection latency. If the rat did not collect the food within 10 s of its delivery, the operant chamber entered the intertrial state, though collection latencies were still recorded up to the start of the next trial. The chamber was then in the intertrial state and remained so until the next trial. There was no mechanism to remove uneaten pellets, but failure to collect the reward was an extremely rare event. The large-reinforcer probability was varied systematically across the session as follows. A session consisted of 5 blocks, each comprising 16 trials in which only one lever was presented (8 trials for each lever, randomized in pairs) followed by 10 free-choice trials. The probability that the large reinforcer was delivered, given that the Large lever had been chosen (preinforcer), varied across blocks: it was initially 1, 0.5, 0.25, 0.125, and 0.0625, respectively, for each block. As trials began every 40 s and there were 130 trials per session, the total session length was ~87 minutes; subjects received one session per day. Choice ratios (percentage choice of the large reinforcer, for each trial block) were calculated using only choice trials on which the subject responded.
Excitotoxic lesions of the AcbC
Subjects were anaesthetized with Avertin (2% w/v 2,2,2-tribromoethanol, 1% w/v 2-methylbutan-2-ol, and 8% v/v ethanol in phosphate-buffered saline, sterilized by filtration, 10 ml/kg intraperitoneally) and placed in a Kopf or Stoelting stereotaxic frame (David Kopf Instruments, Tujunga, California, USA; Stoelting Co., Wood Dale, Illinois, USA) fitted with atraumatic ear bars. The skull was exposed and a dental drill was used to remove the bone directly above the injection sites. The dura mater was broken with the tip of a hypodermic needle, avoiding damage to underlying venous sinuses. Excitotoxic lesions of the AcbC were made by injecting 0.5 μl of 0.09 M quinolinic acid (Sigma, UK) per side through a glass micropipette at coordinates 1.2 mm anterior to bregma, ± 1.8 mm from the midline, and 7.1 mm below the skull surface at bregma; the incisor bar was 3.3 mm below the interaural line [136]. The toxin had been dissolved in 0.1 M phosphate buffer (composition 0.07 M Na2HPO4, 0.028 M NaH2PO4 in double-distilled water, sterilized by filtration) and adjusted with NaOH to a final pH of 7.2–7.4. Toxin was injected over 3 min and the micropipette was left in place for 2 min following injections. Sham lesions were made in the same manner except that vehicle was infused. At the end of the operation, animals were given 15 ml/kg of sterile 5% w/v glucose, 0.9% w/v sodium chloride intraperitoneally. They were given a week to recover, with free access to food, and were handled regularly. Any instances of postoperative constipation were treated with liquid paraffin orally and rectally. At the end of this period, food restriction commenced or was resumed.
Postoperative testing
Subjects were trained preoperatively and tested postoperatively according to the schedule shown in Table 1. In the basic task, used for preoperative training, the probability of large reinforcer delivery declined across trial blocks from 1 to 0.0625 (in the order 1, 0.5, 0.25, 0.125, 0.0625). After subjects had been tested postoperatively for 12 sessions on this schedule, satiety tests were given, to establish the effect of varying primary motivational state on preference for probabilistic reinforcement. Subjects were tested for 4 sessions while alternating between hungry and sated states on consecutive days in counterbalanced fashion (half the subjects experienced hungry and sated days in the order HSHS, and half in the order SHSH). Following a 'hungry' session, animals were placed on free food (maintenance diet) until the start of the next day's 'sated' session, at which time the food was again removed for the 'hungry' session to follow. The comparison was therefore between food deprivation for ~22 h and satiety. Next, subjects were returned to the hungry state and tested for 6 sessions on a schedule in which both the large and small reinforcer were delivered with certainty. Next, the element of uncertainty was reintroduced for another 12 sessions, but this time the probability of large reinforcer delivery (given that the Large lever had been chosen) increased across blocks from 0.0625 to 1 (in the order 0.0625, 0.125, 0.25, 0.5, 1). Finally, subjects were tested for 6 sessions with the large reinforcer always being very unlikely (p = 0.0625), with the small reinforcer remaining certain.
Locomotor activity in a novel environment
Locomotor activity was measured in wire mesh cages, 25 (W) × 40 (D) × 18 (H) cm, each equipped with a water bottle and two horizontal photocell beams situated 1 cm from the floor that enabled movements along the long axis of the cage to be registered. Subjects were placed in these cages, which were initially unfamiliar to them, and their activity was recorded for 2 h. All animals were tested in the food-deprived state. Locomotor hyperactivity and reduced body mass gain have previously been part of the phenotype of AcbC-lesioned rats, though without alterations in the consumption of the reinforcer used in the present experiments [22,24,38,39,41].
Histology
Rats were deeply anaesthetized with pentobarbitone sodium (200 mg/ml, minimum of 1.5 ml i.p.) and perfused transcardially with 0.01 M phosphate-buffered saline (PBS) followed by 4% paraformaldehyde in PBS. Their brains were removed and postfixed in paraformaldehyde before being dehydrated in 20% sucrose for cryoprotection. The brains were sectioned coronally at 60 μm thickness on a freezing microtome and every third section mounted on chromium potassium sulphate/gelatin-coated glass microscope slides and allowed to dry. Sections were passed through a series of ethanol solutions of descending concentration (3 minutes in each of 100%, 95%, and 70% v/v ethanol in water) and stained for ~5 min with cresyl violet. The stain comprises 0.05% w/v aqueous cresyl violet (Raymond A. Lamb Ltd, Eastbourne, UK), 2 mM acetic acid, and 5 mM formic acid in water. Following staining, sections were rinsed in water and 70% ethanol before being differentiated in 95% ethanol. Finally, they were dehydrated and delipidated in 100% ethanol and Histoclear (National Diagnostics, UK) before being cover-slipped using DePeX mounting medium (BDH, UK) and allowed to dry. The sections were used to verify lesion placement and assess the extent of lesion-induced neuronal loss. Lesions were detectable as the absence of visible neurons (cell bodies of the order of 100 μm in diameter with a characteristic shape and appearance), often associated with a degree of tissue collapse (sometimes with consequent ventricular expansion when the lesion was adjacent to a ventricle) and gliosis (visible as the presence of smaller, densely-staining cells).
Data analysis
Data collected by the chamber control programs were imported into a relational database (Microsoft Access 97) for case selection and analysed with SPSS 11. Figures were created with SigmaPlot 2001/v7 and Adobe Illustrator 8. All graphs show group means and error bars are ±1 standard error of the mean (SEM) unless otherwise stated. Count data (e.g. locomotor activity counts), for which variance increases with the mean, were subjected to a square-root transformation prior to any analysis [137]. Homogeneity of variance was verified using Levene's test [138]. General linear models are described as dependent variable = A2 × Bcov × (C5 × Dcov × S) where A is a between-subjects factor with two levels, B is a between-subjects covariate, C is a within-subjects factor with five levels, and D is a within-subjects covariate; S denotes subjects in designs involving within-subjects factors [139]. For repeated measures analyses, Mauchly's test of sphericity of the covariance matrix was applied [140] and the degrees of freedom corrected to more conservative values by multiplying them by the Huynh-Feldt epsilon for any terms involving factors in which the sphericity assumption was violated [141]. Where multiple comparisons were conducted post hoc following a significant overall ANOVA effect for a factor with more than three levels, p values were corrected using the Šidák correction [142], in which pcorrected = 1 - (1 - puncorrected)n for n comparisons.
List of abbreviations used
5-HIAA, 5-hydroxyindoleacetic acid
5-HT, 5-hydroxytryptamine (serotonin)
Acb, nucleus accumbens
AcbC, nucleus accumbens core
AcbSh, nucleus accumbens shell
ANOVA, analysis of variance
CS, conditioned stimulus
CSF, cerebrospinal fluid
DA, dopamine
, Huynh-Feldt epsilon
h, hour
i.p., intraperitoneal
min, minute
NA, noradrenaline
OFC, orbitofrontal cortex
p, probability
preinforcer, probability of delivery of the large reinforcer after it had been chosen
pstatistical, statistical p value (probability of obtaining the observed data, or results more extreme, were the null hypothesis to be true)
PBS, phosphate-buffered saline
PIT, Pavlovian-instrumental transfer
PFC, prefrontal cortex
r2, proportion of variance explained
SED, standard error of the difference between means
SEM, standard error of the mean
SS, sum of squares (sum of squared deviations)
v/v, volume per unit volume
w/v, weight per unit volume
[x, y), a range that includes x but not y
Authors' contributions
RNC conceived and designed the studies, supervised NJH, wrote the software, performed the surgery, and drafted the manuscript. NJH participated in the design of the studies, and tested the animals. The work contributed to NJH's B.A. degree. Both authors analysed the results, and read and approved the final manuscript.
Acknowledgements
Supported by a Wellcome Trust programme grant (to Trevor W. Robbins, Barry J. Everitt, Angela C. Roberts, and Barbara J. Sahakian); conducted within the UK Medical Research Council (MRC) Behavioural and Clinical Neuroscience Centre, Cambridge. We thank three anonymous referees for their helpful comments. Competing interests: none declared.
Figures and Tables
Figure 1 Task schematic: choice between small, certain and large, uncertain reward. Probabilistic choice task, based on similar tasks involving choice between delayed reinforcers [73, 74]. Hungry rats regularly chose between two levers. Responding on one lever led to the certain delivery of a small food reward (1 pellet); responding on the other led to a much larger food reward (4 pellets), but this reward was uncertain, and was delivered with a probability (p) ranging from 1 to 0.0625. The figure shows the format of a single trial. Trials began at regular intervals (every 40 s). Sessions consisted of 5 blocks. In each block, 16 single-lever trials were given (8 trials for each lever, randomized in pairs), to ensure the animals sampled the options available at that time; these were followed by 10 choice trials. The probability of delivery of the large reinforcer was varied systematically across the session: probabilities for each block were initially 1, 0.5, 0.25, 0.125, and 0.0625, respectively (see Table 1).
Figure 2 Schematic of lesions of the nucleus accumbens core. Black shading indicates the extent of neuronal loss common to all subjects; grey indicates the area lesioned in at least one subject. Coronal sections are (from top to bottom) +2.7, +2.2, +1.7, +1.2, +0.7, and +0.2 mm relative (anterior) to bregma. Diagrams are modified from ref. [136].
Figure 3 Choice with probabilistic reinforcement. (a) Preoperative patterns of choice. There were no differences between the groups preoperatively. (b) The first three postoperative sessions. Transiently, AcbC-lesioned rats exhibited relative indifference between the two alternatives; their preference did not differ significantly from 50% at any large-reinforcer probability. As a result, AcbC-lesioned rats preferred the large, unlikely reinforcer more than shams did when its probability was 0.0625 and 0.125 (## p < .01, lesion × probability interaction; * p < .05, ** p < .01, comparison to shams at individual probabilities). However, both groups were influenced by the large-reinforcer probability (p ≤ .004). (c) The last three postoperative sessions on the same basic task. By this point, AcbC-lesioned rats preferred the large reinforcer less when its probability was 0.5 or 1 (## p < .01, interaction; * p < .05, simple effects). Again, both groups were influenced by the large-reinforcer probability (p < .001). (d) When the 4-pellet reinforcer and the 1-pellet reinforcer were both certain, all groups preferred the 4-pellet reinforcer, and when the 4-pellet reinforcer was always very unlikely (delivered with a probability of 0.0625) and the 1-pellet reinforcer was certain, all groups preferred the 1-pellet reinforcer, with no differences between AcbC-lesioned and sham-operated rats. This indicates that both groups discriminated the reinforcers themselves and discriminated their probability of delivery. (e) Choice following further training in which the large-reinforcer probability increased, rather than decreased, across each session. The pattern of choice is similar to c, in that AcbC-lesioned rats were risk-averse compared to shams, i.e. less likely to choose the large, unlikely reinforcer (### p < .001, interaction; * p < .05 and *** p < .001, simple effects). The similarity to c, despite the reversed task order, also indicates that subjects' choice reflected the probabilities in force rather than the order within a session.
Figure 4 Choice, by experienced probability. Choice, replotted by experienced (as opposed to programmed) large-reinforcer probabilities. Panels a-d correspond to panels a-c/e of the previous figure. The statistical patterns of choice remained the same (### p < .001, lesion × experienced delay interaction; ** p < .01, main effect of lesion).
Figure 5 Indifference probabilities. Subjects' behaviour was analysed using a linear regression technique (see text for method of calculation) to estimate the large-reinforcer probability at which they were indifferent between a 4-pellet uncertain large reinforcer and a 1-pellet certain small reinforcer. Rational choice, and optimal choice in this task, would be an indifference probability of 0.25 (that is, it is rational to be indifferent between a certain 1-pellet reinforcer and a 4-pellet reinforcer delivered with a probability of 0.25), shown by the dotted line. Lower indifference probabilities imply risk-prone behaviour; higher indifference probabilities imply risk-averse behaviour. Preoperative and successive postoperative indifference probabilities are shown for AcbC-lesioned and sham-operated control rats (** p < .01, difference from controls).
Figure 6 Amount of food obtained, and effects of satiety on choice. (a) Number of pellets obtained in each trial block; average of the first three postoperative sessions, 13–15 (** p < .01, main effect of lesion). The grey area indicates the expected range of options available to a rat making no omissions: consistent responding on the lever delivering the small, certain reward of a single pellet yields 10 pellets per trial block (horizontal border of the grey area); consistent responding on the lever delivering the large, uncertain reward yields an expected number of pellets that varies with the probability in force (as shown by the diagonal border of the grey area). Optimal behaviour, to maximize the expected amount of food, is to choose the small/certain lever when the large (4-pellet) reinforcer probability is less than 0.25 and to choose the large/uncertain lever when this probability exceeds 0.25. (b) As for (a), but showing data from the final baseline postoperative sessions, 22–24 (# p < .05, lesion × probability interaction; * p < .05, simple effect of lesion). (c) Effects on choice of alternating subjects between states of hunger and satiety. The error bar is twice the SED for the three-way (lesion × hunger × probability) interaction.
Figure 7 Locomotor activity in a novel environment. AcbC-lesioned rats were hyperactive compared to sham-operated controls, being slower to habituate to a novel environment. Analysis using the model lesion2 × (bin12 × S) revealed a lesion × bin interaction (F8.3,133.4 = 2.20, = .758, # p = .029), reflecting a difference in habituation between the groups, and a main effect of bin (F8.34,133.4 = 9.02, p < .001), reflecting habituation, though there was no main effect of lesion (F1,16 = 2.24, p = .154).
Table 1 Testing schedule for probabilistic choice task. Subjects were trained and tested according to the schedule shown here. Initial pre- and postoperative testing was conducted with the probability of large reinforcer delivery declining across trial blocks from 1 to 0.0625 (the steps were p = 1, 0.5, 0.25, 0.125, and 0.0625). Subsequently, subjects were tested alternating between the hungry and sated state (as described in the Methods), before the reinforcement probabilities were manipulated further, as indicated.
Sessions Description Probability (small reward of 1 pellet) Probability (large reward of 4 pellets)
1–12 Preoperative training (12 sessions) p = 1 p = 1 to 0.0625, decreasing
- Surgery and recovery - -
13–24 Postoperative baseline testing (12 sessions) p = 1 p = 1 to 0.0625, decreasing
25–28 Satiety tests (4 sessions) p = 1 p = 1 to 0.0625, decreasing
29–34 Equal probabilities, certain (6 sessions) p = 1 p = 1
35–46 Increasing probabilities (12 sessions) p = 1 p = 0.0625 to 1, increasing
47–52 Large reinforcer always very unlikely (6 sessions) p = 1 p = 0.0625
==== Refs
Ho MY Mobini S Chiang TJ Bradshaw CM Szabadi E Theory and method in the quantitative analysis of "impulsive choice" behaviour: implications for psychopharmacology Psychopharmacology 1999 146 362 372 10550487
Evenden JL Impulsivity: a discussion of clinical and experimental findings Journal of Psychopharmacology 1999 13 180 192 10475725
Eysenck SGB McCown WG, Johnson JL and Shure MB The I7: development of a measure of impulsivity and its relationship to the superfactors of personality The impulsive client: theory, research and treatment 1993 Washington DC, American Psychological Association
Daruna JH Barnes PA McCown WG, Johnson JL and Shure MB A neurodevelopmental view of impulsivity The impulsive client: theory, research and treatment 1993 Washington, DC, American Psychological Association
APA Diagnostic and Statistical Manual of Mental Disorders, fourth edition, text revision (DSM-IV-TR) 2000 Washington DC, American Psychiatric Association
Roy A DeJong J Linnoila M Extraversion in pathological gamblers: correlates with indices of noradrenergic function Archives of General Psychiatry 1989 46 679 681 2751401
Coccaro EF Siever LJ Bloom FE and Kupfer DJ The neuropsychopharmacology of personality disorders Psychopharmacology: The Fouth Generation of Progress 1995 1567-1579 New York, American College of Neuropsychopharmacology / Raven Press [ http://www.acnp.org/content-32.html ]
Holt DD Green L Myerson J Is discounting impulsive?. Evidence from temporal and probability discounting in gambling and non-gambling college students Behav Processes 2003 64 355 367 14580704 10.1016/S0376-6357(03)00141-4
OUP New Shorter Oxford English Dictionary 1997 Oxford, UK, Oxford University Press
Knutson B Adams CM Fong GW Hommer D Anticipation of increasing monetary reward selectively recruits nucleus accumbens J Neurosci 2001 21 RC159 11459880
Breiter HC Aharon I Kahneman D Dale A Shizgal P Functional imaging of neural responses to expectancy and experience of monetary gains and losses Neuron 2001 30 619 639 11395019 10.1016/S0896-6273(01)00303-8
Schultz W Tremblay L Hollerman JR Reward processing in primate orbitofrontal cortex and basal ganglia Cereb Cortex 2000 10 272 284 10731222 10.1093/cercor/10.3.272
Miyazaki K Mogi E Araki N Matsumoto G Reward-quality dependent anticipation in rat nucleus accumbens Neuroreport 1998 9 3943 3948 9875733
Schultz W Apicella P Scarnati E Ljungberg T Neuronal activity in monkey ventral striatum related to the expectation of reward Journal of Neuroscience 1992 12 4595 4610 1464759
Bjork JM Knutson B Fong GW Caggiano DM Bennett SM Hommer DW Incentive-elicited brain activation in adolescents: similarities and differences from young adults J Neurosci 2004 24 1793 1802 14985419 10.1523/JNEUROSCI.4862-03.2004
Martin PD Ono T Effects of reward anticipation, reward presentation, and spatial parameters on the firing of single neurons recorded in the subiculum and nucleus accumbens of freely moving rats Behav Brain Res 2000 116 23 38 11090883 10.1016/S0166-4328(00)00249-7
Cromwell HC Schultz W Effects of expectations for different reward magnitudes on neuronal activity in primate striatum J Neurophysiol 2003 89 2823 2838 12611937
Schultz W Dickinson A Neuronal coding of prediction errors Annual Review of Neuroscience 2000 23 473 500 10845072 10.1146/annurev.neuro.23.1.473
Schultz W Dayan P Montague PR A neural substrate of prediction and reward Science 1997 275 1593 1599 9054347 10.1126/science.275.5306.1593
Schultz W Predictive reward signal of dopamine neurons J Neurophysiol 1998 80 1 27 9658025
Schultz W Tremblay L Hollerman JR Reward prediction in primate basal ganglia and frontal cortex Neuropharmacology 1998 37 421 429 9704983 10.1016/S0028-3908(98)00071-9
Cardinal RN Pennicott DR Sugathapala CL Robbins TW Everitt BJ Impulsive choice induced in rats by lesions of the nucleus accumbens core Science 2001 292 2499 2501 11375482 10.1126/science.1060818
Cardinal RN Robbins TW Everitt BJ Heather N and Vuchinich RE Choosing delayed rewards: perspectives from learning theory, neurochemistry, and neuroanatomy Choice, Behavioral Economics and Addiction 2003 Oxford, Elsevier 183 213, 217-218
Cardinal RN Cheung THC Nucleus accumbens core lesions retard instrumental learning and performance with delayed reinforcement in the rat BMC Neuroscience 2005 6 9 15691387 10.1186/1471-2202-6-9
Balleine B Killcross S Effects of ibotenic acid lesions of the nucleus accumbens on instrumental action Behavioural Brain Research 1994 65 181 193 7718151 10.1016/0166-4328(94)90104-X
Brown VJ Bowman EM Discriminative cues indicating reward magnitude continue to determine reaction time of rats following lesions of the nucleus accumbens European Journal of Neuroscience 1995 7 2479 2485 8845953
Giertler C Bohn I Hauber W Transient inactivation of the rat nucleus accumbens does not impair guidance of instrumental behaviour by stimuli predicting reward magnitude Behav Pharmacol 2004 15 55 63 15075627 10.1097/00008877-200402000-00007
Cardinal RN Winstanley CA Robbins TW Everitt BJ Limbic corticostriatal systems and delayed reinforcement Annals of the New York Academy of Sciences 2004 1021 33 50 15251872 10.1196/annals.1308.004
Salamone JD Cousins MS Bucher S Anhedonia or anergia? Effects of haloperidol and nucleus accumbens dopamine depletion on instrumental response selection in a T-maze cost/benefit procedure Behavioural Brain Research 1994 65 221 229 7718155 10.1016/0166-4328(94)90108-2
Salamone JD Wisniecki A Carlson BB Correa M Nucleus accumbens dopamine depletions make animals highly sensitive to high fixed ratio requirements but do not impair primary food reinforcement Neuroscience 2001 105 863 870 11530224 10.1016/S0306-4522(01)00249-4
Cousins MS Atherton A Turner L Salamone JD Nucleus accumbens dopamine depletions alter relative response allocation in a T-maze cost/benefit task Behavioural Brain Research 1996 74 189 197 8851929 10.1016/0166-4328(95)00151-4
Fiorillo CD Tobler PN Schultz W Discrete coding of reward probability and uncertainty by dopamine neurons Science 2003 299 1898 1902 12649484 10.1126/science.1077349
Tobler PN Fiorillo CD Schultz W Adaptive coding of reward value by dopamine neurons Science 2005 307 1642 1645 15761155 10.1126/science.1105370
Niv Y Duff MO Dayan P Dopamine, uncertainty and TD learning Behavioral and Brain Functions 2005 1 6 15953384 10.1186/1744-9081-1-6
Aron AR Shohamy D Clark J Myers C Gluck MA Poldrack RA Human midbrain sensitivity to cognitive feedback and uncertainty during classification learning J Neurophysiol 2004 92 1144 1152 15014103 10.1152/jn.01209.2003
Ernst M Nelson EE McClure EB Monk CS Munson S Eshel N Zarahn E Leibenluft E Zametkin A Towbin K Blair J Charney D Pine DS Choice selection and reward anticipation: an fMRI study Neuropsychologia 2004 42 1585 1597 15327927 10.1016/j.neuropsychologia.2004.05.011
Matthews SC Simmons AN Lane SD Paulus MP Selective activation of the nucleus accumbens during risk-taking decision making Neuroreport 2004 15 2123 2127 15486494 10.1097/00001756-200409150-00025
Parkinson JA Olmstead MC Burns LH Robbins TW Everitt BJ Dissociation in effects of lesions of the nucleus accumbens core and shell on appetitive Pavlovian approach behavior and the potentiation of conditioned reinforcement and locomotor activity by d-amphetamine Journal of Neuroscience 1999 19 2401 2411 10066290
Cardinal RN Neuropsychology of reinforcement processes in the rat [unpublished PhD thesis; www.pobox.com/~rudolf/publications/2001/PhD ] 2001 Cambridge, UK, University of Cambridge
Cameron AC Windmeijer FAG An R-squared measure of goodness of fit for some common nonlinear regression models Journal of Econometrics 1997 77 329 342 10.1016/S0304-4076(96)01818-0
Maldonado-Irizarry CS Kelley AE Excitotoxic lesions of the core and shell subregions of the nucleus accumbens differentially disrupt body-weight regulation and motor activity in the rat Brain Research Bulletin 1995 38 551 559 8590077 10.1016/0361-9230(95)02030-2
Parkinson JA Limbic corticostriatal circuitry underlying Pavlovian associative learning [unpublished PhD thesis] 1998 Cambridge, UK, University of Cambridge
Kelley AE Ventral striatal control of appetitive motivation: role in ingestive behavior and reward-related learning Neurosci Biobehav Rev 2004 27 765 776 15019426 10.1016/j.neubiorev.2003.11.015
Rachlin H Raineri A Cross D Subjective probability and delay Journal of the Experimental Analysis of Behavior 1991 55 233 244 2037827
Rachlin H Logue AW Gibbon J Frankel M Cognition and behavior in studies of choice Psychological Review 1986 93 33 45 10.1037//0033-295X.93.1.33
Green L Myerson J A discounting framework for choice with delayed and probabilistic rewards Psychol Bull 2004 130 769 792 15367080 10.1037/0033-2909.130.5.769
Rachlin H Siegel E Temporal patterning in probabilistic choice Organizational Behavior and Human Decision Processes 1994 59 161 176 10.1006/obhd.1994.1054
Rachlin H Brown J Cross D Discounting in judgments of delay and probability Journal of Behavioral Decision Making 2000 13 145 149 10.1002/(SICI)1099-0771(200004/06)13:2<145::AID-BDM320>3.0.CO;2-4
Richards JB Zhang L Mitchell SH de Wit H Delay or probability discounting in a model of impulsive behavior: effect of alcohol J Exp Anal Behav 1999 71 121 143 10220927
Kacelnik A Normative and descriptive models of decision making: time discounting and risk sensitivity Ciba Found Symp 1997 208 51 67; discussion 67-70 9386907
Mackintosh NJ The Psychology of Animal Learning 1974 London, Academic Press
Hall J Parkinson JA Connor TM Dickinson A Everitt BJ Involvement of the central nucleus of the amygdala and nucleus accumbens core in mediating Pavlovian influences on instrumental behaviour European Journal of Neuroscience 2001 13 1984 1992 11403692 10.1046/j.0953-816x.2001.01577.x
de Borchgrave R Rawlins JN Dickinson A Balleine BW Effects of cytotoxic nucleus accumbens lesions on instrumental conditioning in rats Exp Brain Res 2002 144 50 68 11976759 10.1007/s00221-002-1031-y
Parkinson JA Robbins TW Everitt BJ Selective excitotoxic lesions of the nucleus accumbens core and shell differentially affect aversive Pavlovian conditioning to discrete and contextual cues Psychobiology 1999 27 256 266
Parkinson JA Willoughby PJ Robbins TW Everitt BJ Disconnection of the anterior cingulate cortex and nucleus accumbens core impairs Pavlovian approach behavior: Further evidence for limbic cortical-ventral striatopallidal systems Behavioral Neuroscience 2000 114 42 63 10718261 10.1037//0735-7044.114.1.42
Parkinson JA Dalley JW Cardinal RN Bamford A Fehnert B Lachenal G Rudarakanchana N Halkerston KM Robbins TW Everitt BJ Nucleus accumbens dopamine depletion impairs both acquisition and performance of appetitive Pavlovian approach behaviour: implications for mesoaccumbens dopamine function Behavioural Brain Research 2002 137 149 163 12445721 10.1016/S0166-4328(02)00291-7
Cardinal RN Parkinson JA Lachenal G Halkerston KM Rudarakanchana N Hall J Morrison CH Howes SR Robbins TW Everitt BJ Effects of lesions of the nucleus accumbens core, anterior cingulate cortex, and central nucleus of the amygdala on autoshaping performance in rats Behavioral Neuroscience 2002 116 553 567 12148923 10.1037//0735-7044.116.4.553
Cardinal RN Parkinson JA Hall J Everitt BJ Emotion and motivation: the role of the amygdala, ventral striatum, and prefrontal cortex Neurosci Biobehav Rev 2002 26 321 352 12034134 10.1016/S0149-7634(02)00007-6
Cardinal RN Everitt BJ Neural and psychological mechanisms underlying appetitive learning: links to drug addiction Curr Opin Neurobiol 2004 14 156 162 15082319 10.1016/j.conb.2004.03.004
Dickinson A Balleine B Motivational control of goal-directed action Animal Learning & Behavior 1994 22 1 18
Dickinson A Mackintosh NJ Instrumental conditioning Animal Learning and Cognition 1994 San Diego, Academic Press 45 79
Kelley AE Smith-Roe SL Holahan MR Response-reinforcement learning is dependent on N-methyl-D-aspartate receptor activation in the nucleus accumbens core Proc Natl Acad Sci U S A 1997 94 12174 12179 9342382 10.1073/pnas.94.22.12174
Baldwin AE Sadeghian K Holahan MR Kelley AE Appetitive instrumental learning is impaired by inhibition of cAMP-dependent protein kinase within the nucleus accumbens Neurobiol Learn Mem 2002 77 44 62 11749085 10.1006/nlme.2000.4002
Smith-Roe SL Kelley AE Coincident activation of NMDA and dopamine D1 receptors within the nucleus accumbens core is required for appetitive instrumental learning J Neurosci 2000 20 7737 7742 11027236
Hernandez PJ Sadeghian K Kelley AE Early consolidation of instrumental learning requires protein synthesis in the nucleus accumbens Nat Neurosci 2002 5 1327 1331 12426572 10.1038/nn973
Corbit LH Muir JL Balleine BW The role of the nucleus accumbens in instrumental conditioning: evidence of a functional dissociation between accumbens core and shell Journal of Neuroscience 2001 21 3251 3260 11312310
Mazur JE Choice between small certain and large uncertain reinforcers Animal Learning & Behavior 1988 16 199 205
Kirby KN Marakovic NN Delay-discounting probabilistic rewards: Rates decrease as amounts increase Psychonomic Bulletin & Review 1996
Giertler C Bohn I Hauber W Involvement of NMDA and AMPA/KA receptors in the nucleus accumbens core in instrumental learning guided by reward-predictive cues Eur J Neurosci 2005 21 1689 1702 15845096
Hauber W Bohn I Giertler C NMDA, but not dopamine D(2), receptors in the rat nucleus accumbens are involved in guidance of instrumental behavior by stimuli predicting reward magnitude Journal of Neuroscience 2000 20 6282 6288 10934279
Giertler C Bohn I Hauber W The rat nucleus accumbens is involved in guiding of instrumental responses by stimuli predicting reward magnitude Eur J Neurosci 2003 18 1993 1996 14622231 10.1046/j.1460-9568.2003.02904.x
Martin-Iverson MT Wilkie D Fibiger HC Effects of haloperidol and d-amphetamine on perceived quantity of food and tones Psychopharmacology 1987 93 374 381 3124167 10.1007/BF00187260
Cardinal RN Robbins TW Everitt BJ The effects of d-amphetamine, chlordiazepoxide, alpha-flupenthixol and behavioural manipulations on choice of signalled and unsignalled delayed reinforcement in rats Psychopharmacology 2000 152 362 375 11140328 10.1007/s002130000536
Evenden JL Ryan CN The pharmacology of impulsive behaviour in rats: the effects of drugs on response choice with varying delays of reinforcement Psychopharmacology 1996 128 161 170 8956377 10.1007/s002130050121
Sozou PD On hyperbolic discounting and uncertain hazard rates [ DOI: 10.1098/rspb.1998.0534 ] Proceedings of the Royal Society of London Series B: Biological Sciences 1998 265 2015 2020 10.1098/rspb.1998.0534
Green L Myerson J Exponential versus hyperbolic discounting of delayed outcomes: Risk and waiting time American Zoologist 1996 36 496 505
Stevenson MK A discounting model for decisions with delayed positive and negative outcomes Journal of Experimental Psychology: General 1986 115 131 154 10.1037//0096-3445.115.2.131
Rachlin H Castrogiovanni A Cross D Probability and delay in commitment Journal of the Experimental Analysis of Behavior 1987 48 347 353
Mazur JE Theories of probabilistic reinforcement Journal of the Experimental Analysis of Behavior 1989 51 87 99 2921590
Mazur JE Conditioned reinforcement and choice with delayed and uncertain primary reinforcers Journal of the Experimental Analysis of Behavior 1995 63 139 150 7714448
Mazur JE Choice, delay, probability, and conditioned reinforcement Animal Learning & Behavior 1997 25 131 147
Mitchell SH Heather N and Vuchinich RE Discounting the value of commodities according to different types of cost Choice, Behavioral Economics and Addiction 2003 Oxford, Elsevier 339 357
Ostaszewski P Green L Myerson J Effects of inflation on the subjective value of delayed and probabilistic rewards Psychonomic Bulletin & Review 1998 5 324 333
Green L Myerson J Ostaszewski P Amount of reward has opposite effects on the discounting of delayed and probabilistic outcomes J Exp Psychol Learn Mem Cogn 1999 25 418 427 10093208 10.1037//0278-7393.25.2.418
Myerson J Green L Hanson JS Hold DD Estle SJ Discounting delayed and probabilistic rewards: Processes and traits Journal of Economic Psychology 2003 24 619 635 10.1016/S0167-4870(03)00005-9
Deakin J Aitken M Robbins T Sahakian BJ Risk taking during decision-making in normal volunteers changes with age J Int Neuropsychol Soc 2004 10 590 598 15327737 10.1017/S1355617704104104
Reynolds B Richards JB Horn K Karraker K Delay discounting and probability discounting as related to cigarette smoking status in adults Behav Processes 2004 65 35 42 14744545 10.1016/S0376-6357(03)00109-8
Vuchinich RE Calamas ML Does the repeated gambles procedure measure impulsivity in social drinkers? Exp Clin Psychopharmacol 1997 5 157 162 9234053 10.1037//1064-1297.5.2.157
Mitchell SH Measures of impulsivity in cigarette smokers and non-smokers Psychopharmacology 1999 146 455 464 10550496
Evenden JL Varieties of impulsivity Psychopharmacology 1999 146 348 361 10550486
Christakou A Robbins TW Everitt BJ Prefrontal cortical-ventral striatal interactions involved in affective modulation of attentional performance: implications for corticostriatal circuit function J Neurosci 2004 24 773 780 14749421 10.1523/JNEUROSCI.0949-03.2004
Eagle DM Robbins TW Lesions of the medial prefrontal cortex or nucleus accumbens core do not impair inhibitory control in rats performing a stop-signal reaction time task Behav Brain Res 2003 146 131 144 14643466 10.1016/j.bbr.2003.09.022
Taylor JR Robbins TW Enhanced behavioural control by conditioned reinforcers following microinjections of d-amphetamine into the nucleus accumbens Psychopharmacology 1984 84 405 412 6440188 10.1007/BF00555222
Taylor JR Robbins TW 6-Hydroxydopamine lesions of the nucleus accumbens, but not of the caudate nucleus, attenuate enhanced responding with reward-related stimuli produced by intra-accumbens d-amphetamine Psychopharmacology 1986 90 390 397 3097729 10.1007/BF00179197
Cador M Taylor JR Robbins TW Potentiation of the effects of reward-related stimuli by dopaminergic-dependent mechanisms in the nucleus accumbens Psychopharmacology 1991 104 377 385 1924645
Salamone JD Correa M Motivational views of reinforcement: implications for understanding the behavioral functions of nucleus accumbens dopamine Behav Brain Res 2002 137 3 25 12445713 10.1016/S0166-4328(02)00282-6
Salamone JD Correa M Mingote SM Weber SM Nucleus accumbens dopamine and the regulation of effort in food-seeking behavior: implications for studies of natural motivation, psychiatry, and drug abuse J Pharmacol Exp Ther 2003 305 1 8 12649346 10.1124/jpet.102.035063
Ikemoto S Panksepp J The role of nucleus accumbens dopamine in motivated behavior: a unifying interpretation with special reference to reward-seeking 0165-0173 1999 31 6 41 10611493
Salamone JD Correa M Mingote SM Weber SM Beyond the reward hypothesis: alternative functions of nucleus accumbens dopamine Curr Opin Pharmacol 2005 5 34 41 15661623 10.1016/j.coph.2004.09.004
Mingote S Weber SM Ishiwari K Correa M Salamone JD Ratio and time requirements on operant schedules: effort-related effects of nucleus accumbens dopamine depletions Eur J Neurosci 2005 21 1749 1757 15845103
Ostaszewski P Karzel K Discounting of delayed and probabilistic losses of different amounts European Psychologist 2002 7 295 301 10.1027//1016-9040.7.4.295
Brog JS Salyapongse A Deutch AY Zahm DS The patterns of afferent innervation of the core and shell in the "accumbens" part of the rat ventral striatum: immunohistochemical detection of retrogradely transported fluoro-gold Journal of Comparative Neurology 1993 338 255 278 8308171 10.1002/cne.903380209
Bechara A Damasio AR Damasio H Anderson SW Insensitivity to future consequences following damage to human prefrontal cortex Cognition 1994 50 7 15 8039375 10.1016/0010-0277(94)90018-3
Bechara A Tranel D Damasio H Damasio AR Failure to respond autonomically to anticipated future outcomes following damage to prefrontal cortex Cerebral Cortex 1996 6 215 225 8670652
Bechara A Damasio H Tranel D Damasio AR Deciding advantageously before knowing the advantageous strategy Science 1997 275 1293 1295 9036851 10.1126/science.275.5304.1293
Manes F Sahakian B Clark L Rogers R Antoun N Aitken M Robbins T Decision-making processes following damage to the prefrontal cortex Brain 2002 125 624 639 11872618 10.1093/brain/awf049
Clark L Manes F Antoun N Sahakian BJ Robbins TW The contributions of lesion laterality and lesion volume to decision-making impairment following frontal lobe damage Neuropsychologia 2003 41 1474 1483 12849765 10.1016/S0028-3932(03)00081-2
Fellows LK Farah MJ Different underlying impairments in decision-making following ventromedial and dorsolateral frontal lobe damage in humans Cereb Cortex 2005 15 58 63 15217900 10.1093/cercor/bhh108
Rogers RD Owen AM Middleton HC Williams EJ Pickard JD Sahakian BJ Robbins TW Choosing between small, likely rewards and large, unlikely rewards activates inferior and orbital prefrontal cortex Journal of Neuroscience 1999 19 9029 9038 10516320
Rogers RD Ramnani N Mackay C Wilson JL Jezzard P Carter CS Smith SM Distinct portions of anterior cingulate cortex and medial prefrontal cortex are activated by reward processing in separable phases of decision-making cognition Biol Psychiatry 2004 55 594 602 15013828 10.1016/j.biopsych.2003.11.012
Rogers RD Everitt BJ Baldacchino A Blackshaw AJ Swainson R Wynne K Baker NB Hunter J Carthy T Booker E London M Deakin JF Sahakian BJ Robbins TW Dissociable deficits in the decision-making cognition of chronic amphetamine abusers, opiate abusers, patients with focal damage to prefrontal cortex, and tryptophan-depleted normal volunteers: evidence for monoaminergic mechanisms Neuropsychopharmacology 1999 20 322 339 10088133 10.1016/S0893-133X(98)00091-8
Mobini S Body S Ho MY Bradshaw CM Szabadi E Deakin JF Anderson IM Effects of lesions of the orbitofrontal cortex on sensitivity to delayed and probabilistic reinforcement Psychopharmacology 2002 160 290 298 11889498 10.1007/s00213-001-0983-0
Kheramin S Body S Herrera FM Bradshaw CM Szabadi E Deakin JF Anderson IM The effect of orbital prefrontal cortex lesions on performance on a progressive ratio schedule: implications for models of inter-temporal choice Behav Brain Res 2005 156 145 152 15474659 10.1016/j.bbr.2004.05.017
Kheramin S Body S Ho MY Velazquez-Martinez DN Bradshaw CM Szabadi E Deakin JF Anderson IM Effects of orbital prefrontal cortex dopamine depletion on inter-temporal choice: a quantitative analysis Psychopharmacology (Berl) 2004 175 206 214 14991223
Kheramin S Body S Ho M Velazquez-Martinez DN Bradshaw CM Szabadi E Deakin JF Anderson IM Role of the orbital prefrontal cortex in choice between delayed and uncertain reinforcers: a quantitative analysis Behav Processes 2003 64 239 250 14580695 10.1016/S0376-6357(03)00142-6
Halliday G Harding A Paxinos G Paxinos G Serotonin and tachykinin systems The Rat Nervous System 1995 London, Academic Press 929 974
Mehlman PT Higley JD Faucher I Lilly AA Taub DM Vickers J Suomi SJ Linnoila M Low CSF 5-HIAA concentrations and severe aggression and impaired impulse control in nonhuman primates American Journal of Psychiatry 1994 151 1485 1491 7522411
Åsberg M Träskman L Thorén P 5-HIAA in the cerebrospinal fluid: a biochemical suicide predictor Archives of General Psychiatry 1976 33 1193 1197 971028
Linnoila M Virkkunen M Scheinin M Nuutila A Rimon R Goodwin FK Low cerebrospinal fluid 5-hydroxyindoleacetic acid concentration differentiates impulsive from nonimpulsive violent behavior Life Sciences 1983 33 2609 2614 6198573 10.1016/0024-3205(83)90344-2
Brown GL Linnoila M CSF serotonin metabolite (5HIAA) studies in depression, impulsivity and violence. Journal of Clinical Psychiatry 1990 51 (supplement 4) 31 41 1691169
Linnoila M Virkkunen M George T Higley D Impulse control disorders International Clinical Psychopharmacology 1993 8 (Supplement 1) 53 56 7504705
Mann JJ Neurobiology of suicidal behaviour Nat Rev Neurosci 2003 4 819 828 14523381 10.1038/nrn1220
Biggio G Fadda F Fanni P Tagliamonte A Gessa G Rapid depletion of serum tryptophan, brain tryptophan, serotonin and 5-hydroxyindoleacetic acid by a trytophan-free diet Life Sciences 1974 14 1321 1329 4823644 10.1016/0024-3205(74)90440-8
Clemens JA Bennett DR Fuller RW The effect of a tryptophan-free diet on prolactin and corticosterone release by serotonergic stimuli Horm Metab Res 1980 12 35 38 6965657
Delgado PL Charney DS Price LH Landis H Heninger GR Neuroendocrine and behavioral effects of dietary tryptophan restriction in healthy subjects Life Sci 1989 45 2323 2332 2601581 10.1016/0024-3205(89)90114-8
Anderson IM Richell RA Bradshaw CM The effect of acute tryptophan depletion on probabilistic choice J Psychopharmacol 2003 17 3 7 12680734 10.1177/0269881103017001687
Rogers RD Tunbridge EM Bhagwagar Z Drevets WC Sahakian BJ Carter CS Tryptophan depletion alters the decision-making of healthy volunteers through altered processing of reward cues Neuropsychopharmacology 2003 28 153 162 12496952 10.1038/sj.npp.1300001
Mobini S Chiang TJ Ho MY Bradshaw CM Szabadi E Effects of central 5-hydroxytryptamine depletion on sensitivity to delayed and probabilistic reinforcement Psychopharmacology (Berl) 2000 152 390 397 11140331 10.1007/s002130000542
Wade TR de Wit H Richards JB Effects of dopaminergic drugs on delayed reward as a measure of impulsive behavior in rats Psychopharmacology 2000 150 90 101 10867981 10.1007/s002130000402
Winstanley CA Theobald DE Dalley JW Robbins TW Interactions between serotonin and dopamine in the control of impulsive choice in rats: therapeutic implications for impulse control disorders Neuropsychopharmacology 2005 30 669 682 15688093
Rogers RD Lancaster M Wakeley J Bhagwagar Z Effects of beta-adrenoceptor blockade on components of human decision-making Psychopharmacology (Berl) 2004 172 157 164 14716472 10.1007/s00213-003-1641-5
O'Carroll RE Papps BP Decision making in humans: the effect of manipulating the central noradrenergic system J Neurol Neurosurg Psychiatry 2003 74 376 378 12588933 10.1136/jnnp.74.3.376
Aston-Jones G Shipley MT Grzanna R Paxinos G The locus coeruleus, A5 and A7 noradrenergic cell groups The Rat Nervous System 1995 London, Academic Press 183 213
Stroustrup B The C++ Programming Language 1986 Reading, Massachusetts, Addison-Wesley
Cardinal RN Aitken MRF Whisker (version 2) [ www.whiskercontrol.com ] 2001 Cambridge, UK, Cambridge University Technical Services Ltd
Paxinos G Watson C The Rat Brain in Stereotaxic Coordinates 1998 Fourth London, Academic Press
Howell DC Statistical Methods for Psychology 1997 Fourth Belmont, California, Wadsworth
Levene H Oklin I Robust tests for the equality of variance Contributions to probability and statistics 1960 Palo Alto, California, Stanford University Press
Keppel G Design and analysis: a researcher's handbook 1982 Second London, Englewood Cliffs: Prentice-Hall
Mauchly JW Significance test for sphericity of a normal n-variate distribution Annals of Mathematical Statistics 1940 11 204 209
Huynh H Feldt LS Conditions under which mean square ratios in repeated measures designs have exact F-distributions Journal of the American Statistical Association 1970 65 1582 1589
Šidák Z Rectangular confidence regions for the means of multivariate normal distributions Journal of the American Statistical Association 1967 81 826 831
| 15921529 | PMC1177958 | CC BY | 2021-01-04 16:03:48 | no | BMC Neurosci. 2005 May 28; 6:37 | utf-8 | BMC Neurosci | 2,005 | 10.1186/1471-2202-6-37 | oa_comm |
==== Front
BMC NeurosciBMC Neuroscience1471-2202BioMed Central London 1471-2202-6-401594386010.1186/1471-2202-6-40Research ArticleGonadal steroids differentially modulate neurotoxicity of HIV and cocaine: testosterone and ICI 182,780 sensitive mechanism Kendall Sherie L [email protected] Caroline F [email protected] Avindra [email protected] Jadwiga [email protected] Cantey L [email protected] Charles F [email protected] Rosemarie M [email protected] Department of Behavioral Sciences, University of Kentucky, Lexington, Kentucky, USA2 Department of Neurology, Johns Hopkins University, Baltimore, MD, USA3 Department of Anatomy and Neurobiology, University of Kentucky, Lexington, Kentucky, USA4 Behavioral Neuroscience Program, Department of Psychology, University of South Carolina, Columbia, SC, USA2005 8 6 2005 6 40 40 19 12 2004 8 6 2005 Copyright © 2005 Kendall 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
HIV Associated Dementia (HAD) is a common complication of human immunodeficiency virus (HIV) infection that erodes the quality of life for patients and burdens health care providers. Intravenous drug use is a major route of HIV transmission, and drug use is associated with increased HAD. Specific proteins released as a consequence of HIV infection (e.g., gp120, the HIV envelope protein and Tat, the nuclear transactivating protein) have been implicated in the pathogenesis of HAD. In primary cultures of human fetal brain tissue, subtoxic doses of gp120 and Tat are capable of interacting with a physiologically relevant dose of cocaine, to produce a significant synergistic neurotoxicity. Using this model system, the neuroprotective potential of gonadal steroids was investigated.
Results
17β-Estradiol (17β-E2), but not 17α-estradiol (17α-E2), was protective against this combined neurotoxicity. Progesterone (PROG) afforded limited neuroprotection, as did dihydrotestosterone (DHT). The efficacy of 5α-testosterone (T)-mediated neuroprotection was robust, similar to that provided by 17β-E2. In the presence of the specific estrogen receptor (ER) antagonist, ICI-182,780, T's neuroprotection was completely blocked. Thus, T acts through the ER to provide neuroprotection against HIV proteins and cocaine. Interestingly, cholesterol also demonstrated concentration-dependent neuroprotection, possibly attributable to cholesterol's serving as a steroid hormone precursor in neurons.
Conclusion
Collectively, the present data indicate that cocaine has a robust interaction with the HIV proteins gp120 and Tat that produces severe neurotoxicity, and this toxicity can be blocked through pretreatment with ER agonists.
==== Body
Background
A particularly devastating complication of HIV infection is a pervasive form of damage to the brain, HAD [1]. The overall incidence of severe HAD is estimated at about 30% of the HIV infected population [2]. However, HAD occurs more often in HIV-positive IV drug users than HIV-positive non-drug users [3-5]. Neuroimaging and autopsy studies demonstrate that the basal ganglia and frontal lobes are preferentially affected by HAD [5,6]. These structures may degenerate with chronic psychostimulant (methamphetamine, cocaine) abuse [7,8], eventually leading to a Parkinson type syndrome [9]. Injection of abused drugs, such as cocaine, has been noted to accelerate the progression of HIV infection to AIDS status and to HAD [8,10-12].
In the United States, cocaine use plays a larger role in HIV transmission to women than it does to men [13]. HIV infected women have lower initial viral loads than men, progress to AIDS status at the same rate as men, yet have higher mortality and lower life expectancy than men [14-16]. How sex differences contribute to the progression to HAD is largely unknown. However, female gender or estrogenic steroids are recognized as protective against several neurological insults, including animal models of ischemia, oxidative stress, and psychostimulant-induced neurotoxicity [17-22] and human neurodegenerative diseases such as Alzheimer's disease [23-26] and Parkinson's disease [27-29]. Tissue culture studies have found that the estrogenic steroid, 17β-E2, is protective against neurotoxic HIV proteins [30,31]. Collectively, these gender/hormonal effects suggest that gonadal hormones may play a differential role in effects of drug abuse, HIV infection, and HAD.
Neurotoxic interactions between the HIV proteins, Tat and gp120, and abused psychostimulant drugs have been previously reported [30,32]. More recently, 17β-E2 proved neuroprotective in vitro [30,32]; however, it is unknown whether this neuroprotection is specifically estrogenic, or also effected by PROG, and whether it contains an androgenic component. Therefore, the aim of the present study was to determine which gonadal steroids provide neuroprotection against the synergistic neurotoxicity of HIV proteins in the presence of cocaine. We investigated the potential for neuroprotection by T, whether this is mediated through an ER mechanism and the potential concentration-dependent neuroprotection by PROG, DHT and cholesterol.
We report here concentration dependent neuroprotection by T mediated through an ICI 182,780-sensitive mechanism. Incomplete neuroprotection at the concentrations (nM) tested was also provided by PROG, DHT, and cholesterol.
Results
17β-, not 17α-Estradiol, protects against HIV proteins plus cocaine synergistic neurotoxicity
Clear and robust synergistic neurotoxicity of the HIV proteins Tat and gp 120 was repeatedly observed when combined with a physiologically relevant dose of cocaine (Figs. 1C, 1D, 2 Left Panel, 3 Left Panel, 4 Top Panel, 5 Top Panel). False color visualization of the intensity of trypan blue within neurons verified the assay for cell death and demonstrated the synergistic toxicity of cocaine with the HIV proteins (Fig. 1). Furthermore, this neurotoxicity is precluded by pretreatment with 10 nM dose of 17β-E2, but not with 17α-E2. Neither the HIV proteins, Tat plus gp120, nor cocaine alone were more toxic than the Locke's buffer control; however, in combination they produced synergistic neurotoxicity (Fig. 2). The ANOVA confirmed the presence of a significant interaction of the HIV proteins with cocaine (F(1,24) = 15.38, p < 0.0008; n = 6 each point) (Fig. 2 Left Panel). The presence of this significant neurotoxic effect was demonstrated in every experiment at p < 0.005. The stereoisomers of estradiol demonstrated a significant treatment effect against this toxicity (F(2,15) = 6.95, p < 0.007). The 17β-stereoisomer of estradiol demonstrated significant neuroprotection (F(1,24) = 24.71, p < 0.0001; n = at least 3 each point). The percent neuronal death with 17α-E2 treatment was not significantly different from the synergistic toxicity control (Fig. 2 Right Panel).
DMSO at 10 nM (0.1%) served as the solvent for 17α-E2 and for ICI-182,780; therefore, it was tested for toxicity and was not significantly different from Locke's buffer incubation (data not shown). β-Cyclodextrin (100 nM) served as an encapsulating carrier to enhance solubility for the steroids 17β-E2, T, PROG and cholesterol and was therefore tested for neuroprotection. When coincubated with the toxic combination of HIV proteins plus cocaine, this carrier also had no significant effect on cell death (data not shown).
Progesterone provides partial neuroprotection against HIV proteins plus cocaine synergistic neurotoxicity
To study the effect of PROG on HIV protein and cocaine synergistic toxicity we treated cultures with 1 nM, 10 nM, and 100 nM doses of PROG prior to adding the toxic combination of HIV proteins and cocaine (Fig. 3 Right Panel). PROG demonstrated significant neuroprotection (F(1,28) = 11.03, p < 0.0025; n = 5 each point). There was no concentration-dependent effect on percent neuronal death (F(1,28) < 1.0). Further, the neuroprotection was incomplete at these concentrations as the magnitude of cell death was significantly greater than that observed with Locke's Buffer control (ps < 0.01 for all concentrations tested). Fig. 3 Left Panel shows the significant interaction of the HIV proteins with cocaine confirming the presence of synergistic neurotoxicity (F(1,16) = 10.97, p < 0.0044; n = 5 each point;).
Testosterone completely protects against HIV proteins plus cocaine synergistic neurotoxicity
Concentrations of 1 pM, 100 pM, 1 nM, 10 nM, 100 nM and 10 μM of T were tested against the synergistically neurotoxic combination of HIV proteins and cocaine (Fig. 4). A significant interaction of the HIV proteins with cocaine (F(1,20) = 8.47, p < 0.0087; n = at least 6 each point) confirmed the presence of synergistic neurotoxicity (Fig. 4 Top Panel). T treatment provided significant neuroprotection (F(1,57) = 12.71, p < 0.0007; n = at least 4 each point). Specifically, there was a concentration-dependent effect of T with a significant linear decrease in percent neuronal death as a function of increasing dose (F(1,57) = 22.56, p < 0.0001) (Fig. 4 Middle Panel). The neuroprotective effect of T was characterized against the negative control of Locke's buffer vehicle and the positive control of the synergistically toxic combination of HIV proteins with cocaine. A ceiling effect of maximal neuroprotection at T concentrations of 10 nM or greater, was noted; magnitude of cell death was not significantly different from incubation with Locke's buffer vehicle. A floor effect was identified at the lowest concentrations tested, 100 pM or less, with the magnitude of cell death not significantly different from that observed by the toxic synergism of the HIV proteins with cocaine.
Dihydrotestosterone provides limited neuroprotection against HIV proteins plus cocaine synergistic neurotoxicity
DHT, the androgen receptor (AR) active metabolite of T, was tested for neuroprotection in the linear range of the T concentration-effect curve, 1 nM, 10 nM and 100 nM, against the toxic synergism of HIV proteins with cocaine (Fig. 4). DHT treatment was confirmed to provide significant neuroprotection (F(1,57) = 10.66, p < 0.002; n = at least 4 each point). However, there was no concentration-dependent effect on percent neuronal death (F(1,57) <1.0). Further, the neuroprotection was incomplete across these concentrations as the magnitude of cell death was significantly greater than that observed with Locke's Buffer control (F(1,57) = 6.72, p < 0.012) (Fig. 4 Bottom Panel). Because the DHT treatments were run in parallel with the T concentration effects, the control values are the same for both sets of treatments. That is, the presence of synergistic neurotoxicity was indicated by the same significant interaction of the HIV proteins with cocaine (F(1,20) = 8.47, p < 0.0087) as it was for the T treatments (Fig. 4 Top Panel). Ethanol at 10 nM (0.1%) served as the solvent for DHT; therefore, it was tested for toxicity and was found not to differ significantly from Locke's buffer incubation (data not shown).
ICI-182,780 blocks testosterone's neuroprotection against HIV proteins plus cocaine synergistic neurotoxicity
The cultures were pretreated with the specific ER antagonist, ICI-182,780 at 100 nM before addition of T (10 nM) then challenged with the HIV proteins plus cocaine synergistic toxicity (Fig. 5). A significant interaction of ICI-182,780 with T was observed (F(1,22) = 33.71, p < 0.0001; n = at least 6 each point). Importantly, the ICI compound was not more toxic to the cultures than was Locke's buffer; nor was it neuroprotective against the toxic challenge. T at 10 nM provided neuroprotection (F(1,65) = 121.26, p < 0.0001), which was not significantly different from incubation with Locke's buffer vehicle alone. This neuroprotective effect of T was fully blocked by pretreatment of the cultures with ICI-182,780 (100 nM) (Fig. 5 Middle Panel). A significant interaction of the HIV proteins with cocaine indicated the presence of synergistic neurotoxicity (F(1,28) = 61.04, p < 0.0001; n = 8 each point) (Fig. 5 Top Panel).
Cholesterol provides concentration dependent neuroprotection against HIV proteins plus cocaine synergistic neurotoxicity
Cholesterol was added to the cultures at concentrations of 1 nM, 10 nM, and 100 nM prior to treatment with the synergistic toxic combination of HIV proteins plus cocaine (Fig. 5). Significant neuroprotection by cholesterol against this synergistic toxicity was observed (F(1,65) = 55.74, p < 0.0001; n = at least 3 each point) (Fig. 5 Bottom Panel). There was a concentration-dependent effect of cholesterol with a significant linear decrease in percent neuronal death as concentration increased (F(1,65) = 4.4042.73, p < 0.0398). The neuroprotection afforded by the high dose (100 nM) of cholesterol was asymptotic to a ceiling effect; the magnitude of cell death was not significantly different from that observed by Locke's buffer incubation. Because the cholesterol treatments were performed concurrently with the ICI-182,780 treatments, the control values are the same for both sets of treatments. The presence of synergistic neurotoxicity was indicated by the same significant interaction of the HIV proteins with cocaine as it was for the ICI-182,780 treatments (F(1,28) = 61.04, p < 0.0001) (Fig. 5 Top Panel).
Discussion
Neuroprotection is clearly afforded by physiological concentrations of 17β-E2 in this human fetal neuronal model of gp120 and Tat HIV protein neurotoxic synergism with cocaine [30]. In the current studies, we have replicated and extended these findings to examine the ability of other gonadal steroids to provide neuroprotection against this combined insult. We now show robust neuroprotection mediated through the ER is provided by T while PROG, DHT and cholesterol provide only partial or incomplete neuroprotection in the same concentration range.
There are at least two mechanisms for the neuroprotective actions of estradiol: receptor mediated cellular events; and non-receptor mediated anti-oxidant effects. Stereo-specific neuroprotection by estradiol (Fig. 2) suggests a receptor-mediated mechanism rather than global scavenging of reactive oxygen species (ROS) by the phenolic ring. Both of the estradiol stereoisomers, but not PROG, T, DHT, nor cholesterol contain the phenolic moiety of estradiol. In contradiction to our results, neuroprotection by neutralization of ROS would be exhibited by both estradiol isomers but not the other steroid compounds tested. Moreover, the phenolic ring's efficacy in quenching ROS usually requires higher concentrations than needed for receptor-mediated activity [33-35]. Our data suggest that there is a specific ER modulation of the neurotoxic cascade preventing neural death, rather than nonspecific scavenging of ROS by estrogens.
In these studies, the neurons were exposed to estrogens for 15 hours. This time course (15-hour incubation) does not preclude estrogen-mediated genomic activity; however, estrogens are known to act through both genomic and non-genomic pathways [36,37]. Nevertheless, the lack of effect by the transcriptionally inactive stereoisomer of estradiol (17α-E2) provides evidence for a genomic neuroprotective mechanism. Neuroprotection by 17α-E2 has been reported in human neuronal cultures [38,39]. This neuroprotection, which was not seen in our model, was realized in some non-genomic manner such as activation of signal transduction. Research reports also support three other non-genomic actions of steroids that may protect neurons from toxic insults [40]: 1) an indirect effect due to structural and functional perturbation of membrane properties by the intercalation of the cholesterol moiety comprising the steroids; 2) a direct activation of membrane-bound steroid specific protein recognition sites; and 3) modulation of classical neurotransmitter receptor activity. Furthermore, neuroprotection conferred by an individual agent may not be attributable to a single mode of action. All steroid hormones, including 17α-E2, might be expected to modulate neurotransmission by altering the relative steroidal composition of the neuronal membrane. Since 17α-E2 failed to show neuroprotection, these nonspecific membrane effects are ruled out, but activity at a membrane receptor or modulation of neurotransmitter activity is not. We tested six steroids (the two estradiol stereoisomers, PROG, T, DHT and cholesterol) at physiologically relevant nanomolar (nM) concentrations against the synergistic neurotoxic challenge to delineate neuroprotection by steroids in this model of HAD, while considering that each steroid may operate by different as well as pleiotropic means.
Treatments with PROG may be neuroprotective [41-43]. The neuroprotective capability of PROG in the physiological range [44] was also exhibited against our neurotoxic challenge (Fig. 3). Although PROG's neuroprotection was significant, it was incomplete in the nM concentration range tested, as PROG was significantly different from the negative control (i.e., relative to incubation in Locke's buffer). The linear regression analysis confirmed the significance of the neuroprotection. However, PROG as well as some of its metabolites are known to act both as agonists at the intracellular and membrane PROG receptors and as allosteric modulators of some neurotransmission [45].
In contrast, T (10 nM) produced neuroprotection of a similar magnitude to that achieved by 17β-E2 (Fig. 3). A log plot of the data (Fig. 4 Middle Panel) describes a linear function between 100 pM and 10 nM. No protection is seen at 1 pM, whereas concentrations above 100 nM show maximal protection, thus establishing the EC50 for the neuroprotective effect of T between 1 pM and 100 nM. The pharmacokinetics of a concentration-dependent effect are consistent with a receptor-mediated mechanism of neuroprotection. Because T can be aromatized to estradiol by P450 aromatase expressed by astrocytes [46] and by neurons [47] in rat brain, it is not entirely surprising to see a robust neuroprotective effect of T, as was seen with 17β-E2. Such neuroprotection could be produced through activation of the ER. Therefore, the question remained whether an androgenic component to the presumptive steroid receptor-mediated neuroprotection is operating.
T is not specific in identifying the receptor mediating its effects because it is the prohormone for both DHT, the principal ligand at the AR, and 17β-E2, the principal ligand at the ER. DHT is a useful androgen for evaluating the relative contributions of AR and ER-mediated neuroprotective effects because it does not serve as an agonist at the ER nor can it be converted to estradiol. Our toxic challenge test results with DHT suggest that the AR may mediate neuroprotection, but do not exclude the possibility of effects by other means (Fig. 4 Bottom Panel). To clarify whether T was acting at the AR or the ER, we used the specific ER antagonist, ICI-182,780, while treating the cultures with T and the toxic combination of HIV proteins and cocaine. The ER antagonist completely blocked T's neuroprotection thus demonstrating that the effect is mediated through the ER (Fig. 5 Middle Panel) at physiological concentrations [48,49]. If an AR-mediated component were active in T's neuroprotection it would have been revealed when T was tested while the ER was pharmacologically blocked by ICI-182,780. Therefore, we conclude that T's neuroprotection is mediated through the ER with no portion attributable to AR activation or to non-genomic pathways.
As a control for general steroidal structural and receptor activation we tested cholesterol, which is known to modulate membrane fluidity but is neurally inactive. We saw that cholesterol's concentration-dependent neuroprotection against synergistic toxicity of the HIV proteins with cocaine is more efficacious than that of PROG and DHT (Fig. 5 Bottom Panel) in that 100 nM cholesterol was fully neuroprotective whereas the same dose of PROG or DHT was not. However, cholesterol is less potent than the gonadal steroids, 17β-E2 and T in that 10 nM cholesterol was not fully neuroprotective, whereas the same dose of the steroids was. The mechanism of this protection is interesting because the cholesterol moiety is common to all steroids, but 17α-E2 was not neuroprotective. Furthermore, the ICI-182,780-sensitive neuroprotection provided by T does not accommodate a nonspecific cholesterol-mediated component. In this model, cholesterol may not promote neuronal survival by any intrinsic activity of its own, but may be serving as a synthesis precursor for an explicitly protective steroid, e.g., 17β-E2. Others have also shown increased bioavailability of cholesterol to interact with effects mediated by steroids in the brain [50,51]. Evidence for all the necessary enzymes of synthesis in human brain has been reported [52,53] except for one, which has been identified in rat brain [54]. Because we use primary cell cultures derived from human fetal brain, it is likely that the metabolic machinery is present, operational and facilitating cholesterol's neuroprotection through an estrogenic end product. In support of our findings, neuroprotection by locally synthesized estrogen has recently been demonstrated [55], which when considered with the current findings, suggests possibilities for steroid prodrugs serving as neuroprotective therapeutics. However, others have demonstrated that cholesterol significantly reduces the neurotoxicity of gp120 by modulating membrane fluidity. In human neuroblastoma cells enriching cholesterol significantly reduced gp120 binding and consequently, the biochemical events triggered by the viral protein leading to necrotic death, while depleting cholesterol had the opposite effect [56]. Therefore the physiochemical properties of the membrane in mediating neurotoxicity/protection may be as significant as some pharmacological agents.
Synergistic neurotoxicity of the HIV proteins gp120 and Tat with cocaine was demonstrated in each experiment. Although neither the HIV protein combination nor cocaine alone was more neurotoxic than the Locke's buffer media, the simultaneous incubation of cells with both the HIV proteins and cocaine produced significant neuronal death. The mechanism by which cocaine facilitates the HIV protein toxicity is not well characterized. Recombinant variants of Tat protein (Tat 1–72 and Tat 1–86) are known to induce neuronal degeneration by mechanisms that involve triggering of apoptotic cascades [57,58]. Cocaine was recently shown to enhance gp120 induced neuronal apoptosis in vivo via a mechanism implicating glutamate excitotoxicity and iNOS expression with abnormal NO production in the neocortex of rat [59]. In vitro work in our lab with rat hippocampal cells supports a similar mechanism for cocaine's enhancement of Tat mediated toxicity by augmentation of ROS production and mitochondrial depolarization [60]. Likewise cocaine at very high concentrations in vitro may induce neuronal apoptosis [61]. Accordingly, estrogen neuroprotection could be via stimulation of the expression of anti-apoptotic proteins [62,63], or alternatively suppression of pro-apoptotic gene transcripts [21,64], although both of these interesting possibilities await further experimentation.
Conclusion
In conclusion, we demonstrated that T provides neuroprotection through an ICI-182,708-sensitive mechanism against the synergistic toxicity of gp120 and Tat HIV proteins with cocaine. Additionally, we show that PROG, DHT and cholesterol provide some less efficacious neuroprotection than do the steroids, 17β-E2 and T. It is conceivable that gonadal steroids may provide biologically based gender-specific prophylaxis and/or amelioration either in the time of onset or symptomology of HAD and drugs of abuse. These results have implications for the innovation of therapeutic strategies, including steroid based neuroprotective prodrugs. Moreover, given that the neuropathology of HAD may have commonalities with dementias of other etiologies, the development of successful pharmacology for HAD would surely have applications for other equally devastating and costly neurodegenerative conditions.
Methods
Culture of human brain cells
Neuronal cultures were prepared as described previously [65-67]. Human fetal brain specimens of gestational age 12–14 weeks were obtained with the consent of women opting for elective termination of pregnancy. Protocols were carried out with strict adherence to the guidelines of the National Institutes of Health (NIH) and with approval from the University of Kentucky Institutional Review Board. Briefly, after mechanical dissociation, the cells were suspended in Opti-MEM with 5% heat-inactivated fetal bovine serum, 0.2% N2 supplement and 1% antibiotic solution (penicillin G sodium 104 units/ml, streptomycin sulfate10 μg/ml and amphotericin-B 25 μg/ml) (GIBCO). The cells were maintained in culture flasks at 37°C in a humidified atmosphere of 5% CO2 and 95% air for at least a month. At least 3 days before the experiment was performed the cells were plated in flat bottom 96 well plates coated with poly-d-lysine. Sample cultures were stained for the neuron-specific enolase, microtubule-associated protein-2, synaptophysin, and glial fibrillary astrocytic protein (GFAP). The cultures were thus characterized as more than 80% homogenous for neurons. The remaining cells were predominantly astrocytes (GFAP positive) with less than 1% microglia/macrophages, as determined by RCA-1 lectin and CD68 staining [67-69]-. Sample cultures were further characterized for the presence of ER-α and ER-β and dopaminergic neurons. As described earlier, ERs were visualized in 5–10% of cultured astrocytes and neurons in the cytoplasm and nuclei of both by immunostaining. Antigenicity for dopamine and dopamine transporter were co-localized in nearly 60% of neurons. Dopamine receptor D1A was seen in 50% of cells while receptor D2 was seen in 40% of cells. ER co-localized with cells staining for dopamine as well as D1A and D2 receptor containing neurons [30].
Recombinant Tat and gp120 proteins
Recombinant Tat was prepared as described previously [67,70] with minor modifications. The tat gene encoding the first 72 amino acids was amplified from HIVBRU obtained from Dr. Richard Gaynor through the AIDS repository at the NIH and inserted into an E coli vector PinPoint Xa-2 (Promega). This construct allowed the expression the Tat1-72 protein as a fusion protein naturally biotinylated at the N-terminus. The biotinylated Tat protein was purified on a column of soft release avidin resin, cleaved from the fusion protein using factor Xa, eluted from the column, then desalted with a PD10 column. The Tat protein was further purified of endotoxin by passage through a polymycin B, cyanogen bromide, immobilized on cross-linked 4% beaded agarose column (Sigma). All purifications steps contained dithiothreitol to prevent oxidation of the proteins. Tat Protein was more than 95% pure as determined by SDS-PAGE followed by silver staining. Analysis by HPLC using a C4 column showed a single symmetrical peak. Western blot analysis showed that this preparation contained both monomeric and dimeric forms of Tat1-72. The functional activity of Tat1-72 was confirmed using a transactivation assay in HL3T1 cells containing an HIV-1 LTR-CAT construct. Chiron Corporation made a gift of gp120 from HIVSF2 which was described previously [30]. Briefly, recombinant gp120 was made in a Chinese Hamster Ovary cell line. Purification yielded 95% pure gp120 as determined by Western blot analysis. The Tat and gp120 preparations contained less than 1 pg/ml endotoxin as determined using a Pyrochrome Chromogenic test kit (Associates of Cape Cod, Inc., Falmouth, MA). The Tat proteins were stored in a lyophilized form and gp120 as a stock solution in water at -80°C in endotoxin-free siliconized microfuge tubes until taken for experimentation. Tat and gp120 are highly susceptible to degradation and loss of biological activity with each freeze-thaw cycle. Therefore, single aliquots were used for each experiment. Any remaining solution was discarded. Prior work has shown that no significant toxicity is seen with Tat less than 80 nM, gp120 less than 40 pM, or Tat (40 nM) plus gp120 (30 pM) [30]. For these studies, we used subtoxic dosage levels of Tat1-72 at 40 nM and gp120 at 32.5 pM.
Drug levels
Cocaine hydrochloride obtained from the NIDA Drug Supply Program was solvated in Locke's buffer (pH 7.2) immediately prior to use. The dosage of cocaine (500 ng/ml) in culture translates to 1.6 μM, which is several orders of magnitude less than studies demonstrating a direct cytotoxic effect in culture [71-73]-. Fetal brain levels of cocaine of 1750 +/- 250 ng/ml were associated with fetal plasma levels of 500 ng/ml following repeated daily dosing in rat [74]. The plasma levels of cocaine from fetuses of mothers exposed to cocaine using the rat IV model [75] was 500 ng/ml as this represents 1/5 the peak brain levels in the fetus. This is also a concentration within the physiological levels experienced by human IV cocaine drug users [76]. Thus 500 ng/ml (1.6 μM) represents a physiologically relevant dose of cocaine to the neurons. Cocaine levels as high as 16 μM are not neurotoxic in human fetal neuronal cultures [30]. In these studies the cocaine solution was diluted such that addition of 1:10 (v/v) to the culture dishes resulted in a final concentration of 1.6 μM. This dose of cocaine was added concurrently with the Tat and gp120 proteins to the cultures.
Neurotoxicity assay
At the time of the experimental treatment, the culture media was replaced with Locke's buffer containing (in mM) 154 NaCl, 5.6 KCl, 2.3 CaCl2, 1 MgCl2, 3.6 NaHCO3, 5 glucose, 5 N-2-hydroxyethylpiperazine-N'2-ethanesulfonic acid (HEPES) and 1% antibiotic solution (penicillin G sodium 104 units/ml, streptomycin sulfate10 μg/ml and amphotericin-B 25 μg/ml) (pH 7.2). Cells were incubated for 15 hours in Locke's buffer or with subtoxic concentrations of the HIV proteins, Tat1-72 (40 nM) plus gp120 (32.5 pM) [30], or cocaine (1.6 μM), or the HIV proteins plus cocaine to demonstrate the synergistic effect on cell death of HIV proteins with cocaine. To replicate the stereoisomer specific neuroprotective properties of estrogen, the cultures were treated with either 17β-E2 (10 nM), or 17α-E2 (10 nM) immediately preceding addition of the HIV proteins plus cocaine. To investigate the neuroprotective properties of other gonadal steroids, the cells were treated with various concentrations of PROG or T, immediately followed by exposure to the HIV proteins plus cocaine. To determine the mechanism of T's neuroprotection, the cultures were treated with various concentrations of DHT (5α-Androstan-17β-ol-3-one; 17β-Hydroxy-5α-androstan-3-one), or ICI-182,780 (100 nM) plus T (10 nM) before the HIV proteins plus cocaine were added. Various concentrations of cholesterol were used as treatment prior to incubation with the HIV proteins plus cocaine to further characterize the nature of gonadal steroid neuroprotection. ICI-182,780, DMSO (10 nM), β-cyclodextrin (100 nM) and ethanol (10 nM) were also tested for neurotoxic/protective effects. ICI-182,780 and 17α-E2 were solvated in DMSO. 17β-E2, T, PROG and cholesterol were β-cyclodextrin encapsulated to provide water solubility. Ethanol was used to solvate DHT. Tat1-72 was produced as indicated above, whereas gp120 was a gift from Chiron Corporation. The ICI-182,780 compound was obtained from Tocris Cookson, Inc. (Ellisville, MO). All other chemical agents were supplied by Sigma Chemicals (St. Louis, MO). At the end of the incubation period neuronal death was assessed by trypan blue (Sigma) exclusion assay as described previously [42,65,66]. In brief, each experiment was conducted in triplicate and at least three independent squads were treated with each pharmacological agent. Five randomly selected fields from each well were photographed using an inverted microscope (Nikon Diphot, 40X) and coded by a technician blinded to the well treatments. Viable and dead neurons in each photomicrograph were counted before decoding for statistical analysis. In representative fields the neurotoxicity assay was verified by false color visualization using the MCID digital camera-based system interfaced with a computer (Imaging Research, Ontario, Canada).
The neuronal and glial phenotypes seen in these cultures have been well characterized immunohistochemically and morphologically in our hands [65,66] as well as by others [77-79]-. Furthermore, the glia readily attach to the surface of the culture dish, whereas neurons do not [80]. In our hands, astrocytic attachment to the culture dish is requisite for the survival of the neurons, which in turn, attach to the substrate of glia. In this way, the glia are seen in a specific plane of microscopy that differs from the microscopic plane in which the neurons are seen. Therefore, we are confident of our ability to visually recognize and enumerate neurons as distinct from glia in these cultures.
Analysis
Data are expressed as percent neuronal cell death/total viable neurons (means ± SEM). A 2 (presence or absence of the viral proteins gp120 and Tat) × 2 (presence or absence of cocaine) analysis of variance (ANOVA) design was employed to ascertain the potential for synergistic viral protein/cocaine-induced neurotoxicity. A significant interaction of the HIV proteins with Coc, as graphically illustrated by the lines diverging from parallel, provided evidence of toxic synergism. One-way ANOVAs were used to assess the neuroprotective potential of the various gonadal steroid pre-treatments. Regression analysis was used to determine concentration-effect relationships [81]. Planned Tukey-Kramer comparisons were used to determine specific treatment effects. An α level of p < 0.05 was considered significant for all statistical tests employed. Computer assisted analyses utilized BMDP Statistical Software, Release 7, Los Angeles, CA (1993) and SPSS Statistical Software release 11.5.0, SPSS, Inc., Chicago, IL (2004).
Authors' contributions
SLK participated in the conception and design of the study, carried out the cell culture studies, quantified and analyzed the data, and drafted the manuscript. CFA participated in the design and coordination of the study. AN participated in the conception and design of the study and manuscript preparation. JT-C characterized the cell cultures and participated in manuscript preparation. CLL helped interpret the data and contributed to the preparation of the manuscript. CFM performed statistical and graphical analyses and helped draft the manuscript. RMB participated in the conception and design of the study and manuscript preparation and was the primary mentor of SLK. All authors read and approved the final manuscript.
Acknowledgements
This research was supported by NIH research grants to Rosemarie M. Booze (DA014401, DA13137), Avindra Nath (NS039253, P20RR015592, R01NS039253), and Charles F. Mactutus (HD043680, DA09160). Sherie L. Kendall was supported by training grants from the NHLB (K30 HL04163) and NIA (T32 AG00242). We also thank Philip D. Ray for technical assistance in producing the Tat protein used in these studies. Dr. Michael Aksenov and Dr. Marina Aksenova provided technical assistance with imaging of the cultures and very helpful discussions for which we are grateful.
Figures and Tables
Figure 1 True and false color photomicrographs of human neurons (A-C). Control culture after trypan blue exclusion assay in true color (A) and false color (B). Cultures exposed to HIV proteins 40 nM Tat plus 32.5 pM gp120 and 1.6 μM cocaine show synergistic toxicity trypan blue exclusion assay in true color (C) and false color (D).
Figure 2 17β-, not 17α-E2, protects human fetal neurons against the synergistically toxic of incubation with Tat [40 nM] plus gp120 [32.5 pM] (HIV Proteins) plus cocaine (Coc) [1.6 μM]. Control conditions of incubation with Locke's Buffer vehicle or the HIV proteins or Coc only produced similar neuronal death. (Left Panel) Significant interaction of the HIV proteins with Coc, as illustrated by the lines diverging from parallel, produced toxic synergism (*p < 0.0008). (Right Panel) Significant neuroprotection by 17β-E2 against HIV proteins w/ Coc toxic synergism (#p < 0.0001).
Figure 3 PROG [M] is partially neuroprotective against synergistic neurotoxicity of HIV proteins with cocaine (Coc). Control conditions of incubation with Locke's Buffer vehicle or the HIV proteins or Coc only produced similar neuronal death. (Left Panel) Significant interaction of the HIV proteins with Coc, as illustrated by the lines diverging from parallel, produced toxic synergism (*p < 0.0044). (Right Panel) PROG provided partial neuroprotection against HIV proteins w/ Coc toxic synergism at all concentrations tested (#p < 0.0025) however, the neuroprotection was not concentration dependent.
Figure 4 T [M] provides concentration-dependent neuroprotection against synergistic neurotoxicity of HIV proteins with cocaine (Coc); DHT's [M] neuroprotection is limited. Control conditions of incubation with Locke's Buffer vehicle or the HIV proteins or Coc only produced similar neuronal death. (Top Panel) Significant interaction of the HIV proteins with Coc, as illustrated by the lines diverging from parallel, produced toxic synergism (*p < 0.0087) (Middle Panel). T provides significant concentration-dependent neuroprotection against HIV proteins w/ Coc toxic synergism (#p < 0.0001). (Bottom Panel). DHT mediated partial protection against synergistic neurotoxicity of HIV proteins w/ Coc (#p < 0.023); however, the neuroprotection was not concentration dependent.
Figure 5 ER mediation of T's protection against synergistic neurotoxicity of HIV proteins with cocaine (Coc); cholesterol's [M] neuroprotection is incomplete. Control conditions of incubation with Locke's Buffer vehicle or the HIV proteins or Coc only produced similar neuronal death. (Top Panel) Significant interaction of the HIV proteins with Coc, as illustrated by the lines diverging from parallel, produces toxic synergism (*p < 0.0001. (Middle Panel) The ER specific antagonist, ICI-182,780 (ICI), completely blocked T-mediated neuroprotection. ICI's toxicity and T's neuroprotection did not differ from vehicle control (#p < 0.0001). (Bottom Panel) Cholesterol produced significant, concentration-dependent protection against synergistic neurotoxicity of HIV proteins w/ Coc (#p < 0.0001); however full protection was demonstrated only at 100 nM.
==== Refs
Buchanan RJ Wang S Huang C Analyses of nursing home residents with HIV and dementia using the minimum data set JAIDS 2001 26 246 255 11242197
Dougherty RH Skolasky RLJ McArthur JC Progression of HIV-associated dementia treated with HAART AIDS Read 2002 12 69 74 11905143
Chiesi A Vella S Dally LG Pedersen C Danner S Johnson AM Schwander S Goebel FD Glauser M Antunes F Lundgren JD X Epidemiology of AIDS dementia complex in Europe by the AIDS in Europe study group J Acquir Immune Defic Syndr Hum Retrovirol 1996 11 39 44 8528731
Davies J Everall IP Weich S McLaughlin J Scaravilli F Lantos PL HIV-associated brain pathology in the United Kingdom: an epidemiological study AIDS 1997 11 1145 1150 9233462 10.1097/00002030-199709000-00010
Bell JE Brettle RP Chiswick A Simmonds P HIV encephalitis, proviral load and dementia in drug users and homosexuals with AIDS. Effect of neocortical involvement. Brain 1998 121 2043 2052 9827765 10.1093/brain/121.11.2043
Berger JR Nath A Greenberg RN Andersen AH Greene RA Bognar A Avison MJ Cerebrovascular changes in the basal ganglia with HIV dementia Neurology 2000 54 921 926 10690987
Holman BL Garada B Johnson KA Mendleson J Hallgring E Teoh SK Worth J Navia B A comparison of brain perfusion SPECT in cocaine abuse and AIDS dementia complex J Nucl Med 1992 33 1312 1315 1613571
Nath A Maragos WF Avison MJ Schmitt FA Berger JR Acceleration of HIV dementia with methamphetamine and cocaine J Neurovirol 2001 7 66 71 11519485 10.1080/135502801300069737
Mirsattari SM Power C Nath A Parkinsonism with HIV infection Movement Disorders 1998 13 684 689 9686775 10.1002/mds.870130413
Webber MR Shoenbaum EE Gourevitch MN Buono D Klein RS A prospective study of HIV disease progression in female and male drug users AIDS 1999 13 257 262 10202832 10.1097/00002030-199902040-00014
Bouwman FH Skolasky RL Hes D Selnes OA Glass JD Nance-Sproson TE Royal W Dal Pan GJ McArthur JC Variable progression of HIV-associated dementia. Neurology 1998 50 1814 1820 9633733
Margolin A Avants SK Warburton LA Hawkins KA Factors affecting cognitive functioning in a sample of human immunodeficiency virus-positive injection drug users AIDS Patient Care and STDs 2002 16 255 267 12133261 10.1089/10872910260066697
Holmberg SD The estimated prevalence and incidence of HIV in 96 large US metropolitan areas Am J Public Health 1996 86 642 654 8629714
Farzadegan JH Hoover DR Astemborski J Lyles CM Margolick JB Markham RB Quinn TC Vlahov D Sex differences in HIV-1 viral load and progression to AIDS. Lancet 1998 352 1510 1514 9820299 10.1016/S0140-6736(98)02372-1
Sterling TA Vlahov D Astermborski J Hoover DR Margolick JB Quinn TC Initial plasma HIV-1 RNA levels and progression to AIDS in women and men N Engl J Med 2001 344 720 725 11236775 10.1056/NEJM200103083441003
Squires KE Treating HIV infection and AIDS in women AIDS Read 2003 13 228 240 12800826
Jover T Tanaka H Calderone A Oguro K Bennett MVL Etgen AM Zukin RS Estrogen protects against global ischemia-induced neuronal death and prevents activation of apoptotic signaling cascades in the hippocampal CA1 J Neurosci 2002 22 2115 2124 11896151
Yang SH Liu R Wu SS Simpkins JW The use of estrogens and related compounds in the treatment of damage from cerebral ischemia Ann N Y Acad Sci 2003 1007 101 107 14993044 10.1196/annals.1286.010
Merchenthaler I Dellovade TL Shugrue PJ Neuroprotection by estrogen in animal models of global and focal ischemia Ann N Y Acad Sci 2003 1007 89 100 14993043 10.1196/annals.1286.009
Rau SW Dubal DB Böttner M Gerhold LM Wise PM Estradiol attenuates programmed cell death after stroke-like injury J Neurosci 2003 23 11420 11426 14673006
Sawada H Ibi M Kihara T Urushitani M Honda K Nakanishi M Akaike A Shimohama S Mechanisms of antiapoptotic effects of estrogens in nigral dopaminergic neurons FASEB J 2000 14 1202 1214 10834942
Gajjar TM Anderson LI Dluzen DE Acute effects of estrogen upon methamphetamine induced neurotoxicity of the nigrostriatal dopaminergic system J Neural Transm 2003 110 1215 1224 14628187 10.1007/s00702-003-0045-3
Tang MT Jacobs D Stern Y Marder K Schofield P Gurland B Andrews H Mayeux R Effect of oestrogen during menopause on risk and age at onset of Alzheimer's disease Lancet 1996 348 429 432 8709781 10.1016/S0140-6736(96)03356-9
Kawas C Resnick S Morrison A Brookmeyer R Corrada M Zonderman A Bacal C Lingle DD Metter E A prospective study of estrogen replacement therapy and the risk of developing Alzheimer's disease: the Baltimore Longitudinal Study of Aging Neurology 1997 48 1517 1521 9191758
Manly JJ Merchant CA Jacobs DM Small SA Bell K Ferin M Mayeux R Endogenous estrogen levels and Alzheimer's disease among postmenopausal women Neurology 2000 54 833 837 10690972
Asthana S Baker LD Craft S Stanczyk FZ Veith RC Raskind MA Plymate SR High-dose estradiol improves cognition for women with AD Neurology 2001 57 605 612 11524467
Saunders-Pullman R Gordon-Elliott J Parides M Fahn S Saunders HR Bressman S The effect of estrogen replacement on early Parkinson's disease Neurology 1999 52 1417 1421 10227628
Tsang KL Ho SL Lo SK Estrogen improves motor disability in parkinsonian postmenopausal women with motor fluctuations Neurology 2000 54 2292 2298 10881255
Wooten GF Curie LJ Bovbjerg VE Lee JK Patrie J Are men at greater risk for Parkinson's disease than women? J Neurosurg Psychiatry 2004 75 637 639 10.1136/jnnp.2003.020982
Turchan J Anderson C Hauser KF Sun Q Zhang J Liu Y Wise PM Kruman I Maragos W Mattson MP Booze R Nath A Estrogen protects against the synergistic toxicity by HIV proteins, methamphetamine and cocaine BMC Neurosci 2001 2 3 11252157 10.1186/1471-2202-2-3
Zemlyak I Brooke SM Sapolsky RM Protection against gp120-induced neurotoxicity by an array of estrogenic steroids Brain Res 2002 958 272 276 12470862 10.1016/S0006-8993(02)03558-8
Nath A Haughey NJ Jones M Anderson C Bell JE Geiger JD Synergistic neurotoxicity by human immunodeficiency virus proteins Tat and gp120: protection by memantine Ann Neurol 2000 47 186 194 10665489 10.1002/1531-8249(200002)47:2<186::AID-ANA8>3.0.CO;2-3
Behl C Skutella T Lezoualc'h F Post A Widman M Newton C Newton C Holsboer F Neuroprotection against oxidative stress by estrogens: structure-activity relationship Mol Pharmacol 1997 51 535 541 9106616
Culmsee C Vedder H Ravati A Junker V Otto D Ahlemeyer B Krieg JC Krieglstein J Neuroprotection by estrogens in a mouse model of focal cerebral ischemia and in cultured neurons: Evidence for a receptor-independent antioxidative mechanism J Cereb Blood Flow Metab 1999 19 1263 1269 10566973 10.1097/00004647-199911000-00011
Moosmann B Behl C The antioxidant neuroprotective effects of estrogens and phenolic compounds are independent from their estrogenic properties Proc Natl Acad Sci U S A 1999 96 8867 8872 10430862 10.1073/pnas.96.16.8867
Moss RL Gu J Wong M Estrogen: Nontranscriptional signaling pathway Recent Progress in Hormone Research 1997 The Endocrine Society 33 69 9238847
Woolley CS Effects of estrogen in the CNS Curr Opin Neurbiol 1999 9 349 354 10.1016/S0959-4388(99)80051-8
Green PS Bishop J Simpkins JW 17a-Estradiol exerts neuroprotective effects on SK-N-SH cells J Neurosci 1997 17 511 515 8987774
Hammond J Quynh L Goodyer C Gelfand M Trifiro M LeBlanc A Testosterone-mediated neuroprotection through the androgen receptor in human primary neurons J Neurochem 2001 77 1319 1326 11389183 10.1046/j.1471-4159.2001.00345.x
Golden GA Mason PE Rubin RT Mason RP Biophysical membrane interactions of steroid hormones: A potential complementary mechanism of steroid action Clinical Neuropharmacology 1998 21 181 189 9617510
Jiang N Chopp M Stein D Feit H Progesterone is neuroprotective after transient middle cerebral artery occlusion in male rats Brain Res 1996 735 101 107 8905174 10.1016/0006-8993(96)00605-1
Vongher JM Frye CA Progesterone in conjunction with estradiol has neuroprotective effects in an animal model of neurodegeneration Pharmacol Biochem Behav 1999 64 777 785 10593201 10.1016/S0091-3057(99)00140-9
Ciriza I Axcoitia I Garcia-Segura LM Reduced progesterone metabolites protect rat hippocampal neurones from kainic acid excitotoxicity in vivo J Neuroendocrin 2004 16 58 63 10.1111/j.1365-2826.2004.01121.x
Chernecky CC Berger BJ Chernecky CC and Berger BJ Laboratory Tests and Diagnostic Procedures 2001 3 Philadelphia, WB Saunders
Rupprecht R Friess E Holsboer F Baulieu EE, Robel P and Schumacher M The Neuropsychopharmacological Potential of Neurosteroids Neurosteroids: A new regulatory function in the nervous system Contemporary Endocrinology Conn PM 1999 20 1 Totowa, NJ, Humana Press 349 364
Zwain IH Yen SSC Cheng CY Astrocytes cultured in vitro produce estradiol-17b and express aromatase cytochrome P-450 (P-450 AROM) mRNA Biochim Biophys Acta 1997 1334 338 348 9101730
Hojo Y Hattori T Enami T Furukawa A Suzuki K Ishii H Mukai H Morrison JH Janssen WGM Kominami S Harada N Kimoto T Kawato S Adult male rat hippocampus synthesizes estradiol from pregnenolone by cytochromes P45017a and P450 aromatase localized in neurons Proc Natl Acad Sci U S A 2003 PNAS Early Edition 1 6
Daly RC Su TP Schmidt PJ Pagliaro M Pickar D Ruiz-Lozano P Neuroendocrine and behavioral effects of high-dose anabolic steroid administration in male normal volunteers Psychoneuroendocrinology 2003 28 317 331 12573299 10.1016/S0306-4530(02)00025-2
Labrie F Bélanger A Cusan L Candas B Physiological changes in dehydroepiandrosterone are not reflected by serum levels of active androgens and estrogens but of their metabolites: Intracrinology J Clin Endocrinol Metab 1997 82 2403 2409 9253308 10.1210/jc.82.8.2403
Papadopoulos V Guarneri P Krueger KE Guidotti A Pregnenolone biosynthesis in C6-2B glioma cell mitochondria: Regulation by a mitochondrial diazepam binding inhibitor receptor Proc Natl Acad Sci U S A 1992 89 5113 5117 1317581
Korneyev A Pan BS Plol A Romeo E Guidotti A Costa E Stimulation of brain pregnenolone synthesis by mitochondrial diazepam binding inhibitor receptor ligands in vivo J Neurochem 1993 61 1515 1524 8397297
Stoffel-Wagner B Neurosteroid metabolism in the human brain Eur J Endocrinology 2001 145 669 679 11720889 10.1530/eje.0.1450669
Weill-Engerer S David JP Sazdovitch V Liere P Schumacher M Delacourte A Baulieu EE Akwa Y In vitro metabolism of dehydroepiandrosterone (DHEA) to 7a-hydroxy-DHEA and D5-androstene-3b,17b-diol in specific regions of the aging brain from Alzheimer's and non-demented patients Brain Res 2003 969 117 125 12676372 10.1016/S0006-8993(03)02288-1
Zwain IH Yen SSC Dehydroepiandrosterone: Biosynthesis and metabolism in the brain Endocrinology 1999 140 880 887 9927319 10.1210/en.140.2.880
McCullough LD Blizzard K Simpson ER Ohran K Hurn PD Aromatase cytochrome P450 and extragonadal estrogen play a role in ischemic neuroprotection J Neurosci 2003 23 8701 8705 14507969
Maccarrone M Navarra M Catani V Corasaniti MT Bagetta G Finazazi-Agrò A Cholesterol-dependent modulation of the toxicity of HIV-1 coat protein gp120 in human neuroblastoma cells J Neurochem 2002 82 1444 1452 12354292 10.1046/j.1471-4159.2002.01072.x
Bonavia R Bajetto A Barbero S Albini A Noonan DM Schettini G HIV-1 Tat causes apoptotic death and calcium homeostasis alterations in rat neurons Biochem Res Commun 2001 288 301 308 10.1006/bbrc.2001.5743
Song L Nath A Geiger JD Moore A Hochman S Human immunodeficiency virus type 1 Tat protein directly activates neuronal N-methyl-D-aspartate receptors at an allosteric zinc-sensitive site J Neurovirol 2003 9 399 403 12775422
Bagetta G Piccirilli S Del Duca C Morrone LA Rombolà L Nappi G De Alba J Knowles RG Corasaniti MT Inducible nitric oxide synthase is involved in the mechanisms of cocaine enhanced neuronal apoptosis induced by HIV-1 gp120 in the neocortex of rat Neurosci Lett 2004 356 183 186 15036625 10.1016/j.neulet.2003.11.065
Aksenova M Aksenov MY Nath A Mactutus C Booze R Cocaine enhances oxidative stress-dependent toxicity of HIV-1 Tat in rat hippocampal cell cultures Society for Neuroscience 2004 573.8
Nassogne MC Louahed J Evrard P Courtoy PJ Cocaine induced apoptosis in cortical neurons in fetal mice. J Neurochem 1997 68 2442 2450 9166738
Singer C Rogers K Dorsa D Modulation of Bcl-2 expression: A potential component of estrogen protection in NT2 neurons. Neuroreport 1998 9 2565 2568 9721933
Pike DJ Estrogen modulates neuronal Bcl-xl expression and beta-amyloid induced apoptosis relevance to Alzheimer's disease. J Neurochem 1999 72 1552 1563 10098861 10.1046/j.1471-4159.1999.721552.x
Garnier M DiLorenzo D Albertini A Maggi A Identification of estrogen-responsive genes in neuroblastoma SK-ER3 cells. J Neurosci 1997 17 4591 4599 9169520
Magnuson DS Knudsen BE Geiger JD Brownstone RM Nath A Human immunodeficiency virus type 1 tat activates non-N-methyl-D-aspartate excitatory amino acid receptors and causes neurotoxicity Ann Neurol 1995 37 373 380 7695237 10.1002/ana.410370314
Nath A Psooy K Martin C Knudsen B Magnuson DS Haughey N Geiger JD Identification of a human immunodeficiency virus type 1 Tat epitope that is neuroexcitatory and neurotoxic J Virol 1996 70 1475 1480 8627665
New DR Ma M Epstein LG Nath A Gelbard HA Human immunodeficiency virus type 1 Tat protein induces death by apoptosis in primary human neuron cultures J Neurovirol 1997 3 168 173 9111179
Gelbard HA Nottet HS Swindells S Jett M Dzenko KA Genis P White R Wang L Choi YB Zhang D Lipton SA Tourtellotte WW Epstein LG Gendelman HE Platelet-activating factor: a candidate human immunodeficiency virus type 1-induced neurotoxin J Virol 1994 68 4628 4635 8207837
New DR Maggirwar SB Epstein LG Dewhurst S Gelbard HA HIV-1 Tat induces neuronal death via tumor necrosis factor-a and activation of non-N-methyl-D-aspartate receptors by a NFkB-independent mechanism The Journal of Biological Chemistry 1998 273 17852 17858 9651389 10.1074/jbc.273.28.17852
Ma M Nath A Molecular determinants for cellular uptake of Tat protein of human immunodeficiency virus type 1 in brain cells J Virol 1997 71 2495 2499 9032389
Gu J Yassini P Goldberg G Shu W Konat GW Wiggins RC Cocaine cytotoxicity in serum-free environment: C6 glioma cell culture Neurotoxicology 1993 14 19 22 8361674
Nassogne MC Evrard P Courtoy PJ Selective neuronal toxicity of cocaine in embryonic mouse brain cocultures Proc Natl Acad Sci U S A 1995 92 11029 11033 7479930
Bennet BA Hyde CE Pecora JR Clodfelter JE Long-term cocaine administration is not neurotoxic to cultured fetal mesencephalic dopamine neurons Neurosci Lett 2003 153 210 214 10.1016/0304-3940(93)90324-E
Robinson SE Enters EK Jackson GF Chinchilli VM Maher JR McDowell KP Allen HM Guo H Maternal and fetal brain and plasma levels of cocaine and benzoylecgonine after acute or chronic maternal intravenous administration of cocaine J Pharmacol Exp Ther 1994 271 1234 1239 7996432
Mactutus CF Herman AS Booze RM Chronic intravenous model for studies of drug (Ab)use in the pregnant and/or group-housed rat: an initial study with cocaine Neurotoxicology and Teratology 1994 16 183 191 8052193 10.1016/0892-0362(94)90116-3
Evans SM Cone EJ Henningfield JE Arterial and venous cocaine plasma concentration in humans: Relationship to route of administration, cardiovascular effects and subjective effects J Pharmacol Exp Ther 1996 279 1345 1356 8968359
Chiu FC Rozental R Bassallo C Lyman WD Spray DC Human fetal neurons in culture: intercellular communication and voltage- and ligand-gated responses J Neurosci Res 1994 38 687 697 7807584 10.1002/jnr.490380611
Sanfeliu C Sebastiá J Ki SU Methylmercury neurotoxicity in cultures of human neurons, astrocytes, neuroblastoma cells Neurotoxicology 2001 22 317 327 11456333 10.1016/S0161-813X(01)00015-8
Messam CA Hou J Berman JW Major EO Analysis of the temporal expression of nestin in human fetal brain derived neuronal and glial progenitor cells Developmental Brain Research 2002 134 87 92 11947939 10.1016/S0165-3806(01)00325-X
Fried G Andersson E Csöregh L Enmark E Gustafsson JA Aanesen A Österlund C Estrogen receptor beta is expressed in human embryonic brain cells and is regulated by 17b-estradiol Eur J Neurosci 2004 20 2345 2354 15525276 10.1111/j.1460-9568.2004.03693.x
Winter BJ Statistical Principles in Experimental Design 1971 2 New York, McGraw Hill
| 15943860 | PMC1177959 | CC BY | 2021-01-04 16:03:47 | no | BMC Neurosci. 2005 Jun 8; 6:40 | utf-8 | BMC Neurosci | 2,005 | 10.1186/1471-2202-6-40 | oa_comm |
==== Front
BMC OphthalmolBMC Ophthalmology1471-2415BioMed Central London 1471-2415-5-101594388910.1186/1471-2415-5-10Case ReportVitrectomy with complete posterior hyaloid removal for ischemic central retinal vein occlusion: Series of cases Leizaola-Fernández Carlos [email protected]árez-Tatá Luis [email protected] Hugo [email protected] Juner [email protected] J [email protected]énez-Sierra Juan M [email protected] Jose L [email protected]ón Virgilio [email protected] Department of Retina. Hospital "Dr. Luis Sánchez Bulnes" Asociación Para Evitar la Ceguera en México (APEC), Distrito Federal, México2 Department of Electrophisiology. Hospital "Dr. Luis Sánchez Bulnes" Asociación Para Evitar la Ceguera en México (APEC), Distrito Federal, México2005 20 5 2005 5 10 10 25 10 2004 20 5 2005 Copyright © 2005 Leizaola-Fernández 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
Central retinal vein occlusion (CRVO) is a common retinal vascular disorder with potentially complications: (1) persistent macular edema and (2) neovascular glaucoma. No safe treatment exists that promotes the return of lost vision. Eyes with CRVO may be predisposed to vitreous degeneration. It has been suggested that if the vitreous remains attached to the macula owing to a firm vitreomacular adhesion, the resultant vitreous traction can cause inflammation with retinal capillary dilation, leakage and subsequent edema6. The roll of vitrectomy in ischemic CRVO surgical procedures has not been evaluated.
Case presentation
This is a non comparative, prospective, longitudinal, experimental and descriptive series of cases. Ten eyes with ischemic CRVO. Vitrectomy with complete posterior hyaloid removal was performed. VA, rubeosis, intraocular pressure (IOP), and macular edema were evaluated clinically. Multifocal ERG (m-ERG), fluorescein angiography (FAG) and optic coherence tomography (OCT) were performed. Follow-up was at least 6 months. Moderate improvement of visual acuity was observed in 60% eyes and stabilized in 40%. IOP changed from 15.7 ± 3.05 mmHg to 14.9 ± 2.69 mmHg post-operative and macular edema from 976 ± 196 μm to 640 ± 191 μm to six month. The P1 wave amplitude changed from 25.46 ± 12.4 mV to 20.54 ± 11.2 mV.
Conclusion
A solo PPV with posterior hyaloid removal may help to improve anatomic and functional retina conditions in some cases. These results should be considered when analyzing other surgical maneuvers.
==== Body
Background
Central retinal vein occlusion (CRVO) is a common retinal vascular disorder with potentially complications like reduced vision resulting from extensive intraretinal hemorrhage, retinal ischemia and persistent macular edema and neovascular glaucoma secondary to iris neovascularization. Severe visual loss may develop, including blindness. Capillary nonperfusion and retinal ischemia develop in 34% of patients with CRVO. Iris neovascularization and neovascular glaucoma may occur in 45–85% of ischemic CRVO [15,20,23,24]. No safe treatment exists that promotes the return of lost vision. Treatments that target the secondary effects of venous occlusion, such as grid laser photocoagulation for macular edema and prophylactic panretinal laser photocoagulation for nonperfused CRVO, were shown to be ineffective in improving visual acuity in the Central Vein Occlusion Study (CVOS) [21,22]. Attempts to bypass venous obstruction in CRVO using laser chorioretinal anastomosis have been performed with limited success and significant complications [3,11]. Radial optic neurotomy has been shown to be beneficial for the treatment of CRVO [13]. Most patients who develop CRVO (50 – 70%) have associated hypertension, cardiovascular disease, or diabetes mellitus [5]. The treatment of these systemic conditions is not effective in the management of the ocular complications. It has been postulated that vitreous traction may play a role in the pathogenesis of cystoid macular edema [16,17]. Eyes with CRVO may be predisposed to vitreous degeneration. It has been suggested that if the vitreous remains attached to the macula owing to a firm vitreomacular adhesion, the resultant vitreous traction can cause inflammation with retinal capillary dilation, leakage and subsequent edema [16]. Pars plana vitrectomy is believed to eliminate this mechanical traction and increase oxygenation to the retina and to reduce the risk of macular edema and neovascularization. The purpose of this study was to perform an assessment of the safety and efficacy of pars plana vitrectomy combined with complete posterior hyaloid removal in ischemic CRVO.
Case presentation
Non comparative, prospective, longitudinal, experimental and descriptive series of cases of patients with ischemic CRVO referred to Hospital Luis Sanchez Bulnes, Asociación para Evitar la Ceguera en México, México City. The study was approved by Hospital Review Committee The inclusion criteria were: confirmed presence of central retinal vein occlusion, less than 6 months evolution with visual acuity (VA) of <20/100 or those whose VA decreased more than 50% during the follow-up, an over 10 disc diameter area of nonperfused retina, intraocular pressure ≤ 21 mmHg, age ≥ 21 years old, afferent pupillary defect, ability to obtain good quality fundus photographs and angiograms and absence of neovascularization. The exclusion criteria in the study eye were: intercurrent eye disease that is likely to affect VA over study period, presence of any diabetic retinopathy, other retinal vascular disease, vitreous hemorrhage, presence of neovascularization (iris, angle, retina and disk) or previous laser treatment. Best corrected ETDRS visual acuity (BCVA), relative afferent pupillary defect (RAPD), slit-lamp examination, indirect ophthalmoscopy, fundus photography, fluorescein angiography (FA), multifocal electroretino*graphy (mERG) and optic coherence tomography (OCT) were performed preoperatively and at 1, 3, and 6 months postoperatively. Follow-up was at least 6 months.
Surgical procedure
After informed consent was obtained and a full explanation of the risks and benefits and the experimental nature of this procedure were conveyed to the patient (in accordance with the Helsinki declaration), patients underwent a standard three port pars plana vitrectomy was performed, during which the posterior hyaloid was removed using active aspiration. The sclerostomy sites and conjunctiva were closed in the usual fashion. No panretinal photocoagulation and gas tamponade were used.
OCT and mERG
Macular thickness was analyzed with OCT (Humphrey 1000; Humphrey Instruments, San Leandro, California). Measurement of central macular thickness was evaluated according to Hee protocol [7], which in each eye, six consecutive tomographic cuts were obtained in a radial pattern, surrounding the fovea at similar angular distances among them. Each one of these cuts was oriented along a line that intersects the foveal center. The central foveal thickness was registered at the intersection of the 6 cuts.
The electrical function of the macular area was determined by multifocal electroretinography (mERG). The RETI-scan™ multifocal system, produced by Roland Consult, was used for this purpose. The stimulation and recording of the mERG were performed using the m-sequence technique. Contact lens ERG-JET electrodes as well as one ground electrode in the center of forehead and two reference electrodes in the temporal region of the patient were placed. The stimulus, consisting of 61 hexagons covering a visual field of 30°, was presented on a monitor (ELSA 20"-VGA-monitor) with a frame rate of 75 Hz at a distance of 28 cm from the patient's eye. Each element alternated between black and white (93% contrast, mean luminance 51.8 cd/m2). The patient was instructed to maintain fixation on longitudinal axes intersecting one focal point. The amplifier setting was 100 μV; the lower cut off frequency was 10 Hz and the upper cut off frequency was 100 Hz. No notch filter was used. Each recording session was subdivided into 8 recording segments of approximately 47 seconds. The signals were registered with sampling intervals of 83 mseg. The results were distributed in 5 concentric rings obtaining for each ring, the N1 amplitude, P1 amplitude, and the implicit times of N1 and P1. The amplitude of the positive component P1 was measured, and the response density (nV/deg2) was calculated by dividing the response amplitude (nV) by the retinal area (deg2).
Results
The study data are displayed in Table 1 and 2. Patient's age ranged from 57 to 74 years, with a mean of 65.13 years. There were 4 females and 6 males. All patients were Hispanic. Initial treatment was attempted between January 2002 and February 2003. The time interval range from diagnosis to treatment was 37 to 270 days, with a mean of 103.13 days. In all patients, the post treatment follow-up examination was at least 6 months. Four patients had diabetes mellitus type 2 and six patients had systemic arterial hypertension.
The median preoperative BCVA was 20/600 (range, 20/1600 to 20/300). The median postoperative BCVA was 20/300 (range, 20/1600 to 20/100). VA was unchanged in 4 of 10 eyes (40%) and improved in 6 of 10 eyes (60%). None eye had no light perception. The median preoperative PIO was 15.7 ± 3.05 mmHg and the median postoperative PIO was 14.9 ± 2.68 mmHg. Central macular edema registered by OCT changed from 976 ± 196 μm to 640 ± 191 μm to six month. The P1 wave amplitude of the central ring varied from 25.46 ± 12.4 mV to 20.54 ± 11.2 mV to final follow up.
None eye suffered retinal o disc neovascular membrane. The iris neovascularization was presented in 3 eyes (30%). All these eyes were treated with extensive photocoagulation. None of these eyes suffered neovascular glaucoma. Eight eyes presented middle opacity of lens.
Discussion
Fine and Brucker [4] and Wallow and 3 colleagues [25] demonstrated that cystoid macular edema began with ischemia and intracytoplasmic swelling of Müller cells, and speculated that an intrinsic metabolic agent could cause these changes. The vitreous is believed to store pharmacologic agents that may cause macular edema. Hikichi and 2 colleagues[8] proposed that in eyes with vitreous macular attachment, the macular edema is persistent because those metabolic agents remain in contact with the macula. Also they observed that a complete posterior vitreous detachment seems to protect against retinal or optic disc neovascularization. Another theory asserts that because vitreous fibres extending from the vitreous base to the macula and Müller cell attachment to the internal limiting lamina are prominent in this region [17,20] centripetal traction caused by vitreous contraction after CRVO and transmitted to Müller cells by the vitreous fibres attached to the macula may cause swelling of these cells. Kado and 6 colleagues [9] showed that macular edema lasted longer in eyes with vitreous macular attachment than in those whose vitreous was detachment from the macula, due to centripetal traction transmitted to the Müller cells by vitreous fibres inserted into the macula. Based on these facts, we proposed that a vitrectomy could eliminate the metabolic agents present in the vitreous as well as the mechanical traction, and in this way might protect against retinal and optic disc neovascularization. Also, pars plana vitrectomy may beneficial effects upon retinal ischemic by allowing circulation in the vitreous cavity of fluid oxygenated by unaffected retina or others sites in the eye such as the ciliary body [10,19]. In our patients, a standard three port pars plana vitrectomy was performed followed by creation of posterior hyaloid separation and removal of the posterior cortical vitreous. The central macular edema registered by OCT changed from 976 ± 196 μm pre operative to 640 ± 191 μm to six month. All patients presented persistent cystoid macular edema to six month.
The level of visual improvement seems relatively good when compared with the results of conventional treatment such as retinal photocoagulation, or only observation. Although 40% of eyes maintained stable visual acuity and 60% of eyes had visual improvement, there was still considerable long-term visual morbidity in this series. Only two eyes had final VA better than 20/200 (20/100 in both eyes). The rest of eyes had a final VA of 20/200 or less. Long-term visual morbidity may be corresponding to residual perfused cystoid macular edema or to degenerative squealer of previous chronic edema. None patient presented significant complication post treatment such as neovascular glaucoma or retinal detachment. Most patients with CRVO lose their visual acuity permanently because they develop macular edema or ischemia. The degree of retinal hypoxia and infarction and probably the extent of abnormal vascular permeability may be are reflected in electroretinogram (ERG) amplitude and timing changes. In our patients, the P1 wave amplitude changed from 25.46 ± 12.4 mV pre operative to 20.54 ± 11.2 mV to six month. Once the VA of these patients is severely reduced, 80% of them cannot expect any improvement spontaneously [2,14,23]. In this study, the pre treatment visual acuity in all eyes ranged from 20/800 to 20/300 (20/300). It is recognized that earlier intervention probably is favourable to later to prevent macular scarring from longstanding edema [12,18]. In our series, in eyes with early treatment (≤90 days), 4 eyes (67%) improved VA and 2 eyes (33%) were unchanged. In eyes with later treatment (>90 days), 2 eyes (50%) improved VA and 2 eyes (50%) were unchanged.
In severe cases of ischemic CRVO, the role of the vitreous was not observed. Increased venous pressure, which leads to damage of the capillary endothelium and breakdown of the blood-retinal barrier, was reported to be the main cause of macular edema in CRVO [6,25]. Because capillary damage may be aggravated sufficiently to negate the effect of complete posterior vitreous detachment on macular edema resolution in severe cases, the vitreoretinal relationship can only influence the prognosis of macular conditions when there is mild capillary damage in the macular region.
In the ischemic type of CRVO, the reported prevalence of posterior segment and iris neovascularization ranges from 0% to 55% and from 16% to 67%, respectively [15,20,23,24]. In our series, none eyes presented retinal or optic neovascularization membrane. The iris neovascularization was presented in 3 patients (30%). The posterior vitreous detachment seems to play a protective role against retinal or optic disc neovascularization in central retinal vein occlusion [1,8], although the vitreoretinal relationship may not influence the development of iris neovascularization.
The efficacy and safety of vitrectomy in these patients is still in question. Further studies such as a randomized controlled clinical trial are needed to determine whether vitrectomy is of benefit in ischaemic CRVO. Our results suggest that a solo PPV with posterior hyaloid removal may help to improve anatomic and functional retina conditions in some cases. These results should be considered when analyzing other surgical maneuvers.
Conclusion
A solo PPV with posterior hyaloid removal may help to improve anatomic and functional retina conditions in some cases. These results should be considered when analyzing other surgical maneuvers.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
CLF: participated in the sequence alignment and drafted the manuscript
LST: participated in the sequence alignment and drafted the manuscript
HQM: participated in the sequence alignment
JCL: participated in the sequence alignment, sending the manuscript and doing the corrections
JFG: participated in the design of the study and performed the statistical analysis
JMJS: participated in doing and analysis of electrophysiological studies
JLGN participated in doing all surgical procedures
VMC: participated in its design and coordination and helped to draft the manuscript
Pre-publication history
The pre-publication history for this paper can be accessed here:
Figures and Tables
Table 1 Patients Characteristics
Case Age (years) Sex HBP DM Duration (days)* VA initial VA 3 month VA final Follow-Up (month)
1 57 M N N 120 20/1600 20/1600 20/1600 12
2 66 F N Y 66 20/1600 20/1600 20/800 10
3 72 M N Y 75 20/1600 20/1600 20/1600 8
4 70 M Y N 80 20/1600 20/1600 20/1600 8
5 59 M N Y 37 20/300 20/100 20/100 7
6 74 F Y Y 98 20/800 20/100 20/100 7
7 65 M Y N 65 20/300 20/300 20/200 6
8 58 M N N 80 20/1600 20/800 20/600 6
9 64 F Y N 180 20/1600 20/1600 20/1600 6
10 60 F Y N 90 20/600 20/300 20/300 6
HBP: high blood pressure; DM: diabetes mellitus; VA: visual acuity; IOP: intraocular pressure; M: masculine; F: feminine; Y: yes; N: no
* Time interval range from diagnosis to treatment
Table 2 IOP, macular thickness and mERG values
IOP mmHg Macular thickness (μm) OCT P1 wave amplitude (mV) mERG
Case Preop Postop Preop Postop Preop Postop
1 12 14 1100 800 44.2 16.4
2 16 18 1176 800 13.8 12.3
3 12 11 1200 750 7.0 30.2
4 18 16 1000 375 23.0 6.1
5 14 18 920 477 38.0 40.0
6 12 10 780 320 9.0 24.9
7 15 16 1000 700 23.8 14.7
8 20 18 1300 750 17.7 6.01
9 18 14 730 650 36.7 29.2
10 18 16 850 700 25.0 30.5
IOP: intraocular pressure; OCT: optic coherence tomography; mERG: multifocal electroretinogram; Preop: preoperative; Postop: postoperative
==== Refs
Akiba J Kado M Kakehashi A Trempe C Role of the vitreous in posterior segment neovascularization in central retinal vein occlusion Ophthalmic Surg 1991 22 498 502 1945273
Chen JC Klein ML Watzke RC Handelman IL Robertson JE Natural course of perfused central retinal vein occlusion Can J Ophthalmol 1995 30 21 24 7780892
Fekrat S Goldberg MF Finkelstein D Laser-induced chorioretinal venous anastomosis for nonischemic central or branch retinal vein occlusion Arch Ophthalmol 1998 116 43 52 9445207
Fine BS Brucker AJ Macular edema and cystoid macular edema Am J Ophthalmol 1981 92 466 481 7294109
Gutman FA Evaluation of a patient with central retinal vein occlusion Ophthalmology 1983 90 481 483 6877780
Hayrech SS Classification of central retinal vein occlusion Ophthalmology 1983 90 458 474 6877778
Hee MR Puliafito CA Duker JS Reichel E Coker JG Wilkins JR Schuman JS Swanson EA Fujimoto JG Topography of diabetic macular edema with optical coherence tomography Ophthalmology 1998 105 360 370 9479300 10.1016/S0161-6420(98)93601-6
Hikichi T Konno S Trempe CL Role of the vitreous in central retinal vein occlusion Retina 1995 15 29 33 7754244
Kado M Jalkh AE Yoshida A Takahashi M Wazen N Trempre CL Schepens CL vitreous changes and macular edema in central retinal vein occlusion Ophthalmic Surg 1990 21 544 549 2234801
Maeda N Tano Y Intraocular oxygen tension in eyes with proliferative diabetic retinopathy with and without vitreous Graefes Arch Clin Exp Ophthalmol 1996 234 S66 S69 8871152
McAllister IL Constable IJ Laser-induced chorioretinal venous anastomosis for treatment of nonischemic central retinal vein occlusion Arch Ophthalmol 1995 113 456 462 7710396
McAllister IL Douglas JP Constable IJ Yu DY Laser-induced chorioretinal venous anastomosis for nonischemic central vein occlusion: evaluation of the complications and their risk factors Am J Ophthalmol 1998 126 219 229 9727516 10.1016/S0002-9394(98)00156-1
Opremcak EM Bruce RA Lomeo MD Rindenour CD Lestón AD Rehmar AJ Radial optic neurotomy for central retinal vein occlusion. A retrospective Pilot Study of 11 Consecutive Cases Retina 2001 21 408 415 11642369 10.1097/00006982-200110000-00002
Quinlan PM Elman MJ Bhatt AK Mardesich P Enger C The natural course of central retinal vein occlusion Am J Ophthalmol 1990 110 118 123 2378375
Rath EZ Frank RN Shin DH Kim C Risk factors for retinal vein occlusion: a case controlled study Ophthalmology 1992 99 509 514 1584567
Schepens CL Avila MP Jalkh AE Trempe CL Role of the vitreous in cystoid macular edema Surv Ophthalmol 1984 28 499 504 6463849 10.1016/0039-6257(84)90232-7
Sebag J Balazs EA Pathogenesis of cystoid macular edema: an anatomic consideration of vitreoretinal adhesions Surv Ophthalmol 1984 28 493 498 6379948 10.1016/0039-6257(84)90231-5
Sebag J The vitreous: structure, function, and pathobiology 1989 133 New York, NY Springer-Verlag 35 58
Stefansson E Novack RI Hatchell DL Vitrectomy prevents retinal hypoxia in branch retinal vein occlusion Invest Ophthalmol Vis Sci 1990 31 284 289 2303330
The Central Vein Occlusion Study Group Baseline and early natural history report: The Central Vein Occlusion Study Arch Ophthalmol 1993 111 1087 1095 7688950
The Central Vein Occlusion Study Group Evaluation of grid pattern photocoagulation for macular edema in central vein occlusion: the Central Vein Occlusion Study Group M report Ophthalmology 1995 102 1425 1433 9097788
The Central Vein Occlusion Study Group A randomized clinical trial of early panretinal photocoagulation for ischemic central vein occlusion: the Central Vein Occlusion Study Group N report Ophthalmology 1995 102 1434 1444 9097789
The Central Vein Occlusion Study Group Natural history and clinical management of central retinal vein occlusion Arch Ophthalmol 1997 115 486 491 9109757
The Eye Disease Case-Control Study Group Risk factors for central retinal vein occlusion Arch Ophthalmol 1996 114 545 554 8619763
Wallow IHL Danis RP Bindley C Neider M Cystoid degeneration in experimental branch vein occlusion Ophthalmology 1998 95 1371 1379 3226685
| 15943889 | PMC1177960 | CC BY | 2021-01-04 16:03:49 | no | BMC Ophthalmol. 2005 May 20; 5:10 | utf-8 | BMC Ophthalmol | 2,005 | 10.1186/1471-2415-5-10 | oa_comm |
==== Front
BMC PediatrBMC Pediatrics1471-2431BioMed Central London 1471-2431-5-151594386910.1186/1471-2431-5-15Research ArticleMotives for choosing growth-enhancing hormone treatment in adolescents with idiopathic short stature: a questionnaire and structured interview study Visser – van Balen Hanneke [email protected] Rinie [email protected] Gerdine A [email protected] Jaap [email protected] Jan M [email protected] Gerben [email protected] Department of Pediatric Psychology, University Medical Center Utrecht, Wilhelmina Children's Hospital, P.O. Box 85090, 3508 AB Utrecht, The Netherlands2 Department of Developmental Psychology, Utrecht University, P.O. Box 80140, 3508 TC Utrecht, The Netherlands3 Department of Health Psychology, Utrecht University, P.O. Box 80140, 3508 TC Utrecht, The Netherlands4 Department of Pediatrics, Gooi-Noord Hospital, P.O. Box 900, 1250 CA Laren, The Netherlands5 Department of Medical Psychology, VU University Medical Center, P.O. Box 7057, 1007 MB Amsterdam, The Netherlands6 Department of Pediatrics, Leiden University Medical Center, P.O. Box 9600, 2300 RC Leiden, The Netherlands2005 8 6 2005 5 15 15 17 12 2004 8 6 2005 Copyright © 2005 Visser – van Balen et al; licensee BioMed Central Ltd.2005Visser – van Balen 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
Growth-enhancing hormone treatment is considered a possible intervention in short but otherwise healthy adolescents. Although height gain is an obvious measure for evaluating hormone treatment, this may not be the ultimate goal for the person, but rather a means to reach other goals such as the amelioration of current height-related psychosocial problems or the enhancement of future prospects in life and society. The aim of our study was to clarify the motives of adolescents and their parents when choosing to participate in a growth-enhancing trial combining growth hormone and puberty-delaying hormone treatment.
Methods
Participants were early pubertal adolescents (25 girls, 13 boys) aged from 11 to 13 years (mean age 11.5 years) with a height standard deviation score (SDS) ranging from -1.03 to -3.43. All had been classified as idiopathic short stature or persistent short stature born small for the gestational age (intrauterine growth retardation) on the basis of a height SDS below -2, or had a height SDS between -1 and -2 and a predicted adult height SDS below -2. The adolescents and their parents completed questionnaires and a structured interview on the presence of height-related stressors, parental worries about their child's behavior and future prospects, problems in psychosocial functioning, and treatment expectations. Questionnaire scores were compared to norms of the general Dutch population.
Results
The adolescents reported normal psychosocial functioning and highly positive expectations of the treatment in terms of height gain, whereas the parents reported that their children encountered some behavioral problems (being anxious/depressed, and social and attention problems) and height-related stressors (being teased and juvenilized). About 40% of the parents were worried about their children's future prospects for finding a spouse or job. The motives of the adolescents and their parents exhibited rather different profiles. The most prevalent parental worries related to the current or future functioning of their children, while a few cases were characterized by no observed motives or by psychosocial problems only reported by the adolescents themselves.
Conclusion
The motives for participating in a growth-enhancing hormone trial are more obvious in the parents than in the adolescents themselves. Two out of three parents report worries about the future opportunities or observe modest current psychosocial problems in their children. The adolescents want to gain height, but the motivation underlying this remains unclear. Few of the adolescents experience psychosocial problems. Our analyses revealed differences among individuals in terms of motives, which implies that in an evaluation of hormone treatment, the importance of divergent outcome variables will also differ among individuals. Effectiveness evaluations of hormone treatment to increase height and the consequential fulfillment of other goals must be awaited.
==== Body
Background
Growth-enhancing hormone treatment is considered a possible intervention in short but otherwise healthy adolescents. Although height gain is an obvious measure for evaluating hormone treatment, this may not be the ultimate goal of the person, but rather a means to reach other goals such as to ameliorate current height-related psychosocial problems or to enhance future prospects in life and society.
Psychosocial functioning has been emphasized in considerations of the motives of children with idiopathic short stature (ISS) and their parents seeking hormone treatment to achieve a height gain. Children who have never been medically referred do not seem to suffer from their stature [1-4]. Most studies involving medically referred children have revealed reduced social competence [5-14], while internalizing and externalizing behavior, body image, self-esteem, and scholastic competence have been shown, on average, to be between normal and below normal [5-8,10-17]. In addition to the possibility of improving psychosocial functioning, several other factors are hypothesized to motivate children and their parents to choose hormone treatment, including the experience of height-related psychosocial stressors (such as being teased or juvenilized), future expectations for finding a spouse and a desired career commensurate with intellectual potential and interests, and the expectation of a treatment-induced height gain.
Insight into the divergent motives at the time of choosing hormone treatment is important, because it will help in the choice of proper outcome measures for the evaluation of a trial. Insight into motivational factors is especially important in treatments combining biosynthetic growth hormone (GH) with a puberty-delaying hormone, because the possible benefits of enhancing growth must be balanced against the possible negative psychosocial consequences of delaying pubertal development. Moreover, such insight will clarify whether other treatments, such as psychosocial counseling, should be employed to solve current psychosocial problems and reach future goals.
Our study examined 38 medically referred early pubertal, short adolescents who agreed to participate in an experimental trial of a combined treatment of GH and a puberty-delaying hormone, and compared the results with short adolescents who received no intervention. Participants had a height standard deviation score (SDS) below -2, or between -1 and -2 with a predicted adult height SDS below -2, all without apparent underlying pathology. Twenty-six of the subjects had a normal birth weight and length and were classified as ISS [18], while 12 were born small for gestational age (SGA) [19,20]. Both ISS and SGA have been associated with problems in behavior, social functioning, school competence, and attention [21-23], while SGA has also been associated with lowered intelligence and minor neurologic dysfunctions [21,22,24].
With the aim of clarifying the motives of parents and children for participation in the trial, we examined the baseline scores of the adolescents on the following psychosocial factors before initiating the treatment: presence of height-related stressors, parental worries about the child's behavior and future prospects, self-reported problems in psychosocial functioning, and treatment expectations.
Methods
Study population and procedures
The inclusion criteria for participation in the study were as follows: being in early puberty as documented by Tanner breast stage 2 or 3 for girls and Tanner genital stage 2 or 3 for boys; an actual height SDS for the same age and sex of less than -2.0, or a height SDS between -1.0 and -2.0 with a predicted adult height of more than -2.0 SDS; a chronological age and bone age of less than 12 years for girls and 13 years for boys; a documented GH response of >20 mU/L (>10 ug/L) after a standard provocation test and/or during a sleep test; a ratio of sitting height/subischial leg height between the 3rd and 97th percentiles; and normal blood tests and urinalysis.
The study population consisted of 38 early pubertal adolescents (25 girls,13 boys) aged from 11 to 13 years (mean age 11.5 years). Most of the cohort (17 girls, 9 boys) had a normal birth size and were classified as ISS. Twelve adolescents (8 girls, 4 boys) were known to have had a birth length of more than 2 SD below the mean for gestational age and were classified as short-stature-born SGA. The most likely reason for the girls substantially outnumbering the boys is that the combination of short stature and early puberty is more common in girls than boys. The mean height of the adolescents was 134.9 (5.8) cm, with a range of 120.0–148.5 cm, corresponding to a height SDS ranging from -1.03 to -3.43. There were nine adolescents with an actual height SDS between -1.0 and -2.0 and a predicted adult height SDS of less than -2.0.
The following four Dutch hospitals participated in this multi-center study: the University Medical Center Utrecht-Wilhelmina Children's Hospital, the VU University Medical Center Amsterdam, the Sophia Children's Hospital Rotterdam, and the Catharina Hospital Eindhoven [19]. The first 40 referred adolescents who fulfilled the inclusion criteria and agreed to participate after having been informed about the study participated in our randomized controlled trial with puberty-delaying gonadotropin-releasing hormone agonist (GnRHa) therapy in combination with GH therapy or no therapeutic intervention. Twenty of the adolescents were randomly selected to receive the combination therapy for three years, while the other 20 adolescents were to receive no treatment. Immediately after randomization, two participants (one from the treatment group and one from the control group) decided to stop participation.
The experimental trial examined the hypothesis that the combination of GH and GnRHa for three years would increase the final height in comparison with untreated controls [19]. For three years the adolescents in the treatment group were given GH daily by subcutaneous injection of 4 IU (1.33 mg) per square meter body surface; (Genotropin, Pharmacia, Sweden; now Pfizer, New York, USA) and a depot preparation of 3.75 mg of GnRHa (Decapeptyl-CR, Ferring, Sweden) every four weeks intramuscularly. The adolescents in the control group were followed yearly to document their growth and pubertal development [19]. Before treatment, and at one, two, and three years after beginning the treatment, the adolescents and the parents of both the treatment and control groups filled out questionnaires to assess the psychosocial functioning of the adolescents. The parents were also interviewed.
A skilled psychologist examined the adolescents and their parents in the interval between randomization and starting treatment. This interval was chosen for ethical reasons to avoid the suggestion that the answers of the parents or adolescents might affect the group allocation. Moreover, providing knowledge of the result of the randomization procedure to both parents and adolescents was assumed to make them better able to give unbiased responses to the questionnaires and interview. The adolescents filled out questionnaires on perceived competence, psychological distress, body image, self-image, and personality characteristics. The parents filled out questionnaires regarding emotional and behavioral problems of their children, and were interviewed by a psychologist.
The medical ethics committees at the four participating centers approved the study. The parents of all children provided written informed consent before participation.
Measures
Parental reports
A structured interview was conducted with parents to assess the adolescents' health-related development, current height-related stressors, and parental concerns about their child's future. Parents were asked about health problems, the child's age at which a physician was consulted for the first time because of the growth retardation, and who initiated the referral (school doctor, family doctor, or pediatrician). In order to determine the height-related psychosocial stressors encountered by the adolescents, the parents were asked whether their children were teased or juvenilized by other children. To find out parental worries about the future of their child, parents were asked whether they considered that the prospects of their child were equal to that of persons of normal height in the labor market and finding a spouse (yes, doubtful, no).
To assess behavioral and emotional problems, the parents filled out the Child Behavior Check List (CBCL) [25]. This questionnaire consists of the following eight syndrome areas (the so-called narrow-band scales): withdrawn behavior, somatic complaints, anxious/depressed behavior, social problems, thought problems, attention problems, delinquent behavior, and aggressive behavior. The scores of the subscales were summarized into a score for internalizing and externalizing problems, and a total problem score (the so-called broad-band scales).
The Silhouette Apperception Technique (SAT) has been shown to be a valid and useful instrument for assessing perceptions of the present and future heights of adolescents [7,26-28]. The parents were shown drawings of people of different heights, corresponding proportionally with the heights of the 3rd, 25th, 50th, 75th, and 97th percentiles, and asked to identify the drawings that they thought best fit the current and future heights of their child.
Adolescents' self-reports
The adolescents completed a global intelligence test and a series of questionnaires regarding perceived competence, psychological distress, and personality characteristics.
The Dutch versions of the Self-Perception Profile for Children (CBSK) and the Self-Perception Profile for Adolescents (CBSA) were used [29,30]. The CBSK consists of six perceived competence subscales: scholastic competence, social acceptance, athletic competence, physical appearance, behavior and conscience, and global self-worth. The CBSA consists of these six scales plus the scales of friendship and romantic love. In our study we only used the six scales that the CBSK and the CBSA have in common.
Anxiety was measured using the Dutch version of the State-Trait Anxiety Inventory for Children (ZBV-K), with the subscales of state anxiety and trait anxiety [31]. Depressive mood was assessed by the KDVK (Dutch Short Depression Questionnaire for Children) [32].
Personality characteristics of the adolescents were investigated using the Dutch Personality Questionnaire-Junior (NPV-J), with five subscales: inadequacy, perseverance, social inadequacy, recalcitrance, and dominance [33].
The SAT was used to measure perceptions of the present and future heights of the adolescents [7,26-28].
Intelligence was assessed by the Dutch short version of the Wechsler Intelligence Scale for Children (Revised) (WISC-Rn) [34].
Data analyses
Standard deviation scores (SDS) were used to compare the results of the quantitative questionnaires with the means of the general Dutch population. This was performed by subtracting the mean score of the norm group of the same age and sex from the score of the participants, with this difference divided by the SD of the norm group. This score expresses how much each study subject deviates from the norm in SD units. T-tests examined whether these age-and sex-adjusted scores significantly deviated from zero (i.e., from the norm).
To categorize groups of adolescents with divergent motives for wanting hormone treatment, we formed subgroups based on the presence of height-related stressors (being teased or juvenilized), parental worries about future prospects (regarding finding a spouse or job), parental worries about their child's behavior (internalizing or externalizing problems), and self-reported problems in psychosocial functioning (low self-esteem, anxiety, or depressive mood). The cut-off score for having externalizing or internalizing problems was a norm deviation score of 1 SD higher than the mean for the same age and sex on the CBCL scales for internalizing and externalizing behavioral problems. This score corresponds with a CBCL T-score of 60 (84th percentile) in the healthy norm group, and conforms with CBCL standards for the borderline range of clinical problems [25]. The cut-off score for anxiety and low self-esteem was a norm deviation score of at least 1 SD higher (anxiety) or lower (self-esteem) than the means for the same age and sex on the trait anxiety and the global self-worth scale, respectively. The cut-off score for indicating clinical signs of depression was a raw score of 4 points or more on the depression scale [32].
Our study group was small and significant differences were generally not evident in the results between boys and girls, between adolescents with ISS and those born SGA, and between adolescents who were to receive hormone treatment and those who were not. Hence our report focuses primarily on the results for the whole group.
Results
Medical referral
The mean age of the child when a physician was consulted for the first time because of the growth retardation was 7.3 years (SD = 3.5 years). Medical referral was initiated by the family physician (53%; seven adolescents with SGA, thirteen with ISS) or school physician (23.5%; two adolescents with SGA, seven with ISS), whereas some of the adolescents were already under supervision of a pediatrician as young children because of health problems other than short stature (23.5%; three adolescents with SGA, six with ISS). Ninety percent of the parents reported that their children had never experienced serious health problems.
Height-related psychosocial stressors and future expectations
Approximately one-quarter (28%) of parents reported that their child was teased by peers because of short stature (15% of boys, 35% of girls), and 14% of parents reported that their child was juvenilized by peers (23% of boys, 9% of girls). All but one adolescent took part in gymnastic lessons at school (97%).
Regarding the effects of short stature on future expectations, 44.5% of the parents expected their child to have a lower prospect in the labor market as an adult (39% of boys, 48% of girls), and 39% expected their child to have a lower prospect of finding a spouse (77% of boys, 17% of girls). This difference between boys and girls was significant (p < 0.01). No other significant differences between boys and girls were found.
Psychosocial functioning
Parental reports
On parental ratings of behavioral difficulties as measured by the CBCL, the short adolescents exhibited higher scores than the general Dutch norm group on the broad-band scales of internalizing problems (p < 0.05) and total behavioral problems (p < 0.05) (Table 1). The elevated score for withdrawn behavior was marginally significant (p < 0.10), while the higher score for anxious/depressed behavior was significant (p < 0.05). These narrow-band scales are summarized in the internalizing scale. The short adolescents also exhibited higher than normal scores on the social problems (p = 0.01) and attention problems (p < 0.05) scales. The scores deviated by 0.50–0.73 SDS from the norm (see Table 1), which reflects moderate effects. Three adolescents were in the clinical range for internalizing problems (CBCL T-score of 63; 90th percentile). The short adolescents did not score significantly higher than the norm group on the narrow-band scales of somatic complaints, thought problems, delinquent behavior and aggressive behavior, or the broad-band scale of externalizing problems. One adolescent scored in the clinical range for externalizing behavior.
Table 1 Emotional and behavioral problems of 34 adolescents with short stature as judged by their parents.
Variable d SD t p
Withdrawn behavior 0.34 1.10 1.82 0.08
Somatic complaints 0.38 1.34 1.66 0.11
Anxious/depressed behavior 0.54 1.35 2.35 0.03 *
Social problems 0.73 1.44 2.96 0.01 **
Thought problems 0.26 1.00 1.49 0.15
Attention problems 0.59 1.36 2.54 0.02 *
Delinquent behavior 0.29 1.46 1.18 0.25
Aggressive behavior 0.26 1.03 1.47 0.15
Internalizing problems 0.54 1.31 2.42 0.02 *
Externalizing problems 0.28 1.17 1.40 0.17
Total behavioral problems 0.50 1.25 2.35 0.03 *
Mean scores (d), standard deviations (SD), and t and p values. The d values reflect the deviations from the Dutch normative population in standard deviation units, where a positive score indicates that the adolescents with short stature are judged to have more problems than the norm group.
The d values have the following common effect sizes: a value smaller than 0.2 reflects no deviation from the norm, while values between 0.2 and 0.5, between 0.5 and 0.8, and greater than 0.8 reflect small, moderate, and large deviations, respectively [42]. T-tests examined whether norm deviation scores deviated from zero (the norm).
* p < 0.05, ** p < 0.01
To clarify the nature of the problems perceived by the parents, we examined specific items on the CBCL scales where parents observed significant problems. Items for which the adolescents deviated at least moderately from the norm included generic items on the scale of anxious/depressed behavior, height-related items on the scale of social problems, and both generic and height-related items on the scale of attention problems (for the specific items, see Table 2).
Table 2 Scores at separate items of the CBCL scales anxious/depressed behavior, social problems, and attention problems.
Item d SD t p
Anxious/depressed behavior:
112. Worries 0.83 1.36 3.55 0.001 **
45. Nervous, highstrung, or tense 0.65 1.25 3.03 0.005 **
35. Feels worthless or inferior 0.63 1.66 2.22 0.03 *
50. Too fearful or anxious 0.42 1.53 1.59 0.12
12. Complaints of loneliness 0.37 1.49 1.44 0.16
32. Feels he/she has to be perfect 0.35 1.03 1.98 0.06
103. Unhappy, sad, or depressed 0.33 1.21 1.57 0.13
14. Cries a lot 0.19 1.30 0.86 0.40
52. Feels too guilty 0.14 1.47 0.54 0.59
33. Feels or complaints that no one loves him/her 0.11 1.08 0.60 0.55
71. Self-conscious or easily embarrassed 0.11 1.19 0.53 0.60
89. Suspicious 0.08 0.98 0.45 0.65
31. Fears he/she might think or do something bad -0.03 0.89 -0.21 0.84
34. Feels others are out to get him/her -0.11 0.80 -0.83 0.41
Social problems:
64. Prefers being with younger kids 1.06 1.67 3.71 0.001 **
38. Gets teased a lot 0.91 1.38 3.86 0.000 **
1. Acts too young for his/her age 0.65 1.39 2.73 0.010 **
11. Clings to adults or too dependent 0.30 1.31 1.33 0.19
55. Overweight 0.03 1.00 0.20 0.85
48. Not liked by other kids -0.08 0.81 -0.58 0.57
62. Poorly coordinated or clumsy -0.10 0.75 -0.75 0.46
25. Doesn't get along with other kids -0.15 0.57 -1.56 0.13
Attention problems
1. Acts too young for his/her age 0.65 1.39 2.73 0.010 **
45. Nervous, highstrung, or tense 0.65 1.25 3.03 0.005 **
8. Can't concentrate, can't pay attention for long 0.42 1.20 2.07 0.047 *
46. Nervous movements or twitching 0.39 1.55 1.48 0.15
41. Impulsive or acts without thinking 0.31 1.23 1.48 0.15
13. Confused or seems to be in a fog 0.24 1.27 1.12 0.27
61. Poor school work 0.17 1.23 0.82 0.42
10. Can't sit still, restless, or hyperactive 0.15 0.99 0.90 0.37
17. Day-dreams or gets lost in his/her thoughts 0.14 1.05 0.77 0.45
80. Stares blankly 0.06 1.11 0.31 0.76
62. Poorly coordinated or clumsy -0.10 0.75 -0.75 0.46
Mean scores (d), standard deviations (SD), and t and p values. The d values reflect the deviations from the Dutch normative population in standard deviation units, where a positive score indicates that the adolescents with short stature are judged to have more problems than the norm group.
The d values have the following common effect sizes: a value smaller than 0.2 reflects no deviation from the norm, while values between 0.2 and 0.5, between 0.5 and 0.8, and greater than 0.8 reflect small, moderate, and large deviations, respectively [42]. T-tests examined whether norm deviation scores deviated from zero (the norm). * p < 0.05, ** p < 0.01
Note that some items of the CBCL load on more than one scale
Adolescents' self-reports
Virtually none of the scores of the short adolescents on self-reported questionnaires differed significantly from the norm group (Table 3). Adolescents with ISS or SGA did not deviate from the norm group with respect to perceived competence and psychological distress. Indeed, their perceived social acceptance was higher than that in the norm group (p = 0.05). With respect to personality characteristics, adolescents with ISS or SGA described themselves to be more persistent (quiet, conscientious, and diligent) than the norm group (p < 0.05). Two adolescents were in the clinical range (deviation of more than 2SDs) for global self-worth. One adolescent was at the clinical level for trait anxiety.
Table 3 Psychological functioning and distress as reported by the adolescents with short stature.
Variable n d SD t p
Perceived competence (CBSK):
Scholastic competence 31 0.15 1.06 0.78 0.45
Social acceptance 31 0.37 1.00 2.04 0.05 *
Athletic competence 31 0.22 1.11 1.08 0.29
Physical appearance 31 -0.21 0.98 -1.22 0.23
Behavior/conscience 31 0.17 1.05 0.90 0.38
Global self-worth 31 0.02 1.08 0.07 0.94
Psychological distress (ZBV-K):
State anxiety 37 -0.01 1.06 -0.08 0.94
Trait anxiety 37 0.09 1.12 -0.52 0.61
Personality characteristics (NPV-J):
Inadequacy 38 -0.08 0.88 -0.52 0.61
Perseverance 38 0.34 0.95 2.20 0.03 *
Social inadequacy 38 0.05 0.82 0.36 0.72
Recalcitrance 38 -0.05 0.97 -0.32 0.75
Domination 38 0.23 1.07 1.34 0.19
Mean scores (d), standard deviations (SD), and t and p values. The d values reflect the deviations from the Dutch normative population in standard deviation units, where a positive score indicates that the adolescents with short stature judged themselves to have higher perceived competence, more anxiety, and a higher score on personality scales than the norm group, respectively.
The d values have the following common effect sizes: a value smaller than 0.2 reflects no deviation from the norm, while values between 0.2 and 0.5, between 0.5 and 0.8, and greater than 0.8 reflect small, moderate, and large deviations, respectively [42]. T-tests examined whether norm deviation scores deviated from zero (the norm).
* p < 0.05, ** p < 0.01
Cognitive functioning
Intelligence scores (corrected total WISC-Rn scores) ranged from 66 to 128, with a mean score of 99.8. Six adolescents (three ISS, three SGA) had an intelligence lower than 80. The intelligence did not differ significantly between adolescents with ISS and those with SGA.
Expectations of hormone treatment: perception of current and future heights
Perceptions of current and future heights were analyzed separately in adolescents who were to receive hormone treatment and those who were not (Table 4). Most parents and adolescents in both the treatment and control groups estimated the current height of the adolescent at the 3rd percentile of the population references, which corresponds well to their actual height. Parents of the adolescents in the treatment group expected a greater future height than the parents of the adolescents in the control group (Z = -2.68, p = 0.007). None of the other comparisons between the experimental and control groups revealed significant differences.
Table 4 Current and future heights as perceived by adolescents and their parents on the Silhouette Apperception Technique
Treatment group (n = 19) Current height (%) Future height (%)
Percentile Adolescents Parents Adolescents Parents
3rd 52.6 88.9 5.3 0.0
25th 36.8 5.6 15.8 38.9
50th 5.3 5.6 26.3 38.9
75th 0.0 0.0 42.1 11.1
97th 5.3 0.0 10.5 11.1
Control group (n = 19) Current height (%) Future height (%)
Percentile Adolescents Parents Adolescents Parents
3rd 47.4 77.8 15.8 50.0
25th 47.4 16.7 26.3 22.2
50th 5.3 5.6 15.8 11.1
75th 0.0 0.0 42.1 16.7
97th 0.0 0.0 0.0 0.0
Groups classified according to motives
To provide a summary description, four groups of adolescents were distinguished based on the presence of height-related stressors, parental worries about future prospects, parental worries about their children's behavior, and self-reported problems in psychosocial functioning (Table 5).
Table 5 Classification of adolescents based on motives
Parental reports Self-reports
Height-related stressors Worries about future prospects Behavioral problems Psychosocial problems
Group 1 (n = 4) - - - -
Group 2a (n = 3) - + - -
Group 2b (n = 4) + - - -
Group 2c (n = 4) + + - -
Group 3a (n = 4) - + + -
Group 3b (n = 7) + + + -
Group 4a (n = 6) - - - +
Group 4b (n = 2) - - + +
The presence or absence of a motive is indicated by '+' and '-', respectively; motives include parental reports of the presence of height-related stressors (being teased or juvenilized), worries about future prospects (finding a spouse or job), and behavioral problems (internalizing or externalizing problems), and self-reporting of psychosocial problems (anxiety, low self-worth, or depressive mood). The four missing cases are due to one of the classifying variables being missing.
Group 1 consisted of four adolescents and their parents (12%) who did not report any psychosocial problems. However, all adolescents in this group reported having high expectations of the treatment in terms of height gain (these data are not listed in Table 5).
Group 2 consisted of 11 adolescents (32%) whose parents reported height-related psychosocial stressors or worries about future prospects, but no problems in psychosocial functioning (parental or adolescents' reports).
Group 3 consisted of 11 adolescents (32%) whose parents reported problems in psychosocial functioning as well as worries about future prospects and in most cases the presence of height-related psychosocial stressors. The adolescents themselves did not report problems in psychosocial functioning.
Group 4 consisted of eight adolescents (24%) who reported problems in psychosocial functioning, while their parents did not report height-related psychosocial stressors and worries about future prospects, and in most cases did not report behavioral problems.
Discussion
In a medically referred group of early pubertal adolescents with ISS or SGA, the motivation of the adolescents and their parents for choosing hormone treatment was investigated before initiating a combined GH and puberty delaying hormone treatment.
Parental reports revealed that current height-related psychosocial stressors were not the main reason for wanting growth-enhancing hormone treatment. Some of the parents reported that their children were teased (28%) or juvenilized (14%) because of their stature. These findings are close to those from another Dutch study [7], but in contrast with an American study that showed teasing and juvenilizing rates of 50% and 70%, respectively [14]. According to their parents, most of the adolescents in the present study were relatively free of current stressors. More than 40% of the parents expected that their children would have a decreased prospect in the labor market or difficulties in finding a spouse. This suggests that the motivation of providing opportunities for the future of the adolescents was a compelling reason for parents to choose hormone treatment.
Another possible reason for wanting hormone treatment is parental worries about the psychosocial functioning of their children. On average, as judged by their parents, the adolescents encountered internalizing symptoms, such as anxious or depressed tendencies, as well as social and attention problems. These problems were of a moderate magnitude compared to Dutch norms. It is likely that our analysis overestimated the actual psychosocial dysfunctioning, because the normative criteria were based on a very healthy group: any child who had been referred to a mental-health professional in the past 12 months, or who was currently receiving special educational services, was excluded from the normative sample [25,35]. Moreover, some of these problems are height-related issues that need not be a behavioral problem, such as preferring to be with younger kids or acting too young for his/her age.
The perceptions of adolescents about their own psychosocial functioning did not confirm the parental worries. The adolescents reported normal competence and personality, and even higher competence on social acceptance and perseverance, and little distress. This raises the question of whether adolescents or their parents are the best judges about well-being and functioning of adolescents. Adolescents may be unreliable informants because they are too young to give an adequate assessment of their own functioning, lack a time perspective, or have a tendency toward denial, while parents may be unreliable because of unrealistic anxieties about the health, future, and behavior of their children [36,37]. Our observation of more psychosocial problems being reported by parents than adolescents suggests that the perception of psychosocial problems is a stronger motive for parents than for adolescents when choosing to participate in the hormone treatment trial.
The perception of the current height of the adolescents was accurate in the majority of parents, while several adolescents had a tendency to overestimate their current height. Consistent with previous observations [26,38], several adolescents who were to receive hormone treatment had unrealistic expectations of their future height (as did their parents). Even those who were not to receive hormone treatment had high expectations of their future height. Optimism, even when unrealistic, has been shown to motivate the choice for a treatment and its adherence once started in several diseases [39,40]. However, unrealistic expectations may also be associated with a poor psychosocial outcome, as has been demonstrated in persons seeking cosmetic surgery, for example [41].
The tentative breakdown of subgroups provides a descriptive summary of four rather different profiles of motives for hormone treatment in the adolescents and their parents. A small group of adolescents and their parents reported no psychosocial problems. Highly positive treatment expectations of the adolescents in terms of height gain was the only detected motive, with the underlying reason remaining unclear. Perhaps it predominantly reflects the developmental wish of any child to want to grow (up). The parents reported height-related stressors or psychosocial problems and in most cases these worries about current problems were accompanied by worries about future prospects. The final group consisted of adolescents who reported problems in their psychosocial functioning, while their parents did not necessarily observe problems and were not worried. The cause of these problems and the relation with height remain unclear. In choosing such an intensive hormone treatment involving daily injections, pubertal delay, and possible side effects, we would have expected at least a subgroup of cases to show a motivation in parents as well as adolescents. However, none of the adolescents exhibited elevated scores on all motives, and only two pairs of parents and adolescents were congruent with respect to the observation of psychosocial problems. In the majority of cases it was either the parents or the adolescents who reported one or more motives.
Our sample size was sufficiently large to allow conclusions to be drawn regarding the comparison with normative data, but a larger sample size is needed for examining with sufficient power the possible roles of gender, type of short stature, and risks and protective factors that may modulate the psychosocial functioning of these adolescents [42].
Conclusion
Our study demonstrates that the motives of parents to let their children participate in a growth-enhancing hormone trial are more obvious than the motives of the early pubertal adolescents themselves. Two out of three parents reported worries about the future opportunities or observed modest current psychosocial problems in their children. The adolescents wanted to gain height, but the underlying motivation remains unclear. Few of the adolescents experienced psychosocial problems. Our analyses showed that motives varied among individuals. This result implies that when evaluating hormone treatment, the importance of divergent outcome variables will also differ among individuals. Effectiveness evaluations of hormone treatment to increase height and the consequential fulfillment of other goals must be awaited.
Abbreviations
ISS idiopathic short stature
SGA small for gestational age
SDS standard deviation score
GH growth hormone
GnRHa gonadotropin-releasing hormone agonists
CBCL Child Behavior Check List
SAT Silhouette Apperception Technique
CBSK/A Dutch version of the Self-Perception Profile for Children/Adolescents
ZBV-K Dutch version of the State-Trait Anxiety Inventory for Children
KDVK Dutch Short Depression Questionnaire for Children
NPV-J Dutch Personality Questionnaire-Junior
WISC-Rn Dutch short version of the Wechsler Intelligence Scale for Children (Revised)
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
HVB and RG analyzed the data and wrote the manuscript together with GS (the project supervisor) who, along with JH, was also involved in the study design and acquisition of data. GAK and JMW (the medical supervisor) were responsible for the design and execution of the medical part of the study. All authors critically revised the manuscript, and read and approved the final version.
Pre-publication history
The pre-publication history for this paper can be accessed here:
==== Refs
Voss LD Short normal stature and psychosocial disadvantage: a critical review of the evidence J Pediatr Endocrinol Metab 2001 14 701 711 11453518
Kranzler JH Rosenbloom AL Proctor B Diamond FBJ Watson M Is short stature a handicap? A comparison of the psychosocial functioning of referred and nonreferred children with normal short stature and children with normal stature J Pediatr 2000 136 96 102 10636982 10.1016/S0022-3476(00)90057-X
Voss LD Mulligan J Bullying in school: are short pupils at risk? Questionnaire study in a cohort BMJ 2000 320 612 613 10698879 10.1136/bmj.320.7235.612
Sandberg DE Bukowski WM Fung CM Noll RB Height and social adjustment: are extremes a cause for concern and action? Pediatrics 2004 114 744 750 15342848 10.1542/peds.2003-1169-L
Gordon M Crouthamel C Post EM Richman RA Psychosocial aspects of constitutional short stature: social competence, behavior problems, self-esteem, and family functioning J Pediatr 1982 101 477 480 7108676
Holmes CS Karlsson JA Thompson RG Stabler B and Underwood LE Longitudinal evaluation of behavior patterns in children with short stature Slow grows the child: psychosocial aspects of growth delay 1986 New Jersey, Lawrence Erlbaum Associates 1 12
Huisman J Medisch-psychologische bijdragen aan nieuwe behandelingswijzen in de kindergeneeskunde (The contribution of behavioral medicine to interventions in pediatrics) 1993 Amsterdam, Free University
Siegel P Stabler B and Underwood LE The psychological adjustment of short children and normal controls Growth, stature, and adaptation 1994 Chapel Hill, University of North Carolina 123 134
Sandberg DE Eiholzer U, Haverkamp F and Voss LD Experiences of being short: should we expect problems of psychosocial adjustment? Growth, stature, and psychosocial well-being 1999 Seattle/Toronto, Hogrefe & Huber Publishers 15 26
Frankel SA Psychological complications of short stature in childhood. Some implications of the role of visual comparisons in normal and pathological development Psychoanal Study Child 1996 51 455 474 9029971
Skuse D Gilmour J Tian CS Hindmarsh P Psychosocial assessment of children with short stature: a preliminary report Acta Paediatr Suppl 1994 406 11 17 7734805
Sandberg DE Stabler B and Underwood LE Short stature in middle childhood: a survey of psychosocial functioning in a clinic-referred sample Growth, stature, and adaptation 1994 Chapel Hill, University of North Carolina 19 34
Sandberg DE Brook AE Campos SP Short stature: a psychosocial burden requiring growth hormone therapy? Pediatrics 1994 94 832 840 7970998
Sandberg DE Michael P Drotar D Psychosocial stress related to short stature: does their presence imply psychological dysfunction? Measuring health-related quality of life in children and adolescents; implications for research and practice 1998 Mahwah, New Jersey, Lawrence Erlbaum Associates 287 312
Richman RA Gordon M Tegtmeyer P Crouthamel C Post EM Stabler B and Underwood LE Academic and emotional difficulties associated with constitutional short stature Slow grows the child: psychosocial aspects of growth delay 1986 Hillsdale, New Jersey, Lawrence Erlbaum Associates
Gilmour J Skuse D Short stature--the role of intelligence in psychosocial adjustment Arch Dis Child 1996 75 25 31 8813866
Sandberg DE Voss LD The psychosocial consequences of short stature: a review of the evidence Best Pract Res Clin Endocrinol Metab 2002 16 449 463 12464228 10.1053/beem.2002.0211
Ranke MB Towards a consensus on the definition of idiopathic short stature Horm Res 1996 45 Suppl 2 64 66 8805048
Kamp GA Mul D Waelkens JJ Jansen M Delemarre-van de Waal HA Verhoeven-Wind L Frolich M Oostdijk W Wit JM A randomized controlled trial of three years growth hormone and gonadotropin-releasing hormone agonist treatment in children with idiopathic short stature and intrauterine growth retardation J Clin Endocrinol Metab 2001 86 2969 2975 11443153 10.1210/jc.86.7.2969
Chatelain P Children born with intra-uterine growth retardation (IUGR) or small for gestational age (SGA): long term growth and metabolic consequences Endocr Regul 2000 34 33 36 10808251
Lundgren EM Cnattingius S Jonsson B Tuvemo T Intellectual and psychological performance in males born small for gestational age Horm Res 2003 59 Suppl 1 139 141 12638528 10.1159/000067850
van Pareren YK Duivenvoorden HJ Slijper FS Koot HM Hokken-Koelega AC Intelligence and psychosocial functioning during long-term growth hormone therapy in children born small for gestational age J Clin Endocrinol Metab 2004 89 5295 5302 15531473 10.1210/jc.2003-031187
van der Reijden-Lakeman I Slijper FM van Dongen-Melman JE de Waal WJ Verhulst FC Self-concept before and after two years of growth hormone treatment in intrauterine growth-retarded children Horm Res 1996 46 88 94 8871187
Lundgren M Cnattingius S Jonsson B Tuvemo T Intellectual performance in young adult males born small for gestational age Growth Horm IGF Res 2004 14 Suppl A S7 8 15135769 10.1016/j.ghir.2004.03.004
Verhulst FC Van der Ende J Koot HM Handleiding voor de CBCL 4-18 (Manual for the Child Behavior Check List) 1996 Assen/Maastricht, Van Gorcum
Grew RS Stabler B Williams RW Underwood LE Facilitating patient understanding in the treatment of growth delay Clin Pediatr (Phila) 1983 22 685 690 6617047
Mul D Versluis-den Bieman HJ Slijper FM Oostdijk W Waelkens JJ Drop SL Psychological assessments before and after treatment of early puberty in adopted children Acta Paediatr 2001 90 965 971 11683207 10.1080/080352501750064905
Ross JL Sandberg DE Rose SR Leschek EW Baron J Chipman JJ Cassorla FG Quigley CA Crowe BJ Roberts K Cutler GBJ Psychological adaptation in children with idiopathic short stature treated with growth hormone or placebo J Clin Endocrinol Metab 2004 89 4873 4878 15472178 10.1210/jc.2004-0791
Veerman JW Straathof MAE Treffers PDA van den Bergh BRH ten Brink LT Handleiding bij de Competentie Belevingsschaal voor Kinderen, CBSK (Manual of the Self Perception Profile for Children - Dutch version) 1997 Lisse, Swets & Zeitlinger
Treffers PDA Goedhart AW Veerman JW Van den Bergh BRH Ackaert L Rycke L Competentiebelevingsschaal voor adolescenten (CBSA), handleiding (Manual of the Self perception Profile for Adolescents - Dutch version) 2002 Lisse, Swets & Zeitlinger
Bakker FC Van Wieringen PCW Van der Ploeg HM Spielberger CD Handleiding bij de Zelf-Beoordelings-Vragenlijst voor kinderen, ZBV-K (Manual of the Stait-Trait Anxiety for children - Dutch version) 1989 Lisse, Swets & Zeitlinger
de Wit CAM Handleiding bij de depressievragenlijst voor kinderen, DVK en KDVK (Manual of the short depression questionnaire for children - Dutch version) 2001 Amersfoort/Leuven, Acco
Luteijn F van Dijk H van der Ploeg FAE Handleiding bij de Junior Nederlandse Persoonlijkheidsvragenlijst, NPV-J (Manual of the Dutch Personality Questionnaire - junior) 1989 Lisse, Swets & Zeitlinger
Van Haassen PP De Bruyn EEJ Pijl YL Poortinga YH Lutjespelberg HC van der Steene G Coetsier P Speolders-Claes R Stinissen J Handleiding bij de WISC-R (Manual of the Wechsler Intelligence Scale for Children - revised) 1986 Lisse, Swets & Zeitlinger
Sandberg DE Meyer-Bahlburg HF Yager TJ The Child Behavior Checklist nonclinical standardization samples: should they be utilized as norms? J Am Acad Child Adolesc Psychiatry 1991 30 124 134 2005047
Connolly MA Johnson JA Measuring quality of life in paediatric patients Pharmacoeconomics 1999 16 605 625 10724790
Theunissen NC Vogels TG Koopman HM Verrips GH Zwinderman KA Verloove-Vanhorick SP Wit JM The proxy problem: child report versus parent report in health-related quality of life research Qual Life Res 1998 7 387 397 9691719 10.1023/A:1008801802877
Erling A Wiklund I Albertsson-Wikland K Prepubertal children with short stature have a different perception of their well-being and stature than their parents Qual Life Res 1994 3 425 429 7866360 10.1007/BF00435394
Gudas LJ Koocher GP Wypij D Perceptions of medical compliance in children and adolescents with cystic fibrosis J Dev Behav Pediatr 1991 12 236 242 1939678
Fournier M De Ridder D Bensing J Optimism and adaptation to chronic disease: The role of optimism in relation to self-care options of type 1 diabetes mellitus, rheumatoid arthritis and multiple sclerosis Br J Health Psychol 2002 7 409 432 12614494 10.1348/135910702320645390
Honigman RJ Phillips KA Castle DJ A review of psychosocial outcomes for patients seeking cosmetic surgery Plast Reconstr Surg 2004 113 1229 1237 15083026 10.1097/01.PRS.0000110214.88868.CA
Cohen J Statistical power analysis for the behavioral sciences 1977 New York, Academic Press
| 15943869 | PMC1177961 | CC BY | 2021-01-04 16:31:07 | no | BMC Pediatr. 2005 Jun 8; 5:15 | utf-8 | BMC Pediatr | 2,005 | 10.1186/1471-2431-5-15 | oa_comm |
==== Front
BMC PediatrBMC PediatrBMC Pediatrics1471-2431BioMed Central 1471-2431-5-161595525110.1186/1471-2431-5-16Research ArticleThe clinical course of acute otitis media in high-risk Australian Aboriginal children: a longitudinal study Gibney Katherine B [email protected] Peter S [email protected] Jonathan R [email protected] Susan A [email protected] Heidi C [email protected] Elizabeth [email protected] Amanda J [email protected] Department of Medicine, Royal Darwin Hospital, Australia2 Ear Health and Education Unit, Menzies School of Health Research, Darwin, Australia3 Institute of Advanced Studies, Charles Darwin University, Australia4 Northern Territory Clinical School, Darwin, Australia5 Department of Paediatrics, University of Melbourne and Royal Children's Hospital, Melbourne, Australia6 Murdoch Children's Research Institute, Melbourne, Australia2005 14 6 2005 5 16 16 1 11 2004 14 6 2005 Copyright ©2005 Gibney et al; licensee BioMed Central Ltd.2005Gibney et al; licensee BioMed Central Ltd.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
It is unclear why some children with acute otitis media (AOM) have poor outcomes. Our aim was to describe the clinical course of AOM and the associated bacterial nasopharyngeal colonisation in a high-risk population of Australian Aboriginal children.
Methods
We examined Aboriginal children younger than eight years who had a clinical diagnosis of AOM. Pneumatic otoscopy and video-otoscopy of the tympanic membrane (TM) and tympanometry was done every weekday if possible. We followed children for either two weeks (AOM without perforation), or three weeks (AOM with perforation), or for longer periods if the infection persisted. Nasopharyngeal swabs were taken at study entry and then weekly.
Results
We enrolled 31 children and conducted a total of 219 assessments. Most children had bulging of the TM or recent middle ear discharge at diagnosis. Persistent signs of suppurative OM (without ear pain) were present in most children 7 days (23/30, 77%), and 14 days (20/26, 77%) later. Episodes of AOM did not usually have a sudden onset or short duration. Six of the 14 children with fresh discharge in their ear canal had an intact or functionally intact TM. Perforation size generally remained very small (<2% of the TM). Healing followed by re-perforation was common. Ninety-three nasophyngeal swabs were taken. Most swabs cultured Streptococcus pneumoniae (82%), Haemophilus influenzae (71%), and Moraxella catarrhalis (95%); 63% of swabs cultured all three pathogens.
Conclusion
In this high-risk population, AOM was generally painless and persistent. These infections were associated with persistent bacterial colonisation of the nasopharynx and any benefits of antibiotics were modest at best. Systematic follow up with careful examination and review of treatment are required and clinical resolution cannot be assumed.
==== Body
Background
Today, the majority of episodes of acute otitis media (AOM) in developed countries will resolve even if they are not treated with antibiotics. This conclusion is based on the findings of randomised placebo-controlled trials, longitudinal studies of initially withholding antibiotic treatment, and meta-analyses of randomised controlled trials. [1-3]. However, the outcomes in children from populations where there is an increased risk of complications remain uncertain.
We have described the onset of early and severe otitis media in Australian Aboriginal children[4,5]. This is associated with dense colonisation of the nasopharynx with respiratory bacteria. In previous studies in this population, AOM (defined as the presence of an effusion plus bulging of the tympanic membrane (TM) or recent perforation) was frequently asymptomatic. AOM was often present on otoscopic examination 4 weeks after onset, despite antibiotic treatment [6]. Children with frequent episodes of bulging of their TM were those most likely to develop new perforations. However, we could not determine whether AOM in high-risk populations initially responded to treatment and then recurred, or persisted despite treatment. The aim of this study was to describe the clinical course of AOM and the associated bacterial nasopharyngeal colonization in Aboriginal children from a remote community in the period immediately after diagnosis.
Methods
Setting
The local Human Research Ethics Committee and the community-controlled Tiwi Health Board approved the study. It took place in the year 2000 in a remote Aboriginal community situated 70 km north of Darwin (population 1300). The community has an average of 30 births per year and an infant mortality rate of 30 per 1000 live births. The standard of housing is poor and overcrowding is common [7].
Participants
We enrolled Aboriginal children younger than eight years who lived in the community if: i) they had AOM; ii) they were resident in the community; and iii) parents provided written consent for their participation. All children in this community develop otitis media by 12 weeks of age and around 50% experience a perforated TM in the first year of life [4]. Children in this study were similarly prone to severe infections and all had been treated for otitis media in the past. Staff at the community health centre and otitis media researchers working on other projects made the initial diagnosis. Health centre staff examined children's ears when they presented to the children's clinic either unwell or for follow-up of a medical condition. Researchers examined children's ears as part of a regular 4 weekly surveillance program. Children with otorrhoea that had persisted for longer than six weeks were not eligible unless they had a diagnosis of AOM in the other ear. We attempted to assess children every weekday they were in the study. The planned duration of follow-up was either two weeks (AOM without perforation), or three weeks (AOM with perforation). We followed children with persistent AOM until the AOM resolved, or the study period ended.
Clinical assessments
We used a questionnaire and review of the clinic notes to collect information about each child's current and past ear health. Wax and pus were removed from the ear canal under direct vision using a voroscope (WelchAllyn LumiView). A Siegel's speculum was used for pneumatic otoscopy. Images of the TM were video recorded and classified on a standardised data collection form. The position of the TM was described as retracted (R), neutral (N), mild bulging (B1), moderate bulging (B2), and marked bulging (B3). TMs that had a perforation that was associated with mild, moderate, or marked bulging (P1-P3) were considered to be functionally intact. Examples of these different positions and "pinhole perforations" are available in a training video [8]. Tympanometry (Grason Stadler GSI-38, Madison, Wisconsin, USA) was used in ears without any discharge. Tympanograms were classified according to a modified Jerger's classification [9]. Tympanograms with a canal volume of <0.3 cm3 were excluded.
We classified children according to the ear with the more severe acute infection. When an assessment was only successful in one ear, we used the diagnosis of that ear and assumed the disease was unilateral. Generally, the person who made the initial diagnosis started antibiotic treatment before referring the child to this study. Medical or nursing research staff prescribed additional antibiotics during the study according to population-specific, evidence-based guidelines [10]. We calculated compliance with prescribed antibiotics on a weekly basis (the number of doses reported as taken divided by the number of doses prescribed for that week). All families were reminded about their medication at each assessment and some had help in dispensing medicine from Aboriginal research assistants.
Definitions
Our classification of otitis media was based on current population-specific, evidence-based guidelines [10]. In this study (where close follow-up was ensured), we used the following criteria: i) Aerated middle ear-normal TM mobility and Type A or C tympanogram; ii) Otitis media with effusion (OME)-middle ear effusion behind an intact TM identified by an air-fluid level or bubble seen through a translucent TM or decreased mobility of the TM or type B tympanogram (admittance <0.2 mmho); iii) AOM without perforation-clinical diagnosis by health staff or moderate to marked bulging of an intact TM plus decreased TM mobility or Type B tympanogram; iv) AOM with perforation-evidence of recent TM perforation provided by clinical history and visualisation of fresh pus in the ear canal. The TM was intact if the perforation had already healed by the time of the assessment, and 'functionally intact' if the TM was bulging and the perforation could only be identified by pneumatic otoscopy; v) Chronic suppurative otitis media (CSOM)-TM perforation with otorrhoea present for more than six weeks; vi) Dry perforation-TM perforation with no discharge seen in the ear canal or within the middle ear space.
AOM was cured when there was no bulging and no discharge present on examination. Similarly, we described an episode of AOM as improved if: i) the ear discharge had resolved but bulging of the TM persisted in AOM with perforation, or ii) the bulging of the TM was reduced to mild in AOM without perforation. Examples of mild, moderate, and marked bulging of the TM are available in our training video [8]. The outcome at "1 week" was determined by the examination that was closest to day 7 (at least 5 days after the diagnosis). The outcome at "2 weeks" was determined by the examination that was closest to day 14 (at least 10 days after diagnosis).
Microbiology: specimen collection and processing
We took swabs of the nasopharynx (and ear discharge if present) on up to five occasions: day 0 (the day of enrolment) and days 4–7, 10–14, 17–21 and 24–28. All swabs were then smeared for gram staining and frozen in 1.0 ml skim-milk-glucose-glycerol broth (SMGGB). We processed swabs (after completion of clinical observations) using standard methods that have been previously published [4]. We tested isolates of pneumococcus for sensitivity to oxacillin, penicillin, erythromycin, sulphamethoxazole, tetracycline, and chloramphenicol using a disc diffusion method (calibrated dichotomous susceptibility, CDS) [11]. Colonies resistant to oxacillin or penicillin were classified as penicillin resistant. Colonies resistant to three or more classes of antibiotics were classified as multi-resistant. Penicillin minimum inhibitory concentration (MIC) was determined by E-test and categorised according to the following breakpoints: susceptible <0.064 μg; intermediate resistance 0.064 – 1.0 μg/ml; and high resistance >1.0 μg/ml.
Results
Features of participants and examinations
We enrolled 31 children in this study and completed 219 assessments. Clinic staff referred 13 children and other researchers referred 18 children. Children referred from the clinic were more likely to have otorrhea (8/13 versus 6/18). The mean duration of follow up was 21 days (range 3–57 days). Most children were less than 2 years of age and had a past history of otorrhea (see Table 1). Overall, 230 tympanograms were recorded. Nearly all of these were Type B (223/230, 97%). Tympanograms were not done if ear discharge was present.
Table 1 Characteristics of study participants.
Characteristic No. (%) of children
Number of children 31 (100%)
Male 15 (48%)
Age < 2 years 22 (71%)
Previous history of otorrhoea 22 (71%)
Otoscopy at start of study (per child):
Redness 24 (78%)
Bulging (intact, no discharge) 15 (49%)
Otorrhoea (<6 weeks) 14 (45%)
Perforation seen (<6 weeks) 12 (39%)
Bilateral AOM 13 (42%)
Initial diagnosis of AOM
Seventeen children had an initial diagnosis of AOM without perforation and 14 children had AOM with perforation. Nearly all children (27/31, 87%) had at least moderate bulging of the TM or recent discharge when they were first seen in this study; 13 had bilateral disease (42%). Symptoms of AOM could be assessed in 24/31 (77%) of study participants. Mothers reported otorrhea in 11/24 (46%) and ear pain in 7/24 (29%). Of the 14 children with AOM with perforation, two had a TM that appeared to have healed by the time of our examination and four had a TM that was functionally intact. All perforations seen were initially tiny in size (less than 2% of the area of the TM) and just antero-inferior to the centre of the TM.
Antibiotic treatment
Six children were known to be receiving antibiotics effective against respiratory pathogens at the time of the initial AOM diagnosis (amoxicillin 50 mg/kg/day). All the other children had received antibiotics for otitis media in the past but the time since their last treatment was not recorded. Following diagnosis, antibiotic treatment was started in an additional 22 children. Two of the three children not treated initially developed moderate bulging of the TM and started treatment on day 6–8 (A31 and A32). One child (A19) did not receive any antibiotics. The antibiotics prescribed were twice daily amoxicillin (25), amoxicillin-clavulanate (1), trimethoprim-sulphamethoxazole (3), or daily intramuscular procaine penicillin (1). The mean dose of penicillin or amoxicillin was 52 mg/kg/day (standard deviation 15, range 25–100). Recommended treatment duration was at least 7 days for AOM without perforation and at least 14 days for AOM with perforation. Compliance with prescribed treatment was poor. Overall, we documented 462/1151 prescribed doses (40%) as taken, and estimated that 17/30 participants (57%) took less than half their recommended treatment.
Clinical outcomes
AOM was persistent in most children from this high-risk population (see Table 2 and Table 3). Overall, 77% of children still signs of ongoing inflammation at the 7 day and 14 day assessments. Very few ears returned to normal. Of the 438 ear examinations attempted during this study, only 5 (1%) were consistent with an aerated, intact TM (A03 on 1 occasion, A19 on 1 occasion, and A29 on 3 occasions). Even when the acute infection resolved, recurrence was common. Children who received more than 50% of the prescribed antibiotics had similar rates of treatment failure to the children who did not (75% vs 71%, Risk difference 4%, 95%CI -33, 41). Similarly, children with a past history of otorrhea had similar rates of treatment failure to the children who did not (68% vs 75%, Risk difference -7%, 95%CI -43, 30). There were also no obvious differences in the clinical course of the six children who were already on antibiotics when they were enrolled in the study.
Table 2 Clinical outcomes at 7 days after diagnosis of acute otitis media (AOM). AOM was considered cured when there was no bulging and no discharge present on examination. An episode of AOM was considered improved if: i) the ear discharge had resolved but bulging of the TM persisted in AOM with perforation; or ii) the bulging of the TM was reduced to mild in AOM without perforation. Despite improvement, these children still had persistent signs of suppurative infection.
Cure (%) Improved (%) Not Improved (%)
AOM without Perforation (n = 16) 6 (38%) 1 (6%) 9 (56%)
AOM with Perforation (n = 14) 1 (7%) 3 (21%) 10 (71%)
TOTAL (n = 30) 7 (23%) 4 (13%) 19 (63%)
Table 3 Clinical outcomes at 14 days after diagnosis of acute otitis media (AOM). See Table 2 for explanation of diagnostic categories.
Cure (%) Improved (%) Not Improved (%)
AOM without Perforation (n = 14) 3 (21%) 1 (7%) 10 (71%)
AOM with Perforation (n = 12) 3 (25%) 1 (8%) 8 (67%)
TOTAL (n = 26) 6 (23%) 2 (8%) 18 (69%)
Seventeen children (55%) had a TM perforation documented at some point during the study. Of the 4 children with AOM with perforation who were followed for 6 weeks, 3 had healing and recurrence of their TM perforation documented. None of these four children had the typical otoscopic features of CSOM. In all cases the perforation was too small (usually a "pinhole") to allow adequate delivery of topical antibiotics to the middle ear space.
Microbiological outcomes
Overall, 93 nasopharyngeal swabs (31 children) and 38 swabs of ear discharge (13 children) were taken during this study. Each child was eligible for a maximum of five swabs (mean 3). Most of the nasopharyngeal swabs were positive for S. pneumoniae (76/93, 82%), H. influenzae (66/93, 71%) or M. catarrhalis (86/93, 95%). All but two of the nasopharyngeal swabs were positive for at least 1 pathogen (98%), and 59/93 (63%) were positive for all three otitis media pathogens. Antibiotic sensitivity testing was done in 74 of the 76 nasopharyngeal swabs positive for S. pneumoniae. Of these, 59/74 (80%) were penicillin resistant and 23/74 (31%) were multi-resistant. E-tests were available for 44 of the 59 penicillin resistant isolates: 3/44 (7%) were highly penicillin resistant.
We were able to take an initial swab in 27/31 children. All but one of these swabs were positive for at least one pathogen, and 18/27 (67%) were positive for all three otitis media pathogens-see Table 4. The prescription of antibiotics did not reduce carriage of these pathogens at follow up assessments. Nasopharyngeal carriage of penicillin resistant pneumococci was common at diagnosis (16/27, 59%). This did not change substantially following treatment-see Figure 3. The carriage of penicillin resistant pneumococci was not associated with an increased rate of clinical failure (62% vs 80%, Risk difference -18%, 95%CI -62, 25).
Table 4 The proportion of children with bacterial pathogens isolated from their nasopharynx.
Spn (%) PRSpn (%) Hi (%) Mc (%)
Day 0 22/27 (81%) 16/27 (59%) 19/27 (70%) 26/27 (96%)
Week 1 20/24 (83%) 16/24 (67%) 17/24 (71%) 24/24 (100%)
Week 2 18/21 (86%) 13/21 (62%) 17/21 (81%) 20/21 (95%)
Spn = Streptococcus pneumoniae; PRSpn = Penicillin Resistant Streptococcus pneumoniae; Hi = Haemophilus influenzae; Mc = Moraxella catarrhalis.
Identification of nasopharyngeal pathogens from swabs of ear discharge was less common: S. pneumoniae 11/38 (29%), H. influenzae 12/38 (32%), or M. catarrhalis 2/38 (5%). Nineteen ear discharge swabs (50%) were positive for at least one of these pathogens and 1/38 (3%) was positive for all three otitis media pathogens.
Onset and progression of AOM with perforation
The changes in TM position and diagnosis over the 219 examinations are shown in Figure 1. Of the 14 children who entered the study with a diagnosis of AOM with perforation, two TMs had already healed and six were functionally intact (ie. pinhole perforation only seen on pneumatic otoscopy). Both children who initially had fresh discharge obscuring an intact TM subsequently experienced recurrent perforations (A02 and A03). In the other 12 children with identifiable perforations, discharge continued seeping through the perforation almost immediately after completing the cleaning or swabbing of the middle ear canal. Eight TMs were initially perforated and in the neutral position. Over time, five of these also became functionally intact (ie. the perforation size reduced and the TM became bulging). AOM without perforation progressed to perforation in three children (A01, A10 and A26). In each case, the TM was observed to be bulging prior to perforation. On one occasion when a new perforation was observed, the TM had perforated and healed again within a 24 hour period (A26).
Figure 1 Position and integrity of tympanic membrane in 31 Australian Aboriginal children following a new diagnosis of acute otitis media. Shaded = discharge present; C = chronic perforation; P = acute perforation; B = bulging; n = neutral; R = retracted; 0 = no bulging; 1 = mild bulging; 2 = moderate bulging; 3 = marked bulging. * = past history of otorrhea.
There were two patterns of resolution of AOM with perforation. In four children, the suppurative process resolved and the perforation became dry (A07, A18, A22, A29). Two of these TMs subsequently healed during the study period. More commonly, the perforation appeared to be healing while the suppurative process was ongoing (A01, A02, A03, A10, A12, A14, A20, A26, and A28). Perforations were observed to heal and re-perforate in 4 children (A02, A03, A12, and A14). In one child, the right and left TM healed and re-perforated 8 times over a 6 week period (A02), confirming that healing and re-perforation can occur frequently.
Discussion
This was the first study describing the clinical course of AOM in Australian Aboriginal children. It is also the first study providing a detailed description of otoscopy findings in a population of high-risk of TM perforation. AOM (or suppurative OM) was common, usually not associated with ear pain, frequently bilateral, and often associated with perforation of the TM. In this population, AOM was generally persistent. Infections with a sudden onset and short duration were uncommon.
Strengths and limitations of the study
We used a standardised clinical assessment that included tympanometry and video-otoscopy. Since definitions of AOM vary [12,13],. the detailed description of the position and integrity of the TM over time is especially useful (see Figure 1). This information has not been reported previously.
Limitations of the study include the small number of children enrolled and the number of scheduled examinations that were missed. This is a consequence of performing research in remote Aboriginal communities. It reflects the challenges associated with research in remote locations with small populations. Participating families are culturally different, highly mobile, and have competing priorities that are not easily predicted. These factors make daily follow-up difficult. The small sample size and relative homogeneity of the study population mean that potentially important factors that predict outcome may no be identified. However, a larger study (or observation over a longer duration) is unlikely to change our conclusion that "persistent AOM" is common in this population (overall risk of persistent AOM 77% 95%CI 58, 90). Since there are currently no data available in the world on the clinical course of AOM in a population at high-risk of AOM, the information contained within Figure 1 is both unique and clinically important.
The high rates of nasopharyngeal carriage of all three respiratory pathogens are striking. The microbiological methods used in this small study cannot determine i) which pathogens extend from the nasopharynx to the middle ear space; ii) the relative importance of concurrent infection with multiple organisms; iii) the rate of acquisition of new pathogens while on treatment; or iv) the role of antibiotic resistant bacteria. While we did not find pneumococcal penicillin resistance of organisms in the nasopahrynx to be a good predictor of outcome, the small sample size mean that important effects cannot be excluded.
Comparison with outcomes in clinical trials
Should we be surprised that "persistent AOM" is so common in this population? In published prospective studies where children have not been treated with antibiotics, the rate of clinical failure in the first week ranged from 2% to 83% [14-23]. The median failure rate was 24%. The most striking influence on reported rates of clinical failure was the definition used for this outcome. Authors that defined clinical failure in terms of persistent symptoms (fever or otalgia) reported the lowest rates of clinical failure (all <25% at the end of the first week) [14,16-19]. The median clinical failure rate for these studies was 15%. In contrast, the five studies with the highest clinical failure rates (range 38–83%, median 73%) defined clinical failure as persistence of otoscopic signs [15,20-23]. Only one study described both the rate of persistent symptoms (13%) and persistent otoscopic signs (73%) after seven days [22]. Overall, these studies suggest that symptoms resolve quickly in most children with AOM not receiving antibiotics while otoscopic signs do not. Consequently, we should probably not be surprised by our findings of persistent otoscopic signs in a high-risk population where compliance with recommended antibiotic treatment is poor and antibiotic resistance is common.
Unfortunately, in nearly all published clinical trials, there is insufficient information to determine whether persistent bulging of the TM was a common finding [24]. Persistent middle ear discharge was probably unusual (as it is a readily identified complication of AOM). None of the studies described the outcomes specifically for the subgroup of children who had AOM with perforation at the time of diagnosis. Similarly, none of these studies described the associated nasopharyngeal colonisation during episodes of AOM. However, other studies in developed countries have found that dense colonisation with multiple bacterial pathogens is unusual [25]. This may be an important risk factor for the children included in this study.
Implications of the study
For populations where perforation of the TM and CSOM are uncommon, this study provides important information about: i) appropriate definitions for different types of OM; and ii) identification of individual children most at risk of CSOM. For high-risk populations, we believe our description of "persistent AOM" is likely to be generalisable. This persistence may be related to the severity of clinical presentation, the bacterial load (multiplicity of species and strains and their density of infection), the frequency of exposure to multiple pathogenic strains, or poor compliance with antibiotics. The early age of onset of suppurative ear infections and the low rates of reported symptoms make early recognition difficult. Consequently, active surveillance in all infants in this population is recommended.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
All the authors contributed to the development of the study protocol and the writing of the paper. As part of her BMedSci, KG initiated the study protocol, collected and analysed the data, and wrote the first draft of the manuscript. PM, JC, SS, and AL supervised KG and assisted with the design and analysis of the study. PM, JC, SS, and AL were responsible for early revisions of the manuscript. PM, AL, and ES assisted with data collection. AL, ES, and HSV supervised the microbiological aspects of the study that were completed by their technical staff. All authors read and approved the final manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
http://www.biomedcentral.com/1471-2431/5/16/prepub
Acknowledgements
We are extremely grateful to the Tiwi Health Board, Julanimawu Health Clinic, Murrupurttiyanawu Catholic School, and the Xavier Education Centre for supporting this study. We thank Kim Hare, Jemima Beissbarth, and Beverley Hayhurst for their assistance with data collection and laboratory procedures. Funding was provided by the National Health and Medical Research Council and the Cooperative Research Centre for Aboriginal and Tropical Health.
==== Refs
Takata GS Chan LS Shekelle P Morton SC Mason W Marcy SM Evidence assessment of management of acute otitis media: I. The role of antibiotics in treatment of uncomplicated acute otitis media Pediatrics 2001 108 239 247 10.1542/peds.108.2.239 11483783
Glasziou PP Del Mar CB Sanders SL Hayem M Antibiotics for acute otitis media in children Cochrane Database Syst Rev 2005 CD000219
O'Neill P Roberts T Acute otitis media Clin Evid 2004 324 337 15865642
Leach AJ Boswell JB Asche V Nienhuys TG Mathews JD Bacterial colonization of the nasopharynx predicts very early onset and persistence of otitis media in Australian aboriginal infants Pediatr Infect Dis J 1994 13 983 989 7845752
Boswell JB Nienhuys TG Patterns of persistent otitis media in the first year of life in aboriginal and non-aboriginal infants Ann Otol Rhinol Laryngol 1996 105 893 900 8916866
Morris PS Yonovitz A Leach AJ Mathews JD Tos M Are respiratory mucosal diseases the same in developing countries? Diagnosing acute otitis media in a high risk population Otitis Media Today 1999 The Hague: Kugler Publications 47 53
Hoy WE Norman RJ Hayhurst BG Pugsley DJ A health profile of adults in a Northern Territory aboriginal community, with an emphasis on preventable morbidities Aust N Z J Public Health 1997 21 121 126 9161065
McCallum G Wilson C Thomsen P Angeles G Leach A Morris P The Ear Video Workbook-Trainer Version. Improving the diagnosis and treatment of otitis media in young Aboriginal children Darwin, Australia. Cooperative Research Centre in Aboriginal and Tropical Health 2002
Tos M Poulsen G Hancke AB Screening tympanometry during the first year of life Acta Otolaryngol 1979 88 388 394 532614
Morris P Ballinger D Leach A Koops H Hayhurst B Stubbs L Scott J Anand A Daby J Paterson B Yonovitz A Recommendations for Clinical Care Guidelines on the Management of Otitis Media in Aboriginal and Torres Strait Islander Populations Canberra: Office of Aboriginal andTorres Strait Islander Health 2001
Bell SM Additions and modifications to the range of antibiotics tested by the CDS method of antibiotic sensitivity testing Pathology 1988 20 303 304 3205601
Hayden GF Acute suppurative otitis media in children. Diversity of clinical diagnostic criteria Clin Pediatr (Phila) 1981 20 99 104 7460467
Froom J Culpepper L Grob P Bartelds A Bowers P Bridges-Webb C Diagnosis and antibiotic treatment of acute otitis media: report from International Primary Care Network BMJ 1990 300 582 586 2108756
Howie VM Ploussard JH Efficacy of fixed combination antibiotics versus separate components in otitis media. Effectiveness of erythromycin estrolate, triple sulfonamide, ampicillin, erythromycin estolate-triple sulfonamide, and placebo in 280 patients with acute otitis media under two and one-half years of age Clin Pediatr (Phila) 1972 11 205 214 4537206
Damoiseaux RA van Balen FA Hoes AW Verheij TJ de Melker RA Primary care based randomised, double blind trial of amoxicillin versus placebo for acute otitis media in children aged under 2 years BMJ 2000 320 350 354 10.1136/bmj.320.7231.350 10657332
Halsted C Lepow ML Balassanian N Emmerich J Wolinsky E Otitis media. Clinical observations, microbiology, and evaluation of therapy Am J Dis Child 1968 115 542 551 4384577
van Buchem FL Dunk JH van't Hof MA Therapy of acute otitis media: myringotomy, antibiotics, or neither? A double-blind study in children Lancet 1981 2 883 887 10.1016/S0140-6736(81)91388-X 6117681
Appelman CL Claessen JQ Touw OF Hordijk GJ de MR Co-amoxiclav in recurrent acute otitis media: placebo controlled study BMJ 1991 303 1450 1452 1773154
Kaleida PH Casselbrant ML Rockette HE Paradise JL Bluestone CD Blatter MM Reisinger KS Wa;d ER Supance JS Amoxicillin or myringotomy or both for acute otitis media: results of a randomized clinical trial Pediatrics 1991 87 466 474 2011422
Laxdal OE Merida J Jones RH Treatment of acute otitis media: a controlled study of 142 children Can Med Assoc J 1970 102 263 268 4190002
Lorentzen P Haugsten P Treatment of acute suppurative otitis media J Laryngol Otol 1977 91 331 340 404375
van Buchem FL Peeters van MF Acute otitis media: a new treatment strategy Br Med J (Clin Res Ed) 1985 290 1033 1037
Engelhard D Cohen D Strauss N Sacks TG Jorczak-Sarni L Shapiro M Randomised study of myringotomy, amoxycillin/clavulanate, or both for acute otitis media in infants Lancet 1989 2 141 143 10.1016/S0140-6736(89)90192-X 2567903
Wald ER Acute otitis media: more trouble with the evidence Pediatr Infect Dis J 2003 22 103 104 10.1097/00006454-200302000-00001 12586970
Faden H Waz MJ Bernstein JM Brodsky L Stanievich J Ogra PL Nasopharyngeal flora in the first three years of life in normal and otitis-prone children Ann Otol Rhinol Laryngol 1991 100 612 615 1908199
| 15955251 | PMC1177962 | CC BY | 2021-01-04 22:25:38 | no | BMC Pediatr. 2005 Jun 14; 5:16 | utf-8 | BMC Pediatr | 2,005 | 10.1186/1471-2431-5-16 | oa_comm |
==== Front
BMC Public HealthBMC Public Health1471-2458BioMed Central London 1471-2458-5-591593510110.1186/1471-2458-5-59Research ArticleMeasles vaccine coverage and factors related to uncompleted vaccination among 18-month-old and 36-month-old children in Kyoto, Japan Matsumura Takayo [email protected] Takeo [email protected] Shigeru [email protected] Hideko [email protected] Department of Health Informatics, Kyoto University School of Public Health, Kyoto, Japan2 Kami-gyo Public Health Center, City of Kyoto, Kyoto, Japan3 Department of General Medicine and Clinical Epidemiology, Kyoto University Graduate School of Medicine, Kyoto, Japan4 Fushimi Public Health Center, City of Kyoto, Kyoto, Japan2005 4 6 2005 5 59 59 5 10 2004 4 6 2005 Copyright © 2005 Matsumura et al; licensee BioMed Central Ltd.2005Matsumura 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
Due to low vaccine coverage, Japan has not only experienced outbreaks of measles but has also been exporting it overseas. This study aims to survey measles vaccine coverage and the factors uncompleted vaccination among community-living children.
Methods
Subjects were the parents whose children had undergone either an 18-month or a 36-month checkup publicly provided by Kyoto City during November 2001 to January 2002. An anonymous self-administered questionnaire survey was conducted.
Results
The coverage was 73.2% among the 18-month-old children (n = 2707) and 88.9% among the 36-month-old children (n = 2340), respectively. The following characteristics of mothers were related to uncompleted measles vaccination: aged below 30, working, concerned about the adverse events of the vaccine, and had insufficient knowledge. Similarly, the following characteristics among children were related to uncompleted measles vaccination: not the first-born child, interacting with other children in group settings. The coverage was the lowest among the children whose mothers were concerned about the adverse events of the vaccine without proper knowledge of measles and its vaccination.
Conclusion
To increase vaccine coverage among children, parents' awareness about measles and vaccination against it should be promoted, especially for working mothers. Efforts to enhance access to vaccination services and to communicate with parents about changing vaccination schedules are necessary.
==== Body
Background
Despite the existence of a safe, effective, and inexpensive vaccine, measles is still not being controlled in many parts of the world. The use of measles vaccine over the last 30 years has reduced global measles morbidity and mortality by 74 and 85%, respectively, compared with the prevaccine era. Regional elimination goals have been established in the Americas (by 2000), Europe (by 2007) and the Eastern Mediterranean (by 2010) [1].
Vaccine coverage in excess of 95% interrupts endemic transmission of measles in many countries, but achievement of such coverage almost always requires great collaborative efforts. [2] The global strategy began in the 1990s and many countries implemented immunization and enhanced measles surveillance policies related to this during the late 1990's. Some (e.g. the Americas) achieved measles elimination quite early [2,3]. Since 2000, the WHO and UNICEF have recommended that, in addition to achieving high coverage with the first dose of measles vaccine, all children should be offered a second opportunity for measles vaccination to increase the total number of people who are immune [4].
In the United States, measles was targeted for elimination, and persisted at low incidence until 1989, when an epidemic swept the country. To prevent spread among school-age populations, a second dose of MMR vaccine is recommended. To increase immunization coverage among inner-city preschool populations, a number of activities have been undertaken to improve the immunization delivery system[5,6]. Today, measles has been eliminated in the United States [7].
During 1997 and 2001, a total of 156 (82%) of 191 countries provided a second opportunity to receive a vaccination through supplementary immunization activities or through routine health services. Japan remains as one of the remaining 35 countries that do not give a second opportunity for measles vaccination [4]. Of the 26 imported measles cases detected in the US in 2000, the largest number of these cases was from Japan (seven cases). In 2000, Japan is reported to be the principal exporting country of measles to the US [7].
The Immunization Law in Japan has been providing children with measles vaccination since 1978. Since Measles-mumps-rubella (MMR) vaccine was introduced into Japan in 1989, a number of cases of post-vaccination aseptic meningitis have been reported and these have been attributed to the use of Urabe Am9 mumps vaccine [8]. In 1993, the Ministry of Health and Welfare (MHW) withdrew the domestically produced MMR vaccine [9]. As of 1994, an amendment to the Immunization Law made vaccination voluntary and not mandatory. According to the present law, a single dose of measles vaccine is recommended for children over one year of age. Children are eligible to receive measles vaccination after 12 months following birth but not beyond 90 months. Until January 2004, adminisiration of measles vaccine was recommended between 12 and 24 months of age, instead of between 12 and 15 months when children have the greatest risk of contracting measles [10]. In Japan, measles vaccine coverage has remained low, and either small or moderate outbreaks have occurred repeatedly in communities. According to an infectious disease surveillance (2000), total measles cases were estimated to be from 180,000 to 210,000, and total deaths were estimated to be 88 [11,12]. Measles cases are most frequently observed among non-immunized children, particularly between 12 to 24 months. However, a nation-wide survey conducted in 2000 showed that measles vaccine coverage in Japan was 81.4%, which is not enough to prevent outbreaks [13]. In this context, the Japan Pediatric Association, the Japan Child Health Association, and the Japanese Association of Pediatrics jointly appealed to the Ministry of Health, Labor and Welfare to promote vaccination against measles in July 2001 [14]. Although secondary vaccine failure (SVF) has increased in Japan and discussion on two dose measles immunization has begun [15,16], actual increase in vaccine coverage among children between 12 to 24 months-old become a priority in the present setting. The level of vaccine coverage within a community is necessary to develop appropriate measures. Still, no public system in Japan is capable of providing such information on a regular basis.
The purpose of the present study is to examine the following: measles vaccine coverage among 18-month-old and 36-month-old children; parents' knowledge of and perceptions about measles and its vaccination; and factors related to uncompleted measles vaccination in Japanese local communities.
Methods
The present survey was conducted in Kyoto City, an area that covers 610 km2 and is situated in the middle of the Kinki district of Japan and has a population of 1.47 million as of October 2001. The number of one-year-olds and three-year-olds in Kyoto City are approximately 12,000 respectively. The survey covers a period of three months extending from November 2001 to January 2002. Subjects were parents of children who had received either an 18-month or a 36-month checkup at either one of the 11 health centers or the three substations in Kyoto City. In Japan, parents are encouraged to have their children receive 18-month and 36-month checkups at public health centers. The present survey was conducted in cooperation with the Kami-gyo Public Health Center located in Kyoto City. An anonymous self-administered questionnaire was sent out to each subject, enclosed with a pre-consultation form, which is usually posted to the parents prior to the health checkup. Questionnaire, which subjects completed at home, were collected on the day of the checkup. The questionnaire included items on relevant attributes of parents (e.g. gender, age, job status); relevant attributes of children (e.g. birth order, whether or not they interacted with other children in group settings (e.g. attending nursery schools), and presence/absence of allergy); the child's vaccination history; whether or not the child had contracted measles in the past; reasons for not receiving measles vaccination (if the relevant child had not been vaccinated against measles); knowledge regarding measles and its vaccination, and sources of information regarding measles vaccination.
Questionnaire items regarding knowledge of measles and its vaccination were developed based on a survey conducted in 1991, which investigated parents' awareness of the plausible grave consequences of contracting measles and the importance of measles vaccination [17].
This study's outcome measure is defined as measles vaccine coverage among 18 – and 36-month-old children. A logistic regression analysis was carried out to assess the factors related to uncompleted measles vaccination. To investigate how different levels of knowledge regarding measles and its vaccination, and varying degrees of concern for possible adverse events could influence uncompleted measles vaccination, the respondents were categorized into four groups according to their levels of knowledge and degree of concern. Respondents were also divided into a "low knowledge" group and a "high knowledge" group based on the median split of their total scores on a set of questions regarding knowledge of measles and its vaccination. A score of 0–6 points among the 18-month-old group and a score of 0–5 points among the 36-month-old group was deemed as indicative of "low knowledge", while a score of 7–12 points among the 18-month-old group and a score of 6–12 points among the 36-month-old group was deemed as indicative of "high knowledge." Results of the logistic regression analysis carried out using three dummy variables, namely X1, X2 and X3, are shown in Table 1.
Table 1 Characteristics of group 1~4
knowledge concern dummy variable Number
18-month group 36-month group
Group 1 high Low 690 748
Group 2 low Low X1 1145 813
Group 3 high high X2 367 379
Group 4 low high X3 418 306
Statistical software (JMP4J, SAS institute) was used to carry out the analysis. For the validation study, the subject's answer to the question regarding the child's vaccination history was checked against the child's Personal Health Record Book (this control was conducted only for a certain number of subjects).
Results
The number of children who received a health checkup and the response rate
During the survey period, 90.6% of 18-month-old children (3072 out of 3391) and 85.7% of 36-month-old children (2836 out of 3308) in the city received a health checkup. The response rate was 88.1% (2707 respondents out of 3072 patients who received a health checkup) and 82.5% (2340 respondents out of 2836 patients who received a health checkup). Because the majority of parents who responded to the questionnaire were mothers (98.1% of respondents among the 18-month-old group and 99.8% of respondents among the 36-month-old group), the analysis was conducted under the assumption that all respondents were mothers.
Measles vaccine coverage and incidence of measles
Measles vaccine coverage was calculated by dividing number of respondents whose children had been vaccinated against measles by total number of respondents. Coverage was 73.2% among the 18-month-old group and 88.9% among the 36-month-old group. In both groups, there were 44 children with a history of measles. The cumulative incidence of measles was calculated as 1.6% for the 18-month-old group and 1.9% for the 36-month-old group. In both groups, measles incidence was significantly higher among children who interacted with other children in group settings compared to children who did not. (18-month-old group: 3.3% (95%CI: 2.1–5.1) vs. 1.2% (95%CI: 0.8–1.7), 36-month-old group: 2.9% (95%CI: 2.0–4.3) vs. 1.3% (95%CI: 0.8–2.0))
To test validity, we compared responses about their children's history of measles vaccination from all respondents whose children received a health checkup at Kami-gyo Public Health Center during this survey period with the data registered in their children's Personal Health Record Books. Results showed that 98.2% of respondents (108 out of 110 respondents) in the 18-month-old group and 98.3% of respondents (117 out of 119 respondents) in the 36-month-old group provided valid answers regarding their children's history of measles vaccination.
The reasons for not receiving measles vaccine
Respondents whose children had not been vaccinated against measles (698 18-month-old children and 211 36-month-old children) were asked about the reasons for not having their child be vaccinated. (Table 2) In both groups, the majority of respondents replied that either they had not received vaccination but were planning to in the near future or that they had wanted to receive vaccination but had missed the opportunity to do so. The percentage of respondents who indicated concern about adverse events from a vaccination was 2.9% among the 18-month-old group and 9.5% among the 36-month-old group, with the percentage being slightly higher for the latter group.
Table 2 Reasons for not receiving measles vaccination (multiple answers)
18-month-old group 36-month-old group
Number of non-vaccinated children n = 698 n = 211
% %
has not yet received measles vaccination but is going to receive it in the near future 76.2 46.9
had a cold at the time of vaccination 32.5 24.2
had to receive other vaccinations first 23.5 11.4
wanted to receive measles vaccination but missed the chance to receive it 22.2 34.1
wasn't aware of/forgot about the vaccination schedule 6.9 9.5
was sick at the time of vaccination 4.9 1.9
has already been infected with measles 4.9 11.8
was concerned about adverse events of measles vaccination 2.9 9.5
naturally acquired immunity seems more effective than artificially acquired immunity 2.4 4.7
vaccination is no longer mandatory 2.4 6.6
vaccination does not seem to be sufficiently effective 0.7 1.4
Others 5.4 3.3
Knowledge about measles
The percentage of respondents who were aware that all children should be vaccinated against measles once they have reached the age of one was 78.7% among the 18-month-old group and 71.7% among the 36-month-old group. While approximately half of respondents were aware that measles is a highly contagious disease with symptoms of fever and rashes, only about 30% were aware that there have been cases of death from measles in Japan or that measles-infected children can suffer from the sequelae of measles. (Table 3) There were a total of 12 questions regarding knowledge of measles, with one point assigned to each question. The median of the total score for the 12 questions were 6 points in the 18-month-old group and 5 points in the 36-month-old group.
Table 3 Knowledge about measles among the mothers of children eligible for 18-month or 36-month checkup
Percentage of respondents answering "yes"
18-month group 36-month group
Participants n = 2707 n = 2340
All children should be vaccinated against measles once they have reached the age of one 78.7 71.7
Measles is accompanied by high fever 74.1 76.2
Measles is accompanied by rashes 72.6 73.7
Measles is highly contagious 66.4 67.9
Most people can be immunized against measles through vaccination 50.4 51.0
The best way to prevent measles is through vaccination 45.1 45.7
Severe cases of measles are more likely to occur in younger children 41.6 34.4
There are some reported cases of death from measles in Japan 40.7 37.6
Measles can cause encephalitis 37.6 38.4
Some measles-infected children may suffer from the sequelae of measles 30.7 30.7
Measles can cause pneumonia 26.0 24.0
Measles is accompanied by severe coughing 10.0 8.8
Predictors of uncompleted measles vaccination
A logistic regression analysis was conducted considering the following variables: mother's age (< 30 years), mother working, not first born child, interaction with other children in group settings, presence of allergies in child, mother having concern about the adverse events of measles vaccination, "low knowledge" about measles. (Table 4) Respondents were divided into a "low knowledge" group and a "high knowledge" group based on the median split of their scores on a set of questionnaire items related to the knowledge of measles and its vaccination. Among the variables related to mothers' characteristics, working and concern about adverse events of measles vaccination, and "low knowledge" of measles were significantly associated with uncompleted measles vaccination. Among variables related to child's characteristics, not first born child and interaction with other children were significantly associated with uncompleted measles vaccination.
Table 4 Factors related to uncompleted measles vaccination (univariate analysis)
variable Prevalence (%) Odds Ratio (95% CI) P value
18-month group Mother's age (< 30 years) 36.7 1.11 (0.93, 1.32) 0.26
Mother working 25.1 1.66 (1.37, 2.01) <0.0001
Not first born child 45.5 1.93 (1.62, 2.30) <0.0001
Interaction with other children 21.2 1.76 (1.44, 2.15) <0.0001
Presence of allergy 17.8 0.90 (0.71, 1.13) 0.37
Concern about adverse events 30.0 1.30 (1.07, 1.56) 0.007
Low knowledge 60.7 2.09 (1.73, 2.53) <0.0001
36-month group Mother's age (< 30 years) 22.5 1.32 (0.95, 1.81) 0.09
Mother working 33.3 2.60 (1.95, 3.49) <0.0001
Not first born child 46.5 2.04 (1.53, 2.74) <0.0001
Interaction with other children 38.3 2.64 (1.98, 3.54) <0.0001
Presence of allergy 27.9 1.31 (0.96,1.76) 0.09
Concern about adverse events 30.5 1.52 (1.12, 2.05) 0.01
Low knowledge 51.1 1.91 (1.43,2.58) <0.0001
CI: Confidence Intervals
To investigate how different levels of knowledge regarding measles and its vaccination and varying degrees of concern about possible adverse events could influence uncompleted vaccination, respondents were categorized into four groups according to their levels of knowledge (cutoff points = the median values) and degree of concern. This logistic regression analysis was carried out using three dummy variables, namely, X1, X2 and X3. (Table 1) Results showed that the odds ratio of uncompleted measles vaccination was highest for Group 4 (i.e. "low knowledge" and high concern about adverse events), followed by Group 2 ("low knowledge" and low concern about adverse events), Group 3 ("high knowledge" and high concern about adverse effects), and Group 1 ("high knowledge" and low concern about adverse events). (Table 5) Among the 18-month-old group, measles vaccine coverage was 73.2%, the odds ratio may be inflated as an approximate value of the risk ratio.
Table 5 Factors related to uncompleted measles vaccination (multivariate analysis)
18 month group 36 month group
Variable Odds Ratio (95% CI) P value Odds Ratio (95% CI) P value
Mother's age (< 30 years) 1.29 (1.06, 1.59) 0.01 1.67 (1.15, 2.40) 0.007
Mother working 1.38 (1.00, 1.89) 0.05 1.75 (1.16, 2.66) 0.008
Not first born child 2.24 (1.84, 2.73) <0.0001 2.38 (1.71, 3.34) <0.0001
Interaction with other children 1.57 (1.13, 2.18) 0.008 1.90 (1.24, 2.89) 0.003
Presence of allergy 0.84 (0.66, 1.07) 0.17 1.23 (0.88, 1.71) 0.22
High knowledge, low concern (Group 1) 1.00 1.00
Low knowledge, low concern (Group 2) 2.14 (1.67, 2.75) <0.0001 2.08 (1.38, 3.19) 0.001
High knowledge, high concern (Group 3) 1.36 (0.97, 1.90) 0.07 1.69 (1.01, 2.79) 0.04
Low knowledge, high concern (Group 4) 3.67 (2.74, 4.93) <0.0001 3.69 (2.30, 5.94) <0.0001
CI: Confidence Intervals
Discussion
Since 1978, measles vaccination has become part of routine vaccination schedules in Japan. Following the amendment to the Immunization Law in 1994, vaccination has become voluntary. According to this law, a single dose measles vaccine schedule is recommended for children over one year of age. Children are eligible to receive measles vaccination after 12 months following birth but not exceeding 90 months. In January 2004, a recommendation was issued that children should be vaccinated between 12 and 15 months of age, not between 12 and 24 months as it was previously. This change was proposed as a measure to prevent increased incidence of measles among non-vaccinated one-year-old children. Optional vaccination is available for individuals who are now older than the conventional age of vaccination. Due to low vaccine coverage, small or moderate outbreaks of measles are quite common in Japan. Although most infections occur in non-vaccinated young children around the age of one-year-old, cases of adult measles has also been on the rise [18-20]. Approximately 60% of reported measles cases in Japan involve people under the age of five years without a history of vaccination, and an increase in vaccination coverage at age one may be decreasing by approximately half, the total number of all measles patients [21].
To increase vaccine coverage against measles, an accurate estimate of the current rate of vaccine coverage is required. Presently, the coverage rate is calculated by dividing total number of persons vaccinated per year by total number of children between the ages of 12 and 24 months [11]. This formula does not accurately represent the actual number of children who have or have not been vaccinated against measles. In Japan, local governments are responsible for vaccination in their respective populations. Vaccination policy thus differs across different cities, towns and villages (namely, individual or mass vaccination, provided with or without charge, and vaccination age ranging from 12 to 18 months). Approximately 90% of local governments in Japan, including Kyoto City, provide vaccination to children who are older than one-year on an individual basis without charge [13]. According to the nation-wide survey conducted in 2000 (total number of persons vaccinated against measles per year over total number of persons who were going to receive vaccination per year), measles vaccine coverage in Japan is 81.4%, surpassing 80% for the first time since the survey initially began [13]. This survey's methodology, however, is problematic. The definition of denominator (total number of persons who were going to receive vaccination per year) differs across individual local governments. To obtain vaccine coverage is important for each local government.
Kyoto City experienced moderate-scale measles outbreaks in 1984, followed by smaller outbreaks in 1987, 1992 and 1997 at five-year intervals. The number of reported cases of measles at one point in 2001 was 205, with no reported cases of further spread. A further report identified that 44.4% of all measles cases occur in children ages two or below [22].
In the present survey, measles vaccine coverage, which was calculated by simply dividing the number of respondents whose children had been vaccinated by total number of respondents, was 73.2% among the 18-month-old group and 88.9% among the 36-month-old group. The percentage of children in Kyoto City that actually received health checkup was 90.6% among 18-month-olds and 85.7% among 36-month-olds. Because it can be assumed that vaccine coverage rate may be lower among children who do not receive health checkups, actual coverage rate may also be lower than that indicated in the present study.
Several previous surveys in Japan report vaccine coverage among 18-month children to be around 70% and that among 36-month children to be around 90% [23-25]. These findings are consistent with the present study. Several studies have also reported that factors related to a child-bearing environment, such as young motherhood, mother working, not being the first-born child, and attending nursery school, may contribute to low vaccine coverage [23-25]. Our findings agree. The present study also confirmed that factors such as concern about adverse events and low levels of knowledge were significantly associated with uncompleted measles vaccination. We checked for colinearity in the multivariate regression model. The correlation coefficients except among "mother working" and "interaction with other children" was 0.01–0.29. Although the correlation coefficients among "mother working" and "interaction with other children" was high (0.75 in 18-month group and 0.66 in 36-month group), these are different matters and we are interested in both of them. We put both of them into the final model, and got the result that we can interpret.
Lieu textitet al. examined the MMR vaccination history among 15-month-old children from affluent families and the factors that may delay vaccination. These factors were: having a large number of children, not having a regular doctor, not knowing when the shot was due, and not worrying about the risk of shots [26]. The present study showed that measles vaccine coverage was lowest among the group of children whose mothers were concerned about the adverse events of the vaccine without proper knowledge of measles and its vaccination (the coverage: was 18-month-old group: 61.9%, 36-month-old group: 85.0%). Even among mothers who indicated their concern about adverse events, the odds ratio of uncompleted measles vaccination was higher in those who had insufficient knowledge regarding measles and its vaccination. These results suggest that the level of knowledge is more highly correlated with immunization status than the level of concern about adverse events; this highlights the importance of promoting awareness among parents regarding measles vaccination. Effort is needed to improve benefit and risk communication [27]. Most important to vaccination is a balanced approach to the potential risks of natural infection, vaccine benefits, and adverse reactions to the vaccine. Given the potential risk of natural measles infection and the vaccine's efficiency and safety, the benefits are obvious [21].
As part of a campaign to raise awareness of measles vaccination in Japan among parents whose children are about to be enrolled in nursery schools, kindergartens and elementary schools, the Japanese Medical Association and the Japanese Association of Pediatrics have jointly launched a Vaccination Week for Children in March, 2004. During this Vaccination Week, consultation regarding measles vaccination was available to parents, and vaccination was carried out both on Saturday and Sunday.
Conclusion
In order to promote measles vaccination, further consideration should be given to the following measures: to promote awareness about measles, the benefits of measles vaccine, and the possibility of adverse events from vaccination among parents, especially among those who are aged below 30 or those who have two or more children; to establish a system to enhance access to vaccination services by, for example, providing vaccination services at nights and on weekends, and; to encourage and support nursery school staff in promoting vaccination as well as in collecting vaccination history data about the children.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
TM conceived of the study and completed the analysis. TN supervised all aspects of its implementation and was in charge of its writing. SO and HI assisted with the study and analysis. All authors read and approved the final manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
We appreciate the cooperation of the Kami-gyo Public health Center, City of Kyoto and its staff.
==== Refs
Cutts FT Henao-Restrepo A Olive JM Measles elimination: progress and challenges Vaccine 1999 17 S47 52 10559534 10.1016/S0264-410X(99)00309-6
Duke T Mgone CS Measles: not just another viral exanthem Lancet 2003 361 763 773 12620751 10.1016/S0140-6736(03)12661-X
de Quadros CA Olive JM Hersh BS Strassburg MA Henderson DA Branding-Bennett D Alleyne GA Measles elimination in the Americas. Evolving strategies JAMA 1996 275 224 229 8604176 10.1001/jama.275.3.224
Centers for Disease Control and Prevention Update: Global measles control and mortality reduction-worldwide, 1991–2001 MMWR 2003 52 471 475 12807081
Wood DL Brunell PA Measles control in the United States: problems of the past and challenges for the future Clin Microbiol Rev 1995 8 260 267 7621401
American Academy of Pediatrics. Committee on Infectious Diseases Age for routine administration of the second dose of measles-mumps-rubella vaccine Pediatrics 1998 101 129 133 11345975 10.1542/peds.101.1.129
Centers for Disease Control and Prevention Measles-United States, 2000 MMWR 2002 51 120 123 11898926
Fujinaga T Motegi Y Tamura H Kuroume T A prefecture-wide survey of mumps meningitis associated with measles, mumps and rubella vaccine Pediatr Infect Dis 1991 10 204 209
Ueda K Miyazaki C Hidaka Y Okada K Kusuhara K Kadoya R Aseptic meningitis caused by measles-mumps-rubella vaccine in Japan Lancet 1995 346 701 702 7658837 10.1016/S0140-6736(95)92311-X
Gomi H Takahashi H Why is measles still endemic in Japan? Lancet 2004 364 328 329 15276387 10.1016/S0140-6736(04)16715-9
Nakatani H Sano T Iuchi T Development of Vaccination Policy in Japan: Current issues and policy directions Jpn J Infect Dis 2002 55 101 111 12403906
Nagai M Second warning report on epidemics and patient estimation the Study Group on the Evaluation of Sentinel Surveillance Annual Report granted by the Ministry of Health, Labour and Welfare 2002 (in Japanese)
Isomura S Yamamoto N A study on effective procedures for measles vaccination. Vaccine Working Group Report granted by the Ministry of Health, Labour and Welfare 2002 189 196 (in Japanese)
President of the Japan Pediatric Society, The Japanese Society of Child Health and the Japan Pediatric Association For the improvement of measles vaccine coverage and towards its eradication A petition submitted to the Ministry of Health, Labour and Welfare 2002 (in Japanese)
Hirose M Hidaka Y Miyazaki C Ueda K Yoshikawa H Five cases of measles secondary vaccine failure with confirmed seroconversion after live measles vaccination Scand J Infect Dis 1997 29 187 190 9181657
Kobune T Katayama M Okada H Takeda M Sato T Tashiro M Measles vaccine failures and control of measles in Japan Clinical virology 2000 28 10 14 (in Japanese)
Isomura S Factors affecting acceptance rate of measles vaccine Virus 1998 48 183 188 (in Japanese)
Infectious Disease Surveillance Center Measles 1999~2001 Infectious Agents Surveillance Report 2001 22 273 277 (in Japanese)
Nakano T Ihara T Kamiya H Measles Outbreak among non-immunized children in a Japanese hospital Scand J Infect Dis 2002 34 426 429 12160169 10.1080/00365540110080638
Takayama N Change in the age-distribution of measles patients admitted to our hospital from 1981 to 2002 Kansenshogaku Zasshi 2003 77 488 492 (in Japanese) 12931574
Nakayama T Zhou J Fujino M Current status of measles in Japan J Infect Chemother 2003 9 1 7 12673398 10.1007/s10156-002-0206-6
Bureau of Public Health and Welfare, City of Kyoto Infectious Diseases Annual Report. Kyoto-City 2001 98 99 (in Japanese)
Yasui Y Imamura J Imai T Okazawa A Kida K Ichii M Sunagawa T Ohyama T Okabe N Okuno Y Results of the CAP study on measles vaccine and anti-measures for measles in Sakai, Japan Infectious Agents Surveillance Report 2001 22 280 282 (in Japanese)
Kato M Takahashi H An analysis of factors related to immunization rates The journal of child health 1999 58 373 378 (in Japanese)
Koshida R Inoue N Asaka K Sakurai N Vaccination coverage at 36-month child health check-ups in the western area of Kanazawa station, Japan Japanese journal of pediatrics 2000 53 1685 1690 (in Japanese)
Lieu TA Black SB Ray P Chellino M Shinefield HR Adler NE Risk factors for delayed immunization among children in an HMO Am J Public Health 1994 84 1621 1625 7943481
Dittmann S Vaccine safety: risk communication-a global perspective Vaccine 2001 19 2446 2456 11257376 10.1016/S0264-410X(00)00470-9
| 15935101 | PMC1177963 | CC BY | 2021-01-04 16:28:56 | no | BMC Public Health. 2005 Jun 4; 5:59 | utf-8 | BMC Public Health | 2,005 | 10.1186/1471-2458-5-59 | oa_comm |
==== Front
BMC Public HealthBMC Public Health1471-2458BioMed Central London 1471-2458-5-611593874710.1186/1471-2458-5-61Research ArticleSex- and age-specific trends in mortality from suicide and undetermined death in Germany 1991–2002 Baumert Jens J [email protected] Natalia [email protected] Karl-Heinz [email protected] Institute of Epidemiology, GSF National Research Center for Environment and Health, Ingolstaedter Landstr. 1, 85764 Neuherberg, Germany2 Department for Psychosomatic Medicine, Psychotherapy und Medical Psychology, University Hospital of the Technical University of Munich, Langerstr. 3, 81675 München, Germany2005 6 6 2005 5 61 61 21 2 2005 6 6 2005 Copyright © 2005 Baumert et al; licensee BioMed Central Ltd.2005Baumert 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
Over the last decade, significant downward linear time trends in suicide mortality were observed in most Western countries. To date, it is not established whether those favourable time trends developed homogeneously for sex and age groups and how they were affected by the number of undetermined deaths.
Methods
Data on suicide mortality and undetermined death from 1991 to 2002 in Germany were obtained from the German Federal Statistical Office. For each year, the age-standardised suicide rate (SR), undetermined death rate (UDR) and total rate (SR+UDR) was calculated by direct standardisation separately for men and women. Time trends were analyzed by Poisson regression estimating the average annual percentage change (AAPC) of the rates for sex and four age groups (15–24, 25–44, 45–74, ≥ 75 years).
Results
A significant decline of the SR was observed in all age groups but was less pronounced among the younger ages, particularly among men aged 15–24 years (AAPC -0.7%, p = 0.041). The SR in the oldest male age group (≥ 75 years) declined much stronger (AAPC -3.5%, p < 0.001). In women, the AAPC of the SR ranged from -1.7% to -4.6%. The average annual percentage changes in the age groups 25 – 74 years did not differ substantially for SR and SR+UDR. In contrast, due to an increase of undetermined deaths for subjects ≥ 75 years, time trends in this age group were affected by the number of undetermined deaths, especially in women.
Conclusion
Observing downward trends in suicide mortality with lower declines for younger subjects, prevention strategies should focus in particular on younger subjects.
==== Body
Background
Over the last two decades, significant downward trends in overall suicide rates have been observed in North America and in most countries of Europe [1-4]. Chishti et al. on behalf of the EUROSAVE working group [1] analysed data from 15 countries in the European Union (EU) and revealed downward trends in most EU countries except Ireland and Spain.
However, reasons for those favourable trends have not been ascertained so far. Moreover, it is not established whether those favourable time trends developed homogeneously for sex and age groups. It is not unlikely that the general decrease of suicide mortality may hide rises in certain age and sex groups. Guaiana et al. analysed age- and sex-specific suicide rates in Italy from 1986 to 1996 and found that in men and women over 45 years of age, rates progressively decreased while in men between 15 and 44 years of age, suicide rates progressively rose [5]. Gunnell and Middleton [6] revealed an overall decline of age-standardised suicide rates in England and Wales by 18% between 1981 and 1998. However, an increase of suicide rates for men aged 15–44 years was embedded in the downward trend.
Inconsistencies of death certification practices and the changes in the estimated number of unknown cases may bias the accuracy of national suicide figures. Therefore, analyses of suicide trends often include "undetermined deaths" in their calculations to take account of misclassifications. Chishti et al. [1] negated misclassification as reason for changing suicide rates for most EU member states, however, showed borderline significant upward linear trends in death due to undetermined causes for Belgium for the years 1994–1995 and Germany for the years 1984 – 1997.
The aim of the present study was to investigate whether time trends of suicide mortality in Germany over the 12-year observation period from 1991 to 2002 developed homogeneously for men and women older than ≥ 15 years in four age groups. Another major issue of the present study was a sex- and age-specific investigation whether trends of deaths due to undetermined causes affected suicide mortality trends.
Methods
Mortality data were provided by the Federal Statistical Office in Germany and allowed to calculate suicide and undetermined death rates for each year, standardised by sex and age [7]. Death by suicide was defined as "intentional self-harm" according to the ICD-9 (International Classification of Diseases, 9th revision; WHO, 1977) categories E950 to E959 in the time period from 1991 to 1997 and to the ICD-10 (10th revision; WHO, 1992) categories X60 to X84 and Y87 from 1998 to 2002 [8,9]. Deaths due to undetermined causes were defined according to ICD-9 (E980-E989) and ICD-10 (Y10-Y34, Y89).
Based on these data, three different "event rates" were defined: Suicide rate (SR), undetermined death rate (UDR) and the total rate (SR+UDR) which included deaths by suicide and undetermined causes. To adjust these event rates for annual changing population figures and age structures, direct standardization with 10-year age interval groups (15–24, ..., ≥ 75 years) were used to estimate age-adjusted event rates for each sex [10]. The population of Germany in 1991 was chosen as standard population. For four age groups (15–24, 25–44, 45–74, ≥ 75 years), event rates were estimated for 100,000 inhabitants aged ≥ 15 years and per year leading to a data set with 96 observations (12 years per two sexes per four age groups).
Time trends in the 12-year observation period from 1991 to 2002 for the events suicide and undetermined death as well as for the event suicide or undetermined death were analyzed by Poisson regression models using the event rate under concern as outcome variable [11]. Estimating the regression equation
Log (event rate) = α + β * YEAR + ε
(containing YEAR as the year of event as linear term), the average annual percentage change (AAPC) of the event rate was given by
AAPC = [exp() - 1] * 100
and calculated separately for each sex and age group (four categories).
Modifications of time trends of the age-standardized suicide, undetermined and total rate by sex were tested by additionally including variables sex and the interaction term year of event * sex in the regression equation:
Log (event rate) = α + β1 * YEAR
+ β2 * SEX
+ γ * YEAR * SEX + ε
A rejection of the test hypotheses γ = 0 indicates a significant modification of time trends by sex. Moreover, for each sex, modifications of time trends by age group was tested by including variables age group with interaction terms age group*year of event in the regression equation:
Log (event rate) = α + β1 * YEAR
+ β2 * AGE(25–44) + β3 * AGE(45–74) + β4 * AGE(≥ 75)
+ γ1 * YEAR * AGE(25–44) + γ2 * YEAR * AGE(45–74)
+ γ3 * YEAR * AGE(≥ 75) + ε
A rejection of the test hypotheses γ1 = 0, γ2 = 0, γ3 = 0 indicates a significant modification of time trends by age group.
To correct for over-dispersion of the Poisson model, the dispersion parameter was estimated by the ratio of the deviance to its associated degrees of freedom [11].
For all statistical analyses, a P value less than 0.05 was considered to be statistically significant. All analyses were performed with the statistical software package SAS for Unix, Version 8.2 [12].
Results
Overall time trend
From 1991 until 2002, a total of 146 709 completed suicides for subjects aged ≥ 15 years were recorded (104 675 men, 41 404 women). During this 12-year observation period, a decline of the age-standardised suicide rate (SR) was observed from 29.80 in 1991 to 22.78 in 2002 for men (Figure 1) and from 12.25 in 1991 to 8.09 in 2002 for women (Figure 2). The percentage change of the suicide rate was -23.6% in men and -34.0% in women leading to an impaired men-to-women-ratio in suicide rate which changed from 2.4 (1991) to 2.8 (2002). On average, age-standardised suicide rates declined significantly per year in men (AAPC -2.4%, 95% CI -4.1 to -0.7) and women (AAPC -4.0%, 95% CI -6.6 to -1.4). The downward trend was significantly stronger for women than for men (p < 0.001).
Figure 1 Age-standardized suicide rate (SR), undetermined death rate (UDR) and total rate (SR+UDR) from 1991 to 2002 for men aged ≥ 15 years.
Figure 2 Age-standardized suicide rate (SR), undetermined death rate (UDR) and total rate (SR+UDR) from 1991 to 2002 for women aged ≥ 15 years.
A number of 26 121 events (16 632 men, 9489 women) were classified as undetermined death. No substantial change of the age-standardised undetermined death rate (UDR) during the observation period was observed for men. In contrast, the UDR increased remarkably in women from 2.10 in 1991 to 2.86 in (2002) caused mainly by a strong "leap" from 1997 (with use of ICD-9) to 1998 (with use of ICD-10). Therefore, the total rate SR+UDR ran approximately parallel with SR in men but not in women.
Age-specific time trends
The decline in suicide rates was observed in all age groups but was less pronounced among the younger ages, particularly among men aged 15 – 24 years with an AAPC of -0.7 (95% CI: -1.3 to 0.0, p = 0.041) (Table 1). Compared to these, the suicide rate in the oldest male age group (≥ 75 years) declined by 3.5-times. In women, the modification of trends by age group was similar and the downward trend in each age interval was markedly higher compared to men (Table 2). The SR time trends were significantly modified by age group in men (p < 0.001) and women (p = 0.005).
Table 1 Average annual percentage changes (AAPC) of the suicide rate (SR), undetermined death rate (UDR) and total rate (SR+UDR) in men aged ≥ 15 years over a 12-year observation period in Germany, 1991–2002
Men 1991–1997 1998–2002 1991–2002
AAPC (95% CI) P value AAPC (95% CI) P value AAPC (95% CI) P value
SR
15–24 years -0.9 (-2.7 – 1.0) 0.361 -0.3 (-1.5 – 0.9) 0.596 -0.7 (-1.3 – 0.0) 0.041
25–44 years -1.5 (-2.5 – -0.5) 0.002 -2.1 (-3.3 – -0.9) 0.001 -2.2 (-2.6 – -1.8) <.001
45–74 years -2.5 (-3.5 – -1.5) <.001 -1.4 (-3.3 – 0.4) 0.135 -2.2 (-2.7 – -1.8) <.001
≥ 75 years -2.6 (-3.9 – -1.2) <.001 -2.3 (-4.2 – -0.3) 0.024 -3.5 (-4.2 – -2.9) <.001
p value for interaction* 0.395 0.596 <.001
UDR
15–24 years -1.9 (-4.2 – 0.3) 0.095 -3.0 (-9.1 – 3.4) 0.350 -0.6 (-2.0 – 0.7) 0.353
25–44 years -1.4 (-3.7 – 1.0) 0.245 -2.9 (-6.4 – 0.7) 0.109 -2.7 (-3.7 – -1.7) <.001
45–74 years -4.9 (-8.0 – -1.6) 0.004 3.1 (-0.9 – 7.3) 0.125 -0.5 (-2.3 – 1.4) 0.620
≥ 75 years -5.8 (-10.1 – -1.2) 0.014 -0.1 (-4.0 – 3.8) 0.945 7.3 (2.9 – 11.9) 0.001
p value for interaction* 0.248 0.367 <.001
SR+UDR
15–24 years -1.0 (-2.4 – 0.3) 0.136 -0.8 (-2.4 – 0.9) 0.369 -0.6 (-1.2 – -0.1) 0.013
25–44 years -1.5 (-2.2 – -0.8) <.001 -2.2 (-3.0 – -1.4) <.001 -2.3 (-2.6 – -1.9) <.001
45–74 years -2.8 (-3.6 – -1.9) <.001 -0.8 (-2.6 – 1.0) 0.371 -2.0 (-2.5 – -1.6) <.001
≥ 75 years -2.9 (-4.1 – -1.7) <.001 -1.9 (-4.2 – 0.5) 0.117 -2.1 (-2.7 – -1.6) <.001
p value for interaction* 0.118 0.763 0.009
* p value for test for interaction year of event * age group
Overall sex-specific time trends of the total rate (SR+UDR) revealed significant declines with an AAPC of -2.1 (95% CI: -2.4 to -1.9) in men and -2.4 (95% CI: -3.2 to -1.5) in women. The trends for undetermined death by age groups revealed significant increases of the rates in both sexes for subjects ≥ 75 years leading to a dilution of the stronger decline of the total rate SR+UDR in men aged ≥ 75 years (AAPC -2.1, p < 0.001). For women in the same age group, no decline could be observed for the total rate SR+UDR (AAPC 0.6, p = 0.597). The average annual percentage changes in the other age groups did not differ substantially for SR and SR+UDR.
Discussion
The present study provides an analysis of trends in suicide mortality and undetermined death rates over a recent 12-years observation period (1991–2002) in Germany and confirms the general downward trend for suicide mortality which reached a change of -23.6% for men and -34.0% for women from 1991 to 2002. These trends were not affected by deaths of undetermined causes in men and women aged 15 to 74 years.
Similar results for time trends of suicide rates were found in other studies. Levi et al. [2] analyzed trends in mortality from suicide excluding undetermined death over a much longer term period of 1965 to 1999 in 47 countries by applying the WHO database and revealed that age-standardized mortality from suicide in most European countries peaked in 1980–84 but decreased thereafter until 1995–98. On average, the downward trend was more pronounced in women compared to men. A similar finding was provided by Chishti et al. (2003) who found significant downward linear time trends in suicide mortality in most European countries for the years 1984 – 1998 [1]. In England and Wales, an overall decline of age-standardised suicide rates in men and women by 18% between 1981 and 1998 was observed [6]. Other studies, analysing data from Italy, Spain, Belgium, and Canada, also showed peaks of prevalence in suicide mortality in the 8th decade of the last century [5,13-15]. In contrast, Stark et al. observed increases in suicide mortality including undetermined deaths for all male age groups in Scotland in 1981 – 1999 while for women a downward trend was confirmed over the same observation period [16].
The present study adds new findings to time trends in suicide mortality and undetermined death by analysing sex- and age-specific developments. In brief, the analysis revealed more pronounced downward trends in women compared to men in all age groups even. Downward suicide mortality trends were less pronounced in age groups 15 – 24 years. This is in agreement with all other studies analysing age effects on suicide rates [5,13-15] except for Greece where increasing suicide trends 1980–1995 were most pronounced in the 45–54 male age group [16]. It is of major concern that especially young people had the least advantage from the general downward trend. In contrast to Gunnell and Middleton [6] who revealed an increase of suicide rates for men aged 15 – 44 years in England and Wales from 1981 to 1998, we found no "hidden" upward trends in particular age groups in our data base. To further maintain and stabilize the favourable downward trend and to tailor preventive measures according to actual requirements, further research highlighting possible causal explanations is urgently required. National prevention strategies should set a particular focus on younger, predominately male subjects.
To date, possible causal explanations remain highly speculative. However, among reasons which may account for the favourable findings in particular among the elderly and the much less favourable findings in the younger age groups are recent improvements in anti-depressive treatment. Recently, Gunnell and Ashby [18] studied associations of antidepressant prescribing and suicide rates in Britain. They revealed an increasing antidepressant prescribing in the elderly from 1991 to 1998 which was associated with a reduction in suicide rate in the observation period [18] Similar results were found in two other studies [19,20]. At the same time, palliative care has substantially improved over the last decade which may have contributed to findings which showed that a diagnosis of cancer in more recent data sets was no longer associated with increased risk of suicide [21,22]. A study investigating trends in risk factors for suicides revealed unfavourable trends with rises in divorce, declines in marriage and increases of unemployment and income inequality which are more related to suicides in younger subjects, especially in men, than in older subjects [23-25].
Informational value of the analysis is primarily restricted to Germany. However, results may be generalized because Germany holds a medial position in average suicide rates among western highly industrialized countries [1]. The time trend analysis was restricted to the last available data base with a 12-years observation period which may be considered as short to identify recently changing time patterns. However, number of cases included by each year was sufficient enough to allow valid calculations. Moreover, we restricted the data set to this time period because substantial changes in political and social structure had occurred beforehand in Germany.
Conclusion
In the present study, time trends of the suicide rate (SR) and the total rate (SR+UDR) ran similar over a 10-years observation period in the age groups 15 – 74 years indicating that the assessment of suicide mortality in these age groups was not affected by the number of undetermined deaths. Undetermined deaths are mainly considered as probable suicides and were included in several suicide investigations in the past. A study by Ohberg and Lonnquist estimated that almost 9 from 10 undetermined suicides can be regarded as suicides [26]. However, the assessment of time trends in the oldest age group (≥ 75 years) revealed to be rather difficult due to different AAPCs for the suicide rate and the total rate, especially in women. The rate of undetermined deaths increased for both sexes in the oldest age group. Further research is urgently recommended to ascertain reasons for the "leap" of the number of undetermined deaths from 1997 to 1998 for women aged ≥ 75 years.
List of abbreviations
AAPC Average annual percentage change
SR Suicide rate
SR+UDR Total rate
UDR Undetermined death rate
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
JJB performed the statistical analysis of the data, contributed to the design of the study and interpretation of the results and drafted the paper. NE contributed to the design of the study and the interpretation of the results. KHL had the idea of the study, contributed to the interpretation of the paper and revisited the paper critically for important intellectual content. All authors read and approved the final draft of the paper.
Table 2 Average annual percentage changes (AAPC) of the suicide rate (SR), undetermined death rate (UDR) and total rate (SR+UDR) in women aged ≥ 15 years over a 12-year observation period in Germany, 1991–2002
Women 1991–1997 1998–2002 1991–2002
AAPC (95% CI) P value AAPC (95% CI) P value AAPC (95% CI) P value
SR
15–24 years 0.2 (-2.7 – 3.1) 0.906 -3.7 (-8.3 – 1.1) 0.131 -1.7 (-3.0 -0.4) 0.012
25–44 years -2.7 (-4.7 – -0.7) 0.007 -1.3 (-4.2 – 1.6) 0.372 -2.9 (-3.7 -2.1) <.001
45–74 years -4.6 (-5.4 – -3.7) <.001 -1.2 (-1.8 – -0.5) <.001 -4.2 (-4.7 -3.6) <.001
≥ 75 years -5.7 (-7.8 – -3.4) <.001 0.5 (-1.8 – 2.8) 0.682 -4.6 (-5.8 -3.5) <.001
p value for interaction* 0.013 0.255 0.005
UDR
15–24 years -1.7 (-7.1 – 4.1) 0.559 -6.2 (-21.5 – 12) 0.478 -0.2 (-3.5 – 3.2) 0.913
25–44 years -5.6 (-8.1 – -3.0) <.001 -2.5 (-7.6 – 2.9) 0.358 -3.1 (-4.5 – -1.7) <.001
45–74 years -6.0 (-8.9 – -3.0) 0.001 -1.0 (-3.5 – 1.7) 0.467 -1.9 (-3.5 – -0.2) 0.030
≥ 75 years -8.6 (-12.1 – -5.0) <.001 1.6 (-0.4 – 3.7) 0.116 14.7 (7.6 – 22.3) <.001
p value for interaction* 0.144 0.399 <.001
SR+UDR
15–24 years -0.2 (-2.2 – 1.9) 0.865 -4.3 (-8.4 – 0.1) 0.053 -1.4 (-2.4 – -0.3) 0.011
25–44 years -3.2 (-5.0 – -1.3) 0.001 -1.5 (-4.6 – 1.7) 0.361 -3.0 (-3.7 – -2.2) <.001
45–74 years -4.7 (-5.5 – -3.9) <.001 -1.1 (-1.8 – -0.5) 0.001 -3.9 (-4.5 – -3.4) <.001
≥ 75 years -6.2 (-8.1 – -4.2) <.001 1.0 (-0.3 – 2.3) 0.125 0.6 (-1.6 – 2.8) 0.597
p value for interaction* 0.002 0.016 <.001
* p value for test for interaction year of event * age group
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
This study had no specific funding source. We are indebted to the Federal Statistical Office for providing the data.
==== Refs
Chishti P Stone DH Corcoran P Williamson E Petridou E EUROSAVE Working Group Suicide mortality in the European Union European J Public Health 2003 13 108 114 12803408 10.1093/eurpub/13.2.108
Levi F La Vecchia C Lucchini F Negri E Saxena S Maulik PK Saraceno B Trends in mortality from suicide, 1965–99 Acta Psychiatica Scandinavica 2003 108 341 349 10.1034/j.1600-0447.2003.00147.x
World Health Organization (WHO)
Mann J A current perspective of suicide and attempted suicide Annals Internal Med 2002 136 302 311
Guaiana G D'Avanzo B Barbui C Update of suicide trends in Italy from 1986 to 1996 Soc Psychiatry Psychiatr Epidemiol 2002 37 267 270 12111031 10.1007/s001270200018
Gunnell D Middleton N National suicide rates as an indicator of the effect of suicide on premature mortality Lancet 2003 362 961 962 14511931 10.1016/S0140-6736(03)14367-X
Statistisches Bundesamt
World Health Organization (WHO) The Ninth Revision of the International Classification of Diseaeses (ICD-9) 1977 Geneva: World Health Organization
World Health Organization (WHO) The Tenth Revision of the International Classification of Diseaeses (ICD-10) 1992 Geneva: World Health Organization
Breslow NE Day NE Statistical Methods in Cancer Research The Design and Analysis of Cohort Studies 1987 II Lyon: IARC
McCullagh P Nelder JA Generalized linear models 1989 2 Cambridge: University Press
SAS SAS/STAT User's Guide, Version 8 2000 Cary, NC: SAS Institute Inc
Granizo JJ Guallar E Rodriguez-Artalejo F Age-period-cohort analysis of suicide mortality rates in Spain, 1959–1991 Int J Epidemiol 1996 25 814 820 8921461
Moens GF The rise of suicide mortality in Belgium between 1968–72 and 1978–81: age, sex and geographical characteristics Int J Epidemiol 1984 13 243 245 6735572
Huchcroft SA Tanney BL Sex-specific suicide trends in Canada, 1971–1985 Int J Epidemiol 1988 17 839 843 3225093
Stark C Hopkins P Gibbs D Rapson T Belbin A Hay A Trends in suicide in Scotland 1981 – 1999: age, method and geography BMC Public Health 2004 4 49 15496228 10.1186/1471-2458-4-49
Zacharakis CA Madianos MG Papadimitriou GN Stefanis CN Suicide in Greece 1980–1995: patterns and social factors Soc Psychiatry Psychiatr Epidemiol 1998 33 471 476 9780809 10.1007/s001270050081
Gunnell D Ashby D Antidepressants and suicide: what is the balance of benefit and harm BMJ 2004 329 34 38 Review. 15231620 10.1136/bmj.329.7456.34
Morgan OWC Griffiths C Majeed A Association between mortality from suicide in England and antidepressant prescribing: an ecological study BMC Public Health 2004 4 63 15613234 10.1186/1471-2458-4-63
Kelly CB Ansari T Rafferty T Stevenson M Antidepressant prescribing and suicide rate in Northern Ireland Eur Psychiatry 2003 18 325 328 14643559 10.1016/j.eurpsy.2003.03.005
Lawrence D Almeida OP Hulse GK Jablensky AV Holman CD Suicide and attempted suicide among older adults in Western Australia Psychol Med 2000 30 813 821 11037089 10.1017/S0033291799002391
Akechi T Okamura H Yamawaki S Uchitomi Y Why do some cancer patients with depression desire an early death and others do not? Psychosomatics 2001 42 141 145 11239127 10.1176/appi.psy.42.2.141
Charlton J Trends and patterns in suicide in England and Wales Int J Epidemiol 1995 24 45 52
Gunnell D Lopatatzidis A Dorling D Wehner H Southall H Frankel S Suicide and unemployment in young people. Analysis of trends in England and Wales, 1921–1995 Br J Psychiatry 1999 175 263 270 10645329
Gunnell D Middleton N Whitley E Dorling D Frankel S Why are suicide rates rising in young men but falling in the elderly? – A time- series analysis of trends in England and Wales 1950–1998 Soc Sci Med 2003 57 595 611 12821009 10.1016/S0277-9536(02)00408-2
Ohberg A Lonnqvist J Suicides hidden among undetermined deaths Acta Psychiatr Scand 1998 98 214 218 9761408
| 15938747 | PMC1177964 | CC BY | 2021-01-04 16:28:56 | no | BMC Public Health. 2005 Jun 6; 5:61 | utf-8 | BMC Public Health | 2,005 | 10.1186/1471-2458-5-61 | oa_comm |
==== Front
BMC Public HealthBMC Public Health1471-2458BioMed Central London 1471-2458-5-651595524910.1186/1471-2458-5-65Research ArticleMunicipal health expectancy in Japan: decreased healthy longevity of older people in socioeconomically disadvantaged areas Fukuda Yoshiharu [email protected] Keiko [email protected] Takehito [email protected] Health Promotion/International Health, Division of Public Health, Graduate School of Tokyo Medical and Dental University, 1-5-45 Yushima, Bunkyo-ku, Tokyo 113-8519, Japan2005 14 6 2005 5 65 65 7 9 2004 14 6 2005 Copyright © 2005 Fukuda et al; licensee BioMed Central Ltd.2005Fukuda 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
Little is known about small-area variation in healthy longevity of older people and its socioeconomic correlates. This study aimed to estimate health expectancy at 65 years (HE65) at the municipal level in Japan, and to examine its relation to area socio-demographic conditions.
Methods
HE65 of municipalities (N = 3361) across Japan was estimated by a linear regression formula with life expectancy at 65 years and the prevalence of those certificated as needing nursing care. The relation between HE65 and area socio-demographic indicators was examined using correlation coefficients.
Results
The estimated HE65 (years) ranged from 13.13 to 17.39 for men and from 14.84 to 20.53 for women. HE65 was significantly positively correlated with the proportion of elderly and per capita income, and negatively correlated with the percentage of households of a single elderly person, divorce rate, and unemployment rate. These relations were stronger in large municipalities (with a population of more than 100,000) than in small and medium-size municipalities.
Conclusion
A decrease in healthy longevity of older people was associated with a higher percentage of households of a single elderly person and divorce rate, and lower socioeconomic conditions. This study suggests that older people in urban areas are susceptible to socio-demographic factors, and a social support network for older people living in socioeconomically disadvantaged conditions should be encouraged.
==== Body
Background
The ageing of the population has been progressing throughout the world, and Japan is the leading country with population ageing: the proportion of elderly (aged 65 years or over) was 17.4% in 2000 and is estimated to be over 35% in 2050 [1]. The quality of the years lived and well-being in an ageing society with declining mortality and increased life expectancy (LE) have been addressed [2].
To measure health levels including the aspects of quality of life and well-being, several comprehensive health indicators have been developed, and among them health expectancy (HE) has been commonly used [2-5]. HE is calculated by application of prevalence of morbidity data such as disability prevalence on age-specific person years in a life table [2,6]. The national health plan in Japan, "Health Japan 21", accompanied by local action plans, aims to prolong HE at the national and local levels through disease prevention and health promotion [7]. However, it is difficult to calculate and monitor HE for health policy planning and evaluation, especially at the local level, because of a lack of data on morbidity. Although previous studies have calculated HE at the prefectural level in Japan, these calculations used inconsistent and various databases of disability prevalence, and could not be applied at the municipal level because of limited availability of data [8-10].
In Japan, long-term care insurance (LTCI) was introduced in 2000 for older people needing nursing care, and citizens aged 40 years or over can receive insurance benefits after application and certification that they need nursing care. Since LTCI unions are managed by municipalities, it is theoretically possible to calculate HE using the municipal data of LTCI as morbidity. However, municipal HE has been calculated in only one prefecture [11], because the sex- and age-specific prevalence of those certificated as needing nursing care for LTCI by municipalities is not routinely published.
It was confirmed by previous studies that socio-demographic factors influence mortality, morbidity, and other health conditions of older people [12-15]. Socioeconomic status such as education level, occupational social class, and income has been demonstrated to be a critical factor predicting the health status of older people [13,15]. Family structure and living arrangement have been elicited as other critical factors influencing the health of older people, and conditions such as living alone and being widowed/divorced are related to poorer health status [16,17].
There was little relation between HE and socio-demographic indicators related to income and family structure at the prefectural level in Japan [18], while a municipal-level study in one prefecture showed a weak but significant relation between HE and socio-demographic indicators such as the percentage of nuclear households [11]. The relation between health status and regional characteristics might have been obscured due the higher level analysis which included heterogeneous smaller level regions, as a previous study demonstrated that the relation differed according to the area classification by urban-rural difference [19]. Analysis at different levels could lead to an inconsistent relation, and the smaller-area analysis more directly reflects the relation at the individual and household levels [20,21]. Therefore, studies examining the relation between socio-demographic factors and health including not only death but also morbidity at a smaller level will provide reliable evidence of the influence of socio-demographic factors on health in older people.
The purposes of this study were to estimate HE of older people at the municipal level across Japan, and to elucidate the association between HE and area socio-demographic factors.
Methods
Study units
Japan consists mainly of four islands (a map is shown in Figures 1 and 2). Hokkaido Island is located in the northern part, and includes only one prefecture (Hokkaido prefecture). The mainland (Honshu) includes several metropolitan areas such as Tokyo, Nagoya, and Osaka. The other islands are Shikoku and Kyushu.
Figure 1 Mapping of health expectancy for men. Map of Japan (a) and health expectancy at 65 years (HE65) of men by municipality (b). Municipalities (N = 3361) are classified into quintiles according to HE65, and are colored accordingly.
Figure 2 Mapping of health expectancy for women. Map of Japan (a) and health expectancy at 65 years (HE65) of women by municipality (b). Municipalities (N = 3361) are classified into quintiles according to HE65, and are colored accordingly.
According to Local Autonomy Law, local public entities in Japan are divided into two categories. The first category consists of cities, towns, and villages. The second category consists of prefectures (N = 47). All districts in the country belong to one of the municipalities and at the same time fall within the boundaries of one of the prefectures. As specific cases, the Tokyo prefecture (Tokyo Metropolis) includes 23 special wards ("ku") in addition to cities, towns and villages, and 12 large cities ("cities designated by ordinance") such as Nagoya and Osaka consist of wards ("ku"). In total, there were 3361 units of cities, towns and villages as well as Tokyo special wards and wards of cities designated by ordinance in December 2001.
Estimation of health expectancy
Calculation of HE by the Sullivan method requires life-table data and the sex- and age-specific morbidity [6]. However, municipal data of the sex- and age-specific number of those certificated as needing nursing care for LTCI are not available, and only the total number was available by insurance union. Therefore, this study estimated HE at 65 years (HE65) by a linear regression formula using life expectancy at 65 years (LE65) and the certificated rate in LTCI (P): HE65 = a + b × LE65 + c × P.
The formula was drawn from prefectural data (N = 47) of precise HE65, which was calculated by the Sullivan method with the prefectural life-table in 2000 [22] and the sex- and age-specific certification rate of LTCI in 2002 [23]. The coefficient of determinant (R2) for the regression with LE65 and the crude certification rate was 0.913 for men and 0.906 for women. When the standardized certification ratio (SCR), which was age-standardised by the indirect method, was used as a predictor, R2 was increased to 0.986 for men and 0.982 for women.
The validity of this estimation was confirmed by the municipal data (N = 59) of one prefecture (Shimane prefecture), in which the precise HE65 had been calculated by the Sullivan method [11]. The correlation coefficient between the precise HE65 and estimated HE65 predicted by the formula calculated from prefectural data was 0.976 for men and 0.956 for women.
Municipal LE65 was drawn from the municipal life-table in 2000, using the number of deaths during 1999–2001 and the 2000 census population, and the empirical Bayesian method was applied [24]. The number of those certificated as needing nursing care for LTCI by insurance union in May, 2002, was obtained from the database published by the Welfare and Medical Statistical Agency [25]. SCR was calculated using the municipal population and the age-specific certification rates of the national data from the All-Japan Federation of National Health Insurance Organizations, in which the age categories were 65–69, 70–74, 75–79, and 80+ years [23]. Although most municipalities (N = 2823, 84.0%) had their own insurance unions, there were 61 insurance alliances managed with neighbouring municipalities. In these cases, SCR of the alliance was calculated and used as that of the municipalities for HE estimation.
There was substantial variation in the population size among municipalities, ranging from only a few hundred to one million, and the certification rate in municipalities with a small population size showed extreme fluctuation. To correct the fluctuation of SCR, Bayesian SCR was estimated using hierarchical Poisson regression and Markov chain Monte Carlo (MCMC) method [26-31]. The levels of hierarchy were municipalities for the lower level and secondary medical care zones (SMCZs) for the higher level. The details of calculation of Bayesian SCR are described in the additional file [see Additional file 1]. The range of SCR was decreased from crude SCR of 24.8 to 178.0 (S.D. = 18.3) to Bayesian SCR of 46.8 to 161.3 (S.D. = 16.5).
The regression formula obtained from prefectural data and used for estimation of municipal HE65 was HE65 = 4.772 + 0.706 × LE65 – 1.776 × SCR for men and HE65 = 12.024 + 0.453 × LE65 – 4.576 × SCR for women. Statistical analysis was conducted using SPSS 11.0 for linear regression analysis and MLwiN 1.10 for Bayesian hierarchical Poisson regression. Geographic mapping of HE was conducted using ArcGIS 8.3, with the Japanese map (geographic coordinate system: GRS 1980) obtained from ESRI Japan .
Analysis of relation to socio-demographic indicators
The socio-demographic indicators used in this study were population, population density, proportion of elderly (aged 65 years or over), percentage of households of a single elderly person (aged 65 years or over), percentage of nuclear households, divorce rate, per capita income, and unemployment rate. The data were drawn from the database based on the population census and other national surveys of 1999 and 2000 [32]. Since crude divorce rate showed a large difference in variance within the three categories of municipalities according to the population size mentioned later, Bayesian method with hierarchical Poisson regression was applied to predict the value adjusted for the statistical fluctuation due to small population size, as well as SCR. A summary of the socio-demographic indicators is shown in Table 1.
Table 1 Characteristics of socio-demographic indicators among municipalities in Japan (N = 3361).
Indicator Mean ± S.D. (range)
Population (thousands) 37.8 ± 77.3 (0.2 – 1024.1)
Population density (per square kilometer) 928.9 ± 2257.6 (1.5 – 19854.1)
Proportion of elderly (%) a 23.7 ± 7.3 (7.6 – 50.6)
Percentage of households of a single elderly (%)a 8.0 ± 4.2 (0.7 – 31.6)
Percentage of nuclear households (%) 54.4 ± 8.7 (20.2 – 78.8)
Divorce rate (per 1000) 1.7 ± 0.4 (0.9 – 4.0)
Per capita income (thousand yen) 1225 ± 299 (419 – 3926)
Unemployment rate (%) 3.88 ± 1.64 (0.0 – 18.1)
a65 years or over
The relation of HE65 and LE65 to socio-demographic indicators was examined using correlation coefficient (Spearman's) for total municipalities. Then, we examined the correlation according to the population size of municipalities, in which municipalities were divided into three categories according to population size: less than 10,000 (small, N = 1554), 10,000 to 100,000 (medium-size, N = 1469), and more than 100,000 (large, N = 338). Statistical analysis was conducted using SPSS 11.0.
Results
Estimated health expectancy
Table 2 shows LE65 and estimated HE65. HE65 in men (years) ranged from 13.13 to 17.39 with a mean (S.D.) of 15.53 (0.52), and HE65 in women (years) ranged from 14.84 to 20.53 with a mean (S.D.) of 18.01 (0.77). For both men and women, large municipalities showed significantly (p < 0.001) lower HE65 than small and medium-size municipalities. HE65 for the overall Japanese population, which was estimated using the regression formula, was 15.39 for men and 17.62 for women, while HE65 estimated by the Sullivan method with sex- and age-specific long-term care prevalence was 15.40 for men and 17.62 for women, and LE65 in 2000 was 17.56 for men and 22.46 for women [22].
Table 2 Life expectancy at 65 years (LE65) and health expectancy at 65 years (HE65) among municipalities in Japan.
Sex Population size a LE65 HE65
Mean ± S.D. (range) Mean ± S.D. (range)
Men Total 17.56 ± 0.58 (14.9 – 20.3) 15.53 ± 0.52 (13.13 – 17.39)
Small 17.61 ± 0.59 (15.0 – 19.8) 15.56 ± 0.53 (13.63 – 17.39)
Medium 17.46 ± 0.58 (15.8 – 20.3) 15.53 ± 0.51 (13.48 – 17.08)
Large 17.58 ± 0.57 (14.9 – 19.5) 15.42 ± 0.49 (13.13 – 16.84)
Women Total 22.54 ± 0.68 (20.1 – 27.2) 18.01 ± 0.77 (14.84 – 20.53)
Small 22.64 ± 0.66 (20.1 – 26.3) 18.03 ± 0.78 (14.83 – 20.53)
Medium 22.48 ± 0.72 (20.2 – 27.3) 18.09 ± 0.77 (15.45 – 20.11)
Large 22.40 ± 0.57 (20.6 – 24.5) 17.62 ± 0.62 (14.84 – 18.94)
a Population size: total = total municipalities (N = 3361); small = less than 10,000 (N = 1554), medium = 10,000 to 100,000 (N = 1469), large = more than 100,000 (N = 338)
The geographic mapping of HE65 in men is shown in Figure 1. The central part of the mainland showed accumulation of municipalities with longer HE65, while the northern part of the mainland showed accumulation of municipalities with shorter HE65. Metropolitan areas, especially Osaka, showed relatively shorter HE65. Other regions such as Hokkaido, Kyushu and Shikoku showed a miscellaneous pattern of municipalities with shorter and longer HE65. The geographical pattern in women, shown in Figure 2, was similar to that in men, and the shorter HE65 in the metropolitan areas was more remarkable.
The relation to socio-demographic indicators
Table 3 shows the correlation coefficients of LE65 and HE65 with socio-demographic indicators. As the number of samples was extremely large, most of the correlation coefficients showed statistical significance (p < 0.001). Compared to LE65, HE65 showed a stronger correlation with the percentage of households of a single elderly person and per capita income for men, and with the percentage of nuclear households, divorce rate and unemployment for men and women, while a weaker correlation with the proportion of elderly for men and women. The correlation with the percentage of households of a single elderly person and per capita income for women was reversed between LE65 and HE65.
Table 3 Correlation of health expectancy at 65 years with socio-demographic indicators. Correlation coefficient of life expectancy at 65 years (LE65), health expectancy at 65 years (HE65) with socio-demographic indicators by size of municipalitiesa in Japan.
Men Women
Indicator LE65 HE65 LE65 HE65
Total Total Small Medium Large Total Total Small Medium Large
Population -0.11 -0.10 -0.12 -0.08 0.05 -0.18 -0.09 0.00 -0.13 -0.06
Population density -0.11 -0.10 -0.14 -0.03 -0.02 -0.20 -0.10 -0.01 -0.10 -0.27
Proportion of elderly 0.10 0.05 0.08 -0.02 -0.27 0.20 0.05 0.01 0.02 -0.15
Percentage of households of a single elderly 0.00 -0.18 -0.14 -0.27 -0.43 0.17 -0.25 -0.22 -0.33 -0.50
Percentage of nuclear households -0.01 -0.12 -0.13 -0.09 0.05 -0.03 -0.20 -0.21 -0.23 0.06
Divorce rate -0.17 -0.26 -0.25 -0.26 -0.44 -0.15 -0.29 -0.29 -0.26 -0.44
Per capita income 0.06 0.17 0.29 0.22 0.46 -0.15 0.16 0.34 0.21 0.26
Unemployment rate -0.28 -0.35 -0.36 -0.35 -0.52 -0.20 -0.33 -0.24 -0.38 -0.57
Life expectancy at 65 year for some sex 0.80 0.82 0.79 0.86 0.21 0.24 0.17 0.30
a Total = total municipalities (N = 3361); Small = municipalities with a population of less than 10,000 (N = 1554); Medium = municipalities with a population of 10,000 to 100,000 (N = 1469); Large = municipalities with a population of more than 100,000 (N = 338).
Statistical significant: p < 0.001 for coefficients of more than 0.05 or less than 0.05 in total municipalities and for coefficients of more than 0.1 or less than 0.1 in small, medium, and large municipality groups.
Concerning the correlation coefficients according to population size of municipalities, the socio-demographic indicators showed a stronger correlation in large municipalities (with a population of more than 100,000) than in small and medium-size municipalities, except for population, population density, and the percentage of nuclear households. Especially, the percentage of households of a single elderly person, divorce rate, per capita income and unemployment rate showed correlation coefficients of more than 0.4 (or less than -0.4) in large municipalities.
Discussion
This study estimated HE of older people and demonstrated its relation to socio-demographic indicators at the municipal level across Japan. As a result, the estimated HE65 (years) ranged from 13.13 to 17.39 for men and from 14.84 to 20.53 for women, and it was closely related to area socio-demographic conditions such as the percentage of households of a single elderly person, divorce rate, per capita income and unemployment rate. The relation of HE65 to socio-demographic indicators was stronger than that of LE65, and municipalities with a larger population showed a stronger relation between HE65 and socio-demographic factors.
Lower income and education are strongly related to higher morbidity of older people and shorter LE and HE in other countries [13,15,33]. In a previous study in Japan, however, area indicators related to income, education and unemployment showed little association with mortality of older people [26]. This finding was consistent with the result of this study that LE65 showed a significant but weak correlation with socio-demographic indicators. HE65 was more strongly correlated with socio-demographic indicators than LE65. These findings indicated that the health status measured by not only mortality but also morbidity is more sensitive for area socioeconomic conditions than that measured by only mortality, and the health level of older people would be substantially decreased by socioeconomic conditions in disadvantaged areas.
A notable finding of this study was the negative relation between HE65 and the percentage of households of a single elderly person and divorce rate. Shorter life expectancy might cause an increase in older people living alone because of an increase in widows/widowers. However, considering that the proportion of a single elderly person did not showed a negative relation to LE65, it is likely that the shorter HE was caused by an increase of older people living alone. It was confirmed by previous studies that living alone including divorce and loss of a spouse has a negative impact on the health of older people [16]. In addition, since most divorced persons were in the young- and middle-aged population, and divorces among older people are more unusual than among the middle-aged population [34], it seems plausible that the increased divorce rate influence the health status of older people through decreased support in families and communities.
In this study, we found a difference in the relation between HE65 and socio-demographic factors according to the size of municipalities. A previous study, also, showed that urban areas showed a stronger relation between health status and area socioeconomic conditions than rural areas [35]. The first possible reason for this urban-rural difference is the choice of indicator/index for area socioeconomic conditions. A previous study demonstrated that the traditional deprivation index (Townsend index) was related to health only in urban areas, while an index including more multiple domains was related to health in both urban and rural areas [19]. The second possible explanation is the existence of other factors mediating the relations, such as the social support network. The social support network has been demonstrated to influence the health status, and less social support is closely linked to poorer health in older people [12,36]. Also, social factors are less associated with mortality in a socially cohesive area [37], and the relation between socioeconomic inequality and mortality is mediated by so-called social capital [38]. Since urban areas show less social support network than rural areas in Japan [39,40], it is likely that in urban areas, less social support network causes a decline in health status and enhances the influence of other social factors on health status in the elderly population.
A few weaknesses of this study should be mentioned, especially concerning the estimation of disability prevalence and HE. First, the disability prevalence used for HE estimation was based on LTCI data. As a previous study demonstrated, utilization of LTCI services was dependent on not only the level of disability but also socioeconomic conditions such as household income and family structure [41]. However, LTCI in Japan provides consistent insurance across the nation and universally covers all elderly persons aged over 65 years [42], and thus, the influence of socioeconomic conditions on service utilization and regional variation would seem to be relatively small. In addition, LTCI data are the sole source for routine estimation of HE at present. Second, the disability prevalence (SCR) was estimated by hierarchal regression and MCMC procedure (fully Bayes method) [27-31]. Bayes estimation is commonly used for data smoothing in small area analysis and disease mapping, while alternative methods including empirical Bayes estimation can be applied [29-31,43-46]. Although SMCZ is generally used as a higher-level regional unit for Bayes estimation of municipal mortality data in Japan [24,26,47,48], an alternative hierarchical setting of regional levels would result in different figures of SCR, and consequently of HE. Finally, HE65 was estimated by a simple linear regression formula based on prefectural data. The validity of the estimated HE was confirmed by the municipal data of one prefecture, and the correlation coefficients between the precise and estimated HE were large enough to use the estimated HE for geographical comparison, instead of the precise HE. However, the application of this HE estimation for individual municipalities should be thought out. To expand the use of HE, the routine publication of detailed data of LTCI including the sex- and age-specific number of those certificated as needing nursing care is expected, which would enable calculation of the precise HE.
The results of this study have several implications in prolonging healthy longevity and supporting older people. The strong association between HE65 and socio-demographic factors indicates the need for efforts in the social context. Especially, there should be a focus on older people living alone. Also, the strong relation between HE65 and socio-demographic factors in large municipalities indicates that older people living in urban areas lack a social support network and are vulnerable to socioeconomically disadvantaged conditions. The social support network including an informal network and community activities will contribute to improving the health status of older people and prolongation of healthy longevity, especially in urban areas.
Although we addressed HE in this study, the results suggest that disability prevalence (SCR) is a useful indicator independently of mortality data including LE. Since HE was estimated using LE and SCR, the difference in the relation with socio-demographic indicators between LE and HE implies that SCR includes quite different dimensions from LE. The use of long-term care prevalence, in combination with mortality indicators (e.g., LE) and comprehensive indicators (e.g., HE), will contribute to concrete health policy making and evaluation for improvement of health, well-being, and quality of life in elderly population.
Conclusion
This study established a method to estimate HE available at the local level using a linear regression formula with life expectancy and long-term care prevalence. Municipal HE65 (years) ranged from 13.13 to 17.39 for men and from 14.84 to 20.53 for women. As a result of correlation analysis, a decreased HE was associated with higher percentage of households of a single elderly person and divorce rate, and lower socioeconomic conditions. This study suggests that older people in urban areas are susceptible to socio-demographic factors, and a social support network for older people living in socioeconomically disadvantaged conditions should be encouraged.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
YF designed the study, analyzed the data, and drafted the article. KN helped to interpret the results and edited the draft. TT supervised the data analysis and writing the article.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Supplementary Material
Additional File 1
Calculation of Bayesian SCR. Calculation of standardized certification ratio (SCR) using Bayesian hierarchical Poisson regression
Click here for file
Acknowledgements
This study was supported by a Grant-in-Aid for Scientific Research by the Japan Society for the Promotion of Science (Grant No. 14570326 and 16590497).
The authors wish to thank Dr Masahiro Umezaki (Graduate School of Tokyo Medical and Dental University) for technical advice on geographic mapping and Dr Tomoki Nakaya (Ritsumeikan University) for helpful comments on statistics.
==== Refs
Health and Welfare Statistics, Association Kokumin Eiseino Doko 2003 Tokyo: Health and Welfare Statistics Association
Robine JM Romieu I Cambois E Health life expectancy indicators Bull World Health Organ 1999 77 181 185 10083720
Mathers CD Sadana R Salomon JA Murray CJL Lopes AD Healthy life expectancy in 191 countries, 1999 Lancet 2001 357 1685 1691 11425392 10.1016/S0140-6736(00)04824-8
Bone MR Bebbington AC Nicholaas G Policy application of health life expectancy J Aging Health 1998 10 136 153
Pamuk ER Wagener DK Molla MT Achieving national health objectives: the impact on life expectancy and on healthy life expectancy Am J Public Health 2004 94 378 383 14998799
Sullivan DF A single index of mortality and morbidity HSMHA Health Report 1971 86 347 354
Kawahara K Health Promotion Law and Community Health Law 2003 Tokyo: Sunlife Kikaku
Miyashita M Hashimoto S Ojima T Nakamura Y Hayashi M Kato M Fukudome K Prevalence of the elderly requiring nursing care and health expectancy Koseino Shihyo 1999 46 25 29
Segami K Indicators toward the 21st century 2000 Tokyo: Health and Welfare Statistics Association
Hasegawa T Estimation of indicators of premature death and health expectancy for evaluation of Health Japan 21 2001 Tokyo: Ministry of Health, Labour and Welfare
Shimane Prefectural Institute of Public Health and Environmental Science Geographical variation of healthy life expectancy in Shimane Prefecture 2002 Matsue: Shimane Prefectural Institute of Public Health and Environmental Science
Stuck AE Walthert JM Nikolaus T Bula CJ Hohmann C Beck JC Risk factors for functional status decline in community-living elderly people: a systematic literature review Soc Sci Med 1999 48 445 469 10075171 10.1016/S0277-9536(98)00370-0
Huisman M Kunst AE Mackenbach JP Socioeconomic inequalities in morbidity among the elderly; a European overview Soc Sci Med 2003 57 861 873 12850111 10.1016/S0277-9536(02)00454-9
Bentham G Eimermann J Haynes R Lovett A Brainard J Limiting long term illness and its associations with mortality and indicators of social deprivation J Epidemiol Community Health 1995 49 S57 S64 8594136
Bassuk SB Berkman LF Amick BC 3rd Socioeconomic status and mortality among the elderly: findings from four US communities Am J Epidemiol 2002 155 520 533 11882526 10.1093/aje/155.6.520
Katz SJ Kabeto M Langa KM Gender disparities in the receipt of home care for elderly people with disability in the United States JAMA 2000 284 3022 3027 11122589 10.1001/jama.284.23.3022
Lin G Regional assessment of elderly disability in the U.S Soc Sci Med 2000 50 1015 1024 10714923 10.1016/S0277-9536(99)00351-2
Yamaguchi F Kakehashi M An analysis of factors concerned with the active life expectancy and the duration in care of the elderly J Population Problems 2001 57 51 67
Jordan H Roderick P Martin D The index of multiple deprivation 2000 and accessibility effects on health J Epidemiol Community Health 2004 58 250 257 14966241 10.1136/jech.2003.013011
Krieger N Chen JT Waterman PD Soobader MJ Subramanian SV Carson R Geocoding and monitoring of US socioeconomic inequalities in mortality and cancer incidence: does the choice of area-based measure and geographic level matter?: the Public Health Disparities Geocoding Project Am J Epidemiol 2002 156 471 482 12196317 10.1093/aje/kwf068
Pickett KE Pearl M Multilevel analysis of neighbourhood socioeconomic context and health outcomes: a critical review J Epidemiol Community Health 2001 55 111 122 11154250 10.1136/jech.55.2.111
Ministry of Health, Labour and Welfare Prefectural life table 2000 2003 Tokyo: Health and Welfare Statistics Association
All-Japan Federation of National Health Insurance Organizations Statistics for the long-term care insurance 2002 Tokyo: All-Japan Federation of National Health Insurance Organizations
Ministry of Health, Labour and Welfare Municipal life table 2000 2003 Tokyo: Health and Welfare Statistics Association
Welfare and Medical Service Agency Number of the certificated people of long term care insurance 2003 Tokyo: Welfare and Medical Service Agency
Fukuda Y Nakamura K Takano T Municipal socioeconomic status and mortality in Japan: sex and age differences, and trends in 1973–1998 Soc Sci Med 2004 59 2435 2445 15474199 10.1016/j.socscimed.2004.04.012
Leyland AH McLeod A Mortality in England and Wales, 1979–1992 2000 Glasgow: MRC Social & Public Health Sciences Unit
Rashash J Browne W Goldstein H Yang M Plewis I Healy M Woodhouse G Draper D Langford I Lewis A user's guide to MLwiN 2002 London: Centre for Multilevel Modelling, Institute of Education, University of London
Nakaya T Nakaya T, Tanimura S, Nihei N, Horikoshi Y Spatial epidemiology Health GIS 2004 Tokyo: Kokon Shoin 74 121
Wakefield JC Best NG Waller L Elliot JC, Wakefield NG, Best NG, Briggs DJ Bayesian approaches to disease mapping Spatial Epidemiology 2000 New York: Oxford University Press 104 127
Bernardinelli L Montomoli C Empirical Bayes versus fully Bayesian analysis of geographical variation in disease risk Stat Med 1992 11 983 1007 1496200
Statistical Information Institute for Consulting and Analysis Municipal data of social indicators 2003 Tokyo: Statistical Information Institute for Consulting and Analysis
Gutierrez-Fisac JL Gispert R Sola J Factors explaining the geographical differences in Disability Free Life Expectancy in Spain J Epidemiol Community Health 2000 54 451 455 10818121 10.1136/jech.54.6.451
Ministry of Health, Labour and Welfare Vital statistics of Japan 2001 2003 Tokyo: Health and Welfare Statistics Association
Barnett S Roderick P Martin D Diamond I Wrigley H Interactions between three proxies of health care need at the small area level: an urban/rural comparison J Epidemiol Community Health 2002 56 754 761 12239201 10.1136/jech.56.10.754
Kishi R Horikawa N Role of the social support network which influences age of death and physical function of elderly people study of trends in and outside of Japan and future problems Nippon Koshu Eisei Zasshi 2004 51 79 93 15058098
Iwasaki M Otani T Ohta A Sasazawa Y Kuroiwa M Suzuki S Rural-urban differences in sociodemographic, social network and lifestyle factors related to mortality of middle-aged Japanese men from the Komo-Iso cohort study J Epidemiol 2002 12 93 104 12033534
Kawachi I Kennedy BP Lochner K Prothrow-Stith D Social capital, income inequality, and mortality Am J Public Health 1997 87 1491 1498 9314802
Kishi R Eguchi T Sasatani H Yaguchi T Health status, social networks and support systems of elderly men and women in a large city and in former coal-mining areas – a population-based comparative study of 70-year-old residents of Sapporo and Yubari Nippon Koshu Eisei Zasshi 1994 41 474 488 8049516
Iwasaki M Otani T Sunaga R Miyazaki H Xiao L Wang N Sasazawa Y Suzuki S Social networks and mortality based on the Komo-Ise cohort study in Japan Int J Epidemiol 2002 31 1208 1218 12540724 10.1093/ije/31.6.1208
Sugisawa H Fukaya T Sugihara Y Ishikawa H Nakatani Y Kim HK Factors related to under-utilization of in-home services under the long term insurance system Nippon Koshu Eisei Zasshi 2002 49 425 436 12087769
Ikegami N Public long-term care insurance in Japan JAMA 1997 278 1310 1314 9343449 10.1001/jama.278.16.1310
Best N Richardson S Thomson A A comparison of Bayesian spatial models for disease mapping Stat Methods Med Res 2005 14 35 59 15690999 10.1191/0962280205sm388oa
Leyland AH Davies CA Empirical Bayes methods for disease mapping Stat Methods Med Res 2005 14 17 34 15690998 10.1191/0962280205sm387oa
Lawson AB Biggeri AB Boehning D Lesaffre E Viel JF Clark A Schlattmann P Divino F Disease mapping models: an empirical evaluation. Disease Mapping Collaborative Group Stat Med 2000 19 2217 2241 10960849 10.1002/1097-0258(20000915/30)19:17/18<2217::AID-SIM565>3.0.CO;2-E
Ghosh M Rao JNK Small area estimation an appraisal Statistical Science 1994 9 55 76
Nakaya T An information statistical evaluation of areal units for mortality mapping: toward constructing an analyzing system of disease maps on geographical information system Theory and Applications of GIS 1996 4 53 60
Hirako T Sakeki N Nakaya T Application of Bayesian methods on municipal health statistics Koseino Shihyo 1999 46 3 11
| 15955249 | PMC1177965 | CC BY | 2021-01-04 16:28:56 | no | BMC Public Health. 2005 Jun 14; 5:65 | utf-8 | BMC Public Health | 2,005 | 10.1186/1471-2458-5-65 | oa_comm |
==== Front
BMC Public HealthBMC Public Health1471-2458BioMed Central London 1471-2458-5-661595815610.1186/1471-2458-5-66Research ArticleSeroprevalence and risk factors for toxoplasma infection among pregnant women in Aydin province, Turkey Ertug Sema [email protected] Pinar [email protected] Munevver [email protected] Hasan [email protected] Department of Parasitology, Adnan Menderes University, School of Medicine, Aydin, Turkey2 Department of Public Health, Adnan Menderes University, School of Medicine, Aydin, Turkey3 Department of Child Health, Adnan Menderes University, School of Medicine, Aydin, Turkey4 Department of Gynecology, Adnan Menderes University, School of Medicine, Aydin, Turkey2005 15 6 2005 5 66 66 1 3 2005 15 6 2005 Copyright © 2005 Ertug et al; licensee BioMed Central Ltd.2005Ertug 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 aims of the present study were to determine the prevalence of toxoplasmosis in pregnant women at first trimester of their pregnancy and to follow up the seroconversion for next two trimesters, and to identify the risk factors and possible contamination routes in Aydin province, Turkey.
Method
The sample size was calculated as 423 on a prevalence of 50%, d=0.05 at a confidence level of 95% with 10% addition. It was a cross-sectional study with multistage sampling. After a questionnaire applied to the pregnant women, anti-Toxoplasma IgG antibodies were studied with ELISA and IFA, values in conflict with DA test, where IgM antibodies were studied with ELISA and for borderline or positive values of IgM avidity test was used.
Results
The mean age of 389 (92.9%) of pregnant women in the study was 24.28+/-4.56 years, the seroprevalence of anti-Toxoplasma IgG antibodies for toxoplasmosis was 30.1%. Seroprevalence was increased with age (p=0.001) and with drinking water consumption other than bottled water (p=0.042). No significant relations were observed between anti-Toxoplasma IgG antibodies and education level, being native or migrant, abortion history, consumption of meat, vegetable and milk/milk products, personal or kitchen hygiene habits, cat owning at home of the pregnant women. No IgM antibody was detected.
Conclusion
One of every three pregnant women in Aydin was at risk of toxoplasmosis at the first trimester of their pregnancy. Increased seroprevalance with age was a predictable result because of increasing time of exposure. Increased seroprevalence with consumption of municipal and uncontrolled water (well/spring water) supplies was similar with latest epidemiological findings.
==== Body
Background
Toxoplasma gondii (T.gondii), an obligate intracellular parasite found in many species throughout the world, causes a variety of clinical syndromes in human and animals [1]. Toxoplasmosis during pregnancy can cause congenital infection and manifest as mental retardation and blindness in the infant. The severity of fetal disease varies inversely with the gestational age at which maternal infection occurs [2].
Seroprevalence estimated for human population varies greatly among different countries, among different geographical areas within one country, and among different ethnic groups living in the same area [2]. Seroprevalence of T.gondii infection in women at childbearing age is found to be between 4%-100%. Incidence of primary maternal infection during pregnancy varies in a range of 1 to 310 per 10.000 pregnancies in the populations in Europe, Asia, Australia, and the Americas [2]. The serological screening of pregnant women for toxoplasmosis and the follow-up until delivery are not routine procedures in Turkey. In a few studies performed in our country, seroprevalence of T.gondii infection in women at childbearing age is found to be between 19.2% to 85%; and it is estimated that incidence of congenital toxoplasmosis is 0.1% [3-5].
Major routes of infection are: a) ingestion of oocysts through close contact with infected cat or cat's faeces, b) ingestion of water or food contaminated with the oocysts, c) eating raw or undercooked meat from infected animals that contain the tissue cysts, d) transplantation of infected organs, and e) congenital infection [1]. The following risk factors have also been identified in recent epidemiological studies: owning cats [6], eating raw or uncooked pork, lamb, mutton, beef, game or mincemeat products [6-10], eating raw or unwashed vegetables or fruits [8], frequent consumption of raw vegetables outside the home [6], travelling outside of Europe, the United States and Canada [7], having poor hand hygiene [6], washing kitchen knives infrequently [8], cleaning the cat litter box [8], contact with soil [7], a history of working in soil-related occupations [11]. Similarly, outbreaks of toxoplasmosis have been reported to be related with eating raw or uncooked pork, lamb, mutton or beef products [12] and additionally, with oocyst contaminated water [13] or soil [14].
The aims of the present study were to determine the prevalence of toxoplasmosis in pregnant women at the first trimester of their pregnancy and to follow up the seroconversion for the next two trimesters, and to identify the risk factors and possible contamination routes in the Aydin province, Turkey.
Methods
Study population
The study was performed in Aydin, one of the major cities in the Aegean region of Turkey with 8.007 km2 and 903677 population in 2004. Apart from the industry, the main agricultural products of the province are figs, olives, strawberries and cotton; especially figs are known throughout the world. The ancient name of the province of Aydin was Tralleis. It was celebrated as the center of sculpture, with a well known sculpture school. The data was acquired from health centers in urban and rural areas of Aydin. The study design was cross-sectional. The sample size was calculated as 384 on a prevalence of 50%, d = 0.05 at a confidence level of 95%. A total of 10% of the sample population was added to the sample size; so, the final study population size was 423. Multistage sampling was used in the selection of the study group. Aydin was separated into four regions according to the socio-economic and health data taken from Directory of Health. Four health centers (Two urban and two rural) were randomly selected from each region. The pregnant women were admitted into the study at their first visit to health center or when first detected as pregnant in the field by the midwife responsible for the selected population (approximately 2500 people per midwife) during a six month period. Permission was taken from Directorate of Health. A questionnaire was performed. There were questions eliciting socio-demographic data including age, education (illiteracy, primary school, high school, university or more), occupation, parity, residency and related risk factors including source of drinking water (general network, bottled water, well, spring, tap water, other), obstetrical history (total number of pregnancy, stillbirths, abortions and live births), frequency (one meal a day, a few meals in a week, a few meals in a month, seldom/never) and type of meat (beef, lamb, chicken, game, pork, delicatessen), vegetables and fruits (raw or not, at home or outside), egg, mayonnaise and milk (pasteurised, unpasteurised) consumption, cooking preferences (raw, rare-if the center is still raw, medium-if the center is still pink, well-done-if no pink meat is seen), kitchen hygiene (hand washing/kitchen utensils infrequently), owning cat (outdoor or indoor), history of cleaning cat litter box or feeding raw meat scraps, eating out (never, frequency- often: at least once a week, sometimes: a few a months, rare: a few in a year), travelling abroad, soil exposure (occupation or hobby). Informed consent was obtained from the participant.
Serum samples
In the first trimester, 389 (92.0%) women could be reached. During follow-up period, 257 (66.1 %) of them in the second trimester and 124 (31.9%) of them in the third trimester could be reached. Only one serum sample for each pregnant woman was studied at each trimester, so a total of 770 serum samples were studied at the end of the study.
Each blood sample was taken by a nurse at the health center and then centrifuged to separate serum and kept at -20°C in the laboratory of the department of parasitology. Serological study was performed within one week.
Serological methods
The Toxoplasma specific IgG antibodies in 770 serum samples of three trimesters were studied by Enzyme Linked Immunosorbant Assay (ELISA) and methods. In the first trimester (n = 389), there was a disagreement between ELISA and Indirect Immunofluorescence Antibody (IFA) methods (IFA negative, ELISA positive) in six (1.54%) serum samples, then these samples were also studied by direct agglutination test (DA) and no antibody answer was observed. No disagreement between ELISA and IFA methods was observed in other samples.
The Toxoplasma specific IgM antibodies were studied for 770 serum samples of three trimesters by ELISA method. Ten border results (1.30%) and three positive results (0.39%) were also assessed by avidity test. There was high avidity at 13 (1.69%) serum samples.
a. ELISA
The Toxoplasma specific IgG antibodies were studied by commercial kit (Biokit-Bioelisa Toxo IgG/Italy) according to the manufacturer's instructions. All specimens were analysed using enzyme immunoassay test. The results more then 10 IU/ml were taken as positive results in the current study. The consistency between IFAT and bioELISA IgG kit was found high in an earlier study performed in the parasitology laboratory (kappa = 0.941, p = 0000) (15).
The Toxoplasma specific IgM antibodies were studied by commercial kit (Organon-Toxonostika IgM II MikroELISA kit) according to the manufacturer's instructions. The consistency of this commercial kit was found to be high by comparisons with many other commercial kits [16].
b. IFA test
Antigen preparations were made from tachyzoites of the TRH strain of T.gondii. Tachyzoites were obtained from the peritoneal exudates of mice infected 2 days earlier. These antigens have been routinely used for the serological tests in our laboratory [17]. Dilutions of 1:16 and higher were evaluated as positive in the current study.
c. DA test
Serum samples were analysed for antibodies to T. gondii by the direct agglutination test using formalin fixed tachyzoites as antigen. Titers ≥ 20 were considered positive [18,19].
d. Avidity Test
EIAgen Toxoplasma IgG Avidity test by Adaltis (Italy) was used in the study and avidity was considered as high avidity for results more then 35%, as low avidity for less then 30% and borderline avidity between 30%-35%.
Statistical assessment
A statistical software package was used for data analysis. The descriptive data was given as mean ± standard deviation (SD). The chi-squared test was used for the analytic assessment. The differences were considered to be statistically significant when the p value obtained was less than 0.05.
Results
There were 389 pregnant women at the beginning of the study. The mean age of the women was 24.28 ± 4.56 years. The mean gestation week was 9.85 ± 2.21. The means of total number of pregnancies, total abortion and stillbirth, and total live births were 2.01 ± 1.37, 0.36 ± 0.80 and 0.65 ± 0.82, respectively.
The seroprevalence of Toxoplasma specific IgG was 30.1% among pregnant women at the first trimester of their pregnancy (n = 389). During the follow-up during pregnancy, we could reach 257 (66.1 %) women in the second trimester and 124 (31.9%) women in the third trimester. No Toxoplasma specific IgM in total 770 serum samples of three trimesters was detected.
Socio-demographic characteristics and obstetrical history are given in Table 1. Risk factors related to eating preferences and hygienic habits are given in Table 2.
Table 1 Toxoplasmosis and sociodemographic/obstetric factors in the study population
Risk factors Toxoplasma IgG
Negative Positive
n % n % p
Sociodemographic factors
Age (n = 356)
15–29 223 72.2 86 27.8 0.001
30–40 23 48.9 24 51.1
Education (n = 357)
Less than primary school 171 69.0 77 31.0 0.978
Secondary school and more 75 68.8 34 31.2
Occupation (n = 358)
Housewife 230 68.9 104 31.1 0.840
Other 17 70.8 7 29.2
Residence (n = 389)
Urban 218 68.6 100 31.4 0.213
Rural 54 76.1 17 23.9
Environmental factors
Drinking water (n = 358)
General network 136 67.0 67 33.0 0.042
Bottled 43 84.3 8 15.7
Other 69 66.3 110 30.7
Obstetric history
Abortion history (n = 357)
None 189 70.8 78 29.2 0.186
One or more 57 63.3 33 36.7
Table 2 Toxoplasmosis and risk factors related to eating preferences and hygienic habits.
Risk factors Toxoplasma IgG
Negative Positive
n % n % p
Frequency of meat consumption (n = 338)
Every day 12 54.5 10 45.5 0.099
A few times per week or less 225 71.2 91 28.8
Cooking preferences (n = 333)
Undercooked 38 66.7 19 33.3 0.707
Well-done 191 69.2 85 30.8
Eating raw meat* (n = 341)
Yes 32 68.9 13 28.9 0.767
No 204 68.9 92 31.1
Type of meat
Beef (n = 330)
Yes 181 68.8 82 31.2 0.493
No 49 73.1 18 26.9
Poultry(n = 352)
Yes 233 69.1 104 30.9 0.839
No 10 66.7 5 33.3
Game(n = 301)
Yes 27 84.4 5 15.6 0.080
No 187 69.5 82 30.5
Delicatessen (n = 340)
Yes 162 68.9 73 31.1 0.522
No 76 72.4 29 27.6
Milk and milk products(n = 355)
Yes 193 69.9 83 30.1 0.958
No 55 69.6 24 30.4
Eating raw vegetables
Yes 235 69.5 103 30.5 0.982
No 9 69.2 4 30.6
Washing kitchen utensils after cutting vegetables (n = 345)
Yes 241 70.5 101 29.5 0.886
No 2 66.7 1 33.3
Washing hands before meals (n = 351)
Yes 239 68.9 108 31.1 0.180
No 4 100.0 0 -
Eating outside of the home
Yes 141 66.8 70 33.2 0.245
No 106 72.6 40 27.4
Current cat ownership(n = 353)
Yes 5 62.5 3 37.5 0.695
No 238 69.9 107 31.0
• A Turkish delicacy made of spiced raw meat
Discussion
Seroprevalence of T.gondii infection range between 15%-77% in different countries [9,11]. In the current study, the seroprevalence of Toxoplasma specific IgG was 30.1% in pregnant women at the first trimester. In outpatient women at childbearing age it is found to be between 21.8% to 85% in our country [4,20-23]. The only population-based study that could be reached was performed in a village. In that study, the seroprevalence of toxoplasmosis was found to be 30% in 600 people aged 7–50 years [24]. Latter seroprevalence is similar with the current study.
As in the study of Bobic et al. (1998) that have found that prevalence increases as the age increases [9], age effect was observed in the current study. The reason might be increasing risk of exposure with age.
In the current study, no statistical meaningful difference was observed between urban and rural areas for toxoplasmosis seroprevalence. Baril et al.(1999) [6] have found similar results with the current study; however, Ades et al (1993) [25] have found higher seroprevalences in urban areas. There is a need to assess the nature of infection chain for Aydin.
Seroprevalence was found to be changed according to usage of different drinking water sources in the current study. The highest prevalence (33.7%) was in the people using general network water, followed by uncontrolled water sources (30.7%). The municipal network water in Aydın is collected to processing pools from open springs next to a few villages. The high seroprevalence in general network water users may be in accordance with the latest articles that have showed the presence of oocysts in chlorinated network water. A study is needed to investigate oocysts in domestic and outdoor life for those whose water sources are the same. Lower prevalences in bottled water users might be explained in that the containers are filled soon after the water surfaces. Oocyst form of T.gondii seems to be the major factor in the infection of water resources. Bowie et al (1997) assessed an outbreak in the western Canadian province of British Columbia and found that a municipal water system that used unfiltered, chloraminated surface water had been the likely cause of the large community-wide outbreak of toxoplasmosis [26].
Frequent consumption and type of meat (pig, sheep and goat) were identified as the principle risk factor in several recent studies of T.gondii infections in humans [6-9,27]. No relation was observed between seroprevalence and type of meat consumed in the current study. In Turkey, beef and lamb are commonly used, especially mixed together. Pork is never used due to religious ban. However, it is known that pork is mixed with other type of meat and served without knowledge of consumers.
The cooking temperature of meat is an important issue in the infection of T.gondii. Thorough cooking is always preferred in Turkey. However, raw meat is also consumed in traditional cuisine as "cig kofte" which is a mixture of raw meats. In the current study, no statistical difference was observed between T.gondii prevalence, and cooking preferences and consumption of raw meat.
The association of cats and human toxoplasmosis is difficult to assess by epidemiological surveys because soil, not the cats, is the main culprit. Oocysts are not found on cat fur [28] and are often buried in soil along with cat faeces, and soil contact is universal and difficult to avoid [29]. In Turkey, cats and dogs are commonly stray animals. There is always a possibility that oocysts of the stray cats may gain infectious ability outside the animal. In the current study, the number of cat owners was just a few, and no relation was detected.
In the previous studies, lower educational level, soil-related occupations [11], eating raw or unwashed vegetables or fruits, cleaning the cat litter box [8], having poor hand hygiene [6,8], eating raw vegetables outside the home [6], contact with soil [7], travelling outside of Europe, the United States and Canada [7] were all found as risk factors for toxoplasmosis. These factors were assessed in the current study; however, no relation was found.
The importance of the current study is that it is one of the few – to our knowledge only one -population based studies performed in Turkey. However, there are some restrictions in the study. First of all, dye test could not be used as a gold standard. In order to overcome this restriction, different test were used together. Secondly, although health centers as primary care units cover all the population which they are responsible for, and the data obtained from them is the most appropriate for population-based studies, there may be a few women that could not be reached. Maybe one of the most important restrictions is that the assessment of risk factors was done according to information given by the patient. Some issues such as hygienic behaviour might be misrepresented as higher because of shaming or other factors. However, an improvement in the quality of data was attempted by using district midwifes who were familiar and trusted by the women during data collection.
Conclusion
In conclusion, 69.9 % of pregnant women in the first trimester in the Aydin Province-Turkey, are susceptible to acute infection, and no Toxoplasma specific IgM was detected during the follow-up of pregnant women for three trimesters. This may be due to lower ratios of serum assessment in the second (66.1 %) and in the third (31.9%) trimesters. At the current moment, there is no legal obligation for education about means of minimizing exposure to T.gondii. in our country. Health authorities, especially primary health care givers should be sensitive to the importance of the issue. Hygiene of water supplies is also important. Future studies on possibility of contamination of network water by oocysts should be performed and used for prevention activities. The data found in the current study point out that due to a high risk of toxoplasmosis in pregnant women, a debate on developing a general screening program for toxoplasmosis in pregnancy in Turkey should be begun.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
SE planned the research; performed parasitic examinations and contributed discussing the results and writing manuscript. PO performed the sampling and statistical analyzes and contributed discussing the results and writing manuscript. MT participated in initial study design, and revised the manuscript. HY participated in initial study design, and revised the manuscript. All authors 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 two grants (No: TPF 01002 and No: TPF 02012) from Adnan Menderes University Research Fund. The researchers also thank to Health Directorate of Aydin for the field work.
==== Refs
Ashburn D Ho Yen DO, Joss AWL History and general epidemiology. Human toxoplasmosis 1992 3 Oxford University Press Chapter 56 76
Tenter AM Heckerroth AR Weiss LM Toxoplasma gondii: from animals to human Int J Parasitol 2000 30 1217 1258 11113252 10.1016/S0020-7519(00)00124-7
Gun H Tanyuksel M Haznedaroglu T Erdal N Gursoy HG Seropositivity of toxoplasmosis in health college students Acta Parasitologica Turcica 1993 17 15 19
Kara H Ozcan K Tanriverdi S Koltas S Demonstration of IgG and IgM antibodies against Toxoplasma in maternal blood, cord blood and amniotic fluid Acta Parasitologica Turcica 1999 23 115 118
Petersen E Pollak A Reiter-Owona I Recent trends in research on congenital toxoplasmosis Int J Parasitol 2001 31 115 144 11239933 10.1016/S0020-7519(00)00140-5
Baril L Ancelle T Goulet V Thulliez P Tirard-Fleury V Carme B Risk factors for Toxoplasma infection in pregnancy: a case-control study in France Scand J Infect Dis 1999 31 305 9 10482062 10.1080/00365549950163626
Cook AJ Gilbert RE Buffolano W Sources of Toxoplasma infection in pregnant women: a European multicentre case-control study Br Med J 2000 15 142 7 10894691
Kapperud G Jenum PA Stray-Pedersen B Melby KK Eskild A Eng J Risk factors for Toxoplasma gondii infection in pregnancy: results of a prospective case-control study in Norway Am J Epidemiol 1996 144 405 412 8712198
Bobic B Jevremovic I Marinkovic J Sibalic D Djurkovic-Djakovic O Risk factors for Toxoplasma infection in a reproductive age female population in the area of Belgrade, Yugoslavia Eur J Epidemiol 1998 14 605 10 9794128 10.1023/A:1007461225944
Buffolano W Gilbert RE Holland FJ Fratta D Palumbo F Ades AE Risk factors for recent Toxoplasma infection in pregnant women in Naples Epidemiol Infect 1996 116 347 51 8666080
Jones JL Kruszon-Moran D Wilson M McQuillan G Navin T McAuley JB Toxoplasma gondii infection in the United States: seroprevalence and risk factors Am J Epidemiol 2001 154 357 65 11495859 10.1093/aje/154.4.357
Choi WY Nam HW Kwak NH Huh W Kim YR Kang MW Cho SY Dubey JP Foodborne outbreaks of human toxoplasmosis J Infect Dis 1997 175 1280 1282 9129105
Benenson MW Takafuji ET Lemon SM Greenup RL Sulzer AJ Oocyst-transmitted toxoplasmosis associated with ingestion of contaminated water N Engl J Med 1982 307 666 9 7110216
Stagno S Dykes AC Amos CS Head RA Juranek DD Walls K An outbreak of toxoplasmosis linked to cats Pediatrics 1980 65 706 12 7189277
Sarı C Okyay P Ertug S Agreement of the commercial ELISA kit with IFA-ELISA kits prepared in our laboratory for detecting Toxoplasma specific IgG antibodies and comparison of the cost of the methods Acta Parasitologica Turcica 2003 27 1 3
Kuman HA Ertug S Ozensoy S Guruz AY Comparison of different commercial diagnostic kits in detection of anti-Toxoplasma IgM antibodies in the diagnosis of acute toxoplasmosis Acta Parasitologica Turcica 1999 23 227 232
Ertug S Uner A Aksoy U Gunduz C Guruz Y Correlation between ELISA, IFA and IHA techniques in the diagnosis of toxoplasmosis Acta Parasitologica Turcica 2000 24 4 8
Kuman HA Ertug S Bayram S Ertabaklar H The evaluation of antibodies against Toxoplasma gondii using ELISA, IFA and DA tests Acta Parasitologica Turcica 2000 24 333 336
Desmonts G Remington JS Direct agglutination test for diagnosis of Toxoplasma infection: Method for increasing sensitivity and specificity J Clin Microbiol 1980 11 562 568 7000807
Kaleli B Kaleli I Aktan E Akalın H Aksit F Toxoplasma IgG and IgM seropositivity in pregnancy Acta Parasitologica Turcica 1997 21 241 243
Demirci M Cicioglu Aridogan B Can R Kaya S Seroprevalans of toxoplasmosis in various groups in Isparta Acta Parasitologica Turcica 2001 25 107 109
Yazar S Altunoluk B Akman AA Şahin İ Investigation of anti-Toxoplasma gondii antibodies in women during pregnancy Acta Parasitologica Turcica 2000 24 343 345
Kocabeyoglu O Yergok YZ Emekdas G Koşan E Birinci I Diler H Toxoplasma IgG and IgM antibodies prevalence in pregnant women Acta Parasitologica Turcica 1996 20 149 153
Yolasigmaz A Sakru N Yazar S Akisu C Guruz AY Kuman HA Altintas N Investigation of anti-Toxoplasma antibodies in residence of urban and rural areas Acta Parasitologica Turcica 2003 27 81 84
Ades AE Parker S Gilbert R Tookey PA Berry T Hjelm M Wilcox AH Cubitt D Peckham CS Maternal prevalence of Toxoplasma antibody based on anonymous neonatal serosurvey: a geographical analysis Epidemiol Infect 1993 110 127 33 8432316
Bowie WR King AS Werker DH Isaac-Renton JL Bell A Eng SB Marion SA Outbreak of toxoplasmosis associated with municipal drinking water Lancet 1997 350 173 177 9250185 10.1016/S0140-6736(96)11105-3
Dubey JP Thulliez P Powell EC Toxoplasma gondii in Iowa sows: comparison of antibody titers to isolation of T.gondii by bioassays in mice and cats J Parasitol 1995 81 48 53 7876977
Dubey JP Duration of immunity to shedding of Toxoplasma gondii oocysts by cats J Parasitol 1995 81 410 415 7776126
Dubey JP Sources of Toxoplasma gondii infection in pregnancy BMJ 2000 321 127 128 10894674 10.1136/bmj.321.7254.127
| 15958156 | PMC1177966 | CC BY | 2021-01-04 16:28:56 | no | BMC Public Health. 2005 Jun 15; 5:66 | utf-8 | BMC Public Health | 2,005 | 10.1186/1471-2458-5-66 | oa_comm |
==== Front
BMC Public HealthBMC Public Health1471-2458BioMed Central London 1471-2458-5-671595817010.1186/1471-2458-5-67Research ArticleA prospective study of cumulative job stress in relation to mental health Godin Isabelle [email protected] France [email protected] Yves [email protected] Johannes [email protected] School of Public Health Université Libre de Bruxelles, HealthPsychology Unit CP 596, 808 Route de Lennik, 1070 Brussels, Belgium2 School of Public Health Université Libre de Bruxelles, Epidemiology Department, 808 Route de Lennik, 1070 Brussels, Belgium3 Department of Medical Sociology, University of Duesseldorf, Universitaetsstr. 1, 40225 Duesseldorf, Germany2005 15 6 2005 5 67 67 19 1 2005 15 6 2005 Copyright © 2005 Godin et al; licensee BioMed Central Ltd.2005Godin 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 study tests associations between psychosocial stress at work measured by the effort-reward imbalance model in a dynamic perspective, and multiple indicators of poor mental health, in a prospective design.
Methods
1986 male and female employees from four Belgian enterprises were followed-up over one year within the framework of the Somstress study. Based on two consecutive measurements, an index of cumulative job stress was constructed and its associations with five indicators of mental health were studied, excluding caseness at entry (for depression, anxiety, somatisation, chronic fatigue and psychotropic drug consumption respectively). Taking into account the longitudinal design, four categories of job stress are defined: 1) employees free from stress at both measures, 2) job stress present at first measure but not at the second one, 3) recent onset of job stress as evidenced by second measure 4) workers exposed to stress at both measures. Multivariate logistic regression with appropriate adjustments was applied.
Results
In bivariate analysis, a clear graded association of cumulative job stress with all five mental health indicators is observed, both in men and women. In multivariate logistic regression analysis, recent onset of stress is strongly associated with poor mental health among men (odds ratios ranging from 1.8 to 4.6), while cumulative stress shows strongest effects on mental health in women (odds ratios ranging from 1.4 to 7.1).
Conclusion
Cumulative experience and recent onset of job stress in terms of high effort spent and low reward received is associated with elevated risk of all five indicators of poor mental health at follow-up in a large cohort of employees.
==== Body
Background
Instability of employment, rapid change of demands and intensification of work pressure are widely prevalent consequences of economic globalization and technological change [1]. Even in established sectors of industrial production, administration and services of advanced societies experiences of downsizing, mergers and outsourcing are increasingly shared by employees [2]. Surveys of working conditions in Europe indicate that stressful experience recently increased in the European workforce although variations between countries and sectors are observed [3]. Chronic stressful experience at work can adversely affect physical and mental health. This has been documented in a large number of epidemiological studies based mainly on two complementary theoretical concepts, the demand-control model [4-6] and the effort-reward imbalance model [[7,8]; see also [9-12]]. The demand-control model posits that jobs characterised by high quantitative demands in combination with low decision latitude adversely affect health. The focus of the effort-reward imbalance model is put on contractual non-reciprocity where high efforts at work are not met by adequate rewards in terms of money, esteem, promotion prospects and job security.
Both models have been tested in the frame of prospective epidemiological studies, but in a majority of these investigations the measurement of exposure was restricted to baseline assessment. Thus, effects of cumulative job stress on incident disease have not been sufficiently explored so far. However, with regard to the demand-control model, there are a few important exceptions indicating, firstly, that recurrent job stress is indeed associated with elevated risk of ill health [13,14], morbidity [15,16], and mortality [17], and, secondly, that reduction of job stress over time results in improved health [18]. No comparable data on health effects of cumulative stress are available from the effort-reward imbalance model. This is particularly critical as this model documents close links with macroeconomic changes, suggesting that downsizing, mergers and outsourcing result in increases of effort-reward imbalance [19] and, thus, may indirectly affect health.
This study analyses the dynamics of stressful work experience over time, based on the effort-reward imbalance model, in relation to mental health, using longitudinal data of a large cohort. We test the hypothesis that the risks of poor mental health after one year are higher among employees who either continuously experience high job stress or who experience an increase in job stress from the first to the second measurement, compared to the remaining employees with either continuously low levels or decreasing levels of job stress over time. Both conditions, continuous exposure and incident exposure to job stress, are more likely to occur under conditions of downsizing and related macroeconomic constraints.
For two additional reasons, the effort-reward imbalance model is chosen to measure stressful experience at work. First, as mentioned, two of the three reward components specified in this model provide a direct link with the labour market dynamics that are becoming increasingly relevant in a globalized economy: promotion prospects including job security and level of salary or wage [20]. In addition the model is composed by an extrinsic component (perceived demands, perceived rewards) and an intrinsic component (coping with demands at work; overcommitment as a motivational risk factor), thus allowing for a differentiation of 'subjectively perceived' situational and personal characteristics of stressful experience at work. Secondly, the standard measurement of this model, a Likert-scaled self-administered questionnaire (see Methods) was shown to be highly sensitive to change of exposure over time [21]. This is an essential prerequisite of studying dynamics of stressful experience at work in a reliable way.
Methods
Study design
Somstress is a Belgian research project based on a prospective protocol with a repeated measure of exactly the same self-administered questionnaire in a one year time interval. Its main purpose is to assess mental health and psychological well-being in relation to working conditions by combining organizational, psychosocial and behavioral data. Depression, anxiety, somatisation, chronic fatigue and, as a more objective correlate, psychotropic drug consumption, were chosen as indicators of mental health according to the research protocol. A previous publication restricted to cross-sectional baseline data already documented associations of psychosocial stress at work with these indicators of mental health [22]. Yet, the current report is the first to analyse these associations prospectively.
Study sample
Four enterprises were selected according to their economic stability in order to include contextual variation into the study design. To this end, a special index of economic instability was constructed (for description see [23,24]). This index reflects the extent and change over the past three years in the employment and unemployment rates in each one of a broad spectrum of economic activities, based on a national coding system. A larger number of companies was evaluated accordingly. In the final selection process, four enterprises were included. The four selected companies differ gradually on this variable, ranging from a very stable company (Firm 1) to a very unstable company (Firm 4). All four companies belong to the private or public service sector where the majority of employees are white collars.
At time 1 (T1, 2000) as at time 2 (T2, 2001), all workers of the four enterprises were invited to participate in the study. Participation was voluntary. At baseline, 9634 questionnaires were sent out, corresponding to the total number of workers, and 3804 were returned (global participation rate 40 per cent). A similar participation rate was achieved at follow up (T2) with 2709 questionnaires returned (global participation rate 37 per cent). However, full data from identical subjects obtained from both surveys was restricted to 1986 employees (paired sample). Compared to other studies, this participation rate is relatively low, and could be partially explained by the uncertainty and the feeling of threat induced by the possibility of merging and downsizing in several workplaces.
At the end of the study, 3 different combinations of samples of workers can be analyzed: all participants to the first measure (N = 3804), all participants to the second measure (N = 2707), and participants to both measures (N = 1986). In this paper, it is this last sample (paired sample) that will be analyzed because of its prospective properties and therefore the possibility of testing the study hypothesis. This sub-sample is well comparable with the larger sample (N = 3804 and 2709 respectively) in terms of major socio-demographic characteristics, such as age, sex, educational level and professional qualification.
Participants to the first measure only are very similar to the participants to both measures (paired sample). In other terms, those who were lost during the follow up do not differ in terms of socio-economic, demographic conditions or health status (self-rated health). We can therefore exclude a bias due to selective attrition. Moreover, both populations are representative of the whole population of workers, in each enterprise, for the available criteria: gender, age, occupation and department or service.
Data collection
The fully standardized questionnaires contain data on socio-demographic characteristics of the respondent, on psychosocial stress at work and on indicators of mental health and psychological well-being (at T1 and T2). The same questionnaire was submitted at both measures (T1 and T2).
Effort-reward imbalance at work was measured by the original questionnaire [20] containing the three scales 'effort' (5 Likert scaled items; the 6th item measuring physical load was omitted as there were mainly white collars), 'reward' (11 Likert scaled items with three subscales 'esteem', 'salary and promotion prospects' and 'job security') and 'overcommitment' (short version with 6 Likert scaled items defining a one-dimensional scale). A score reflecting the extent of imbalance was constructed by a ratio of the two scales 'effort' (nominator) and 'reward' (denominator, adjusted for unequal number of items by a correction factor). In this study, as in other reports (e.g. [25]), the upper quartile of the distribution of the ratio defines the risk condition of chronic psychosocial stress at work. Similarly, a group at risk in terms of the intrinsic component of the model, overcommitment, was defined by scores in the upper tertile of the respective scale, according to the established procedure [8,20].
In order to evaluate the dynamics of stressful experience at work over time, the sample was divided into four groups based on values of a summary variable, the ratio between effort and reward scores: group 1 (the reference group) was composed of employees who were free from job stress at either occasion (scores on the ratio were lower than those in the upper quartile); in group 2, job stress was present at first, but no longer at second measurement; conversely, group 3 was characterized by an absence of job stress at first, but a demonstration of it at second measurement; finally, group 4 was composed of employees who continuously reported a high level of job stress at either occasion (for group description see Table 2). It is important to note that about 25 percent of the sample are considered at risk (group 3 and 4) in terms of our research hypothesis.
Table 2 Sociodemographic indicators of the four job stress groups
Effort-reward imbalance at work (ERI) N (%)
T1 no-T2 no T1 yes-T2no T1 no-T2 yes T1 yes-T2 yes
Sex (ns)
Men 672 (64.7) 108 (10.4) 91 (8.8) 168 (16.2)
Women 561 (65.5) 88 (10.3) 95 (11.1) 112 (13.1)
Education (**)
Lowest (vocational school or less) 416 (67.4) 60 (9.7) 43 (7.0) 98 (15.9)
Secondary school 222 (67.9) 31 (9.5) 24 (7.3) 50 (15.3)
College 409 (62.5) 81 (12.4) 83 (12.7) 81 (12.4)
Highest (university) 172 (63.9) 19 (7.1) 32 (11.9) 46 (17.1)
Age (ns)
18–34 yrs 309 (67.9) 52 (11.4) 45 (9.9) 49 (10.8)
35–49 yrs 745 (63.7) 116 (9.9) 120 (10.3) 189 (16.2)
50 yrs and + 157 (66.2) 22 (9.3) 19 (8.0) 39 (16.5)
* p < 0.05, ** p < 0.01, *** p < 0.001
In addition to this model, job dissatisfaction (5 items) and threat perceived from global economy (3 items) were assessed using respective items from the Job Content Questionnaire [26].
We used validated mental health measures derived from the Symptom Check List SCL90 [27] for (1) depression (16 items), (2) anxiety (10 items) and (3) somatisation (12 items). Each one of these mental health indicators represents a distinct, psychometrically tested scale [22,27].
Chronic fatigue was included as a further indicator of impaired mental health as assessed by a 4-item-scale that was developed in a Dutch study [28]. Internal consistencies of the SCL90 scales give the following Cronbach's alpha values: 0.93 (T1 and T2) for depression, 0.86 (T1) and 0.89 (T2) for anxiety, and 0.86 (T1) and 0.87 (T2) for somatisation. For chronic fatigue, the respective value is 0.86.
In addition, a more objective correlate of mental health problems, amount of psychotropic drug consumption was assessed by a scale summarizing type and frequency of the consumption of tranquillizers, antidepressants and/or sleeping tablets during the last 4 weeks [29].
The description of the SCL90 depression variable gives a range of 58 (min. 16, max. 74) and a mean score of 24.2 (9.9 st. dev.). Those values are for anxiety: range 35 (min. 10, max. 45), mean 14.6 (5.8 st. dev.) and for somatisation: range 45 (min. 12, max. 57), mean 18.9 (7.0 st. dev.). For chronic fatigue, the range is 24 (min. 4, max. 28), and the mean 14.8 (6.7 st. dev.). Intercorrelations T1-T2 for the SCL90 scores give r2 values of 45 per cent for depression, 40 per cent for somatisation and 48 per cent for anxiety.
Due to non-normal distribution of core variables we applied logistic instead of linear regression analysis. In order to identify groups at risk of impaired mental health, all mental health indicators were dichotomized at the upper quartile of each score distribution [22]. While we are aware of the loss of information due to dichotomization of core variables, we nevertheless maintain that the statistical models applied and the large sample size may give us conservative estimates of the hypothesized associations.
Statistical analysis
Data analysis is mainly based on the longitudinal aspects, i.e. on the sample of 1986 participants with complete data from both surveys. We apply Mc Nemar tests for paired sample in order to test the evolution between the two measures (univariate analysis). Identification of predictors of mental health problems at T2 is done by logistic regression analysis.
The two components of the effort-reward imbalance model (effort-reward ratio and overcommitment) are introduced separately in order to assess their relative contribution to the estimation of mental health problems. In this study, further refined analyses are not conducted, such as a test of possible effect modification (e.g. degree of job instability, socio-economic status) in order not to loose statistical power. Rather, these variables, together with age, job dissatisfaction and threat from global economy are introduced as confounders into the multivariate analysis. All models are calculated separately for men and women and for the five mental health indicators. Interactions between independent variables were tested and a significant interaction between stress and gender was found, implicating separate analysis for gender.
Because people and work characteristics vary quite a lot across workplaces, this latter variable could play a confounding role, and, as such, was introduced in the multivariate analysis.
To measure incident mental health problems (onset of new cases), individuals presenting at T1 the mental health problem studied at T2 were excluded from the analysis. This was done for obvious methodological reasons although the overall sample size was reduced by some 25 per cent. As a consequence, about 75 per cent of the total paired sample will be included in each logistic regression model.
Results
Socio-demographic, psychosocial and health-related characteristics of the sample are shown in Table 1, separately for the four enterprises. As specified above, the different combinations of sub-samples (all participants to T1, all participants to T2 and the paired sample, i.e. participants to both measures) do not differ in terms of sex, age, education, professional qualification as they are very similar in their mental health indicators.
Table 1 Sample description (paired sample, N = 1986)
Enterprise 1 N (%) Enterprise 2 N (%) Enterprise 3 N (%) Enterprise 4 N (%) Total N (%)
Sex (***)
Men 131 (23.9) 206 (42.8) 200 (68.5) 529 (79.7) 1066 (53.7)
Women 418 (76.1) 275 (47.2) 92 (31.5) 135 (20.3) 920 (46.3)
Education (***)
Lowest (vocational school or less) 93 (17.0) 160 (33.7) 61 (21.0) 412 (64.1) 726 (37.1)
Secondary school 35 (6.4) 58 (12.2) 58 (19.9) 121 (18.8) 272 (13.9)
College 277 (50.5) 174 (36.6) 125 (43.0) 105 (16.3) 681 (34.8)
Highest (university) 143 (26.1) 83 (17.5) 47 (16.2) 5 (0.8) 278 (14.2)
Age(***)
Mean (standard dev.) 38.9 (8.14) 39.1 (8.75) 39.6 (9.0) 43.2 (7.2) 40.5 (8.4)
18–34 yrs 162 (29.5) 150 (31.2) 92 (31.5) 69 (11.0) 473 (24.2)
35–49 yrs 338 (61.6) 279 (58.0) 154 (52.7) 457 (72.7) 1228 (62.9)
50 and + 49 (8.9) 52 (10.8) 46 (15.8) 103 (16.4) 250 (12.8)
Effort-reward imbalance (*)
T1 no, T2 no 350 (68.2) 289 (63.2) 170 (60.3) 424 (65.9) 1233 (65.1)
T1 yes, T2 no 50 (25.5) 46 (10.1) 39 (13.8) 61 (9.5) 196 (10.3)
T1 no, T2 yes 56 (10.9) 55 (12.0) 27 (9.6) 48 (7.5) 186 (9.8)
T1 yes, T2 yes 57 (11.1) 67 (14.7) 46 (16.3) 110 (17.1) 280 (14.8)
Overcommitment (high) (**) 164 (30.1) 171 (35.9) 108 (37.2) 272 (41.1) 715 (36.3)
Health (highest quartiles)
Depression (n.s.) 141 (26.0) 105 (22.0) 59 (20.3) 174 (26.5) 479 (24.3)
Anxiety (*) 125 (20.3) 127 (26.6) 55 (18.9) 178 (27.1) 485 (24.6)
Somatisation (**) 149 (27.4) 131 (27.3) 55 (18.9) 194 (29.6) 529 (26.9)
Chronic fatigue 141 (26.6) 132 (28.3) 74 (26.1) 183 (28.4) 530 (27.5)
Psychotropic drug consumption 119 (22.8) 91 (19.6) 43 (15.5) 162 (25.6) 415 (21.9)
* p < 0.05, ** p < 0.01, *** p < 0.001
While mean age of the sample is about 40 years and while about half of the sample is composed of women, we observe a considerable variation for those variables across the four enterprises. In particular, participants in enterprise 4 are significantly older and there are less women. In addition, their educational degree is also lower, and the rate of employees with poor health indicators is remarkably high compared to the other workplaces. These differences are partly explained by the fact that this latter enterprise is a telecommunication company which hires mainly (male) technicians and workers with low degree of qualification.
As mentioned, the dynamics of job stress are explored by defining four subgroups. The proportions of employees in each category of this summary variable are indicated in Table 2, together with socio-demographic characteristics. Besides education, a proxy measure of socioeconomic status, the groups did not differ with regard to these characteristics.
Relationships between the four job stress categories and mental health indicators show a steep gradient for depression, anxiety and somatisation, both in men and women (Figures 1 and 2). Less clear patterns were observed for chronic fatigue and psychotropic drug consumption (data not shown). Men and women reporting continuous or incident job stress during the observation period show higher proportions of mental health problems compared to those with low or decreasing job stress. Interestingly, the prevalence varies from about ten per cent in the group without stressful experience at work to about fifty per cent in the continuously stressed group, particularly among men.
Figure 1 Dynamics of stressful experience at work and prevalence of mental health (SCL90) at T2 (men). For description of categories see text.
Figure 2 Dynamics of stressful experience at work and prevalence of mental health (SCL90) at T2 (women). For description of categories see text.
As a final step, multivariate analysis is conducted to test the research hypothesis.
As indicated, logistic regression analysis is performed separately for men and women, with adjustment for age, education, threat from global economy, job dissatisfaction and work place instability. The dynamics of effort-reward imbalance at work over time are related to risks of poor mental health as suggested by our hypothesis. We do not observe a single effect in the group of employees who experienced a decrease in stressful work over time (group 2). In contrast, almost all odds ratios of mental health problems are significantly elevated in the two groups with newly emerging or continuously high job stress. In only 3 cases out of 20, odds ratios do not reach significance. There are interesting differences between gender. For men, recent onset of stressful experience at work is associated with a relatively highest risk of poor mental health, whereas for women relatively highest risks are observed in the group with continuously high job stress level for four indicators out of five. Odds ratios range between 1.4 and 7.1 for women and between 1.8 and 4.6 for men. When comparing the extrinsic versus the intrinsic component of the effort-reward imbalance model, it is obvious that associations of overcommitment with mental health are less consistent and generally weaker than those observed with the extrinsic component.
Discussion
This study documents consistent associations of stressful experience at work over time with newly emerging mental health problems, using five mental health indicators. Employees with continuous job stress over a one year observation period and those with recently evolving job stress were at higher risk of developing poor mental health compared to the remaining groups, after exclusion of men and women with manifest mental problems at the study onset. One employee out of four in this large cohort belongs to one of the two groups with critical experience of job stress where, overall, a threefold elevated risk of incident poor mental health is observed. As the measurement of stressful experience at work is sensitive to change over time, it is unlikely that the observed systematic differences are spurious. Moreover, results remain true for men and women and after adjustment for relevant confounders, including the level of contextual job instability. While the gradient of mental health according to job stress is similar between men and women in bivariate analysis (Figure 1 and 2), an interesting gender difference results from multivariate analysis. Men are more reactive to a recent stress exposure, whereas women are more responsive to cumulative job stress (Table 3). This new finding needs further confirmation before being interpreted in a broader context.
Table 3 Poor mental health at T2 in relation to dynamics of stressful experience at work Multivariate logistic regression analysis (odds ratios (OR), 95% confidence intervals (CI))
men (N = 836): OR$ (95% CI)
depression anxiety somatisation chronic fatigue psychotropic
Effort-reward imbalance1 *** ** ** ** **
T1 yes-T2 no 1.2 (0.5–2.9) 0.8 (0.3–2.1) 1.5 (0.7–3.4) 1.4 (0.7–3.1) 1.5 (0.6–4.1)
T1 no-T2 yes 4.6 (2.3–9.2) 3.7 (1.7–7.8) 4.1 (2.0–8.5) 3.4 (1.7–6.7) 3.2 (1.5–7.0)
T1 yes-T2 yes 2.8 (1.3–5.7) 2.3 (1.1–4.8) 2.0 (0.9–4.4) 1.8 (0.9–3.6) 3.4 (1.5–7.7)
Overcommitment ** ** * n.s. n.s.
yes 2.4 (1.4–4.1) 2.5 (1.5–4.4) 1.8 (1.1–3.1) 1.2 (0.8–2.0) 1.1 (0.6–2.1)
women (N = 700): OR$ (95% CI)
depression anxiety somatisation chronic fatigue psychotropic
Effort-reward imbalance1 *** ** ** *** *
T1 yes-T2 no 1.3 (0.5–3.2) 1.1 (0.4–3.1) 1.6 (0.5–4.5) 1.2 (0.5–3.1) 1.0 (0.3–3.0)
T1 no- T2 yes 3.2 (1.6–6.4) 2.3 (1.1–4.8) 3.5 (1.7–7.2) 2.0 (0.9–4.1) 2.7 (1.3–5.6)
T 1 yes-T2 yes 4.6 (2.3–9.0) 4.5 (2.1–9.8) 3.6 (1.6–8.2) 7.1 (3.4–14.5) 1.4 (0.5–3.5)
Overcommitment * n.s. n.s. n.s. n.s
yes 1.8 (1.0–3.0) 1.6 (0.9–2.9) 0.8 (0.4–1.4) 1.1 (0.6–2.0) 1.4 (0.7–2.6)
1reference group: effort-reward imbalance T1 0, T2 0 : OR = 1.0
$ adjusted for age, education, threat from global economy, job dissatisfaction, workplace instability
* p < 0.05, ** p < 0.01, *** p < 0.001
Despite this evidence, this study suffers from several limitations. First, although employees with manifest mental health problems at T1 were excluded from multivariate analysis, we cannot rule out the possibility that a deterioration of mental health during the observation period has affected the measurement of job stress at T2. Secondly, even with a prospective protocol, two measurement waves only were conducted, and time of exposure was limited. However, structural and organizational changes occurred in all four enterprises under study even during this short observation period, thus reflecting the accelerated dynamics of work-related stress in current economic situation.
Thirdly, as the five indicators of mental health were not independent, some of the reported effects in multivariate analysis may have been overestimated. On the other hand, excluding employees with a respective mental health problem at study onset results in a conservative estimate, as this group may have suffered from previous job stress and that controlling for 'caseness' attenuates the associations under study. A further limitation concerns our decision of categorizing both the predicting and criterion variables, thus loosing information available from continuous data. Linear regressions performed with the normally distributed dependent variables did not yield different results. As this is not a clinical study, the decision of using scores in the upper quartile of the distribution of mental health indicators seems justified. Similarly, using the upper quartile of the ratio between effort and reward as a measure of stressful experience at work is in line with the evidence from several epidemiological studies although analysis of a log-transformed continuous ratio might reveal even stronger effects [25,30,31]. The same remains true with regard to a further test of main effects and interaction terms of the variables 'effort' and 'reward' which formed the basis for the ratio. Studies repeatedly revealed that the effect size of the ratio exceeds the effect sizes of the single variables [8,30,31]. Finally, we cannot exclude a 'healthy worker' effect of the final sample with full data (although this effect would point to a conservative estimate of the observed effect), and we cannot solve the methodological problem of reporting bias as both types of variables, the measure of stressful experience at work and the measures of mental health, were based on self-reported data.
Nevertheless, this study has remarkable strengths. It tests a theoretical model, effort-reward imbalance, that captures some of the core aspects of dynamics of stressful experience at work in a globalized, rapidly changing economy. Moreover, this model was shown to predict a variety of health problems in prospective observational studies in different occupational groups of several countries. Health outcomes include coronary heart disease [25,32], cardiovascular mortality [9] mild-to-moderate psychiatric disorder (mainly affective disorder) [33], alcohol dependence [34], type 2-diabetes [35] and poor self-rated health or poor mental and physical functioning [30,36].
This prospective evidence is supplemented by findings from cross-sectional studies testing associations of effort-reward imbalance at work with mental health, in particular depression [31,37,38]. Having controlled for different socio-demographic and workplace characteristics permits to better identify the evolution of the stress component in a one-year interval and its effect on the worker's mental health. Obviously, the evolution of the stress component in terms of effort-reward imbalance reflects some of the organizational changes in the enterprises.
The validity of reported results is further supported by two observations. First, in a cross-sectional analysis of the baseline data from the Somstress study, effort-reward imbalance at work was found to be associated not only with mental health indicators, but also with sickness absence, an indicator that is less vulnerable to reporting bias [22]. Secondly, the psychometric properties of the Belgian (French) version of the effort-reward imbalance questionnaire were tested in a comparative methodological data analysis of five different European countries. The values of internal consistency, discriminant validity of the scales and goodness of fit of the three model components (effort, reward, overcommitment) were well comparable across the five data sets and met psychometric criteria with satisfactory degree in all samples [20]. Finally, to our knowledge, this is the first report on psychosocial occupational health research with reference to the effort-reward imbalance model that explores the dynamics of stressful experience at work over time in a large sample of male and female employees from enterprises with different degrees of economic stability, using a variety of established mental health indicators.
Conclusion
In conclusion, continuous experience of stress at work over time and recently emerging job stress experience are both associated with elevated risk of mental health problems. Results underline the importance of studying health effects of a globalized, rapidly changing economy and of developing appropriate measures to reduce the respective burden of disease in working populations.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
IG: study design, data collection and manuscript writing
FK: study design, data collection and manuscript writing
YC: manuscript writing
JS: creation of the measurement instrument, manuscript writing
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
This Somstress project is funded by the Belgian Federal Office for Scientific, Technical and Cultural Affairs.
==== Refs
Landsbergis P The changing organization of work and the safety and health of working people: a commentary J Occup Environ Med 2003 45 61 72 12553180
International Labor Organisation Bureau for Workers' Activities Labour Market Trends and Globalization's Impact on Them 2004 Geneva: International Labor Organisation
Merllié D Paoli P Dix ans de conditions de travail dans l'Union européenne 2000 Dublin: Fondation Européenne pour l'amélioration des conditions de vie et de travail
Belkic K Landsbergis P Schnall PL Baker D Is job strain a major source of cardiovascular disease risk? Scand J Work Environ Health 2004 30 85 128 15127782
Hemingway H Marmot M Evidence based cardiology: psychosocial factors in the aetiology and prognosis of coronary heart disease. Systematic review of prospective cohort studies BMJ 1999 318 1460 1467 10346775
Schnall P Belkic K Landsbergis P Baker D (Eds) The workplace and cardiovascular disease Occup Med: State of the Arts Reviews 2000 15 1 334
Marmot M Theorell T Siegrist J Stansfeld SA, Marmot MG Work and coronary heart disease Stress and the Heart 2002 London: BMJ Books 50 71
van Vegchel N de Jonge J Bosma H Schaufeli W Reviewing the effort-reward imbalance model: drawing up the balance of 45 empirical studies Soc Sci Med 2005 60 1117 1131 15589679 10.1016/j.socscimed.2004.06.043
Kivimäki M Leino-Arjas P Luukkonen R Riihimäki H Vahtera J Kirjonen J Work stress and risk of cardiovacular mortality: prospective cohort study of industrial employees BMJ 2002 325 857 860 12386034 10.1136/bmj.325.7369.857
Ferrie J Shipley M Stansfeld SA Marmot M Effects of chronic job insecurity and change in job security on self reported health, minor psychiatric morbidity, physiological measures, and health related behaviours in British civil servants: the Whitehall II study J Epidemiol Community Health 2002 56 450 454 12011203 10.1136/jech.56.6.450
Westerlund H Ferrie J Hagberg J Jeding K Oxenstierna G Theorell T Workplace expansion, long-term sickness absence, and hospital admission Lancet 2004 363 1193 1197 15081652 10.1016/S0140-6736(04)15949-7
Kivimäki M Ferrie J Head J Shipley M Vahtera J Marmot M Organisational justice and change in justice as predictors of employee health: the Whitehall II study J Epidemiol Community Health 2004 58 931 937 15483310 10.1136/jech.2003.019026
Bourbonnais R Comeau M Vezina M Job strain and evolution of mental health among nurses J Occup Health Psychol 1999 4 95 107 10212863 10.1037//1076-8998.4.2.95
de Lange AH Taris TW Kompier M Houtman I Bongers P Effects of stable and changing demand-control histories on worker health Scand J Work Environ Health 2002 28 94 108 12019593
Johnson J Stewart W Hall E Fredlund P Theorell T Long-term psychosocial work environment and cardiovascular mortality among Swedish men Am J Public Health 1996 86 324 331 8604756
Schwartz J Imputation of job characteristic scores Occup Med, State of the Art Reviews 2000 15 172 175
Amick B McDonough P Chang H Rogers WH Pieper C Duncan G Relationship between all-cause mortality and cumulative working life course psychosocial and physical exposures in the United States labor market from 1968 to 1992 Psychosom Med 2002 64 370 381 12021412
Schnall P Schwartz J Landsbergis P Warren K Pickering T A longitudinal study of job strain and ambulatory blood pressure: results from a three-year follow-up Psychosom Med 1998 60 697 706 9847028
Dragano N Verde PE Siegrist J Organizational downsizing and work stress: testing synergistic health effects in employed men and women J Epidemiol Community Health
Siegrist J Starke D Chandola T Godin I Marmot M Niedhammer I Peter R The measurement of effort-reward imbalance at work: European comparisons Soc Sci Med 2004 58 1483 1499 14759692 10.1016/S0277-9536(03)00351-4
Tsutsumi A Nagami M Morimoto K Matoba T Responsiveness of measures in the effort-reward imbalance questionnaire to organizational changes: a validation study J Psychosom Research 2002 52 249 256 11943243 10.1016/S0022-3999(02)00291-X
Godin I Kittel F Differential economic stability and psychosocial stress at work: associations with psychosomatic complaints and absenteeism Soc Sci Med 2004 58 1543 1553 14759697 10.1016/S0277-9536(03)00345-9
Godin I Desmarez P Kittel F Stress and job precarity: classification index of economic sectors Arch Public Health 2002 60 143 151
Godin I Bullying, workers' health and labour instability J Epidemiol Community Health 2004 58 258 259 14966242 10.1136/jech.2003.009456
Kuper H Singh-Manoux A Siegrist J Marmot M When reciprocity fails: effort-reward imbalance in relation to CHD and health functionning within the Whitehall II study Occup Environ Med 2002 59 777 784 12409537 10.1136/oem.59.11.777
Karasek R Brisson C Kawakami N Houtman I Bongers P Amick B The Job Content Questionnaire (JCQ): an instrument for internationally comparative assessments of psychosocial job characteristics J Occup Health Psychol 1998 3 322 355 9805280 10.1037//1076-8998.3.4.322
Derogatis L Administration, Scoring and Procedures Manual -I for the R(evised) version 1977 Baltimore: John Hopkins University School of Medicine
Alberts M Smets E Vercoulen J Garssen B Bleijenberg G 'Verkorte vermoeidheidsvragenlijst': een praktisch hulpmiddel bij het scoren van vermoeidheid Ned Tijdschr Geneeskd 1997 141 1526 1530 9543741
Pelfrene E Vlerick P Kittel F Mak R de Backer G Kornitzer M Psychosocial work environment and psychological well-bein: Assessment of the buffering effects in the job-demand-control(-support) model in Belstress Stress and Health 2002 18 43 56 10.1002/smi.920
Niedhammer I Tek ML Starke D Siegrist J Effort-reward imbalance model and self-reported health: cross-sectional and prospective findings from the GAZEL cohort Soc Sci Med 2004 58 1531 1541 14759696 10.1016/S0277-9536(03)00346-0
Pikhart H Bobak M Pajak A Malyutina S Kubinova R Topor R Sebakova H Nikitin Y Marmot M Psychosocial factors at work and depression in three countries of Central and Eastern Europe Soc Sci Med 2004 58 1475 1482 14759691 10.1016/S0277-9536(03)00350-2
Bosma H Peter R Siegrist J Marmot M Two alternative job stress models and the risk of coronary heart disease Am J Public Health 1998 88 68 74 9584036
Stansfeld SA Fuhrer R Shipley M Marmot M Work characteristics predict psychiatric disorder: prospective results from the Whitehall II Study Occup Environ Med 1999 56 302 307 10472303
Head J Stansfeld SA Siegrist J The psychosocial work environment and alcohol dependence: a prospective study Occup Environ Med 2004 61 219 224 14985516 10.1136/oem.2002.005256
Kumari M Head J Marmot M Prospective study of social and other risk factors for incidence of type 2 diabetes in the Whitehall II study Arch Intern Med 2004 164 1873 1880 15451762 10.1001/archinte.164.17.1873
Stansfeld S Bosma H Hemingway H Marmot M Psychosocial work characteristics and social support as predictors of SF-36 health functioning: the Whitehall II study Psychosom Med 1998 60 247 255 9625210
Larisch M Joksimovic L von dem Knesebeck O Starke D Siegrist J Effort-reward imbalance at work and depressive symptoms-a cross-sectional investigation of middle-aged employees Psychother Psychosom Med Psychol 2003 53 223 228 12709890 10.1055/s-2003-38867
Tsutsumi A Kayaba K Theorell T Siegrist J Association between job stress and depression among Japanese employees threatened by job loss in a comparison between two complementary job-stress models Scand J Work Environ Health 2001 27 146 153 11409598
| 15958170 | PMC1177967 | CC BY | 2021-01-04 16:28:56 | no | BMC Public Health. 2005 Jun 15; 5:67 | utf-8 | BMC Public Health | 2,005 | 10.1186/1471-2458-5-67 | oa_comm |
==== Front
BMC Public HealthBMC Public Health1471-2458BioMed Central London 1471-2458-5-681596085610.1186/1471-2458-5-68Research ArticleThe incidence of varicella and herpes zoster in Massachusetts as measured by the Behavioral Risk Factor Surveillance System (BRFSS) during a period of increasing varicella vaccine coverage, 1998–2003 Yih W Katherine [email protected] Daniel R [email protected] Susan M [email protected] Aisha O [email protected] Zi [email protected] Karen M [email protected] Jane F [email protected] Department of Ambulatory Care and Prevention, Harvard Medical School and Harvard Pilgrim Health Care, Boston, USA2 Department of Epidemiology, Boston University School of Public Health, Boston, USA3 Division of Epidemiology and Immunization, Bureau of Communicable Disease Control, Massachusetts Department of Public Health, Boston, USA4 Health Investigation Branch, Division of Health Studies, Agency for Toxic Substance and Disease Registry, Centers for Disease Control and Prevention, Atlanta, USA5 Health Survey Program; Center for Health Information, Statistics, Research and Evaluation; Massachusetts Department of Public Health; Boston, USA6 Applied Statistics, Evaluation and Technical Services; Bureau of Family and Community Health; Massachusetts Department of Public Health; Boston, USA7 Viral Vaccine-Preventable Disease Branch, Epidemiology and Surveillance Division, National Immunization Program, Centers for Disease Control and Prevention, Atlanta, USA2005 16 6 2005 5 68 68 16 2 2005 16 6 2005 Copyright © 2005 Yih et al; licensee BioMed Central Ltd.2005Yih 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 authors sought to monitor the impact of widespread varicella vaccination on the epidemiology of varicella and herpes zoster. While varicella incidence would be expected to decrease, mathematical models predict an initial increase in herpes zoster incidence if re-exposure to varicella protects against reactivation of the varicella zoster virus.
Methods
In 1998–2003, as varicella vaccine uptake increased, incidence of varicella and herpes zoster in Massachusetts was monitored using the random-digit-dial Behavioral Risk Factor Surveillance System.
Results
Between 1998 and 2003, varicella incidence declined from 16.5/1,000 to 3.5/1,000 (79%) overall with ≥66% decreases for all age groups except adults (27% decrease). Age-standardized estimates of overall herpes zoster occurrence increased from 2.77/1,000 to 5.25/1,000 (90%) in the period 1999–2003, and the trend in both crude and adjusted rates was highly significant (p < 0.001). Annual age-specific rates were somewhat unstable, but all increased, and the trend was significant for the 25–44 year and 65+ year age groups.
Conclusion
As varicella vaccine coverage in children increased, the incidence of varicella decreased and the occurrence of herpes zoster increased. If the observed increase in herpes zoster incidence is real, widespread vaccination of children is only one of several possible explanations. Further studies are needed to understand secular trends in herpes zoster before and after use of varicella vaccine in the United States and other countries.
==== Body
Background
Prior to the licensure of varicella vaccine in the U.S. in 1995, varicella (chickenpox) was a highly prevalent disease of childhood, resulting in approximately four million cases and an average of 10,630–13,500 hospitalizations and 90 deaths in the U.S. each year [1-4]. After primary infection, the varicella zoster virus (VZV) lies dormant in sensory ganglia and can reactivate later, producing herpes zoster (shingles) and sometimes persistent, debilitating pain and other serious complications [5]. Associated with a decline in specific cell-mediated immunity, herpes zoster affects some 600,000–900,000 mostly elderly or immunocompromised people in the U.S. each year, its incidence increasing with age [1,6].
Establishing and/or enhancing varicella and herpes zoster surveillance has been essential in order to assess the impact of widespread varicella vaccination [7,8]. Despite the fact that national vaccination coverage among children 19–35 months of age had risen to 85 percent by 2003 [9], varicella did not become a nationally notifiable disease in the U.S. until that year. Herpes zoster has never been a nationally notifiable disease, and no states require reporting of cases. Studies of herpes zoster using different methods have reported unadjusted incidence rates ranging from 1.3 to 4.8 per 1,000 population [10-13], but there is little documentation of secular trends over time. Some studies suggest that re-exposure to VZV after primary infection may decrease the risk of herpes zoster through immunologic boosting [14-16]. If this is the case, widespread use of varicella vaccine could, by reducing circulating VZV, increase the incidence of herpes zoster over the first few decades of universal childhood vaccination. On the other hand, studies in leukemic as well as healthy children have found a lower incidence and severity of herpes zoster among vaccinated than among comparable unvaccinated children [17-21], which would appear to bode well for the more distant future if vaccine coverage reaches high levels. Clearly, consistent and long-term surveillance for herpes zoster will be necessary in order to fully assess the impact of varicella vaccination on the epidemiology of zoster.
Starting in 1998, as uptake of varicella vaccine was increasing (due in part to immunization requirements for daycare and school entry), the Massachusetts Department of Public Health, in collaboration with the Centers for Disease Control and Prevention (CDC), enhanced varicella and herpes zoster surveillance by including questions in the Behavioral Risk Factor Surveillance System (BRFSS) about the occurrence of these diseases. Researchers in Kentucky had used both the BRFSS and a school-based cohort study to estimate age-specific varicella incidence in 1990–1992 and found generally close concordance between results of the two methods [22].
Herein, we analyze varicella and herpes zoster incidence as estimated by the Massachusetts BRFSS in the period 1998–2003.
Methods
Vaccine use and coverage
The Massachusetts Department of Public Health began to distribute varicella vaccine in September 1996 for use in 12–18-month-olds and susceptible sixth-graders. In October 1997, eligibility was expanded to include all children between 1 and 18 years of age. Beginning in August 1998, proof of immunity to varicella (written documentation of age-appropriate immunization or physician-certified reliable history of disease) was required for child-care center attendance for children ≥ 19 months of age who were born after 1996. Finally, in September 1999, proof of immunity became required for entry into kindergarten and seventh grade.
Vaccine coverage estimates were obtained from the National Immunization Survey. The CDC conducts this random-digit-dial telephone survey of households regarding the immunization status of children 19–35 months of age each year. National Immunization Survey methods have been published elsewhere [23].
Varicella and herpes zoster incidence
The BRFSS is an ongoing, random-digit-dial telephone survey of adults aged 18 years and older and gathers information on health characteristics, risks, and preventive behaviors. The survey is conducted in all U.S. states in collaboration with the CDC and state health departments. Once a household is contacted, one adult is randomly selected for interview. Characteristics of the BRFSS are described in more detail elsewhere [24].
The BRFSS includes a core set of questions asked by all states, as well as optional topics added by individual states. Massachusetts added questions about chickenpox and shingles among all household members in 1998–2000 and 2002–2003. Respondents were first asked to enumerate all household members and their ages. Trained interviewers then asked respondents whether any household members had chickenpox in the past 12 months and the current age(s) of the affected household member(s). Respondents were also asked whether any household members had ever had shingles, the current age(s) of affected household member(s), and the age(s) when they had shingles.
For these analyses, we calculated the annual age-specific incidence of varicella and herpes zoster per 1,000 population. In 1998 only, the response to the age-at-shingles question was recorded only in five-year age groups (0–4 years, 5–9 years, etc.), so annual incidence of herpes zoster could not be calculated for that year. We made some simplifying assumptions in estimating age-specific annual incidence of varicella and herpes zoster. For varicella, each case that was reported as occurring in the previous 12 months was assigned to the individual's current age. For herpes zoster, a case was counted as occurring within the previous year if the age at the time of the reported case was the same as or 1 year less than the current age. Our data on herpes zoster combine incident and recurrent cases, as no distinction was made between these in the interview. In presenting the varicella data, we used five standard child/adolescent age groups and grouped adults 20 years and older into a sixth. In the case of herpes zoster, we used four 20–25-year age intervals in order to increase the stability of the data.
BRFSS data were weighted to account for probability of selection and differential participation by age, sex, and race/ethnicity. All data were analyzed using appropriate procedures in SAS and SUDAAN, taking into account the survey sampling scheme, intra-class correlation among household members, and weighting of the data (SAS Institute, Cary, NC; SUDAAN, RTI, Research Triangle Park, NC). Tests for linear trend were done using logistic regression (with incidence as the outcome and year as the exposure variable), adjusting for age in comparisons of overall incidence. Age standardization was based on the standard 2000 U.S. population [25] in 14 mostly five-year age groups: 1–14, 15–19, 20–24, 25–29, 30–34, 35–39, 40–44, 45–49, 50–54, 55–59, 60–64, 65–69, 70–74, and 75+. The oldest group was not further subdivided due to diminishing cell sizes.
Results
Response rate
The response rate for the Massachusetts BRFSS as a whole was 59% in 1998, 55% in 1999, 41% in 2000, 66% in 2002, and 65% in 2003. The large apparent increase in response rate in 2002 was mainly due to the fact that in that year BRFSS expanded the number of case disposition codes in order to more accurately differentiate between different types of response and non-response and to bring the BRFSS disposition codes more in line with those recommended by the American Association for Public Opinion Research (AAPOR). This broader array of codes allows for more accurate accounting of participation rates, such as response rates (pers. comm., Dr. Michael Link, BRFSS, CDC). The number of households responding to the varicella-herpes zoster module ranged between 4,200 and 4,900, representing between 11,000 and 14,000 members.
Vaccine coverage
Using National Immunization Survey estimates for Massachusetts children 19–35 months of age [26], coverage increased from 23 percent in 1997 to 48 percent in 1998 and reached 89 percent in 2003 (Figure 1).
Figure 1 Varicella immunization levels and proportion of Massachusetts youth with varicella, 1998–2003. Annual National Immunization Survey estimates of varicella immunization levels and percent of Massachusetts residents < 20 years of age with varicella in the past year by date of interview, 1998–2003.
Varicella incidence
In 1998, the highest incidences of varicella were observed in children 1–4 years of age (83 per 1,000) and 5–9 years of age (76 per 1,000). Between 1998 and 2003, overall unadjusted varicella incidence declined by 13.0 cases per 1,000 (79%), from 16.5 to 3.5 per 1,000 (p < 0.0001; Table 1 and Figure 1). The overall incidence of varicella was higher in 2003 than in 2002 but not statistically significantly so. Over 1998–2003, incidence declined for all age groups, with the greatest decreases among infants, 1–4-, 5–9-, and 15–19-year olds (100%, 89%, 80%, and 92%, respectively) and the smallest among adults (27%). The trends were highly significant (p < 0.0001) for the 1–4- and 5–9-year olds and significant (p = 0.02) for the 10–14-year olds.
Table 1 Age-specific incidence (cases per thousand) of varicella and trends, Massachusetts, 1998–2003
Year Parameter Age group in years
< 1 1–4 5–9 10–14 15–19 20+ All ages, crude
1998
na 144 762 1025 922 791 8241 11885
Cases 4 66 83 23 9 34 219
Incidence 15.2 82.6 76.4 18.6 30.2 2.2 16.5
95% CI 4.7–49.5 59.9–113.9 55.0–104.5 10.8–32.0 8.3–109.4 1.4–3.5 12.8–21.3
1999
na 318 1603 1131 1094 898 8483 13527
Cases 4 46 41 11 3 17 122
Incidence 6.2 36.5 38.2 10.4 2.9 1.7 10.9
95% CI 1.2–31.6 21.1–63.0 24.8–59.0 4.8–22.5 0.6–13.4 0.9–3.4 7.8–15.1
2000
na 139 671 790 918 693 7518 10729
Cases 1 7 9 9 3 3 32
Incidence 11.4 11.6 10.3 8.8 4.2 0.3 2.7
95% CI 1.4–53.7 4.7–25.6 4.8–20.6 4.0–18.1 0.8–19.8 0.1–1.23 1.8–4.3
2002
na 161 600 780 831 705 7789 10866
Cases 1 4 5 3 1 6 20
Incidence 8.7 4.3 8.0 4.6 1.4 0.5 1.7
95% CI 1.1–44.6 1.2–13.9 3.0–20.0 1.3–15.4 0.2–9.7 0.2–1.1 1.0–2.8
2003
na 155 613 827 914 726 8012 11247
Cases 0 7 15 6 2 10 40
Incidence 0.0 8.8 15.0 6.4 2.3 1.6 3.5
95% CI 3.6–20.2 7.8–27.0 2.7–14.4 0.6–9.2 0.8–3.1 2.3–5.2
1998–2003
% change for period -100% -89% -80% -66% -92% -27% -79%
p for trend test n.a. <0.0001 <0.0001 0.02 0.15 0.16 <0.0001
a N's exclude individuals for whom age or varicella status was missing. The number excluded for these reasons was 494, 233, 478, and 195 in 1999, 2000, 2002, and 2003, respectively. Values for n's and numbers of cases are unweighted, whereas incidence rates and 95% confidence intervals are based on weighted analysis.
Herpes zoster incidence
Overall age-standardized (and crude) incidence per 1,000 of herpes zoster was 2.77 (2.23) in 1999, 3.54 (3.57) in 2000, 6.47 (6.51) in 2002, and 5.25 (5.38) in 2003 (Table 2), and the increasing trend was highly significant for both the age-adjusted (p < 0.0009) and crude (p < 0.0001) incidences. The increase in the age-standardized incidence was 90 percent over the five-year period. Over this period, age-specific incidences increased by 41–161 percent, although the annual age-specific incidence estimates were somewhat unstable (Table 2, Figure 2).
Table 2 Age-specific incidence (cases per thousand) of herpes zoster and trends, Massachusetts, 1999–2003
Year Parameter Age group in years
1–24 25–44 45–64 65+ All ages, crude All ages, standardized a
1999
nb 5655 4230 2214 911 13010
Cases 5 14 11 10 40
Incidence 0.87 1.87 4.82 6.9 2.23 2.77
95% CI 0.25–3.05 0.90–3.88 2.36–9.86 2.97–16.05 1.46–3.43 1.82–4.20
2000
nb 3668 3449 2211 1121 10449
Cases 2 12 13 8 35
Incidence 0.46 3.74 6.45 6.45 3.57 3.54
95% CI 0.11–1.91 2.04–6.76 3.60–11.29 3.04–13.16 2.50–5.08 2.48–5.03
2002
nb 3469 3234 2451 1303 10457
Cases 6 14 20 23 63
Incidence 1.70 5.13 7.42 20.03 6.51 6.47
95% CI 0.71–4.03 2.65–9.70 4.39–12.28 12.46–30.59 4.86–8.67 4.92–8.47
2003
nb 3698 3260 2700 1278 10936
Cases 8 16 18 14 56
Incidence 2.19 4.88 6.79 11.74 5.38 5.25
95% CI 1.05–4.50 2.86–8.20 4.02–11.23 6.57–20.12 4.03–7.14 3.96–6.94
1999–2003
% change for period 152% 161% 41% 70% 141% 90%
p for trend test 0.10 0.03 0.42 0.03 0.0001 0.0009
a Incidence is standardized using the standard 2000 U.S. population. The p-value is from age-adjusted analysis for trend using logistic regression.
b N's exclude individuals for whom age or herpes zoster status was missing. The number excluded for these reasons was 1011, 513, 887, and 506 in 1999, 2000, 2002, and 2003, respectively. Values for n's and numbers of cases are unweighted, whereas incidence rates and 95% confidence intervals are based on weighted analysis.
Figure 2 Annual incidence of herpes zoster by age, 1999–2003. Annual incidence of herpes zoster by age: comparison of Harvard Community Health Plan (HCHP), 1990–1992 [6], and the Massachusetts Behavioral Risk Factor Surveillance System (BRFSS), 1999, 2000, 2002, and 2003.
Discussion
The age-specific incidences of varicella in 1998 were generally within the range of estimates reported in studies conducted using a variety of methods in other parts of the U.S. prior to the introduction of the vaccine [1,22,27-29]. We noted a marked reduction in varicella incidence in all age groups in Massachusetts over the period 1998–2003; the decline was between 66 and 100 percent for all groups except adults, in whom incidence decreased by 27 percent. Vaccine coverage in Massachusetts 19–35-month-olds increased from 48 to 89 percent over the same 6-year period.
Seward et al. [30] observed a similarly consistent decline over the 6-year period of 1995–2000 in three areas of active surveillance, whose range of vaccine coverage as measured by the National Immunization Survey encompassed Massachusetts' coverage each year. Likewise, Jumaan et al. [31] documented a decline in age-adjusted incidence of varicella from 2.63 per 1,000 in 1995 to 0.92 per 1,000 in 2002. Decreases in varicella incidence after vaccine licensure have been noted in Illinois, Michigan, Texas, and West Virginia as well [32]. Similarly, large decreases in incidence among children in childcare centers and among U.S. Navy recruits have been associated with varicella vaccination [33,34].
The overall incidence of varicella was higher in 2003 than in 2002, although the difference was not statistically significant. No such upturn was apparent in Massachusetts Department of Public Health surveillance data, which showed an overall decrease of 20 percent in cases and incidence between 2002 and 2003, in spite of increased efforts to stimulate varicella reporting. (Overall incidence of herpes zoster decreased from about 6 to about 5 per 1,000 from 2002 to 2003, but the change was not statistically significant for either the crude or the age-standardized estimates.)
Age-standardized overall annual herpes zoster incidence increased by 2.5 cases per 1,000 (90%) over the 1999–2003 period, to 5.25 per 1,000, higher than the age-standardized estimates derived using the same standard population from a study in the Netherlands in 1994–1999 (3.34 per 1,000) [35] and from a study in Seattle, Washington in 1992–2002 (range: 3.47–4.05 per 1,000) [31] but not substantially higher than the estimate obtained from a study in France in 1997–1998 (4.55 per 1,000) [13]. The age-specific incidences of herpes zoster found here were consistently higher than in a Minnesota study of 50 years ago [12], and more than half of the age-specific point estimates in four other studies [11-13,6] fell below the 95 percent confidence intervals of our combined 2002–2003 estimates. (Confidence intervals for most of these previous studies were not available for comparison.) The four age groups showed increases of 41–161 percent, and the trend was significant for the 25–44 and 65+ age groups. In spite of the fact that the annual age-specific estimates of herpes zoster incidence were unstable, as can be seen particularly in the 65+ year-old age group (Figure 2), it does appear that there was a generally increasing trend in herpes zoster incidence for all age groups, based on these four years of data over a five-year period. This trend appears to conflict with findings from the Seattle study, a large longitudinal study in a health-maintenance organization in which age-adjusted as well as age-specific herpes zoster incidences remained relatively stable over the period 1992–2002, during the latter four years of which varicella incidence declined [31].
If in fact herpes zoster incidence is rising in the current period, there are several possible explanations, which may be operating in combination. One of these is varicella vaccination. Widespread varicella vaccination may end up increasing herpes zoster incidence among adults who have had varicella. Recent studies have suggested a protective effect against herpes zoster from exposure to varicella and/or children [14-16]. One mathematical model predicts a long-term (equilibrium state) elevated incidence of herpes zoster of 20 percent or more relative to pre-vaccination levels, under the assumption of immunologic boosting by exposure to VZV, if vaccine virus has a 75 percent or greater chance of reactivating to cause herpes zoster [36]. A later model by Brisson et al. [16], which was parameterized by means of a large, population-based survey and assumes that immunity from boosting lasts for 20 years, predicts a substantial increase in herpes zoster cases over the first 30–50 years after the initiation of mass vaccination, peaking about 20 years after the start of mass vaccination at an incidence of 51 percent over the pre-vaccination level and eventually falling below the initial incidence. Those aged 10–44 years at the start of mass vaccination will be most affected, according to the model, their lifetime risk exceeding 50 percent, compared to 33 percent in the pre-vaccination period – most will have been previously infected but will no longer experience boosting by exposure to children with varicella disease.
The Brisson model predicts the greatest rates of increase in herpes zoster incidence in the years just after the start of mass varicella vaccination, with a 40 percent increase over the pre-vaccination level expected by Year 10. Yet, over just a five-year period, we saw an increase of this magnitude in the 45–64-year age group and even larger increases in the other age groups, and it seems unlikely that varicella vaccination alone can explain our herpes zoster results. Other possible explanations include increases in the proportion of people with immunosuppressive conditions and therapies, in the duration of those conditions and treatments, and/or in the prevalence of other triggering factors. Some past studies have noted an increase in zoster incidence well before the availability of varicella vaccine, one finding 28- and 41-percent increases for Minnesota women and men, respectively, between the late 1940s and the late 1950s [12] and one finding a 35-percent increase between 1979 and 1997 in Canada [37]. Similarly, data from the National Health Interview Survey suggest that incidence of herpes zoster rose in the 60+ age groups between 1974 and 1994 [38].
Our study is subject to several limitations. Our method of estimating annual incidence of herpes zoster, by which a case was counted as occurring within the past year if the age at the time of the reported case was the same as or 1 year less than the age at interview, would tend to overestimate annual incidence because this encompasses a period greater than 12 months. The inclusion of recurrent cases with first-time cases would have the same effect. However, any overestimation due to these factors would be felt uniformly over age groups and time and would not be expected to influence year-to-year differences. Nevertheless, this would affect comparisons with other studies and likely contributes to our generally higher rates.
Secondly, varicella and herpes zoster incidences were based on respondent recall, both self- and proxy-report, and we did not validate the accuracy of the reporting. Proxy reporting may or may not introduce reporting error, depending on the health condition in question and the relationship of the respondent to the person about whom information is sought [39,40]. Regarding varicella, in our combined data for 1999 and 2000, a parent or step-parent was the respondent in more than 90 percent of households with children, and restricting the varicella analysis to these parent-respondents did not appreciably change our estimates of annual age-specific incidence. Researchers using a telephone survey to investigate varicella incidence in a city in Minnesota estimated through medical chart abstraction that 13 percent of parental reports of varicella in the previous year were either not varicella or occurred more than 13 months before the interview [28]. Our data could reflect similar over-reporting; however, error of this nature would not likely change from year to year and thus the trends would not be affected. One study found shingles self-report among the elderly to be accurate, with few false-positive and false-negative reports when compared to physician diagnosis [41]. We found no significant difference between shingles incidence in respondents of age 50 and older and shingles incidence in the same age group as reported by proxies.
Also, in this study, herpes zoster was not likely often diagnosed by laboratory testing. If the public or physicians became more aware of zoster in the latter years of the study, there might have been more of a tendency to attribute rash illness to zoster with each successive annual survey, leading to an increase in observed incidence. However, we noted no increasing attention to zoster in the popular media nor a concern among physicians about a projected increase in incidence. Thus, we doubt that the observed trend is due to greater awareness.
We did not collect data on underlying immunosuppressive conditions, which would have permitted us to examine the extent to which changes in herpes zoster incidence were related to changes in the prevalence of such diseases or therapies. Although it is unlikely this prevalence changed much in the five-year study period, it is conceivable that the proportion of respondents with such conditions varied randomly from year to year, which could have affected the results.
Finally, the response rate for the BRFSS has been uneven, hitting a low of 41 percent in 2000. If the incidence of varicella or herpes zoster differs among responders and non-responders, this could introduce some bias in the estimates of both the absolute incidence and of trends over time. However, a comparison of the 2000 BRFSS sample to the 2000 U.S. Census data for Massachusetts concluded that the BRFSS sample reflected its target population in most demographic and socioeconomic variables despite its low response rate [42]. Any differences in characteristics that might be relevant to herpes zoster incidence, such as race/ethnicity and proportion of households with children, were slight. The proportion of households with children in the sample varied little between 2000 and 2003, from a low of 37.1 percent to a high of 38.2 percent. In 1999, it was 33.7 percent, but given that the effect of living with children is thought to be protective, this difference would not explain our results.
Conclusion
As varicella vaccine coverage in children increased, the incidence of varicella decreased and the occurrence of herpes zoster increased. If the observed increase is real, widespread vaccination of children is only one of several possible explanations.
Further research on herpes zoster is clearly warranted, particularly as a vaccine to prevent herpes zoster and its complications is likely to be available in the United States within the next few years [43-45]. A better understanding is needed of risk factors, the possible roles of internal (from subclinical reactivation) and external (from exposure to circulating VZV) boosting in preventing or delaying zoster, and temporal trends in incidence since mid-century. Additional methods for surveillance for herpes zoster should be considered to monitor the impact of widespread varicella vaccination of children as well as the impact of VZV vaccination of adults, should it become generally recommended.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
WKY helped design the questionnaire, interpreted the data, and drafted and revised the manuscript. DB helped design the questionnaire, oversaw implementation of the survey for several years, oversaw analysis of the data through 2002, interpreted the data, and critically reviewed all drafts of the manuscript, providing important insights. Recognizing the importance of carrying out the survey, SL secured funding to add the varicella and herpes zoster questions, and helped design the questionnaire, provided immunization data, and reviewed the manuscript. AJ contributed critical insights and raised questions in the interpretation of the data, provided comparative data and references, and critically reviewed all drafts of the manuscript. ZZ supported staff time for data analysis and himself re-analyzed the data through 2003; in addition, he helped interpret the data, ruling out certain artifactual explanations for the observations. KC analyzed the data through 2002 and provided much of the material for the methods section. JS funded the work, helped design the questionnaire, raised questions and contributed critical insights for the interpretation of the data, and critically reviewed the manuscript, supplying partial rewrites. All authors have approved the current version for publication.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
Funding was provided by the National Immunization Program (grant number H23/CCH104486) and the Behavioral Risk Factor Surveillance System, Centers for Disease Control and Prevention. Dr. Yih receives support from the Vaccine Safety Datalink Project, National Immunization Program, Centers for Disease Control and Prevention. We gratefully acknowledge the bibliographic assistance of Jennifer Mann and critical readings by Paul Gargiullo.
==== Refs
Wharton M The epidemiology of varicella-zoster virus infections Infect Dis Clin North Am 1996 10 571 581 8856352 10.1016/S0891-5520(05)70313-5
Galil K C Brown F Lin J Seward Hospitalizations for varicella in the United States, 1988 to 1999 Pediatr Infect Dis J 2002 21 931 934 12394815 10.1097/00006454-200210000-00009
Meyer PA Seward JF Jumaan AO Wharton M Varicella mortality: trends before vaccine licensure in the United States, 1970–1994 J Infect Dis 2000 182 383 390 10915066 10.1086/315714
Davis MM Patel MS Chem BS Gebremariam A Decline in varicella-related hospitalizations and expenditures for children and adults after introduction of varicella vaccine in the United States Pediatrics 2004 114 786 792 15342855 10.1542/peds.2004-0012
Gilden DH Kleinschmidt-DeMasters BK LaGuardia JJ Mahalingam R Cohrs RJ Neurologic complications of the reactivation of varicella-zoster virus N Engl J Med 2000 342 635 645 10699164 10.1056/NEJM200003023420906
Donahue JG Choo PW Manson JE Platt R The incidence of herpes zoster Arch Intern Med 1995 155 1605 1609 7618983 10.1001/archinte.155.15.1605
Centers for Disease Control and Prevention Evaluation of varicella reporting to the National Notifiable Disease Surveillance System – United States, 1972–1997 Morb Mortal Wkly Rep 1999 48 55 58
Food and Drug Administration Summary for Basis of Approval Reference number: 93-0395
Centers for Disease Control and Prevention National, state, and urban area vaccination levels among children aged 19–35 months – United States, 2003 Morb Mortal Wkly Rep 2004 53 658 661
McGregor RM Herpes zoster, chicken-pox, and cancer in general practice Br Med J 1957 32 84 87 13383200
Hope-Simpson RE The nature of herpes zoster: A long-term study and a new hypothesis Proc R Soc Med 1965 58 9 20 14267505
Ragozzino MW Melton LJ IIIKurland LT Chu CP Perry HO Population-based study of herpes zoster and its sequelae Medicine (Baltimore) 1982 61 310 316 6981045
Chidiac C Bruxelle J Daures JP Hoang-Xuan T Morel P Lepiege A El Hasnaoui A de Labareyre C Characteristics of patients with herpes zoster on presentation to practitioners in France Clin Infect Dis 2001 33 62 69 11389496 10.1086/320884
Gershon AA LaRussa P Steinberg S Mervish N Lo SH Meier P The protective effect of immunologic boosting against zoster: An analysis in leukemic children who were vaccinated against chickenpox J Infect Dis 1996 173 450 453 8568309
Thomas SL Wheeler JG Hall AJ Contacts with varicella or with children and protection against herpes zoster in adults: a case-control study Lancet 2002 360 678 682 12241874 10.1016/S0140-6736(02)09837-9
Brisson M Gay NJ Edmunds WJ Andrews NJ Exposure to varicella boosts immunity to herpes-zoster: implications for mass vaccination against chickenpox Vaccine 2002 20 2500 2507 12057605 10.1016/S0264-410X(02)00180-9
Brunell PA Taylor J Geiser CF Frierson L Lydick E Risk of herpes zoster in children with leukemia: varicella vaccine compared with history of chickenpox Pediatrics 1986 77 53 65 3455669
Lawrence R Gershon AA Holzman R Steinberg SP the NIAID Varicella Vaccine Collaborative Study Group The risk of zoster after varicella vaccination in children with leukemia N Engl J Med 1988 318 543 548 2828948
Hardy I Gershon AA Steinberg SP LaRussa PL the Varicella Vaccine Collaborative Study Group The incidence of zoster after immunization with live attenuated varicella vaccine. A study in children with leukemia N Engl J Med 1991 325 1545 1550 1658650
Plotkin SA Starr SE Connor K Morton D Zoster in normal children after varicella vaccine J Infect Dis 1989 159 1000 1001 2540244
Gershon A Silverstein S Live attenuated varicella vaccine for prevention of herpes zoster Biologicals 1997 25 227 230 9236057 10.1006/biol.1997.0089
Finger R Hughes JP Meade BJ Pelletier AR Palmer CT Age-specific incidence of chickenpox Public Health Reports 1994 109 750 755 7800783
Smith PJ Battaglia MP Huggins VJ Hoagland DC Roden A Khare M Ezzati-Rice TM Wright RA Overview of the sampling design and statistical methods used in the National Immunization Survey Am J Prev Med 2001 20 61 69 10.1016/S0749-3797(01)00285-9
Powell-Griner E Anderson JE Murphy W State and sex-specific prevalence of selected characteristics – Behavioral Risk Factor Surveillance System, 1994 and 1995 CDC Surveillance Summary, August 1, 1997 Morb Mortal Wkly Rep 1997 46 1 31
Klein RJ Schoenborn CA Age adjustment using the 2000 projected U.S. population Statistical Notes 20 1 10
Centers for Disease Control and Prevention, National Immunization Program
Guess HA Broughton DD Melton LJ IIIKurland LT Population-based studies of varicella complications Pediatrics 1986 78 723 727 3763290
Yawn BP Yawn RA Lydick E The community impact of childhood varicella infections J Pediatrics 1997 130 759 765
Choo PW Donahue JG Manson JE Platt R The epidemiology of varicella and its complications J Infect Dis 1995 172 706 712 7658062
Seward JF Watson BM Peterson CL Mascola L Pelosi JW Zhang JX Maupin TJ Goldman GS Tabony LJ Brodovicz KG Jumaan AO Wharton M Varicella disease after introduction of varicella vaccine in the United States, 1995–2000 JAMA 2002 287 606 611 11829699 10.1001/jama.287.5.606
Jumaan AO Yu O Jackson LA Bohlke K Galil K Seward JF Herpes zoster incidence before and after declines in varicella incidence associated with varicella vaccination, 1992–2002 J Infect Dis
Centers for Disease Control and Prevention Decline in annual incidence of varicella – selected states, 1990–2001 Morb Mortal Wkly Rep 2003 52 884 885
Clements DA Zaref JI Bland CL Walter EB Coplan P Partial uptake of varicella vaccine and the epidemiological effect on varicella disease in 11 day-care centers in North Carolina Arch Pediatr Adolesc Med 2001 155 433 461
Ryan MA Smith TC Honner WK Gray GC Varicella susceptibility and vaccine use among young adults enlisting in the United States Navy J Med Virol 2003 S15 19 12627481 10.1002/jmv.10314
Opstelten W Mauritz JW de Wit NJ van Wijck AJM Stalman WAB van Essen GA Herpes zoster and postherpetic neuralgia: incidence and risk indicators using a general practice research database Fam Pract 2002 19 471 475 12356697 10.1093/fampra/19.5.471
Garnett GP Grenfell BT The epidemiology of varicella-zoster virus infection: The influence of varicella on the prevalence of herpes zoster Epidemiol Infect 1992 108 513 528 See pp. 521–522 1318219
Brisson M Edmunds WJ Law B Gay NJ Walld R Brownell M Roos L de Serres G Epidemiology of varicella zoster virus infection in Canada and the United Kingdom Epidemiol Infect 2001 127 305 314 11693508
Jumaan AO Seward JF Wooten K Chen J Singleton JA Varicella and herpes zoster surveillance in the U.S., 1970–1994 Presented at 41st Annual Meeting of IDSA, October 9–12, 2003. San Diego, CA Abstract # 899
Woolsey TD Lawrence PS Balamuth E An evaluation of chronic disease prevalence data from the health interview survey Am J Public Health 1962 52 1631 1637 14001857
Halabi S Zurayk H Awaida R Darwish M Saab B Reliability and validity of self and proxy reporting of morbidity data: a case study from Beirut, Lebanon Int J Epidemiol 1992 21 607 12 1386064
Schmader K George LK Newton R Hamilton JD The accuracy of self-report of herpes zoster J Clin Epidemiol 1994 47 1271 1276 7722563 10.1016/0895-4356(94)90132-5
Zhang Z Cohen B How representative is the Massachusetts BRFSS sample: a comparison to the current census data 2003 Presented at BRFSS Annual Conference, St. Louis, MO
Oxman MN Levin MJ Johnson GR Schmader KE Straus SE Gelb LD Arbeit RD Simberkoff MS Gershon AA Davis LE Weiberg A Boardman KD A vaccine to prevent herpes zoster and postherpetic neuralgia in older adults N Engl J Med 2005 352 2271 2284 15930418 10.1056/NEJMoa051016
Gilden DH Varicella-zoster virus vaccine – grown-ups need it, too N Engl J Med 2005 352 2344 2346 15930426 10.1056/NEJMe058090
Arvin A Aging, immunity, and the varicella-zoster virus N Engl J Med 2005 352 2266 2267 15930416 10.1056/NEJMp058091
| 15960856 | PMC1177968 | CC BY | 2021-01-04 16:28:56 | no | BMC Public Health. 2005 Jun 16; 5:68 | utf-8 | BMC Public Health | 2,005 | 10.1186/1471-2458-5-68 | oa_comm |
==== Front
BMC Public HealthBMC Public Health1471-2458BioMed Central London 1471-2458-5-741600197710.1186/1471-2458-5-74Research ArticleCollege campus smoking policies and programs and students' smoking behaviors Borders Tyrone F [email protected] K Tom [email protected] Donna [email protected] Lee [email protected] Danielle [email protected] Department of Health Policy and Management, University of Arkansas for Medical Sciences, Little Rock, Arkansas, USA2 Division of Health Services Research, Texas Tech University Health Science CenterLubbock, Texas, USA3 Department of Pediatrics and Center for Tobacco Control and Prevention, Texas Tech University Health Science Center, Lubbock, Texas, USA4 Department of Psychology, Texas Tech University, Texas, USA5 Center for Tobacco Control and Prevention, Texas Tech University Health Science Center, Lubbock, Texas, USA2005 7 7 2005 5 74 74 1 4 2005 7 7 2005 Copyright © 2005 Borders et al; licensee BioMed Central Ltd.2005Borders 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
Although tobacco use in the United States has declined over the past 20 years, cigarette use among college students remains high. Additional research is thus needed to determine how university tobacco control policies and preventive education programs affect college students' smoking behaviors.
Methods
Approximately 13,000 undergraduate students at 12 universities or colleges in the state of Texas completed a web-based survey. College smoking policies were obtained from a survey of college administrators and from college websites. Logistic regression analyses were conducted to estimate the effects of individual smoking policies and programs on the odds of cigarette smoking.
Results
Of the individual programs, only having a preventive education program on campus was associated with lower odds of smoking. The existence of smoking cessation programs and designated smoking areas were associated with higher odds of smoking. Policies governing the sale and distribution of cigarettes were insignificantly associated with smoking.
Conclusion
Rather than focusing on policies restricting cigarette sales and use, college administrators should consider implementing or expanding tobacco prevention and education programs to further reduce student smoking rates.
==== Body
Background
Despite reductions in smoking rates over the past 20 years, tobacco use remains a key public health concern. According to recent estimates from the Centers for Disease Control, smoking is the leading cause of preventable death in the United States [1]. Cigarette use among college students is of particular concern as they have a substantially higher prevalence of smoking relative to the general adult population [1,2].
To curb smoking rates on college campuses, several national organizations recommend that colleges and universities enact stricter tobacco control and prevention policies. The American College Health Association (ACHA) and American Cancer Society (ACS) advocate that colleges ban smoking in all campus buildings and residence halls; prohibit the sale, sampling, and advertising of tobacco products; restrict smoking to a minimum of 20 feet from building entrances and air intake units; limit or prohibit spit tobacco use on campus; and implement tobacco prevention/education and cessation programs on campus [3,4].
Many institutions have responded to these recommendations. According to a survey of 50 colleges and universities across the United States, the number of campuses regulating cigarette use in student housing increased from 1% to 54% between 1994–1995 and 2002–2003 [5]. A cross-sectional survey of college students suggests that campus housing smoking bans are effective in decreasing smoking rates [6]. In addition to policies regulating where students can smoke, many universities have implemented public education campaigns aimed at non-smokers and smoking cessation programs targeting current smokers. According to a survey of health directors at 393 four-year universities across the U.S., 40% of schools offered some type of smoking cessation program [6]. Several studies have shown that anti-smoking messages can reduce smoking rates among younger adults [7-10], but little research has been conducted to assess their effectiveness among college students.
While there are some indications that more stringent tobacco policies and greater availability of promotion and prevention programs reduce smoking rates on college campuses, there is a lack of conclusive research in the area. More extensive regulatory policies could have a negative effect if they elicit a rebellious response from students. An additional question which remains insufficiently answered is the degree to which regulatory policies as compared to prevention or education are associated with smoking behaviors. An improved understanding of the relative effects of different types of policies and programs could assist college and university administrators as they consider how to invest resources to curb smoking.
Methods
Survey sample
A web-based survey, which was approved by Texas Tech University Health Sciences Center's institutional review board (IRB), was conducted among college students in Texas. The survey contained questions covering past and current cigarette, cigar, and smokeless tobacco use; knowledge of the risks associated with tobacco use; and responses to tobacco marketing campaigns. The sampling frame and methodology have been discussed elsewhere [11], but are briefly described here. All 72 colleges and universities within the state of Texas were invited to participate in a web-based survey of their undergraduate students' tobacco use. Of the 72 institutions, 27 did not elect to respond to requests to participate in the survey, 3 did not collect e-mail addresses from their students (which was necessary to recruit students), and 29 refused to release students' e-mail addresses. Thus, 13 college and universities agreed to participate in the study, representing 184,559 students.
Students from the participating schools were e-mailed and invited to participate in a web-based survey about tobacco use. To encourage participation, students were offered to enroll in a lottery to receive one of five $500 airline gift certificates upon completion of the survey. They were then directed to a web site where they could complete the survey. Approximately 13% of students from the 13 participating schools completed the survey. For the analyses conducted here, one college was excluded because it had a particularly low response rate (approximately 2%). Thus, the final sample represented 13,041 undergraduate students at 12 four-year colleges and universities in Texas.
Variables
Because we were concerned with the occurrence of smoking over a defined time period, as opposed to the quantity of cigarette consumption, we defined current smoking as having smoked at least one cigarette in the past 30 days. This is the standard method of categorizing persons as current vs. non-current smokers [6,12-14].
College-level campus smoking policies and regulations were obtained through mail surveys of college administrators. College administrators were asked whether their school had specific campus policies which restricted tobacco distribution, prohibited tobacco sales, restricted smoking to 20 feet from building entrances, prohibited smoking in student residence halls or dormitories, clearly identified non-smoking areas, prohibited tobacco ads in campus publications, prohibited events sponsored by tobacco companies, and prohibited smoking in all indoor public areas. They were also asked whether their school provided preventive education related to smoking and smoking cessation courses. Data collected from college administrators were supplemented by information obtained from each college's website.
Student-level control variables included the smoking status of a student's social circle (roommates and friends), academic information (college and academic classification), demographics (age, race/ethnicity, and gender), and memberships in certain groups (fraternity/sorority and intercollegiate sports team).
Analyses
To investigate whether universities' smoking policies and programs were significantly associated with the probability of smoking, multivariate logistic regressions were conducted. Among the ten policies, three (prohibition of tobacco advertisements in campus publications, events sponsored by tobacco companies, and smoking in indoor public areas) were present at all universities in the sample. Due to the lack of variation in these three policies, they were excluded from the analyses.
Univariate logistic regression analyses were first conducted to estimate unadjusted odds ratios for the university-level policies and programs. Student-level variables were included in the final multivariate model to control for their potential confounding effects. Because more than one student from each participating university was present in the sample, a traditional regression analysis was not used as it would provide larger standard errors of the estimated coefficients. To correct for the within-cluster dependence of the observations, a sandwich-like robust estimator of the variance-covariance matrix was used.
Results
The sample is described in Table 1. Seven of the 12 schools (58.3%) had a policy regarding tobacco distribution and prohibited smoking in residence halls. Only 2 (16.7%) universities prohibited tobacco sales on campus and had restrictions on smoking to a minimum of 20 feet from entrances of buildings. Five schools (41.7%) clearly identified non-smoking areas. Five schools (41.7%) provided preventive education and four (33.3%) delivered smoking cessation courses.
Table 1 Descriptive Statistics of Sample
College-level policies and programs % (n = 12)
Restrict tobacco distribution 58.3
Prohibit tobacco sales on campus 16.7
Restrict smoking to 20 ft. from entrances 16.7
Prohibit smoking in residence halls 58.3
Clearly identify non-smoking areas 41.7
Provide prevention and education 41.7
Provide smoking cessation classes 33.3
Student-level control variables % (n = 13,041)
Age group (referent = 18–19)
20–21 38.5
≥22 38.3
Race/Ethnicity (referent = white)
Hispanic 11.7
Black 2.6
Others 11.4
Gender (referent = male)
Female 61.3
Roommate smokes (referent = no roommate)
Yes 27.8
No 50.4
% of friends who smoke (referent = ≤20%)
21–50% 28.0
≥50% 15.5
College (referent = applied science)
Arts 15.7
Business 18.2
Engineering 8.6
Human Sciences 11.8
Mass Communications 11.7
Others 21.7
Academic classification (referent = freshman)
Sophomore 14.7
Junior 20.1
Senior 28.1
Graduate 22.2
Fraternity/sorority member (referent = no)
Yes 16.2
Intercollegiate athlete (referent = no)
Yes 4.6
Approximately 37% of the students in the sample reported smoking in the past 30 days. The majority of the sample (74%) was non-Hispanic white and about 61% was female. Although about half of the sample had roommates who smoked, only 16% reported that more than half of their friends were smokers. Approximately 16% of the students were members of fraternities or sororities. Fewer than 5% were members of intercollegiate sports teams.
Table 2 shows the effects of universities' smoking policies and programs on the probability of smoking, controlling for students' characteristics. Only three of the seven policies or programs showed significant effects. Preventive education programs decreased the odds of smoking by about 23%. In contrast, smoking cessation programs and the identification/designation of smoking areas significantly increased the odds of smoking by 30% and 45%, respectively.
Table 2 Unadjusted and Adjusted Odds Ratios (OR) for Current Smoking
Unadjusted OR (95% CI) p Adjusted OR (95% CI) p
College-level policies and programs
Restrict tobacco distribution 1.12 (0.82–1.52) 0.48 1.00 (0.90–1.11) 0.97
Prohibit tobacco sales on campus 0.82 (0.67–1.02) 0.07 0.95 (0.83–1.08) 0.43
Restrict smoking to 20 ft. from entrances 0.83 (0.67–1.03) 0.09 0.91 (0.79–1.04) 0.15
Prohibit smoking in residence halls 1.21 (0.94–1.54) 0.14 1.02 (0.91–1.36) 0.77
Clearly identify non-smoking areas 1.48 (1.21–1.81) <0.01 1.45 (1.32–1.60) <0.01
Provide prevention and education 0.95 (0.66–1.39) 0.81 0.77 (0.63–0.93) <0.01
Provide smoking cessation classes 0.95 (0.66–1.37) 0.80 1.30 (1.17–1.45) <0.01
Student-level control variables
Age group (referent = 18–19)
20–21 - 0.92 (0.84–1.01) 0.07
≥22 - 0.85 (0.74–0.98) 0.02
Race/Ethnicity (referent = white)
Hispanic - 0.85 (0.72–1.00) 0.05
Black - 0.51 (0.37–0.69) <0.02
Others - 0.94 (0.87–1.01) 0.11
Gender (referent = male)
Female - 0.94 (0.90–0.99) 0.02
Roommate smokes (referent = no roommate)
Yes - 1.42 (1.26–1.59) <0.01
No - 0.67 (0.64–0.70) <0.01
% of friends who smoke (referent = ≤20%)
21–50% - 2.62 (2.46–2.79) <0.01
≥50% - 6.44 (5.45–7.60) <0.01
College (referent = applied science)
Arts - 1.02 (0.83–1.27) 0.84
Business - 0.91 (0.79–1.03) 0.14
Engineering - 0.80 (0.73–0.89) <0.01
Human Sciences - 0.96 (0.82–1.12) 0.62
Mass Communications - 0.99 (0.85–1.15) 0.88
Others - 1.18 (1.02–1.37) 0.03
Academic classification (referent = freshman)
Sophomore - 0.86 (0.74–0.99) 0.04
Junior - 0.79 (0.67–0.93) 0.01
Senior - 0.79 (0.72–0.86) <0.01
Graduate - 0.59 (0.50–0.69) <0.01
Fraternity/sorority member (referent = no)
Yes - 1.09 (0.98–1.22) 0.12
Intercollegiate athlete (referent = no)
Yes - 0.93 (0.80–1.07) 0.30
Estimation results indicated that student characteristics played a more important role in predicting smoking status. The smoking status of a students' social circle was significantly associated with smoking status. Also, maturity decreased the probability of being a smoker. As compared with freshmen, sophomores, juniors, seniors, and graduate students were significantly less likely to smoke. This pattern is mirrored in the age variable.
As compared with non-Hispanic white students, Hispanic and black students were less likely to smoke. Also, females were less likely to smoke, which is consistent with national level observations among adults [1]. Being a member of a fraternity/sorority or intercollegiate sports teams was not found to significantly contribute to the probability of smoking.
Discussion
Restriction of tobacco distribution, prohibition of tobacco sales, restriction of smoking within 20 feet from entrances, and prohibition of smoking in residence halls were not associated with the odds of smoking, suggesting that they are ineffective in influencing college students' smoking behaviors. Yet, other programs do appear to impact college students' smoking behaviors. We found that students enrolled at schools with smoking cessation programs had higher smoking rates, but this should not necessarily lead to the conclusion that smoking cessation classes contribute to smoking. It is reasonable to infer that administrators of colleges and universities that have higher smoking rates implement cessation programs in an attempt to reduce cigarette use. On the other hand, it has been shown that that college campus smoking cessation courses and programs are underutilized and that college students often want to quit on their own [15].
Prevention-oriented education was associated with lower odds of current cigarette use. Students who were exposed to education about smoking were 23 percent less likely to smoke compared their counterparts who were not exposed to campus education. There is little information about the efficacy of such campaigns on college campuses, but a number of studies have shown that anti-tobacco messages have a positive effect on tobacco use rates in other populations [7,9,10]. When Massachusetts evaluated the effects of its comprehensive state-wide tobacco control program on college students who were in high school when the program was initiated, results indicated that those students exposed to the tobacco prevention and education portions of the program had lower smoking rates than college students that were not exposed [14].
Although several of the policies and programs examined in this paper do not appear to affect current cigarette use, it is possible that others do affect college students' smoking status. Taking into consideration that all of the schools in the present sample prohibited tobacco advertisements in campus publications, events sponsored by tobacco companies, and smoking in indoor public areas, it was not possible to test for their effects on the odds of smoking. The latter three regulations may limit smoking to a point where the addition of other regulations has little impact. It should also be noted that the policies and programs discussed in the present paper could yield other, unmeasured benefits, such as reducing smoking rates among faculty and staff.
Despite the notable findings, the present study has a number of methodological limitations. Although the data set includes a large number of respondents, the participation rate could bias the results. The level of distribution and enforcement of policies was not measured, but could influence their effectiveness in controlling tobacco use [16]. Additional research is thus warranted to determine the degree to which college students are aware of published campus tobacco policies. Because we did not obtain detailed descriptions of the nature and content of preventive education and smoking cessation courses available on campus, we can not conclude whether particular programs are more effective than others.
Conclusion
In summary, results from the present study imply a need for additional evaluation of the effects of various approaches to tobacco control and prevention on student tobacco usage rates. Future studies should employ different methodologies whereby distinct conclusions about the effectiveness of specific policies or programs can be determined. In the interim, however, the present results suggest that prevention and education programs may have the strongest protective effect on college students' cigarette use.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
TFB conceived the study design, participated in data analysis, and led the writing of the manuscript. KTX assisted with data analysis and contributed to the writing of the manuscript. DB was responsible for conceiving the survey and overseeing data collection. LC assisted with the survey design and data collection. DSM assisted with data collection. All authors contributed to the writing of the manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
This study was funded by the Texas Department of Public Health. The authors would like to thank Joel West and the Institute for Communications Research, Texas Tech University, for their assistance in conducting the survey.
==== Refs
Centers for Disease Control Cigarette Smoking Among Adults-United States MMWR 2001 50 869 873 11666113
Johnston LD O'Malley PM Bachman JG Monitoring the Future National Survey Results on Drug Use, 1975–2002. Volume II: College students and adults ages 19–40 NIH Publication No 03-5376 2003 Accessed June 28, 2005
American College Health Association Subcommittee on Tobacco Intervention and Prevention Position statement of tobacco on college and university campuses Accessed June 10, 2004
American Cancer Society's Smoke-Free New England Campus Initiative Creating a Smoke-free Campus Accessed June 28, 2005
Halperin AC Rigotti NA US public universities' compliance with recommended tobacco-control policies J Am Coll Health 2003 51 181 188 12822709
Wechsler H Lee JE Rigotti NA Cigarette use by college students in smoke-free housing: results of a national study Am J Prev Med 2001 20 202 207 11275447 10.1016/S0749-3797(00)00313-5
Backinger CL Fagan P Matthews E Grana R Adolescent and young adult tobacco prevention and cessation: current status and future directions Tob Control 2003 12 IV46 53 14645940
McVey D Stapleton J Can anti-smoking television advertising affect smoking behaviour? controlled trial of the Health Education Authority for England's anti-smoking TV campaign Tob Control 2000 9 273 282 10982571 10.1136/tc.9.3.273
Borland R Balmford J Understanding how mass media campaigns impact on smokers Tob Control 2003 12 ii45 52 12878773
Goldman LK Glantz SA Evaluation of antismoking advertising campaigns JAMA 1998 279 772 777 9508154 10.1001/jama.279.10.772
Morrell HER Cohen LM Bacchi D West J Predictors of smoking and smokeless tobacco use in college students: a preliminary study using web-based survey methodology J Am College Health
Rigotti NA Lee JE Wechsler H US college students' use of tobacco products: results of a national survey JAMA 2000 284 699 705 10927777 10.1001/jama.284.6.699
Emmons KM Wechsler H Dowdall G Abraham M Predictors of smoking among US college students Am J Public Health 1998 88 104 107 9584013
Rigotti NA Regan S Majchrzak NE Knight JR Wechsler H Tobacco use by Massachusetts public college students: long term effect of the Massachusetts Tobacco Control Program Tob Control 2002 11 ii20 4 12034976
Wechsler H Kelley K Seibring M Kuo M Rigotti NA College smoking policies and smoking cessation programs: results of a survey of college health center directors J Am Coll Health 2001 49 205 212 11337895
Pinilla J Gonzalez B Barber P Santana Y Smoking in young adolescents: an approach with multilevel discrete choice models J Epidemiol Community Health 2002 56 227 232 11854347 10.1136/jech.56.3.227
| 16001977 | PMC1177969 | CC BY | 2021-01-04 16:28:55 | no | BMC Public Health. 2005 Jul 7; 5:74 | utf-8 | BMC Public Health | 2,005 | 10.1186/1471-2458-5-74 | oa_comm |
==== Front
BMC Plant BiolBMC Plant Biology1471-2229BioMed Central London 1471-2229-5-101595524610.1186/1471-2229-5-10DatabaseConstruction of a bacterial artificial chromosome library from the spikemoss Selaginella moellendorffii: a new resource for plant comparative genomics Wang Wenming [email protected] Milos [email protected] Meizhong [email protected] Nicholas [email protected] Hye Ran [email protected] Jing-Ke [email protected] Dave [email protected] Christopher [email protected] K [email protected] John [email protected] Clint [email protected] Pamphilis Claude [email protected] Dina [email protected] Jeff [email protected] Rod A [email protected] Jo Ann [email protected] Arizona Genomics Institute, Department of Plant Sciences, University of Arizona, Tucson, AZ 85721, USA2 Department of Botany and Plant Pathology, Purdue University, West Lafayette, IN 47907, USA3 Department of Biochemistry, Purdue University, West Lafayette, IN 47907, USA4 Benaroya Research Institute at Virginia Mason, 1201 Ninth Avenue, Seattle, WA 98101, USA5 Department of Biology and Huck Institutes of Life Sciences, Penn State University, University Park, PA 16802, USA6 Department of Biology and Center for Developmental Biology, University of Washington, Seattle, WA 98195, USA7 Department of Genetics, Biochemistry and Life Science Studies, Clemson University, Clemson, SC 29634, USA8 Cold Spring Harbor Laboratory, Cold Spring Harbor, NY 11724, USA9 Center for Biosystems Research, University of Maryland Biotechnology Institute, College Park, MD 20742, USA2005 14 6 2005 5 10 10 6 1 2005 14 6 2005 Copyright © 2005 Wang 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 lycophytes are an ancient lineage of vascular plants that diverged from the seed plant lineage about 400 Myr ago. Although the lycophytes occupy an important phylogenetic position for understanding the evolution of plants and their genomes, no genomic resources exist for this group of plants.
Results
Here we describe the construction of a large-insert bacterial artificial chromosome (BAC) library from the lycophyte Selaginella moellendorffii. Based on cell flow cytometry, this species has the smallest genome size among the different lycophytes tested, including Huperzia lucidula, Diphaiastrum digita, Isoetes engelmanii and S. kraussiana. The arrayed BAC library consists of 9126 clones; the average insert size is estimated to be 122 kb. Inserts of chloroplast origin account for 2.3% of the clones. The BAC library contains an estimated ten genome-equivalents based on DNA hybridizations using five single-copy and two duplicated S. moellendorffii genes as probes.
Conclusion
The S. moellenforffii BAC library, the first to be constructed from a lycophyte, will be useful to the scientific community as a resource for comparative plant genomics and evolution.
==== Body
Background
The lycophytes (class Lycopsida) are an ancient group of vascular plants that dominated the earth's flora during the Carboniferous period. The three orders of lycophytes that remain from this period include the homosporous Lycopodiales, the heterosporous Selaginellales and the heterosporous Isoetales. All of these plants are distinguishable from ferns and flowering plants by the presence of microphylls (as opposed to euphylls), the absence of leaf gaps and the absence of lateral roots. In common with ferns but not flowering plants, all lycophytes produce free-living spores, an independent gametophyte generation and non-integumented sporangia. Based upon the fossil record, the lycophytes are thought to have emerged during the early Devonian about 400 Myr ago prior to the evolution of leaves and roots in vascular plants [1,2]. Based on recent DNA-based phylogenetic analyses, the Lycopsida clade is monophyletic and sister to the fern/seed plant, or euphyllophyte, clade [3]. As representatives of the earliest and still-surviving vascular plant lineage, the lycophytes are an important group of plants for providing insights into the early evolution of land plants.
While genomic resources are available for many species of flowering plants, including the sequences of the Arabidopsis thaliana [4], rice [5,6] and poplar [7] genomes, very few resources exist for plants other than angiosperms. A draft genome sequence is available for Chlamydomonas reinhardtii [8], a chlorophytic green alga that is a distant relative of the charophytic algal group that gave rise to land plants [9]. The huge phylogenetic gap between the characterized genomes of Chlamydomonas reinhardtii and flowering plants greatly limits our ability to study how important features in plants originated and diversified at a genetic level. To help fill this gap, the genome sizes of several species of lycophytes were determined by cell flow cytometry in order to identify a lycophyte with a relatively small genome. Of those surveyed, the spikemoss Selaginella moellendorffii was found to have the smallest, with a nuclear genome size less than 127 Mbp. Here we describe the construction and characterization of a large insert BAC library for this species.
Results and discussion
Genome size estimates
The approximate genome sizes of several lycophyte species were determined by cell flow cytometry using as an internal standard the nuclei of other plants or cells with genomes of known sizes (Table 1). The homosporous Huperzia lucidula and Diphaiastrum digita (Lycopodiales) have the largest genomes of those surveyed, estimated to be 5585 and 2670 Mbp/1C, respectively. Isoetes engelmanii (Isoetales), a heterosporous lycophyte, has a genome size of 1710 Mbp/1C. The heterosporous Selaginella moellendorffii (Selaginellales) has the smallest genome of those surveyed, between 88 and 127 Mbp/1C. The three estimates of S. moellendorffii genome size (Table 1) vary depending on species used as the internal standard. We also confirmed that the genome size of S. moellendorffii is smaller than that of S. kraussiana (Table 1), which was previously reported to have a genome size ranging from 0.32–0.72 pg/2C, or 157–320 Mbp/1C [10]. The decision to construct a BAC library from the heterosporous S. moellendorffii was largely based on its having the smallest genome size of all lycophyte accessions examined here or reported previously.
BAC library construction and characterization
In constructing the BAC library, nuclear DNA was isolated from the growing tips of S. mollendorffii sporophytes, partially digested with Hind III and the library constructed and processed essentially according to Luo and Wing [11] as described in the Methods section. A total of 9126 BAC clones were picked, arrayed and processed for long-term storage. To estimate the insert sizes of this library, BAC DNAs were prepared from 410 randomly selected clones, digested with Not I and size-separated by pulse field gel electrophoresis. The Not I restriction patterns of 43 randomly selected clones are illustrated in Figure 1. Of the 410 BAC DNAs prepared, 23 yielded no DNA and 9 lacked an insert. Of the remaining 378 clones with inserts, the inserts ranged in size from 9 to 292 kb, with 75% having inserts sizes 90–159 kb and 84% with inserts greater than 90 kb. The average insert size of the clones with inserts was 122 +/- 44 kb (SD). The distribution of insert sizes is shown in Figure 2.
To estimate the extent of chloroplast and mitochondrial DNA contamination of the BAC library, the arrayed library was probed with two S. moellendorffi DNA fragments that contain either chloroplast- or mitochondrial-encoded genes (Table 1). The S. moellendorffii DNA fragment containing the chloroplast encoded ribosomal proteins S8, L2 and S19 hybridized to 207 BAC clones. The S. moellendorffii DNA fragment containing the mitochondria-encoded NADH DEHYDROGENASE SUBUNIT 5 gene did not hybridize to any BAC clones but did hybridize to itself (data not shown). It is unlikely that either fragment is of nuclear origin given that these genes are encoded by organelles in every plant species studied so far. The results of these hybridizations demonstrate that a very small proportion (2.3%) of the BAC inserts are of chloroplast origin. This is within the expected range for organellar DNA contamination in large insert DNA libraries [12]. The inability to detect clones that hybridize to mitochondrial DNA might reflect a mitochondrial genome that is small enough to be efficiently removed from the nuclear DNA preparation. However, the sizes of mitochondrial genomes in the lycophytes have yet to be reported.
To determine the number of genome equivalents represented by the BAC library, the arrayed library was probed with the seven S. moellendorffii genes listed in Table 2. All seven genes potentially encode proteins that are at least 43% identical in amino acid sequence to similar genes in angiosperms. The two genes homologous to GIBBERELLIN INSENSITIVE (SmGAI) and CYTOCHROME P450 98 (SmCYP98) were obtained by RT-PCR. The gene homologous to OXYANION TRANSLOCATION PROTEIN (SmOTP) was identified from an EST library (C. Chapple, unpub. obs.), while the genes homologous to the SHORTROOT (SmSHR), MAGNESIUM CHELATASE SUBUNIT H (SmChlH), ZINC-FINGER (SmZNF) and SYNTAXIN (SmSNT) gene fragments used as probes were identified from the sequences generated from the ends a sheared S. moellendorffii genomic library (J. Banks, unpub. obs.). Based upon the genomic DNA hybridization results shown in Figure 3, these genes are present in either one or two copies in the genome. The number of BAC clones that hybridized to the single-copy genes ranged from 7–12, and to about twice that number of clones (16 and 23) using the two duplicated genes as probes (Table 2). Based on the results of these experiments, the estimated number of genome equivalents represented by the BAC library is ~10. Assuming that one nuclear genome equivalent is represented by 871 BAC clones with nuclear DNA inserts and each BAC insert is 122 kb, we estimate that the S. moellendorffii genome is 106 Mbp. This estimate coincides well with the estimated size of the nuclear genome of S. moellendorffii based upon cell flow cytometry.
To our knowledge, this is the first Lycopsida BAC library constructed. A BAC library was recently published for Physcomitrella patens [13], a moss that has become a popular model system for functional genomics, biochemistry, evolutionary and developmental genetic studies (reviewed by [14]). As a representative of the early branching vascular plants, the S. moellendorffii library described here will link genomic resources from algae (Chlamydomonas reinhardtii) and moss (P. patens) to seed plants, including important crop species. The library also will be a useful resource for readily identifying genes that are involved in developmental, physiological and biochemical processes in the lycophytes and provide an important tool for the study of plant evolution. The nuclear genome of S. moellendorffi is currently being sequenced by the Department of Energy Joint Genome Institute using a shotgun sequencing approach [15]. The BAC library described here is currently available to the scientific community through the Arizona Genomics Institute [16].
Conclusion
We have shown that the lycophyte S. moellendorffii has a very small genome size, as small or smaller than that of Arabidopsis thaliana based on cell flow cytometry, and have constructed from this species a large insert BAC library that contains about 10 genome equivalents and has an average insert size of 122 kb.
Methods
Plant material
Selaginella moellendorffii plants were obtained from Plant Delights Nursery, Inc., Raleigh, NC. Huperzia lucidula plants were obtained from Carolina Biological Supply Company (Burlington, NC; referred to there as Lycopodium lucidulum). Diphaiastrum digita plants were obtained from Gar Rothwell (Ohio University, Athens, OH). Isoetes engelmannii plants were obtained from Gerald Gastony (Indiana University, Bloomington, IN). Once obtained, all plants were grown in a local greenhouse under 50% shade cloth.
Nuclear DNA content determination
The procedure used to analyze nuclear DNA content in plant cells was modified from Arumuganathan and Earle [17]. Glycine max, Oryza sativa cv Nipponbare or Arabidopsis thaliana or chicken red blood cell nuclei were used as internal standards. For flow cytometric analysis, 50 mg of fresh leaf tissue was placed on ice in a sterile 35 × 10 mm plastic petri dish. The tissue was sliced into 0.25 mm to 1 mm segments in a solution containing 10 mM MgSO4, 50 mM KCl, 5 mM Hepes, pH 8.0, 3 mM dithiothreitol, 0.1 mg ml-1 propidium iodide, 1.5 mg ml-1 DNase free RNase (Rhoche, Indianapolis, IN) and 0.25% (v/v) Triton X-100. The suspended nuclei were filtered through 30 μm nylon mesh and incubated at 37 C for 30 min before flow cytometric analysis. Suspensions of sample nuclei were each spiked with a suspension of standard nuclei (prepared in above solution) and analyzed with a FAC calibur flow cytometer (Becton-Dickinson, San Jose, CA). For each measurement, the propidium iodide fluorescence area signals (FL2-A) from 1000 nuclei were collected and analyzed by CellQuest software (Becton-Dickinson, San Jose, CA) on a Macintosh computer. The mean position of the G0/G1 (Nuclei) peak of the sample and the internal standard were determined by CellQuest software. The mean nuclear DNA content of each plant sample, measured in pg, was based on 1000 scanned nuclei.
BAC library construction
The growing tips (1 cm) of plants were harvested and flash frozen in liquid nitrogen prior to nuclei preparation. Purified nuclei were prepared according to Luo and Wing [11]. The embedding of nuclei, Hind III restriction enzyme digestion of DNA and the preparation of high molecular weight DNA fragments were performed according to Luo and Wing [11]. The Hind III cloning-ready single copy pIndigoBAC536 vector was prepared from the high copy pCUGIBAC1 plasmid as described by Luo et al. [18]. High molecular weight genomic DNA fragments were ligated to the vector and transformed into E. coli stain DH10B T1 phage-resistant cells (Invitrogen, Carlsbad, CA). Transformed colonies were picked and transferred into individual wells of 384 microtiter plates, grown and then stored at -80C. The BAC library was gridded onto 11.25 × 22.5 cm filters in high density, double spots and 4 × 4 patterns with a Genetix QB (Genetix, UK). To characterize the BAC inserts, BAC DNA samples were prepared with a Tomtec Quadra 96 model 320 (Tomtec, Hamden, CT) in a 96-well format, digested with Not I, separated on 1% agarose CHEF (Bio-Rad, Hercules, CA) gels at 5–15 sec linear ramp time, 6 V/cm, 14C in 0.5 × TBE buffer for 16 hours and stained with ethidium bromide.
DNA hybridizations
Genomic DNA for gel blot analysis and PCR was isolated from S. moellendorffii plants using the Nucleon Phytopure kit (Amersham Biosciences, Piscataway, NJ). RNA was isolated using the RNeasy Plant Mini Kit (Qiagen, Valencia, CA); cDNA was synthesized and RT-PCR performed using the cMaster RTplusPCR System (Eppendorf, Westbury, NY). The 479 bp SmGAI cDNA fragment (GenBank accession AY874058) was initially obtained by RT-PCR using the primers 5'cayttyacigciaaycargci3' and 5' tcraaiarigcrctrtartartg3'. The 473 bp SmCYP98 cDNA fragment (GenBank accession AY843208) was obtained by RT-PCR using the primers 5'gtdgcvttcaacaacatwac3' and 5'ccatnccwgchgtgatcat3'. In all cases, y = c or t, r = a or g, d = a, g or t; v = a, c or g; w = a or t; n = a, t, g or c. All PCR products were cloned into the pGEM-T EASY vector (Promega, Madison, WI) and sequenced. Several genes used as probes were generated by PCR using genomic DNA as template. The 302 bp genomic SmSHR fragment (GenBank accession AY877259) was obtained using the primers 5'ggtggacctctcctctcctc3' and 5'atccaggtttgtagcgcttg3', the 254 bp genomic SmZNF fragment (GenBank accession AY877260) was obtained using the primers 5'gaggtcgtctccttgtcacc3' and 5'cggcgaaagtgtttcttgat3', the 395 bp genomic SmChlH fragment (GenBank accession AY877262) was obtained using the primers 5'ggatcgccttcatatccaaa3' and 5'aaactcgcggtcacagtctt3', and the 375 bp genomic SmSNT fragment (GenBank accession AY877261) was obtained using the primers 5'gatccaggccaagatgaaga3' and 5'tgccagtgaccgtgaagtag3'. The SmOTP gene (accession number AY877263) was identified from a S. moellendorffii EST library; the entire insert was used as a probe. The S. moellendorffii chloroplast (GenBank accession AY877264) and mitochondrial (GenBank accession AY877265) fragment sequences were identified from a partially sequenced, small-insert, sheared genomic library. For DNA blot hybridizations, each cloned insert was gel purified and labeled with 32P using the Megaprime DNA Labelling System (Amersham Biosciences, Piscataway, NJ). For DNA gel blots, 4 μg of S. moellendorffii genomic DNA was digested with restriction enzymes, fractionated and alkaline transferred to nylon membranes according to Sambrook and Russell [19]. All filters were hybridized at 65C in a solution containing 0.5 M phosphate buffer (pH 7.2), 7% (w/v) SDS, and 1 mM EDTA. All membranes were washed under stringent conditions (0.1X SSC, 0.1% SDS, 65C).
Authors' contributions
JAB, JC, CC, CD, DM, JT and RAW participated in the design of this study. KA performed the cell flow cytometry genome size determination. ML and WW constructed the BAC library and DK, CM, HK, NS, ML, MT, WW and JW participated in library characterization. All authors have read and approved of this study.
Acknowledgements
We thank Kiran Rao, Cari Soderlund, Angelina Angelova, Amber Hopf and Luke Gumaelius for their assistance in developing the BAC library resource. This work was supported by the Purdue Agricultural Research Programs and National Science Foundation grant numbers 0207110 (Banks), 0211611 (Wing), and 0208502 (Mandoli). This is manuscript number 2005-17581 of the Purdue Agricultural Research Program.
Figures and Tables
Figure 1 Not I digests of BAC DNA isolated from 43 random BAC clones from the S. moellendorffii BAC library. The vector contains two Not I sites that flank the Hind III cloning sites. The vector is present in all samples with the exception of the first and last lanes, which contain the DNA marker PFG midrange I (New England Biolabs, Beverly, MA).
Figure 2 The distribution of BAC insert sizes from 410 randomly chosen BAC clones.
Figure 3 Results of DNA blots to determine the gene copy number of various DNA fragments in the S. moellendorffii genome. DNA from S. moellendorffii was digested with EcoR V, EcoR I, Nco I or Hind III and probed with the SmSNT fragment (panel A), the SmSHR fragment (panel B), the SmZNF fragment (panel C), the SmChlH fragment (panel D), the SmGAI fragment (panel E), the SmOTP cDNA (panel F), and the SmCYP98 cDNA (panel G). The EcoR V digest is absent in panel A. The left-most lanes in panels B, C and F contain molecular weight markers.
Table 1 Results of cell flow cytometry to determine the nuclear genome sizes of some lycophytes.
Sample species Internal standard (pg/2C) DNA content of sample (pg/2C); n1; SD DNA content of sample species (Mbp/1C)2
Huperzia lucidula Chick red blood cells (2.333) 11.40; 4; 0.036 5586
Diphaiastrum digita Chick red blood cells (2.333) 5.45; 4; 0.017 2670
Isoetes engelmanii Glycine max leaf (2.254) 3.49; 4; 0.005 1710
Selaginella moellendorffii Arabidopsis thaliana flower buds (0.365)
Oryza sativa leaf (1.06)
Glycine max leaf (2.254) 0.18; 8; 0.009
0.25; 4; 0.25
0.26; 4; 0.006 88
123
127
Selaginella kraussiana Oryza sativa leaf (1.06)
Glycine max leaf (2.254) 0.43; 4; 0.006
0.49; 4; 0.003 211
240
1nuclei from sample and internal standard source plants or cells were prepared; n represents the number of times each nuclei preparation was sampled by cell flow cytometry.
2assumes that 1 pg = 980 Mbp
3Galbraith et al. [20]
4Bennett and Leitch [21]
5Bennett et al. [22]
6Bennett et al. [23]
Table 2 Results of DNA hybridizations to determine gene copy number and the number of genome equivalents represented in the arrayed BAC library.
Gene probe: Homologous to following plant genes (e value; % amino acid identity)1: Genome copy number: Number of positive BAC clones from arrayed library:
SmGAI cDNA GIBBERELLIC ACID INSENSITIVE (8e-34; 43) 1 14
SmCYP98 cDNA CYTOCHROME P450 98A3 (1e-40; 66) 2 23
SmOTP OXYANION TRANSLOCATION PROTEIN (1e-36; 53) 2 16
SmSHR SHORTROOT (3e-45; 47) 1 7
SmZNF Zn-FINGER PROTEIN (1e-69; 90) 1 7
SmChlH Mg CHELATASE SUBUNIT H (1e-130; 85) 1 8
SmSNT SYNTAXIN (9e-55; 54) 1 12
S. moellendorffii chloroplast fragment2 Contains:
Ribosomal protein L22 (2e-33; 56)
Ribosomal protein L2 (1e-33; 71)
Ribosomal protein S19 (2e-31; 71) ND 207
S. moellendorffii mitochondrial fragment2 Contains NADH DEHYDROGENASE subunit 5 (6e-61; 71) ND 0
1Gene listed is based on results of blastx using the S. moellendorffii sequence as a query against the NR database and default parameters [24].
2sequence described is part of a genomic fragment <5 kb that was used as a probe.
==== Refs
Kenrick P Crane PR The origin and early diversification of plants on land Nature 1997 389 33 39 10.1038/37918
Stewart W Rothwell GW Paleobotany and the Evolution of Plants 1993 2nd Cambridge, Cambridge University Press
Pryer KM Schneider H Smith AR Cranfill R Wolf PG Hunt JS Sipes SD Horsetails and ferns are a monophyletic group and the closest living relatives to seed plants Nature 2001 409 618 622 11214320 10.1038/35054555
Arabdiopsis Genome Initiative Analysis of the genome sequence of the flowering plant Arabidopsis thaliana. Nature 2000 408 796 815 11130711 10.1038/35048692
Goff SA Ricke D Lan TH Presting G Wang R Dunn M Glazebrook J Sessions A Oeller P Varma H Hadley D Hutchison D Martin C Katagiri F Lange BM Moughamer T Xia Y Budworth P Zhong J Miguel T Paszkowski U Zhang S Colbert M Sun WL Chen L Cooper B Park S Wood TC Mao L Quail P Wing R Dean R Yu Y Zharkikh A Shen R Sahasrabudhe S Thomas A Cannings R Gutin A Pruss D Reid J Tavtigian S Mitchell J Eldredge G Scholl T Miller RM Bhatnagar S Adey N Rubano T Tusneem N Robinson R Feldhaus J Macalma T Oliphant A Briggs S A draft sequence of the rice genome (Oryza sativa L. ssp. japonica) Science 2002 296 92 100 11935018 10.1126/science.1068275
Yu J Hu S Wang J Wong GK Li S Liu B Deng Y Dai L Zhou Y Zhang X Cao M Liu J Sun J Tang J Chen Y Huang X Lin W Ye C Tong W Cong L Geng J Han Y Li L Li W Hu G Li J Liu Z Qi Q Li T Wang X Lu H Wu T Zhu M Ni P Han H Dong W Ren X Feng X Cui P Li X Wang H Xu X Zhai W Xu Z Zhang J He S Xu J Zhang K Zheng X Dong J Zeng W Tao L Ye J Tan J Chen X He J Liu D Tian W Tian C Xia H Bao Q Li G Gao H Cao T Zhao W Li P Chen W Zhang Y Hu J Liu S Yang J Zhang G Xiong Y Li Z Mao L Zhou C Zhu Z Chen R Hao B Zheng W Chen S Guo W Tao M Zhu L Yuan L Yang H A draft sequence of the rice genome (Oryza sativa L. ssp. indica) Science 2002 296 79 92 11935017 10.1126/science.1068037
JGI Populus trichocarpa v1.0
JGI Chlamydomonas reinhardtii v2.0 2005
Qui YL Palmer JD Phylogeny of early land plants: insights from genes and genomes Trends in Plant Science 1999 4 26 30 10234267 10.1016/S1360-1385(98)01361-2
Obermayer R Leitch IJ Hanson L Bennett MD Nuclear DNA C-values in 30 species double the familial representation in pteridophytes Ann Bot (Lond) 2002 90 209 217 12197518 10.1093/aob/mcf167
Luo M Wing R An improved method for plant BAC library construction. Methods in Molecular Biology 2003 236 3 20 14501055
Zhang X Choi S Woo S Li Z RA W Construction and characterization of two rice bacterial artificial chromosome libraries from the parents of a permanent recombinant inbred mapping population Molecular Breeding 1996 2 11 24
Liang C Xi Y Shu J Li J Yang J Che K Jin D Liu X Weng M He Y Wang B Construction of a BAC library of Physcomitrella patens and isolation of a LEA gene Plant Science 2004 167 491 498 10.1016/j.plantsci.2004.04.015
Schaefer DG Zryd JP The moss Physcomitrella patens, now and then Plant Physiology 2001 127 1430 1438 11743086 10.1104/pp.127.4.1430
JGI Community Sequencing Program Sequencing Plans for 2005
Arizona Genomics Institute Ordering Website
Arumuganathan E Earle ED Estimation of nuclear DNA content of plants by flow cytometry. Plant Molecular Biology Reporter 1991 9 229 233
Luo M Wang Y Frisch D Joobeur T Wing RA Dean RA Melon bacterial artificial chromosome (BAC) library construction using improved methods and identification of clones linked to the locus conferring resistance to melon Fusarium Wilt Genome 2001 44 154 162 11341724 10.1139/gen-44-2-154
Sambrook J Russell DW Molecular cloning : a laboratory manual 2001 1 3 Cold Spring Harbor, N.Y, Cold Spring Harbor Laboratory Press
Galbraith DW Harkins KR Maddox JM Ayres NM Sharma DP Firoozabady E Rapid flow cytometric analysis of the cell cycle in intact plant tissues. Science 1983 220 1049 1051
Bennett MD Leitch IJ Nuclear DNA amounts in angiosperms - 583 new estimates Annals of Botany 1997 8 169 196 10.1006/anbo.1997.0415
Bennett MD Leitch IJ Price HJ Johnston JS Comparisons with Caenorhabditis (approximately 100 Mb) and Drosophila (approximately 175 Mb) using flow cytometry show genome size in Arabidopsis to be approximately 157 Mb and thus approximately 25% larger than the Arabidopsis genome initiative estimate of approximately 125 Mb Annals of Botany 2003 91 547 557 12646499 10.1093/aob/mcg057
Bennett MD Smith JB Nuclear DNA amounts in angiosperms Philosophical Transactions of the Royal Society of London B 1991 334 309 345
Altschul SF Gish W Miller W Myers EW Lipman DJ Basic local alignment search tool Journal of Molecular Biology 1990 215 403 410 2231712 10.1006/jmbi.1990.9999
| 15955246 | PMC1177970 | CC BY | 2021-01-04 16:03:52 | no | BMC Plant Biol. 2005 Jun 14; 5:10 | utf-8 | BMC Plant Biol | 2,005 | 10.1186/1471-2229-5-10 | oa_comm |
==== Front
BMC Plant BiolBMC Plant Biology1471-2229BioMed Central London 1471-2229-5-71590720310.1186/1471-2229-5-7Research ArticleLight-induced morphological alteration in anthocyanin-accumulating vacuoles of maize cells Irani Niloufer G [email protected] Erich [email protected] Department of Plant Cellular and Molecular Biology and Plant Biotechnology Center, The Ohio State University, Columbus, OH 43210, USA2005 20 5 2005 5 7 7 3 2 2005 20 5 2005 Copyright © 2005 Irani and Grotewold; licensee BioMed Central Ltd.2005Irani and Grotewold; 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
Plant pigmentation is affected by a variety of factors. Light, an important plant developmental signal, influences the accumulation of anthocyanins primarily through the activation of the transcription factors that regulate the flavonoid biosynthetic pathway. In this study, we utilized maize Black Mexican Sweet (BMS) cells expressing the R and C1 regulators of anthocyanin biosynthesis from a light-insensitive promoter as a means to investigate the existence of additional levels of control of pigmentation by light.
Results
BMS cells expressing the R and C1 regulators from the CaMV 35S constitutive promoter accumulate anthocyanins when grown in complete darkness, suggesting that the transcription factors R and C1 are sufficient for the transcription of the genes corresponding to the structural enzymes of the pathway, with no requirement for additional light-induced regulators. Interestingly, light induces a "darkening" in the color of the purple anthocyanin pigmentation of transgenic BMS cells expressing R and C1. This change in the pigment hue is not associated with a variation in the levels or types of anthocyanins present, or with an alteration of the transcript levels of several flavonoid biosynthetic genes. However, cytological observations show that light drives unexpected changes in the morphology and distribution of the anthocyanins-containing vacuolar compartments.
Conclusion
By uncoupling the effect of light on anthocyanin accumulation, we have found light to induce the fusion of anthocyanin-containing vacuoles, the coalescence of anthocyanic vacuolar inclusion (AVI)-like structures contained, and the spread of anthocyanins from the inclusions into the vacuolar sap. Similar light-induced alterations in vacuolar morphology are also evident in the epidermal cells of maize floral whorls accumulating anthocyanins. Our findings suggest a novel mechanism for the action of light on the vacuolar storage of anthocyanin.
==== Body
Background
Anthocyanins, the coloured end product of the flavonoid pathway, play an important role in attracting insects or animals for pollination and seed dispersal. In addition, they play roles as anti-oxidants and in protecting DNA and the photosynthetic apparatus from high radiation fluxes [1]. Other possible functions of anthocyanins, such as the protection against cold stress or providing drought resistance, are likely to be associated with activities restricted to particular classes of plants [2].
The biosynthesis of the flavonoids, a large phenylpropanoid-derived group of phenolic compounds, provides one of the best described plant metabolic pathways, with many of the structural and regulatory genes in the pathway identified and cloned [3,4]. Less is known regarding the mechanisms by which the water-soluble anthocyanins are transported from their site of synthesis, the cytoplasmic surface of the endoplasmic reticulum [5,6], to the vacuole, where they are usually sequestered [7]. Plant vacuoles are highly dynamic, multifunctional organelles that are the primary storage and turnover sites of macromolecules. These membrane-bound organelles, which can occupy up to 90% of the total cellular volume, are integral part of the endomembrane system, serving as the terminal products of the secretory pathway [8].
Several plant species store anthocyanins within vacuolar inclusions that have been loosely termed anthocyanoplasts which have been observed to start as vesicles in the cytoplasm and appear to be membrane bound [9,10]. More recently, the intravacuolar structures observed in the flower petals of various plants, including carnation and lisianthus, were termed anthocyanic vacuolar inclusions, or AVIs [11]. These inclusions were suggested to be membrane-less, proteinaceous matrixes that acted as anthocyanin traps, preferentially for anthocyanidin 3, 5-diglycosides [11] or acylated anthocyanins [12]. Once in the vacuole, many factors influence the in vivo pigmentation provided by anthocyanins, with important consequences for the eco-physiology of plants [13,14]. Some of these factors include the particular type of anthocyanidin present (for e.g., pelargonidin cyanidin, or myricetin), the shape of the cells in which the pigments accumulate [15], the concentration of the pigment, the formation of complexes between anthocyanins and co-pigments [16], and the vacuolar pH[17,18].
Environmental conditions are known to induce the accumulation of anthocyanin pigments across the major groups of higher plants, of them light being the best studied [2,19,20]. In Arabidopsis, the anthocyanin pathway is regulated in a circadian fashion, with flavonoid genes peaking at the end of the dark cycle, likely preparing plants for daybreak [21]. In maize, members of the PL1/C1 R2R3 MYB and B/R bHLH families of regulatory proteins cooperate for the regulation of anthocyanin pigments [3] and are necessary for the expression of the anthocyanin biosynthetic genes c2 (chalcone synthase), chi1 (chalcone isomerase), f3h (flavanone 3-hydroxylase) a1 (dihydroflavonol 4-reductase), a2 (leucoanthocyanidin dioxygenase/anthocyanin synthase), bz1 (UDP glucose:flavonoid 3-O-glucosyl transferase) and bz2 (glutathione S-transferase) [19]. The light-induced expression of members of these R2R3 MYB and bHLH classes of transcription factors has been proposed to be responsible for the induction of anthocyanins in maize by light [22-26].
Here, we investigated whether light affects the accumulation of maize anthocyanin pigmentation at a level other than through the activation of the known R2R3 MYB and bHLH regulators of the pathway. Towards this goal, we analyzed previously described [27] transgenic Black Mexican Sweet (BMS) maize cells in culture expressing the Zea mays R and C1 regulators from a constitutive, light insensitive CaMV 35S promoter (BMS35S::R+35S::C1). We show here that BMS35S::R+35S::C1 cells are red, even when grown in complete darkness. Upon light treatment, there is a darkening of the color of the BMS35S::R+35S::C1 cells, without an appreciable increase in the quantity of anthocyanins or in the type of anthocyanidins present. Consistent with these findings, the steady-state levels of several anthocyanin biosynthetic genes does not increase upon light treatment. Interestingly, at the subcellular level, light induces an alteration in the way the anthocyanins are distributed within vacuolar compartments. A similar alteration in the morphology of anthocyanin-accumulating vacuoles is observed when maize tassel glumes are irradiated with light, suggesting that the phenomena observed in BMS35S::R+35S::C1 cells in culture also occurs in planta. Together, our findings suggest a novel mechanism for the action of light on the packaging of anthocyanins in the vacuole and in subvacuolar compartments. This effect of light could only be uncovered after making the biosynthesis of anthocyanins independent of light.
Results
BMS cells expressing the transcription factors R and C1 accumulate anthocyanins in the dark
To determine whether the light control of the maize anthocyanin pathway is mediated by the expression of the B/R and C1/PL regulators and/or the biosynthetic genes [23], we investigated the pigmentation of BMS cells expressing the R and C1 genes from the constitutive CaMV 35S promoter (BMS35S::R+35S::C1) [27]. BMS35S::R+35S::C1 cells grown in complete darkness for 30 days were fully pigmented with anthocyanins (Fig. 1A). The bombardment of BMS cells with the R and C1 regulators driven from the 35S promoter (p35SR + p35SC1) resulted in the accumulation of red cells within 15 hours, even when cells were kept in complete darkness after bombardment (compare Fig. 1B and 1C). These results indicate that the constitutive expression of the R and C1 regulators is sufficient for the activation of the pathway, even in the absence of light.
Figure 1 Anthocyanins accumulate in maize BMS35S::R+35S::C1 cells in the dark. (A) Dark-grown BMS35S::R+35S::C1 cells expressing the R and C1 anthocyanin regulators from the CaMV 35S promoter accumulate anthocyanins. Transient expression of the R and C1 regulators in BMS cells by microprojectile bombardment induce anthocyanins in the dark (B) or the light (C). The magnification bars represent 200 μm. The bar in the inset DIC image is 20 μm.
BMS35S::R+35S::C1 cells darken in the light
To investigate whether light has any additional effect on the pigmentation present in BMS35S::R+35S::C1 cells, we compared the color of BMS and BMS35S::R+35S::C1 cells grown for six days in complete darkness or under light conditions (see Methods). BMS35S::R+35S::C1 cells grown in the light showed a visual darkening, when compared to cells grown in the dark (Fig. 2, BMS35S::R+35S::C1). However, light did not induce any visible difference in the white-yellow color of the control BMS cells not expressing the C1 and R regulators (Fig. 2, BMS). We quantified these visual color differences in vivo using a reflectometer and the CIELAB color space value system. L* values, representing the lightness level, were reproducibly not affected by light (Table 1). The a* values for BMS35S::R+35S::C1 cells were positive (+a), consistent with the red color characteristic of these cells. There was however a significant (p < 0.05) reduction in the a-values when the BMS35S::R+35S::C1 cells were exposed to light, which was observed in each of the three times that the experiment was performed (Table 1). Although the b* values, contributing to yellow (+b) or blue (-b), were significantly different (p < 0.05) between the dark and light grown BMS35S::R+35S::C1 cells, they hovered near the zero value, suggesting a low contribution to the overall observed color shift. Thus, the a* and the b* values observed corresponded to a quantifiably red color (dull red), with a decreased degree of redness in the light, which is in agreement with our visual observations (Fig. 2). In the absence of a change in the L* value, the apparent darkening of the cells is likely caused by the spectral shift of the reflected light towards the less red.
Table 1 Reflectance analysis and L* a* b* values of dark- and light-grown BMS35S::R+35S::C1 cells in the CIELAB color scale. The a* value contributes to red (a+) or green (a-), the b* value contributes to yellow (+b) or blue (-b) and L* represent the lightness level.
BMS35S::R+35S::C1 L* a* b*
Dark 21.42 ± 1.29 12.06 ± 1.51 0.45 ± 0.48
Light 21.01 ± 0.93 7.02 ± 0.88 -0.45 ± 0.21
Figure 2 Light induces the darkening of anthocyanin pigmentation. Images of BMS control and BMS35S::R+35S::C1 cells grown for ten days under total darkness (Dark) or light (Light) conditions.
The anthocyanin contents or mRNA steady state levels of biosynthetic genes are not altered by white light in BMS35S::R+35S::C1 cells
To investigate whether the change in color observed between dark- and light grown BMS35S::R+35S::C1 cells is due to an alteration in the quality or quantity of the pigments, anthocyanins were extracted and quantified. Methanolic extracts of the cells, normalized for dry weight, were analyzed spectrophotometrically. The absorption spectra of pigments obtained from acidic methanol extracts of dark- and light-treated cells showed identical profiles and very similar absorbance values at 530 nm (Fig. 3A, B). To determine whether light induced an alteration in the type of anthocyanidin present, total anthocyanidins were extracted and separated by thin layer chromatography (TLC). Similar levels of cyanidin and pelargonidin are present in light- and dark treated BMS35S::R+35S::C1 cells (Fig. 3C), consistent with the two major types of anthocyanidins previously described in BMS35S::R+35S::C1 cells [28].
Figure 3 Similar quantities of cyaniding and pelargonidin accumulate in dark and light-grown BMS35S::R+35S::C1 cells. (A) Spectral profile of methanol-HCL extracts of dark- (black line) or light-grown (green line) BMS35S::R+35S::C1 cells. The red line corresponds to the spectra of control BMS extracts. (B) Quantitation of anthocyanins in dark- and light-grown BMS35S::R+35S::C1 cells. Bars represent the SD of measurements obtained in three independent experiments. (C) Qualitative analysis of the anthocyanins present in dark- and light-grown BMS35S::R+35S::C1 cells by TLC.
In BMS cells, R and C1 are known to activate several flavonoid biosynthetic genes [27]. To investigate whether the activation of these genes was further enhanced by the light treatment, total RNA was extracted from BMS35S::R+35S::C1 cells incubated for six days in the light or the dark and analyzed by RNA gel blots (Northern blots, Fig. 4). No significant alteration in the steady-state mRNA levels for chalcone synthase (c2, Fig. 4), flavanone 3-hydroxylase (f3h, Fig. 4) or dihydroflavonol 4-reductase (a1, Fig. 4) was observed for the light grown BMS35S::R+35S::C1 cells. The control BMS cells showed no mRNA accumulation for these genes (Fig. 4). Together, these results suggest that the darkening of the BMS35S::R+35S::C1 cells upon light treatment is not due to an alteration in the quantity or quality of the anthocyanin pigments present.
Figure 4 Northern analysis of flavonoid biosynthetic genes show no alterations in the steady-state mRNA levels induced by light in BMS35S::R+35S::C1 cells. Total RNA from dark- (D) and light-grown (L) control BMS and BMS35S::R+35S::C1 cells were analyzed by Northern with probes corresponding to the c2 (chalcone synthase), f3h (flavanone 3-hydroxylase) and a1 (dihyroflavonol 4-reductase) genes. Ubiquitin (ubq) was used as a normalizing control. The numbers indicate the relative hybridization signal (in arbitrary units) obtained with each probe, normalized for the corresponding signal obtained with ubq.
White light induces alterations in the sub-cellular distribution and vacuolar organization of anthocyanins in BMS35S::R+35S::C1 cells
Alterations in rose flower pigmentation were associated previously with the formation of AVI-like structures [29]. To investigate whether a similar alteration in the packing of anthocyanins could be with the light-induced hue alteration of BMS35S::R+35S::C1 cells, we investigated the sub-cellular morphology of dark- and light-treated BMS35S::R+35S::C1 cells. To unequivocally visualize the vacuole(s), BMS and BMS35S::R+35::C1 were stained with the cell permeable, acetoxymethyl derivative of the fluorescent vacuolar dye, 2',7'-bis(2-carboxyethyl)-5(6)-carboxyfluorescein (BCECF-AM). In the control BMS cells, there are typically one to a few large vacuolar compartments (Fig. 5A, C). In contrast, BMS35S::R+35::C1 cells are always multi-vacuolated (Fig. 5B, D). Unfortunately, anthocyanins concentrated in the vacuolar inclusions quench the fluorescence of the BCECF-AM dye, as observed in some of the larger vacuoles (Fig. 5B).
Figure 5 BMS35S::R+35S::C1 cells have multiple vacuoles. Laser scanning confocal microscopy (false colored) images of (A) BMS and (B) BMS35S::R+35S::C1 cells loaded with 10 μM of the vacuolar dye BCECF-AM. Laser scanning confocal 'light transmitted' images of (C) BMS and (D) BMS35S::R+35S::C1 cells are shown in black and white. The bar represent 20 μm.
Within the vacuole, anthocyanins accumulate in inclusions that, when observed under polarized light, appear round and regular in shape or aggregated, like an intertwine of fine strings with blebs (Fig. 6). The number of vacuoles and inclusions per vacuole are dramatically affected by light in BMS35S::R+35S::C1 cells. In dark-treated BMS35S::R+35::C1 cells, the major representative cell type (Fig. 6A, representing 34% of all the 239 cells observed), had a range of 20 to 30 observable vacuoles (Table 2). In these cells, the anthocyanins were present mainly in the rounded vacuolar inclusions with a characteristic pale pink coloration of the vacuolar sap, a phenomenon called here "anthocyanin spread". The next two most abundant cell types present in dark-treated BMS35S::R+35::C1 cells, corresponding to 16% and 11% of all the cells, had a range of 10–20 or 20–30 visually observable vacuoles per cell, respectively (Table 2). Cells of these groups are characterized by no observable anthocyanin spread in the vacuolar sap under the light microscope, yet had red or pale red anthocyanin inclusions in the vacuole (Fig. 6C, E).
Table 2 Vacuole distribution in dark- and light-grown BMS35S::R+35S::C1 cells. The classification is based on the number of vacuoles and the presence or absence of anthocyanins in the vacuolar sap. The A-F letters correspond to the panels in Fig. 6, and the presence (+) or absence (-) of anthocyanins in the vacuolar sap (spread) is indicated.
% Representation No. of vacuoles (range) Anthocyanin spread
35S::R+C1 Dark (n = 239)
A 34% 20–30 +
C 16% 10–20 -
E 11% 20–30 -
35S::R+C1 Light (n = 400)
B 35% 1–10 ++
D 26% 10–20 +
F 12% 1–10 +++
Figure 6 Light induces alterations in the distribution of anthocyanins within vacuolar compartments. Representative DIC images of six day old (A, C, E) dark-grown and (B, D, F) light-grown BMS35S::R+35S::C1 cells. BMS cells grown in the (G) dark or (H) light show no significant alterations. Numbers correspond to the percentage values indicated in Table 2. The bar represents 20 μm.
In the light, the majority of the BMS35S::R+35::C1 cells (35%, Table 2) show one to ten pigmented vacuoles with red inclusions that appeared most of time diffuse and like "tangled strings" (Fig, 6B). These structures were determined to be 0.1 to 0.3 μm in diameter, appeared in certain cells to be branched with bleb-like structures at their ends. A significant population of cells (26%, Table 2) had 10–20 vacuoles, lightly colored with discrete deeply pigmented, spherical inclusions (Fig. 6D). The third most abundant class of BMS35S::R+35::C1 cells present in the light (12%, Table 2) showed deeply pigmented vacuoles and enlarged inclusions (Fig. 6F).
Comparison of the size of the vacuolar inclusions in the dark- and light-grown BMS35S::R+35::C1 cells (Fig. 7) shows that the majority of the inclusions in the dark grown samples are in the size range of 0.1 μm – l μm, with a noticeable absence of larger ones. The modal range of the vacuolar inclusions in light grown cells was 2 μm – 3 μm, with some as large as 14 μm. This may reflect the fusion of smaller vacuolar inclusions to give rise to larger ones (see below). The control BMS cells, grown in the dark or light, did not show any evident sub-cellular morphological changes (not shown).
Figure 7 Comparison of the size distribution of vacuolar inclusion containing anthocyanins in BMS35S::R+35S::C1 cells grown in the light or the dark. Round vacuolar inclusions were measured and classified according to the size range measured in microns (μm) that they fell into. Blue and red bars represent the AVI sizes in dark- and light-grown cells, respectively.
To further understand the changes induced by light in anthocyanin-accumulating cells, we performed laser scanning confocal microscopy of BCECF-AM loaded BMS35S::R+35::C1 cells. Dark-grown cells showed the presence of multiple vacuoles (Fig. 8G), which, upon light treatment, appeared to coalesce to form fewer, much larger vacuoles (Fig. 8H). Light-grown cells display a decrease in fluorescence, likely because of quenching by the anthocyanins released from the vacuolar inclusions. Dark- or light-grown BMS cells showed no distinctive differences in visual fluorescence intensity or vacuolar morphology (Fig. 8 C, D), suggesting that the observed morphological alterations are either a consequence of the expression of the R and C1 regulators, of the accumulation of anthocyanins or of distinctive properties of the vacuoles in which anthocyanins accumulate.
Figure 8 Morphology of vacuoles of dark and light grown BMS and BMS35S::R+3ss::C1 cells loaded with BCECF-AM.
Together, these results show that light-exposed BMS35S::R+35::C1 cells have a significant reduction in the number of vacuoles with a associated increase in their size, a change in the number, shape and size of the AVIs and a release of anthocyanins from the AVIs into the vacuole.
Light-induced vacuolar morphological alterations in anthocyanin-accumulating maize floral organs
To establish whether the light-induced vacuolar alterations observed in BMS cells could be observed also in planta, we looked at the tassels of maize B-I Pl plants that accumulated large quantities of anthocyanins (Fig. 9). The inner, light-protected lemma and palea (Fig. 9B) were the choice of material to observe the light-induced alterations in vacuolar morphology. The epidermal cells of these appendages in a C2-Idf mutant lacking anthocyanins (Fig. 9C) have one to a few large, observable, colorless, central vacuoles (Fig. 9D). In contrast, depending on the physiological and developmental stage of the florets, the epidermal cells of the lemma or palea of the B-I Pl florets (Fig. 9B, E) were either already filled with anthocyanins, or were in the initial stages of accumulation (Fig 9F). These cells show a distinctive multi-vacuolar morphology, and the vacuoles were often heavily loaded with anthocyanins and AVI-like structures (Fig. 9F).
Figure 9 Morphology of anthocyanin-accumulating cells in maize floral organs. The maize male inflorescence, the tassel, is a panicle made of numerous spikes, each formed by numerous, paired spikelets. (A) Single spike with paired spikelets. The spikelet has outer and inner glumes (bracts of the florets) and each floret has a lemma, a palea, a highly reduced lodigule and 3 stamens (B) Spikelet dissection: the two florets with an outer glume, an inner glume; each floret with a lemma, palea, highly reduced lodigule and three stamens. (C) Digital macro images of open florets from a C2-Idf (chalcone synthase) mutant that accumulates no anthocyanins. (D) DIC light micrographs of the lemma from male flowers of C2-Idf plant (E). Digital macro images of open florets from a B-I Pl plant. (F) DIC light micrographs of the lemma from male flowers of a B-I Pl plant. The bar represents 10 μm
Light from the microscope was sufficient to induce dramatic alterations in the morphology and distribution of the anthocyanin-containing vacuolar structures, as seen in the time-lapse images taken at four-second intervals (Fig. 10, see Additional file 1: Movie 1 for the original data used to perform this analysis). In these series, the initially thin, tubular anthocyanin-filled structures (average diameter of 0.6 μm, Fig. 10, green arrows) expand to a thickness of about 1.4 μm, dynamically filling the entire cell (Fig. 10A, blue arrows). These thick, finger-like projections became swollen, sheet-like structures (~3.3 μm in diameter, Fig 10, black arrows), to then become rounded (Fig 10A, B, C; orange arrows) and fuse with each other. These rounded/oval compartments measuring one 1–9 μm in diameter displayed, just like the tubular structures, dynamic morphological changes. Swollen "blebs" were observed moving along fine tubules and the ends of the thick tubules swelled up into round structures. Fusion events, once initiated, were very rapid, which resulted in the formation of large fusion bodies (Fig. 10C, D, E; red arrows) containing numerous clear (i.e., no anthocyanin-containing) structures (Fig 10C, D, E; yellow arrows). These clear inclusions were also observed initially in the sheet-like structures and were formed as the size of the tubules grew. The fusion bodies progressed rapidly to fill the entire cell, finally coalescing together resulting in the anthocyanin spread throughout the compartment (Fig 10E). The defined margins around the large central AVI-like structure (~15 μm across) become more diffuse with a lighter translucent red halo around the opaque, dark body (Fig. 10E).
Figure 10 Sub-cellular morphology of BI Pl maize floral cells accumulating anthocyanins. DIC images of a maize lemma from B-I/B-Peru plant over-accumulating anthocyanins. The above are extracted images from a time-lapse series (See Additional file 1: Movie 1). The time points on the images indicate the period from time 0' i.e. when the sample was mounted onto the stage and exposed to the microscope light. (A) and (B) occur earlier in the series while (D), (C) and (E) are in rapid succession (24 seconds apart). The large central inclusion corresponds to a vacuolar inclusion containing anthocyanins measuring 15 μm in diameter. The green, blue and black arrows indicate, in that order, sequential stages in the conversion of thin tubular anthocyanin-filled structures to thick sheet-like structures. The orange arrows indicate the next step, which is the conversion into round structures. The red arrows indicate large fusion bodies resulting from the fusion of the swollen round structures. The yellow arrows point to clear spherical structures devoid of anthocyanins. The bar represents 10 μm.
In contrast to the BMS35S::R+35S::C1 cells, in which anthocyanin production was uncoupled from the light-induced morphological alterations, the accumulation of anthocyanins in these B-I Pl cells is light induced [23]. Thus, the observed alterations in vacuolar morphology could be a consequence of the light-induced expression of the transcription factors, of the light-induced accumulation of anthocyanin, of light-induced alterations in vacuolar morphology observed in BMS35S::R+35S::C1 cells, or a combination of them.
Discussion
Anthocyanin pigments play many important eco-physiological roles in plants. While the biosynthesis and regulation of anthocyanins has been extensively described, little is known on how these pigments are sequestered in the vacuole and to what extent their modes of storage affect color. We describe here cytological changes of vacuoles and sub-vacuolar compartments containing anthocyanins in maize cells exposed to light.
Our studies in BMS35S::R+35S::C1 cells show anthocyanins to accumulate even in total darkness, without any effect of light on the levels of anthocyanins nor on the amount of transcripts of the various flavonoid biosynthetic genes. From these results, we conclude that, when the R and C1 regulators are expressed, there is no need for additional light-induced factors to influence the control of the pathway. This provides strong support to previous findings suggesting that the flavonoid pathway is regulated by light at the level of transcription of the known R2R3 MYB and/or bHLH transcriptional activators and not at the level of the pathway structural genes [23,25,26]. Thus, the BMS35S::R+35S::C1 cells provide a powerful tool to uncouple the effect of light on anthocyanin accumulation and pigmentation, something not feasible in most plant systems, where anthocyanins are induced by light.
The similar levels of anthocyanins in the BMS35S::R+35S::C1 cells under dark or light conditions allowed us to uncover a second effect of light on pigmentation. Light-treated BMS35S::R+35S::C1 cells were darker in color, when compared to identical cells grown in the dark (Fig. 2). Significant quantifiable reflectance differences are observed between BMS35S::R+35S::C1 cells grown for six days in continuous light or dark (Table 1), changes that are not associated with a variation in the amount or type of anthocyanins present (Fig. 3). Microscopy studies established that extensive vacuolar morphological alterations (Fig. 6) correlate with the color darkening. BMS cells not expressing the anthocyanin regulators usually have one or a few vacuolar compartments (Fig. 5). In contrast, constitutive expression of R and C1 in these cells results in a remarkable increase in vacuolar number. It remains to be established whether the accumulation of flavonoids/anthocyanins, the expression of the regulators, or both are necessary and sufficient to trigger the biogenesis of new vacuolar compartments. Within the vacuoles, anthocyanins accumulate in maize cells in red spherical bodies that resemble vacuolar anthocyanoplasts [27]. Although we have not yet unequivocally established whether the anthocyanin inclusions present in BMS35S::R+35S::C1 cells are membrane-bound or not, they have very similar characteristics to the recently-described yellow auto-fluorescent bodies (YFB) present in the vacuoles of BMS cells, induced by the expression of the P1 regulator of 3-deoxy flavonoid biosynthesis [30].
The organization of the anthocyanin inclusions present in BMS35S::R+35S::C1 cells undergo dramatic modifications in the presence of light. These alterations include a reduction in their numbers and an enlargement of their size (Fig. 6). In addition, light-treated vacuoles often showed a spreading of the anthocyanin pigment within the vacuolar lumen, which may be a result of release from the AVI-like structures. The morphological changes observed in the vacuoles of BMS35S::R+35S::C1 cells are not a "curiosity" of cells in culture. We uncovered similar light-induced alterations in the vacuolar structure of maize floral tissues accumulating high levels of anthocyanins (Fig. 8, 9). Time-lapse experiments (see Additional file 1: Movie 1) illustrate the formation and fusions of tubular and globular anthocyanin-filled structures that ultimately coalesce to give one or a few large central vacuoles characteristic of pigmented maize cells. These tubular-vesicular structures, some of which are filled with clear vesicles, are reminiscent of observations of the tubular provacuoles found in vacuolating Euphorbia root cells [8,31]. The anatomical identity of these structures remains to be established, but similar to what was reported for the Euphorbia root cells, the ontogenesis of larger anthocyanin-containing vacuoles from smaller ones is reminiscent of vacuolar fusion and/or autophagy.
It is possible that these light-induced morphological alterations in the anthocyanin-containing structures are directly associated, if not responsible, for the observed color changes. However, alterations in vacuolar pH or the association of anthocyanins with co-pigments can also result in changes in the hue of anthocyanin pigmentation [13,14]. To investigate whether light induces a change in the vacuolar pH of BMS cells, we utilized the BCECF-AM fluorescent dye, for which the ratio of fluorescence emission at the dual excitation wavelengths of 490 nm/440 nm can be calibrated to an in vivo generated pH curve [32]. Using this method, we established that the vacuoles of dark and light grown BMS cells were acidic at pH 5.8 and showed no significant pH differences (see Additional file 2: Fig. 1). Similarly, pH measurement of crude homogenates of BMS or BMS35S::R+35S::C1 cells in water did not yield significant pH value differences between each other or between light- and dark-grown cells (data not shown). Although unlikely, based on the absence of a shift of the λmax of the pigments (Fig. 3 A, B), we also investigated whether light participated in the induction of phytochemicals that could serve as co-pigments, and hence contribute to the color change. Reverse phase HPLC analyses exhibited no significant differences in the peak profiles of phenolic compounds in the dark and light grown BMS and BMS35S::R+35S::C1 cells (not shown).
Although these results do not rule out the possibility of a local and/or transient light-induced pH change or the induction of a minor co-pigment responsible for the color shift, one possible explanation is that the alterations in vacuolar morphology are the cause for the light induced darkening phenomenon of the anthocyanin containing cells. Similar light-induced effects were previously described for the anthocyanoplasts of red cabbage, an observation that was attributed to an increased accumulation of anthocyanins enhanced by light [9]. Our results, however, offer the alternative explanation that light itself can affect the packaging and distribution of anthocyanins within the vacuole, independently of variations in the levels of anthocyanins. Similarly, flowers of the "Rhapsody in Blue" rose cultivar show a change in color induced by age, from red-purple to bluish-purple, and this variation was associated with a progressive accumulation of anthocyanins into AVI-like structures [29]. Lisianthus flowers also show a correlation between the packaging of anthocyanins into AVIs, the presence of "blackish-purple" pigmentation at the base of the petal, and the reduction or absence of AVIs in the outer zones, associated with a lighter purple color of this region [11]. It is thus possible that the observed alteration in the hue of light-grown BMS35S::R+35S::C1 cells reflects a much more general mechanism of light on the packaging of anthocyanins within the vacuole, and hence on pigmentation. If so, the BMS35S::R+35S::C1 cells provide a convenient system for the dissection of the mechanisms of this process because of the molecular and cellular tools available.
Conclusion
The results presented here provide evidence that light affects anthocyanin pigmentation by mechanisms beyond the transcriptional regulation of genes encoding pathway enzymes. In maize floral organs and cultured cells, light induces dramatic morphological alterations in the packing of anthocyanins and distribution of vacuolar and sub-vacuolar compartments. Similar phenomena have been observed before, but the difficulties associated in uncoupling anthocyanin production with morphological alterations in their packaging prevented to draw conclusive cause-consequence relationships.
Methods
Growth, maintenance and treatment of BMS cells
BMS cells were maintained in conditions previously described [33]. In brief, BMS cells were sub-cultured every seven days in liquid MS media supplemented with 2,4,D (0.5 g/L), 3% sucrose (BMS media) on a rotatory shaker (150 rpm) in the dark at 25 ± 2°C. For dark and light treatments, cells from suspension cultures were plated on filter paper overlaid on BMS solid media containing 0.3% phytagel, and were allowed to establish for 20 days in darkness at 25 ± 2°C. Plates were shifted to total darkness (covered with alumnum foil) or light at 50 ± 5 μmol.m-2 .s-1 (Cool white. 215W, F96T12/CW/VHO, Sylvania, Canada).
Transient expression experiments
BMS suspension cells (3 g of cells in 25 ml of BMS media) were treated overnight with 1.7% PEG in BMS media. One ml of cells was overlaid on pre-soaked filter papers in Petri plates. Ten micrograms of 35S::R+35S::C1 plasmid (pPHP687 in [27]) was coated onto gold microprojectiles according to the manufacturers recommendations (Bio-Rad Laboratories, Inc., USA). Coated gold particles were bombarded into PEG-treated BMS cells using a Biolistic PDS-1000/He particle gun (Bio-Rad Laboratories, Inc. USA) at 1,100 psi. The plates were kept in the dark (covered with foil) or exposed to light for a period of 24 hr, after which cells were analyzed microscopically.
Reflectance analysis
In vivo reflectance measures were taken with a Minolta CR-300 reflectometer/colorimeter (Minolta, Japan). The color was represented as CIEL*a*b* values (for the CIE D65/10° illuminant/observer condition). The L* value represents the lightness level, ranging from 100 (white) to 0 (black), the a* (+a red; -a green) and b* (+b yellow; -b blue). The instrument was normalized to standard white tile provided with the instrument before performing analysis on cells grown on solid BMS media in Petri plates.
Extraction and analysis of anthocyanin pigments
BMS35S::R+35S::C1 or control BMS cells after a light or dark treatment were lyophilized for 36 hrs. Anthocyanins and other phenolics were extracted in 50% methanol overnight at a ratio of 50 μg of dry tissue per μl of methanol. Methanol extracts were diluted in 1% HCL in 50% methanol and absorption spectra were collected between 400 to 700 nm with 5 nm intervals at 0.5 s with a Cary 50 UV-VIS spectrophotometer (Varian, Inc. USA). Graphs were generated using the Cary WinUV software. Anthocyanins were measured spectrophotometrically at 530 nm. For the generation of the anthocyanidins from the corresponding anthocyanins, methanol extracts were hydrolyzed by the addition of an equal volume of 2 M HCL (37% v/v) and heated in a boiling water bath for 20 minutes. Hydrolyzed samples were extracted with isoamyl alcohol. Chromatographic separation of the anthocyanidins was performed by thin layer chromatography (TLC) on cellulose plates (5730/7, EM Science, Germany) with HCL/formic acid/H2O, 3:30:10, v/v as the mobile phase. Twenty μL of methanolic extracts of non-hydrolyzed and hydrolyzed samples were injected into a Waters Alliance® 2695 Separations module (Waters Corporation, Milford, MA) in conditions as described [34]. The HPLC profiles were obtained at 280 nm using the Waters 2996 Photodiode Array Detector and analyzed with the Empower software (Waters Corporation, Milford, MA).
Extraction and analysis of RNA
Dark and light grown BMS cells were homogenized in liquid nitrogen and total RNA was extracted using the TRIzol reagent following the manufactures recommendations (Invitrogen, Life Technologies, USA). For Northern analyses, 25 μg of total RNA was separated on a formaldehyde-containing 1% agarose gel and blotted onto a nitrocellulose membrane (Bio-Rad Laboratories, Inc., USA). The blot was hybridized with cDNA probes corresponding to c2 [35], f3h [36] and a1 [37]. Ubiquitin [38] was used as a normalization control. Comparison of the hybridization signals was performed on a BioRad phosphorimager (BioRad Laboratories, Inc., USA) and ratios of the dark and light grown callus hybridization signal to the ubiquitin normalization control were compared.
Plant material
Maize kernels for C2-Idf the genetic stock 418D C2-Idf1 (Active-1); A1 A2 C1 R1 and BI Pl- 219I B1-I; A1 A2 C1 C2 Pll-Rhoades rl-r, 219J B1-I; A1 A2 C1 C2 Pl1-Rhoades r1-g were obtained from the Maize Coop . The kernels were planted in the field in the summer and just before anthesis, male tassels were collected for observation of vacuolar structure of the lemma or the palea.
Microscopy analysis and vacuolar staining
Digital images of the maize floral whorls and callus cells were captured with a Nikon COOLPIX 5700 camera. Macroscopic images of the transient experiments were visualized with an Olympus SZH10 Research Stereo microscope (Olympus, Japan) and images were captured with a Olympus DP10 digital camera. Light and dark grown BMS cells were examined under a Nikon Eclipse E600 microscope. Differential interference contrast (DIC) pictures were taken with a SPOT, RT-Slider digital camera and analyzed using the SPOT imaging software (Diagnostic Instruments, Inc., USA). DIC time lapse images were taken every 4 seconds for 2 hours and were converted into a movie using the SPOT imaging software. For vacuolar staining, transformed and control BMS cells were incubated with 10 μM BCECF-AM (Molecular Probes, USA) in BMS media for 40 minutes at room temperature. Cells were spun down, washed twice and re-suspended in BMS media. Laser scanning confocal microscopy with a PCM 2000/Nikon Eclipse600 system (Nikon Bioscience Confocal Systems, NY) was used to capture digitized images of the BCECF stained cells using the Nikon Plan Fluor 40X/0.75 air objective (1 pixel = 0.3 μm) as described [39]. The 488 nm excitation wavelength of the argon laser was used in conjunction with a 515/30 nm bandpass emission filter (EM515/30HQ). Images were captured using the SIMPLEPCI software (Compix Imaging Systems, PA) and assembled using Adobe PHOTOSHOP (Adobe Systems, Mountain View, CA).
pH measurement
BMS cells (6 ml) were grown in 35 mm Petri plates at a concentration of 0.lg/ml (fresh weight/vol.) for 6 days under light (50 μmol.m-2.s-1) and dark (foil covered) at 100 rpm. Cells were filtered, weighed and resuspended at a concentration of 0.1 g/ml. One ml of cells were loaded with 10 μM BCECF-AM as described above. Hundred μl of loaded and washed cells were pipetted into a 96 well microtitre plate. An in situ calibration curve was generated separately for each of the replicates for the dark and light grown BMS cells. 100 μl of 0.1 M of various pH buffers from a range of 5.0 to 7.0 with 0.005% digitonin [40] was added to 100 μl of cells, and incubated for 10 min. Fluorescence emission was measured at 535 nm with excitation at 440 nm and 490 nm using the FLEX station™ and data analysis program SOFTmax PRO 4.3 (Molecular Devices, CA). The emission ratio at 490/440 was used for calculation of the pH, where irregularities due to unequal loading are eliminated. These measurements were carried out in triplicate.
Authors' contributions
NGI carried out all the molecular and cellular experiments described. EG conceived the project and participated in the design and coordination of the study.
Supplementary Material
Additional File 1
Time-lapse DIC images of a maize lemma from B-I Pl plant over-accumulating anthocyanins. Images were taken every four seconds. The time-lapse series was converted into a movie at one frame per second, therefore a speedup of 4X real time. The last four images represented in Fig. 10(C, D and E) are from this time-lapse series. Note the clear round inclusions in the tubular structures and the large vacuoles at the end. Refer to the results section for a detailed description. Viewing Instructions: The movie can be visualized using either Quicktime player or Windows Media Player. The plug-ins can be downloaded from (Quicktime Player) or Player). The movie size is ~2 MB, therefore it is more efficient to view by downloading the movie to your hard drive.
Click here for file
Additional File 2
Figure 1. In situ measurement of vacuolar pH using BCECF-AM. Equal amounts (0.01 g/100 μl fresh wt) of BCECF AM loaded cells were placed in microtiter plates, and the emission measured at 535 nm, 440 nm and 490 nm excitation wavelengths (A). The 490/440 ratio was calculated (B). An in situ calibration curve was generated separately for each of the dark and light samples with various pH buffers with 0.005% digitonin (C). Vacuoles of both dark and light grown BMS cells were acidic at pH 5.8 and showed no significant pH differences (D).
Click here for file
Acknowledgements
The authors thank Steven Schwarz for assistance with the reflectometer experiments, Biao Ding and JC Jang for assistance with the use of microscopes, Mike Zhu for use of the plate reader, Annkatrin Rose and Asuka Itaya for critical reading of this manuscript, J. Marcela Hernandez for the original observation that light darkened the cells, and the OSU Plant-Microbe Genomics Facility for supporting the establishment of the Metabolomics Laboratory. This research was supported by a grant from the National Science Foundation (MCB-0139962) to EG.
==== Refs
Gould KS Nature's swiss army knife: The diverse protective roles of anthocyanins in leaves J Biomed Biotechnol 2004 2004 314 320 15577195
Chalker-Scott L Environmental significance of anthocyanins in plant stress response Photochem and Photobiol 1999 70 1 9
Mol J Grotewold E Koes R How genes paint flowers and seeds Trends Plant Sci 1998 3 212 217
Winkel-Shirley B Flavonoid biosynthesis. A colorful model for genetics, biochemistry, cell biology and biotechnology Plant Physiol 2001 126 485 493 11402179
Hrazdina G Wagner GJ Metabolic pathways as enzyme complexes: evidence for the synthesis of phenylpropanoids and flavonoids on membrane associate enzyme complexes Arch Biochem Biophys 1985 237 88 100 3970546
Winkel-Shirley B Evidence of enzyme complexes in the phenylpropanoid and flavonoid pathways Physiol Plant 1999 107 142 149
Grotewold E The challenges of moving chemicals within and out of cells: insights into the transport of plant natural products Planta 2004 219 906 909 15278457
Marty F Plant vacuoles Plant Cell 1999 11 587 599 10213780
Pecket CR Small CJ Occurrence, location and development of anthocyanoplasts Phytochemistry 1980 19 2571 2576
Nozzolillo C Ishikura N An investigation of the intracellular site of anthocyanoplasts using isolated protoplasts and vacuoles Plant Cell Rep 1988 7 389 392
Markham KR Gould KS Winefield CS Mitchell KA Bloor SJ Boase MR Anthocyanic vacuolar inclusions – their nature and significance in flower colouration Phytochemistry 2000 55 327 36 11117881
Conn S Zhang W Franco C Anthocyanic vacuolar inclusions (AVIs) selectively bind acylated anthocyanins in Vita vinifera L. (grapevine) suspension culture Biotech Ltrs 2003 25 835 839
Forkmann G Flavonoids as flower pigments: The formation of the natural spectrum and its extension by genetic engineering Plant Breeding 1991 106 1 26
Markham KR Ofman DJ Lisianthus flavonoid pigments and factors influencing their expression in flower colour Phytochemistry 1993 34 679 85 7764146
Noda K-I Glover BJ Linstead P Martin C Flower colour intensity depends on specialized cell shape controlled by a Myb-related transcription factor Nature 1994 369 661 664 8208293
Kondo T Yoshida K Nakagawa A Kawai T Tamura H Goto T Structural basis of blue-colour development in flower petals from Commelina communis Nature 1992 358 515 518
Yoshida K Kondo T Okazaki Y Katou K Cause of blue petal colour Nature 1995 373 291
Fukada-Tanaka S Inagaki Y Yamaguchi T Saito N lida S Colour- enhancing protein in blue petals Nature 2000 407 581 11034195
Irani NG Hernandez JM Grotewold E Regulation of anthocyanin pigmentation Rec Adv Phytochem 2003 38 59 78
Winkel-Shirley B Biosynthesis of flavonoids and effects of stress Curr Opin Plant Biol 2002 5 218 23 11960739
Harmer SL Hogenesch JB Straume M Chang H-S Han B Zhu T Wang X Kreps JA Kay SA Orchestrated transcription of key pathways in Arabidopsis by the circadian clock Science 2000 290 2110 2113 11118138
Taylor LP Briggs WR Genetic regulation and photocontrol of anthocyanin accumulation in maize seedlings Plant Cell 1990 2 115 127 2136630
Procissi A Dolfini S Ronchi A Tonelli C Light-dependent spatial and temporal expression of pigment regulatory genes in developing maize seeds Plant Cell 1997 9 1547 1557 12237395
Petroni K Cominlli E Consonni G Gusnaroli G Gavazzi G Tonelli C The developmental expression of the maize regulatory gene Hopi determines germination-dependent anthocyanin accumulation Genetics 2000 115 323 336 10790406
Piazza P Procissi A Jenkins GI Tonelli C Members of the c1/pl1 regulatory gene family mediate the response of maize aleurone and mesocotyl to different light qualities and cytokinins Plant Physiol 2002 128 1077 86 11891262
Tonelli C Consonni G Dolfini SF Dellaporta SL Viotti A Genetic and molecular analysis of Sn, a light-inducible, tissue specific regulatory gene in maize Mol Gen Genet 1991 225 401 410 1673220
Grotewold E Chamberlin M Snook M Siame B Butler L Swenson J Maddock S Clair GS Bowen B Engineering secondary metabolism in maize cells by ectopic expression of transcription factors Plant Cell 1998 10 721 40 9596632
Dong X Braun EL Grotewold E Functional conservation of plant secondary metabolic enzymes revealed by complementation of Arabidopsis flavonoid mutants with maize genes Plant Physiol 2001 127 46 57 11553733
Gonnet JF Origin of the color of Cv. Rhapsody in Blue rose and some other so-called "blue" roses J Agric Food Chem 2003 51 4990 4994 12903958
Lin Y Irani NG Grotewold E Sub-cellular trafficking of phytochemicals using auto-fluorescent compounds in maize cells BMC Plant Biology 2003 2003 10 14687417
Marty F Cytochemical studies on GERL, provacuoles, and vacuoles in meristematic cells of Euphorbia Proc Natl Acad Sci USA 1978 75 852 856 16592499
Swanson SJ Jones RL Gibberellic acid induces vacuolar acidification in barley aleurone Plant Cell 1996 8 2211 2221 12239377
Dias AP Brown J Bonello P Grotewold E Grotewold E Metabolite profiling as a functional genomics tool Plant Functional Genomics: Methods & Protocols 2003 Totowa, NJ: The Humana Press 415 425
Dias AP Grotewold E Manipulating the accumulation of phenolics in maize cultured cells using transcription factors Biochem Engin J 2003 14 207 216
Wienand U Weydemann U Niesbach-Klîsgen U Peterson P Saedler H Molecular cloning of the c2 locus of Zea mays, the gene encoding for chalcone synthase Mol Gen Genet 1986 203 202 207
Deboo GB Albertsen MC Taylor LP Flavanone 3-hydroxylase transcripts and flavonol accumulation are temporally coordinated in maize anthers Plant J 1995 7 703 713 7773305
Schwarz-Sommer Z Shepherd N Tacke E Gierl A Rohde W Leclercq L Mattes M Berndtgen R Peterson PA Saedler H Influence of transposable elements on the structure and function of the A1 gene of Zea mays EMBO J 1987 6 287 294 15981326
Christensen AH Sharrok RA Quail PH Maize polyubiquitin genes: Structure, thermal perturbation of expression and transcript splicing, and promoter activity following transfer to protoplasts by electroporation Plant Mol Biol 1992 18 675 689 1313711
Rose A Meier I A domain unique to plant RanGAP is responsible for its targeting to the plant nuclear rim Proc Natl Acad Sci USA 2001 98 15377 82 11752475
Moseyko N Feldman LJ Expression of pH-sensitive green fluorescent protein in Arabidopsis thaliana Plant Cell Environ 2001 24 557 63 11706851
| 15907203 | PMC1177971 | CC BY | 2021-01-04 16:37:14 | no | BMC Plant Biol. 2005 May 20; 5:7 | utf-8 | BMC Plant Biol | 2,005 | 10.1186/1471-2229-5-7 | oa_comm |
==== Front
BMC Plant BiolBMC Plant Biology1471-2229BioMed Central London 1471-2229-5-71590720310.1186/1471-2229-5-7Research ArticleLight-induced morphological alteration in anthocyanin-accumulating vacuoles of maize cells Irani Niloufer G [email protected] Erich [email protected] Department of Plant Cellular and Molecular Biology and Plant Biotechnology Center, The Ohio State University, Columbus, OH 43210, USA2005 20 5 2005 5 7 7 3 2 2005 20 5 2005 Copyright © 2005 Irani and Grotewold; licensee BioMed Central Ltd.2005Irani and Grotewold; 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
Plant pigmentation is affected by a variety of factors. Light, an important plant developmental signal, influences the accumulation of anthocyanins primarily through the activation of the transcription factors that regulate the flavonoid biosynthetic pathway. In this study, we utilized maize Black Mexican Sweet (BMS) cells expressing the R and C1 regulators of anthocyanin biosynthesis from a light-insensitive promoter as a means to investigate the existence of additional levels of control of pigmentation by light.
Results
BMS cells expressing the R and C1 regulators from the CaMV 35S constitutive promoter accumulate anthocyanins when grown in complete darkness, suggesting that the transcription factors R and C1 are sufficient for the transcription of the genes corresponding to the structural enzymes of the pathway, with no requirement for additional light-induced regulators. Interestingly, light induces a "darkening" in the color of the purple anthocyanin pigmentation of transgenic BMS cells expressing R and C1. This change in the pigment hue is not associated with a variation in the levels or types of anthocyanins present, or with an alteration of the transcript levels of several flavonoid biosynthetic genes. However, cytological observations show that light drives unexpected changes in the morphology and distribution of the anthocyanins-containing vacuolar compartments.
Conclusion
By uncoupling the effect of light on anthocyanin accumulation, we have found light to induce the fusion of anthocyanin-containing vacuoles, the coalescence of anthocyanic vacuolar inclusion (AVI)-like structures contained, and the spread of anthocyanins from the inclusions into the vacuolar sap. Similar light-induced alterations in vacuolar morphology are also evident in the epidermal cells of maize floral whorls accumulating anthocyanins. Our findings suggest a novel mechanism for the action of light on the vacuolar storage of anthocyanin.
==== Body
Background
Anthocyanins, the coloured end product of the flavonoid pathway, play an important role in attracting insects or animals for pollination and seed dispersal. In addition, they play roles as anti-oxidants and in protecting DNA and the photosynthetic apparatus from high radiation fluxes [1]. Other possible functions of anthocyanins, such as the protection against cold stress or providing drought resistance, are likely to be associated with activities restricted to particular classes of plants [2].
The biosynthesis of the flavonoids, a large phenylpropanoid-derived group of phenolic compounds, provides one of the best described plant metabolic pathways, with many of the structural and regulatory genes in the pathway identified and cloned [3,4]. Less is known regarding the mechanisms by which the water-soluble anthocyanins are transported from their site of synthesis, the cytoplasmic surface of the endoplasmic reticulum [5,6], to the vacuole, where they are usually sequestered [7]. Plant vacuoles are highly dynamic, multifunctional organelles that are the primary storage and turnover sites of macromolecules. These membrane-bound organelles, which can occupy up to 90% of the total cellular volume, are integral part of the endomembrane system, serving as the terminal products of the secretory pathway [8].
Several plant species store anthocyanins within vacuolar inclusions that have been loosely termed anthocyanoplasts which have been observed to start as vesicles in the cytoplasm and appear to be membrane bound [9,10]. More recently, the intravacuolar structures observed in the flower petals of various plants, including carnation and lisianthus, were termed anthocyanic vacuolar inclusions, or AVIs [11]. These inclusions were suggested to be membrane-less, proteinaceous matrixes that acted as anthocyanin traps, preferentially for anthocyanidin 3, 5-diglycosides [11] or acylated anthocyanins [12]. Once in the vacuole, many factors influence the in vivo pigmentation provided by anthocyanins, with important consequences for the eco-physiology of plants [13,14]. Some of these factors include the particular type of anthocyanidin present (for e.g., pelargonidin cyanidin, or myricetin), the shape of the cells in which the pigments accumulate [15], the concentration of the pigment, the formation of complexes between anthocyanins and co-pigments [16], and the vacuolar pH[17,18].
Environmental conditions are known to induce the accumulation of anthocyanin pigments across the major groups of higher plants, of them light being the best studied [2,19,20]. In Arabidopsis, the anthocyanin pathway is regulated in a circadian fashion, with flavonoid genes peaking at the end of the dark cycle, likely preparing plants for daybreak [21]. In maize, members of the PL1/C1 R2R3 MYB and B/R bHLH families of regulatory proteins cooperate for the regulation of anthocyanin pigments [3] and are necessary for the expression of the anthocyanin biosynthetic genes c2 (chalcone synthase), chi1 (chalcone isomerase), f3h (flavanone 3-hydroxylase) a1 (dihydroflavonol 4-reductase), a2 (leucoanthocyanidin dioxygenase/anthocyanin synthase), bz1 (UDP glucose:flavonoid 3-O-glucosyl transferase) and bz2 (glutathione S-transferase) [19]. The light-induced expression of members of these R2R3 MYB and bHLH classes of transcription factors has been proposed to be responsible for the induction of anthocyanins in maize by light [22-26].
Here, we investigated whether light affects the accumulation of maize anthocyanin pigmentation at a level other than through the activation of the known R2R3 MYB and bHLH regulators of the pathway. Towards this goal, we analyzed previously described [27] transgenic Black Mexican Sweet (BMS) maize cells in culture expressing the Zea mays R and C1 regulators from a constitutive, light insensitive CaMV 35S promoter (BMS35S::R+35S::C1). We show here that BMS35S::R+35S::C1 cells are red, even when grown in complete darkness. Upon light treatment, there is a darkening of the color of the BMS35S::R+35S::C1 cells, without an appreciable increase in the quantity of anthocyanins or in the type of anthocyanidins present. Consistent with these findings, the steady-state levels of several anthocyanin biosynthetic genes does not increase upon light treatment. Interestingly, at the subcellular level, light induces an alteration in the way the anthocyanins are distributed within vacuolar compartments. A similar alteration in the morphology of anthocyanin-accumulating vacuoles is observed when maize tassel glumes are irradiated with light, suggesting that the phenomena observed in BMS35S::R+35S::C1 cells in culture also occurs in planta. Together, our findings suggest a novel mechanism for the action of light on the packaging of anthocyanins in the vacuole and in subvacuolar compartments. This effect of light could only be uncovered after making the biosynthesis of anthocyanins independent of light.
Results
BMS cells expressing the transcription factors R and C1 accumulate anthocyanins in the dark
To determine whether the light control of the maize anthocyanin pathway is mediated by the expression of the B/R and C1/PL regulators and/or the biosynthetic genes [23], we investigated the pigmentation of BMS cells expressing the R and C1 genes from the constitutive CaMV 35S promoter (BMS35S::R+35S::C1) [27]. BMS35S::R+35S::C1 cells grown in complete darkness for 30 days were fully pigmented with anthocyanins (Fig. 1A). The bombardment of BMS cells with the R and C1 regulators driven from the 35S promoter (p35SR + p35SC1) resulted in the accumulation of red cells within 15 hours, even when cells were kept in complete darkness after bombardment (compare Fig. 1B and 1C). These results indicate that the constitutive expression of the R and C1 regulators is sufficient for the activation of the pathway, even in the absence of light.
Figure 1 Anthocyanins accumulate in maize BMS35S::R+35S::C1 cells in the dark. (A) Dark-grown BMS35S::R+35S::C1 cells expressing the R and C1 anthocyanin regulators from the CaMV 35S promoter accumulate anthocyanins. Transient expression of the R and C1 regulators in BMS cells by microprojectile bombardment induce anthocyanins in the dark (B) or the light (C). The magnification bars represent 200 μm. The bar in the inset DIC image is 20 μm.
BMS35S::R+35S::C1 cells darken in the light
To investigate whether light has any additional effect on the pigmentation present in BMS35S::R+35S::C1 cells, we compared the color of BMS and BMS35S::R+35S::C1 cells grown for six days in complete darkness or under light conditions (see Methods). BMS35S::R+35S::C1 cells grown in the light showed a visual darkening, when compared to cells grown in the dark (Fig. 2, BMS35S::R+35S::C1). However, light did not induce any visible difference in the white-yellow color of the control BMS cells not expressing the C1 and R regulators (Fig. 2, BMS). We quantified these visual color differences in vivo using a reflectometer and the CIELAB color space value system. L* values, representing the lightness level, were reproducibly not affected by light (Table 1). The a* values for BMS35S::R+35S::C1 cells were positive (+a), consistent with the red color characteristic of these cells. There was however a significant (p < 0.05) reduction in the a-values when the BMS35S::R+35S::C1 cells were exposed to light, which was observed in each of the three times that the experiment was performed (Table 1). Although the b* values, contributing to yellow (+b) or blue (-b), were significantly different (p < 0.05) between the dark and light grown BMS35S::R+35S::C1 cells, they hovered near the zero value, suggesting a low contribution to the overall observed color shift. Thus, the a* and the b* values observed corresponded to a quantifiably red color (dull red), with a decreased degree of redness in the light, which is in agreement with our visual observations (Fig. 2). In the absence of a change in the L* value, the apparent darkening of the cells is likely caused by the spectral shift of the reflected light towards the less red.
Table 1 Reflectance analysis and L* a* b* values of dark- and light-grown BMS35S::R+35S::C1 cells in the CIELAB color scale. The a* value contributes to red (a+) or green (a-), the b* value contributes to yellow (+b) or blue (-b) and L* represent the lightness level.
BMS35S::R+35S::C1 L* a* b*
Dark 21.42 ± 1.29 12.06 ± 1.51 0.45 ± 0.48
Light 21.01 ± 0.93 7.02 ± 0.88 -0.45 ± 0.21
Figure 2 Light induces the darkening of anthocyanin pigmentation. Images of BMS control and BMS35S::R+35S::C1 cells grown for ten days under total darkness (Dark) or light (Light) conditions.
The anthocyanin contents or mRNA steady state levels of biosynthetic genes are not altered by white light in BMS35S::R+35S::C1 cells
To investigate whether the change in color observed between dark- and light grown BMS35S::R+35S::C1 cells is due to an alteration in the quality or quantity of the pigments, anthocyanins were extracted and quantified. Methanolic extracts of the cells, normalized for dry weight, were analyzed spectrophotometrically. The absorption spectra of pigments obtained from acidic methanol extracts of dark- and light-treated cells showed identical profiles and very similar absorbance values at 530 nm (Fig. 3A, B). To determine whether light induced an alteration in the type of anthocyanidin present, total anthocyanidins were extracted and separated by thin layer chromatography (TLC). Similar levels of cyanidin and pelargonidin are present in light- and dark treated BMS35S::R+35S::C1 cells (Fig. 3C), consistent with the two major types of anthocyanidins previously described in BMS35S::R+35S::C1 cells [28].
Figure 3 Similar quantities of cyaniding and pelargonidin accumulate in dark and light-grown BMS35S::R+35S::C1 cells. (A) Spectral profile of methanol-HCL extracts of dark- (black line) or light-grown (green line) BMS35S::R+35S::C1 cells. The red line corresponds to the spectra of control BMS extracts. (B) Quantitation of anthocyanins in dark- and light-grown BMS35S::R+35S::C1 cells. Bars represent the SD of measurements obtained in three independent experiments. (C) Qualitative analysis of the anthocyanins present in dark- and light-grown BMS35S::R+35S::C1 cells by TLC.
In BMS cells, R and C1 are known to activate several flavonoid biosynthetic genes [27]. To investigate whether the activation of these genes was further enhanced by the light treatment, total RNA was extracted from BMS35S::R+35S::C1 cells incubated for six days in the light or the dark and analyzed by RNA gel blots (Northern blots, Fig. 4). No significant alteration in the steady-state mRNA levels for chalcone synthase (c2, Fig. 4), flavanone 3-hydroxylase (f3h, Fig. 4) or dihydroflavonol 4-reductase (a1, Fig. 4) was observed for the light grown BMS35S::R+35S::C1 cells. The control BMS cells showed no mRNA accumulation for these genes (Fig. 4). Together, these results suggest that the darkening of the BMS35S::R+35S::C1 cells upon light treatment is not due to an alteration in the quantity or quality of the anthocyanin pigments present.
Figure 4 Northern analysis of flavonoid biosynthetic genes show no alterations in the steady-state mRNA levels induced by light in BMS35S::R+35S::C1 cells. Total RNA from dark- (D) and light-grown (L) control BMS and BMS35S::R+35S::C1 cells were analyzed by Northern with probes corresponding to the c2 (chalcone synthase), f3h (flavanone 3-hydroxylase) and a1 (dihyroflavonol 4-reductase) genes. Ubiquitin (ubq) was used as a normalizing control. The numbers indicate the relative hybridization signal (in arbitrary units) obtained with each probe, normalized for the corresponding signal obtained with ubq.
White light induces alterations in the sub-cellular distribution and vacuolar organization of anthocyanins in BMS35S::R+35S::C1 cells
Alterations in rose flower pigmentation were associated previously with the formation of AVI-like structures [29]. To investigate whether a similar alteration in the packing of anthocyanins could be with the light-induced hue alteration of BMS35S::R+35S::C1 cells, we investigated the sub-cellular morphology of dark- and light-treated BMS35S::R+35S::C1 cells. To unequivocally visualize the vacuole(s), BMS and BMS35S::R+35::C1 were stained with the cell permeable, acetoxymethyl derivative of the fluorescent vacuolar dye, 2',7'-bis(2-carboxyethyl)-5(6)-carboxyfluorescein (BCECF-AM). In the control BMS cells, there are typically one to a few large vacuolar compartments (Fig. 5A, C). In contrast, BMS35S::R+35::C1 cells are always multi-vacuolated (Fig. 5B, D). Unfortunately, anthocyanins concentrated in the vacuolar inclusions quench the fluorescence of the BCECF-AM dye, as observed in some of the larger vacuoles (Fig. 5B).
Figure 5 BMS35S::R+35S::C1 cells have multiple vacuoles. Laser scanning confocal microscopy (false colored) images of (A) BMS and (B) BMS35S::R+35S::C1 cells loaded with 10 μM of the vacuolar dye BCECF-AM. Laser scanning confocal 'light transmitted' images of (C) BMS and (D) BMS35S::R+35S::C1 cells are shown in black and white. The bar represent 20 μm.
Within the vacuole, anthocyanins accumulate in inclusions that, when observed under polarized light, appear round and regular in shape or aggregated, like an intertwine of fine strings with blebs (Fig. 6). The number of vacuoles and inclusions per vacuole are dramatically affected by light in BMS35S::R+35S::C1 cells. In dark-treated BMS35S::R+35::C1 cells, the major representative cell type (Fig. 6A, representing 34% of all the 239 cells observed), had a range of 20 to 30 observable vacuoles (Table 2). In these cells, the anthocyanins were present mainly in the rounded vacuolar inclusions with a characteristic pale pink coloration of the vacuolar sap, a phenomenon called here "anthocyanin spread". The next two most abundant cell types present in dark-treated BMS35S::R+35::C1 cells, corresponding to 16% and 11% of all the cells, had a range of 10–20 or 20–30 visually observable vacuoles per cell, respectively (Table 2). Cells of these groups are characterized by no observable anthocyanin spread in the vacuolar sap under the light microscope, yet had red or pale red anthocyanin inclusions in the vacuole (Fig. 6C, E).
Table 2 Vacuole distribution in dark- and light-grown BMS35S::R+35S::C1 cells. The classification is based on the number of vacuoles and the presence or absence of anthocyanins in the vacuolar sap. The A-F letters correspond to the panels in Fig. 6, and the presence (+) or absence (-) of anthocyanins in the vacuolar sap (spread) is indicated.
% Representation No. of vacuoles (range) Anthocyanin spread
35S::R+C1 Dark (n = 239)
A 34% 20–30 +
C 16% 10–20 -
E 11% 20–30 -
35S::R+C1 Light (n = 400)
B 35% 1–10 ++
D 26% 10–20 +
F 12% 1–10 +++
Figure 6 Light induces alterations in the distribution of anthocyanins within vacuolar compartments. Representative DIC images of six day old (A, C, E) dark-grown and (B, D, F) light-grown BMS35S::R+35S::C1 cells. BMS cells grown in the (G) dark or (H) light show no significant alterations. Numbers correspond to the percentage values indicated in Table 2. The bar represents 20 μm.
In the light, the majority of the BMS35S::R+35::C1 cells (35%, Table 2) show one to ten pigmented vacuoles with red inclusions that appeared most of time diffuse and like "tangled strings" (Fig, 6B). These structures were determined to be 0.1 to 0.3 μm in diameter, appeared in certain cells to be branched with bleb-like structures at their ends. A significant population of cells (26%, Table 2) had 10–20 vacuoles, lightly colored with discrete deeply pigmented, spherical inclusions (Fig. 6D). The third most abundant class of BMS35S::R+35::C1 cells present in the light (12%, Table 2) showed deeply pigmented vacuoles and enlarged inclusions (Fig. 6F).
Comparison of the size of the vacuolar inclusions in the dark- and light-grown BMS35S::R+35::C1 cells (Fig. 7) shows that the majority of the inclusions in the dark grown samples are in the size range of 0.1 μm – l μm, with a noticeable absence of larger ones. The modal range of the vacuolar inclusions in light grown cells was 2 μm – 3 μm, with some as large as 14 μm. This may reflect the fusion of smaller vacuolar inclusions to give rise to larger ones (see below). The control BMS cells, grown in the dark or light, did not show any evident sub-cellular morphological changes (not shown).
Figure 7 Comparison of the size distribution of vacuolar inclusion containing anthocyanins in BMS35S::R+35S::C1 cells grown in the light or the dark. Round vacuolar inclusions were measured and classified according to the size range measured in microns (μm) that they fell into. Blue and red bars represent the AVI sizes in dark- and light-grown cells, respectively.
To further understand the changes induced by light in anthocyanin-accumulating cells, we performed laser scanning confocal microscopy of BCECF-AM loaded BMS35S::R+35::C1 cells. Dark-grown cells showed the presence of multiple vacuoles (Fig. 8G), which, upon light treatment, appeared to coalesce to form fewer, much larger vacuoles (Fig. 8H). Light-grown cells display a decrease in fluorescence, likely because of quenching by the anthocyanins released from the vacuolar inclusions. Dark- or light-grown BMS cells showed no distinctive differences in visual fluorescence intensity or vacuolar morphology (Fig. 8 C, D), suggesting that the observed morphological alterations are either a consequence of the expression of the R and C1 regulators, of the accumulation of anthocyanins or of distinctive properties of the vacuoles in which anthocyanins accumulate.
Figure 8 Morphology of vacuoles of dark and light grown BMS and BMS35S::R+3ss::C1 cells loaded with BCECF-AM.
Together, these results show that light-exposed BMS35S::R+35::C1 cells have a significant reduction in the number of vacuoles with a associated increase in their size, a change in the number, shape and size of the AVIs and a release of anthocyanins from the AVIs into the vacuole.
Light-induced vacuolar morphological alterations in anthocyanin-accumulating maize floral organs
To establish whether the light-induced vacuolar alterations observed in BMS cells could be observed also in planta, we looked at the tassels of maize B-I Pl plants that accumulated large quantities of anthocyanins (Fig. 9). The inner, light-protected lemma and palea (Fig. 9B) were the choice of material to observe the light-induced alterations in vacuolar morphology. The epidermal cells of these appendages in a C2-Idf mutant lacking anthocyanins (Fig. 9C) have one to a few large, observable, colorless, central vacuoles (Fig. 9D). In contrast, depending on the physiological and developmental stage of the florets, the epidermal cells of the lemma or palea of the B-I Pl florets (Fig. 9B, E) were either already filled with anthocyanins, or were in the initial stages of accumulation (Fig 9F). These cells show a distinctive multi-vacuolar morphology, and the vacuoles were often heavily loaded with anthocyanins and AVI-like structures (Fig. 9F).
Figure 9 Morphology of anthocyanin-accumulating cells in maize floral organs. The maize male inflorescence, the tassel, is a panicle made of numerous spikes, each formed by numerous, paired spikelets. (A) Single spike with paired spikelets. The spikelet has outer and inner glumes (bracts of the florets) and each floret has a lemma, a palea, a highly reduced lodigule and 3 stamens (B) Spikelet dissection: the two florets with an outer glume, an inner glume; each floret with a lemma, palea, highly reduced lodigule and three stamens. (C) Digital macro images of open florets from a C2-Idf (chalcone synthase) mutant that accumulates no anthocyanins. (D) DIC light micrographs of the lemma from male flowers of C2-Idf plant (E). Digital macro images of open florets from a B-I Pl plant. (F) DIC light micrographs of the lemma from male flowers of a B-I Pl plant. The bar represents 10 μm
Light from the microscope was sufficient to induce dramatic alterations in the morphology and distribution of the anthocyanin-containing vacuolar structures, as seen in the time-lapse images taken at four-second intervals (Fig. 10, see Additional file 1: Movie 1 for the original data used to perform this analysis). In these series, the initially thin, tubular anthocyanin-filled structures (average diameter of 0.6 μm, Fig. 10, green arrows) expand to a thickness of about 1.4 μm, dynamically filling the entire cell (Fig. 10A, blue arrows). These thick, finger-like projections became swollen, sheet-like structures (~3.3 μm in diameter, Fig 10, black arrows), to then become rounded (Fig 10A, B, C; orange arrows) and fuse with each other. These rounded/oval compartments measuring one 1–9 μm in diameter displayed, just like the tubular structures, dynamic morphological changes. Swollen "blebs" were observed moving along fine tubules and the ends of the thick tubules swelled up into round structures. Fusion events, once initiated, were very rapid, which resulted in the formation of large fusion bodies (Fig. 10C, D, E; red arrows) containing numerous clear (i.e., no anthocyanin-containing) structures (Fig 10C, D, E; yellow arrows). These clear inclusions were also observed initially in the sheet-like structures and were formed as the size of the tubules grew. The fusion bodies progressed rapidly to fill the entire cell, finally coalescing together resulting in the anthocyanin spread throughout the compartment (Fig 10E). The defined margins around the large central AVI-like structure (~15 μm across) become more diffuse with a lighter translucent red halo around the opaque, dark body (Fig. 10E).
Figure 10 Sub-cellular morphology of BI Pl maize floral cells accumulating anthocyanins. DIC images of a maize lemma from B-I/B-Peru plant over-accumulating anthocyanins. The above are extracted images from a time-lapse series (See Additional file 1: Movie 1). The time points on the images indicate the period from time 0' i.e. when the sample was mounted onto the stage and exposed to the microscope light. (A) and (B) occur earlier in the series while (D), (C) and (E) are in rapid succession (24 seconds apart). The large central inclusion corresponds to a vacuolar inclusion containing anthocyanins measuring 15 μm in diameter. The green, blue and black arrows indicate, in that order, sequential stages in the conversion of thin tubular anthocyanin-filled structures to thick sheet-like structures. The orange arrows indicate the next step, which is the conversion into round structures. The red arrows indicate large fusion bodies resulting from the fusion of the swollen round structures. The yellow arrows point to clear spherical structures devoid of anthocyanins. The bar represents 10 μm.
In contrast to the BMS35S::R+35S::C1 cells, in which anthocyanin production was uncoupled from the light-induced morphological alterations, the accumulation of anthocyanins in these B-I Pl cells is light induced [23]. Thus, the observed alterations in vacuolar morphology could be a consequence of the light-induced expression of the transcription factors, of the light-induced accumulation of anthocyanin, of light-induced alterations in vacuolar morphology observed in BMS35S::R+35S::C1 cells, or a combination of them.
Discussion
Anthocyanin pigments play many important eco-physiological roles in plants. While the biosynthesis and regulation of anthocyanins has been extensively described, little is known on how these pigments are sequestered in the vacuole and to what extent their modes of storage affect color. We describe here cytological changes of vacuoles and sub-vacuolar compartments containing anthocyanins in maize cells exposed to light.
Our studies in BMS35S::R+35S::C1 cells show anthocyanins to accumulate even in total darkness, without any effect of light on the levels of anthocyanins nor on the amount of transcripts of the various flavonoid biosynthetic genes. From these results, we conclude that, when the R and C1 regulators are expressed, there is no need for additional light-induced factors to influence the control of the pathway. This provides strong support to previous findings suggesting that the flavonoid pathway is regulated by light at the level of transcription of the known R2R3 MYB and/or bHLH transcriptional activators and not at the level of the pathway structural genes [23,25,26]. Thus, the BMS35S::R+35S::C1 cells provide a powerful tool to uncouple the effect of light on anthocyanin accumulation and pigmentation, something not feasible in most plant systems, where anthocyanins are induced by light.
The similar levels of anthocyanins in the BMS35S::R+35S::C1 cells under dark or light conditions allowed us to uncover a second effect of light on pigmentation. Light-treated BMS35S::R+35S::C1 cells were darker in color, when compared to identical cells grown in the dark (Fig. 2). Significant quantifiable reflectance differences are observed between BMS35S::R+35S::C1 cells grown for six days in continuous light or dark (Table 1), changes that are not associated with a variation in the amount or type of anthocyanins present (Fig. 3). Microscopy studies established that extensive vacuolar morphological alterations (Fig. 6) correlate with the color darkening. BMS cells not expressing the anthocyanin regulators usually have one or a few vacuolar compartments (Fig. 5). In contrast, constitutive expression of R and C1 in these cells results in a remarkable increase in vacuolar number. It remains to be established whether the accumulation of flavonoids/anthocyanins, the expression of the regulators, or both are necessary and sufficient to trigger the biogenesis of new vacuolar compartments. Within the vacuoles, anthocyanins accumulate in maize cells in red spherical bodies that resemble vacuolar anthocyanoplasts [27]. Although we have not yet unequivocally established whether the anthocyanin inclusions present in BMS35S::R+35S::C1 cells are membrane-bound or not, they have very similar characteristics to the recently-described yellow auto-fluorescent bodies (YFB) present in the vacuoles of BMS cells, induced by the expression of the P1 regulator of 3-deoxy flavonoid biosynthesis [30].
The organization of the anthocyanin inclusions present in BMS35S::R+35S::C1 cells undergo dramatic modifications in the presence of light. These alterations include a reduction in their numbers and an enlargement of their size (Fig. 6). In addition, light-treated vacuoles often showed a spreading of the anthocyanin pigment within the vacuolar lumen, which may be a result of release from the AVI-like structures. The morphological changes observed in the vacuoles of BMS35S::R+35S::C1 cells are not a "curiosity" of cells in culture. We uncovered similar light-induced alterations in the vacuolar structure of maize floral tissues accumulating high levels of anthocyanins (Fig. 8, 9). Time-lapse experiments (see Additional file 1: Movie 1) illustrate the formation and fusions of tubular and globular anthocyanin-filled structures that ultimately coalesce to give one or a few large central vacuoles characteristic of pigmented maize cells. These tubular-vesicular structures, some of which are filled with clear vesicles, are reminiscent of observations of the tubular provacuoles found in vacuolating Euphorbia root cells [8,31]. The anatomical identity of these structures remains to be established, but similar to what was reported for the Euphorbia root cells, the ontogenesis of larger anthocyanin-containing vacuoles from smaller ones is reminiscent of vacuolar fusion and/or autophagy.
It is possible that these light-induced morphological alterations in the anthocyanin-containing structures are directly associated, if not responsible, for the observed color changes. However, alterations in vacuolar pH or the association of anthocyanins with co-pigments can also result in changes in the hue of anthocyanin pigmentation [13,14]. To investigate whether light induces a change in the vacuolar pH of BMS cells, we utilized the BCECF-AM fluorescent dye, for which the ratio of fluorescence emission at the dual excitation wavelengths of 490 nm/440 nm can be calibrated to an in vivo generated pH curve [32]. Using this method, we established that the vacuoles of dark and light grown BMS cells were acidic at pH 5.8 and showed no significant pH differences (see Additional file 2: Fig. 1). Similarly, pH measurement of crude homogenates of BMS or BMS35S::R+35S::C1 cells in water did not yield significant pH value differences between each other or between light- and dark-grown cells (data not shown). Although unlikely, based on the absence of a shift of the λmax of the pigments (Fig. 3 A, B), we also investigated whether light participated in the induction of phytochemicals that could serve as co-pigments, and hence contribute to the color change. Reverse phase HPLC analyses exhibited no significant differences in the peak profiles of phenolic compounds in the dark and light grown BMS and BMS35S::R+35S::C1 cells (not shown).
Although these results do not rule out the possibility of a local and/or transient light-induced pH change or the induction of a minor co-pigment responsible for the color shift, one possible explanation is that the alterations in vacuolar morphology are the cause for the light induced darkening phenomenon of the anthocyanin containing cells. Similar light-induced effects were previously described for the anthocyanoplasts of red cabbage, an observation that was attributed to an increased accumulation of anthocyanins enhanced by light [9]. Our results, however, offer the alternative explanation that light itself can affect the packaging and distribution of anthocyanins within the vacuole, independently of variations in the levels of anthocyanins. Similarly, flowers of the "Rhapsody in Blue" rose cultivar show a change in color induced by age, from red-purple to bluish-purple, and this variation was associated with a progressive accumulation of anthocyanins into AVI-like structures [29]. Lisianthus flowers also show a correlation between the packaging of anthocyanins into AVIs, the presence of "blackish-purple" pigmentation at the base of the petal, and the reduction or absence of AVIs in the outer zones, associated with a lighter purple color of this region [11]. It is thus possible that the observed alteration in the hue of light-grown BMS35S::R+35S::C1 cells reflects a much more general mechanism of light on the packaging of anthocyanins within the vacuole, and hence on pigmentation. If so, the BMS35S::R+35S::C1 cells provide a convenient system for the dissection of the mechanisms of this process because of the molecular and cellular tools available.
Conclusion
The results presented here provide evidence that light affects anthocyanin pigmentation by mechanisms beyond the transcriptional regulation of genes encoding pathway enzymes. In maize floral organs and cultured cells, light induces dramatic morphological alterations in the packing of anthocyanins and distribution of vacuolar and sub-vacuolar compartments. Similar phenomena have been observed before, but the difficulties associated in uncoupling anthocyanin production with morphological alterations in their packaging prevented to draw conclusive cause-consequence relationships.
Methods
Growth, maintenance and treatment of BMS cells
BMS cells were maintained in conditions previously described [33]. In brief, BMS cells were sub-cultured every seven days in liquid MS media supplemented with 2,4,D (0.5 g/L), 3% sucrose (BMS media) on a rotatory shaker (150 rpm) in the dark at 25 ± 2°C. For dark and light treatments, cells from suspension cultures were plated on filter paper overlaid on BMS solid media containing 0.3% phytagel, and were allowed to establish for 20 days in darkness at 25 ± 2°C. Plates were shifted to total darkness (covered with alumnum foil) or light at 50 ± 5 μmol.m-2 .s-1 (Cool white. 215W, F96T12/CW/VHO, Sylvania, Canada).
Transient expression experiments
BMS suspension cells (3 g of cells in 25 ml of BMS media) were treated overnight with 1.7% PEG in BMS media. One ml of cells was overlaid on pre-soaked filter papers in Petri plates. Ten micrograms of 35S::R+35S::C1 plasmid (pPHP687 in [27]) was coated onto gold microprojectiles according to the manufacturers recommendations (Bio-Rad Laboratories, Inc., USA). Coated gold particles were bombarded into PEG-treated BMS cells using a Biolistic PDS-1000/He particle gun (Bio-Rad Laboratories, Inc. USA) at 1,100 psi. The plates were kept in the dark (covered with foil) or exposed to light for a period of 24 hr, after which cells were analyzed microscopically.
Reflectance analysis
In vivo reflectance measures were taken with a Minolta CR-300 reflectometer/colorimeter (Minolta, Japan). The color was represented as CIEL*a*b* values (for the CIE D65/10° illuminant/observer condition). The L* value represents the lightness level, ranging from 100 (white) to 0 (black), the a* (+a red; -a green) and b* (+b yellow; -b blue). The instrument was normalized to standard white tile provided with the instrument before performing analysis on cells grown on solid BMS media in Petri plates.
Extraction and analysis of anthocyanin pigments
BMS35S::R+35S::C1 or control BMS cells after a light or dark treatment were lyophilized for 36 hrs. Anthocyanins and other phenolics were extracted in 50% methanol overnight at a ratio of 50 μg of dry tissue per μl of methanol. Methanol extracts were diluted in 1% HCL in 50% methanol and absorption spectra were collected between 400 to 700 nm with 5 nm intervals at 0.5 s with a Cary 50 UV-VIS spectrophotometer (Varian, Inc. USA). Graphs were generated using the Cary WinUV software. Anthocyanins were measured spectrophotometrically at 530 nm. For the generation of the anthocyanidins from the corresponding anthocyanins, methanol extracts were hydrolyzed by the addition of an equal volume of 2 M HCL (37% v/v) and heated in a boiling water bath for 20 minutes. Hydrolyzed samples were extracted with isoamyl alcohol. Chromatographic separation of the anthocyanidins was performed by thin layer chromatography (TLC) on cellulose plates (5730/7, EM Science, Germany) with HCL/formic acid/H2O, 3:30:10, v/v as the mobile phase. Twenty μL of methanolic extracts of non-hydrolyzed and hydrolyzed samples were injected into a Waters Alliance® 2695 Separations module (Waters Corporation, Milford, MA) in conditions as described [34]. The HPLC profiles were obtained at 280 nm using the Waters 2996 Photodiode Array Detector and analyzed with the Empower software (Waters Corporation, Milford, MA).
Extraction and analysis of RNA
Dark and light grown BMS cells were homogenized in liquid nitrogen and total RNA was extracted using the TRIzol reagent following the manufactures recommendations (Invitrogen, Life Technologies, USA). For Northern analyses, 25 μg of total RNA was separated on a formaldehyde-containing 1% agarose gel and blotted onto a nitrocellulose membrane (Bio-Rad Laboratories, Inc., USA). The blot was hybridized with cDNA probes corresponding to c2 [35], f3h [36] and a1 [37]. Ubiquitin [38] was used as a normalization control. Comparison of the hybridization signals was performed on a BioRad phosphorimager (BioRad Laboratories, Inc., USA) and ratios of the dark and light grown callus hybridization signal to the ubiquitin normalization control were compared.
Plant material
Maize kernels for C2-Idf the genetic stock 418D C2-Idf1 (Active-1); A1 A2 C1 R1 and BI Pl- 219I B1-I; A1 A2 C1 C2 Pll-Rhoades rl-r, 219J B1-I; A1 A2 C1 C2 Pl1-Rhoades r1-g were obtained from the Maize Coop . The kernels were planted in the field in the summer and just before anthesis, male tassels were collected for observation of vacuolar structure of the lemma or the palea.
Microscopy analysis and vacuolar staining
Digital images of the maize floral whorls and callus cells were captured with a Nikon COOLPIX 5700 camera. Macroscopic images of the transient experiments were visualized with an Olympus SZH10 Research Stereo microscope (Olympus, Japan) and images were captured with a Olympus DP10 digital camera. Light and dark grown BMS cells were examined under a Nikon Eclipse E600 microscope. Differential interference contrast (DIC) pictures were taken with a SPOT, RT-Slider digital camera and analyzed using the SPOT imaging software (Diagnostic Instruments, Inc., USA). DIC time lapse images were taken every 4 seconds for 2 hours and were converted into a movie using the SPOT imaging software. For vacuolar staining, transformed and control BMS cells were incubated with 10 μM BCECF-AM (Molecular Probes, USA) in BMS media for 40 minutes at room temperature. Cells were spun down, washed twice and re-suspended in BMS media. Laser scanning confocal microscopy with a PCM 2000/Nikon Eclipse600 system (Nikon Bioscience Confocal Systems, NY) was used to capture digitized images of the BCECF stained cells using the Nikon Plan Fluor 40X/0.75 air objective (1 pixel = 0.3 μm) as described [39]. The 488 nm excitation wavelength of the argon laser was used in conjunction with a 515/30 nm bandpass emission filter (EM515/30HQ). Images were captured using the SIMPLEPCI software (Compix Imaging Systems, PA) and assembled using Adobe PHOTOSHOP (Adobe Systems, Mountain View, CA).
pH measurement
BMS cells (6 ml) were grown in 35 mm Petri plates at a concentration of 0.lg/ml (fresh weight/vol.) for 6 days under light (50 μmol.m-2.s-1) and dark (foil covered) at 100 rpm. Cells were filtered, weighed and resuspended at a concentration of 0.1 g/ml. One ml of cells were loaded with 10 μM BCECF-AM as described above. Hundred μl of loaded and washed cells were pipetted into a 96 well microtitre plate. An in situ calibration curve was generated separately for each of the replicates for the dark and light grown BMS cells. 100 μl of 0.1 M of various pH buffers from a range of 5.0 to 7.0 with 0.005% digitonin [40] was added to 100 μl of cells, and incubated for 10 min. Fluorescence emission was measured at 535 nm with excitation at 440 nm and 490 nm using the FLEX station™ and data analysis program SOFTmax PRO 4.3 (Molecular Devices, CA). The emission ratio at 490/440 was used for calculation of the pH, where irregularities due to unequal loading are eliminated. These measurements were carried out in triplicate.
Authors' contributions
NGI carried out all the molecular and cellular experiments described. EG conceived the project and participated in the design and coordination of the study.
Supplementary Material
Additional File 1
Time-lapse DIC images of a maize lemma from B-I Pl plant over-accumulating anthocyanins. Images were taken every four seconds. The time-lapse series was converted into a movie at one frame per second, therefore a speedup of 4X real time. The last four images represented in Fig. 10(C, D and E) are from this time-lapse series. Note the clear round inclusions in the tubular structures and the large vacuoles at the end. Refer to the results section for a detailed description. Viewing Instructions: The movie can be visualized using either Quicktime player or Windows Media Player. The plug-ins can be downloaded from (Quicktime Player) or Player). The movie size is ~2 MB, therefore it is more efficient to view by downloading the movie to your hard drive.
Click here for file
Additional File 2
Figure 1. In situ measurement of vacuolar pH using BCECF-AM. Equal amounts (0.01 g/100 μl fresh wt) of BCECF AM loaded cells were placed in microtiter plates, and the emission measured at 535 nm, 440 nm and 490 nm excitation wavelengths (A). The 490/440 ratio was calculated (B). An in situ calibration curve was generated separately for each of the dark and light samples with various pH buffers with 0.005% digitonin (C). Vacuoles of both dark and light grown BMS cells were acidic at pH 5.8 and showed no significant pH differences (D).
Click here for file
Acknowledgements
The authors thank Steven Schwarz for assistance with the reflectometer experiments, Biao Ding and JC Jang for assistance with the use of microscopes, Mike Zhu for use of the plate reader, Annkatrin Rose and Asuka Itaya for critical reading of this manuscript, J. Marcela Hernandez for the original observation that light darkened the cells, and the OSU Plant-Microbe Genomics Facility for supporting the establishment of the Metabolomics Laboratory. This research was supported by a grant from the National Science Foundation (MCB-0139962) to EG.
==== Refs
Gould KS Nature's swiss army knife: The diverse protective roles of anthocyanins in leaves J Biomed Biotechnol 2004 2004 314 320 15577195
Chalker-Scott L Environmental significance of anthocyanins in plant stress response Photochem and Photobiol 1999 70 1 9
Mol J Grotewold E Koes R How genes paint flowers and seeds Trends Plant Sci 1998 3 212 217
Winkel-Shirley B Flavonoid biosynthesis. A colorful model for genetics, biochemistry, cell biology and biotechnology Plant Physiol 2001 126 485 493 11402179
Hrazdina G Wagner GJ Metabolic pathways as enzyme complexes: evidence for the synthesis of phenylpropanoids and flavonoids on membrane associate enzyme complexes Arch Biochem Biophys 1985 237 88 100 3970546
Winkel-Shirley B Evidence of enzyme complexes in the phenylpropanoid and flavonoid pathways Physiol Plant 1999 107 142 149
Grotewold E The challenges of moving chemicals within and out of cells: insights into the transport of plant natural products Planta 2004 219 906 909 15278457
Marty F Plant vacuoles Plant Cell 1999 11 587 599 10213780
Pecket CR Small CJ Occurrence, location and development of anthocyanoplasts Phytochemistry 1980 19 2571 2576
Nozzolillo C Ishikura N An investigation of the intracellular site of anthocyanoplasts using isolated protoplasts and vacuoles Plant Cell Rep 1988 7 389 392
Markham KR Gould KS Winefield CS Mitchell KA Bloor SJ Boase MR Anthocyanic vacuolar inclusions – their nature and significance in flower colouration Phytochemistry 2000 55 327 36 11117881
Conn S Zhang W Franco C Anthocyanic vacuolar inclusions (AVIs) selectively bind acylated anthocyanins in Vita vinifera L. (grapevine) suspension culture Biotech Ltrs 2003 25 835 839
Forkmann G Flavonoids as flower pigments: The formation of the natural spectrum and its extension by genetic engineering Plant Breeding 1991 106 1 26
Markham KR Ofman DJ Lisianthus flavonoid pigments and factors influencing their expression in flower colour Phytochemistry 1993 34 679 85 7764146
Noda K-I Glover BJ Linstead P Martin C Flower colour intensity depends on specialized cell shape controlled by a Myb-related transcription factor Nature 1994 369 661 664 8208293
Kondo T Yoshida K Nakagawa A Kawai T Tamura H Goto T Structural basis of blue-colour development in flower petals from Commelina communis Nature 1992 358 515 518
Yoshida K Kondo T Okazaki Y Katou K Cause of blue petal colour Nature 1995 373 291
Fukada-Tanaka S Inagaki Y Yamaguchi T Saito N lida S Colour- enhancing protein in blue petals Nature 2000 407 581 11034195
Irani NG Hernandez JM Grotewold E Regulation of anthocyanin pigmentation Rec Adv Phytochem 2003 38 59 78
Winkel-Shirley B Biosynthesis of flavonoids and effects of stress Curr Opin Plant Biol 2002 5 218 23 11960739
Harmer SL Hogenesch JB Straume M Chang H-S Han B Zhu T Wang X Kreps JA Kay SA Orchestrated transcription of key pathways in Arabidopsis by the circadian clock Science 2000 290 2110 2113 11118138
Taylor LP Briggs WR Genetic regulation and photocontrol of anthocyanin accumulation in maize seedlings Plant Cell 1990 2 115 127 2136630
Procissi A Dolfini S Ronchi A Tonelli C Light-dependent spatial and temporal expression of pigment regulatory genes in developing maize seeds Plant Cell 1997 9 1547 1557 12237395
Petroni K Cominlli E Consonni G Gusnaroli G Gavazzi G Tonelli C The developmental expression of the maize regulatory gene Hopi determines germination-dependent anthocyanin accumulation Genetics 2000 115 323 336 10790406
Piazza P Procissi A Jenkins GI Tonelli C Members of the c1/pl1 regulatory gene family mediate the response of maize aleurone and mesocotyl to different light qualities and cytokinins Plant Physiol 2002 128 1077 86 11891262
Tonelli C Consonni G Dolfini SF Dellaporta SL Viotti A Genetic and molecular analysis of Sn, a light-inducible, tissue specific regulatory gene in maize Mol Gen Genet 1991 225 401 410 1673220
Grotewold E Chamberlin M Snook M Siame B Butler L Swenson J Maddock S Clair GS Bowen B Engineering secondary metabolism in maize cells by ectopic expression of transcription factors Plant Cell 1998 10 721 40 9596632
Dong X Braun EL Grotewold E Functional conservation of plant secondary metabolic enzymes revealed by complementation of Arabidopsis flavonoid mutants with maize genes Plant Physiol 2001 127 46 57 11553733
Gonnet JF Origin of the color of Cv. Rhapsody in Blue rose and some other so-called "blue" roses J Agric Food Chem 2003 51 4990 4994 12903958
Lin Y Irani NG Grotewold E Sub-cellular trafficking of phytochemicals using auto-fluorescent compounds in maize cells BMC Plant Biology 2003 2003 10 14687417
Marty F Cytochemical studies on GERL, provacuoles, and vacuoles in meristematic cells of Euphorbia Proc Natl Acad Sci USA 1978 75 852 856 16592499
Swanson SJ Jones RL Gibberellic acid induces vacuolar acidification in barley aleurone Plant Cell 1996 8 2211 2221 12239377
Dias AP Brown J Bonello P Grotewold E Grotewold E Metabolite profiling as a functional genomics tool Plant Functional Genomics: Methods & Protocols 2003 Totowa, NJ: The Humana Press 415 425
Dias AP Grotewold E Manipulating the accumulation of phenolics in maize cultured cells using transcription factors Biochem Engin J 2003 14 207 216
Wienand U Weydemann U Niesbach-Klîsgen U Peterson P Saedler H Molecular cloning of the c2 locus of Zea mays, the gene encoding for chalcone synthase Mol Gen Genet 1986 203 202 207
Deboo GB Albertsen MC Taylor LP Flavanone 3-hydroxylase transcripts and flavonol accumulation are temporally coordinated in maize anthers Plant J 1995 7 703 713 7773305
Schwarz-Sommer Z Shepherd N Tacke E Gierl A Rohde W Leclercq L Mattes M Berndtgen R Peterson PA Saedler H Influence of transposable elements on the structure and function of the A1 gene of Zea mays EMBO J 1987 6 287 294 15981326
Christensen AH Sharrok RA Quail PH Maize polyubiquitin genes: Structure, thermal perturbation of expression and transcript splicing, and promoter activity following transfer to protoplasts by electroporation Plant Mol Biol 1992 18 675 689 1313711
Rose A Meier I A domain unique to plant RanGAP is responsible for its targeting to the plant nuclear rim Proc Natl Acad Sci USA 2001 98 15377 82 11752475
Moseyko N Feldman LJ Expression of pH-sensitive green fluorescent protein in Arabidopsis thaliana Plant Cell Environ 2001 24 557 63 11706851
| 15949041 | PMC1177972 | CC BY | 2021-01-04 16:30:02 | no | BMC Urol. 2005 Jun 10; 5:11 | latin-1 | BMC Urol | 2,005 | 10.1186/1471-2490-5-11 | oa_comm |
==== Front
Clin Mol AllergyClinical and molecular allergy : CMA1476-7961BioMed Central London 1476-7961-3-71593510210.1186/1476-7961-3-7ReviewAnother explanation for the low allergy rate in the rural Alpine foothills Wjst Matthias [email protected] Institut für Epidemiologie GSF – Forschungszentrum für Umwelt und Gesundheit Ingolstädter Landstrasse 1 D-85758 Neuherberg / Munich Germany2005 5 6 2005 3 7 7 25 1 2005 5 6 2005 Copyright © 2005 Wjst; licensee BioMed Central Ltd.2005Wjst; 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.
A low allergy rate in coal and wood heated homes has been described in the small villages in the Alpine foothills and subsequently found to be associated with the farming environment. This was interpreted within the framework of the hygiene hypothesis but there are also alternative explanations. Lower air pollution could be one reason, which is, however, unlikely since the differences between the Bavarian countryside and the Munich municipal area were only weak. There could be genetic differences between the urban and rural population by previous isolation or by self-selection. The potential drop-out of allergy genes, however, will also not explain the absent increase of allergies in two generations. More likely, other lifestyle factors are important. Dietary habits are different in farmers and a less frequent vitamin D supplementation of newborns (otherwise expected to be allergy promoting) has been shown recently. The underlying cause for the "non-allergic farm child" remains speculative until the transfer of any farm-associated factor is leading to a similar risk reduction in the general population.
==== Body
Introduction
Allergy prevalence has been on the rise worldwide and nearly hundred years after coining "Allergie" in the "Münchner Medizinische Wochenschrift" [1] the causal risk factors are still unknown.
At the end of the 1980s, air pollution related effects were thought to be responsible for the allergy epidemic. It turned out, however, that at least traffic related combustion was not the main culprit in the Munich municipal area, neither based on the inner city distribution of pollutants [2] nor by comparison with a control region in Upper Bavaria. In this study, located in the South of Munich on the Alpine foothills, I examined nearly two thousand fourth-grade children between October 1989 and July 1990 in more than 50 villages. I already noted at that time a relationship between the farm odour in some of the small classrooms and the nearly absence of any positive skin prick test (the "Ostallgäu" phenomenon). A protective effect of coal heating was eventually published six years later [3] but received little attention as the public interest focused mainly on East- and West German air pollution differences [4]. It was only in 1999 where a long series of studies in the farming environment started [5-12] which lead to the current version of the hygiene hypothesis that allergy develops where the natural high endotoxin level on farms is absent. Endotoxin has already been described in a study in 2000 as the main component protecting against allergic sensitisation [13].
Problems with the hygiene hypothesis
The hygiene hypothesis is based on the the initial observation of "unhygienic" siblings by Golding and Peters 1986 [14]. After more than one decade of research [15], however, Strachan concluded that "an inverse association between infection and allergy has not been confirmed directly by epidemiological studies. The available data are either inconsistent or inconclusive" [16]. This view is supported by several other authors [17-19] as the adaptive immune system "with an array of potential interactions .... is reduced to a single level" [20]. Although even a patent has been filed on components of stable air to treat allergy [21], a task force of the European Academy of Allergology and Clinical Immunology (EAACI) arrived at the opinion that "there is no recently published evidence in favour of a clinical use of so-called bacterial extracts against asthma and allergic diseases." [22]
During the discussion of factors related to hygiene it seemed to be largely neglected, that (viral) infection may even enhance allergic disease [23]. Also the inverse association of hepatitis antibodies and allergy found in Italian military students [24] has not been reproduced in consecutive studies [25-27]. The protection against allergic disease by mycobacteria [29] could also not be reproduced in the following dozen studies [28]. The support for the hygiene hypothesis therefore remains weak.
Unfortunately, all farming studies are based on observational and retrospective data given rise to concerns not mentioned in previous reviews [30,31]. The transition of a farming society into the industrial age neither coincides with the main peak of the allergy prevalence in Western countries nor does it match the geographical distribution of the disease.
Is endotoxin to blame?
Although there are well-designed studies describing the immunological action of endotoxin [32-35] there are no quantitative data in humans how the nanogram exposure on the pulmonary epithelium will supersede the gram-wise exposure on the gut mucosa. The number of bacteria on the human body's surface is more than 10 times greater than all his somatic cells [36]. Even if N-acetyl-muramic acid is found to be significantly higher in dust from farm children's mattresses (+20% [37]) or endotoxin units are being increased (+66% [10]), is is unclear whether this has any biological meaning [38]. There are many reasons why dust deposition on the floor may not be equal to effective exposure as this involves inhalation, deposition, uptake, processing, preservation and target delivery. In the only study available so far, both asthmatic and non asthmatic probands had the same LPS concentrations in their bronchoalveolar lavage [39].
Even if we assume a relevant target exposure, there are effective mechanism to counteract endotoxin [40,41]. Dose and timing [42], even the origin from different bacteria [43,44] as well as host characteristics [38,45] are being important. Lipopolysaccharides from some bacteria may induce even a Th2 type response [44] where the induction of sensitization is an allergen-specific phenomenon that can not be simply attributed to endotoxin [46]. Epidemiological effects of LPS in dust are often found with extremes of the distribution only, either not significant [47], marginally significant [13,6,48-50], non-linear [10], heterogeneous [51] or even in the opposite direction [52-54].
The main contradiction [55], however, stems from the fact that farming is a frequent risk factor for allergy [56] and asthma [57,58]. This might be the explanation why some studies do not find any association between farming and sensitization [59,60] or even opposite results [61].
Research into the biology of endotoxin had many unexpected turns and "has engendered immense curiosity over the years" according to one of its pioneers [41,62]. "Why should diminished exposure to microrganism result in inadequate priming of T regulatory cells?" [63]. Any different LPS exposure effect in early life than later on as suggested by Martinez [64] is contradicted by studies where inhalation of LPS induces airway inflammatory response and wheezing [65-69]. This airway response was dose-dependent in both, healthy and asthmatic subjects [65], genetically determined [70-72] and may be enhanced by concomitant inhalation of allergen challenge [73]. It is therefore not unexpected that endotoxin exposure is still the main determinant of lung function decline in farmers [74,75].
Are other bacterial components relevant?
With the ubiquitous occurrence of LPS, its association also to non-farm settings [10], or other household factors [76,77] the situation is far from being clear. There might be effects by other bacterial products [37,78] but there are even considerable doubts if bacterial co-factors are responsible for the observed effects. The largest study concluded that "environmental changes affecting the whole of society have promoted an increase in asthma, allergic rhinitis and eczema in both farming and non-farming environments ... whereas the protective effect of growing up on a farm on the risk of asthma appears to be a fairly recent phenomenon" [79]. Similar conclusions are reported in the second largest study where "the percentage of subjects with symptoms of rhinitis or allergic sensitization was generally lower in subjects who had lived on a farm than in other subjects but the difference was significant only in subjects born after 1961" [80]. In addition also a study from Switzerland reported only a very recent increase of allergy in children from non-farming households [7]. If we assume that the bacterial universe did not undergo a major change since 1961, direct bateria-related effects are not very likely.
What else could explain the "non-allergic farm child" effect?
Lower air pollution by industry or car traffic could be one reason. Unfortunately, this explanation is rather unlikely as the absolute difference between the Upper Bavarian countryside and the Munich municipal area was weak [3].
Second, there might be a self-selection mechanism leading to the drop-out of allergic people, otherwise known in epidemiology as "healthy worker" effect. This phenomenon can hardly explain the absent increase of allergies during the last generation [80].
As there is a clear genetic influence on the development of allergy [81,82] there might be different genes and variants in farmers due to previous isolation. This may be assumed from the unexpected finding of longer linkage disequilibrium blocks in a recent comparison of rural and urban communities [83]. Again, this observation does not explain the recent generational increase although we have argued earlier that the reduction of newborn respiratory mortality by antibiotics may have changed our gene pool [84]. Also other environmental exposure may influence the gene pool. It could be shown recently that elevated levels of folic acid during the periconceptional period could select human embryos that carry a mutant MTHFR allele (with adverse effect on later vascular disease) [85]. Any differential exposure in farmers might therefore be important on their particular genetic background.
Fourth, the socioeconomic situation in the Alpine farmers is different compared to the major cities. There might be a lower vaccination rate although there is no evidence that early vaccination can cause later allergy [86-90]. Farmer might use less antibiotics (an effect under extensive research [87,91-94]), however, the antibiotic level in farm dust has been reported to be high [95]. In one study farm children had more siblings, were more likely to be breast-fed and to have pets [96]. In another study farm children had again more siblings, were more likely to have a cat or dog, to experience more serious respiratory infections and less likely to have attended daycare [80]. A higher number of siblings is in favour of the traditional hygiene hypothesis [15] but adjustment for family size did not resolve the farming effect. Less daycare attendance even argues against the hygiene hypothesis [97,98].
Do dietary factors play a role?
Finally, food and dietary habits may be different in farmers. For example farmers use less aggressive vitamin supplements (Figure 1, [99]). This observation may be important as vitamin D is widely used in the newborn period to prevent rickets [100] although its main metabolite is known to suppresses dendritic cell function resulting in the inability to mount a sufficient Th1 response [101]. Animal [102,103], genetic [104-106] and epidemiological studies [99,107] now support a role in the development of allergy.
Figure 1 The figure is adapted and drawn from a previous study reported in reference [99]. Included are 10,821 individuals of a Finnish birth cohort, where the percent of individuals with intake of the recommended vitamin D supplementation of 50 μg/day (2000 IU) recorded at the first birthday follow-up in 1967) is plotted against the percent of individuals sensitized against cat, birch, timothy grass or house dust mite at age 31 by profession.
This seems to be particular important as the "non-allergic farm child" effect is observed preferentially in a region only after the general introduction of vitamin D supplementation. The upsurge in allergy and asthma prevalence has been identified as a "post-1960s"-epidemic [80,108] which matches exactly with the time point of a general rickets prophylaxis approach in Bavaria [109]. Furthermore, the farm protection was seen mainly found during the first year of life [30,110] where vitamin D supplementation period is now recommended in Bavaria [111,112].
Farmers consume more local foods and less supplements. The protective effect of farm milk could relate to the avoidance of otherwise fortified milk from supermarkets [113]. Although milk is usually not fortified in the Alpine region, nearly all baby foods contain vitamin supplements. An alternative food related hypothesis has been setup for Crohns' disease [114] where the transition of cold food storage could be leading to different bacterial exposure.
Body height and head circumference, further pieces in the puzzle?
There is also another unpublished observation from our first study 1989 in Upper Bavaria where remote village size was not only associated with less allergic rhinitis but also with decreased body height. An increase in body height is a known effect of vitamin D treatment [115-119]. In a Norwegian study, male farmers were on average 2,3 cm and female famers 1,4 cm smaller (personal communication E. Omenaas 2005 [120]). A more recent German study [121] showed birth weight to be positively associated with later allergic sensitization while in British babies the head circumference was associated with the development of high IgE levels [122-125]. Do vitamin D supplements explain this association?
Relationship between hygiene and vitamin hypothesis
Both, vitamin and hygiene hypotheses are not mutually exclusive. For example there has been a higher frequency of respiratory infections in vitamin D deficient children [126-129], a phenomenon also found in farming children [80]. On a cellular level it is being known that calcitriol pulsed dendritic cells show a blunted response to LPS [130,131], where LPS pulsed IL-12 response [13,132] can override the otherwise blocking effect of calcitriol (giving possibly farming children a higher capacity to tolerate external vitamin D doses). Similar results have been obtained in human monocytic cells where LPS downregulated vitamin D receptor levels and thus inhibited vitamin D action [133].
Conclusion
Many of the clinical and epidemiological observations in the farming populations are neither conclusive nor fully understood. Will further studies in the rural Alpine foothills provide the final answer?
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
The author developed the hypothesis presented here, conducted the literature survey, wrote the paper and approved the final version of the manuscript.
Funders
My salary is paid by GSF FE 73922.
Acknowledgements
None.
==== Refs
Pirquet Allergie Münchn Med Wschr 1906 53 1457
Wjst M Reitmeir P Dold S Wulff A Nicolai T von Loeffelholz-Colberg EF von Mutius E Road traffic and adverse effects on respiratory health in children Bmj 1993 307 596 600 7691304
von Mutius E Illi S Nicolai T Martinez FD Relation of indoor heating with asthma, allergic sensitisation, and bronchial responsiveness: survey of children in south Bavaria Bmj 1996 312 1448 1450 8664621
von Mutius E Martinez FD Fritzsch C Nicolai T Roell G Thiemann HH Prevalence of asthma and atopy in two areas of West and East Germany Am J Respir Crit Care Med 1994 149 358 364 8306030
Braun-Fahrlander C Gassner M Grize L Neu U Sennhauser FH Varonier HS Vuille JC Wuthrich B Prevalence of hay fever and allergic sensitization in farmer's children and their peers living in the same rural community. SCARPOL team. Swiss Study on Childhood Allergy and Respiratory Symptoms with Respect to Air Pollution Clin Exp Allergy 1999 29 28 34 10051699 10.1046/j.1365-2222.1999.00479.x
Von Ehrenstein OS Von Mutius E Illi S Baumann L Bohm O von Kries R Reduced risk of hay fever and asthma among children of farmers Clin Exp Allergy 2000 30 187 193 10651770 10.1046/j.1365-2222.2000.00801.x
Gassner-Bachmann M Wuthrich B [Farmers' children suffer less from hay fever and asthma] Dtsch Med Wochenschr 2000 125 924 931 10967955 10.1055/s-2000-6778
Riedler J Braun-Fahrlander C Eder W Schreuer M Waser M Maisch S Carr D Schierl R Nowak D von Mutius E Exposure to farming in early life and development of asthma and allergy: a cross-sectional survey Lancet 2001 358 1129 1133 11597666 10.1016/S0140-6736(01)06252-3
Kilpelainen M Terho EO Helenius H Koskenvuo M Childhood farm environment and asthma and sensitization in young adulthood Allergy 2002 57 1130 1135 12464040 10.1034/j.1398-9995.2002.23341.x
Braun-Fahrlander C Riedler J Herz U Eder W Waser M Grize L Maisch S Carr D Gerlach F Bufe A Lauener RP Schierl R Renz H Nowak D von Mutius E Environmental exposure to endotoxin and its relation to asthma in school-age children N Engl J Med 2002 347 869 877 12239255 10.1056/NEJMoa020057
Horak FJ Studnicka M Gartner C Veiter A Tauber E Urbanek R Frischer T Parental farming protects children against atopy: longitudinal evidence involving skin prick tests Clin Exp Allergy 2002 32 1155 1159 12190651 10.1046/j.1365-2745.2002.01448.x
Portengen L Sigsgaard T Omland O Hjort C Heederik D Doekes G Low prevalence of atopy in young Danish farmers and farming students born and raised on a farm Clin Exp Allergy 2002 32 247 253 11929489 10.1046/j.1365-2222.2002.01310.x
Gereda JE Leung DY Thatayatikom A Streib JE Price MR Klinnert MD Liu AH Relation between house-dust endotoxin exposure, type 1 T-cell development, and allergen sensitisation in infants at high risk of asthma Lancet 2000 355 1680 1683 10905243 10.1016/S0140-6736(00)02239-X
Golding JPT Eczema and hay fever Butler NR, Golding J, eds From Birth to Five A Study of the Health and Behaviour of Britain's 5-Year-Olds Oxford: Pergamon Press 1986 171 186
Strachan DP Hay fever, hygiene, and household size Bmj 1989 299 1259 1260 2513902
Strachan DP Family size, infection and atopy: the first decade of the "hygiene hypothesis" Thorax 2000 55 Suppl 1 S2 10 10943631 10.1136/thorax.55.suppl_1.S2
Platts-Mills TA Asthma Severity and Prevalence: An Ongoing Interaction between Exposure, Hygiene, and Lifestyle PLoS Med 2005 2 122 126 10.1371/journal.pmed.0020034
Flohr C Pascoe D Williams HC Atopic dermatitis and the 'hygiene hypothesis': too clean to be true? Br J Dermatol 2005 152 202 216 15727630 10.1111/j.1365-2133.2004.06436.x
Forastiere F Sunyer J Farchi S Corbo G Pistelli R Baldacci S Simoni M Agabiti N Perucci CA Viegi G Number of offspring and maternal allergy Allergy 2005 60 510 514 15727585 10.1111/j.1398-9995.2005.00736.x
Maziak W The asthma epidemic and our artificial habitats BMC Pulm Med 2005 5 5 15799786 10.1186/1471-2466-5-5
Mutius von E Braun-Fahrlaender C Nowak D Riedler J Composition containing bacterial antigens, used for the prophylaxis and the treatment of allergic diseases European Patent Office; WO0149319; http://v3.espacenet.com/textdoc?DB=EPODOC&IDX=WO0149319&F=0&QPN=WO0149319&RPN=WO0149319&DOC=deb45b02a96018d3934d05c58d45ecbc84 2001
Matricardi PM Bjorksten B Bonini S Bousquet J Djukanovic R Dreborg S Gereda J Malling HJ Popov T Raz E Renz H Wold A Microbial products in allergy prevention and therapy Allergy 2003 58 461 471 12757444 10.1034/j.1398-9995.2003.00175.x
Dahl ME Dabbagh K Liggitt D Kim S Lewis DB Viral-induced T helper type 1 responses enhance allergic disease by effects on lung dendritic cells Nat Immunol 2004 5 337 343 14973436 10.1038/ni1041
Matricardi PM Rosmini F Ferrigno L Nisini R Rapicetta M Chionne P Stroffolini T Pasquini P D'Amelio R Cross sectional retrospective study of prevalence of atopy among Italian military students with antibodies against hepatitis A virus Bmj 1997 314 999 1003 9112843
Uter W Stock C Pfahlberg A Guillen-Grima F Aguinaga-Ontoso I Brun-Sandiumenge C Kramer A Association between infections and signs and symptoms of 'atopic' hypersensitivity--results of a cross-sectional survey among first-year university students in Germany and Spain Allergy 2003 58 580 584 12823114 10.1034/j.1398-9995.2003.00102.x
Jarvis D Luczynska C Chinn S Burney P The association of hepatitis A and Helicobacter pylori with sensitization to common allergens, asthma and hay fever in a population of young British adults Allergy 2004 59 1063 1067 15355464 10.1111/j.1398-9995.2004.00539.x
Gonzalez-Quintela A Gude F Boquete O Aguilera A Rey J Meijide LM Fernandez-Merino MC Vidal C Association of hepatitis A virus infection with allergic sensitization in a population with high prevalence of hepatitis A virus exposure Allergy 2005 60 98 103 15575938 10.1111/j.1398-9995.2005.00682.x
Shirakawa T Enomoto T Shimazu S Hopkin JM The inverse association between tuberculin responses and atopic disorder Science 1997 275 77 79 8974396 10.1126/science.275.5296.77
Matricardi PM Yazdanbakhsh M Mycobacteria and atopy, 6 years later: a fascinating, still unfinished, business Clin Exp Allergy 2003 33 717 720 12801302 10.1046/j.1365-2222.2003.01695.x
von Mutius E Influences in allergy: epidemiology and the environment J Allergy Clin Immunol 2004 113 373 9; quiz 380 15007331 10.1016/j.jaci.2003.12.040
Kabesch M Lauener RP Why Old McDonald had a farm but no allergies: genes, environments, and the hygiene hypothesis J Leukoc Biol 2004 75 383 387 14704371 10.1189/jlb.1003468
Gerhold K Bluemchen K Franke A Stock P Hamelmann E Exposure to endotoxin and allergen in early life and its effect on allergen sensitization in mice J Allergy Clin Immunol 2003 112 389 396 12897747 10.1067/mai.2003.1646
Velasco G Campo M Manrique OJ Bellou A He H Arestides RS Schaub B Perkins DL Finn PW Toll-like Receptor 4 or 2 Agonists Decrease Allergic Inflammation Am J Respir Cell Mol Biol 2004
Alexis NE Lay JC Almond M Bromberg PA Patel DD Peden DB Acute LPS inhalation in healthy volunteers induces dendritic cell maturation in vivo J Allergy Clin Immunol 2005 115 345 350 15696093 10.1016/j.jaci.2004.11.040
Tulic MK Wale JL Holt PG Sly PD Modification of the inflammatory response to allergen challenge after exposure to bacterial lipopolysaccharide Am J Respir Cell Mol Biol 2000 22 604 612 10783133
Backhed F Ley RE Sonnenburg JL Peterson DA Gordon JI Host-bacterial mutualism in the human intestine Science 2005 307 1915 1920 15790844 10.1126/science.1104816
van Strien RT Engel R Holst O Bufe A Eder W Waser M Braun-Fahrlander C Riedler J Nowak D von Mutius E Microbial exposure of rural school children, as assessed by levels of N-acetyl-muramic acid in mattress dust, and its association with respiratory health J Allergy Clin Immunol 2004 113 860 867 15131567 10.1016/j.jaci.2004.01.783
Singh J Schwartz DA Endotoxin and the lung: Insight into the host-environment interaction J Allergy Clin Immunol 2005 115 330 333 15696090 10.1016/j.jaci.2004.11.021
Dubin W Martin TR Swoveland P Leturcq DJ Moriarty AM Tobias PS Bleecker ER Goldblum SE Hasday JD Asthma and endotoxin: lipopolysaccharide-binding protein and soluble CD14 in bronchoalveolar compartment Am J Physiol 1996 270 L736 44 8967507
Sigsgaard THD On the hygiene hypothesis: Regulation down, up, or sideways? Journal of Allergy and Clinical Immunology 2005 online 23 March, 2005
Beutler B SHIP, TGF-beta, and endotoxin tolerance Immunity 2004 21 134 135 15308094 10.1016/j.immuni.2004.07.014
Langenkamp A Nagata K Murphy K Wu L Lanzavecchia A Sallusto F Kinetics and expression patterns of chemokine receptors in human CD4+ T lymphocytes primed by myeloid or plasmacytoid dendritic cells Eur J Immunol 2003 33 474 482 12645946 10.1002/immu.200310023
Ogura Y Bonen DK Inohara N Nicolae DL Chen FF Ramos R Britton H Moran T Karaliuskas R Duerr RH Achkar JP Brant SR Bayless TM Kirschner BS Hanauer SB Nunez G Cho JH A frameshift mutation in NOD2 associated with susceptibility to Crohn's disease Nature 2001 411 603 606 11385577 10.1038/35079114
Pulendran B Kumar P Cutler CW Mohamadzadeh M Van Dyke T Banchereau J Lipopolysaccharides from distinct pathogens induce different classes of immune responses in vivo J Immunol 2001 167 5067 5076 11673516
Williams LK Ownby DR Maliarik MJ Johnson CC The role of endotoxin and its receptors in allergic disease Ann Allergy Asthma Immunol 2005 94 323 332 15801242
Erwin EA Wickens K Custis NJ Siebers R Woodfolk J Barry D Crane J Platts-Mills TA Cat and dust mite sensitivity and tolerance in relation to wheezing among children raised with high exposure to both allergens J Allergy Clin Immunol 2005 115 74 79 15637550 10.1016/j.jaci.2004.10.030
Merchant JA Naleway AL Svendsen ER Kelly KM Burmeister LF Stromquist AM Taylor CD Thorne PS Reynolds SJ Sanderson WT Chrischilles EA Asthma and farm exposures in a cohort of rural Iowa children Environ Health Perspect 2005 113 350 356 15743727
Chrischilles E Ahrens R Kuehl A Kelly K Thorne P Burmeister L Merchant J Asthma prevalence and morbidity among rural Iowa schoolchildren J Allergy Clin Immunol 2004 113 66 71 14713909 10.1016/j.jaci.2003.09.037
Salam MT Li YF Langholz B Gilliland FD Early-life environmental risk factors for asthma: findings from the Children's Health Study Environ Health Perspect 2004 112 760 765 15121522
Schram D Doekes G Boeve M Douwes J Riedler J Ublagger E Mutius E Budde J Pershagen G Nyberg F Alm J Braun-Fahrlander C Waser M Brunekreef B Bacterial and fungal components in house dust of farm children, Rudolf Steiner school children and reference children - the PARSIFAL Study Allergy 2005 60 611 618 15813805 10.1111/j.1398-9995.2005.00748.x
Downs SH Marks GB Mitakakis TZ Leuppi JD Car NG Peat JK Having lived on a farm and protection against allergic diseases in Australia Clin Exp Allergy 2001 31 570 575 11359424 10.1046/j.1365-2222.2001.01070.x
Wickens K Lane JM Fitzharris P Siebers R Riley G Douwes J Smith T Crane J Farm residence and exposures and the risk of allergic diseases in New Zealand children Allergy 2002 57 1171 1179 12464046 10.1034/j.1398-9995.2002.t01-1-23644.x
Bolte G Bischof W Borte M Lehmann I Wichmann HE Heinrich J Early endotoxin exposure and atopy development in infants: results of a birth cohort study Clin Exp Allergy 2003 33 770 776 12801311 10.1046/j.1365-2222.2003.01665.x
Park JH Gold DR Spiegelman DL Burge HA Milton DK House dust endotoxin and wheeze in the first year of life Am J Respir Crit Care Med 2001 163 322 328 11179100
Liu AH Endotoxin exposure in allergy and asthma: reconciling a paradox J Allergy Clin Immunol 2002 109 379 392 11897980 10.1067/mai.2002.122157
Monso E Magarolas R Radon K Danuser B Iversen M Weber C Opravil U Donham KJ Nowak D Respiratory symptoms of obstructive lung disease in European crop farmers Am J Respir Crit Care Med 2000 162 1246 1250 11029325
Kogevinas M Anto JM Sunyer J Tobias A Kromhout H Burney P Occupational asthma in Europe and other industrialised areas: a population-based study. European Community Respiratory Health Survey Study Group Lancet 1999 353 1750 1754 10347988 10.1016/S0140-6736(98)07397-8
Kronqvist M Johansson E Pershagen G Johansson SG van Hage-Hamsten M Risk factors associated with asthma and rhinoconjunctivitis among Swedish farmers Allergy 1999 54 1142 1149 10604549 10.1034/j.1398-9995.1999.00115.x
Vidal C Boquete O Gude F Rey J Meijide LM Fernandez-Merino MC Gonzalez-Quintela A High prevalence of storage mite sensitization in a general adult population Allergy 2004 59 401 405 15005763 10.1111/j.1398-9995.2003.00430.x
Chatzi L Prokopakis E Tzanakis N Alegakis A Bizakis I Siafakas N Lionis C Allergic rhinitis, asthma, and atopy among grape farmers in a rural population in crete, Greece Chest 2005 127 372 378 15654002 10.1378/chest.127.1.372
Hoppin JA Umbach DM London SJ Alavanja MC Sandler DP Animal production and wheeze in the Agricultural Health Study: interactions with atopy, asthma, and smoking Occup Environ Med 2003 60 e3 12883030 10.1136/oem.60.8.e3
Beutler B Rietschel ET Innate immune sensing and its roots: the story of endotoxin Nat Rev Immunol 2003 3 169 176 12563300 10.1038/nri1004
Rook GA Martinelli R Brunet LR Ernst Schering Res Found Workshop 2005 Modelling gene-environment interactions in Th1- and Th2-dominated diseases of laboratory animals Animal models of T cell-mediated skin diseases Ernst Schering Research Foundation Workshop, 50 (50) Springer Verlag, pp 45-68 ISBN 3540210679 2005
Martinez FD Holt PG Role of microbial burden in aetiology of allergy and asthma Lancet 1999 354 Suppl 2 SII12 5 10507253 10.1016/S0140-6736(99)90437-3
Michel O Nagy AM Schroeven M Duchateau J Neve J Fondu P Sergysels R Dose-response relationship to inhaled endotoxin in normal subjects Am J Respir Crit Care Med 1997 156 1157 1164 9351616
Reed CE Milton DK Endotoxin-stimulated innate immunity: A contributing factor for asthma J Allergy Clin Immunol 2001 108 157 166 11496229 10.1067/mai.2001.116862
Sandstrom T Bjermer L Rylander R Lipopolysaccharide (LPS) inhalation in healthy subjects increases neutrophils, lymphocytes and fibronectin levels in bronchoalveolar lavage fluid Eur Respir J 1992 5 992 996 1426208
Rizzo MC Naspitz CK Fernandez-Caldas E Lockey RF Mimica I Sole D Endotoxin exposure and symptoms in asthmatic children Pediatr Allergy Immunol 1997 8 121 126 9532251
Litonjua AA Milton DK Celedon JC Ryan L Weiss ST Gold DR A longitudinal analysis of wheezing in young children: the independent effects of early life exposure to house dust endotoxin, allergens, and pets J Allergy Clin Immunol 2002 110 736 742 12417882 10.1067/mai.2002.128948
Arbour NC Lorenz E Schutte BC Zabner J Kline JN Jones M Frees K Watt JL Schwartz DA TLR4 mutations are associated with endotoxin hyporesponsiveness in humans Nat Genet 2000 25 187 191 10835634 10.1038/76048
Werner M Topp R Wimmer K Richter K Bischof W Wjst M Heinrich J TLR4 gene variants modify endotoxin effects on asthma J Allergy Clin Immunol 2003 112 323 330 12897738 10.1067/mai.2003.1648
Michel O LeVan TD Stern D Dentener M Thorn J Gnat D Beijer ML Cochaux P Holt PG Martinez FD Rylander R Systemic responsiveness to lipopolysaccharide and polymorphisms in the toll-like receptor 4 gene in human beings J Allergy Clin Immunol 2003 112 923 929 14610481 10.1016/j.jaci.2003.05.001
Michel O Dentener M Corazza F Buurman W Rylander R Healthy subjects express differences in clinical responses to inhaled lipopolysaccharide that are related with inflammation and with atopy J Allergy Clin Immunol 2001 107 797 804 11344345 10.1067/mai.2001.114249
Vogelzang PF van der Gulden JW Folgering H Kolk JJ Heederik D Preller L Tielen MJ van Schayck CP Endotoxin exposure as a major determinant of lung function decline in pig farmers Am J Respir Crit Care Med 1998 157 15 18 9445272
Monso E Riu E Radon K Magarolas R Danuser B Iversen M Morera J Nowak D Chronic obstructive pulmonary disease in never-smoking animal farmers working inside confinement buildings Am J Ind Med 2004 46 357 362 15376214 10.1002/ajim.20077
Heinrich J Gehring U Douwes J Koch A Fahlbusch B Bischof W Wichmann HE Pets and vermin are associated with high endotoxin levels in house dust Clin Exp Allergy 2001 31 1839 1845 11737034 10.1046/j.1365-2222.2001.01220.x
Gereda JE Klinnert MD Price MR Leung DY Liu AH Metropolitan home living conditions associated with indoor endotoxin levels J Allergy Clin Immunol 2001 107 790 796 11344344 10.1067/mai.2001.115245
Roy SR Schiltz AM Marotta A Shen Y Liu AH Bacterial DNA in house and farm barn dust J Allergy Clin Immunol 2003 112 571 578 13679817 10.1016/S0091-6749(03)01863-3
Braback L Hjern A Rasmussen F Trends in asthma, allergic rhinitis and eczema among Swedish conscripts from farming and non-farming environments. A nationwide study over three decades Clin Exp Allergy 2004 34 38 43 14720260 10.1111/j.1365-2222.2004.01841.x
Leynaert B Neukirch C Jarvis D Chinn S Burney P Neukirch F Does living on a farm during childhood protect against asthma, allergic rhinitis, and atopy in adulthood? Am J Respir Crit Care Med 2001 164 1829 1834 11734431
Dold S Heinrich J Wichmann HE Wjst M Ascaris-specific IgE and allergic sensitization in a cohort of school children in the former East Germany J Allergy Clin Immunol 1998 102 414 420 9768582
Hoffjan S Nicolae D Ober C Association studies for asthma and atopic diseases: a comprehensive review of the literature Respir Res 2003 4 14 14748924 10.1186/1465-9921-4-14
Vitart V Carothers AD Hayward C Teague P Hastie ND Campbell H Wright AF Increased level of linkage disequilibrium in rural compared with urban communities: a factor to consider in association-study design Am J Hum Genet 2005 76 763 772 15791542 10.1086/429840
Wjst M Is the increase in allergic asthma associated with an inborn Th1 maturation or with an environmental Th1 trigger defect? Allergy 2004 59 148 150 14763926 10.1046/j.1398-9995.2003.00374.x
Lucock M Yates Z Folic acid - vitamin and panacea or genetic time bomb? Nat Rev Genet 2005 6 235 240 15738964 10.1038/nrg1558
Koppen S de Groot R Neijens HJ Nagelkerke N van Eden W Rumke HC No epidemiological evidence for infant vaccinations to cause allergic disease Vaccine 2004 22 3375 3385 15308362 10.1016/j.vaccine.2004.02.033
Kemp T Pearce N Fitzharris P Crane J Fergusson D St George I Wickens K Beasley R Is infant immunization a risk factor for childhood asthma or allergy? Epidemiology 1997 8 678 680 9345669 10.1097/00001648-199710000-00011
Nilsson L Kjellman NI Bjorksten B Allergic disease at the age of 7 years after pertussis vaccination in infancy: results from the follow-up of a randomized controlled trial of 3 vaccines Arch Pediatr Adolesc Med 2003 157 1184 1189 14662571 10.1001/archpedi.157.12.1184
Anderson HR Poloniecki JD Strachan DP Beasley R Bjorksten B Asher MI Immunization and symptoms of atopic disease in children: results from the International Study of Asthma and Allergies in Childhood Am J Public Health 2001 91 1126 1129 11441744
Alm JS Swartz J Lilja G Scheynius A Pershagen G Atopy in children of families with an anthroposophic lifestyle Lancet 1999 353 1485 1488 10232315 10.1016/S0140-6736(98)09344-1
Wjst M Hoelscher B Frye C Wichmann HE Dold S Heinrich J Early antibiotic treatment and later asthma Eur J Med Res 2001 6 263 271 11432798
Celedon JC Fuhlbrigge A Rifas-Shiman S Weiss ST Finkelstein JA Antibiotic use in the first year of life and asthma in early childhood Clin Exp Allergy 2004 34 1011 1016 15248843 10.1111/j.1365-2222.2004.01994.x
Noverr MC Falkowski NR McDonald RA McKenzie AN Huffnagle GB Development of allergic airway disease in mice following antibiotic therapy and fungal microbiota increase: role of host genetics, antigen, and interleukin-13 Infect Immun 2005 73 30 38 15618138 10.1128/IAI.73.1.30-38.2005
Bashir ME Louie S Shi HN Nagler-Anderson C Toll-like receptor 4 signaling by intestinal microbes influences susceptibility to food allergy J Immunol 2004 172 6978 6987 15153518
Hamscher G Pawelzick HT Sczesny S Nau H Hartung J Antibiotics in dust originating from a pig-fattening farm: a new source of health hazard for farmers? Environ Health Perspect 2003 111 1590 1594 14527837
Adler A Tager I Quintero DR Decreased prevalence of asthma among farm-reared children compared with those who are rural but not farm-reared J Allergy Clin Immunol 2005 115 67 73 15637549 10.1016/j.jaci.2004.10.008
Ball TM Castro-Rodriguez JA Griffith KA Holberg CJ Martinez FD Wright AL Siblings, day-care attendance, and the risk of asthma and wheezing during childhood N Engl J Med 2000 343 538 543 10954761 10.1056/NEJM200008243430803
Kramer U Heinrich J Wjst M Wichmann HE Age of entry to day nursery and allergy in later childhood Lancet 1999 353 450 454 9989715 10.1016/S0140-6736(98)06329-6
Hyppönen E Sovio U Wjst M Patel S Pekkanen J Hartikainen AL Järvelin MR Infant Vitamin D supplementation and allergic conditions in adulthood: Northen Finland birth cohort 1996 Ann NY Acad Sci 2004 1037 84 95 15699498 10.1196/annals.1337.013
Wjst M Dold S Genes, factor X, and allergens: what causes allergic diseases? Allergy 1999 54 757 759 10442534 10.1034/j.1398-9995.1999.00193.x
Feldman D Pike J Wesley Glorreux FH Vitamin D Academic Press 1997
Matheu V Back O Mondoc E Issazadeh-Navikas S Dual effects of vitamin D-induced alteration of TH1/TH2 cytokine expression: enhancing IgE production and decreasing airway eosinophilia in murine allergic airway disease J Allergy Clin Immunol 2003 112 585 592 13679819 10.1016/S0091-6749(03)01855-4
Wittke A Weaver V Mahon BD August A Cantorna MT Vitamin D receptor-deficient mice fail to develop experimental allergic asthma J Immunol 2004 173 3432 3436 15322208
Poon AH Laprise C Lemire M Montpetit A Sinnett D Schurr E Hudson TJ Association of vitamin D receptor genetic variants with susceptibility to asthma and atopy Am J Respir Crit Care Med 2004 170 967 973 15282199 10.1164/rccm.200403-412OC
Raby BA Lazarus R Silverman EK Lake S Lange C Wjst M Weiss ST Association of vitamin D receptor gene polymorphisms with childhood and adult asthma Am J Respir Crit Care Med 2004 170 1057 1065 15282200 10.1164/rccm.200404-447OC
Wjst M Genetic variants in the vitamin D receptor and childhood asthma BMC Genetics 2005 6 2 15651992 10.1186/1471-2156-6-2
Milner JD Stein DM McCarter R Moon RY Early infant multivitamin supplementation is associated with increased risk for food allergy and asthma Pediatrics 2004 114 27 32 15231904 10.1542/peds.114.1.27
Holt PG Sly PD Martinez FD Weiss ST Bjorksten B von Mutius E Wahn U Drug development strategies for asthma: in search of a new paradigm Nat Immunol 2004 5 695 698 15224096 10.1038/ni0704-695
Wjst M The triple T allergy hypothesis Clin Dev Immunol 2004 11 175 15330454 10.1080/10446670410001722258
Radon K Ehrenstein V Praml G Nowak D Childhood visits to animal buildings and atopic diseases in adulthood: an age-dependent relationship Am J Ind Med 2004 46 349 356 15376228 10.1002/ajim.20000
Kersting M Ernährung des gesunden Säuglings, Lebensmittel- und mahlzeitenbezogene Empfehlungen. Monatsschr Kinderheilkd 2001 149 4 10 10.1007/s001120050716
Kersting M Alexy U Vitamin and mineral supplements for the use of children on the German market: products, nutrients, dosages Ann Nutr Metab 2000 44 125 128 11053899 10.1159/000012834
Barnes M Cullinan P Athanasaki P MacNeill S Hole AM Harris J Kalogeraki S Chatzinikolaou M Drakonakis N Bibaki-Liakou V Newman Taylor AJ Bibakis I Crete: does farming explain urban and rural differences in atopy? Clin Exp Allergy 2001 31 1822 1828 11737032 10.1046/j.1365-2222.2001.01240.x
Hugot JP Alberti C Berrebi D Bingen E Cezard JP Crohn's disease: the cold chain hypothesis Lancet 2003 362 2012 2015 14683664 10.1016/S0140-6736(03)15024-6
Brooke OG Butters F Wood C Intrauterine vitamin D nutrition and postnatal growth in Asian infants Br Med J (Clin Res Ed) 1981 283 1024 6794748
Sankaran K Papageorgiou A Ninan A Sankaran R A randomized, controlled evaluation of two commercially available human breast milk fortifiers in healthy preterm neonates J Am Diet Assoc 1996 96 1145 1149 8906139 10.1016/S0002-8223(96)00295-7
Du X Zhu K Trube A Zhang Q Ma G Hu X Fraser DR Greenfield H School-milk intervention trial enhances growth and bone mineral accretion in Chinese girls aged 10-12 years in Beijing Br J Nutr 2004 92 159 168 15230999 10.1079/BJN20041118
Burne TH McGrath JJ Eyles DW Mackay-Sim A Behavioural characterization of Vitamin D receptor knockout mice Behav Brain Res 2005 157 299 308 15639181 10.1016/j.bbr.2004.07.008
Liu Y Albertsson-Wikland K Karlberg J Long-term consequences of early linear growth retardation (stunting) in Swedish children Pediatr Res 2000 47 475 480 10759154
Eduard W Omenaas E Bakke PS Douwes J Heederik D Atopic and non-atopic asthma in a farming and a general population Am J Ind Med 2004 46 396 399 15376208 10.1002/ajim.20088
Bolte G Schmidt M Maziak W Keil U Nasca P von Mutius E Weiland SK The relation of markers of fetal growth with asthma, allergies and serum immunoglobulin E levels in children at age 5-7 years Clin Exp Allergy 2004 34 381 388 15005731 10.1111/j.1365-2222.2004.01890.x
Leadbitter P Pearce N Cheng S Sears MR Holdaway MD Flannery EM Herbison GP Beasley R Relationship between fetal growth and the development of asthma and atopy in childhood Thorax 1999 54 905 910 10491453
Beasley R Leadbitter P Pearce N Crane J Is enhanced fetal growth a risk factor for the development of atopy or asthma? Int Arch Allergy Immunol 1999 118 408 410 10224460 10.1159/000024149
Gregory A Doull I Pearce N Cheng S Leadbitter P Holgate S Beasley R The relationship between anthropometric measurements at birth: asthma and atopy in childhood Clin Exp Allergy 1999 29 330 333 10202339 10.1046/j.1365-2222.1999.00485.x
Katz KA Pocock SJ Strachan DP Neonatal head circumference, neonatal weight, and risk of hayfever, asthma and eczema in a large cohort of adolescents from Sheffield, England Clin Exp Allergy 2003 33 737 745 12801306 10.1046/j.1365-2222.2003.01670.x
Beser E Cakmakci T Factors affecting the morbidity of vitamin D deficiency rickets and primary protection East Afr Med J 1994 71 358 362 7835254
Wayse V Yousafzai A Mogale K Filteau S Association of subclinical vitamin D deficiency with severe acute lower respiratory infection in Indian children under 5 y Eur J Clin Nutr 2004 58 563 567 15042122 10.1038/sj.ejcn.1601845
Najada AS Habashneh MS Khader M The frequency of nutritional rickets among hospitalized infants and its relation to respiratory diseases J Trop Pediatr 2004 50 364 368 15537725 10.1093/tropej/50.6.364
Sachan A Gupta R Das V Agarwal A Awasthi PK Bhatia V High prevalence of vitamin D deficiency among pregnant women and their newborns in northern India Am J Clin Nutr 2005 81 1060 1064 15883429
Griffin MD Lutz W Phan VA Bachman LA McKean DJ Kumar R Dendritic cell modulation by 1alpha,25 dihydroxyvitamin D3 and its analogs: a vitamin D receptor-dependent pathway that promotes a persistent state of immaturity in vitro and in vivo Proc Natl Acad Sci U S A 2001 98 6800 6805 11371626 10.1073/pnas.121172198
Lyakh LA Sanford M Chekol S Young HA Roberts AB TGF-{beta} and Vitamin D3 Utilize Distinct Pathways to Suppress IL-12 Production and Modulate Rapid Differentiation of Human Monocytes into CD83+ Dendritic Cells J Immunol 2005 174 2061 2070 15699136
Kuipers H Hijdra D De Vries VC Hammad H Prins JB Coyle AJ Hoogsteden HC Lambrecht BN Lipopolysaccharide-induced suppression of airway Th2 responses does not require IL-12 production by dendritic cells J Immunol 2003 171 3645 3654 14500662
Pramanik R Asplin JR Lindeman C Favus MJ Bai S Coe FL Lipopolysaccharide negatively modulates vitamin D action by down-regulating expression of vitamin D-induced VDR in human monocytic THP-1 cells Cell Immunol 2004 232 137 43 15876428 10.1016/j.cellimm.2005.03.004
| 15935102 | PMC1177973 | CC BY | 2021-01-04 16:36:25 | no | Clin Mol Allergy. 2005 Jun 5; 3:7 | utf-8 | Clin Mol Allergy | 2,005 | 10.1186/1476-7961-3-7 | oa_comm |
==== Front
CytojournalCytoJournal1742-6413BioMed Central London 1742-6413-2-71594148910.1186/1742-6413-2-7ReviewFine needle aspiration biopsy of the liver: Algorithmic approach and current issues in the diagnosis of hepatocellular carcinoma Wee Aileen [email protected] Professor and Senior Consultant, Department of Pathology, National University of Singapore, National University Hospital, 5 Lower Kent Ridge Road, Singapore, 119074, Republic of Singapore2005 8 6 2005 2 7 7 14 2 2005 8 6 2005 Copyright © 2005 Wee; licensee BioMed Central Ltd.2005Wee; 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 role of fine needle aspiration biopsy (FNAB) in the evaluation of focal liver lesions has evolved. Guided FNAB is still useful to procure a tissue diagnosis if clinical, biochemical and radiologic findings are inconclusive. Major diagnostic issues include: (i) Distinction of benign hepatocellular nodular lesions from reactive hepatocytes, (ii) Distinction of well-differentiated hepatocellular carcinoma (WD-HCC) from benign hepatocellular nodular lesions, (iii) Distinction of poorly differentiated HCC from cholangiocarcinoma and metastatic carcinomas, (iv) Determination of histogenesis of malignant tumor, and (v) Determination of primary site of origin of malignant tumor. This review gives a general overview of hepatic FNAB; outlines an algorithmic approach to cytodiagnosis with emphasis on HCC, its variants and their mimics; and addresses current diagnostic issues. Close radiologic surveillance of high-risk cirrhotic patients has resulted in the increasing detection of smaller lesions with many subjected to biopsy for tissue characterization. The need for tissue confirmation in clinically obvious HCC is questioned due to risk of malignant seeding. When a biopsy is indicated, core needle biopsy is favored over FNAB. The inherent difficulty of distinguishing small/early HCC from benign hepatocellular nodular lesions has resulted in indeterminate reports. Changing concepts in the understanding of the biological behavior and morphologic evolution of HCC and its precursors; and the current lack of agreement on the morphologic criteria for distinguishing high-grade dysplastic lesions (with small cell change) from WD-HCC, have profound impact on nomenclature, cytohistologic interpretation and management. Optimization of hepatic FNAB to enhance the yield and accuracy of diagnoses requires close clinicopathologic correlation; combined cytohistologic approach; judicious use of ancillary tests; and skilled healthcare teams.
Diagnostic algorithmfine needle aspiration biopsyhepatocellular carcinomaimmunohistochemistryliver
==== Body
Review
Until recently, guided fine needle aspiration biopsy (FNAB) was accepted as a safe procedure to procure tissue diagnosis in the management of patients with focal liver lesions. However, the role of FNAB in the evaluation of such lesions, especially hepatocellular carcinoma (HCC), and the diagnostic challenges it poses have evolved [1-4]. Much diagnostic information can be gleaned nowadays from an ever-expanding array of tumor markers and sophisticated imaging modalities. This has obviated the need, in some practices, for tissue confirmation in clinically obvious HCC. Surveillance of high-risk cirrhotic patients, has contributed to an increasing detection of focal liver lesions that are <2 cm in diameter [5,6]. Study of autopsy material and cirrhotic explants, coupled with immunohistochemical and/other ancillary techniques, have led to better understanding of the morphologic evolution and biological behavior of certain hepatocellular nodular lesions [7-10]. This changing concept in the evolution of HCC and its precursor lesions has great impact on nomenclature [11], cytohistologic interpretation, management strategies and treatment outcomes [5,12]. Critical refinement of current histopathologic criteria for the diagnosis of small ("early") HCC is necessary with increasing demands for tissue characterization of small "suspicious" nodules [13-17]. The window of opportunity in the treatment of HCC is limited as patients tend to present with advanced cancers.
The review gives an algorithmic approach to the FNAB diagnosis of focal liver lesions; an overview of current diagnostic issues concerning hepatic FNAB; problems and pitfalls in the cytodiagnosis of well-differentiated hepatocellular nodular lesions; and diagnostic utility of immunohistochemistry.
General approach to focal liver lesions
Focal lesions in the liver range from cysts and inflammatory processes to neoplasms, be they benign or malignant, primary or metastatic [11,18]. This realm of pathology comes with its share of diagnostic challenges and pitfalls. Although imaging and tumor markers can provide a diagnosis in many instances, tissue confirmation may be warranted under circumstances where the clinical, biochemical and imaging profiles are not conclusive [19]. A combined smearing and microhistology approach is strongly recommended [20-24]. Smears are air-dried and stained with Diff-Quik/May-Grunwald-Giemsa as well as fixed in 95% alcohol and stained by the Papanicolaou method. Particulate material is formalin-fixed for preparation of cell blocks for histologic study and for special and immunostains. Use of FNAB needles (21 gauge) with cutting mechanism may provide microbiopsy tissue cores.
Apart from cystic/inflammatory conditions, the major diagnostic issues are: (i) Distinction of benign hepatocellular nodular lesions, namely, macroregenerative nodule (MRN), dysplastic nodule (DN), focal nodular hyperplasia (FNH) and liver cell adenoma (LCA), from reactive hepatocytes; (ii) Distinction of well-differentiated HCC (WD-HCC) from benign hepatocellular nodular lesions; (iii) Distinction of poorly differentiated HCC (PD-HCC) from cholangiocarcinoma (CC) and metastatic carcinomas; (iv) Determination of histogenesis of malignant tumor; and (v) Determination of primary site of origin of malignant tumor.
Algorithmic approach
A diagnostic algorithm for FNAB diagnosis of focal liver lesions is given in Table 1. The emphasis is on HCC, its variants, and their differentiation from benign and malignant mimics [20].
Table 1 Diagnostic algorithm for fine needle aspiration cytology of the liver
Patient with liver mass/es
STEP 1: Establish category of clinical presentation
Imaging
STEP 2: Establish category of radiologic findings
Fine needle aspiration biopsy
STEP 3: Establish nature of cytohistologic findings
Ancillary techniques
(special stains, immunohistochemistry)
STEP 4: Further confirm nature of cytohistologic findings
Clinicopathologic correlation
STEP 5: Establish final diagnosis based on multidisciplinary approach
Step 1: Establish category of clinical presentation
A patient may present with a liver mass/es under the following clinical scenarios: (i) Routine medical check-up, (ii) Chronic liver disease due to, for example, hepatitis B or C virus infection, and alcoholism, (iii) Known cancer cases, (iv) Symptomatic patients, and (v) In childhood. Important relevant data include serum alpha-fetoprotein (AFP) and carcinoembryonic antigen (CEA) levels, hepatitis virus markers, results of liver function tests, presence of cirrhosis, biliary tract disease, calculi, liver fluke infestation, and history of hormone usage. Radiologic correlation is mandatory.
Step 2: Establish category of radiologic findings
In many instances, a preoperative diagnosis can be achieved with a high degree of accuracy based on non-invasive imaging techniques and close clinical correlation. FNAB is useful in defining those lesions without characteristic imaging appearance. The solid or cystic nature of the lesion; number, size and location of the lesion/s; absence or presence of hepatomegaly, cirrhosis, steatosis, regional lymphadenopathy and calculi; and status of the biliary tract are important clues to the final diagnosis. Imaging of liver masses can be divided into two categories, namely, cystic and solid lesions. The focus is on a rational and pragmatic approach to hepatic FNAB. A list of entities is given in Table 2 as a working guide.
Table 2 Cystic and Solid Lesions of the Liver
CYSTIC LESIONS
BENIGN AND MALIGNANT ENTITIES
Congenital cysts
Parasitic cysts: Hydatid cysts
Abscesses: Pyogenic, amebic and fungal abscesses
Granulomas
Inflammatory pseudotumors (can be solid)
Cystic neoplasms: Biliary cystadenoma/cystadenocarcinoma, cholangiocarcinoma associated with cystic liver disease, cystic metastases (from ovary, pancreas)
SOLID LESIONS
BENIGN ENTITIES
Benign hepatocellular nodular lesions
Regenerative nodules (cirrhosis, nodular regenerative hyperplasia)
Macroregenerative nodule
Dysplastic nodule
Focal nodular hyperplasia
Liver cell adenoma
Focal fatty change
Benign non-hepatocellular nodular lesions
Bile duct adenoma, hamartoma
Hemangioma
Angiomyolipoma
Extramedullary hematopioesis
MALIGNANT ENTITIES
Classic primary liver cancers
Hepatocellular carcinoma (HCC)
Intrahepatic cholangiocarcinoma
Variants of primary liver cancers
Variants of hepatocellular carcinoma:
HCC with fatty change
HCC, clear cell type
HCC, small cell type
HCC, undifferentiated type
HCC, spindle cell type
HCC, giant cell type
Fibrolamellar HCC
HCC with biliary differentiation
Combined hepatocellular-cholangiocarcinoma
Variants of cholangiocarcinoma:
Biliary papillary neoplasia / Intraductal papillary cholangiocarcinoma
Mucinous intrahepatic cholangiocarcinoma
Mimics of primary liver cancers
Adenocarcinoma
Squamous cell carcinoma
Small round cell malignancy
Clear cell malignancy
Pleomorphic cell malignancy
Spindle cell malignancy
Giant cell malignancy
Hepatoid carcinoma; alpha-fetoprotein-producing carcinoma
Step 3: Establish nature of cytohistologic findings
FNAB of normal/reactive liver
The liver parenchyma comprises a heterogeneous population of hepatobiliary and related cells, namely, hepatocytes, bile duct and ductular epithelia (Figure 1); and Kupffer, endothelial, mesothelial and inflammatory cells. Hepatocytes often contain intracytoplasmic inclusions, such as, fat vacuoles, Mallory bodies and to a lesser extent, hyaline bodies; as well as intranuclear cytoplasmic inclusions. Pigments, such as lipofuscin (Figure 2), bile and iron, may be present.
Figure 1 Normal and reactive hepatocytes: FNAB. Monolayered clusters of loosely-cohesive, well-defined, polygonal cells with ample dense, granular cytoplasm, round central nuclei and low nuclear-cytoplasmic ratio. Note the double-layered row of ductular epithelial cells with scanty cytoplasm (May-Grunwald-Giemsa).
Figure 2 Normal and reactive hepatocytes: FNAB. (A) Hepatocytes show dense granular cytoplasm, round central nuclei, well-delineated nuclear membrane, distinct nucleoli, granular chromatin and binucleation. Note polymorphism displayed by nonneoplastic hepatocytes. The cells contain brown granules of lipofuscin pigment in the cytoplasm (Papanicolaou). (B) Lipofuscin appears as black granules. The two elongated nuclei are likely to be Kupffer cells (May-Grunwald-Giemsa).
FNAB of liver with large and small cell change
The terms "large cell change" and "small cell change" have replaced large and small cell dysplasia [18]. Hepatocytes with large cell change, considered a low-grade lesion, exhibit both cell and nuclear enlargement with nuclear atypia but retaining the normal nuclear-cytoplasmic ratio (N/C) of ≤ 1/3, by eyeballing the diameters of the nucleus and the cell (Figure 3). On the other hand, in small cell change, considered a high-grade lesion and putative precancerous link with HCC, the hepatocytes are small and monotonous with subtle increase in N/C ratio; they impart an impression of nuclear crowding. Both types of changes are common in cirrhosis. The atypical cells can form dysplastic foci (<1 mm diameter) or nodules (>1 mm diameter) in both cirrhotic and non-cirrhotic livers [18].
Figure 3 Hepatocytes with large cell change: FNAB. There is simultaneous nuclear and cell enlargement of the hepatocytes, thus maintaining the nuclear-cytoplasmic ratio of 1/3. Note mild nuclear atypia (Papanicolaou).
FNAB of a liver mass: Stepwise approach
Liver aspirates can come from malignant or benign conditions of hepatocellular or non-hepatocellular origin [25-27]. Certain entities can give characteristic gross appearances when smeared [20]. Naked eye inspection coupled with low power scanning view of the smears is a helpful adjunctive step to determine: (i) Patterns of spread of smears, (ii) Malignant or benign cell picture, (iii) Hepatocytic nature or otherwise, and (iv) Monotonous hepatocytic population or heterogeneous liver parenchymal components.
Patterns of spread of smears
A. Uniformly granular pattern. Hypercellular. Regularly irregular tissue fragments; evenly distributed in rows with tendency to be equidistant. Most likely HCC (Figure 4).
Figure 4 Naked eye inspection of hepatic aspirate. Uniformly granular pattern of spread of classic hepatocellular carcinoma. Note the regularly irregular tumor fragments which tend to be equidistant (Papanicolaou).
B. Non-uniformly granular pattern. Hypercellular. A mixed picture of irregular tissue fragments of variable size/shape; may include uniform granules. Variable cell picture -malignant to benign lesions of any origin.
C. Scanty pattern. Hypocellular. Streaks or fine clumps of cells. Variable cell picture -desmoplastic malignancies or benign conditions, such as cyst contents and hepatocellular nodular lesions. Pitfalls: Necrotic or cystic neoplasm; non-lesional sample.
D. Fluid pattern. Amorphous look. Thin fluid or thick pus. Usually benign and non-hepatocellular, such as abscess or infected cyst contents. Pitfalls: Dissociated tumor cells, such as lymphomas.
E. Microbiopsy pattern. Short and narrow, difficult-to-focus tissue cores, usually composed of nonneoplastic liver parenchyma with intact architecture. Admixed with other patterns, if the specimen is representative and includes lesional tissue. The clue is from the other patterns. Rarely, tumor tissue.
FNAB of hepatocellular carcinoma
HCC can be small/focal, solitary/large, and multifocal/diffuse; with satellite nodules and large vein involvement. Classic HCC is usually graded into well, moderately or poorly differentiated lesions. Histologic patterns comprise trabecular-sinusoidal, pseudoacinar and solid types; combinations are frequent [28]. Close attention should be paid to architectural details in cell blocks/microbiopsies and smears. Accurate distinction from metastases, especially unresectable lesions, is necessary for appropriate therapy. One should be aware that there are limitations to the cytodiagnosis of HCC [23,29-31].
Smears
• Hypercellular smears with uniformly granular pattern of spread of the cells.
• Cohesive clusters of malignant hepatocytes with arborizing, tongue-like projections of broad cords (>2 cells thick) that may be wrapped by peripheral endothelium (Figures 5, 6).
Figure 5 Well-differentiated hepatocellular carcinoma with trabecular and pseudoacinar patterns: FNAB. (A) Thick arborizing cords of malignant hepatocytes showing cellular monotony, increased nuclear-cytoplasmic ratio, and impression of nuclear crowding. The circular spaces among the cords represent pseudoacini (Papanicolaou). (B) Corresponding histologic section of the tumor shows trabecular-sinusoidal arrangement with pseudoacini. Note the uniformity of the tumor cells and cords 2 to 3 cells thick (H&E).
Figure 6 Moderately differentiated hepatocellular carcinoma: FNAB. Thick cords of malignant hepatocytes are wrapped by peripheral endothelium. They appear to be floating on transverse section view (Papanicolaou).
• Rows of transgressing endothelium in larger aggregates; basement membrane material ("sinusoidal capillarization") best seen in Giemsa preparations.
• Cohesion is the rule; however, tendency to dissociation noted in highly WD-HCC and PD-HCC.
• Pseudoacini containing bile or pale secretions are not uncommon. The spaces are surrounded by polygonal cells with central nuclei similar to adjacent cells (Figure 7).
Figure 7 Moderately differentiated hepatocellular carcinoma with pseudoacinar pattern: FNAB. (A) Pseudoacini filled with bile which appears as dark brown blobs (Papanicolaou) (B) Bile appears black (May-Grunwald-Giemsa). (C) Corresponding histologic section of the tumor shows cystically dilated canaliculi filled with golden-brown bile and surrounded by polygonal cells with central nuclei (H&E).
• Hepatocytic characteristics include polygonal cells with well-defined borders, ample granular cytoplasm, central round nucleus, well-delineated nuclear membrane, prominent nucleolus and fine, irregularly granular chromatin. Mitoses increase with nuclear grade. Cytologic features of malignancy are wanting at the highly WD-HCC end whereas resemblance to hepatocyes is lacking at the PD-HCC end.
• Tumor cells may be smaller, larger or of the same size as normal hepatocytes. WD-HCC cells tend to be conspicuous by their small size, monotony, subtle increase in N/C ratio and nuclear crowding. PD-HCC cells tend to be pleomorphic (Figure 8).
Figure 8 Poorly differentiated hepatocellular carcinoma: FNAB. High-grade tumor shows marked pleomorphism but still retaining some hepatocytic characteristics (Papanicolaou).
• Atypical naked hepatocytic nuclei may abound (Figure 9).
Figure 9 Poorly differentiated hepatocellular carcinoma: FNAB. Atypical naked hepatocytic nuclei exhibit pleomorphism, thin nuclear membrane, nuclear contour irregularities, prominent nucleoli and multinucleation (Papanicolaou).
• Multinucleated tumor giant cells may be of "osteoclastic" or pleomorphic type. The former shows nuclear features akin to adjacent HCC cells. Tumor giant cells may be found in all grades of HCC. Their presence does not necessarily upgrade the tumor.
• Bile may be present within tumor cells or in canaliculi or pseudoacini.
• Intracytoplasmic fat and glycogen vacuoles are common. Intracytoplasmic inclusions include hyaline, pale and Mallory bodies. Intranuclear cytoplasmic inclusions are not specific.
• Bile duct epithelial cells, if present, are few and far apart. Kupffer cells may be seen.
Cell blocks/microbiopsies
• Microhistology provides additional invaluable architectural details such as trabecular-sinusoidal pattern formed by broad trabeculae (>2 cells thick); pseudoacini; unpaired arteries; and deficient/virtually absent reticulin framework.
FNAB of well-differentiated hepatocellular nodular lesions
The accuracy of cytodiagnosis at this end of the spectrum is often an issue with indeterminate reports being rendered [14,32]. The diagnostic dilemmas are: (i) Is the sample representative? (ii) Are the hepatocytes malignant (highly WD-HCC) or benign? (iii) If benign, are they neoplastic (LCA) or nonneoplastic hepatocytes? (iv) If nonneoplastic, are they intralesional (FNH, DN or MRN) or extralesional hepatocytes (cirrhosis or normal liver), with/without fatty change? Cytologic features predictive of HCC include increased N/C ratio, cellular monomorphism, nuclear crowding, trabeculae >2 cells thick, atypical naked hepatocytic nuclei and lack of bile duct cells [23,33]. Cytologic parameters distinguishing highly WD-HCC from cirrhosis include well-defined cytoplasmic borders, scant cytoplasm, monotonous cytoplasm, thick cytoplasm, eccentric nuclei and increased N/C ratio [34].
FNAB of variants of hepatocellular carcinoma
HCC is well known for its histologic variations and subtypes. A large tumor can harbor areas that are more easily recognizable than others. This has significant practical implications on the number of passes and sampling in hepatic FNAB.
The variants of HCC include:
• HCC with fatty change: Fatty change can occur in HCC without associated steatosis. Small (early) lesions are prone to fatty change due to inadequate vascularization. WD-HCC cells with cytoplasmic fat vacuoles can be mistaken for hepatocytes from fatty liver or focal fatty change (Figure 10) [35,36]. PD-HCC cells with fat vacuoles may mimic malignant lipoblasts or signet-ring adenocarcinoma cells.
Figure 10 Fatty change: FNAB. Hepatocytes exhibit polymorphism and contain cytoplasmic fat vacuoles of varying sizes pushing nucleus to one side. Cytologic picture can mimic signet-ring cell adenocarcinoma (Papanicolaou).
• HCC, clear cell type: Clear cell change in hepatocytes is due to abundant cytoplasmic glycogen or lipid content [37]. Tumor cells with empty-looking vacuoles after removal of glycogen during processing may mimic fatty change. Focal clear cell change is frequent. Diffuse clear cell change occurs in <10% of cases of HCC [38].
• HCC, small cell type: This type is reminiscent of neuroendocrine tumors with tendency to dissociation and microacinar formation but no obvious trabecular pattern [39,40]. Scrutiny of the nuclear/chromatin features should reveal the true histogenesis of the small round cells.
• HCC, undifferentiated type: The tumor cells are larger than the small cell category with no obvious hepatocytic characteristics.
• HCC, spindle cell type: This is rare and is more likely to be seen with tumor giant cells as part of a larger tumor [41].
• HCC, giant cell type: This pure variant is rare.
• Fibrolamellar HCC: This occurs in non-cirrhotic livers of young patients and has a good prognosis. It comprises large, discohesive polygonal hepatocytes with abundant oncocytic cytoplasm and lamellar fibrosis. Pale bodies are common [42,43].
• HCC with biliary differentiation: Some HCC are positive for biliary markers (AE1/3, CK19) [44]. The stem cell theory with bipotential progenitor cells capable of developing into either hepatocytes or biliary epithelial cells provides a satisfactory explanation for primary liver cancers arising from different stages of the cell lineage [45].
• Combined hepatocellular-cholangiocarcinoma (CHCC-CC): This is a rare tumor containing unequivocal elements of HCC and CC that are intimately admixed with a transitional component [46,47]. The HCC cells are expected to be AFP and Hep Par 1-positive and show polyclonal CEA (pCEA) canalicular staining. The CC cells are AE1/3-positive and show brush border/diffuse cytoplasmic pCEA reactivity. The intermediate cells exhibit hybrid features with equivocal immunoprofiles.
FNAB of cholangiocarcinoma and its variants
Intrahepatic CC are rare and usually occur in non-cirrhotic livers in close proximity with sizeable bile ducts. Predisposing diseases include clonorchiasis, opisthorchiasis, hepatolithiasis and primary sclerosing cholangitis. CC are usually well to moderately differentiated adenocarcinomas with variable degree of desmoplasia. They can be categorized into papillary and/or tubular adenocarcinomas. Mucin production is minimal. A squamous component may be present. Adjacent biliary epithelial changes include carcinoma-in-situ and biliary ductular proliferation. There are three cytologic smear patterns, namely (i) Scanty cell smear pattern, (ii) Adenocarcinoma with proliferating ductular clusters, and (iii) Adenocarcinoma without prominent ductular clusters [20].
The variants of CC include:
• Biliary papillary neoplasia / intraductal papillary CC [48]
• Mucinous intrahepatic CC [20]
FNAB of non-hepatocellular carcinoma malignancies
That the liver is a common target for metastases makes the separation between primary and secondary malignancies all the more difficult, especially when the particular histologic subtype can arise in the liver as well. Categorizing them based on cytologic patterns is a good start.
• Adenocarcinoma: Most are metastases from stomach, colorectum, pancreas, breast and lungs. Colorectal metastases have much tumor diathesis. Signet-ring cell adenocarcinomas are likely to be gastric in origin. Pancreaticobiliary tract adenocarcinomas can have squamous components. For any adenocarcinoma in hepatic aspirates, CC, HCC with pseudoacini and CHCC-CC have to be considered.
• Squamous cell carcinoma: Most are metastatic or arise in the pancreaticobiliary tract. Large, spindly, "tadpole-shaped" or bizarre cells with dense cytoplasm, keratinization and much necrosis may be seen. Adenosquamous variants are not uncommon.
• Small/intermediate round cell malignancy: This includes neuroendocrine tumors (NET), small cell undifferentiated carcinomas (SCUC), undifferentiated nasopharnygeal carcinomas and lymphomas [40,49]. Other possibilities include melanoma, Merkel cell tumor, metastatic adenocarcinomas from prostate, stomach and breast (lobular carcinoma), CC, certain sarcomas and HCC, small cell type. Most NET are from the gastrointestinal tract (GIT), pancreaticobiliary tract or lung; primary hepatic NET is unusual [50].SCUC usually originates in the lung. Hepatic metastases from nasopharnygeal carcinomas tend to be markedly necrotic mimicking abscesses radiologically. Lymphoma seldom presents as a primary neoplasm although hepatic involvement is common in advanced disease [51]. It has been reported in hepatitis B virus infection, systemic lupus erythematosus, primary biliary cirrhosis and acquired immunodeficiency syndrome. A high index of suspicion is necessary. It can be mistaken for poorly differentiated carcinoma or HCC. Inflammatory processes, such as inflammatory pseudotumors [52-54], and sinusoidal hematopoietic cells have first to be excluded. Nodular extramedullary hematopoiesis can rarely mimic metastases.
• Clear cell malignancy: This can also arise in the kidney, adrenal and ovary [55]. Renal cell carcinoma contains transgressing endothelium and papillary endothelium in fibrovascular cores but lacks peripheral endothelium. Nuclear features also differ. However, renal cell carcinoma has not been called the great mimic for no good reason. Metastases can occur years later. Other pitfalls include hepatocytes with fatty change and metastases from liposarcoma and signet-ring cell adenocarcinoma.
• Pleomorphic cell malignancy: This includes large cell undifferentiated carcinomas (LCUC), large cell lymphomas, germ cell tumors and various sarcomas. LCUC is not a pure or single entity; tumors may show glandular and neuroendocrine differentiation. Common sites are lung, GIT and female genital tract.
• Spindle cell malignancy: Well-differentiated spindle cell tumors include leiomyosarcoma (LS), neurogenic tumors and fibroblastic/stromal tumors, including gastrointestinal stromal tumor (GIST) [56]. At the poorly differentiated end, LS, malignant fibrous histiocytoma, undifferentiated sarcoma or even sarcomatoid HCC or CC with a spindle cell component, have to be considered. Apart from distinguishing LS from leiomyoma, other differential diagnoses include benign reactive processes [57]. The epithelioid variant of LS occurs primarily in the GIT. Hepatic metastases can be mistaken for epithelial tumors, such as, HCC, CC, metastatic carcinoma and melanoma. GIST also has spindle and epithelioid cell types [58]. Metastatic GIST may pose diagnostic problems due to their variable morphologic spectrum and cytologic atypia; c-kit staining is necessary. Hepatic angiosarcoma associated with Thorotrast and epithelioid hemangioendothelioma are rare. In the young, differential diagnoses of spindle cell lesions include inflammatory pseudotumor, infantile hemangioendothelioma, mesenchymal hamartoma and undifferentiated (embryonal) sarcoma.
• Giant cell malignancy: This can be carcinomatous or sarcomatous. Giant cell carcinomas have been noted in lung, pancreas, thyroid, kidney and breast. Hepatic metastases have to be distinguished from HCC, giant cell type.
• Hepatoid carcinoma; AFP-producing carcinoma: Primary hepatic tumors are not the only source of AFP. Neither is the liver the only site of origin for hepatocytic-looking carcinomas. Most of them arise in the lung and GIT. Those with hepatoid features mimic classic HCC in having a proclivity for vascular permeation and distant metastases, in this case to the liver. Others produce AFP but are non-hepatoid; being usually adenocarcinoma, undifferentiated carcinoma and small or large cell neuroendocrine carcinoma [59].
Suffice it to say that when dealing with FNAB of liver masses, one must be fully aware of any past history of malignancy or rule out any hitherto undetected cancer. Some of the limitations in the categorization of tumors obtained by FNAB can be overcome by immunohistochemistry. The advantages of an exact cytodiagnosis are obvious – it may save the patient a diagnostic laparotomy, especially in inoperable cases, and allow for specific chemotherapy to be instituted without delay. However, at best, information gleaned from a precise cytodiagnosis can sometimes only favor a particular primary site.
Step 4: Further confirm nature of cytohistologic findings
The initial cytologic assessment is crucial as it forms the basis upon which ancillary tests are ordered; the results of which should be interpreted in the larger context of the case.
Special stains
• Reticulin
Study of the reticulin framework, stained by Gomori's method, is important in the analysis of well-differentiated hepatocellular nodules [60,61]. HCC have deficient or absent reticulin; the reticulin framework in LCA and FNH may not be well developed either.
• Periodic acid-Schiff with and without diastase; Mucicarmine
This is usually performed to distinguish glycogen from epithelial mucin. Hepatocytes are loaded with glycogen. Cells exhibiting glandular (biliary) differentiation may show intracytoplasmic or intraluminal mucin production.
• Fat
Fat can be demonstrated in nonneoplastic and neoplastic hepatocytes by staining fresh or formalin-fixed tissue with Oil Red O. It has no discriminant value in defining the biological status of hepatocytes but may be of help in deciphering the contents of cytoplasmic vacuoles of cells of unrecognizable histogenesis.
• Iron
Iron appears as black cytoplasmic granules in hepatocytes stained with Giemsa stains and dark brown with Papanicolaou stains. The pigment can be confirmed with Perl's prussian blue method. Iron accumulation in hepatocytes favors a benign process. On the other hand, in populations where hepatic iron accumulation is common, the absence of iron in hepatocytes should alert one to the presence of a proliferative process, be it regenerative or neoplastic.
Immunohistochemistry
A whole battery of antibodies is available for the comparative immunohistochemical study of primary and metastatic liver tumors. The two major diagnostic issues are (i) whether the hepatocytes are malignant or benign, and (ii) what the histogenesis of the malignant cells is [See "Diagnostic utility of immunohistochemistry"].
Step 5: Establish final diagnosis based on multidisciplinary approach
Close clinicopathologic correlation is mandatory for enhancing the yield of FNAB diagnoses and the reduction of indeterminate reports.
Current diagnostic issues
Reappraisal of role of hepatic fine needle aspiration biopsy
1. It is still to procure a tissue diagnosis as part of the evaluation of focal liver lesions but under circumstances where the clinical, biochemical and imaging profiles are not conclusive
A benign cytodiagnosis obviates unnecessary surgery. Surgical resection is indicated for any resectable malignant hepatic mass, be it primary or secondary. In unresectable malignant lesions, a precise cytohistologic typing is crucial for appropriate alternative therapy. The need for biopsy diagnosis in HCC is now a hotly debated topic [2,4,12]. The stand in some practices is that a needle biopsy may be indicated only if it is not possible to diagnose HCC by other means, namely, serum AFP concentration, spiral computed tomography (CT) and magnetic resonance imaging (MRI). In advanced HCC, the need for biopsy may be obviated due to the non-availability of effective therapy. In early/small HCC, even if a biopsy is performed, the diagnostic sensitivity of the procedure may be as low as 60% [3]. If there is a high index of suspicion for HCC despite negative cytologic sampling, specific therapy may be instituted.
2. It is a safe and well-tolerated, minimally invasive procedure with low risk of complications in suitable candidates and in skilled hands
Needle track seeding by malignant cells is the main reason often cited by opponents of FNAB of the liver [2,62-64]. There is no reliable data to establish the risk; the figure of 0.006% is regarded as a gross underestimation by many authors [3,65,66]. Needle biopsies, in general, are usually not indicated in patients deemed suitable for liver transplantation due to possible seeding [5]. The move to institute definitive treatment for classic HCC diagnosed solely on clinical, biochemical and radiologic grounds without tissue confirmation has crept slowly into some practices. The risk of false positive diagnosis of HCC with subsequent aggressive therapy, such as liver transplantation, by far outweighs the risk of seeding.
3. It is a technique with high sensitivity and specificity when practised in a multidisciplinary setting by skilled operators
Tissue procurement by FNAB under radiologic guidance and cytologic interpretation of the aspirated material are both highly operator-dependent. An experienced screener on-site can give a rapid assessment of adequacy.
4. Surveillance for early HCC in high-risk patients has resulted in the detection of "suspicious" nodules in cirrhotic livers that are often <2 cm in diameter
The usual surveillance tools are serum AFP concentration and ultrasonography (US). AFP is not a very good screening test since it has a sensitivity of 39–64%, a specificity of 76–91% and a positive predictive value of between 9–32% [6,67]. On the other hand, US has, as a screening test in hepatitis B surface antigen (HBsAg) carriers, a sensitivity of 71% and specificity of 93%, but its positive predictive value is only 14% [6]. Although a spectrum of hepatocellular nodular lesions can be encountered, especially in a background of cirrhosis, studies have revealed that about half of them are not HCC. Tissue characterization is, therefore, mandatory and FNAB may be the simplest and most practical means to reach small, deep-seated lesions [1].
The European Association for the Study of the Liver (Barcelona-2000 EASL Conference) [5]has drawn up guidelines for the current clinical management of HCC. HCC can be diagnosed in a setting of liver cirrhosis if a focal liver lesion is >2 cm in diameter with arterial hypervascularization, and shows coincident features in at least two imaging techniques (US, spiral CT, MRI and angiography); or characteristic features in one imaging technique associated with serum AFP level of >400 ng/ml. It follows then that there is no necessity to obtain cytologic and/or histologic confirmation for all cases of suspected HCC. It is recommended that punctures be limited to patients with nodules <2 cm diameter for the differentiation of HCC from regenerative nodules. In the future, tumor biopsy may assume a different role by becoming a useful tool for procuring tissue for the molecular profiling of the disease. It is important to note that the above recommendations are not meant for everyone. It should be carried out in highly selected patients by highly skilled healthcare teams in specialized tertiary centers.
Fine needle aspiration biopsy versus core needle biopsy
Fine needle aspiration biopsy is useful for (i) cirrhotic patients with poor liver function with risk of bleeding; (ii) liver masses with obstructive jaundice and risk of bile leakage, those near big vessels, or where there is need to go through bowel; (iii) small (<2 cm diameter), deep-seated and difficult to approach nodules that require close patient co-operation during the procedure; (iv) representative sampling of sizeable lesions by re-direction of the needle and multiple passes; and (v) on-site rapid assessment of adequacy and rendering of provisional diagnosis, as well as for appropriate triage of tissue specimens for ancillary studies (e.g. microbiology, flow cytometry, genetic testing, molecular diagnostics, cell block preparation and electron microscopy) [22].
Core needle biopsy (CNB), with the availability of more material, provides tissue for histologic and immunohistochemical studies, especially in two major areas of diagnostic difficulties, namely, in the (i) differentiation of WD-HCC from benign hepatocellular nodules; and (ii) separation of HCC from CC and metastases. A critical comparative evaluation of the risks of seeding by CNB and FNAB is needed. For the proponents of CNB, the view is that a single pass with larger bore needles (<20 gauge) may be preferable to multiple passes by finer needles needed to obtain sufficient material for cytohistologic examination. The incidence of seeding may increase as a result of the number of punctures of tumors detected at an early stage in patients with a longer life expectancy [66].
Consensus: The diagnostic accuracy in terms of sensitivity, specificity and positive predictive value of FNAB for HCC is almost similar to that of CNB. The accuracy rate is highly operator-dependent and increases with both techniques combined. The specificity and positive predictive value of FNAB in the diagnosis of malignant hepatic lesions has been shown to be close to 100% in most studies [19,22,68,69]. These results are comparable to the accuracy of CNB. In a comparative study, it was reported that both procedures, FNAB and CNB, had the same diagnostic accuracy of 78% when considered separately and of 88% when considered in combination [21]. The conclusion was that the great advantage of combining the two techniques was the reduction in false negative results. Performing both procedures at the same sitting may not be feasible due to medical contraindications and may also not be acceptable clinical practice. Using larger caliber cutting needle biopsies can be associated with a greater number of complications [68]. This quagmire can be overcome by using cutting FNAB needles. At our institution, the utilization of such (21 gauge) aspiration needles has provided us with ample material for smear preparations as well as tissue cores resembling microbiopsies rather than mere cell blocks [20,34]. Our practice has been to retrieve sizeable particulate matter from the material on the glass slides prior to smearing and fixing them in formalin for cell block preparation. Many studies have attested to the improved diagnostic yield and accuracy of FNAB using the combined cytohistologic approach [23,70]
FNAB can provide rapid on-site diagnosis when the smears are stained with Diff-Quik or Ultra-fast Papanicolaou stain [28]. In the era of rising costs in medical practice and higher patient/practitioner/institution expectations of efficiency and faster turn-around time, FNAB can obviate the need to wait for tissue processing if accurate cytologic diagnoses can be rendered. Another cost-saving advantage, especially for less developed countries, is that smears are cheap, convenient and easy to prepare as long as there is an experienced person to interpret them.
Considering the overall advantages and cost-analysis, FNAB can be suggested as the initial method of choice for evaluation of focal liver lesions in most clinical settings. The final choice should be decided on the basis of the working clinical diagnosis and the institutional/personal experience.
Separation of well-differentiated hepatocellular carcinoma from benign hepatocellular nodular lesions
Factors that pose diagnostic problems and pitfalls
1. HCC can be small and focal, solitary and large, multifocal, or diffuse and infiltrating; thereby, mimicking small benign lesions on the one hand and metastases on the other, especially in imaging studies.
2. Serum AFP, though fairly specific, has poor sensitivity for the diagnosis of HCC, regardless of tumor size or degree of differentiation. There is significant elevation in about 50–60% of HCC [24,33]. Small WD-HCC are usually not associated with serum AFP elevation. On the other hand, transient increases may be seen with inflammatory flares in chronic viral hepatitis. Published data at the current moment suggest using values of >400 ng/ml for diagnostic confirmation of HCC [67].
3. The diagnostic dilemma at the highly WD-HCC end is that the hepatocytic histogenesis is obvious but proof of malignancy may be lacking [24,32,33]. The cell cords tend not to be >2 cells thick and the cellular pleomorphism and subtle increase in the N/C ratios of the hepatocytes may not be appreciated under light microscopy. In fact, highly WD-HCC tend to be composed of small, uniform, strikingly monotonous neoplastic hepatocytes with slightly increased N/C ratios, imparting an impression of nuclear crowding [34]. The recognition of polymorphism with variation in cell and nuclear sizes and a normal N/C ratio of 1/3 should alert one to the likelihood of benignity of the hepatocytes [20].
4. Routine surveillance by more than one imaging technique in high-risk patients tends to detect nodule/s of various natures in the cirrhotic livers [10,71]. HCC tend to occur in a cirrhotic background together with MRN and low- and high-grade DN. Pure light microscopic interpretation with immunohistochemical input may no longer suffice in the diagnostic workup towards an accurate tissue diagnosis.
5. Smaller and smaller lesions (<2 cm diameter) are being increasingly detected by imaging and subjected to FNAB for tissue diagnosis.
6. Early HCC tend to be small and highly well-differentiated, making differentiation from MRN and DN difficult. Increasing numbers of equivocal nodular lesions are being detected in cirrhotic livers by various imaging modalities. The histopathologic interpretation of these nodules is highly controversial, let alone cytologic assessment. Eastern pathologists tend to call them early HCC while the Western fraternity tends to go for a diagnosis of high-grade DN [14,17].
7. Small HCC are prone to fatty change due to inadequate neoangiogenesis; poor vascularity makes them difficult to visualize and characterize by imaging methods [8,14]. Increasing degree of heterogeneity is exhibited as the lesion enlarges [15].
8. Well-differentiated hepatocellular nodules of any nature can occur in non-cirrhotic livers and have to be distinguished from each other and from the surrounding liver.
9. Fatty change can occur in WD-HCC, LCA, FNH and DN without associated steatosis in the surrounding liver parenchyma. An entity called focal fatty alteration/change can also mimic these nodules [20,36].
10. Current histologic criteria for the diagnosis of highly WD-HCC include loss of reticulin, thickening of trabeculae (>2 cells thick), sinusoidal capillarization, unpaired arterioles, cellular pleomorphism/heterogeneity, "clonal" growth patterns, pseudoacini, mitotic activity [14,16] and stromal invasion of portal tracts. Some of these criteria are lacking in cytologic material.
11. In high-risk cases, some "non-malignant" nodules with large cell change (low-grade DN) or small cell change (high-grade DN) may be precursor lesions [11,18,72-74]. Morphometric studies suggest that small cell change may be the more sinister lesion biologically. HCC occurs in 5–40% of cirrhotic patients and foci of cancer are found in a third of DN [75]. Some of these precursor lesions may be indistinguishable from malignant hepatocellular nodules by light microscopy alone.
12. Current concepts on how premalignant lesions develop and how HCC may arise within them impact on the accuracy of pathologic diagnosis of hepatocellular nodules [7,9,10]. Inadequacies of FNAB evaluation of hepatocellular nodules cannot be ignored given the focality of proliferative clones of atypical cells within the nodules and the shortcomings of pure morphologic interpretation.
Diagnostic utility of immunohistochemistry
There are two major applications for immunohistochemical markers in the diagnostic workup of focal liver lesions [20,76]. One is to decipher the exact histogenetic origin of obvious tumor nodules – that is, the histologic typing and the primary site [77]. It may not always be possible to distinguish between the poorly differentiated entities of HCC, CC and metastatic carcinomas [25,78]. By the same token, adenocarcinomas occurring in the liver may be metastatic or primary in origin. Of interest lately is the increasing documentation of AFP-producing extrahepatic hepatoid/non-hepatoid carcinomas that have a propensity for vascular invasion and liver metastases [59,79]. The immunoprofile of these tumors, originating mostly in the GIT and lungs, is almost identical to that of HCC. Serum AFP levels tend to be very high. The other application concerns the distinction of the various lesions within the realm of well-differentiated hepatocellular nodular lesions. [See "FNAB of well-differentiated hepatocellular nodular lesions"]. For ascertainment of malignancy in hepatocellular nodules, the antibody panel should comprise at least AFP, pCEA or CD10, and CD34 [80-82]. Additional markers, such as Hep Par 1 and cytokeratins, should be included if the histogenesis of the tumor is to be studied. Markers of cell proliferation, proliferating cell nuclear antigen (PCNA) and Ki67; and p53 are not routinely used.
• AFP is fairly specific but not sensitive for HCC. Tissue AFP immunoreactivity is expected in HCC (Figure 11) but it may be patchy and minimal. Sensitivity is about 50% (range, 20–75%) and is low at both ends of the histologic spectrum of HCC [44,83-86]. A study of 56 patients with small HCC (<2 cm diameter) showed AFP-positivity in 44.6% of the tumors [87]. A variable staining pattern may be encountered with CHCC-CC. MRN in cirrhosis are not associated with elevated serum or stainable tissue AFP [88]; neither are DN or LCA. The specificity of AFP for HCC is being challenged by reports of AFP-producing extrahepatic hepatoid/non-hepatoid carcinomas [59]. There is an increasing tendency to drop this immunomarker from the panel for HCC workup.
Figure 11 Hepatocellular carcinoma: FNAB. The tumor cells stain positively for alpha-fetoprotein (Immunostain).
• pCEA stains bile canaliculi and ductal epithelium but not hepatocytes. A characteristic "chicken-wire fence" canalicular staining pattern is seen in normal liver [89]. Biliary epithelial cells show diffuse cytoplasmic and brush border staining. A normal canalicular pattern is expected for benign hepatocellular nodular lesions. However, some LCA and even, some FNH, may exhibit deficient canaliculi. There are two patterns of staining in HCC – canalicular and/or diffuse cytoplasmic staining. The canalicular pattern is abnormal and deficient in HCC with twisted and often dilated structures; it may not even be appreciable in high-grade HCC [78]. Instead, diffuse cytoplasmic staining may be seen in less differentiated HCC, making distinction from a poorly differentiated adenocarcinoma difficult. In the context of a carcinoma, a canalicular pattern is specific for HCC.
• CD10 is expressed in normal and neoplastic liver, exhibiting a similar canalicular staining pattern to pCEA [90,91]. Although it does not differentiate between benign and malignant hepatocellular nodular lesions, CD10 is very useful in distinguishing HCC from non-HCC malignancies. The sensitivity of CD10 (68.3%) is far better than immuno-staining for AFP (23.8%) but less sensitive than pCEA (95.2%) in the diagnosis of HCC [92].
• Hep Par 1 (Hepatocyte antigen) is a sensitive marker for hepatocytic differentiation and is part of the antibody panel for distinguishing HCC from CC and metastases. However, not all HCC stain uniformly (Figure 12) and not all Hep Par 1-positive tumors are of hepatocellular origin or arise in the liver [86,93-96]. Its variable and heterogeneous staining pattern, which can range from 100% -positive cells in WD-HCC to <5% in some PD-HCC cases, may lead to false negative results in small samples [97]. MRN, DN, FNH and LCA tend to exhibit 100% positivity. Hence, this antibody has no discriminant value in the evaluation of the biological status of well-differentiated hepatocellular nodular lesions.
Figure 12 Hepatocellular carcinoma: Histology. Parts of the tumor show intense granular cytoplasmic reactivity with HepPar1 (Immunostain).
• Cytokeratins (CK 7, 8, 18, 19, 20; CAM 5.2; AE1/AE3). Mature hepatocytes stain with CK 8 and 18 and CAM 5.2 but not with CK 7, 19 or 20 or AE1/AE3. CAM 5.2 is the most reliable cytokeratin antibody for HCC. AE1/AE3 negativity is expected in hepatocellular lesions. Focal CK 7 and 19 positivity can be seen in high-grade HCC. HCC is generally CK 20 negative [98].
• CD34 is an intercellular adhesion protein found in normal endothelium but absent in normal sinusoids. CD34 highlights regions of sinusoidal capillarization where there is basement membrane material deposition. Diffuse sinusoidal CD34 reactivity is seen in HCC, even small WD-HCC [99]. However, significant reactivity is also seen in LCA and some FNH. In cirrhotic and DN, staining is absent or minimal, and confined to the periportal/periseptal regions [100].
Conclusion
Tissue confirmation is recommended in the diagnosis of focal liver lesions as the risk of aggressive therapy is greater than the risk of malignant seeding. FNAB has many advantages that CNB lacks. The lack of tissue can be overcome by using cutting aspiration needles that can provide material for smear cytology and microhistology. Guided FNAB is a highly operator-dependent procedure as is the preparation and interpretation of the cytologic material. Optimization of FNAB in the diagnosis of focal liver lesions, increase in the yield of true positive diagnoses, and rendering of fewer indeterminate reports require close clinicopathologic correlation; combination of smear cytology and microhistology, use of a discriminant panel of special and immunostains; and team work by skilled operators on all fronts.
List of abbreviations
FNAB : Fine needle aspiration biopsy
HCC : Hepatocellular carcinoma
MRN : Macroregenerative nodule
DN : Dysplastic nodule
FNH : Focal nodular hyperplasia
LCA : Liver cell adenoma
WD-HCC : Well-differentiated hepatocellular carcinoma
PD-HCC : Poorly differentiated hepatocellular carcinoma
CC : Cholangiocarcinoma
AFP : Alpha-fetoprotein
CEA : Carcinoembyronic antigen
N/C ratio : nuclear-cytoplasmic ratio
CHCC-CC : Combined hepatocellular-cholangiocarcinoma
pCEA : polyclonal carcinoembyronic antigen
NET : Neuroendocrine tumor
SCUC : Small cell undifferentiated carcinoma
GIT : Gastrointestinal tract
LCUC : Large cell undifferentiated carcinoma
LS : Leiomyosarcoma
GIST : Gastrointestinal stromal tumor
CT : Computed tomography
MRI : Magnetic resonance imaging
US : Ultrasonography
CNB : Core needle biopsy
PCNA : Proliferating cell nuclear antigen
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
This is solely my work.
Acknowledgements
The author would like to thank T. C. Tan for photographic assistance.
"Co-editors of Cytojournal, Vinod B. Shidham, MD, MRCPath, FIAC and Barbara F. Atkinson, MD thank:
The academic editor CAROLYN ERNST GROTKOWSKI, M.D., QUEST DIAGNOSTICS INCORPORATED, Mount Laural, NJ 08054, USA ([email protected]) for completing the peer-review process for this manuscript."
==== Refs
Caturelli E Ghittoni G Roselli P De Palo M Anti M Fine needle biopsy of focal liver lesions: The hepatologist's point of view Liver Transpl 2004 10 S26 S29 14762835 10.1002/lt.20037
Cetta F Zuckermann M De Nisi A Is needle biopsy of the liver necessary to diagnose HCC? Gut 2001 48 440 (Letter to editor) 11256362 10.1136/gut.48.3.440
Durand F Regimbeau JM Belghiti J Sauvanet A Vilgrain V Terris B Moutardier V Farges O Valla D Assessment of the benefits and risks of percutaneous biopsy before surgical resection of hepatocellular carcinoma J Hepatol 2001 35 254 258 11580148 10.1016/S0168-8278(01)00108-8
Llatjos M Muns R Tallada N Need for biopsy in hepatocellular carcinoma J Hepatol 2001 35 874 (Letter to editor).
Bruix J Sherman M Llovet JM Beaugrand M Lencioni R Burroughs AK Christensen E Pagliaro L Colombo M Rodes J Clinical management of hepatocellular carcinoma. Conclusions of the Barcelona – 2000 EASL Conference J Hepatol 2001 35 421 430 11592607 10.1016/S0168-8278(01)00130-1
Collier J Sherman M Screening for hepatocellular carcinoma Hepatology 1998 27 273 278 9425947 10.1002/hep.510270140
Borzio M Borzio F Croce A Sala M Salmi A Leandro G Bruno S Roncalli M Ultrasonography-detected macroregenerative nodules in cirrhosis: a prospective study Gastroenterology 1997 112 1617 1623 9136841
Kutami R Nakashima Y Nakashima O Shiota K Kojiro M Pathomorphologic study on the mechanism of fatty change in small hepatocellular carcinoma of humans J Hepatol 2000 33 282 289 10952246 10.1016/S0168-8278(00)80369-4
Terasaki S Kaneko S Kobayashi K Nonomura A Nakanuma Y Histological features predicting malignant transformation of nonmalignant hepatocellular nodules: a prospective study Gastroenterology 1998 115 1216 1222 9797377
Theise ND Macroregenerative (dysplastic) nodules and hepatocarcinogenesis: Theoretical and clinical considerations Semin Liver Dis 1995 15 360 371 8578320
International Working Party Terminology of nodular hepatocellular lesions Hepatology 1995 22 983 993 7657307 10.1016/0270-9139(95)90324-0
Torzilli G Minagawa M Takayama T Inoue K Hui AM Kubota K Ohtomo K Nakuuchi M Accurate preoperative evaluation of liver mass lesions without fine-needle biopsy Hepatology 1999 30 889 893 10498639 10.1002/hep.510300411
Kimura H Nakajima T Kagawa K Deguchi T Kakusui M Katagishi T Okanoue T Kashima K Ashihara T Angiogenesis in hepatocellular carcinoma as evaluated by CD 34 immunohistochemistry Liver 1998 18 14 19 9548262
Kojiro M Focus on dysplastic nodules and early hepatocellular carcinoma: An Eastern point of view Liver Transpl 2004 10 S3 S8 14762831 10.1002/lt.20042
Nakashima Y Nakashima O Hsia CC Kojiro M Tabor E Vascularization of small hepatocellular carcinomas: Correlation with differentiation Liver 1999 19 12 18 9928760
Quaglia A Bhattacharjya S Dhillon AP Limitations of the histopathological diagnosis and prognostic assessment of hepatocellular carcinoma Histopathology 2001 38 167 174 11207830 10.1046/j.1365-2559.2001.01000.x
Roncalli M Hepatocellular nodules in cirrhosis: Focus on diagnostic criteria on liver biopsy. A Western experience Liver Transpl 2004 10 S9 S15 14762832 10.1002/lt.20047
Ishak KG Goodman ZD Stocker JT Tumors of the liver and intrahepatic bile ducts Atlas of Tumor Pathology 3rd series Fascicle 31 2001 Armed Forces Institute of Pathology, Washington, DC
Hertz G Reddy VB Green L Spitz D Massarani-Wafai R Selvaggi SM Kluskens L Gattuso P Fine-needle aspiration biopsy of the liver: A multicenter study of 602 radiologically guided FNA Diagn Cytopathol 2000 23 326 328 11074627 10.1002/1097-0339(200011)23:5<326::AID-DC8>3.0.CO;2-1
Wee A Sampatanukul P Fine needle aspiration cytology of the liver. Diagnostic algorithms. A Southeast Asian perspective 2004 Bangkok: Year Book Publisher Co Ltd
Franca AVC Valerio HMG Trevisan M Escanhoela C Seva-Pereira T Zucoloto S Martinelli A Soares EC Fine needle aspiration biopsy for improving the diagnostic accuracy of cut needle biopsy of focal liver lesions Acta Cytol 2003 47 332 336 12789910
Jain D Diagnosis of hepatocellular carcinoma. Fine needle aspiration cytology or needle core biopsy J Clin Gastroenterol 2002 35 S101 S108 12394213 10.1097/00004836-200211002-00006
Longchampt E Patriache C Fabre M Accuracy of cytology vs. microbiopsy for the diagnosis of well-differentiated hepatocellular carcinoma and macroregenerative nodule. Definition of standardized criteria from a study of 100 cases Acta Cytol 2000 44 515 523 10934942
Wee A Nilsson B Tan LKA Yap I Fine needle aspiration biopsy of hepatocellular carcinoma: Diagnostic dilemma at the ends of the spectrum Acta Cytol 1994 38 347 354 7514830
Das DK Cytodiagnosis of hepatocellular carcinoma in fine-needle aspirates of the liver: Its differentiation from reactive hepatocytes and metastatic adenocarcinoma Diagn Cytopathol 1999 21 370 377 10572267 10.1002/(SICI)1097-0339(199912)21:6<370::AID-DC2>3.0.CO;2-K
Ferrell L Malignant liver tumors that mimic benign lesions: Analysis of five distinct lesions Semin Diagn Pathol 1995 12 64 76 7770675
Pisharodi LR Lavoie R Bedrossian CWM Differential diagnostic dilemmas in malignant fine-needle aspirates of liver: A practical approach to final diagnosis Diagn Cytopathol 1995 12 364 371 7544724
Yang GC Yang GY Tao LC Cytologic features and histologic correlations of microacinar and microtrabecular types of well-differentiated hepatocellular carcinoma in fine-needle aspiration biopsy Cancer (Cancer Cytopathol) 2004 102 27 33
Sole M Calvet X Cuberes T Maderuelo F Bruix J Bru C Rey MJ Serna N Cardesa A Value and limitations of cytologic criteria for the diagnosis of hepatocellular carcinoma by fine needle aspiration biopsy Acta Cytol 1993 37 309 316 8388606
Soyuer I Ekinci C Kaya M Genc Y Bahar K Diagnosis of hepatocellular carcinoma by fine needle aspiration cytology: Cellular features Acta Cytol 2003 47 581 589 12920750
Wee A Nilsson B Chan-Wilde C Yap I Fine needle aspiration biopsy of hepatocellular carcinoma: Some unusual features Acta Cytol 1991 35 661 670 1659093
de Boer WB Segal A Frost FA Sterrett GF Cytodiagnosis of well differentiated hepatocellular carcinoma: Can indeterminate diagnoses be reduced? Cancer 1999 87 270 277 10536352 10.1002/(SICI)1097-0142(19991025)87:5<270::AID-CNCR6>3.0.CO;2-L
Wee A Nilsson B Highly well differentiated hepatocellular carcinoma and benign hepatocellular lesions: Can they be distinguished on fine needle aspiration biopsy? Acta Cytol 2003 47 16 26 12585026
Takenaka A Kaji I Kasugai H Sasaki Y Ishiguro S Wada A Horai T Otani T Ishikawa H Usefulness of diagnostic criteria for aspiration cytology of hepatocellular carcinoma Acta Cytol 1999 43 610 616 10432883
Layfield LJ Focal fatty change of the liver: Cytologic findings in a radiographic mimic of metastases Diagn Cytopathol 1994 11 385 389 7895578
Zeppa P Anniciello A Vetrani A Palombini L Fine needle aspiration biopsy of hepatic focal fatty change: A report of two cases Acta Cytol 2002 46 567 570 12040655
Gupta SK Das DK Rajwanshi A Bhusnurmath SR Cytology of hepatocellular carcinoma Diagn Cytopathol 1986 2 290 294 2431848
Singh HK Silverman JF Geisinger KR Fine-needle aspiration cytomorphology of clear-cell hepatocellular carcinoma Diagn Cytopathol 1997 17 306 310 9316789 10.1002/(SICI)1097-0339(199710)17:4<306::AID-DC13>3.0.CO;2-S
Piatti B Caspani B Giudici C Ferrario D Fine needle aspiration biopsy of hepatocellular carcinoma resembling neuroendocrine tumor: A case report Acta Cytol 1997 41 583 586 9100803
Wee A Nilsson B Yap I Fine needle aspiration biopsy of small/intermediate cell tumors in the liver: Considerations in a Southeast Asian population Acta Cytol 1996 40 937 947 8842170
Maeda T Adachi E Kajiyama K Takenaka K Sugimachi K Tsuneyoshi M Spindle cell hepatocellular carcinoma: A clinicopathologic and immunohistochemical analysis of 15 cases Cancer 1996 77 51 57 8630939 10.1002/(SICI)1097-0142(19960101)77:1<51::AID-CNCR10>3.0.CO;2-7
Davenport RD Cytologic diagnosis of fibrolamellar carcinoma of the liver by fine needle aspiration Diagn Cytopathol 1990 6 275 279 1698598
Perez-Guillermo M Masgrau NA Garcia-Solano J Sola-Perez J de Agustin y de Agustin P Cytologic aspect of fibrolamellar hepatocellular carcinoma in fine-needle aspirates Diagn Cytopathol 1999 21 180 187 10450103 10.1002/(SICI)1097-0339(199909)21:3<180::AID-DC7>3.0.CO;2-O
Wu PC Fang JWS Lau VKT Lai CL Lo CK Lau JYN Classification of hepatocellular carcinoma according to hepatocellular and biliary differentiation markers: Clinical and biological implications Am J Pathol 1996 149 1167 1175 8863666
Vessey CJ de la Hall PM Hepatic stem cells: A review Pathology 2001 33 130 141 11358043 10.1080/00313020120038719
Jarnagin WR Weber S Tickoo SK Koea JB Obiekwe S Fong Y DeMatteo RP Blumgart LH Klimstra D Combined hepatocellular and cholangiocarcinoma: demographic, clinical, and prognostic factors Cancer 2002 94 2040 2046 11932907 10.1002/cncr.10392
Wee A Nilsson B Combined hepatocellular-cholangiocarcinoma: Diagnostic challenge in hepatic fine needle aspiration biopsy Acta Cytol 1999 43 131 138 10097699
Tsui WM Lam PW Mak CK Pay KH Fine-needle aspiration cytologic diagnosis of intrahepatic biliary papillomatosis (intraductal papillary tumor): Report of three cases and comparative study with cholangiocarcinoma Diagn Cytopathol 2000 22 293 298 10790236 10.1002/(SICI)1097-0339(200005)22:5<293::AID-DC7>3.0.CO;2-E
Pisharodi LR Bedrossian C Diagnosis and differential diagnosis of small-cell lesions of the liver Diagn Cytopathol 1998 19 29 32 9664180 10.1002/(SICI)1097-0339(199807)19:1<29::AID-DC6>3.0.CO;2-J
Prosser JM Dusenbery D Histocytologic diagnosis of neuroendocrine tumors in the liver: A retrospective study of 23 cases Diagn Cytopathol 1997 16 383 391 9143838 10.1002/(SICI)1097-0339(199705)16:5<383::AID-DC2>3.0.CO;2-D
Collins KA Geisinger KR Raab SS Silverman JF Fine needle aspiration biopsy of hepatic lymphomas: Cytomorphology and ancillary studies Acta Cytol 1996 40 257 262 8629408
Malhotra V Gondal R Tatke M Sarin SK Fine needle aspiration cytologic appearance of inflammatory pseudotumor of the liver: A case report Acta Cytol 1997 41 1325 1328 9990267
Nakama T Hayashi K Komada N Ochiai T Hori T Shiori S Tsubouchi H Inflammatory pseudotumor of the liver diagnosed by needle liver biopsy under ultrasonographic tomography guidance J Gastroenterol 2000 35 641 645 10955605 10.1007/s005350070066
Wee A Nilsson B Yap I Chong SM Aspiration cytology of liver abscesses: With an emphasis on diagnostic pitfalls Acta Cytol 1995 39 453 462 7762332
Hughes JH Jensen CS Donnelly AD Cohen MB Silverman JF Geisinger KR Raab SS The role of fine-needle aspiration cytology in the evaluation of metastatic clear cell tumors Cancer 1999 87 380 389 10603192 10.1002/(SICI)1097-0142(19991225)87:6<380::AID-CNCR9>3.0.CO;2-1
Powers CN Berardo MD Frable WJ Fine-needle aspiration biopsy: Pitfalls in the diagnosis of spindle-cell lesions Diagn Cytopathol 1994 10 232 241 8050331
Tao LC Davidson DD Aspiration biopsy cytology of smooth muscle tumors: A cytologic approach to the differentiation between leiomyosarcoma and leiomyoma Acta Cytol 1993 37 300 308 8498132
Padilla C Saez A Catala I Vidal A Garcia L Tolosa F Javier Andreu F Combalia N Fine-needle aspiration cytology diagnosis of metastatic gastrointestinal stromal tumor in the liver Diagn Cytopathol 2002 27 298 302 12411997 10.1002/dc.10191
Wee A Thamboo TP Thomas A Alpha-fetoprotein-producing liver carcinomas of primary extrahepatic origin: Fine needle aspiration biopsy experience in 2 cases Acta Cytol 2003 47 799 808 14526683
Bergman S Graeme-Cook F Pitman MB The usefulness of the reticulin stain in the differential diagnosis of liver nodules on fine-needle aspiration biopsy cell block preparations Mod Pathol 1997 10 1258 1264 9436973
Gagliano EF Reticulin stain in the fine needle aspiration differential diagnosis of liver nodules Acta Cytol 1995 39 596 598 7539204
Cedrone A Rapaccini GL Pompili M Grattagliano A Aliotta A Trombino C Neoplastic seeding complicating percutaneous ethanol injection for treatment of hepatocellular carcinoma Radiology 1992 183 787 788 1316621
Smith EH Complications of percutaneous abdominal fine-needle biopsy Radiology 1991 178 253 258 1984314
Takamori R Wong LL Dang C Wong L Needle-tract implantation from hepatocellular carcinoma: is needle biopsy of the liver always necessary? Liver Transpl 2000 6 67 72 10648580
Huang GT Sheu JC Yang PM Lee HS Wang TH Chen DS Ultrasound-guided cutting biopsy for the diagnosis of hepatocellular carcinoma – a study based on 420 patients J Hepatol 1996 25 334 338 8895013 10.1016/S0168-8278(96)80120-6
Schotman SN De Man RA Stoker J Zondervan PE IJzermans JNM Subcutaneous seeding of hepatocellular carcinoma after percutaneous needle biopsy Gut 1999 45 626 627 10486377
Sherman M Alphafetoprotein: an orbituary J Hepatol 2001 34 603 605 11394662 10.1016/S0168-8278(01)00025-3
Buscarini L Fornari F Bolondi L Colombo P Livraghi T Magnolfi F Rapaccini GL Salmi A Ultrasound-guided fine-needle biopsy of focal liver lesions: techniques, diagnostic accuracy and complications. A retrospective study on 2091 biopsies J Hepatol 1990 11 344 348 2290025 10.1016/0168-8278(90)90219-H
Fornari F Civardi G Cavanna L Rossi S Buscarini E Di Stasi M Sbolli G Buscarini L Ultrasonically guided fine-needle aspiration biopsy: a highly diagnostic procedure for hepatic tumors Am J Gastroenterol 1990 85 1009 1013 2165350
Caturelli E Bisceglia M Fusilli S Squillante MM Castelvertere M Siena DA Cytological vs microhistological diagnosis of hepatocellular carcinoma: Comparative accuracies in the same fine-needle biopsy specimen Dig Dis Sci 1996 41 2326 2331 9011437
Zhou H Wolff M Pauliet D Fischer HP Pfeifer U Multiple macroregenerative nodules in liver cirrhosis due to Budd-Chiari syndrome: Case reports and review of the literature Hepatogastroenterology 2000 47 522 527 10791228
Ferrell LD Crawford JM Dhillon AP Scheuer PJ Nakanuma Y Proposal for standardized criteria for the diagnosis of benign, borderline, and malignant hepatocellular lesions arising in chronic advanced liver disease Am J Surg Pathol 1993 17 1113 1123 8214256
Hytiroglou P Thiese ND Differential diagnosis of hepatocellular nodular lesions Semin Diagn Pathol 1998 15 285 299 9845429
Wanless IR Nodular regenerative hyperplasia, dysplasia, and hepatocellular carcinoma Am J Gastroenterol 1996 91 836 837 8633567
Terada T Ueda K Nakanuma Y Histopathological and morphometric analysis of atypical adenomatous hyperplasia of human cirrhotic livers Virchows Arch (A) Pathol Anat Histopathol 1993 422 381 388 8391735 10.1007/BF01605457
Goldstein NS Silverman JF Dabbs DJ Immunohistochemistry of the gastrointestinal tract, pancreas, bile ducts, gallbladder, and liver Diagnostic Immunohistochemistry 2002 New York. Churchill Livingstone 374 378
Gupta RK Kenwright DN Naran S Lallu S Fauck R Aspiration cytodiagnosis of small cell malignancies found in fine needle aspirate (FNA) of the liver: An immunocytochemical study Cytopathology 2000 11 262 267 10983726 10.1046/j.1365-2303.2000.00258.x
Lau SK Prakash S Geller SA Alsabeh R Comparative immunohistochemical profile of hepatocellular carcinoma, cholangiocarcinoma, and metastatic adenocarcinoma Hum Pathol 2002 33 1175 1181 12514785 10.1053/hupa.2002.130104
Ishikura H Kishimoto T Andachi H Kakuta Y Yoshiki T Gastrointestinal hepatoid adenocarcinoma: venous permeation and mimicry of hepatocellular carcinoma, a report of four cases Histopathology 1997 31 47 54 9253624 10.1046/j.1365-2559.1997.5740812.x
Saad RS Luckasevic TM Noga CM Fukuda H Tanikawa K Diagnostic value of HepPar1, pCEA, CD10, and CD34 expression in separating hepatocellular carcinoma from metastatic carcinoma in fine-needle aspiration cytology Diagn Cytopathol 2004 30 1 6 14696137 10.1002/dc.10345
Stahl J Voyvodic F Biopsy diagnosis of malignant versus benign liver "nodules": New helpful markers. An update Adv Anat Pathol 2000 7 230 239 10907808
Tsuji M Kashihara T Terada N Mori H An immunohistochemical study of hepatic atypical adenomatous hyperplasia, hepatocellular carcinoma, and cholangiocarcinoma with alpha-fetoprotein, carcinoembryonic antigen, CA 19-9, epithelial membrane antigen, and cytokeratins 18 and 19 Pathol Int 1999 49 310 317 10365850 10.1046/j.1440-1827.1999.00865.x
Brumm C Schulze C Charels K Morohoshi T Kloppel G The significance of alpha-fetoprotein and other tumor markers in differential immunohistochemistry of primary liver tumors Histopathology 1989 14 503 513 2472345
Hurrlimann J Gardiol D Immunohistochemical characterization of 130 cases of primary hepatic carcinomas Am J Surg Pathol 1991 15 280 288 1847609
Ma CK Zarbo RJ Frierson HF Lee MW Comparative immunohistochemical study of primary and metastatic carcinomas of the liver Am J Clin Pathol 1993 99 551 557 7684185
Minervini MI Demetris AJ Lee RG Carr BI Madariaga J Nalesnik MA Utilization of hepatocyte-specific antibody in the immunocytochemical evaluation of liver tumors Mod Pathol 1997 10 686 692 9237179
Sato K Tanaka M Kusaba T Fukuda H Tanikawa K Immunohistochemical demonstration of alpha-fetoprotein in small hepatocellular carcinoma Oncol Rep 1998 5 355 358 9468556
Theise ND Fiel IM Hytiroglou P Ferrell L Schwartz M Miller C Thung SN Macroregenerative nodules in cirrhosis are not associated with elevated serum or stainable tissue alpha-fetoprotein Liver 1995 15 30 34 7539880
Wee A Nilsson B pCEA canalicular immunostaining in fine needle aspiration biopsy diagnosis of hepatocellular carcinoma Acta Cytol 1997 41 1147 1155 9250313
Borscheri N Roessner A Rocken C Canalicular immunostaining of neprilysin (CD10) as a diagnostic marker for hepatocellular carcinomas Am J Surg Pathol 2001 25 1297 1301 11688465 10.1097/00000478-200110000-00011
Lin F Abdallah H Meschter S Diagnostic utility of CD10 in differentiating carcinoma from metastatic carcinoma in fine-needle aspiration biopsy (FNAB) of the liver Diagn Cytopathol 2004 30 92 97 14755758 10.1002/dc.10419
Morrison C Marsh W JrFrankel WL A comparison of CD10 to pCEA, MOC-31, and Hepatocyte for the distinction of malignant tumors in the liver Mod Pathol 2002 15 1279 1287 12481008 10.1097/01.MP.0000037312.69565.24
Chu PG Ishizawa S Wu E Weiss LM Hepatocyte antigen as a marker of hepatocellular carcinoma: an immunohistochemical comparison to carcinoembryonic antigen, CD 10, and alpha-fetoprotein Am J Surg Pathol 2002 26 978 988 12170084 10.1097/00000478-200208000-00002
Fan Z van de Rijn M Montgomery K Rouse RV Hep Par 1 antibody stain for the differential diagnosis of hepatocellular carcinoma: 676 tumors tested using tissue microarrays and conventional tissue sections Mod Pathol 2003 16 137 144 12591966 10.1097/01.MP.0000052103.13730.20
Siddiqui MT Saboorian MH Gokaslan ST Ashfaq R Diagnostic utility of the HepPar1 antibody to differentiate hepatocellular carcinoma from metastatic carcinoma in fine-needle aspiration samples Cancer (Cancer Cytopathol) 2001 96 49 52
Zimmerman RL Burke MA Young NA Solomides CC Bibbo M Diagnostic value of hepatocyte paraffin 1 antibody to discriminate hepatocellular carcinoma from metastatic carcinoma in fine needle aspiration biopsies of the liver Cancer (Cancer Cytopathol) 2001 93 288 291
Leong AS-Y Sormunen RT Tsui WMS Liew CT Hep Par 1 and selected antibodies in the immunohistological distinction of hepatocellular carcinoma from cholangiocarcinoma, combined tumours and metastatic carcinoma Histopathology 1998 33 318 324 9822920 10.1046/j.1365-2559.1998.00522.x
Maeda T Kajiyama K Adachi E Takenaka K Sugimachi K Tsuneyoshi M The expression of cytokeratins 7, 19, and 20 in primary and metastatic carcinomas of the liver Mod Pathol 1996 9 901 909 8878022
de Boer WB Segal A Frost FA Sterrett GF Can CD 34 discriminate between benign and malignant hepatocytic lesions in fine-needle aspirates and thin core biopsies? Cancer (Cancer Cytopathol) 2000 90 273 278
Kong CS Appenzeller M Ferrell LD Utility of CD 34 reactivity in evaluating focal nodular hepatocellular lesions sampled by fine needle aspiration biopsy Acta Cytol 2000 44 218 222 10740609
| 15941489 | PMC1177974 | CC BY | 2021-01-04 16:36:23 | no | Cytojournal. 2005 Jun 8; 2:7 | utf-8 | Cytojournal | 2,005 | 10.1186/1742-6413-2-7 | oa_comm |
==== Front
Dyn MedDynamic medicine : DM1476-5918BioMed Central London 1476-5918-4-61593509710.1186/1476-5918-4-6ResearchReplacing a Swiss ball for an exercise bench causes variable changes in trunk muscle activity during upper limb strength exercises Lehman Gregory J [email protected] Trish [email protected] Jo [email protected] Patricia [email protected] Sara [email protected] Department of Graduate Studies, Canadian Memorial Chiropractic College, Toronto, ON, Canada2 Undergraduate Department, Canadian Memorial Chiropractic College, Toronto, ON, Canada2005 3 6 2005 4 6 6 18 11 2004 3 6 2005 Copyright © 2005 Lehman et al; licensee BioMed Central Ltd.2005Lehman 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 addition of Swiss balls to conventional exercise programs has recently been adopted. Swiss balls are an unstable surface which may result in an increased need for force output from trunk muscles to provide adequate spinal stability or balance. The aim of the study was to determine whether the addition of a Swiss ball to upper body strength exercises results in consistent increases in trunk muscle activation levels.
Methods
The myoelectric activity of four trunk muscles was quantified during the performance of upper body resistance exercises while seated on both a stable (exercise bench) and labile (swiss ball) surface. Participants performed the supine chest press, shoulder press, lateral raise, biceps curl and overhead triceps extension. A repeated measures ANOVA with post-hoc Tukey test was used to determine the influence of seated surface type on muscle activity for each muscle.
Results & Discussion
There was no statistically significant (p < .05) difference in muscle activity between surface conditions. However, there was large degree of variability across subjects suggesting that some individuals respond differently to surface stability. These findings suggest that the incorporation of swiss balls instead of an exercise bench into upper body strength training regimes may not be justified based only on the belief that an increase spinal stabilizing musculature activity is inherent. Biomechanically justified ground based exercises have been researched and should form the basis for spinal stability training as preventative and therapeutic exercise training regimes.
Conclusion
Selected trunk muscle activity during certain upper limb strength training exercises is not consistently influenced by the replacement of an exercise bench with a swiss ball.
EMGexercisespine stabilityswiss ballsrehabilitationlow back pain
==== Body
Background
The use of physioballs/Swiss balls in strength and conditioning programs has become ubiquitous. Swiss balls have been incorporated into strength training regimes and touted as a means to more effectively train the musculoskeletal system. Performing strength exercises on Swiss balls has been advocated on the belief that a labile surface will provide a greater challenge to the trunk musculature, increase the dynamic balance of the user and possibly train users to stabilize their spines to prevent and treat injury.
Despite a few studies, the research supporting these ideas is sparse. Vera-Garcia et al [1] documented increases in rectus abdominis and external oblique activity during curl ups when performed on a Swiss ball compared with a stable surface. Mori [2] documented trunk muscle activation levels during a variety of trunk muscle exercises showing that substantial levels of trunk muscle activity occurs. However, because the exercise tasks were not also performed on a stable surface it is unknown how much the Swiss ball's instability contributed to a demand for muscle activation. Andersen [3] investigated the influence of a Swiss ball on upper limb muscle activation and force production during a chest press. The study found that while muscle activation in the primary movers was not different between surfaces, the amount of force generated was significantly less on a Swiss ball. These results were mirrored in a previous study [4] investigating force and muscle activation of the lower extremity on unstable surfaces.
Swiss balls are currently used to replace stable benches during the performance of upper body strength training exercises. While previous work has documented the myoelectric activity of the trunk muscles during exercises specifically designed to train the trunk muscles, no study has documented the effect of an unstable surface on trunk muscle activity during resistance exercises for the upper limbs. Due to common use of Swiss balls this lack of research is significant for both performance and safety concerns (i.e. Swiss balls may increase the risk of falling without providing an exercise benefit). Adequate spinal stability is important in the prevention and treatment of low back injuries [5]. Stability is achieved through the coactivation of trunk muscles; therefore, endurance training has been postulated to be beneficial in training trunk muscles to provide stability. It is possible that performing upper body strength exercises on a swiss ball can increase trunk muscle activity to a sufficient extent to adequately stress the spinal stabilizing musculature to achieve beneficial endurance training effects. This may render conventional trunk resistance exercises superfluous and increase the efficiency of rehabilitation and prophylactic exercise programs. Contrarily, an elevated muscle activation level may be contraindicated in subjects with low back injury or unstable spines. Co-activation of the trunk muscles has a compressive loading cost that may outweigh the benefits of trunk muscle training. Safe exercises on stable ground have been advocated and thoroughly investigated with a detailed biomechanical model [6] which provide an excellent balance between muscle stress and low compressive/shear penalty, the same can not be argued for the majority of exercises incorporating the use of Swiss balls.
In light of the popularity of Swiss balls and the lack of research investigating their influence on trunk muscle activity during upper limb strength exercises, it was the aim of this study to determine if the use of a Swiss ball instead of an exercise bench results in consistent increases across subjects in trunk muscle activation levels during upper body strength training exercises.
Methods
Participant Characteristics
Seven healthy males (average age (standard deviation) 28 (3.8), average height in cm (standard deviation) 179.7 (7.13) and average mass in kg 84.6 (8.09) and five females (age = 23.6 kg(.8), height 168.3 cm(5.04) and mass 61 kg(5.2)) with weight lifting and abdominal exercise experience, were recruited from a convenience sample consisting of college students. Subjects currently experiencing low back pain or a history of low back pain within 3 months were excluded from the study. Participants read and signed an information and consent form approved by the institutions Internal Review Board.
Experimental Design
A single factor repeated measures design was used to analyze the effect of trunk muscle activity during common weight training exercises on a Swiss ball compared to the trunk muscle activity found during the performance of the exercises on an exercise bench. All subjects performed 6 different exercises on two different support surfaces for a total of 12 separate movement tasks during a single testing session. The order of the tasks was kept constant with every subject.
Instrumentation
EMG data was collected using disposable bipolar Ag-AgCl disc surface electrodes with a diameter of one cm, adhered bilaterally over the muscle groups with a centre-to-centre spacing of 2 cm. Raw EMG was amplified between 1000 and 20,000 times, depending on the subject. The amplifier had a CMRR of 10,000:1 (Bortec EMG, Calgary AB, Canada). Raw EMG was band pass filtered (10 and 1000 HZ) and A/D converted at 2000 Hz using a National Instruments data acquisition system.
Electrode Placement
Skin preparation included shaving (when necessary) and cleansing and abrading the skin with alcohol solution prior to applying the electrodes to reduce skin impedance. Four sites on participants' right sides were chosen for electrode placement: (1) rectus abdominus (RA) 3 cm lateral to the umbilicus, (2) External Oblique (EO) 15 cm lateral to the umbilicus oriented in the direction of the muscle fibres (3) internal oblique (IO) 10 cm lateral to midline (inferior (2 cm) to the ASIS angled superomedial to inferolateral parallel to the underlying muscle fibers and (4) erector spinae 2 cm lateral to L4–L5 interspinous space in a superomedial to an inferolateral orientation over the muscle fibers. A reference electrode was placed over the eleventh rib. The investigators may have made slight modifications and adjustments for any anatomical variations between subjects.
Normalization task procedure
Subjects were required to perform maximum voluntary contractions for the trunk musculature. Subjects were required to perform a 3 second maximal supine isometric trunk curl up and bilateral twist against an immovable resistance to maximally recruit the rectus abdominis, external oblique and internal oblique. Subjects, performed this movement supine with the spine in neutral. Second, the subjects performed an isometric prone trunk extension against a fixed resistance to maximally recruit the lumbar erector spinae. All subjects performed the normalization tasks in the same order. Participants practiced the exercises before the collection of data. The muscle activity during all subsequent exercise tasks was expressed as a percent of the peak activity found during the normalization procedure. The peak activity was found visually after the signal had been processed in an identical manner to the exercise tasks.
Exercises Performed
Subjects performed six exercise tasks. The six exercises were modified to be performed while seated on the labile surface of an exercise ball resulting in a total of 12 separate movement tasks:
1.a) Supine abdominal curls on a flat bench with feet flat on floor and folded across chest.
1.b) Modified; supine abdominal curl on a Swiss ball with feet flat on floor and Swiss ball positioned under the low back.
2.a) Supine dumbell chest press on a flat bench, with feet flat on the floor. Subjects started with weight at chest level and hands shoulder width apart. Subjects pressed weight up until elbows were extended.
2.b) Modified; supine chest press on a Swiss ball with feet flat on the floor and Swiss ball positioned under the shoulders and thoracic spine.
3.a) Seated shoulder press on flat bench with no back support and feet on the floor. Subjects started with weights at shoulder level and pressed up to full elbow extension.
3.b) Modified; seated shoulder press on Swiss ball.
4.a) Seated lateral shoulder raise on flat bench. The weighted straight arm was abducted to 90 degrees from 0 degrees of abduction.
4.b) Modified; seated lateral shoulder raise on a Swiss ball.
5.a) Seated two arm biceps curl on flat bench. Subjects started in the anatomical position and flexed the arms bilaterally.
5.b) Seated two arm biceps curl on Swiss ball.
6.a) Seated double arm overhead triceps extension on a flat bench. Subjects began with their shoulders and elbows fully flexed and the weight behind their head. The elbow was then extended to raise and lower the weight.
6.b) Modified; seated double arm overhead triceps extension on a Swiss ball.
Exercises 2–6 were performed with dumbbells. Each subject selected a weight such that they are able to complete 4 repetitions of each task without reaching fatigue. The exercise task saw the subject perform slow controlled concentric contractions followed by eccentric lowering of the weight. This was considered one repetition. Without pause 3 repetitions were repeated during a trial. The same weight was used for both the original and modified versions of each exercise task. The weight varied from 10 – 40 lbs. All six exercises were performed in order from 1 to 6 (a) on a hard flat bench then repeated in the same order on a Swiss ball (1–6b). Two sets occurred for each exercise task. It should be noted that the curl up exercise on the ground and on the Swiss ball does not control for posture, muscle length or other factors which can influence the myoelectric signal. This exercise was mainly included to give the reader a biologically significant reference for the amount of muscle activity occurring during the exercises. An inference of whether a Swiss ball influences muscle activity during a curl up can not be made due the significant differences in posture. For the other exercises studied a neutral lumbar curve could be maintained through out either condition.
Description of Exercise Movement
After being instrumented, subjects performed the normalization tasks and then the 12 movement tasks. The subjects were instructed to perform the concentric phase for 2 seconds and the eccentric phase for 4 seconds (under the count and supervision of two examiners). Data was collected for 25 seconds during each exercise.
EMG Processing
The raw myoelectric signal of all trials for both the MVCs and the exercise tasks was processed identically. A linear envelop was calculated by first full wave rectifying (the absolute value of each data point) and then smoothing using a 100 ms moving average with a 50 ms overlap. The average activity over the course of the movement was then calculated for each trial and expressed as a percentage of the activity found during the MVC for each specific muscle and participant.
Statistical Analysis
A repeated measures ANOVA was then used to determine if a difference in type of supporting surface influenced trunk muscle activity for each muscle. A post hoc Tukey test was used to examine if statistically significant (p < .05) differences exist in trunk muscle activity for the different exercises performed.
Qualitative Analysis
The difference in muscle activity (expressed as a %MVC) was also calculated for each subject (12), muscle (4) and exercise (6) between the two surface conditions for a total of 240 differences in muscle activity. A change greater than 5% MVC was considered significant. The number of occurrences of a significant increase or decrease in muscle activity was recorded.
Results
There was no significant difference found between performing each of the six exercises on the Swiss ball and the flat bench for any of the four muscle groups investigated. Tables one to six details the average muscle activity for each muscle group during the six different exercises studied. Twenty six muscles showed significant increases in muscle activity and 22 muscles showed a significant decrease in muscle activity across all conditions.
While there was not a statistically significant increase in any muscle studied, the Internal Oblique muscle tended to have the greatest number of increases in muscle activity when on the Swiss ball compared with the bench. The grouped average internal oblique muscle activity difference between the two conditions tended to be greater than 5% MVC during the curl up, bench press and shoulder press. The bench press exercise showed a trend for muscles to increase their myoelectric signal on the Swiss ball (14 occurrences), while there were only two instances when a muscle's activity decreased more than 5% MVC.
Absolute increases of greater than 11% MVC were seen in the internal oblique in three subjects when performing the bench press on a Swiss ball compared with performing the bench press on the more stable bench. One subject's average activity was approximately 3% MVC on the bench and increased to more than 17% MVC when on the Swiss ball. The biceps curl exercise had 6 occurences of trunk muscles increasing their activity on the Swiss ball while the triceps extension had only one incidence of an increase in muscle activity yet seven occurrences where the average muscle activity decreased more than 5% MVC. The lateral raise exercise showed no trends with only 3 instances of activity increases and 3 instances where muscle activity decreased for all muscles studied. While the shoulder press showed a trend of increased internal oblique activity on the Swiss ball (average absolute difference of 6.52% MVC) there was only 2 instances where any trunk muscle increased its activity on the Swiss ball more than 5% MVC and 5 instances of a muscle showing a decrease in activity of more than 5% MVC.
Discussion
Replacing an exercise bench with a Swiss ball is not a guarantee for increased trunk muscle activation during upper body strength exercises. There does not appear to be a consistent, generalized response to the addition of a Swiss ball. Statistically, there is no difference between conditions, however the study population showed large variability. This suggests that individuals respond differently to unstable surfaces. Health and fitness professionals who advocate the addition of Swiss balls into exercise programs for the upper limbs can not support this change via the argument that the spinal musculature system is stressed to a greater extent (i.e. increased muscle activity) for all individuals. Importantly, this study does not dismiss the use of Swiss balls for exercises designed to train the trunk muscles. Increases in trunk muscle activity have been documented for these exercises [1] but a general increase in trunk muscle activity was not seen for the upper limb strength exercises studied in the current study. The study's findings also suggest that performing upper limb strength exercises on a Swiss ball does not cause excessive compressive loading due to increased trunk muscle co-activation and therefore may be safe for the low back injured. Changes in compressive or shear loading may different due to postural factors but this was not measured in the present study. What needs to be questioned is what benefit exists in performing upper limb strength exercises on a Swiss ball. An injury risk may still be present because Swiss balls are unstable and may increase the risk of falling and subsequent injury. If the justification is to "train the core" (i.e. recruit agonist-antagonist trunk muscles) then this can't be supported by the results of this study. If other justifications are made (increases in balancing ability, recruitment of secondary hip and leg muscles) then this exercise modification may be reasonable. To minimize injury risk, like any exercise program, exercise difficulty progression should be used in incorporating Swiss balls into an exercise program. Participants who wish to use a Swiss ball in their exercise program should learn the basic upper body strength moves on a stable surface first.
Advocating the use of the Swiss ball in exercise or rehabilitation programs may be justified via other benefits. A recent study has documented short term gains in one legged stance following an exercise program of abdominal curl ups and trunk extensions on a Swiss ball [7]. Swiss balls are often more portable and affordable than a traditional weight bench and may therefore increase exercise compliance and adoption. Anecdotally, Swiss ball classes are popular and enjoyable.
Exercise prescription should be goal dependent. If a therapist merely wants variety in an exercise program and increased exercise compliance and enjoyment then the adoption of a Swiss ball appears reasonable but not justified biomechanically. If the aim of a therapist is to rehabilitate or prevent low back injury then sound biomechanically justified or clinically proven rehabilitation protocols should be advocated. Kavcic et al [6] provides biomechanical support for ground based simple exercises (curl up, side bridge, four point kneeling with leg extension) to adequately train the spinal stabilizers while minimizing the compressive/shear penalty and ensuring adequate spinal stability.
This study is limited to the exercises investigated and weights used. For many of the exercises the weight was not near the maximum load the participant could use. The weight levels were chosen based on the rationale that the same low weight is used during "FitBall" classes geared toward a novice exerciser. Challenging each subject with a greater load may influence trunk muscle activity. Future work should address this limitation. Additionally, only surface electromyographic activity was recorded. The muscles studied are considered global stability muscles and may not adequately represent the muscle activation levels in smaller intersegmental spinal muscles. These muscles have a greater proprioceptive function and if the Swiss ball stresses these muscles to a greater extent this may form the basis for an improved balance effect following training. As well, no measures of range of motion occurred. It is possible that different ranges of motion were seen which would alter the myoelectric signal amplitude without a change in force production due to the length tension properties of muscle. Conversely, more advantageous postures would allow greater force production and hence spinal stability without changes in muscle activity. This study is also limited to conclusions regarding the stability of the surface examined. Other labile surfaces (wobble boards) may result in differences in trunk muscle activity recruitment. These results can not be generalized to all unstable surfaces and all strength training exercises.
Conclusion
A consistent generalized trend was not seen across subjects (subjects did not uniformly increase trunk muscle activity) when replacing an exercise bench with a Swiss ball during upper limb strength exercises. Individual responses were variable. This suggests that participants respond differently to surface stability modifications.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
GL: Conception, design, data collection, data analysis, manuscript preparation
TG, JL, PP, ST: design, data collection, manuscript preparation
Table 1 Abdominal curl up exercise muscle activation levels (% MVC, standard deviation in brackets), average difference between surfaces and number of participants whose change in activity was greater than 5% MVC during
Surface Muscles studied during abdominal curl up exercise
Rectus Abdominis External Oblique Internal Oblique Erector Spinae
Ball 29.0 (33.1) 23.2 (20.6) 32.9 (27.6) 3.2 (3.5)
Bench 25.7 (16.6) 21.1 (13.4) 27.1 (16.4) 6.2 (10.6)
Difference 3.36(20.6) 2.17 (11.65) 5.82 (18.2) 3 (7.61)
Increase 2 2 4 0
Decrease 2 3 2 1
Table 2 Bench Press Exercise muscle activation levels (% MVC, standard deviation in brackets), average difference between surfaces and number of participants whose change in activity was greater than 5% MVC.
Surface Muscles studied during bench press exercise
Rectus Abdominis External Oblique Internal Oblique Erector Spinae
Ball 7.4 (6.3) 5.7 (6.8) 13.5 (9.2) 6.06 (5.9)
Bench 4.7 (6.2) 3.1 (2.8) 8.2 (6.8) 3.1 (1.6)
Difference 2.68 (6.49) 2.52 (4.7) 5.2 (6.26) 2.93 (5.9)
Increase 4 3 4 3
Decrease 2 0 0 0
Table 3 Biceps Curl Exercise muscle activation levels (% MVC, standard deviation in brackets), average difference between surfaces and number of participants whose change in activity was greater than 5% MVC.
Surface Muscles studied during biceps curl exercise
Rectus Abdominis External Oblique Internal Oblique Erector Spinae
Ball 5.0 (5.8) 3.0 (4.4) 9.0 (8.4) 8.7 (7.4)
Bench 4.2 (5.7) 2.2 (1.9) 6.9 (6.0) 6.5 (6.3)
Difference .83 (2.65) .74 (2.86) 2.14 (5.79) 2.23 (4.67)
Increase 1 1 1 3
Decrease 0 0 0 0
Table 4 Lateral raise exercise muscle activation levels (% MVC, standard deviation in brackets) average difference between surfaces and number of participants whose change in activity was greater than 5% MVC.
Surface Muscles studied during lateral raise exercise
Rectus Abdominis External Oblique Internal Oblique Erector Spinae
Ball 5.2 (6.2) 3.0 (3.3) 7.8 (7.6) 3.0 (2.0)
Bench 5.3 (8.5) 2.0 (1.9) 6.5 (6.1) 4.0 (4.2)
Difference -.07 (3.21) .99 (2) 1.23 (1.97) -1 (4.26)
Increase 1 1 1 0
Decrease 2 0 0 1
Table 5 Shoulder press exercise muscle activation levels (% MVC, standard deviation in brackets) average difference between surfaces and number of participants whose change in activity was greater than 5% MVC.
Surface Muscles studied during shoulder press exercise
Rectus Abdominis External Oblique Internal Oblique Erector Spinae
Ball 6.01 (6.29) 4.1 (5.4) 21.7 (31.5) 3.7 (3.3)
Bench 6.9 (9.6) 3.5 (3.6) 15.2 (15.2) 13.4 (30.3)
Difference -.98 (4.29) .6 (2.59) 6.52 (30.23) -1.07 (4.62
Increase 0 1 1 0
Decrease 1 0 3 1
Table 6 Triceps extension exercise muscle activation levels (% MVC, standard deviation in brackets) average difference between surfaces and number of participants whose change in activity was greater than 5% MVC.
Surface Muscles Studied during triceps extension exercise
Rectus Abdominis External Oblique Internal Oblique Erector Spinae
Ball 4.3 (3.6) 3.7 (4.3) 13.5 (12.5) 3.4 (3.3)
Bench 9.8 (16.6) 4.3 (4.0) 16.3 (16.5) 3.1 (2.0)
Difference -5.51 (14.04) .67 (1.91) 2.76 (5.8) .31 (3.42)
Increase 0 0 0 1
Decrease 3 1 3 0
==== Refs
Vera-Garcia FJ Grenier SG McGill SM Abdominal muscle response during curl-ups on both stable and labile surfaces Physical Therapy 2000 80 564 9 10842409
Mori A Electromyographic activity of selected trunk muscles during stabilization exercises using a gym ball Electromyogr Clin Neurophysiol 2004 44 57 64 15008027
Anderson KG Behm DG Maintenance of EMG activity and loss of force output with instability J Strength Cond Res 2004 18 637 40 15320684 10.1519/1533-4287(2004)18<637:MOEAAL>2.0.CO;2
Behm DG Anderson K Curnew RS Muscle force and activation under stable and unstable conditions J Strength Cond Res 2002 16 416 22 12173956 10.1519/1533-4287(2002)016<0416:MFAAUS>2.0.CO;2
McGill SM Grenier S Kavcic N Cholewicki J Coordination of muscle activity to assure stability of the lumbar spine J Electromyogr Kinesiol 2003 13 353 9 12832165 10.1016/S1050-6411(03)00043-9
Kavcic N Grenier S McGill SM Quantifying tissue loads and spine stability while performing commonly prescribed low back stabilization exercises Spine 2004 29 2319 29 15480148 10.1097/01.brs.0000142222.62203.67
Cosio-Lima LM Reynolds KL Winter C Paolone V Jones MT Effects of physioball and conventional floor exercises on early phase adaptations in back and abdominal core stability and balance in women J Strength Cond Res 2003 17 721 5 14636114 10.1519/1533-4287(2003)017<0721:EOPACF>2.0.CO;2
| 15935097 | PMC1177975 | CC BY | 2021-01-04 16:37:29 | no | Dyn Med. 2005 Jun 3; 4:6 | utf-8 | Dyn Med | 2,005 | 10.1186/1476-5918-4-6 | oa_comm |
==== Front
Filaria JFilaria Journal1475-2883BioMed Central London 1475-2883-4-31591670810.1186/1475-2883-4-3Short PaperThe subcutaneous movements of filarial infective larvae are impaired in vaccinated hosts in comparison to primary infected hosts Babayan Simon A [email protected] Tarik [email protected] Phat N [email protected] Goff Laetitia [email protected] Jean-Charles [email protected] Odile [email protected] Parasitologie Comparée et Modèles Expérimentaux, Muséum National d'Histoire Naturelle, Paris, France2 Institutes of Evolution, Immunology and Infection Research, University of Edinburgh, Edinburgh, UK3 Unité d'Anatomie et de Cytologie pathologiques, Hôpital St Michel, Paris, France4 Laboratoire de Parasitologie, Faculté de Pharmacie, Chatenay-Malabry, France2005 25 5 2005 4 3 3 27 8 2004 25 5 2005 Copyright © 2005 Babayan et al; licensee BioMed Central Ltd.2005Babayan 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.
Our aim in this study was to observe the movements of filarial infective larvae following inoculation into the mammalian host and to assess the effect of vaccination on larval migration, in situ. Here we present recordings of larvae progressing through the subcutaneous tissues and inguinal lymph node of primary infected or vaccinated mice. We used the filaria Litomosoides sigmodontis in BALB/c mice that were necropsied 6 hours after the challenge inoculation of 200 larvae. Subcutaneous tissue sections were taken from the inoculation site and larvae were filmed in order to quantify their movements. Our analyses showed that the subcutaneous larvae were less motile in the vaccinated mice than in primary-infected mice and had more leucocytes attached to the cuticle. We propose that this reduced motility may result in the failure of a majority of larvae to evade the inflammatory reaction, thereby being a possible mechanism involved in the early vaccine-induced protection.
==== Body
Findings
Effective protection induced by vaccination with irradiated larvae has been assessed in all filarial systems studied [1-3]. Previous studies have also shown that a characteristic feature of the immune response generated by irradiated larvae vaccines is a strong Th2 polarization [4-8]. With the model Litomosoides sigmodontis in mice [9], protection has been shown to occur within the first 48 hours post-inoculation and that the killing of larvae takes place in the subcutaneous tissue as in primary infected mice [10]. However, whereas only neutrophils infiltrate this tissue in primary infected mice, eosinophils are also present in vaccinated mice, and in conjunction with B cells account for the protection [11]. In both primary-infected and vaccinated mice, the larvae that survive are those that penetrate into the lymphatic vessels [11-14].
Inflammatory processes, coupled or not with the Th2 bias, affect the cellular and extracellular composition of the subcutaneous tissue, thus modifying its physical properties [15-21]. Therefore, it appeared relevant to investigate the effects of vaccination on this early but decisive phase of the larvae's migration, more precisely, on the motility of the larvae passing through the connective tissues of the mice. Indeed, filarial worms, like their free-living ancestors, have a powerful musculature and sensorial organs consisting of amphids and head papillae that allow autonomous and directional movements. In two non-filarial nematode parasites, the motility has been studied in naive and immunized hosts, and a reduction assessed [22,23].
The analysis of the mechanical activity of filarial nematodes in subcutaneous tissues at the onset of infection required a tailored protocol.
L. sigmodontis infective larvae were obtained as previously described [24]. Six week-old BALB/c mice, from Charles River, Cléon, France, were vaccinated with three inoculations of irradiated infective larvae as previously described [11]. Mice were challenged with 200 infective larvae (L3) inoculated subcutaneously in the right lumbar region. All experiments and procedures conformed to the French Ministry of Agriculture regulations for animal experimentation (1987).
Six hours after the challenge inoculation, the mice were culled, their skin separated from the animal and placed inner side up on a dissection board. The inguinal and iliac lymph nodes were then sampled, placed between a microscope slide and cover slip, lightly crushed then observed under the microscope.
A total of ten ~50 μm-thick slices of subcutaneous tissue were taken from the zone where the worms were inoculated with a razor blade, progressively spiraling away from that zone. The sections were then mounted for examination under a microscope, in a drop of RPMI 1640 (Fig. 1). Larvae that were found on the edge of the subcutaneous sample or cut by the razor blade were not included in the study. Each mouse was studied for no longer than an hour. Preparations that contained live worms were digitally recorded directly under the microscope, with Adobe Premiere ® 5.0 at 25 frames per second. Three magnifications were used: ×10, ×25 and ×100. Recordings lasted 30 seconds for each magnification and the movement of the worms assessed over that period with the following criteria: type of movements (undulating or uncoordinated), forward or backward progression, speed, oscillations of the head. The non-parametric Mann & Whitney's U test was used to compare groups.
Figure 1 A depiction of the subcutaneous tissue of mice inoculated with L. sigmodontis observed 6 hours post challenge: thin tissue section mounted in RPMI 1640, with dispersed infective larvae (asterisks, or arrowhead if damaged by the razorblade), a few nerves and adipose cells (dotted area) are present.
Five primary infected mice and 5 vaccinated mice were necropsied 6 hours post-challenge. At the dissection microscope, the inoculated area was easily identified because the subcutaneous tissue had a gelatinous aspect, was swollen by oedema and was densely infiltrated by leukocytes particularly in vaccinated mice. A total of 2000 L3 were inoculated, of which 1.95 % were identified and filmed: 39 in the subcutaneous tissue and 4 in the inguinal lymph node. This low ratio of recovery indicates that the larvae were outside the area studied. Larvae extracted from lymph nodes were all from a primary-infected mouse. They undulated and sometimes progressed rapidly, though with apparently poor coordination (Additional file 1); however, good coordination was obtained when pressure was exerted on the cover slip.
The observations of infective larvae in subcutaneous tissues were reported according to the following criteria, which readily discriminate those in primary-infected from those in vaccinated mice: the intensity of movements, the undulations and progression of the worm's head, and host cells attached to or up against the larva. Table 1 gives details about the larvae filmed and the numbers of larvae in each category.
In both groups of mice, the subcutaneous tissue preparation in which larvae undulated had a gel-like consistency. In primary-infected mice the larvae's head oscillated from one side to another at an estimated angle of 20° from the body axis. The rest of the body undulated: well-coordinated S-shaped curves moved from head to tail when the larvae moved forward, and from tail to head when the larvae moved backwards (Additional file 2). The larvae were generally vigorous, often progressing forward in the subcutaneous tissues as observed in 74% of the worms studied. In this environment, the larvae pushed themselves forward by taking advantage of denser parts of the tissues (Additional file 2). In primary-infected mice their movements reached 2.87 on a scale of 5, corresponding to fast movements (Fig. 2). We managed to measure the distance some larvae covered: 8 mm in 30 seconds, corresponding to 1.5 cm per minute.
In vaccinated mice, the larvae's movements were reduced as compared to what had been observed in the primary infections (average of 2 on 5, Additional file 3). A convenient measure of the larvae's movements was the amplitude of their head's side-to-side movements (Additional file 4). Only 8% (1 out of 13 filariae) of the infective larvae moved their head forward in vaccinated mice whereas 73% (11 out of 15 filariae) did in primary-infected mice. Additionally, we more often observed cells attached to the larvae's cuticle, on the anterior third of the body (3 out of 17 worms vs. 1 out of 10, Additional files 3 and 5). Interestingly, we observed that these granulomas form first in the front third of the worm (Additional file 5).
In conclusion, this preliminary study shows that by vaccinating the hosts, the movements of the worms are significantly impaired in comparison to primary-infected hosts and that more cells attach to their cuticle, thus contributing to their killing. Therefore, we suggest that this would prevent or delay entry into the lymphatic vessels through which they evade the inflammatory reaction and migrate to the pleural cavity.
Table 1 Distribution of infective larvae (L3) observed in the subcutaneous tissues of vaccinated (Vacc, mice 1 to 5) and primary-infected (PI, mice 6 to 10) mice 6 hours after the challenge inoculation. The larvae that broke out of the tissue sample during the procedure were not included in the rest of the study.
Mouse Number of L3 in challenge Number of larvae recovered Number of non-motile or abnormal larvae Number of larvae recorded Number of larvae in sub cut tissue Number of larvae escaped during procedure Number of larvae in lymph node
1 200 1 0 1 1 1 0
2 200 6 1 4 4 1 0
3 200 3 0 3 3 1 0
4 200 6 1 5 5 0 0
5 200 12 1 11 11 5 0
TOTAL VACC. 1000 28 3 24 24 8 0
6 200 4 4 0 0 0 0
7 200 5 2 3 3 2 0
8 200 8 1 7 3 1 4
9 200 7 1 0 0 0 0
10 200 13 4 9 9 1 0
TOTAL PI 1000 37 12 19 15 4 4
Figure 2 Assessements of the motility of infective larvae. The movements were assessed on recordings made of larvae in the subcutaneous tissue of vaccinated (Vacc) or primary infected (PI) mice and were scored on a discrete scale from 0 (non-motile) to 5 (maximum motility seen in larvae released in media only).
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
SAB was responsible for carrying out the laboratory experiments described in the manuscript, including mouse preparation, dissections, sample processing and recording and writing the manuscript. TA and PNV carried out the histological studies. LL recorded the video sequences presented as additional files and JCG provided the microscope, software and hardware and his expertise for filming the preparations. OB is the Principal Investigator on the project, and was responsible for obtaining grant support for the project, for the experimental design, the data analysis and the preparation of the manuscript.
Acknowledgements
This study was supported by the European Community grant ICA4-CT-1999-10002.
Supplementary Material
Additional File 1
Infective larva extracted from a lymph node crushed in RPMI media. The larva's movements are rapid and uncoordinated, as it has no grip on its immediate environment. Many cells (mostly lymphocytes) are present in the media around the worm. Magnification: × 100.
Click here for file
Additional File 2
The infective larva in subcutaneous tissue has a well-coordinated, large-amplitude S-shaped movement by which it leans on the surrounding tissue to progress forward. The fibrous structure of the subcutaneous tissue is visible, with many infiltrated cells (mainly neutrophils). Magnification: × 200
Click here for file
Additional File 3
Larva from a vaccinated mouse: it has reduced movements with low amplitude oscillations and a slow forward progression. The tissue has a dense appearance, with many infiltrated cells some of which can be seen right against the cuticle of the worm. The end of the sequence shows the absence of head side-to-side oscillations. Magnification: × 100 and × 500.
Click here for file
Additional File 4
Detailed recording of the head of an infective larva from a vaccinated mouse. The head is only slightly oscillating from side-to-side, and is trapped in a dense gel-like structure with few discrete elements. This larva will certainly not have reached a lymphatic vessel. Magnification: × 500.
Click here for file
Additional File 5
Example of how cells and subcutaneous fibrous structures attach to the cuticle of infective larvae. In this preparation, the larva was nearly freed into the dissection media, but remained attached by granulomatous formations at two points, near the head and around the last third of the body. Its movements are rapid, nearly frantic. Magnification: × 200.
Click here for file
==== Refs
Storey DM Al-Mukhtar AS Vaccination of jirds, Meriones unguiculatus, against Litomosoides carinii and Brugia pahangi using irradiated larvae of L. carinii Tropenmed Parasitol 1982 33 23 24 7101438
Hayashi Y Noda K Shirasaka A Nogami S Nakamura M Vaccination of BALB/c mice against Brugia malayi and B. pahangi with larvae attenuated by gamma irradiation Jpn J Exp Med 1984 54 177 181 6513101
Lucius R Textor G Kern A Kirsten C Acanthocheilonema viteae: vaccination of jirds with irradiation- attenuated stage-3 larvae and with exported larval antigens Exp Parasitol 1991 73 184 196 1889473 10.1016/0014-4894(91)90021-N
Bancroft A Devaney E The analysis of the humoral response of the BALB/c mouse immunized with radiation attenuated third stage larvae of Brugia pahangi Parasite Immunol 1993 15 153 162 8316409
Devaney E Bancroft A Egan A The effect of irradiation on the third stage larvae of Brugia pahangi Parasite Immunol 1993 15 423 427 8414645
Lange AM Yutanawiboonchai W Scott P Abraham D IL-4- and IL-5-dependent protective immunity to Onchocerca volvulus infective larvae in BALB/cBYJ mice J Immunol 1994 153 205 211 8207236
Taylor MJ van Es RP Shay K Folkard SG Townson S Bianco AE Protective immunity against Onchocerca volvulus and O. lienalis infective larvae in mice Trop Med Parasitol 1994 45 17 23 8066376
Le Goff L Loke P Ali HF Taylor DW Allen JE Interleukin-5 is essential for vaccine-mediated immunity but not innate resistance to a filarial parasite Infect Immun 2000 68 2513 2517 10768938 10.1128/IAI.68.5.2513-2517.2000
Petit G Diagne M Marechal P Owen D Taylor D Bain O Maturation of the filaria Litomosoides sigmodontis in BALB/c mice; comparative susceptibility of nine other inbred strains Ann Parasitol Hum Comp 1992 67 144 150 1295407
Le Goff L Maréchal P Petit G Taylor DW Hoffmann W Bain O Early reduction of the challenge recovery rate following immunization with irradiated infective larvae in a filaria mouse system Trop Med Int Health 1997 2 1170 1174 9438473 10.1046/j.1365-3156.1997.d01-218.x
Martin C Saeftel M Vuong PN Babayan S Fischer K Bain O Hoerauf A B-cell deficiency suppresses vaccine-induced protection against murine filariasis but does not increase the recovery rate for primary infection Infect Immun 2001 69 7067 7073 11598082 10.1128/IAI.69.11.7067-7073.2001
Bain O Babayan S Behaviour of filariae: morphological and anatomical signatures of their life style within the arthropod and vertebrate hosts Filaria J 2003 2 16 14675490 10.1186/1475-2883-2-16
Le Goff L Martin C Oswald IP Vuong PN Petit G Ungeheuer MN Bain O Parasitology and immunology of mice vaccinated with irradiated Litomosoides sigmodontis larvae Parasitology 2000 120 271 280 10759085 10.1017/S0031182099005533
Martin C Al-Qaoud KM Ungeheuer MN Paehle K Vuong PN Bain O Fleischer B Hoerauf A IL-5 is essential for vaccine-induced protection and for resolution of primary infection in murine filariasis Med Microbiol Immunol (Berl) 2000 189 67 74 11138639
Clark RA An JQ Greiling D Khan A Schwarzbauer JE Fibroblast migration on fibronectin requires three distinct functional domains J Invest Dermatol 2003 121 695 705 14632184 10.1046/j.1523-1747.2003.12484.x
Stein MP Mold C Du Clos TW C-reactive protein binding to murine leukocytes requires Fc gamma receptors J Immunol 2000 164 1514 1520 10640769
Suankratay C Mold C Zhang Y Potempa LA Lint TF Gewurz H Complement regulation in innate immunity and the acute-phase response: inhibition of mannan-binding lectin-initiated complement cytolysis by C-reactive protein (CRP) Clin Exp Immunol 1998 113 353 359 9737662 10.1046/j.1365-2249.1998.00663.x
Shipley JM Wesselschmidt RL Kobayashi DK Ley TJ Shapiro SD Metalloelastase is required for macrophage-mediated proteolysis and matrix invasion in mice Proc Natl Acad Sci U S A 1996 93 3942 3946 8632994 10.1073/pnas.93.9.3942
Okada S Kita H George TJ Gleich GJ Leiferman KM Migration of eosinophils through basement membrane components in vitro: role of matrix metalloproteinase-9 Am J Respir Cell Mol Biol 1997 17 519 528 9376127
Faveeuw C Preece G Ager A Transendothelial migration of lymphocytes across high endothelial venules into lymph nodes is affected by metalloproteinases Blood 2001 98 688 695 11468168 10.1182/blood.V98.3.688
Sandler NG Mentink-Kane MM Cheever AW Wynn TA Global gene expression profiles during acute pathogen-induced pulmonary inflammation reveal divergent roles for Th1 and Th2 responses in tissue repair J Immunol 2003 171 3655 3667 14500663
Appleton JA Kennedy MW, Harnett W New insights into the intestinal niche of Trichinella spiralis Parasitic nematodes: molecular biology, biochemistry and immunology 2001 CABI Publishing 103 120
Fioretti L Porter A Haydock PJ Curtis R Monoclonal antibodies reactive with secreted-excreted products from the amphids and the cuticle surface of Globodera pallida affect nematode movement and delay invasion of potato roots Int J Parasitol 2002 32 1709 1718 12464417 10.1016/S0020-7519(02)00178-9
Diagne M Petit G Liot P Cabaret J Bain O The filaria Litomosoides galizai in mites; microfilarial distribution in the host and regulation of the transmission Ann Parasitol Hum Comp 1990 65 193 199 2085265
| 15916708 | PMC1177976 | CC BY | 2021-01-04 16:38:24 | no | Filaria J. 2005 May 25; 4:3 | utf-8 | Filaria J | 2,005 | 10.1186/1475-2883-4-3 | oa_comm |
==== Front
Front ZoolFrontiers in Zoology1742-9994BioMed Central London 1742-9994-2-81592705010.1186/1742-9994-2-8ResearchImmune response to sympatric and allopatric parasites in a snail-trematode interaction Osnas Erik E [email protected] Curtis M [email protected] Department of Biology, Indiana University, 1001 E. Third Street, Bloomington, IN 47405, USA2 Department of Wildlife Ecology, University of Wisconsin, 209 Russell Labs, 1630 Linden Drive, Madison, WI 53706, USA2005 31 5 2005 2 8 8 9 3 2005 31 5 2005 Copyright © 2005 Osnas and Lively; licensee BioMed Central Ltd.2005Osnas and Lively; 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 outcome of parasite exposure depends on the (1) genetic specificity of the interaction, (2) induction of host defenses, and (3) parasite counter defenses. We studied both the genetic specificity for infection and the specificity for the host-defense response in a snail-trematode interaction (Potamopyrgus antipodarum-Microphallus sp.) by conducting a reciprocal cross-infection experiment between two populations of host and parasite.
Results
We found that infection was greater in sympatric host-parasite combinations. We also found that the host-defense response (hemocyte concentration) was induced by parasite exposure, but the response did not increase with increased parasite dose nor did it depend on parasite source, host source, or host-parasite combination.
Conclusion
The results are consistent with a genetically specific host-parasite interaction, but inconsistent with a general arms-race type interaction where allocation to defense is the main determinant of host resistance.
coevolutionlocal adaptationimmune defensetrematode
==== Body
Background
Studies of host-parasite interactions can be seen as split between two different approaches [1,2]. One approach tends to emphasize the induction and cost of defense against parasites, while the other approach tends to emphasize the genetic basis and specificity required for successful infection. Both avenues have been productive, but they need not be seen as mutually exclusive [1-3]. The induction of an immune defense, for example, might be required to eliminate parasites once detected by the host [4]. On the other hand, the relative effectiveness of host defense is expected to decline as the diversity of parasite genotypes increases, provided these genotypes show a high degree of host specificity [1]. Understanding the relative importance and possible interactions between immune defense and genotypic specificity requires that both aspects are studied simultaneously in the same system. For example, Kurtz et al. [5] showed that the immune response of grasshoppers was reduced in foreign environments, even though body mass, a measure of general condition, was not reduced. They interpreted this to mean that the grasshoppers require less immune defense in the face of foreign, and presumably locally adapted, parasites. In order to more fully address the interaction between genetic specificity and immune defense, one could first test for local adaptation, and then measure the immune response in both sympatric and allopatric host-parasite combinations.
In the present study, we exposed host snails from two lake populations to two different doses of eggs produced by either sympatric or allopatric trematodes in a reciprocal cross-infection experiment. We then compared the two different egg doses to each other and to no-egg controls to determine whether the host immune system could be induced, and, if so, whether it depended on parasite dose and/or source (i.e., sympatric vs. allopatric). We found that the immune system could be induced to increase the number of hemocytes, but that the induction did not depend on dose or the source of parasites. Nonetheless, the prevalence of infection was greater for sympatric combinations of host and parasite. Thus the induction of defense was not as effective at combating the apparently co-evolved parasites as it was in combating the remote sources of parasites.
Results
There were no significant host-source or parasite-source main effects on parasite prevalence, but a significant main effect of parasite dose was observed (F = 136.08, d.f. = 1, 40, P < 0.001; Table 1; Fig. 1). The two-way interaction between host source and parasite source was significant (F = 213.66, d.f. = 1, 40, P < 0.001), as was the three-way interaction between host source, parasite source, and parasite dose (F = 120.09, d.f. = 1, 40, P < 0.001; Table 1; Fig. 1). No other interactions were significant. The form of the two-way interaction between host and parasite is of the crossing type, indicating that parasites produced higher prevalence of infection in sympatric hosts than in allopatric hosts (Fig. 1a and 1b). The significant three-way interaction between dose, host, and parasite indicates that the strength of the two-way interaction (local adaptation of the parasite) increased with dose (Fig. 1a vs. 1b).
Table 1 Fixed-effect analysis of variance for prevalence of infection with parasite source, host source, and parasite dose as factors. R-squared for the model was 0.92.
Source s.s. d.f. m.s. F P
Parasite source 0.002 1 0.002 2.45 0.125
Host source 0.001 1 0.001 0.96 0.333
Dose 0.133 1 0.133 136.08 <0.001
Parasite * Host 0.210 1 0.210 213.66 <0.001
Parasite * Dose <0.001 1 <0.001 0.12 0.727
Host * Dose <0.001 1 <0.001 0.24 0.629
Parasite * Host * Dose 0.118 1 0.118 120.09 <0.001
Error 0.039 40 0.001
Total 0.782 48
Figure 1 Mean prevalence Microphallus infection (1 s.e.) of the snail host P. antipodarum in relation to parasite source, host source, and dose of parasite: (a) low dose, and (b) high dose. Vertical standard error bars were estimated from the mean squared error in the ANOVA in Table 1.
There was a significant effect of parasite exposure on hemocyte count (F = 3.71, d.f. = 2, 57, P = 0.03; Table 2; Fig. 2). Hemocyte count increased with exposure to parasites (t = 2.66, P = 0.01), but there was not a significant difference between low and high doses of parasites (t = 0.66, P = 0.55; Table 2; Fig. 2). When the controls (no exposure) were deleted and the data set was analyzed as a three-factor experiment, there was only a marginally significant effect of parasite source on hemocyte count (F = 3.69, d.f. = 1, 40, P = 0.06); no other effects or interactions were significant (Fig. 3, Table 3). This marginally significant result suggests that parasites from L. Alexandrina tended to induce a slightly higher hemocyte count than did parasites from L. Mapourika (Fig. 3). Observed power estimates for the effects of dose, parasite source, host source, and the interaction between host and parasite were low (0.08, 0.48, 0.27, and 0.12, respectively). Observed power estimates for the other interactions were also low (< 0.08). However, observed effect sizes were also small (<10%) for all effects. When larger effect sizes were used to estimate power (10% and 20%), the power to detect an effect was moderately high (0.55 and 0.88, respectively). Therefore, it is likely that the true effect sizes were less than 10%.
Table 2 One-way analysis of variance for average hemocyte count across three parasite dose treatments (zero, low, and high). The first contrast tests for a difference between exposed (low and high dose) and not exposed treatments (zero dose), and the second contrast tests for a difference between low and high dose treatments.
s.s. d.f. m.s. F p
Between Groups 0.496 2 0.248 3.705 0.031
Within Groups 3.819 57 0.067
Total 4.316 59
Contrast t d.f. p
Control vs. Exposed 2.66 57 0.010
Low dose vs. High dose 0.600 57 0.551
Figure 2 Mean hemocyte count (1 s.e.) in 0.1 μl hemolymph of the snail host P. antipodarum in relation to dose of parasite. Vertical standard error bars were estimated from the mean squared error in the ANOVA in Table 2.
Figure 3 Mean hemocyte count (1 s.e.) in 0.1 μl hemolymph of the snail host P. antipodarum in relation to parasite source, host source, and dose of parasite: (a) low dose, and (b) high dose. Vertical standard error bars were estimated from the mean squared error in the ANOVA in Table 3
Table 3 Fixed-effect analysis of variance for hemocyte count with parasite source, host source, and parasite dose as factors. R-squared for the model was 0.15.
Source s.s. d.f. m.s. F P
Parasite source 0.257 1 0.257 3.690 0.062
Host source 0.138 1 0.138 1.979 0.167
Dose 0.024 1 0.024 0.347 0.559
Parasite * Host 0.043 1 0.043 0.618 0.436
Parasite * Dose 0.020 1 0.020 0.292 0.592
Host * Dose <0.001 1 <0.001 0.000 0.989
Parasite * Host * Dose 0.006 1 0.006 0.090 0.766
Error 2.787 40 0.070
Total 156.99 48
Discussion
The results suggest that sympatric, locally adapted parasites (Microphallus sp.) induce a similar defense response (as measured by hemocyte concentration) in P. antipodarum snails as do allopatric parasites. Although there was a marginally significant main effect for parasite source, there was clearly no host-by-parasite interaction effect on hemocyte count (Fig. 2, Table 2). In contrast, however, the interaction effect was highly significant for infection success; but there were no significant main effects for either host source or parasite source (Fig. 1, Table 1). Therefore, even though the defense was induced to the same degree by both sympatric and allopatric parasites, the defense was less effective against the locally adapted sympatric parasites. One implication of this result is that the snails cannot increase resistance to infection by coevolved trematodes by increasing the allocation to defense.
Hemocyte concentration is only a rough measure of the immune response. However, hemocyte concentration did show a response to parasite exposure in our study (Fig. 2), and defense mechanisms in other snail-trematode systems are mediated through hemocytes [6,7]. For example, hemocytes encapsulate trematodes and excrete reactive oxygen molecules that are cytotoxic [6,8]. In addition, hemocytes up-regulate genes when exposed to trematodes [9,10]. More detailed observation on the reactions of hemocytes in future studies may lead to different results. However, there was no indication that a sympatric or allopatric source of parasites changed the induction of hemocytes in our study (Fig. 3), as was found in Kurtz et al. [6] for grasshoppers.
Usually, allocation of resources to defense mechanisms has been found to be costly in terms of survival or reproduction [11-16], and it has been assumed that hosts are trading-off fitness for protection against the negative effects of infection. However, when coevolution changes the effectiveness of the host defense, predictions about trade-offs between parasite defense and other fitness components are more complicated [1]. If local adaptation by parasites decreases the probability that hosts will mount a successful defense, then the optimal defense strategy for the host will depend on both the cost and effectiveness of the defense as well as the virulence of the parasite [1]. In the present study, costs of defense were not measured, but a strong defense response was observed (Fig. 2). Previous work in this system has shown that growth and survival did not decrease with increasing parasite exposure [17], so it is likely that the defense response is not costly or that the cost is small. On the other hand, resources in the lab environment may have been sufficient to compensate for the increased demands of the defense resource. This may have obscured the detection of a difference between host-parasite combinations in the hemocytes response.
If defense is not costly, then mounting a defense would be advantageous whenever the host encounters parasites regardless of the effectiveness of the defense mechanisms. Nonetheless, defense might only be effective against some proportion of the parasite population, as would be expected if there is tight genetic specificity for infection. In such a situation, some level of defense allocation is required to defend against some parasite genotypes, but no level of allocation can defend against other parasite genotypes. This later alternative seems likely given that there is good evidence for local adaptation in this system (Fig. 1) [18-21] and the fact that both allopatric and sympatric combinations induced the same defense response (Fig. 3).
In the present study, hosts seemed to increase defense, perhaps to a maximum, following exposure to a very low dose of parasite eggs, yet they did not increase defense at higher dose (Fig. 2). Therefore, even small exposures may be a sufficient cue that increases defense allocation in environments with a high risk of parasitism. One might expect a dose effect under the assumption that hosts would increase defense levels as more and more parasites were encountered, especially in sympatric host-parasite combinations that are more difficult to defend against. However, there was no effect of parasite dose or any higher order interactions involving dose, host source, or parasite source on hemocyte induction (Fig. 3, Table 3). There was a marginally significant parasite main effect for hemocyte induction (Table 3), meaning that some parasite populations might induce greater hemocyte responses than others. Potentially, coevolution may have increased counter-defenses in some trematode populations; thus, larger allocations to defense are required in the host when confronted with these parasites. Counter-defenses of parasites are known to occur in many systems [22,23]. For example, interference of host hemocyte function has been shown in echinostomes (Trematoda) [24-26], and parasitoids of Drosophila bury into the fat body in order to avoid circulating hemocytes [27].
Conclusion
The interaction is highly specific for infection, but not for defense response as measured by hemocyte concentration. These results are most consistent with a highly specific genetic matching mechanism where successful parasites evade or interfere with defenses of specific host genotypes. There was also no effect of parasite dose on hemocyte concentration after initial exposure, which implies that defense needs do not increase with additional parasite exposures. Although measuring specific aspects of the hemocyte reactions may lead to other results, local adaptation of the parasite to sympatric hosts does not seem to produce variation in the induction of hemocyte concentration across sympatric and allopatric exposures.
Methods
Host source
Snail hosts (Potamopyrgus antipodarum) were collected from the Isoetes habitat at the "Camp" site at Lake Alexandrina, New Zealand, [28] by sweeping a net through the underwater vegetation. Snails from Lake Mapourika were collected by the same methods. Previous studies have shown that parasites from these same two lakes are adapted to infect their sympatric host populations [19]. In addition, in a meta-analysis of multiple studies on this system, we found that 32 of 33 parasite populations (in either space or time) were more successful at infecting local snails than snails collected from allopatric locations [21]. Thus, the pattern of parasite adaptation to local host populations appears to be extremely robust in space and time. Lake Alexandrina is on the eastern side of the Southern Alps at about 750 m above sea level, and Lake Mapourika is on the western side of the Southern Alps at 75 m above sea level [29] on the South Island of New Zealand. Hosts from both sources were collected in January 2002 and transported under permit to the lab at Indiana University in cool and wet paper towels. In the lab the snails were kept in large aquaria and fed Spirulina algae until the experiment began on 2 March 2002. During the experimental trials, each experimental unit consisted of 100 snails kept in a 2 l plastic container. The snails were fed Spirulina algae 1–2 times per week and the water in the containers was changed once a week.
Parasite system and culture
Microphallus sp. (Trematoda) is a widespread and locally common undescribed parasite in New Zealand lakes and streams [28,30]. Multilocus allozyme genotype data show that Microphallus is a single outbred species with high levels of gene flow among South Island populations [31]. The parasite exclusively uses P. antipodarum as the intermediate host, and the final hosts are waterfowl. Embryonated Microphallus eggs are ingested from sediment and hatch in the snail's gut, penetrate the intestine, and migrate to the gonads and digestive gland. Following successful establishment, the parasite then undergoes asexual reproduction, replacing much of the host's reproductive tissue and digestive gland, which results in complete sterilization of the snail. The first visible parasite developmental stages (blastocercariae) are detectable after approximately 75 days post-exposure and metacercariae are common by 90 days post-exposure at 16°C in the lab. The life cycle is completed when snails containing metacercariae are consumed by waterfowl.
Although the natural host for the parasite used in this experiment are birds, the parasite is successfully cultured in mouse [32]. To produce parasite eggs that are infective to snails, three mice were each fed 35 Microphallus-infected snails collected from Lake Alexandrina; and similarly, three mice were fed 35 Microphallus-infected snails collected from Lake Mapourika. Three additional mice were not fed parasites, and they were used to provide parasite-free feces in one of our control treatments (see below). Starting 24 hours after feeding, the litter in the mice cages was changed, and fecal pellets were collected three times each day for four days. After collection, the control feces and parasite feces were separately placed in fresh pond water in order to store the parasite eggs and dilute soluble waste. The pond water was aerated and changed twice a day.
Experimental design and procedures
A factorial design was used to study the effects of host source, parasite source, and parasite dose on infection and hemocyte concentration. Thirty 2 l containers each housed 100 snails from Lake Alexandrina, and thirty 2 l containers each housed 100 snails from Lake Mapourika. Within each host source, six containers were given a low dose of Alexandrina parasites, six containers were given a high dose of Alexandrina parasites, six containers were given a low dose of Mapourika parasites, and six containers were given a high dose of Mapourika parasites ("high" and "low" dose were as defined below). The remaining six containers for each host source received no parasites – three containers received mouse feces without parasite eggs and three containers received no parasites or feces. Because there was no difference in hemocyte counts between the no-parasite treatments with feces and the no-parasite treatment without feces (F = 0.001, d.f. = 1, 10, P = 0.975), these treatments were pooled in all analyses below. This result shows that the contents of the mouse feces were, by themselves, not sufficient to increase the hemocyte counts of snails.
Parasite doses were determined by separately condensing the mixtures of mouse feces and water to 1 l volume. The mixture was then split into 100 ml (low dose) and 900 ml (high dose) amounts and then each was diluted up to 1 l. A serological pipette was then used to administer mouse feces equally among containers for each dose treatment. Thus the high dose was expected to have 9 times as many parasite eggs as the low dose. The feces from control mice were handled in the same way, and they were distributed so that snails in the parasite control with feces and low-dose treatments were exposed to the same amount of mouse feces as the high-dose experimental treatments.
Hemocyte quantification
Snails were sampled for hemolymph on days 13 and14 post-exposure. From previous experiments (E. Osnas, unpublished data), it was known that hemocyte concentration increases after parasite exposure by this time. At 14 days post-exposure the parasite is visually undetectable in the snail (see above). To collect hemolymph, three individuals were randomly sub-sampled from each experimental container. For each snail, hemolymph was collected by tapping on the operculum until hemolymph was expelled through the hemal pore [13,33,34]. A 10 μl pipette was then used to draw up the hemolymph and transfer it to a hemocytometer. In P. antipodarum, this technique results in approximately 1–2 μl of hemolymph. Hemocytes in 0.1 μl of the hemocytometer grid surface were then visually counted using a compound microscope at 400 total magnification.
Parasite sampling and statistical analysis
After 90 days post-exposure, snails were examined for infection by dissection under a microscope. At 90 days post-exposure, parasites can be easily detected using a dissecting microscope, but they have not reached the fully mature metacercariae stage. Parasite species and developmental stage were recorded [17,35]. Snails with infections of species other than Microphallus were excluded from the analysis. Similarly, snails infected by mature Microphallus metacercariae were also deleted from the analysis because they were assumed to be already infected when collected in the field [17,36].
For each container, we calculated the percent infection (prevalence) of the snails sampled after 90 days post-exposure, and we calculated the average hemocyte count for the three individuals sampled after 13 days post-exposure. These means were then used as independent data points for statistical analysis. We used a three-factor fixed effect analysis of variance with host source, parasite source, and parasite dose as independent factors to test for local adaptation by the parasite. For this analysis, the control treatments (no parasites) were excluded. Local adaptation of the parasite to the host is indicated by a significant crossing-type interaction between host and parasite source.
We used one-way analysis of variance in order to test for a change in hemocyte concentration with parasite exposure. For this analysis, dose was the independent variable and log-base 10 of hemocyte count was the dependant variable. We used three levels of dose: no exposure, low dose, and high dose, as describe above. We then used independent contrasts to test for an overall increase in hemocyte count with exposure (control verses low and high) and for a difference between low and high exposure levels.
To test for an effect of parasite source, host source, and dose on hemocyte count, we used a full factorial three-factor fixed effect model analysis of variance, just as above for infection. For this analysis, the control treatments (no parasites) were also removed from the data set. This produced a balanced design. All analyses were done in SPSS 12.0 (2003).
Authors' contributions
EEO and CML developed the concept and experimental design for this study. EEO preformed the experiment, collected the data, and wrote a first draft of the manuscript. CML helped perform the experiment and substantially revised the manuscript. Both authors read and approved of the final manuscript.
Acknowledgements
D. Zysling and members of the Lively Lab gave helpful comments on a previous version of the manuscript. This research was funded by grants from the Center for the Integrative Study of Animal Behavior at Indiana University to EEO and by NSF grants DEB-9904840 and DEB-0128510 to CML.
==== Refs
Jokela J Schmid-Hempel P Rigby MC Dr. Pangloss restrained by the Red Queen - steps towards a unified defense theory Oikos 2000 89 267 274 10.1034/j.1600-0706.2000.890207.x
Schmid-Hempel P Ebert D On the evolutionary ecology of specific immune defense Trends Ecol Evol 2003 18 27 32 10.1016/S0169-5347(02)00013-7
Rigby MC Hechinger RF Stevens L Why should parasite resistance be costly? Trends in Parasitology 2002 18 116 120 11854088 10.1016/S1471-4922(01)02203-6
Agrawal AF Lively CM Modelling infection as a two-step process combining gene-for- gene and matching-allele genetics Proc R Soc Lond Ser B-Biol Sci 2003 270 323 334 10.1098/rspb.2002.2193
Kurtz J Klappert K Schneider W Reinhold K Immune defense, dispersal and local adaptation Evol Ecol Res 2002 4 431 439
Bayne CJ Hahn UK Bender RC Mechanisms of molluscan host resistance and of parasite strategies for survival Parasitology 2001 123 S159 S167 11769280 10.1017/S0031182001008137
Loker ES Bayne CJ Molecular studies of the molluscan response to digenean infection Phylogenetic Perspectives on the Vertebrate Immune System Advances in Experimental Medicine and Biology 2001 484 209 222
Zelck UE Von Janowsky B Antioxidant enzymes in intramolluscan Schistosoma mansoni and ROS-induced changes in expression Parasitology 2004 128 493 501 15180317 10.1017/S0031182004004895
Zhang SM Loker ES The FREP gene family in the snail Biomphalaria glabrata: additional members, and evidence consistent with alternative splicing and FREP retrosequences Developmental and Comparative Immunology 2003 27 175 187 12590969 10.1016/S0145-305X(02)00091-5
Zhang SM Loker ES Representation of an immune responsive gene family encoding fibrinogen-related proteins in the freshwater mollusc Biomphalaria glabrata, an intermediate host for Schistosoma mansoni Gene 2004 341 255 266 15474308 10.1016/j.gene.2004.07.003
Langand J Jourdane J Coustau C Delay B Morand S Cost of resistance, expressed as a delayed maturity, detected in the host-parasite system Biomphalaria glabrata Echinostoma caproni Heredity 1998 80 320 325 10.1038/sj.hdy.6882910
Raberg L Nilsson JA Ilmonen P Stjernman M Hasselquist D The cost of an immune response: vaccination reduces parental effort Ecol Lett 2000 3 382 386 10.1046/j.1461-0248.2000.00154.x
Rigby MC Jokela J Predator avoidance and immune defense: costs and trade-offs in snails Proc R Soc Lond Ser B-Biol Sci 2000 267 171 176 10.1098/rspb.2000.0983
Fellowes MDE Kraaijeveld AR Godfray HCJ Trade-off associated with selection for increased ability to resist parasitoid attack in Drosophila melanogaster Proc R Soc Lond Ser B-Biol Sci 1998 265 1553 1558 10.1098/rspb.1998.0471
Kraaijeveld AR Godfray HCJ Trade-off between parasitoid resistance and larval competitive ability in Drosophila melanogaster Nature 1997 389 278 280 9305840 10.1038/38483
Moret Y Schmid-Hempel P Survival for immunity: The price of immune system activation for bumblebee workers Science 2000 290 1166 1168 11073456 10.1126/science.290.5494.1166
Osnas EE Lively CM Parasite dose, prevalence of infection and local adaptation in a host-parasite system Parasitology 2004 128 223 228 15030009 10.1017/S0031182003004360
Dybdahl MF Lively CM Host-parasite coevolution: evidence for rare advantage and time-lagged selection in a natural population Evolution 1998 52 1057 1066
Lively CM Adaptation by a parasitic trematode to local-populations of its snail host Evolution 1989 43 1663 1671
Lively CM Dybdahl MF Parasite adaptation to locally common host genotypes Nature 2000 405 679 681 10864323 10.1038/35015069
Lively CM Dybdahl MF Jokela J Osnas EE Delph LF Host sex and local adaptation by parasites in a snail-trematode interaction American Naturalist 2004 164 S6 S18 15540142 10.1086/424605
Loker ES On being a parasite in an invertebrate host - a short survival course J Parasitol 1994 80 728 747 7931908
Vanderknaap WPW Loker ES Immune mechanisms in trematode-snail interactions Parasitol Today 1990 6 175 182 15463334 10.1016/0169-4758(90)90349-9
Humbert E Coustau C Refractoriness of host haemocytes to parasite immunosuppressive factors as a putative resistance mechanism in the Biomphalaria glabrata-Echinostoma caproni system Parasitology 2001 122 651 660 11444618 10.1017/S003118200100782X
DeGaffe G Loker ES Susceptibility of Biomphalaria glabrata to infection with Echinostoma paraensei: correlation with the effect of parasite secretory-excretory products on host hemocyte spreading Journal of Invertebrate Pathology 1998 71 64 72 9446739 10.1006/jipa.1997.4710
Loker ES Cimino DF Hertel LA Excretory-secretory products of Echinostoma paraensei sporocysts mediate interference with Biomphalaria glabrata hemocyte functions Journal of Parasitology 1992 78 104 115 1738052
Kraaijeveld AR Hutcheson KA Limentani EC Godfray HCJ Costs of counterdefenses to host resistance in a parasitoid of Drosophila Evolution 2001 55 1815 1821 11681736
Jokela J Lively CM Spatial variation in infection by digenetic trematodes in a population of fresh-water snails (Potamopyrgus antipodarum) Oecologia 1995 103 509 517 10.1007/BF00328690
Livingston ME Biggs BJ Gifford JS Inventory of New Zealand lakes Part II South Island 1986 , Water and Soil Directorate, Ministry of Works and Development. Wellington North, New Zealand
Lively CM Evidence from a New-Zealand snail for the maintenance of sex by parasitism Nature 1987 328 519 521 10.1038/328519a0
Dybdahl MF Lively CM The geography of coevolution: comparative population structures for a snail and its trematode parasite Evolution 1996 50 2264 2275
Lively CM McKenzie JC Experimental infection of a freshwater snail, Potamopyrgus antipodarum, with a digenetic trematode, Microphallus sp. New Zealand Journal of Zoology 1991 18 59 62
Amen RI Tijnagel JMG Vanderknaap WPW Meuleman EA Delangedeklerk ESM Sminia T Effects of Trichobilharzia ocellata on hemocytes of Lymnaea stagnalis Dev Comp Immunol 1991 15 105 115 2060680 10.1016/0145-305X(91)90001-F
Russo J Lagadic L Effects of parasitism and pesticide exposure on characteristics and functions of hemocyte populations in the freshwater snail Lymnaea palustris (Gastropoda, Pulmonata) Cell Biol Toxicol 2000 16 15 30 10890503 10.1023/A:1007640519746
Krist AC Lively CM Levri EP Jokela J Spatial variation in susceptibility to infection in a snail- trematode interaction Parasitology 2000 121 395 401 11072902 10.1017/S0031182099006538
Krist AC Effect of the digenean parasite Proterometra macrostoma on host morphology in the freshwater snail Elimia livescens J Parasitol 2000 86 262 267 10780543
| 15927050 | PMC1177977 | CC BY | 2021-01-04 16:38:34 | no | Front Zool. 2005 May 31; 2:8 | utf-8 | Front Zool | 2,005 | 10.1186/1742-9994-2-8 | oa_comm |
==== Front
Front ZoolFrontiers in Zoology1742-9994BioMed Central London 1742-9994-2-91593874910.1186/1742-9994-2-9ReviewProspects for surviving climate change in Antarctic aquatic species Peck Lloyd S [email protected] Natural Environment Research Council, British Antarctic Survey, High Cross, Madingley Road, Cambridge CB3 0ET, UK2005 6 6 2005 2 9 9 11 2 2005 6 6 2005 Copyright © 2005 Peck; licensee BioMed Central Ltd.2005Peck; 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.
Maritime Antarctic freshwater habitats are amongst the fastest changing environments on Earth. Temperatures have risen around 1°C and ice cover has dramatically decreased in 15 years. Few animal species inhabit these sites, but the fairy shrimp Branchinecta gaini typifies those that do. This species survives up to 25°C daily temperature fluctuations in summer and passes winter as eggs at temperatures down to -25°C. Its annual temperature envelope is, therefore around 50°C. This is typical of Antarctic terrestrial species, which exhibit great physiological flexibility in coping with temperature fluctuations. The rapidly changing conditions in the Maritime Antarctic are enhancing fitness in these species by increasing the time available for feeding, growth and reproduction, as well as increasing productivity in lakes. The future problem these animals face is via displacement by alien species from lower latitudes. Such invasions are now well documented from sub-Antarctic sites.
In contrast the marine Antarctic environment has very stable temperatures. However, seasonality is intense with very short summers and long winter periods of low to no algal productivity. Marine animals grow slowly, have long generation times, low metabolic rates and low levels of activity. They also die at temperatures between +5°C and +10°C. Failure of oxygen supply mechanisms and loss of aerobic scope defines upper temperature limits. As temperature rises, their ability to perform work declines rapidly before lethal limits are reached, such that 50% of populations of clams and limpets cannot perform essential activities at 2–3°C, and all scallops are incapable of swimming at 2°C. Currently there is little evidence of temperature change in Antarctic marine sites. Models predict average global sea temperatures will rise by around 2°C by 2100. Such a rise would take many Antarctic marine animals beyond their survival limits. Animals have 3 mechanisms for coping with change: they can 1) use physiological flexibility, 2) evolve new adaptations, 3) migrate to better sites. Antarctic marine species have poor physiological scopes, long generation times and live on a continent whose coastline covers fewer degrees of latitude than all others. On all 3 counts Antarctic marine species have poorer prospects than most large faunal groups elsewhere.
==== Body
Background
A current major focus for science is the prospect of global climate change. It is now internationally accepted that we are in a period of global change, that for some measures such as atmospheric carbon dioxide content, is faster than at any time in recent geological history [1,2]. Much effort has been put into producing models of predicted change [1,3,4], and identifying the evidence for change. These models and their outputs have had profound effects on the science conducted and the focus of research in both physical and biological disciplines.
Environmental change poses a range of problems for species that vary between site and species. Abilities to cope with change differ from individual to species levels. Research into abilities to cope has focussed primarily on temperature change, but other environmental factors including humidity, precipitation and windflow patterns on land and ocean currents and local salinity in the sea are also likely to be affected [1,2]. We are clearly still in the early stages of understanding how biotas around the world will or can respond to change, but such evaluations are essential for the optimum management or mitigation of deleterious effects, both for mankind and the environment.
In this review I evaluate the major characteristics of Antarctic species and contrast the threats and abilities to cope with change for aquatic organisms from terrestrial and marine environments. Emphasis is placed on current environmental variation, physiological flexibility (where flexibility refers to the ability to cope either within immediate tolerance limits or to acclimate over a period of weeks or months), and future threats. If environmental conditions deteriorate organisms can respond by: (1) coping within their physiological tolerance range, or within their ability to acclimate; (2) evolving new characters to allow survival; and (3) migrating to sites where conditions have remained or become favourable [5,6]. The review ends with a general overview of the possible responses that organisms can employ to survive change, and how Antarctic species compare with communities elsewhere in these characteristics.
Antarctic Freshwater Environments
More than 99.5% of continental Antarctica is covered by ice [7] that is, in places, over 4 km thick, and open areas are restricted to the continental margin and mountaintops. Even though there is significant melting of ice in some areas in summer, very few sites have habitable aquatic sites. However, in the maritime Antarctic, along the Antarctic peninsula and offshore islands, significant pools and lakes exist that are inhabited by microbial, algal and invertebrate communities, some of which can be substantial. These pools and lakes are usually seasonal, being frozen in winter, although some of the larger water masses only freeze at the surface. The extent of the period of open water, and temperatures that any of them experience in summer are a function of many factors including air temperature, altitude and aspect. Pools that sustain significant populations of fairy shrimps, Branchinecta gaini, occur near Rothera station on the Antarctic peninsula (Fig. 1). In these pools summer temperatures are positive, but fluctuate between 0°C and 20°C, whereas in winter temperatures fall to around -20°C, but are more stable. The fairy shrimps living there thus experience an annual temperature range of 40°C, or more, and in summer can experience daily temperature fluctuations of up to 20°C. Similar, or more extreme temperature regimes are common in maritime Antarctic terrestrial systems [8].
Figure 1 Water and adjacent air temperature from Jan to Dec 2002 inclusive for two pools containing dense populations of the fairy shrimp Branchinecta gaini on Anchorage Island, Ryder Bay, Adelaide Island (67°34'07"S, 68°07'30"W). Temperatures were logged at 5 min intervals, and are shown as weekly maxima (●) and minima (○). Time points indicated are mid-month.
Ecological conditions in maritime Antarctic (which includes coastal regions of the Antarctic peninsula and several offshore islands [9]) freshwater environments vary markedly over short distances, with nutrients such as orthophosphate and ammonia, as well as water alkalinity varying by more than an order of magnitude between adjacent lakes [10]. Algal communities also vary markedly between localities, with Chlorophyll a concentrations varying by factors of 5 or more between lakes on Signy Island [10]. These marked variations between pools and lakes over short distances are the product of a range of factors including: aspect effects on insolation; small variations in temperature; hydrology and erosion in the catchment area; and in-lake biological activity and biogeochemical cycling.
Thus over the majority of the Antarctic continent there are very few sites where aquatic systems are available for colonisation. However, at the continental margins and in the maritime Antarctic significant numbers of pools and lakes exist where biological communities are established, and these environments are generally both cold and frozen in winter and highly variable in ecological characters.
Physiological capacities of Antarctic Freshwater species
Species inhabiting pools and lakes in Antarctica exhibit great physiological flexibility in coping with variations in environmental temperature. B. gaini populations can survive temperatures up to 10°C for 1 week with no mortality, and temperatures up to 25°C for 48 h with only 50% mortality [11]. Oxygen consumption rates also rise in an expected fashion with temperature, doubling for every 10°C rise in temperature in this species, with rates at 20°C four times higher than those at 1°C [11]. This indicates a metabolic scope for B. gaini of at least × 4 (where a metabolic, or physiological scope refers to the maximum ability to raise a function over the minimum required for long-term survival). Ventilation rate on the other hand only increased by a factor of 2.2 between 1°C and 20°C. This mismatch between capacity to raise ventilation rate and oxygen consumption suggests that, although oxygen delivery does not appear limiting between 1°C and 20°C, the upper lethal limit may be dictated by an inability to raise oxygen supply mechanisms at a sufficient rate, as seen in marine ectotherms [12].
Summer temperature variations are survived as adults. Adults cannot, however, survive freezing, and winter periods are passed as eggs, or cysts, which remain unfrozen to temperatures around -25°C. Antarctic terrestrial species, predominantly Collembola and Acari survive even wider temperature fluctuations as adults, using antifreezes, cryoprotectants and supercooling to maintain viability at temperatures down to -30°C, or below [13]. These groups often endure annual temperature ranges in excess of 50°C [13]. Terrestrial invertebrates in Antarctica appear to have very wide physiological flexibility, and show cold adapted metabolic rates that are faster than metabolic rates of similar species from lower latitudes at any given temperature [14]. Growth rates are also reported to show some temperature compensation, and the optimum temperatures for feeding and growth are low [15].
Terrestrial and freshwater species, therefore, exhibit large physiological flexibility, giving them the ability to function normally at temperatures between 0°C and 15–20°C. They also possess adaptations allowing short-term survival in the range -25°C to +25°C for freshwater species and -30°C to +30°C for some terrestrial Acari and Collembola.
Antarctic marine environments
The Southern Ocean is characterised by low but highly stable temperatures. At the most variable sites, such as Signy Island in the maritime Antarctic, temperatures usually range between -1.8°C in winter and around +1.0°C in summer (Fig 2). At the highest Antarctic sites, typified by McMurdo Sound, temperatures range between-1.7°C and -2.0°C yearly [16], although recent data suggest temperatures may be slightly more variable than this at McMurdo, with temperatures approaching -0.5°C in some years [17]. Thus the total annual fluctuation in sea temperature rarely exceeds 3°C in the region of the Antarctic Peninsula and Scotia Sea and is half this or less at high Antarctic (continental coast), making this one of the most thermally stable environments on earth. Temperatures have been low and stable around Antarctica for a long evolutionary period, at least 10 million years.
Figure 2 Annual cycle of seawater temperature for two sites in Antarctica. (●) 10 m depth, Orwell Bight, Signy Island (10 year mean values); (○) McMurdo Sound. Figure updated from [16].
Other characters of the marine environment vary markedly with season around Antarctica. Annually light varies between no direct sunlight and overall radiative loss from the surface to 24 h direct sunlight and incident radiation levels at or above tropical values over a 24 period in summer. Between 10 and 15 million km2 of sea-ice forms in winter and melts in summer. These factors result in intense seasonality of phytoplankton productivity, especially in nearshore waters. Blooms around Signy Island, South Orkney Islands (60°43'S, 45°36'W) and Rothera Point, Adelaide Island (67°36'S, 68°12'W) are both restricted to approximately 2–3 month periods and at peak reach chlorophyll standing stock levels in excess of 25 mg Chl a m-3 [18,19], values that would be high for any site on earth.
In the open ocean productivity is often associated with the edge of the sea-ice, and a significant portion of overall oceanic productivity occurs in these areas [20]. Micro-algal productivity associated with sea-ice can also be extensive [21], as can blooms of benthic communities on the seabed [22]. When taken together these 3 areas of productivity can be a more significant source of resource supply to secondary consumers than open ocean blooms in the Southern Ocean.
Physiological capacities of Antarctic marine species
Antarctic marine ectotherms are characterised by low rates of growth, development, metabolism and activity with little or no evidence of temperature compensation [23,24] (Fig. 3). In general growth rates are 2–5 times slower in Antarctic species compared to similar temperate organisms, and larval development is 5–10 times slower, with echinoderm embryos taking up to 150 h from fertilisation to hatching in Antarctica [23,25] (Fig. 4). The metabolic rate of adult animals is lowered by a factor of 2 to 3 times for every 10°C drop in temperature compared with lower latitude species [26,27], and metabolic rates of larvae may be even more affected by temperature than adult animals, being around 10 times lower in larvae of Antarctic gastropod molluscs than temperate species [28].
Figure 3 Rates of accomplishing various activities for a range of Antarctic marine invertebrates compared with related or ecologically similar temperate species. The hatched line indicates representative rates for temperate species and is set at a value of 1. Boxes show the range of values, with the midline indicating the mean. Figure reproduced from [24].
Figure 4 Time to hatching for echinoderm embryos from Antarctic, temperate and tropical sites. □ = temperate and tropical echinoderms [25]; ● = Antarctic species [25]; ■ = Antarctic species [31]. Figure modified from [25, 31].
As well as exhibiting low rates for most life history characters, Antarctic marine species are also highly stenothermal, dying in experiments when temperatures are raised to between +5°C and +10°C [26]. There is now strong evidence that these low upper temperature limits are set by a limitation to aerobic scope, or poor capacity to raise metabolic rates to perform work [12,29,30]. In the bivalve mollusc Laternula elliptica oxygen consumption and heartbeat rate rise with temperature between 0°C and 9°C, where both collapse and animals begin to die. Both oxygenated and deoxygenated blood is obtained from samples at 0°C, 3°C and 6°C, but no oxygenated blood is present at 9°C, and death ensues from a deficiency in oxygen supply to the tissues [32]. At around 5°C anaerobic end products of metabolism such as succinate and acetate begin to accumulate, indicating a long-term physiological limit, termed the critical limit [29].
That Antarctic marine species have poor capacities to raise metabolism for work has been shown for several species. The brachiopod Liothyrella uva can only elevate metabolism by × 1.7 over resting rates at 0°C when temperature is raised and the limpet Nacella concinna, by a factor of 2.6 [33]. In bivalves the value for Limopsis marionensis × 1.8 [34], and for Laternula elliptica is × 2.7 [32]. These values compare with the factor of over 4 that the Antarctic freshwater fairy shrimp B. gaini raises its rate of oxygen consumption with increased temperature. One difference here is that in experiments on the marine species temperatures were only raised to between +4°C and +10°C, where animals died, compared to the 25°C for B. gaini .
Limited aerobic scopes have further consequences for Antarctic marine ectotherms. Their ability to perform activity or work is severely temperature restricted. Thus, 50% of individuals of populations of burrowing bivalves, Laternula elliptica lose the ability to bury when temperatures are raised to 2.5°C, there is complete loss of burrowing ability in all animals at 5°C, and large reproductive individuals lose the ability to bury first [35] (Fig. 5). Populations of the limpet Nacella concinna similarly lose the ability to right themselves when turned over, with 50% loss of ability at temperatures between 2°C and 2.5°C. The scallop Adamussium colbecki is even more severely constrained, with no animals capable of swimming when temperatures are raised to 2°C [35]. Essential biological functions are, therefore rapidly lost with rising temperature in Antarctic marine species. Such limited capacity will clearly have marked effects on abilities to obtain and process food, and to reproduce, and hence animal fitness. This factor may be more important ecologically in terms of survival of populations and species than the physiological limits described earlier, although the physiological and ecological restrictions caused by extreme temperature sensitivity will be intimately linked. Current maritime Antarctic summer sea temperatures are usually around 1°C. Reproductive sized L. elliptica and N. concinna lose the ability to perform significant activities at, or below 2.5°C, only 1.5°C above current summer maxima, when animals need to exploit the short phytoplankton bloom. Scallops are even more tightly constrained, with all individuals losing the ability to perform important biological activities at temperatures less than 1°C above current summer maxima. These data are all for bivalve molluscs. However, the effect is likely to be widespread and suggests that a 1°C rise in summer sea temperatures around Antarctica could have severe ecological consequences for populations and communities of marine ectotherms.
Figure 5 The effect of temperature on activity in Antarctic marine species. a. The proportion of animals in a population of the limpet Nacella concinna capable of righting when turned over. b. The proportion of the infaunal bivale Laternula elliptica capable of reburying when removed from sediment. c. The proportion of scallops, Adamussium colbecki capable of swimming. A regression was fitted to data for temperatures above 0.3°C, where a clear relationship exists. Below this value no temperature restriction is apparent. In all 3 graphs regressions were fitted following arcsin transforms of the square root of % values. Dotted lines indicate 95% confidence intervals. Figure modified from [35].
Predicted environmental change and potential consequences
Freshwater habitats
The Antarctic peninsula is currently one of the fastest changing environments on earth, with air temperatures having risen by over 2°C in the last 50 years at some sites [4]. Aquatic environments may be changing even faster than this, with temperatures in lakes on Signy Island having risen by nearly 1°C through the 1980's and 1990's, at a site where air temperatures rose by around 1°C in 40 years [10]. Lakes and pool temperatures are changing faster than local air temperatures because of an amplification effect of reduced ice cover throughout the year. As ice cover decreases a longer period of the year is available for sunlight to penetrate to sediments, which are then heated and higher temperatures are maintained through the winter [10]. Current models predict that subantarctic and maritime Antarctic terrestrial sites are likely to warm significantly over the next 100 years, by possibly as much as 5°C [36]. The result of this will be to make sites where freshwater invertebrates currently exist better habitats for colonisation, and populations of species like the fairy shrimp B. gaini are likely to expand. New sites further south are also likely to become available for colonisation, as pools further south become unfrozen and algal mats build up. The overall consequence, therefore for freshwater species will be an extension of range further south and an improvement of habitat increasing overall population sizes.
As seen earlier species inhabiting terrestrial and freshwater sites have wide physiological flexibilities to cope with environmental changes of the type predicted. The major problem facing the limited number of species currently inhabiting Antarctic freshwater environments is displacement by alien species that are not capable of colonising current habitats. Alien invasions of subantarctic island sites are well documented [37,38], and many of the invasions have used humans as vectors. This is clearly the case for both rats and cats, as well as grasses and insects on a range of islands [37]. However, invasions using natural vectors have occurred, and under predicted future climate change, at some stage conditions will have reached levels where alien species can invade, and current endemics will be displaced. The likely outcome is that current endemic species will move south to areas where alien species are not capable of establishing populations, and lower latitude maritime Antarctic sites will be colonised by alien species, or mixed communities of current endemics and alien invaders.
Marine environments
There is little evidence that sea temperatures are changing around Antarctica, although one report has indicated Antarctic mid water increased in temperature by 0.17°C between the 1950's and the 1980's [39]. Current models predict global sea temperature will rise on average by 2°C in the next 100 years [1,36]. However, there will be marked regional variation, with some areas warming more than 2°C, and others not changing, or possibly cooling following the movement of large-scale ocean currents. Temperature changes in polar oceans are particularly difficult to predict, because of a lack of knowledge concerning how sea-ice is likely to change, and the effects this will have on the underlying water temperatures.
Other factors that are likely to change with rising temperatures are sea-ice and ice shelf cover and productivity. Satellite data indicate sea-ice extent has not changed markedly over the last 25 years [40], and models vary in their predictions of sea-ice cover over the next 100 years. The IPCC panel prediction was that there is likely to be a decrease of around 25% in sea-ice extent over the next 100 years [41]. As significant productivity is associated with the ice edge, a reduction in ice cover will reduce the extent of these blooms, reducing overall productivity.
Responses that enhance survival when the environment changes fall into three main categories, (1) using physiological capacities or scopes, (2) adapting to new conditions, or (3) migrating to areas that provide a better environment. From the earlier analysis, the Antarctic marine fauna is characterised by limited physiological capacities, and has a poor ability to cope with change using physiological capacities, possibly the poorest of any fauna so far described. The Antarctic marine fauna is characterised by slow growth, increased longevity and deferred maturity [23]. Of the species studied, several live in excess of 40–50 years [23,35,42]. Generation times are often in the range 10–20 years. Abilities to produce new adaptations to changing conditions are, therefore, poor. In excess of these considerations most continents have long coastlines covering many degrees of latitude and this, in many cases, allows for migration to more hospitable conditions when environments deteriorate. The coastline of Antarctica covers relatively few degrees of latitude, and less than any other continent. Species inhabiting the Southern Ocean seabed, therefore, have less scope to migrate away from poor conditions than faunas elsewhere. The combination of very poor physiological capacities, with extended life histories producing slow rates of adaptation and restricted available dispersal ranges make Antarctic marine species amongst the least capable of responding to environmental change on earth.
Conclusion
Terrestrial Antarctic species are adapted to very variable conditions and have good prospects for surviving in changing environments. It is likely that species diversity, community biomass and complexity will increase in coming years. There is a strong likelihood that new species will eventually appear on the Antarctic continent as warming continues, and these new species could displace the current endemic groups.
Marine species face a different problem. They have evolved in an environment with possibly the lowest temperature variation on Earth, and have lost the ability to cope with temperature change. A rise in summer sea temperatures of only 2°C could potentially compromise survival for many Antarctic benthic species. They may be one of the most vulnerable faunal groups to environmental change on Earth.
Acknowledgements
Several people have played a significant role in helping with the development of the ideas in this review, including A. Clarke, H.O. Pörtner, D.K.A. Barnes, P. Convey, K. Fraser, R. Worland, P. Tyler, S. Brockington and D. Bailey. I also thank the members of Signy and Rothera stations of the British Antarctic Survey, as well as the technical and aquarium support staff in Cambridge for assistance with animal collections and experimental support,
==== Refs
IPCC Houghton JT, Ding Y, Griggs DJ, Noguer M, van der Linden PJ, Dai X, Maskell K, Johnson CA Climate Change2001: The Scientific Basis Contribution of Working Group I to the Third Assessment Report of the Intergovernmental Panel on Climate Change 2001 Cambridge University Press, Cambridge, and New York, NY
UNEP Global Environment Outlook 3 2002 Earthscan Publications, London
Meehl GA Boer GJ Covey C Latif M Stouffer RJ The coupled model intercomparison project (CMIP) Bull Am Met Soc 2000 81 313 318 10.1175/1520-0477(2000)081<0313:TCMIPC>2.3.CO;2
Mitchell JFB Johns TC Senior CA Transient response to increasing greenhouse gases using models with and without flux adjustment Hadley Centre Technical Note 1998 2 26
Peck LS Bolis CL, Keines R, Maddrell SHP, Elia M, Weibel E Coping with Change: Stenothermy, Physiological Flexibility and Environmental Change in Antarctic Seas Climate Changes:Effects on Plants, Animals and Humans 14th ICCP, September 2000 2002 Troina, Italy (CD)
Peck LS Webb KE Bailey D Extreme sensitivity of biological function to temperature in Antarctic marine species Funct Ecol 2004 18 625 630 10.1111/j.0269-8463.2004.00903.x
Fox AJ Cooper PR Measured properties of the Antarctic ice sheet derived from the SCAR Antarctic digital database Pol Rec 1994 30 201 206
Davey MC Pickup J Block W Temperature variation and its biological significance in fellfield habitats on a maritime Antarctic island Ant Sci 1992 4 383 388
Peck LS Convey P Barnes DKA Environmental constraints on life histories in Antarcitc ecosystems: tempos, timings and predictability Biol Rev 2005
Quayle WC Convey P Peck LS Ellis-Evans JC Butler HG Peat HJ Ecological responses of maritime Antarctic lakes to regional climate change Antarct Res Ser 2003 76 335 347
Peck LS Physiological flexibility: the key to success and survival for Antarctic fairy shrimps in highly fluctuating extreme environments Freshwater Biol 2004 49 1195 1205 10.1111/j.1365-2427.2004.01264.x
Pörtner HO Climate change and temperature dependent biogeography: oxygen limitation of thermal tolerance in animals Naturwissenschaften 2001 88 137 146 11480701 10.1007/s001140100216
Convey P Overwintering strategies of terrestrial invertebrates in Antarctica – the significance of flexibility in extremely seasonal environments Eur J Entomol 1996 93 489 505
Chown SL Gaston KJ Exploring links between physiology and ecology at macro-scales: the role of respiratory metabolism in insects Biol Rev 1999 74 87 120 10.1017/S000632319800526X
Convey P How does cold constrain life cycles of terrestrial plants and animals? Cryo-Letters 2000 21 73 82 12148051
Clarke A Seasonality in the Antarctic marine environment Comp Biochem Physiol 1988 90B 461 473
Hunt BM Hoefling K Cheng CHC Annual warming episodes in seawater temperatures in McMurdo Sound in relationship to endogenous ice in notothenioid fish Ant Sci 2003 15 333 338 10.1017/S0954102003001342
Clarke A Holmes LJ White MG The annual cycle of temperature, chlorophyll and major nutrients at Signy Island, South Orkney Islands, 1969–82 Brit Antarct Surv Bull 1988 80 65 86
Grange L Tyler PA Peck LS Cornelius N Long-term interannual cycles of the gametogenic ecology of the Antarctic brittle star Ophionotus victoriae Mar Ecol Prog Ser 2004 278 141 155
Comiso JC Maynard NG Smith WO JnrSullivan CW Satellite ocean color studies of Antarctic ice edges in summer and autumn J Geophys Res 1990 95 9481 9496
Garrison DL Jeffries MO Gibson A Coale SL Neenan D Fritsen C Okolodkov YB Gowing MM Development of sea ice microbial communities during autumn ice formation in the Ross Sea Mar Ecol Prog Ser 2003 259 1 19
McMinn A Runcie JW Riddle M Effect of seasonal sea ice breakout on the photosynthesis of benthic diatom mats at Casey, Antarctica J Phycol 2004 40 62 69
Peck LS Ecophysiology of Antarctic marine ectotherms: limits to life Pol Biol 2002 25 31 40 10.1007/s003000100308
Peck LS Ansell AD Webb KE Hepburn L Burrows M Burrowing in Antarctic bivalve molluscs Pol Biol 2004 27 357 367 10.1007/s00300-003-0588-7
Bosch I Beauchamp KA Steele ME Pearse JS Development, metamorphosis and seasonal abundance of embryos and larvae of the Antarctic sea urchin Sterechinus neumayeri Biol Bull 1987 173 126 135
Peck LS Conway LZ Harper EM, Taylor JD, Crame JA The myth of metabolic cold adaptation: oxygen consumption in stenothermal Antarctic bivalves The Evolutionary Biology of the Bivalvia 2000 177 Geological Society, London, Special Publications 441 45
Clarke A Johnston N Scaling of metabolic rate and temperature in teleost fish J Anim Ecol 1999 68 893 905 10.1046/j.1365-2656.1999.00337.x
Peck LS Clarke A Chapman A unpublished observation
Pörtner HO Climate variations and the physiological basis of temperature dependent biogeography: systematic to molecular hierarchy of thermal tolerance in animals Comp Biochem Physiol A Physiol 2002 132 739 761
Mark FC Bock C Pörtner HO Oxygen-limited thermal tolerance in Antarctic fish investigated by MRI and P-31-MRS Am J Physiol – Reg Int Comp Physiol 2002 283 r1254 r1262
Stanwell-Smith DP Peck LS Temperature and embryonic development in relation to spawning and field occurrence of larvae of 3 Antarctic echinoderms Biol Bull 1998 194 44 52
Peck LS Pörtner HO Hardewig I Metabolic demand, oxygen supply, and critical temperatures in the Antarctic bivalve Laternula elliptica Physiol Biochem Zool 2002 75 123 133 12024288 10.1086/340990
Peck LS Temperature and basal metabolism in two Antarctic marine herbivores J Exp Mar Biol Ecol 1989 127 1 2 10.1016/0022-0981(89)90205-0
Pörtner HO Peck LS Zielinski S Conway LZ Temperature and metabolism in the highly stenothermal bivalve mollusc Limopsis marionensis from the Weddell Sea, Antarctica Pol Biol 1999 22 17 30 10.1007/s003000050386
Peck LS Webb KE Bailey D Extreme sensitivity of biological function to temperature in Antarctic marine species Funct Ecol 2004 18 625 630 10.1111/j.0269-8463.2004.00903.x
King JC Harangozo SA Climate change in the western Antarctic Peninsula since 1945: observations and possible causes Annals Glaciol 1998 27 571 575
Bergstrom DM Chown SL Life at the front: history, ecology and change on southern ocean islands TREE 1999 14 472 477 10542453
Gabriel AGA Chown SL Barendse J Marshall DJ Mercer RD Pugh PJA Smith VR Biological invasions of Southern Ocean islands: the Collembola of Marion Island as a test of generalities Ecography 2001 24 421 430 10.1034/j.1600-0587.2001.d01-198.x
Gille ST Warming of the Southern Ocean since the 1950s Science 2002 295 1275 1277 11847337 10.1126/science.1065863
Bjørgo E Johannesen OM Miles MW Analysis of merged SMMR-SSMI time series of Arctic and Antarctic sea ice parameters 1978–1995 Geophys Res Let 1997 24 413 416 10.1029/96GL04021
IPCC The Regional Impacts of Climate Change: An Assessment of Vulnerability 1998 Cambridge: Cambridge University Press
Arntz WE Brey T Gallardo VA Antarctic zoobenthos Oceanogr Mar Biol Annu Revs 1994 32 241 304
| 15938749 | PMC1177978 | CC BY | 2021-01-04 16:38:34 | no | Front Zool. 2005 Jun 6; 2:9 | utf-8 | Front Zool | 2,005 | 10.1186/1742-9994-2-9 | oa_comm |
==== Front
Health Qual Life OutcomesHealth and Quality of Life Outcomes1477-7525BioMed Central London 1477-7525-3-311587174610.1186/1477-7525-3-31ResearchCoping, quality of life, and hope in adults with primary antibody deficiencies Sigstad Hanne Marie Høybråten [email protected] Asbjørg [email protected]øland Stig S [email protected] Centre for Rare Disorders, Rikshospitalet University Hospital, Oslo, Norway2 The Department of Special Needs Education, University of Oslo, Oslo, Norway3 Department of Medical Genetics, Rikshospitalet University Hospital, Oslo, Norway4 Section of Clinical Immunology and Infectious Diseases, Department of Medicine, Rikshospitalet University Hospital, Oslo, Norway2005 4 5 2005 3 31 31 26 1 2005 4 5 2005 Copyright © 2005 Sigstad et al; licensee BioMed Central Ltd.2005Sigstad 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
Living with a chronic disease, such as primary antibody deficiency, will often have consequences for quality of life. Previous quality-of-life studies in primary antibody deficiency patients have been limited to different treatment methods. We wanted to study how adults with primary antibody deficiencies manage their conditions and to identify factors that are conducive to coping, good quality of life and hope.
Methods
Questionnaires were sent to all patients ≥20 years of age with primary antibody deficiencies who were served by Rikshospitalet University Hospital. The questionnaires consisted of several standardized scales: Ferrans and Powers Quality of Life Index (QLI), Short Form-36 (SF-36), Jalowiec Coping Scale (JCS), Nowotny Hope Scale (NHS), and one scale we devised with questions about resources and pressures in the past. Of a total of 91, 55 patients (aged 23–76 years) answered the questionnaires. The questionnaire study were supplemented with selected interviews of ten extreme cases, five with low and five with high quality of life scores.
Results
Among the 55 patients, low quality of life scores were related to unemployment, infections in more than four organs, more than two additional diseases, or more than two specific occurrences of stress in the last 2–3 months. Persons with selective IgA deficiency had significantly higher QLI scores than those with other antibody deficiencies. An optimistic coping style was most frequent used, and hope values were moderately high. Based on the interviews, the patients could be divided into three groups: 1) low QLI scores, low hope values, and reduced coping, 2) low QLI scores, moderate hope values, and good coping, and 3) high QLI scores, moderate to strong hope values, and good coping. Coping was related to the patients' sense of closeness and competence.
Conclusion
Low quality of life scores in adults with primary antibody deficiencies were linked to unemployment and disease-related strains. Closeness and competence were preconditions for coping, quality of life and hope. The results are valuable in planning care for this patient group.
primary immunodeficiency diseases
==== Body
1. Background
Primary immunodeficiency diseases represent a heterogeneous group of rare disorders characterized by an increased susceptibility to infections and autoimmune diseases. Primary antibody deficiencies (PAD) constitute the largest subgroup and include: Common Variable Immunodeficiency, X-linked (Brutons) Agammaglobulinemia, Selective IgA deficiency, IgG subclass deficiency, and Hyper IgM syndrome [1]. Some patients need lifelong replacement therapy with immunoglobulins and/or frequent courses of antibiotics as treatment and/or prophylaxis. Patients with PAD have increased incidence of auto-immune diseases and experience long-term complications of infections and/or treatment [2]. Living with a chronic disease, such as PAD, will often have consequences for quality of life. Previous quality-of-life studies in PAD patients have been limited to different treatment methods. After initiation of subcutaneous replacement therapy, increased health-related function and improved self-rated health have been reported [3]. We wanted to study wider aspects of quality of life among adults with PAD: How do they manage their condition? Which factors are conducive to coping, good quality of life, and hope?
Coping, quality of life, and hope are important aspects when the effects of a disease from infancy to old age are examined. There are various partially overlapping perspectives on, and definitions of coping, quality of life, and hope [4]. Coping reflects a process and includes active involvement over a period of time [5,6]. Hope and quality of life describe outcomes rather than processes. Hope and quality of life are concepts which have several dimensions. Coping also includes different strategies, but the total sum of the strategies does not constitute a global definition of the concept. Choice of strategies can influence outcome variables such as hope or quality of life positively or negatively. Coping is of importance for quality of life, and hope can be regarded as a coping strategy [7]. Hope can be seen as a variable that positively contributes to the experience of quality of life.
Coping is defined by Lazarus and Folkman [[5]; p.141] as "Constantly changing cognitive and behavioural efforts to manage, reduce or tolerate external and/or internal demands that are appraised as taxing or exceeding the resources of the person". The coping process depends on the situational context in which it occurs [5]. According to Lazarus and Folkman's theory [5,6], resources and pressures are linked to coping. We used resources and pressures as concepts in the present study. Resources can be divided in two groups; personal and socio-ecological resources. Pressures, such as disease-related experiences, may lead to stress and to reduced coping ability.
Locus of control is seen as a crucial factor in coping [8]. An internal locus of control is present when a person explains events by referring to causes within themselves. The person perceives that the event is contingent upon his/her own behavior or his/her own relatively permanent characteristics. An external locus of control is present when a person explains events by referring to causes in the situation or environment. Events and circumstances are typically perceived as the result of luck, chance, fate, under the control of powerful others, or unpredictable because of the great complexity of the forces surrounding the person given an external locus of control.
Resilience predisposes for successful coping [9,10]. Longitudinal studies in high risk children have showed positive correlation between resilience and overcoming difficult social circumstances as adolescents. Sommerschild [11] developed a theoretical model which sums up of the key points of resilience theory (see Figure 1). Resilience refers to the person's latent resources which can be mobilized in defense of the self in stressful situations. Resilience is based on self confidence. Closeness and competence contribute to self confidence. Figure 1 shows the concept closeness explained at three levels; the dyadic relationship with one competent adult, family, and the social network. Competence encompasses the person's skills and experience of usefulness.
Figure 1 Closeness and competence: preconditions for resilience against stress. (Sommerschild, 1998 [11]).
In the present study, quality of life was defined as a person's overall satisfaction with life: "A person's sense of well-being that stems from satisfaction or dissatisfaction with areas of life that are important to him/her" [[12]; p. 15]. The global concept of quality of life is represented by four domains: A health and functioning domain, a socio-economic domain, a psychological/spiritual domain, and a family domain. Health-related quality of life is defined as a person's satisfaction or happiness within areas of life that are affected by health or health care. Health-related quality of life includes eight domains: physical functioning, role physical, bodily pain, general health perceptions, vitality, social functioning, role emotional, and mental health.
Hope is future-oriented and described as a feeling, an emotion, an experience, a need and a dynamic attribute [13,14]. "A six-dimensional, dynamic attribute of the person which orients to the future includes: active involvement by the individual, comes from within, is possible, relates to or involves others or a higher being, and relates to meaningful outcomes to the individual" [[15]; p.89].
Two questions have been central in the present study:
1. How do adults with PAD manage their condition?
2. What kinds of factors influence their coping, quality of life and hope?
2. Methods
The survey was the main investigation in this study, and was analyzed quantitatively. The interviews were included as a supplement to the survey, and did not represent a traditional qualitative study.
The survey
Sample
As of 2000, 122 patients with PAD were registered in Norway [1]. All PAD patients ≥ 20 years and served by Rikshospitalet, in total 91 persons, received the questionnaires, after excluding one cognitively impaired person. The cohort included 50 men and 41 women, aged 20–82 years, with various PAD diagnoses: Common variable immunodeficiency (n = 66), X-linked Agammaglobulinemia (n = 8), Selective IgA deficiency (n = 16) and Hyper IgM syndrome (n = 1).
55 of the 91 adults we approached completed the questionnaires (60% response rate), 31 men and 24 women. The mean age was 41.6 (median 38, range 20–76) years. The sample included 29 young adults, aged 20–39 years and 26 older adults, aged 40–76 years. Distribution of the specific PAD diagnoses was: Common Variable Immunodeficiency (n = 43), X-linked Agammaglobulinemi (n = 3), Selective IgA deficiency (n = 8) and Hyper IgM syndrome (n = 1). Nine of the responders reported previous infection with Hepatitis C virus (HCV). Specific PAD diagnoses, gender, and mean age, 47.2 (44, 20–82) years, were similar and not significant different, when responders and non-responders were compared. Responders (n = 55) were reasonably representative of the original cohort (n = 91).
Measures
The survey included questions about the individual's past and present situations, and about his/her thought concerning plans for the future. Demographic variables including age, gender, education, employment and marital status were requested. Information was elicited concerning the following disease-related variables: specific diagnosis, duration of clinical immunodeficiency state, frequency of infections, in which organs the infections (acute and chronic) occurred, other medical complications (including Hepatitis C virus infection), additional diseases (including auto-immune diseases), treatment with subcutaneous (ScIg) versus intravenous immunoglobulin (IVIg), other treatments (antibiotics) and stressful events in the previous 2–3 months. Disease-related strains were defined as the most problematic strains linked to the PAD diagnosis, medical complications or additional diseases.
Five different scales were incorporated into one comprehensive 30-page questionnaire. Four of the standardized scales had previously been translated and tested in Norwegian populations [7,16,17]. The standardized scales were: Ferrans and Powers Quality of Life Index (QLI) [18], Short Form-36 (SF-36) [19], Jalowiec Coping Scale (JCS) [20], and Nowotny Hope Scale (NHS) [13]. An additional scale designed for this project (RPP Scale), focused on resources and pressures in the past. Factor analyses were used to assess the empirical support for each subscale in all instruments. Internal consistency was estimated using Cronbach's alpha coefficient.
Ferrans and Powers Quality of Life Index (QLI) has been designed to measure quality of life in both ill and healthy individuals, and this was based on Ferrans' definition of quality of life [12,18]. The global construct quality of life is represented by four underlying subscale domains/subscales:
• Health/Functioning
• Socio-economic
• Psychological/Spiritual
• Family
The QLI consists of two sections. One section measures satisfaction within various domains. The other section measures the importance of each domain for the subject. The items are scored according to a 6-point Likert scale ranging from "very satisfied" to "very dissatisfied" for the satisfaction items, and from "very important" to very unimportant" for the importance items. The overall score is the product of the satisfaction responses and the importance responses. The possible range for the overall and subscale scores is 0–30, the higher the score the better quality of life.
The validity and reliability of QLI has previously been evaluated. Content validity of the original version was assessed on the basis of a review of the literature [12,18]. Concurrent validity of the QLI was provided by a correlation (r = 0.80) between the QLI and a measure of satisfaction with life [12]. Construct validity was found to be satisfactory in different patient populations, and was confirmed by factor analysis ("the maximum-likelihood method" and "the direct oblimin method of rotation") [12,21]. A four-factor solution had the best fit with the data. Internal consistency reliability was 0.95 for the global score, ranging from 0.66 to 0.93 for the subscales. The test-retest reliability varied from 0.87 at two weeks to 0.81 at one month [18].
QLI has been translated and tested in a Norwegian population of newly diagnosed cancer patients [7]. 131 cancer patients participated in the test, 103 in the retest. The Cronbach's alpha coefficient for the QLI was 0.93 for test and 0.95 for retest. The coefficients for the subscales in both tests ranged from 0.79 to 0.91 [7]. Test-retest-reliability was r = .78 within three-four weeks (Pearsons correlation coefficient). The correlation coefficients ranged from r = .65 to r = .83 for the different subscales. Construct validity was analysed by "the maximum-likelihood method" and "direct oblimin method" of rotation (factor analysis). Eight factors had an "Eigenvalue greater than 1". Four factors explained only 45.4% of the variance in this cancer patient cohort, in contrast to 91% in the study of Ferrans and Power [21].
Reliability analyses in the present study with 55 PAD patients showed a Cronbach's alpha of 0.81 for the total QLI, and ranged from 0.54 to 0.92 for the four subscales. The Family subscale had the lowest alpha value, and the Health/Functioning subscale had the highest. Factor analyses based on the four subscales were done by "the maximum-likelihood method", non-rotated method and "direct oblimin method", rotated. Non-rotated method with "Eigenvalue greater than 1", resulted in one cluster where only one of the factors appear, (2.571). All the subscales could be related to this factor. The factor explained 64.3% of the variance in our sample. This result supported a total scale with a common component.
The Short Form-36 (SF-36), one of several generic questionnaires developed to assess health-related quality of life [19], consists of 36 items which measure eight conceptual domains:
• Self-reported General Health (GH)
• Physical Functioning (PF)
• Bodily Pain (BP)
• Mental Health (MH)
• Role limitations (Physical) (RP)
• Role limitations (Emotional) (RE)
• Vitality (VT)
• Social Functioning (SF)
In addition, one item assesses change in health in the past year (HT). The scores in each domain are transformed into 0–100 scales. The higher score the better health-related quality of life.
The reliability of the eight scales has been estimated using both internal consistency and test-retest methods [22]. Reliability coefficients for each of the eight scales were equal or greater than .80 (ranging from .81 in General Health to .93 in Physical Functioning) with the exception of Social Functioning, which had a reliability of .68. The content validity of the SF-36 has been compared to that of other widely used generic health surveys. Systematic comparisons reveal that the SF-36 includes eight of the most frequently represented health concepts.
SF-36 has previously been translated and tested in 2323 persons from the general Norwegian population [17,23]. Reliability analyses (Cronbach's alpha) showed values from 0.80 to 0.93 for the eight subscales, Role limitations (Emotional) had the lowest and Bodily Pain the highest value. Correlations between the SF-36-scales ranged from r = .29 (Mental Health and Physical Functioning) to r = .68 (Mental Health and Vitality).
Reliability analysis (Cronbach's alpha) of the SF-36 in the present study (n = 55) yielded values from 0.74 to 0.92, Role limitations (Emotional) had the lowest and Social Functioning the highest value. Factor analysis is not usually performed when using SF-36. In spite of a relatively small sample size, factor analysis was done in this study by "Principal Component Analysis", non-rotated method and "Varimax with Kaiser Normalization" rotated method of the eight subscales. The analyses revealed two main factors. The scales which contributed the sum score of Physical Health, were one factor. The scales which contributed the sum score of Mental Health, constituted the other factor. Compared with the original SF-36, there was one finding of note in the present study: the Social Functioning subscale was correlated with the sum score of Physical Health. The Social Functioning subscale in the SF-36 is originally included in the sum score of Mental Health.
The Jaloviec coping scale (JCS) is based on Lazarus and Folkman's theory of stress and coping [5,6,20]. The JCS has been designed to measure how people cope with various types of physical, emotional and social stressors. The JCS measures the use and effectiveness of 60 cognitive and behavioural coping strategies in a stressful situation. The items describe cognitive and behavioural efforts in response to stress. In our questionnaire, stress was specified as stress induced by living with PAD. The strategies are grouped into eight coping dimensions:
• Confrontive – "tried to change the situation"
• Evasive – " put off facing up to the problem"
• Optimistic – "tried to think positively"
• Fatalistic – "accepted the situation because very little could be done"
• Emotive – "worried about the problem"
• Palliative – "tried to keep busy and work harder"
• Supportive – "depended on others to help out"
• Self-reliant – "preferred to work things out yourself"
Item responses are rated on a 4-point scale from 0 (never used) to 3 (often used), and a scale of helpfulness from 0 (not helpful) to 3 (very helpful). The higher score, the more coping effort involved. The higher total coping score the more alternation between different coping strategies.
The JCS has previously been tested in several studies [20,24]. Its content validity has been assessed by an expert panel and is supported by a broad theoretical and empirical base. Construct validity has been evaluated. The 60 items are classified into eight subscales, with an agreement ranging from 94% on the Supportive subscale to 54% on the Emotive subscale. Reliability of the JCS, assessed with Cronbach's alpha coefficients and based on results from 24 different studies ranged from 0.48 to 0.81 for the use subscales and from 0.48 to 0.82 for the effectiveness subscales.
JCS has previously been translated and tested in a Norwegian population of 273 patients with psoriasis [16]. Correlations between the eight subscales in JCS ranged from r = .39 (p < .001) to r = .73 (p < .001). Reliability analyses (Cronbach's alpha) of the eight subscales ranged from 0.55 to 0.88. The construct validity of the JCS was analysed by "Principal Component Analysis" with orthogonal rotation (factor analysis). The analyses resulted in three coping dimensions with sufficient internal consistency: confrontive problem-solving coping, normalizing / optimistic coping and combined emotive engagement. 37 % of the total variation in the Norwegian version of JCS was attributed to these three factors [16].
Because few patients responded to the coping effectiveness part of the JCS in the present study, we have only included the coping strategy use part. Reliability analyses (Cronbach's alpha) ranged from 0.41 to 0.75 for the eight subscales (use-scores). The Confrontive, the Evasive and the Optimistic subscales yielded the highest values from 0.73 to 0.75, and the Fatalistic subscale the lowest alpha at 0.41. Our finding of three subscales with the strongest internal consistency is in keeping with the results of Jaloviec et. al. [20]. Factor analyses based on subscales, done by "Principal Component Analysis", non-rotated method and "Varimax with Kaiser Normalization", rotated method, resulted in a three-factor-solution with factor 1: Evasive, Fatalistic and Self-reliant coping; factor 2: Confrontive, Emotive, Palliative coping; and factor 3: Optimistic coping. These analyses revealed that the Optimistic scale could be considered a separate contributing factor in the present study.
Nowotny Hope Scale (NHS) is designed to measure hope in a general adult population after a stressful event [13,15], and has been employed primarily in cancer patients. NHS is a 29-item scale with items scored on a 4-point Likert Scale ranging from 4, strongly agree, to 1, strongly disagree. It consists of six subscales:
• Confidence
• Relates to others
• Future is possible
• Spiritual beliefs
• Active involvement
• Comes from within
The total score range is from 29 to 116, with a high score indicating high hope. Cut-off scores are developed for four levels of hope.
The content validity of the NHS has been evaluated by an expert panel [13]. The concurrent validity was established with the Beck Hopelessness Scale (r = -.47). The construct validity of NHS was analysed by "Principal Components Analysis" with orthogonal rotation (factor analysis). The result supported the six dimensions and subscales of hope. Cronbach alpha's reliability coefficient for this instrument was 0.90. The concurrent validity has been found to be satisfactory [13].
NHS has been translated and tested in the above-mentioned Norwegian population of newly diagnosed cancer patients [25]. Correlations between different subscales ranged from r = -.16 to r = .73. Factor analysis done by "Principal Components Analysis", showed that the items of the "Spiritual beliefs" subscale appeared as one factor, and the "Comes from within" items as another factor analogous to Nowotny's subscale items. With the exception of these two factors, the results of the NHS factor analysis of Rustøen [25] diverged from Nowotny's original six dimensions. At three-four week test-retest of NHS correlation was high, Pearson's r = .81. Correlation coefficients ranged from r = .59 to r = .92 for the various subscales. Cronbach's alpha for NHS was 0.89 both in the test and the retest. The alpha coefficients for the subscales ranged between 0.53 and 0.95.
Reliability analysis of NHS in the present study (n = 55) showed a total Cronbach's alpha of 0.87. Cronbach's alpha of the six subscales ranged from 0.54 to 0.94. Our results were similar to Rustøen and Moum's results [25]. Factor analyses based on subscales were done by "Prinicipal Component Analysis", non-rotated method, and "Varimax with Kaiser Normalization", rotated method. Both the non-rotated and the rotated variant of factor analysis seemed to confirm only two explicit contributing factors: factor 1; the Spiritual subscale, and factor 2; the five other subscales.
The Resources and Pressures in the Past Scale (RPP Scale) was theoretically founded on Lazarus and Folkman's theory of coping [5,6] and consisted of 64 items divided into two main categories: Resources and Pressures. The past was defined as the period from adolescence to present time. The distinction between different domains within the person's resources was based on previous studies of coping [9,10,26-28]. The concept of Resources included both personal characteristics / temperament and social support resources. Pressures were defined as the person's individual perception of his/her experiences, including immunodeficiency-related experiences and general events.
Resources consisted of four subscales:
• Personal characteristics / temperament
• Family and supporting adults
• Supporting persons in school and social network
• Public Health Service
Pressures in our scale consisted of four subscales:
• Immunodeficiency-related events (for example: many hospitalizations because of the disease)
• General events
• Immunodeficiency-related experiences in school (for example: significant absence from school because of the disease)
• General experiences in school
The items were scored on a 5-point Likert Scale ranging from 5, strongly agree, to 1, strongly disagree. The total score range of Resources was from 29 to 145, and the total score range of Pressures was from 35 to 175. A high score indicated either a good availability of resources or a high level of strain.
In the present study, missing was handled by the same procedure as in the standardized scales JCS and NHS: when more than 50% of the subscale is answered, the missing value is replaced by the mean score of the rest of the subscale.
This scale was evaluated by reliability analyses (Cronbach's alpha) and factor analyses. The aim behind using reliability analyses was to evaluate in what degree the individual items correlated with the main concepts in the scale. Some items were excluded to attain highest possible consistency. The Cronbach's alpha was 0.89 for Resources and 0.90 for Pressures, with a range from 0.63 (Personal characteristics/temperament) to 0.90 (Family and supporting adults) in Resources, and a range from 0.59 (Immunodeficiency-related events) to 0.89 (Immunodeficiency-related experiences in school) in Pressures. Validity was tested by "Principal Component Analysis", non-rotated method and "Varimax with Kaiser Normalization" with rotation (factor analysis) based on the subscales. With "Eigenvalue greater than 1", the non-rotated method made only one contributing factor appear for the Pressures scale to which 55% of the total variance could be attributed. Likewise, the factor analysis (non-rotated method) yielded only one contributing factor from the Resources' scale, to which 46% of the total variance could be attributed.
Statistical analyses of the survey data
The SPSS-PC statistical program (v 9.0) was used for data analyses. Descriptive analyses were performed to assess the characteristics of the sample. The impact of demographic and clinical variables on the dependent variables was assessed by t-test for independent samples (two-tailed). The effect sizes were measured by the difference between the means of the samples divided by the mean of the standard deviations of the samples [29]. The effect size was defined by qualitative standardized values (small = .25SD, medium = .50SD, large = 1.00SD). Both correlation analyses and multiple regression analyses were performed to assess the relationship between variables.
The interviews
The interviews were included as a supplement to the survey to elucidate preconditions for coping, good quality of life, and hopefulness. The interview study was designed to probe and to aid in the interpretation of some of the results from the questionnaire. Cases were selected for interviews to detect possible patterns within two groups: patients with high QLI scores and patients with low QLI scores. The selected cases represented a strategic sample of patients with the lowest and highest QLI scores (n = 21). Originally, we wanted to interview all these extreme cases (n = 21). Ten patients consented to participate in the interview study. Ten cases were regarded as sufficient to detect patterns within the different groups. The qualitative interviews were based on significant results related to QLI in the survey. The interviews were semi-structured with a pre-written interview guide, and lasted nearly two hours.
The interview guide was based on the most central issues in the survey: previous resources and pressures, the interviewees' experience of coping and quality of life, and their hope for the future. Questions about coping included present challenges and choice of coping strategies. Questions about experience of quality of life were related to the four dimensions in QLI: Health/Functioning, Socio-economic, Psychological/Spiritual and Family. Questions concerning hope, dealt with the patients' general experience of hope. All concepts were related to a lifespan perspective as the interviewees were asked to evaluate their present situation related to their past. All interviews were done by the same person and tape-recorded.
In accordance with Kvale's methodology [30], the interviews were analyzed on a thematic and a theoretical level. Kvale distinguishes between three different contexts of interpretation of the interview statements. The thematic level implies a condensed form of what the interviewees themselves understand to be the content of their statements. The interpretation is more or less based on the interviewees' self-understanding as understood by the researcher. The common sense level represents a critical common sense understanding. The interpretations may include a wider frame of understanding than those of the interviewees themselves. The interviewer should be critical of what is said, and may focus on the content of the statement. The theoretical level is a framework for interpreting the meaning of a statement. These interpretations are likely to go beyond the interviewees' self-understanding and to exceed common sense understanding. In this study, the theoretical level included the common sense understanding.
The interviews were analyzed to identify interesting and important themes. New themes appearing during the interviews were included in the analyses (thematic level). Similarities and differences were described within and between the extreme groups. The expressed meanings were summarized into shorter terms. Individual texts were further analyzed with respect to meaning of the texts, and to their respective categories. The categories were divided into groups defined by contrasting elements within and across the groups denoting high and low quality of life [31]. When thematic analyses showed differences within or between the extreme groups, further analyses were done. Since some of these variants seemed to be in accordance with previous studies in coping research, the results of the first thematic analyses were reanalyzed according to relevant theory about the constructs of coping, quality of life, and hope (theoretical level).
According to Kvale [30] reliability pertains to the consistency of the research findings, and validity to the truth and correctness of a statement. Kvale emphasizes that issues of verification do not belong to a separate stage of an investigation, but should be addressed throughout the entire process. Validation is done at seven stages in the interview process: 1. Thematizing based on the logic of derivations from theory to the research questions of the study, 2. Designing dependent on the adequacy of the design and the methods used for the purpose of the study, 3. Interviewing based on trustedness of the interviewees reports and the quality of the interviewing itself, 4. Transcribing dependent on the quality of the translation from oral to written language, 5. Analyzing dependent on whether the questions in the interview text are valid and whether the interpretations are logical, 6. Validitating, based on reflective consideration of what forms are relevant to a specific study and 7. Reporting dependent on to what degree a given report is a valid account of the main findings of a study.
3. Ethical aspects
The study was approved by the Norwegian Regional Committee for Medical Research Ethics and the Norwegian Social Science Data Services. Participants were guaranteed anonymity and the right to withdraw from the study at any time. An information letter to respondents provided information about the potentially sensitive items.
4. Results
The survey
t-tests were done on selected demographic and disease-related variables and the results are presented in table 1(see Additional file 1). Other results and comparisons are only presented in the text. Table 1 includes results of total scores (means and standard deviations) of all scales in the present study with one exception: the most significant differences in SFscores were found in four out of the eight domains. Only these are presented in the table (BP, MH, RP and SF).
On the RPP, the 55 adults with PAD reported good availability of resources in the past (personally and support from others) (mean 3.7, range 2.03–4.93, out of a possible total score of 5). Parents and other supporting adults had been of major importance as social support (mean 3.8, range 1.69–5.00 out of a possible score of 5). Adults with PAD had experienced moderate pressures (2.6, 1.27–4.66, out of a possible total score of 5). Pressures related to their immunodeficiency were the most burdensome in the school context (3.09, 1.00–5.00). Four conditions in present time implicated significantly more pressures in the past: younger age (20–39 years) (p = .024), (Effect Size (ES) = .77SD), living alone (p = .015), (ES = .84SD), having more than two additional diseases (p = .005), (ES = 1.07SD), or suffering from infections in more than four organs (p = .038), (ES = .69SD) (t-tests, two-tailed) (Table 1).
The mean score in global QLI (quality of life) was moderate at 20.0 (range 12.3–27.6, out of a possible total score of 30.0). In addition to immunodeficiency, the following conditions were associated with significantly lower QLI scores: unemployment (p = .008), (ES = .80SD), infections in more than four organs (p = .020), (ES = .79SD), the presence of more than two other diseases (p = .001), (ES = 1.55SD), or more than two specific occurrences of stress in the last 2–3 months (p = .007), (ES = 1.15SD) (t-tests, two-tailed). These results are presented in table 1. Unemployed men had lower QLI scores compared to employed men (p= .020). The term "unemployed" was defined to include currently/previously unemployed, never employed and recipient of disability pension. Men working full-time achieved significantly higher QLI scores than men working part-time or unemployed men (p = .016). These differences did not exist among the women. Variables without impact on QLI were: length of education, type of treatment (subcutaneous Ig versus intravenous Ig), frequency of treatments, self-administration of treatment at home (ScIg), hospital based treatment (IVIg), and HCV infection.
Compared to a Norwegian sample of newly diagnosed cancer patients (n = 131) [7], the PAD patients (n = 55) had a significantly lower total QLI score (p < .05), along two dimensions: Health and Functioning (p < .05) and Socio-economic (p < .01).
Health-related quality of life in different conceptual domains on the SF-36, revealed that the adults with PAD had their lowest mean score in General Health (37.8, range 5.0–87.0, of a possible total score of 100.0) and their highest mean score in Physical Functioning (81.1, 10.0–100.0, of a possible total score of 100.0) (SF-36). The most significant differences in SF scores were found in four of the eight domains, and these are presented in table 1. Gender, employment and disease-related pressures/strains had significant influence. Men had a significantly higher score than women in Bodily Pain, Social Functioning and Vitality, respectively (p = .029), (ES = .64SD); (p = .016), (ES = .68SD); and (p = .004), (ES = .85SD) (t-tests, two-tailed). Unemployed men and women, had significantly lower health-related quality of life, compared to employed adults. Low health-related quality of life was found in Bodily Pain (p = .006), (ES = .81SD); General Health (p = .002), (ES = .88SD); Mental Health (p = .021), (ES = .68SD); Physical Functioning (p = .001), (ES = 1.01SD); Role limitations (Physical) (p = .000), (ES = 1.16SD); and Social Functioning (p = .004), (ES = .91SD). The disease-related strains were infections in more than four organs, infections more than eight times yearly, more than two other diseases and/or more than two specific occurrences of stress in the last 2–3 months. Hepatitis C infection did not have a negative influence on health-related quality of life.
Compared to a control group with a normal distribution (n = 2323) [17], our PAD patients (n = 55) showed significantly lower functional ability scores in all areas of health-related quality of life (SF-36). The finding reached statistical significance (.001) in four areas: General Health, Role limitations (Physical), Social Functioning and Vitality. Compared to a sample of psoriasis patients (n = 283) [4], these PAD patients showed different scores in two areas (SF-36): Bodily Pain, where adults with PAD scored higher (p < .01), and General Health, where adults with PAD scored lower (p < .001).
Of the eight coping strategies measured by the JCS, an optimistic coping strategy was most frequently used (mean item rating 2.28, range 0.44–3.00, on a 4-point scale of 0–3). A palliative coping strategy was rarely used (1.18, 0.20–2.29). The total score for all coping strategies used showed a mean item rating of 1.64 (1.13–2.32). This reflects the extent of use of all coping strategies measured [20]. Being unemployed was associated with high coping scores among adults with PAD (p = .013), (ES = .78SD) (t-tests, two-tailed). Full-time employment was associated with lower coping scores compared to part-time employment, housework or unemployment (p = .021), (ES = .67SD) (Table 1).
The PAD patients had moderate hope values on the NHS with a mean score of 84.9 (range 52–102, of a possible total score of 116). Having more than two additional diseases in addition to PAD was associated with a lower hope value among responders (p = .015), (ES = 1.02SD) (t-tests, two-tailed). The results are presented in table 1. There was positive correlation between being hopeful about the future and quality of life (QLI) in the present, r = .454 (p < .001) (Pearson correlation). Regression analysis with quality of life (QLI) as the dependent variable and hope (NHS) as one of the independent variables, showed R2 = .206 (p < .001), which suggests that hope explains 20.6% of the total variance in the quality of life.
Compared to a sample of newly diagnosed cancer patients (n = 131) [7], our PAD patients (n = 55) had a significantly lower total hope value (p < .05) visualized in two dimensions of NHS: Relates to others (p < .01), and Future is possible (p < .05).
The interviews
During thematic analysis of the ten interviews, certain categories appeared to be of particular importance. These categories were used as the main categories in the theoretical analysis: quality of life, closeness and competence as resilience, locus of control, and hope. The five responders with high QLI scores showed more homogeneous results than the five responders with low QLI scores for the interview themes resources and pressures in past, coping ability, quality of life, and hope for future. The latter group was split into two subgroups based on criteria related to experience of closeness and competence, and locus of control. Locus of control was determined by evaluating the responders' answers about their own experience of internal control over external occurrences. In spite of a low QLI score, three of the responders seemed to have strong resilience combined with an internal locus of control.
Based on the theoretical analysis, the subjects with a low QLI score were divided into two groups (Group 1 and Group 2). The subjects with a high QLI score were defined as Group 3. Persons in Group 1 (n = 2) had low scores in all four subscales of QLI (Health/Functioning, Socio-economic, Psychological/Spiritual, Family). They had problematic psychological bonds to their mothers, and less experience of closeness or/and competence (Fig 1). They experienced difficulties in coping (self-reported), they had low hope values and either an internal or an external locus of control. In addition, persons in Group 1 had needed various forms of social support. However, they expressed reluctance to receive such support. Persons in Group 2 (n = 3) had low scores in two subscales of QLI: Health/Functioning and Socio-economic. They had especially close relationship to their mothers, but a positive experience of closeness and competence. They were coping successfully (self-reported), had a moderate hope values and an internal locus of control. The persons in Group 2 also needed additional social support, but received such help according to their own wishes. Persons in Group 3 (n = 5) had high scores in all four subscales of QLI. They had experienced closeness and competence, and they were coping successfully. They had moderate to strong hope values, an internal locus of control, and reported no need for additional support.
5. Discussion
The purpose of the present study was to study how adults with PAD manage their condition and to identify factors that are conducive to coping, good quality of life, and hopefulness. Low scores in quality of life were linked to unemployment and disease-related strains among adults with PAD. Closeness and competence were preconditions for coping, good quality of life and hope.
The survey showed that parents and other supporting adults were the most important caregivers (Resources) in adolescence. This is in accordance with findings in previous coping studies [32]. Not surprisingly, the interviews confirmed the family as the best caregivers during childhood. In cases where the parents did not fulfill their function as caregivers, other people such as neighbors and health personnel functioned as caregivers. In addition, social support was not only associated with positive experiences among the responders with low QLI score. Those with a low QLI score reported a complicated relationship to their mothers. They wanted to be accepted as adults, but did not experience that they were.
Experiences related to immunodeficiency (Pressures) were of major importance, for example: episodes of illness, absence from school, psychosocial consequences of the disease, self-respect and respect from other people. These results came from the survey. The interviews confirmed that occurrences related to the immunodeficiency were the most chronic problems. This is in accordance with Ogden's conclusions [33]. Ogden classified painful school experiences as a long-term element of risk. Many studies have emphasized the importance of the impact of previous pressures on later development [32,34-36]. In accordance with their findings, the results of the present study point to previous resources and pressures as crucial factors for future coping ability and maturation.
A high degree of immunodeficiency-related strain as well as unemployment had a negative impact on Health and Functioning on the QLI in this sample (n = 55) (Quality of life). In order to achieve a high total QLI score, a low score in one dimension has to be compensated by higher scores in the other dimensions. To be satisfied with one's own achievement as an experience of coping is seen as crucial to achieving a high quality of life [37]. Consequently, unemployment requires that one is able to compensate for lack of employment with another meaningful activity. This may be interest in interpreting the finding in this study that unemployed men reported lower quality of life than men with a steady job.
Persons with Selective IgA-deficiency achieved a higher global QLI score than other PAD patients. Patients with symptomatic selective IgA deficiency are usually healthier than other PAD patients [2]. Surprisingly, the QLI differences were not found in the health/function domain, but in the socioeconomic, family and psychological/spiritual domains.
We observed a difference in QLI between those who treated themselves (ScIg) compared to those who were treated by others (IVIg), but the difference was not statistically significant. However, the PAD patients who treated themselves (ScIg) had a significantly higher score in Social Functioning (SF-36) compared to the others. SF-36 measured health-related quality of life. Our study focused on global quality of life, not specific in relation to treatment, and the results did not elucidate all aspects of these different treatment methods. Gardulf [3] found a significantly increased health-related function, and improved self-rated health among patients with PAD after initiation of ScIg infusions. However, our study was not designed to detect differences before and after introduction of a specific treatment method.
Nine of the 55 responders had experienced Hepatitis C virus infection due to contaminated IVIg [38]. Surprisingly, HCV infection had no influence on the QLI scores or scores of SF-36 in this study.
The interviewees with a low QLI score were in poor health and reported some limitations in daily life functioning. Still, these patients showed obvious differences within the group related to other quality of life conditions, as some of them (Group 2) seemed to fully adjust their experience of quality of life. Well-being and satisfactory social support were reported. This was not anticipated because of low QLI scores from the questionnaire. Wilson and Cleary's research studies [39] suggest that there is no direct correlation between serious limitation in health and loss of quality of life. On the other hand, positive self-esteem, ability to be active and to use one's abilities are elements of crucial significance [37,40,41] in achieving a high quality of life score. In our interview study, there was a lack of such characteristics in Group 1 who had low scores in all domains of the QLI. Insufficient involvement, dependence on others, low self-esteem, and lack of happiness and well-being were characteristic interview responses in this group.
There were some differences between the findings in the survey and the interviews: The global QLI scores and SF-scores did not give a good description of social network. Interviews indicated that a good network was important for resilience. The interviews may have described more comprehensively the responders' experience of different types of support. Mutual relationships were identified as important by interview, but had not been included among the specific items in the survey. Psychological characteristics, external living conditions and relationships have been evaluated by others [37] as essential factors for quality of life. The responders with a low QLI score in Group 2 (the interview study) confirmed that a good social network had contributed to increasing their quality of life.
Compared to a Norwegian sample of newly diagnosed cancer patients [7], our 55 PAD patients had a lower global QLI score, along dimensions: Health and functioning and socio-economic. By nature of their disease, the PAD patients have had a life-long course, in contrast to the cancer patients. Perhaps the chronicity of PAD explains some of the divergence.
Compared to the cancer cohort [17], our PAD patients scored lower on functional ability in all areas of health-related quality of life (SF-36). Other quality-of-life studies confirm significant negative implications of physical health limitations [42,43]. Compared to a sample of psoriasis patients [16], our PAD patients scored differently in two areas (SF-36): Bodily Pain, where adults with PAD had a higher score, and General Health, where adults with PAD had a lower score. Bodily pain is not characteristic for PAD patients and can be a possible explanation for the high score in SF-36. Adults with psoriasis have better general health, compared with adults with PAD.
An optimistic coping strategy was most frequently used in dealing with the illness (survey). This is a consistent finding in other studies of groups with other chronic diseases [16,44-46]. Employment is linked to competence and may predispose for successful coping (Fig 1) [11]. According to Sommerschild [11], coping depends on competence in various areas, e.g. perceived self-efficacy, usefulness, and problem-solving skills. Unemployment requires that one is able to compensate in other spheres in order to achieve a feeling of competence [32]. Competence and closeness are areas which may be amenable to psychosocial interventions aimed at increasing quality of life in adults with PAD. Prospective interventions can be designed for patients with low quality of life scores or may include all PAD patients.
The PAD patients had moderate hope scores on the NHS. Having more than two other diagnoses in addition to PAD was associated with a lower hope score. Less hopefulness correlated positively with a high degree of disease-related strain. We found a definite positive correlation between being hopeful about the future and quality of life (QLI) in the present. Hope seemed to be responsible for 1/5 of the total variance in the quality of life. Increasing hope may have an impact on enhancing quality of life.
Compared to the cohort of newly diagnosed cancer patients [7], our PAD patients had a significantly less global hope score in two dimensions of NHS: "Relates to others", and "Future is possible". Studies confirm that persons with a cancer diagnosis, thought to be in a hopeless situation, often have a positive and hopeful vision of the future [47]. The PAD patients include people with a congenital chronic disease for which there is no curative treatment. All of the newly diagnosed cancer patients had been diagnosed within the previous year. They were aware that cancer is a terminal illness, but they may have retained hope for curative treatment.
Another study, which focused on sources of hope among people with chronic diseases, emphasized hope as a main coping strategy [48]. Evangelista [49] found that hope strongly correlated with quality of life in a cohort of female heart transplant recipients. The Group 2 and Group 3 patients in our interviews had plans and dreams, and were optimistic about the future, and their plans were related to other people. But in Group 1, thoughts about the future were characterized by fatalism, less faith in the future and lack of involvement with others.
The low response rate (60%) may be considered a limitation of the external validity of the survey. Including subjects with other immunodeficiency disorders would have increased sample size at the cost of homogeneity. Homogeneity was chosen to enhance validity. Simple statistical analyses were employed. We were able to use the t-test after controlling for statistical assumptions (independent samples, normal distribution, and homogeneity of variance). The findings reported as significant reached, with one exception, levels of significance of .01 and .001. Effect size varied between moderately large to a large [29]. In spite of some concerns regarding statistical power, we consider that our results have relatively good statistical validity [50].
Five scales were used; four of them standardized and well tested by others, and in Norwegian populations. One scale was specially designed for this study. Seen in a lifespan perspective, related to previous pressures/strains and availability of resources, we showed that the results of this scale also were important for the evaluation of the results from the other scales. In spite of some variation between the subscales, the five scales showed good reliability generally (Cronbach's alpha) in keeping with the main constructs of the study. The four standardized scales were internally consistent and test-retest reliability was satisfactory. Despite a thorough testing of the RPP Scale, the construct validity of this scale may be more uncertain. Nevertheless, in spite of some limitations, the total construct validity appears robust.
If the group is too special, it may not be correct to generalize the results to other kinds of persons or other diseases. However, this sample was characterized by a wide age span (23–76 years), different subgroups of PAD, various types of treatment, age at diagnosis and number of disease-related strains. Analysis of the non-responders showed that the responders in this survey (n = 55) were representative for the entire cohort (n = 91), and they constitute 75% of all patients registered with PAD in Norway [1]. Therefore, this survey was has relatively good external validity relative to this patient group.
The results of the interviews were evaluated in according to Kvale's methodology [30] for qualitative research interviewing, and the reliability and validity were found to be reasonably good. An evaluation during the analysis process of the interviews concluded that the size of the sample was sufficient to detect different patterns in these two groups of extreme cases. And similar patterns have been found in previous studies on coping [9,10,51]. Such a confirmation is expected to strengthen the results, and to some extent compensate for the limited sample size. However, the low number necessitates a tentative and explorative use of the study findings.
Since the sample of the ten interview subjects was in accordance with the strategic sample concerning the most central variables impacting on quality of life, the generalizability of the interview study could be evaluated as relatively good.
The interviews were included to elucidate nuances in the knowledge from the survey results. Triangulation of different methods is considered an appropriate strategy for strengthening the validity of research findings [52,53]. The interview study supplemented the survey, supported the constructs in the survey and, as a result, strengthened the validity of the findings. In this project, the results from the survey were thoroughly examined during the interviews. However, the interview study also raised some new hypotheses, which were not registered in the survey. The interviews have added new aspects, and functioned as a supplement to the survey, increasing the relevance of the project and its results. Combining the findings from both the survey and the interviews strengthens the validity of the project's end result.
There may be ethical concerns when inviting patients to complete questionnaires about coping, quality of life and hope. In this study, the information letter gave an opportunity to prepare for potentially sensitive questionnaire items. Moreover, the interviews made ethical demands on the interviewer. In an interview, the interviewer is a part of the method, and has a responsibility to manage separating roles, having focus on the interview and taking care of the interviewees in a professional way. In the present study, to get the interviewees' confidence, the interviewer had to define the content of the interview precisely, and explain the interviewers' role. The interviewer referred the interviewees to other health personnel when needed.
6. Conclusion
Low scores in quality of life were linked to unemployment and disease-related strain among adults with PAD. Coping was closely linked to the patients' sense of closeness and competence. The results are in accordance with previous studies of other groups with chronic diseases. Closeness and competence are areas where psychosocial interventions may contribute to better quality of life in adults with PAD. The findings are relevant also for other groups of patients. Medical interventions should reduce the patient's strain, and support his or her ability to be employed.
Supplementary Material
Additional File 1
Table 1. Mean scores (SD) in PAD patients on resources, strains, quality of life, functioning, coping, and hope
Click here for file
Acknowledgements
The authors thank Trine Prescott, Department of Medical Genetics, Rikshospitalet; Prof. Thorleif Lund, Department of Special Needs Education, University of Oslo; Eva Brox, The Norwegian Immune Deficiency Association; and Heidi Sandersen, Department of Paediatrics, Rikshospitalet for carefully reading and correcting the manuscript.
==== Refs
Stray-Pedersen A Abrahamsen TG Frøland SS Primary Immunodeficiency Diseases in Norway J Clin Immunol 2000 20 477 485 11202238 10.1023/A:1026416017763
Hammarström L Smith CIE Ochs HD, Smith CIE, Puck JM Genetic Approach to Common Variable Immunodeficiency and IgA Deficiency Primary Immunodeficiency Diseases A molecular and genetic approach 1999 New York: Oxford University Press 250 262
Gardulf A Björvell H Gustafson R Hammarström L Smith CIE The life situations of patients with primary antibody deficiencies untreated or treated with subcutaneous gammaglobulin infusions Clin Exp Immunol 1993 92 200 204 7683584
Spilker B Quality of Life Assessments in Clinical Trials 1990 New York: Raven Press
Lazarus R Folkman S Stress, appraisal, and coping 1984 New York: Springer
Lazarus R Folkman S Monat A, Lazarus R The Concept of Coping Stress & Coping 1991 New York: Columbia University Press
Rustøen T Hope and Quality of life in newly diagnosed cancer patients PhD thesis 1998 University of Bergen, Department of Public Health and Primary Health care, Division for Nursing Science
Rotter JB Generalized expectancies for internal versus external control of reinforcement Psychol Monogr 1966 80 1 280 5340840
Werner EE Shonkoff JP, Meisels SJ Protective factors and individual resilience Handbook of early childhood intervention 2000 2 New York: Cambridge University Press
Werner EE Smith RS Overcoming the Odds High Risk Children from Birth to Adulthood 1992 Ithaca and London: Cornell University Press
Sommerschild H Gjærum B, Grøholt B, Sommerschild H Mestring som styrende begrep (Coping as a leading concept) Mestring som mulighet i møte med barn, ungdom og foreldre 1998 Oslo: Tano Aschehoug 21 63
Ferrans CE Development of a quality of life index for patients with cancer Oncol Nurs Forum 1990 17 15 21 2342979
Nowotny M Assessment of hope in patients with cancer: Development of an instrument Oncol Nurs Forum 1989 16 57 61 2911528
Kylmä J Vehviläinen-Julkunen K Hope in nursing research: a meta-analysis of the ontological and epistemological foundations of research on hope J Adv Nurs 1997 25 364 371 9044012 10.1046/j.1365-2648.1997.1997025364.x
Nowotny ML Measurement of hope as exhibited by general adult after a stressful event 1986 47 Dissertation Abstracts International 3296-B (University Microfilms no. 8626494)
Wahl AK Quality of life and coping among patients suffering from psoriasis PhD thesis 1999 University of Bergen, Department of Public Health and Primary Health Care, Division for Nursing Science
Loge JH Kaasa S Short form 36 (SF-36) health survey: normative data from the general Norwegian population Scand J Soc Med 1998 26 250 258 9868748
Ferrans CE Power MJ Quality of Life Index: development and psychometric properties ANS Adv Nurs Sci 1985 8 15 24 3933411
Ware JE JrGandek B IQOLA Project Group The SF-36 Health Survey: Development and Use in Mental Health Research and the IQOLA Project Int J Ment Health 1994 23 49 73
Jalowiec A Murphy SP Powers MJ Psychometric assessment of the Jalowiec Coping Scale Nurs Res 1984 33 157 161 6563533
Ferrans CE Power MJ Psychometric assessment of the Quality of Life Index Res Nurs Health 1992 15 29 38 1579648
Ware JE Jr Spilker B The SF-36 Health Survey Quality of Life and Pharmacoeconomics in Clinical Trials 1996 2 Philadelphia: Lippincott-Raven Publishers 337 345
Loge JH Kaasa S Hjermstad MJ Kvien TK Translation and performance of the Norwegian SF- 36. Health Survey in patients with reumatoid arthritis Data quality assumptions, reliability, and construct validity J Clin Epidemio 1998 51 1069 1076
Jalowiec A Psychometric results on the 1987 Jalowiec coping scale School of Nursing 1991 Chicago, IL: Loyola University of Chicago
Rustøen T Moum T Reliability and Validity of The Norwegian Version of The Nowotny Hope Scale. A Nursing Tool For Measuring Hope in Cancer Patients Scand J Caring Sci 1997 11 33 41 9275820
Sutton C A handbook of research for the helping professions 1987 London and New York: Routledge & Kegan Paul
Beresford BA Resources and Strategies: How Parents Cope with care of a disabled Child J Child Psychol Psychiatry 1994 35 171 209 8163627
Fonagy P Steele M Steele H Higgit A Target M "The Emanuel Miller Memorial Lecture 1992: The Theory and Practice of Resilience." J Child Psychol Psychiatry 1994 35 231 257 8188797
Cohen J Wolman BB Some Statistical Issues in Psychological Research Handbook of Clinical psychology 1965 New York: McGraw-Hill 95 121
Kvale S InterViews An Introduction to Qualitative Research Interviewing 1996 Thousand Oaks: Sage Publications
Strauss A Corbin J Basics of Qualitative Research Grounded Theory Procedures and Techniques 1990 Newbury Park, CA: Sage Publications
Masten AS Best KM Garmezy N Resilience and development: Contributions from the study of children who overcome adversity Dev Psychopathol 1990 2 425 444
Ogden T Kompetanse i kontekst En studie av risiko og kompetanse hos 10- og 13- åringer (Competence in context A study of risks and competence among children, 10- and 13- years old) Rapportserie fra barnevernets Utviklingssenter 1995 Report 3
Garmezy N Rutter M Rutter M, Hersov L "Acute reaction to stress." Child Psychiatry: Modern approaches 1985 Oxford: Blackwell Scientific 152 176
Rutter M Rolf J, Masten AJ, Cicchetti D, Neuchterlein KH, Wintraub S "Psychosocial resilience and protective mechanisms." Risk and protective factors in the development of psychopathology 1990 New York: Cambridge University Press 52 66
Sperling MB Berman WH Attachment in Adults 1994 New York, London: The Guilford Press
Næss S Mastekaasa A Sørensen T Moum T Livskvalitet som psykisk velvære (Quality of life as mental well-being) NOVA Report 3 Oslo 2001
Bjøro K Skaug K Haaland T Frøland SS Long-term outcome of chronic hepatitis C virus infection in primary hypogammaglobulinaemia QJM 1999 92 433 441 10627859 10.1093/qjmed/92.8.433
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
Diener E Assessing subjective well-being Psychol Bull 1994 95 542 575 6399758
Schwarz N Strack F Strack F, Argyle M, Schwarz N Evaluating one's life: a judgement model of subjective well-being Subjective well-being 1991 New York: Pergamon Press 27 47
Nesbitt BJ Heidrich SM Sense of coherence and illness appraisal in older women's quality of life Res Nurs Health 2000 23 25 34 10686570 10.1002/(SICI)1098-240X(200002)23:1<25::AID-NUR4>3.0.CO;2-M
Loge JH Abrahamsen AF Ekeberg O Kaasa S Reduced health-related quality of life among Hodgkin's disease survivors: a comparative study with general population norms Ann Oncol 1999 10 71 77 10076725 10.1023/A:1008315525997
Heidrich SM Mechanisms related to psychological well-being in older women with chronic illnesses: age and disease comparisons Res Nurs Health 1996 19 225 235 8628911 10.1002/(SICI)1098-240X(199606)19:3<225::AID-NUR6>3.0.CO;2-M
Lindqvist R Sjoden PO Coping strategies and quality of life among patients on continuous ambulatory peritoneal dialyseis (CAPD) J Adv Nurs 1998 27 312 319 9515641 10.1046/j.1365-2648.1998.00523.x
Chiou A Potempa K Buschmann MB Anxiety, depression and coping methods of hospitalised patients with myocardial infarction in Taiwan Int J Nurs Stud 1997 32 305 311 9306165 10.1016/S0020-7489(97)00018-7
Yates P Towards a reconceptualization of hope for patients with a diagnosis of cancer J Adv Nurs 1993 18 701 706 8514925 10.1046/j.1365-2648.1993.18050701.x
Raleigh ED Sources of Hope in Chronic Illness Oncol Nurs Forum 1992 19 443 448 1594466
Evangelista LS Doering LV Vassilakis ME Kobashigawa J Hope, mood states and quality of life in female heart transplant recipients J Heart Lung Transplant 2003 22 681 686 12821165 10.1016/S1053-2498(02)00652-6
Cook TD Campbell DT Quasi-experimentation: design & analysis issues for field settings 1979 Chicago: Rand McNally
Werner EE Smith RS Vulnerable but invincible A Longitudinal Study of Resilient Children and Youth 1982 New York: McGraw-Hill Book Company
Patton MQ Qualitative evaluation and research methods 1990 2 Newbury Park, CA: Sage Publications
Lund T The Qualitative-Quantitative Distinction: Some Comments Scand J Educ Res 2005 49 115 132 10.1080/00313830500048790
| 15871746 | PMC1177979 | CC BY | 2021-01-04 16:38:14 | no | Health Qual Life Outcomes. 2005 May 4; 3:31 | utf-8 | Health Qual Life Outcomes | 2,005 | 10.1186/1477-7525-3-31 | oa_comm |
==== Front
Health Qual Life OutcomesHealth and Quality of Life Outcomes1477-7525BioMed Central London 1477-7525-3-401600461210.1186/1477-7525-3-40ResearchThe Herdecke questionnaire on quality of life (HLQ): Validation of factorial structure and development of a short form within a naturopathy treated in-patient collective Ostermann Thomas [email protected]üssing Arndt [email protected] Andre-Michael [email protected] Peter F [email protected] Chair of Medical Theory and Complementary Medicine, Faculty of Medicine, University of Witten/Herdecke, Germany2 Department of Naturopathy, Blankenstein Hospital, Hattingen, Germany2005 8 7 2005 3 40 40 10 6 2005 8 7 2005 Copyright © 2005 Ostermann et al; licensee BioMed Central Ltd.2005Ostermann 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
Quality of life (QoL) of patients has become a central evaluation parameter that also acts as an aid for decisions related to treatment strategies particularly for patients with chronic illnesses. In Germany, one of the newer instruments attempting to measure distinct QoL aspects is the "Herdecke Questionnaire for Quality of Life" (HLQ). In this study, we aimed to validate the HLQ with respect to its factorial structure, and to develop a short form. The validation has been carried out in relation to other questionnaires including the SF-36 Health Survey, the Mood-Scale Bf-S, the Giessen Physical Complaints Questionnaire GBB-24 and McGill's Pain Perception Scale SES.
Methods
Data for this study derived from a model project on the treatment of patients using naturopathy methods in Blankenstein Hospital, Hattingen. In total, 2,461 patients between the ages of 16 and 92 years (mean age: 58.0 ± 13.4 years) were included in this study. Most of the patients (62%) suffered from rheumatic diseases. Factorial validation of the HLQ, it's reliability and external consistency analysis and the development of a short form were carried out using the SPSS software.
Results
Structural analysis of the HLQ-items pointed to a 6-factor model. The internal consistency of both the long and the short version is excellent (Cronbach's α is 0.935 for the HLQ-L and 0.862 for the HLQ-S). The highest reliability in the HLQ-L was obtained for the "Initiative Power and Interest" scale, the lowest for the 2-item scales "Digestive Well-Being" and the "Physical Complaints". However, the scales found by factor analysis herein were only in part congruent with the original 5-scale model which was approved a multitrait analysis approach. The new instrument shows good correlations with several scales of other relevant QoL instruments. The scales "Initiative Power and Interest", "Social Interaction", "Mental Balance", "Motility", "Physical Complaints", "Digestive Well-Being" sufficiently differentiate the diagnostic groups, particularly between the patients suffering on connective tissue and soft tissue disorders from those with metabolic and nutritional disorders or hypersensitivity reactions.
Conclusion
Both the factorial validation and the development of a consistent short-form of the HLQ are important steps forward for researchers in the field of QoL who wish to use the HLQ as a reliable and valid instrument. The results indicate that the HLQ is a unique QoL-instrument that can be used for both in-patient and out-patient-treatment. However, to improve to profile of the HLQ, there is still the need for strengthening the Questionnaire in the dimensions of physical well-being. This is the subject of a separate ongoing study.
QuestionnairesQuality of lifeshort formrheumatic diseasesnaturopathyHerdecke Questionnaire for Quality of Life;
==== Body
Background
The consideration of "quality of life" (QoL) in clinical studies and various attempts to make this construct measurable to determine therapeutic success is an ongoing process. This is particularly the case in those therapeutic attempts that focus on integrative aspects of disease management that in turn offer holistic care including a variety of therapeutic directions. Here, the QoL has become a central evaluation parameter. It simultaneously acts as an aid for decisions on the choice of treatment strategy for chronically ill patients [1], which is obviously a challenging therapeutic aim, and is at least as significant as somatic parameters [2]. QoL has therefore become a leading criteria in many outcome studies alongside somatic and economic factors. In the course of this development, the concept of QoL is explicitly listed as outcome parameter in many medical societies' guidelines [3].
However, there are a variety of opinions regarding the factors that contribute to QoL. According to a WHO-definition, QoL relates to the physical, psychological and social well-being of an individual as laid out by formal health terms [4]. According to this definition, it is necessary to differentiate between a general and a health related QoL [5]. The former relates to aspects that exist independently from any particular disease (e.g. items such as "being spontaneous", or "feeling exhausted"), whereas the later focuses on particular characteristics of specific diseases (e.g. factors such as "walking distance" or "pain" in rheumatic diseases)
Despite the methodological difficulties involved in making QoL measurable, we have seen the development of numerous instruments for measuring disease specific aspects of QoL [6-8] in the recent past. An advantage of disease specific instruments is precise registration regarding strains and limitations of specific diseases rather than those of diseases in general. In addition, the course of clinical diseases can be more easily registered because of the development of disease-related questionnaires ("course of disease sensitivity" of questionnaires). The majority of current recommendations by health economists and clinical pharmacological associations include suggestions regarding the use of disease specific and general QoL questionnaires [9].
In Germany, one of the newer instruments attempting to measure general QoL with a distinct focus is the "Herdecke Questionnaire for Quality of Life" (HLQ is the German acronym of the phrase "Herdecke Questionnaire for Quality of Life") [10,11]. Clinical research projects have been reluctant to employ the HLQ although it was evaluated on a sample of healthy subjects, and that some reference values of clinical studies on different diseases do exist, and also despite of the fact that the HLQ has a very comprehensive understanding of the QoL problematic [12],. This is mainly because conclusive validation based on a large sample is still missing. To improve this situation, this study aimed to show the characteristics of the HLQ, to describe its external validation using other test instruments, and to develop a short form of the questionnaire.
Methods
Data for this study derive from a model project on the treatment of patients using naturopathy methods in Blankenstein Hospital, Hattingen. To investigate the benefits and limits of naturopathic treatment in the field of in-patient care, the Department of Naturopathy was established as a model at the Blankenstein Hospital in Hattingen and was scientifically evaluated by the Chair of Medical Theory and Complementary Medicine of Witten/Herdecke University. This evaluation began on July 1st 1999 and was completed on March 31th 2003. It focused on the following question: "How does a three-week in-patient treatment with naturopathic methods affect the QoL of the patients, regarding a pre-post-comparison and a follow-up carried out after 6 months? Detailed information concerning this model project and its' scientific evaluation can be found in [13] and [14].
In total, 2,461 patients between 16 and 92 years (mean age 58.0 ± 13.4 years) were included in this study. The socio-demographic characteristics of the patients are shown in Table 1.
Table 1 Socio-demographic data of the patient population
male (n = 507) female (n = 1954) total (n = 2461)
age mean 58,6 57,9 58,0
standard deviation 13,4 13,4 13,4
range 17–92 16–92 16–92
n % n % n %
age group under 18 years 1 0.2 3 0.2 4 0.2
18–45 years 89 17.6 346 17.7 435 17.7
45–60 years 162 32.0 676 34.6 838 34.1
60–65 years 81 16.0 308 15.8 389 15.8
65 and older 172 33.9 619 31.7 791 32.1
no details available 2 0.4 2 0.1 4 0.2
diagnostic groups connective tissue and soft tissue disorders 267 52.7 1305 66.8 1572 63.9
chronic disorders of the respiratory system 35 6.9 60 3.1 95 3.9
metabolic and nutritional disorders 92 18.1 133 6.8 225 9.1
hypersensitivity reactions 8 1.6 46 2.4 54 2.2
other indications 85 16.8 364 18.6 449 18.2
no details available 20 3.9 46 2.4 66 2.7
marital status single 56 11.0 178 9.1 234 9.5
married 352 69.4 1055 54.0 1407 57.2
living separated 7 1.4 34 1.7 41 1.7
divorced 37 7.3 227 11.6 264 10.7
widowed 36 7.1 399 20.4 435 17.7
second marriage 12 2.4 28 1.4 40 1.6
no details available 7 1.4 33 1.7 40 1.6
education still at school 2 0.4 7 0.4 9 0.4
no final exam 15 3.0 33 1.7 48 2.0
special school exams 1 0.2 5 0.3 6 0.2
secondary school exams other than GCSE 338 66.7 1191 61.0 1529 62.1
GCSE ? 80 15.8 431 22.1 511 20,8
A levels 60 11.8 201 10.3 261 10.6
other 1 0.2 27 1.4 28 1.1
no details available 10 2.0 59 3.0 69 2.8
most recent profession worker 206 40.6 338 17.3 544 22.1
employee/civil servant 212 41.7 941 48.2 1153 46.9
self employed 42 8.3 94 4.8 136 5.5
not working 15 3.0 242 12.4 257 10.4
unclear 1 0.2 23 1.2 24 1.0
no details available 31 6.1 316 16.2 347 14.1
professional situation full-time professional 142 280 311 15.9 453 18.4
part-time professional 19 3.8 288 14.8 307 12.4
housewife/husband 12 2.4 484 24.8 496 20.2
in training 4 0.8 13 0.7 17 0.7
retired pre retired state 267 49.6 670 33.3 937 38.2
unemployed 46 9.1 115 5.9 161 6.6
no details available 17 3.4 73 3.7 90 3.7
Alongside the HLQ, other standardized questionnaires were used. These included the MOS-SF-36 Health Survey [15], Zerssen's Mood-Scale Bf-S [16], the Giessener Physical Complaints Questionnaire GBB-24 [17] and McGill's Pain Perception Scale SES [18].
The HLQ as referred to in this study uses 39 five-point likert scales ranging from 0 to 4 (agreement/disagreement or often/never). In contrast to the SF-36, the items are not defined by situations related to daily life and household situations (shopping, career situations, physical activity). As a result, the HLQ is very suitable for registering QoL particularly in monitoring the course of a disease or therapeutic intervention [19]. As an evaluation scheme, Schulte et al. [10] described 5 scales of the 39 item HLQ, unfortunately without any confirmation by factor analysis of the following areas: Physical Well-being (4 items), Vitality (9 items), Mental behavior (10 items), Presence of Personality (9 items), Social Environment (7 items). All scales are expressed in percentage values from 0 = lowest to 100 = highest QoL.
The main question of this study relates to the re-examination of the HLQ by means of a factor and reliability analysis and the explorative evaluation of the factors. External validation was performed by correlating the HLQ scales with those of the external test instruments: MOS-SF-36 Health Survey [15], Zerssen's Mood-Scale Bf-S [16], the Giessener Physical Complaints Questionnaire GBB-24 [17] and McGill's Pain Perception Scale SES [18].
Factor analysis was performed using principal components analysis with Varimax rotation on 35 of the 39 items. The items, #13 (avoided conflicts), #14 (behavior of others was unclear to me), #15 (was glad) and #29 (reduced sexual activity) were omitted following the positive preliminary results on the reliability of the HLQ by Kroez et al. [20]. To determine the internal consistency of the questionnaire, reliability analysis was performed using Cronbach's alpha. Both factor analysis and reliability analysis were performed for the long and the short version of the HLQ.
For the short form, only relevant items with a factorial weight of >0.6 were selected. This method of selection was originally suggested by Grimley [21] and has successfully been applied elsewhere [22,23].. Coefficients of determination (R-square) of short and long form scales were calculated to evaluate the proportion of variance of the original HLQ which can be explained by the short form.
Evaluation of responsiveness of the HLQ over a course of time was achieved by analyzing the change of HLQ-total score from the time of admission to the time of discharge by using a dependent t-test and calculation of Cohen's effect size (ES). Cohen's guidelines were used to classify the magnitude of effect sizes: 0.2 represents a small effect, 0.5 a moderate effect, and 0.8 a large effect.
The statistical data evaluation was performed using the SPSS Version 10.0 program packet.
Results
The descriptive statistics of each item, the reliability parameters and the difficulty index are given in Table 2. Considering the high percentage of patients with chronic rheumatic diseases, an item-difficulty index between 0.26 (Item: "I suffered from physical pain") and 0.73 (Items "Family life was a burden" and "I felt over directed") can be regarded as sufficient. This also holds for item-total correlations with values between 0.27 and 0.69 (median: 0.55) for the original HLQ and between 0.32 and 0.61 (median: 0.46) for the short version (HLQ-S), These correlations are considered to be optimal ranges for psychological test instruments.
Table 2 Descriptive statistics and reliability parameters of HLQ-Items
ItemNo. Item Mean SD Item-Diff. Index loading Cronbach's α rItem-total old Scale
Total Scale-wise
long short long short
Initiative Power & Interest 0.885
10* good ideas 2.26 0.94 0.57 0.747 0.933 0.852 0.875 0.54 0.49 3
07* reacted spontaneously 2.13 1.08 0.53 0.640 0.934 0.855 0.880 0.47 0.42 3
11* concerned 2.72 1.02 0.68 0.630 0.932 0.846 0.871 0.67 0.61 3
08* decisive 2.44 0.91 0.61 0.617 0.934 0.852 0.876 0.54 0.48 4
12 put plans into action 2.09 0.91 0.52 0.572 0.933 0.877 0.58 4
25 difficult to take the initiative 2.33 1.12 0.58 0.535 0.932 0.873 0.65 4
36 enhanced personally 2.02 1.11 0.51 0.531 0.934 0.881 0.47 4
34 adapt to other people and situations 2.79 0.81 0.70 0.530 0.934 0.878 0.50 3
33 asserted in the environment 2.48 0.96 0.62 0.519 0.933 0.876 0.55 4
17 felt secure 2.36 0.95 0.59 0.481 0.932 0.874 0.66 4
21 future was clear 2.21 1.14 0.55 0.455 0.933 0.879 0.56 4
06 sought contact to others 2.27 1.10 0.57 0.453 0.935 0.884 0.41 5
30 felt enterprising/energetic 2.01 1.05 0.50 0.431 0.932 0.876 0.66 3
Social Interaction 0.812
16* felt left out 2.79 1.07 0.70 0.697 0.933 0.850 0.775 0.57 0.52 5
27* felt over-directed 2.90 1.10 0.73 0.636 0.933 0.853 0.780 0.57 0.51 4
20* abandoned community life 2.51 1.12 0.63 0.631 0.932 0.847 0.766 0.66 0.59 5
18 family life was a burden 2.92 1.18 0.73 0.546 0.934 0.792 0.52 5
32 didn't feel comfortable in the company of others 2.31 1.10 0.58 0.532 0.934 0.808 0.46 5
28 convey feelings to other 2.63 0.92 0.66 0.496 0.933 0.788 0.57 5
05 anxious/fearful 2.42 1.20 0.61 0.461 0.933 0.797 0.56 3
Mental Balance 0.812
35* nervous / irascibly 1.99 1.06 0.50 0.640 0.934 0.858 0.803 0.47 0.40 3
26* well-balanced 2.01 0.97 0.50 0.600 0.932 0.850 0.770 0.67 0.61 3
19* exhausted 1.24 0.94 0.31 0.567 0.933 0.849 0.782 0.59 0.56 2
09 could recover myself 1.57 0.99 0.39 0.557 0.933 0.784 0.55 2
31 tired 1.38 0.92 0.35 0.554 0.933 0.784 0.57 2
39 I happy 2.10 0.90 0.53 0.499 0.932 0.781 0.69 3
04 sleep was refreshing 1.55 1.05 0.39 0.356 0.935 0.809 0.42 2
Motility 0.781
22* physically agile 1.95 1.02 0.49 0.789 0.935 0.854 0.708 0.40 0.43 1
24* movement was light 1.84 1.07 0.46 0.786 0.934 0.851 0.673 0.47 0.46 1
38* arms and legs felt heavy 1.36 1.07 0.34 0.668 0.935 0.855 0.571 0.39 0.42 1
37 powerful 1.47 0.95 0.37 0.509 0.933 0.482 0.56 2
Physical Complaints 0.692
02* suffered from physical pain 1.04 0.97 0.26 0.726 0.936 0.859 * 0.27 0.32 1
01* felt ill 1.09 0.91 0.27 0.705 0.935 0.853 * 0.42 0.44 2
Digestive Well-Being 0.621
03* good appetite 2.71 1.12 0.68 0.763 0.935 0.857 * 0.37 0.38 2
23* mealtimes were a burden 2.87 1.10 0.72 0.734 0.934 0.853 * 0.45 0.46 2
number of answered items ranged from 2,227 [min.] and 2,430 [max.]
* Short form Item
The results of the structural analysis of the HLQ-items yielded surprising results. The scales found by factor analysis (Table 2) were only partly congruent with the scalesin the original publication [10]. Instead, we found a new and stable 6-factor-model which fits better with the original data than the original 5-scale model derived by Schulte et al., which used a multitrait analysis approach (developed by Hays et al. [24]). This is underlined by a Kaiser-Meyer-Olkin measure of sampling adequacy of 0.957 and a highly significant Bartlett test of sphericity (p < 0.001). The cumulative variance explained by this model is 54.7%.
Correlation analysis (Table 3) of the earlier HLQ scale with the new scale revealed significant correlations between the scales "Social Environment (SOC)" and "Social Interaction (SOCI)" (r = 0.923 for the HLQ-L and r = 0.860 for the HLQ-S). Unfortunately, such clear correlation between an old HLQ scale with the unique factor of our current analysis was not found with the other scales. However, "physical well-being (PWB)" of the old HLQ correlated well with the new "motility (MOT)" scale (HLQ-L r = 0.929 resp. HLQ-S r = 0.958), while the old "vitality (VIT)" scale correlates with the "mental balance scale (MB)" (HLQ-L r = 0.840 resp. HLQ-S r = 0.681). The old scales "presence of personality (PERS)" and "mental balance (MEB)" are represented well by the new scale "initiative power and interest (IPI)" (See Table 3).
Table 3 Partial Correlation of HLQ-Scales with other instruments and with the HLQ-Scales (old adjusted for Gender and Age. Abbrev.: SF-36: PF-physical function, RP-role physical, BP-bodily pain, GH-general health, VT-vitality, SF-social function, RE-role emotional, MH-mental health, MCS-mental component summary, PCS-physical component summary; GBB: SE-severity of exhaustion, GS-gastric symptoms, LP-limb pain, and HS-heart symptoms;, Zerssen's Mood Scale Bf-S; SES: AFF-Affective Pain, SENS-SenSory Pain; HLQ-OLD: PWB-physical well-being, VIT-vitality, MEB-mental behaviour, PERS-presence of personality, SOC-Social Environment.
Initiative Power and Interest Social Interaction Mental Balance Motility Physical Complaints Digestive Well-Being
long short long short Long short long short long Short long short
SF-36 PF 0.200 0.130 0.219 0.188 0.233 0.173 0.551 0.580 0.432 * 0.167 *
RP 0.234 0.178 0.252 0.226 0.294 0.257 0.461 0.453 0.403 * 0.184 *
BP 0.182 0.147 0.213 0.198 0.301 0.242 0.449 0.466 0.688 * 0.183 *
GH 0.358 0.288 0.363 0.308 0.361 0.322 0.353 0.332 0.350 * 0.208 *
VT 0.562 0.470 0.532 0.479 0.668 0.582 0.541 0.478 0.408 * 0.353 *
SF 0.543 0.437 0.632 0.589 0.550 0.486 0.392 0.342 0.337 * 0.324 *
RE 0.452 0.380 0.497 0.418 0.428 0.423 0.264 0.215 0.259 * 0.243 *
MH 0.650 0.543 0.686 0.605 0.719 0.690 0.378 0.316 0.324 * 0.360 *
MCS 0.650 0.548 0.698 0.615 0.667 0.640 0.294 0.213 0.245 * 0.348 *
PCS 0.020 -0.012 0.022 0.025 0.087 0.024 0.489 0.530 0.497 * 0.089 *
GBB 24 SE 0.458 0.363 0.513 0.457 0.626 0.546 0.559 0.496 0.387 * 0.323 *
GS 0.216 0.189 0.266 0.227 7 0.303 0.247 0.205 0.166 0.238 * 0.356 *
LP 0.249 0.201 0.296 0.263 0.386 0.312 0.490 0.498 0.514 * 0.190 *
HS 0.305 0.249 0.360 0.285 0.372 0.342 0.282 0.252 0.248 * 0.309 *
Bf-S 0.655 0.564 0.652 0.599 0.630 0.581 0.409 0.362 0.312 * 0.339 *
SES AFF 0.289 0.231 0.314 0.277 0.371 0.325 0.374 0.372 0.506 * 0.183 *
SENS 0.181 0.143 0.228 0.215 0.254 0.209 0.288 0.299 0.417 * 0.147 *
HLQ-OLD PWB 0.357 0.270 0.335 0.333 0.489 0.412 0.929 0.958 0.638 * 0.259 *
VIT 0.577 0.502 0.590 0.523 0.840 0.681 0.611 0.517 0.560 * 0.661 *
MEB 0.894 0.822 0.772 0.702 0.797 0.770 0.471 0.401 0.278 * 0.412 *
PERS 0.926 0.779 0.760 0.734 0.628 0.584 0.404 0.354 0.226 * 0.342 *
SOC 0.748 0.697 0.923 0.860 0.600 0.573 0.358 0.346 0.248 * 0.346 *
* values of the long version are identical with short version
The internal consistency of the instruments (HLQ-L and HLQ-S), both for the total score (Cronbach's α is 0.935 for the HLQ-L and 0.862 for the HLQ-S) as well as for the subscales of the HLQ-L (Cronbach's α between 0.621 and 0.885) can be regarded as being excellent. The highest alpha reliability in the HLQ-L was obtained for the "Initiative Power and Interest" scale, the lowest for the 2-item scales "Digestive Well-Being" (0.621) and "Physical Complaints" (0.692).
The mean difference between the scales of the HLQ-S and the HLQ-L for all patients is between 1.20 ("Initiative Power and Interest ") and 2.24 points ("Social interaction") on a percentage scale. The absolute differences are clustered in groups and are given in Table 4. Although there is a low overall mean difference, absolute differences greater than 10 percent range between 17.9% ("Initiative Power and Interest") and 26.8% ("Social interaction"). However, with correlation coefficients ranging from 0.899 to 0.964, the proportion of variance of the HLQ-L can be explained by the short form ranges between 79% and 93% and thus can be regarded as an adequate proportion for a short version.
Table 4 Comparison of the HLQ-L and the HLQ-S.
Difference of means Percentage of Patients with a mean difference D Correlation Explained Variance
< 3 3< D <7 7< D <10 10<D< 20 >20
Initiative Power and Interest 1.20 32.9% 33.6% 15.6% 15.8% 2.1% 0.899 81%
Social Interaction 2.24 25.9% 27.3% 20.1% 22.2% 4.6% 0.909 83%
Mental Balance 1.43 27.8% 28.3% 19.6% 20.4% 3.9% 0.888 79%
Motility 1.43 46.5% 29.7% 11.9% 11.1% 0.8% 0.964 93%
Physical Complaints * * * * * * * *
Digestive well-Being * * * * * * * *
* values of the long version are identical with short version
The correlation of the HLQ with other test instruments is shown in Table 5. There are acceptable correlations with r> 0.5 between the mental-health associated scales from the HLQ with those of the other instruments, for example, the "mental health"-Scale of the SF-36. In detail, the scales "Initiative Power and Interest", "Social Interaction" and "Mental Balance" of the HLQ correlate well with "mental health" and the "mental component summary", "Social Functionand "Vitality" of the SF-36 and Zerssens Bf-S Mood-Scale. The "motility" scale of the HLQ correlates with "physical function" and "vitality" of the SF-36, with the "severity of exhaustion" of the Giessener Physical Complaints Questionnaire GBB 24, and somewhat weaker with the "role physical", "bodily pain" and "physical component summary" scales of SF-36 and "limp pain" of the GBB 24. The "physical complaints" subscale of the HLQ correlates well with "bodily pain" of the SF-36 and its "physical component summary" scale, and also with the "affection pain" subscale of McGill's Pain Perception Scales SES. Among the SF-36 scales, the factor "general health" is not represented by the HLQ scales. The factors, "gastric symptoms" and the "heart symptoms" from the GBB 24 scales and "sensory pain" from the SES are not represented by the HLQ.
Table 5 HLQ-scales (Mean ± SD)) of patients separated into diagnostic-, age-and gender specific groups.
age gender n Initiative Power and Interest Social Interaction Mental Balance Motility Physical Complaints Digestive Well-Being
Connective tissue and soft tissue disorders 18–45 M 43 56.4 ± 15.2 67.4 ± 18.9 42.8 ± 17,2 42.0 ± 18.4 25.3 ± 15.8 73.3 ± 20.3
F 167 54.0 ± 15.2 64.2 ± 19.4 36.8 ± 15.8 38.7 ± 19.0 25.6 ± 17.4 67.4 ± 22.4
45–60 M 100 60.7 ± 16.9 69.9 ± 18.8 46.4 ± 16.0 42.8 ± 17.9 25.9 ± 14.0 77.0 ± 19.5
F 483 54.5 ± 15.1 60.2 ± 18.9 37.9 ± 15.1 37.0 ± 18.4 21.6 ± 16.9 67.6 ± 23.3
60–65 M 36 64.9 ± 16.0 73.1 ± 17.9 47.4 ± 19.4 41.3 ± 19.6 29.9 ± 17.7 79.9 ± 21.6
F 208 61.0 ± 15.5 66.9 ± 18.0 41.9 ± 15.0 40.3 ± 17.9 23.7 ± 16.3 74.3 ± 19.8
> 65 M 86 60.7 ± 16.9 70.7 ± 18.6 48.0 ± 16.5 36.9 ± 20.9 20.1 ± 19.2 75.1 ± 23.8
F 437 61.0 ± 16.2 69.2 ± 18.0 46.1 ± 16.8 38.3 ± 21.4 19.4 ± 17.6 70.0 ± 24.7
Chronic disorders of the respiratory system 18–45 M 5 60.4 ± 19.2 70.7 ± 14.6 46.4 ± 20.0 42.5 ± 24.6 27.5 ± 31.1 85.0 ± 22.4
F 12 50.2 ± 9.0 62.2 ± 23.2 34.8 ± 12.1 49.0 ± 17.6 35.4 ± 19.1 59.4 ± 30.7
45–60 M 2 58.7 ± 14.9 73.2 ± 12.6 58.9 ± 7.6 43.8 ± 8.9 37.5 ± 0.00 81.3 ± 26.5
F 14 61.9 ± 11.0 66.6 ± 14.4 37.9 ± 10.3 39.4 ± 18.7 27.9 ± 17.0 65.2 ± 18.5
60–65 M 6 76.6 ± 6.9 72.0 ± 6.5 47.6 ± 10.5 48.9 ± 28.1 25.0 ± 17.7 85.4 ± 12.3
F 12 55.1 ± 17.6 67.7 ± 16.4 37.9 ± 16.1 41.7 ± 16.7 31.8 ± 20.4 59.4 ± 35.0
> 65 M 22 57.0 ± 18.3 67.7 ± 20.1 45.9 ± 18.2 43.2 ± 23.1 29.4 ± 22.7 65.3 ± 27.0
F 22 56.0 ± 15.3 67.2 ± 18.8 42.9 ± 14.2 48.6 ± 22.4 31.8 ± 24.9 69.3 ± 21.4
Metabolic and nutritional disorders 18–45 M 15 54.4 ± 16.2 64.8 ± 20.9 48.1 ± 18.5 53.9 ± 18.4 56.7 ± 29.8 77.5 ± 19.0
F 24 58.1 ± 14.9 69.3 ± 21.8 40.7 ± 13.6 45.3 ± 20.5 42.4 ± 28.4 64.1 ± 25.4
45–60 M 25 55.9 ± 16.3 71.4 ± 18.3 45.8 ± 14.8 40.8 ± 16.4 43.0 ± 27.3 75.5 ± 21.5
F 40 59.4 ± 16.7 65.9 ± 18.4 44.1 ± 17.6 46.3 ± 20.4 39.1 ± 24.1 72.8 ± 24.0
60–65 M 20 70.9 ± 12.6 81.6 ± 13.0 56.7 ± 14.8 53.7 ± 18.4 38.1 ± 21.3 80.0 ± 15.9
F 16 67.9 ± 10.4 74.3 ± 15.2 54.2 ± 10.7 44.9 ± 21.8 39.1 ± 28.5 83.6 ± 15.6
> 65 M 32 66.2 ± 15.0 75.8 ± 14.8 53.0 ± 14.6 49.5 ± 22.3 41.8 ± 19.7 79.3 ± 19.5
F 53 62.1 ± 17.3 70.6 ± 16.6 50.9 ± 19.3 45.3 ± 23.2 30.9 ± 25.2 76.2 ± 20.3
Hypersensitivity and allergic reactions 18–45 M 4 59.1 ± 10.5 68.8 ± 14.7 53.6 ± 28.1 64.1 ± 16.4 28.1 ± 12.0 68.7 ± 26.0
F 16 63.3 ± 17.1 73.8 ± 17.8 38.3 ± 16.2 46.9 ± 20.9 45.3 ± 29.5 67.2 ± 26.6
45–60 M 2 55.8 ± 35.4 62.5 ± 42.9 51.8 ± 37.9 43.8 ± 44.2 25.0 ± 35.4 87.5 ± 17.7
F 15 58.1 ± 14.6 69.9 ± 16.2 42.3 ± 16.1 48.8 ± 17.9 37.5 ± 22.2 61.7 ± 25.6
60–65 M - - - - - - -
F 8 51.8 ± 18.1 62.1 ± 22.4 32.5 ± 11.1 52.3 ± 17.0 28.1 ± 16.0 62.5 ± 25.0
> 65 M - - - - - - -
F 6 71.3 ± 12.9 77.0 ± 7.1 55.2 ± 22.3 60.4 ± 21.2 22.9 ± 12.3 83.3 ± 18.8
other indications 18–45 M 18 47.6 ± 17.5 62.9 ± 15.6 40.7 ± 16.1 48.6 ± 18.4 35.4 ± 16.7 63.2 ± 23.3
F 122 53.8 ± 17.0 60.1 ± 19.8 35.2 ± 16.3 46.6 ± 21.5 35.9 ± 24.0 59.8 ± 24.2
45–60 M 28 57.2 ± 16.8 64.3 ± 25.0 41.8 ± 16.1 40.6 ± 19.1 34.8 ± 19.1 67.4 ± 27.1
F 111 53.3 ± 17.0 58.1 ± 18.4 35.1 ± 14.1 43.9 ± 19.5 31.3 ± 19.6 59.8 ± 25.1
60–65 M 15 55.3 ± 16.8 69.5 ± 14.1 44.8 ± 16.8 45.4 ± 18.8 40.8 ± 28.9 71.7 ± 21.9
F 50 53.4 ± 18.1 58.4 ± 17.3 40.4 ± 14.7 49.6 ± 19.0 31.5 ± 20.7 63.0 ± 26.5
> 65 M 23 65.8 ± 18.8 72.5 ± 16.7 52.8 ± 16.5 52.5 ± 24.6 34.2 ± 27.5 83.2 ± 19.4
F 78 59.0 ± 17.6 69.0 ± 18.9 47.8 ± 16.9 50.6 ± 21.8 30.2 ± 27.0 70.7 ± 26.6
According to the diagnostic spectrum (Table 5), the values of the scale "Motility (MOT)" and "Physical Complaints (PHY)" show particularly low values in patients suffering from rheumatic diseases. Also, in contrast with other scales of the HLQ, these two appear to have little correlation with age, which indicates a suitable discriminatory power of the HLQ considering age and different types of disease.
The results from the responsiveness analysis are presented in Table 6. We found a high sensitivity of the HLQ-scales to change within the treatment with particularly high significant changes in the mean and calculated effect sizes between 0.39 (Digestive Well Being) and 0.92 (Mental Balance).
Table 6 Responsiveness of HLQ-scales measured with Cohen's effect size.
Mean Difference [95% CI] (Admission-Discharge) t-value N Effect-Size ES
Initiative Power and Interest 8.1 [7.4; 8.7] 24.89 2064 0.55
Social Interaction 11.0 [10.3; 11.7] 30.18 2062 0.67
Mental Balance 15.8 [15.1; 16.5] 41.75 2066 0.92
Motility 11.4 [10.6; 12.3] 25.49 2050 0.57
Physical Complaints 21.7 [20.6; 22.8] 39.96 2022 0.89
Digestive well-Being 9.8 [8.7; 10.9] 17.78 2053 0.39
Discussion
The aim of our study was to confirm the structure and consistency of the HLQ. Surprisingly, we found that the original scales presented earlier [10] were not in accordance with the results of this factor analysis. However, the scales "IPI-Initiative Power and Interest", "SOCI – Social Interaction", "MB – Mental Balance", "MOT – Motility", "PHY-Physical Complaints", "DWB – Digestive Well-Being" show a good reliability and sufficiently differentiate the diagnostic groups, especially between those patients suffering with connective tissue and soft tissue disorders from those with metabolic and nutritional disorders or hypersensitivity reactions.
Although the HLQ sub-scales "Initiative Power and Interest", "Social Interaction" and "Mental Balance" of the HLQ correlate well with the corresponding SF-36 scales and with Zerssens Bf-S Mood-Scale, and thus indicate that these qualities share several interconnections, our findings also showed that the HLQ provides several aspects of health such as "Appetite and Digestive Affections" which are not well covered by existing QoL-measures. Nevertheless, with only two items, the subscale "digestive well-being" has to be strengthened by additional items. This is also true for the scale related to physical complaints and pain. With correlation values of 0.11 (physical total scale of the SF-36) and 0.29 (sensory pain SES), it is quite obvious that this scale is deficient and needs an upgrade in respect to quality and number of items.
As, according to [25] internal consistency reliability is a poor predictor of responsiveness, we measured the responsiveness of the HLQ directly using Cohen's effect size. Together with the highly significant results of the t-test statistics and being aware of the methodological limitations which are immanent in obtaining results on a questionnaires responsiveness by means of effect sizes [26], we can nevertheless conclude that the HLQ shows sufficient responsiveness for the use in a clinical setting.
In our opinion, the HLQ is more sensitive to health changes brought about by Complementary Therapies including anthroposophic medicine or homeopathy. This does not mean that the HLQ is only suitable for such therapies. Although, there is a trend to consider QoL-questionnaires being specific for special complementary therapies such as mistletoe treatment in cancer patients [27], we do not favor such labels, as this might result in an inflation of "new" QoL-measures for each new therapeutic situation [28].
QoL is a multidimensional construct composed of functional, physical, emotional, social and spiritual well-being [29,30] with, several interconnections between distinct constructs of well-being. The HLQ scales "Social Interaction", "Mental Balance", "Motility", and "Physical Complaints" share similarities with the these constructs, but highlights two further significant topics, i.e. "Initiative Power and Interest" and "Digestive Well-Being". The highly relevant topic of spirituality and illness is addressed in another instrument developed by our group, the SpREUK questionnaire, with its sub-scales "Search for meaningful support", "Positive interpretation of disease", "Trust in external guidance", "Support through spirituality/religiosity" [22,31,32].
Our evaluation indicates an adequate representation of aspects like "mental well-being" and "depression" which are essential in defining QoL, and shows special features of the HLQ that highlights its' uniqueness in the group of generic QoL-measures. Particularly in clinical studies in which, because of feasibility or patient compliance the use of huge psychometric test batteries is inappropriate, the HLQ now serves as a economic test-instrument. To conclude, we can state that this study presents necessary foundations and developments for existing and future studies that wish to use the HLQ as a reliable and valid instrument.
Authors' contributions
TO performed the statistical analysis, was responsible for the methodological setting of the study and has written most parts of the manuscript, AB has written some parts of manuscript, AMB is director of the study centre and was the clinical supervisor, PFM conceived and designed the study. All authors have read and approved the final manuscript.
==== Refs
Spitzer WO State of science 1986: Quality of life and functional status as target variables for research Journal of Chronic Diseases 1987 40 465 471 3597652 10.1016/0021-9681(87)90002-6
Frost MH Sloan JA Quality of life measurements: a soft outcome – or is it? Am J Manag Care 2002 8 574 9
Lorenz W Ollenschläger G Geraedts M Gerlach FM Gandjour A Helou A Kirchner H Koller M Lauterbach K Reinauer H Sitter H Thomeczek C Das Leitlinien-Manual. Entwicklung und Implementierung von Leitlinien in der Medizin ZaeFQ 2001 95 1 84
Bradley C Gamsu DS Guidelines for encouraging psychological well-being: Report of a working group of the World Health Organisation Regional Office for Europe and International Diabetes Federation European Region St Vincent Declaration Action Programme for Diabetes Diabetic Medicine 1994 11 510 516 8088133
Patrick DL Deyo RA Generic and disease-specific measures in assessing health status and quality of life Medical Care 1989 27 217 232
Gold S Heesen C Schulz H Guder U Mönch A Gbadamosi J Disease specific quality of life instruments in multiple sclerosis: Validation of the Hamburg Quality of Life Questionnaire in Multiple Sclerosis (HAQUAMS) Multiple Sclerosis 2001 7 119 130 11424632 10.1191/135245801678227649
Finlay AY Khan GK Dermatology Life Quality Index (DLQI) – A simple practical measure for routine clinical use Clinical and Experimental Derm 1994 19 210 16
Colwell HH Mathias SD Pasta DJ Henning JM Hunt RH Development of a health-related quality-of-life questionnaire for individuals with gastroesophageal reflux disease: a validation study Dig Dis Sci 1999 44 1376 83 10489923 10.1023/A:1026647701477
Hannoveraner Konsens Empfehlungen zur gesundheitsökonomischen Evaluation Zeitschrift für Allgemeinmedizin 1996 72 485 490
Schulte M Kümmell HC Entwicklung eines Fragebogens zur Lebensqualität auf der Grundlage des anthroposophischen Menschenbildes Der Merkurstab 1996 49 109 122
Schulte M Heckmann C Kümmell HC Heusser P Der Herdecker Fragebogen zur Lebensqualität (HLQ): Entwicklung, Ergebnisse, Verwendungsmöglichkeiten Akademische Forschung in der Anthroposophischen Medizin 2000 Bern: Peter Lang Verlag 249 253
Doerfler W Wisniowski C Gerhard I Behandlung des klimakterischen Symptomenkomplexes mit der Traditionellen Chinesischen Medizin (TCM). 169. Tagung der Mittelrheinischen Gesellschaft für Geburtshilfe und Gynaekologie, Homburg/Saar, München: Alete Wissenschaftl Dienst 1997 153
Beer AM Ostermann T Matthiessen PF Evaluation of in-hospital treatment with naturopathic methods – the Blankenstein model. Part I: Patients and therapeutic concepts Forsch Komplementarmed Klass Naturheilkd 2001 8 6 13 11340309 10.1159/000057189
Ostermann T Beer AM Matthiessen PF Evaluation of inpatient naturopathic treatment – the Blankenstein model. Part II: Effective strength and health status of patients over the course of time Forsch Komplementarmed Klass Naturheilkd 2002 9 269 76 12417803 10.1159/000067520
Bullinger M Kirchberger I SF-36 Fragebogen zum Gesundheitszustand, Handanweisung 1998 Göttingen: Hogrefe Verlag für Psychologie
Zerssen D Die Befindlichkeitsskala 1976 Beltz-Test GmbH
Brähler E Scheer JW Der Gießener Beschwerdebogen (GBB) 1995 Bern: Verlag Hans Huber
Geissner E Die Schmerzempfindungsskala SES – Ein differenziertes und veränderungssensitives Verfahren zur Erfassung chronischer und akuter Schmerzen Rehabilitation 1995 34 XXXV XLIII 8570898
Beer AM Ostermann T Matthiessen PF Changes in quality of life during acute inpatient naturopathic treatment – results of the Blankenstein model Gesundheitswesen 2001 63 242 247 11367955 10.1055/s-2001-12917
Kroz M von Laue HB Zerm R Girke M Development of a questionnaire for endogenous regulation – a contribution for salutogenesis research Forsch Komplementarmed Klass Naturheilkd 2003 10 70 77 12808365 10.1159/000071665
Grimley DM Bellis JM Prochaska JO Riley GE Assessing the stages of change and decision making for contraceptive use for the prevention of pregnancy, sexually transmitted diseases, and acquired immunodeficiency syndrome Health Educ Q 1993 20 455 470 8307766
Büssing A Ostermann T Matthiessen PF Role of Religion and Spirituality in Medical patients – Confirmatory results with the SpREUK questionnaire Health Quality Life Outcomes 2005 3 10 10.1186/1477-7525-3-10
Schumacher J Leppert K Gunzelmann T Strauss B Braehler E Die Resilienzskala – Ein Fragebogen zur Erfassung der psychischen Widerstandsfähigkeit als Personmerkmal Z f Klinische Psychologie, Psychiatrie und Psychotherapie 2004 53 16 39
Hays RD Hayashi T Beyond internal consistency reliability: rationale and user's guide for the Multitrait Analysis Program (MAP) on the microcomputer Behav Res Methods Instrum Comput 1990 22 167 175
Puhan MA Bryant D Guyatt GH Heels-Ansdell D Schunemann HJ Internal consistency reliability is a poor predictor of responsiveness Health and Quality of Life Outcomes 2005 3 33 15877824 10.1186/1477-7525-3-33
Middel B van Sonderen E Statistical significant change versus relevant or important change in (quasi) experimental design: Some conceptual and methodological problems in estimating magnitude of intervention-related change in health services research International Journal of Integrated Care 2002 2 1 22
Kirchberger I Wetzel D Finger T Development and validation of an instrument to measure the effects of a mistletoe preparation on quality of life of cancer patients: the Life Quality Lectin-53 (LQL-53) Questionnaire Qual Life Res 2004 13 463 479 15085919 10.1023/B:QURE.0000018481.33943.97
Ostermann T Fragen zur Mistel. Journal Club Forsch Komplementärmed 2004 11 361 62
Peterman AH Fitchett G Brady MJ Hernandez L Cella D Measuring spiritual well-being in people with cancer: the functional assessment of chronic illness therapy – Spiritual Well-being Scale (FACIT-Sp) Ann Behav Med 2002 24 49 58 12008794 10.1207/S15324796ABM2401_06
Brady MJ Peterman AH Fitchett G Mo M Cella D A case for including spirituality in quality of life measurement in oncology Psychooncology 1999 8 417 428 10559801 10.1002/(SICI)1099-1611(199909/10)8:5<417::AID-PON398>3.3.CO;2-W
Ostermann T Bussing A Matthiessen PF Pilot study for the development of a questionnaire for the measuring of the patients' attitude towards spirituality and religiosity and their coping with disease (SpREUK) Forsch Komplementarmed Klass Naturheilkd 2004 11 346 353 15604625 10.1159/000082816
Büssing A Ostermann T Matthiessen PF Search for meaningful support and the meaning of illness in German cancer patients Anticancer Res 2005 25 1449 55 15865104
| 16004612 | PMC1177980 | CC BY | 2021-01-04 16:38:13 | no | Health Qual Life Outcomes. 2005 Jul 8; 3:40 | utf-8 | Health Qual Life Outcomes | 2,005 | 10.1186/1477-7525-3-40 | oa_comm |
==== Front
Int J Health GeogrInternational Journal of Health Geographics1476-072XBioMed Central London 1476-072X-4-131594386310.1186/1476-072X-4-13ResearchAn information value based analysis of physical and climatic factors affecting dengue fever and dengue haemorrhagic fever incidence Nakhapakorn Kanchana [email protected] Nitin Kumar [email protected] Remote Sensing and GIS field of study, Asian Institute of Technology, Pathumthani, Thailand2005 8 6 2005 4 13 13 24 12 2004 8 6 2005 Copyright © 2005 Nakhapakorn and Tripathi; licensee BioMed Central Ltd.2005Nakhapakorn and Tripathi; 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
Vector-borne diseases are the most dreaded worldwide health problems. Although many campaigns against it have been conducted, Dengue Fever (DF) and Dengue Haemorrhagic Fever (DHF) are still the major health problems of Thailand. The reported number of dengue incidences in 1998 for the Thailand was 129,954, of which Sukhothai province alone reported alarming number of 682. It was the second largest epidemic outbreak of dengue after 1987. Government arranges the remedial facilities as and when dengue is reported. But, the best way to control is to prevent it from happening. This will be possible only when knowledge about the relationship of DF/DHF with climatic and physio-environmental agents is discovered. This paper explores empirical relationship of climatic factors rainfall, temperature and humidity with the DF/DHF incidences using multivariate regression analysis. Also, a GIS based methodology is proposed in this paper to explore the influence of physio-environmental factors on dengue incidences. Remotely sensed data provided important data about physical environment and have been used for many vector borne diseases. Information Values (IV) method was utilised to derive influence of various factors in the quantitative terms. Researchers have not applied this type of analysis for dengue earlier. Sukhothai province was selected for the case study as it had high number of dengue cases in 1998 and also due to its diverse physical setting with variety of land use/land cover types.
Results
Preliminary results demonstrated that physical factors derived from remotely sensed data could indicate variation in physical risk factors affecting DF/DHF. A composite analysis of these three factors with dengue incidences was carried out using multivariate regression analysis. Three empirical models ER-1, ER-2 and ER-3 were evaluated. It was found that these three factors have significant relation with DF/DHF incidences and can be related to the forecast expected number of dengue cases. The results have shown significantly high coefficient of determination if applied only for the rainy season using empirical relation-2 (ER-2). These results have shown further improvement once a concept of time lag of one month was applied using the ER-3 empirical relation. ER-3 model is most suitable for the Sukhothai province in predicting possible dengue incidence with 0.81 coefficient of determination. The spatial statistical relationship of various land use/land cover classes with dengue-affected areas was quantified in the form of information value received from GIS analysis. The highest information value was obtained for the Built-up area. This indicated that Built-up area has the maximum influence on the incidence of dengue. The other classes showing negative values indicate lesser influence on dengue epidemics. Agricultural areas have yielded moderate risk areas based on their medium high information values. Water bodies have shown significant information value for DF/ DHF only in one district. Interestingly, forest had shown no influence on DF/DHF.
Conclusion
This paper explores the potential of remotely sensed data and GIS technology to analyze the spatial factors affecting DF/DHF epidemic. Three empirical models were evaluated. It was found that Empirical Relatrion-3 (ER-3) has yielded very high coefficient of determination to forecast the number of DF/DHF incidence. An analysis of physio-environmental factors such as land use/ land cover types with dengue incidence was carried out. Influence of these factors was obtained in quantitative terms using Information Value method in the GIS environment. It was found that built-up areas have highest influence and constitute the highest risk zones. Forest areas have no influence on DF/DHF epidemic. Agricultural areas have moderate risk in DF/DHF incidences. Finally the dengue risk map of the Sukhothai province was developed using Information Value method. Dengue risk map can be used by the Public Health Department as a base map for applying preventive measures to control the dengue outbreak. Public Health Department can initiate their effort once the ER-3 predicts a possibility of significant high dengue incidence. This will help in focussing the preventive measures being applied on priority in very high and high-risk zones and help in saving time and money.
==== Body
Background
Vector borne diseases are the most common worldwide health hazard and represent a constant and serious risk to a large part of the world's population. Among these, dengue fever especially is sweeping the world in majority of the tropical and arid zones. It is transmitted to the man by the mosquito of the genus Aedes and exists in two forms: the Dengue Fever (DF) or classic dengue and the Dengue Haemorrhagic Fever (DHF), which may evolve into Dengue Shock Syndrome (DSS)[1]. Dengue infection occurs due to the bite of the mosquito Aedes aegypti, that is infected with one of the four dengue virus serotypes [2]. The infection, earlier restricted to urban/semi-urban centres, can now be seen in rural areas as well [3]. Land use/ land cover types and climate play significant role in dengue cases as reported by several researchers [3]. Remotely sensed data can be used to identify, monitor and evaluate environmental factors between vector and environment relationships. Recently, Geographic Information Systems (GIS) and Remotely Sensed data are being used to evaluate and model the relationships between climatic and environmental factors with the incidences of viral diseases. Spatial analysis involves the use of Geographic Information Systems (GIS) for health that has been reviewed by several authors [4-9]. Both spatial and temporal changes in environmental condition may be important determinants of vector-borne disease transmission. Remote sensing data can be used to provide information on the spatial distribution of the vector-borne diseases and the physical environment [10-12]. It is mentioned by a researcher that remote sensing and geodesy have the potential to revolutionize the discipline of epidemiology and its application in human health [13].
Remote sensing and other types of data were used in GIS to identify villages at high risk for malaria transmission in the southern area of Chiapas, Mexico [12]. In Nigeria, a temporal analysis of Landsat Thematic Mapper (TM) satellite data was carried out to test the significance of the guinea worm eradication program based on changes in agricultural production [11]. And, it was also employed to predict and map the location of some of the major diseases affecting human health as well [10,14]. Land use/ land cover types are the critical variables in epidemiology and can be characterized by remote sensing [15]. At the same time, it was observed that use of remote sensing and GIS in health sector is quite limited in Thailand.
Dengue Fever (DF) and Dengue Haemorrhagic Fever (DHF) has become a major international public health concern. Many countries/areas in Asia have been experiencing unusually high levels of dengue/dengue haemorrhagic fever activity in 1998 [16]. The reported number of dengue incidences in 1998 for the Thailand was 129,954, which was the second largest epidemic outbreak of dengue after 1987. In 1998, Sukhothai province, Thailand alone reported alarming number of 682 dengue cases.
A GIS based methodology is proposed in this paper to explore the influence of physio-environmental and climatic factors on dengue incidences. Information Values (IV) method was utilised to derive influence of various physio-environmental factors in the quantitative terms. Researchers have not applied this method of analysis for dengue earlier. Sukhothai province was selected for the case study as it had high number of dengue cases in 1998 and also due to its' diverse physical setting with variety of land use and land cover types.
Data and methods
Study area and data used
Sukhothai province, located in the northern Thailand, was selected as the study area (Figure 1). This province consists of 9 Districts: Muang Sukhothai (D1), Ban Dan Lan Hoi (D2), Khiri Mat (D3), Kong Krailat (D4), Sawankhalok (D5), Si Nakhon (D6), Si Samrong (D7), Si Satchanalai (D8) and Thung Saliam (D9). GIS database was developed for each of the districts. People are predominantly involved in the agriculture such as sugarcane, cassava, and corn. The province has a population of about 521,219. Climate of this area is subtropical with extreme high temperatures rising to 42°C in April and dipping low up to 13.2°C in December. In 1998, the average annual temperature was 28°C, which was higher than the 30 Year average normal temperatures of 26°C. Rainy season in Thailand normally occurs from May to September. The daily maximum and annual rainfall are 62.5 mm and 917.7 mm respectively. The average relative humidity is 94.8 percent.
Figure 1 Location map of study area, Sukhothai, Thailand.
Medical data
Medical data were collected from the Provincial Health Office, Thailand, which collects monthly district level data. In 1998, 682 DF/DHF cases were reported in Sukhothai Province. Morbidity rate was observed at 131 per 100,000 people. The DHF incidences were recorded at the village level. Highest numbers of dengue incidence were recorded in districts Muang, Si Satchanalai and Sawan Khalok (Table 1). It was found that highest number of cases occurred during March and August. This indicated the seasonal dependence in occurrence of DF/DHF cases, which generally starts just before the rainy season and continues till the end of rainy season.
Table 1 Monthly distribution of DF/DHF cases in Sukhothai province in 1998
District Name Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Total
Muang Sukhothai Thani 18 31 40 7 3 1 3 15 16 5 2 0 141
Kong Krai Lat 3 5 6 2 2 3 6 4 4 0 0 0 35
Khili Mat 15 8 21 12 6 14 11 8 5 2 0 0 102
Thung Saliam 5 1 4 4 3 1 3 14 4 3 4 6 52
Ban Dan Lan Hoi 6 5 3 6 6 9 8 9 3 4 0 0 59
Si Satchanalai 5 4 7 9 23 14 20 22 6 2 7 4 123
Si Samrong 5 3 7 3 4 9 9 4 7 0 0 0 51
Sawan Khalok 12 4 15 12 5 2 11 15 7 7 9 7 106
Si Nakhon 1 1 3 3 0 0 0 4 1 0 0 0 13
Factors influencing dengue
Major factors considered for analysis of the occurrence of DF/DHF cases were rainfall, temperature, humidity, and land use/land cover types. DF/DHF outbreaks in Sukhothai, Thailand occurred in 1997, 1998 and 2001. It was noticed that the dengue outbreak coincided with El Nino years. El Nino events in Thailand are normally associated with high temperature and low precipitation. The monthly rainfall, temperature and relative humidity data were collected from the Department of Meteorology, Ministry of Information and Communication Technology, Thailand. Thailand experiences rains from May to September and remaining part of the year remains mostly dry. In addition to the rainfall, temperature, and humidity also influence dengue transmission [17]. Due to high humidity during rainy season mosquito survival is longer and growth is facilitated [18]. The average temperatures in Sukhothai province were between 22°C and 33°C in 5 years (1997–2001). Higher than 20°C is the favourable temperature for Aedes aegypti mosquitoes [2]. The average relative humidity observed in 5 years (1997–2001) was 95.6 percent. The 30 years (1969–2000) average monthly rainfall was 1226.5 mm in Northern Regions of Thailand. As the season in Thailand can be divided in broader context into two types: rainy and non-rainy, the effect of these three factors on dengue was analysed for these two parts of the year.
Physio-environmental factors
The area of Sukhothai province is 6,694.54 km2. Digital remote sensing data from Landsat (Thematic Mapper) were employed to produce the land cover type map using the Maximum Likelihood Classifier (MLC). Various output classes generated were subsequently verified based on the field observations. Maximum Likelihood Classifier offered 86 percent classification accuracy. The satellite data were geo-referenced using the Royal Thai Survey Department (RTSD) base map. The GIS analysis revealed the land use areas as: agricultural (74.7%), forest (21%), water bodies (0.3%) and Built-up (4.0%) for the year 1998. Resultant map from the classification is shown in the figure 2.
Figure 2 Land use/land cover map showing DF/DHF affected settlement in Sukhothai, Thailand.
Methodology
Dependence of the climatic variables was evaluated using multiple regression analysis. In addition, as suggested by many researchers, physio-environmental factors also affect dengue incidence. Information value approach was utilized to explore which physical and environmental factors are more crucial in dengue incidences.
Relationship of dengue incidence with climatic factors using multiple regression analysis
The general purpose of multiple regressions is to learn more about the relationship between several independent or predictor variables and a dependent or criterion variable. The general computational problem that needs to be solved in multiple regression analysis is to fit a straight line to a number of points. In the multivariate case, when there is more than one independent variable, the regression line cannot be visualised in the two-dimensional space, but can be computed just as easily. It is possible to construct a linear equation containing all variables. In general multiple regression procedures will estimate a linear equation of the form:
Y = b0+b1X1+b2X2+...+bkXk
Where k is the number of predictors. Note that in this equation, the regression coefficients (or b0, b1, b2...bk coefficients) represent the independent contributions of each independent variable to the prediction of the dependent variable [19]. In present study climatic independent variables such as rainfall, temperature and humidity were related to the dengue cases in the Sukhathai province. The application of this method is detailed in following section of Results.
Information value method
An understanding of spatial relationship of Dengue epidemic with affecting factors is essential before applying any statistical models to find the influence of factors in dengue epidemic. The simple technique to understand the statistical relationship is conditional analysis, which attempts to assess the probabilistic relationship between relevant factors affecting dengue epidemic and environmental factors. The technique is based on Baye's theorem (Bayesian classifier) according to which frequency data can be used to calculate the probabilities that depends upon the knowledge of previous events or dengue epidemic/outbreak [20-22].
The information value equation is expressed as the natural logarithmic ratio of conditional probability of the thematic feature.
Where, Ij is the predictive information of occurrence of event D if feature A is present under state j
P{D/Aj} is the conditional probability of event D to occur under the condition of feature A and state j.
This is the same as the conditional probability of dengue epidemic (D) to occur because of spatial feature (A) present in one of the thematic(jth) layer.
P{D} is the probability that event D will happen in the selected area irrespective of any evidence.
Map crossing results in a cross table showing the number of pixels per class occupied by dengue epidemic and total number of pixels in each class. The remaining values, necessary to calculate information value, area obtained from these values, actually can represent in term of information value process (Figure 2), therefore
where,
i-value = information value
ndclass = area with dengue epidemic in a class(i.e. land use/land cover types in the buffer zone)
nclass = area in the class (i.e. land use/cover types by district)
ndmap = total area of dengue epidemic in the map (thematic layer)
nmap = total area in the map
Information value of various parameters can be utilized to interpret the relationship of the parameters and dengue incidences (Figure 3).
Figure 3 Methodology for mapping DF/DHF risk zones using Information Value approach.
Buffering
Buffering operation was carried out to determine extended neighbourhood of place of occurrence. Results from the human settlement i.e. urban or sub-urban centre in Sukhothai province were obtained from maps obtained from Ministry of Interior, Thailand. Buffers were created to specify distance of the point pattern of disease from urban/ village locations and to identify the geographic environment conditions such as land cover, water bodies, surrounding village affected by DHF. The buffer distance was considered owing to two factors: flight distance covered during the life span and average distance travelled per day by the Aedes aegypti mosquito. The average lifespan of female is about 8–15 days and the female mosquito can fly about 30–50 m per day on an average. This indicates that in general female mosquito would move around 240–600 m range in their lifetime [23]. These buffer zones were divided into two groups according to the distance from life span, and average distance travelled per day by the Aedes aegypti mosquito. Group 1 was composed of areas within 500 m and Group 2 composed of an area of 1000 m around the dengue incidence. Therefore, the buffer zones of 500 m and 1000 m were created using GIS for this study.
Results and discussion
Regression analysis was used to explore the relationship between the monthly climatic parameters and the number of incidences of DF/DHF in Sukhothai province.
Model development
Multiple regression analysis is employed to develop an empirical model to predict the dengue incidences. The independent variables were used to predict changes in the dependent variable in the rainy and non-rainy seasons. This model was verified using the R2 statistics. The variables used in the models are explained in the Table 2.
Table 2 Variable in the multiple regression models
Dependent variable
Dt Number of DF/DHF cases based on monthly report from the Sukhothai Provincial Public Health office, Ministry of Public Health
Independent variables
T Maximum monthly temperature(°C)
R Total monthly rainfall (mm./month)
H Maximum monthly relative humidity(%)
t time in month
t-1 one month before t
Number of peoples affected by DF/DHF was used as the dependent variable and the rainfall(R), temperature(T) and relative humidity(H) were considered as the independent variables. Multiple regression analysis was carried out for each of the observations of the occurrence of DF/DHF cases and monthly climatic data of 5 years (1997–2001). The Empirical Relationship-1 (ER-1) between number of DF/DHF cases and the climatic data at time t (Tt, Rt and Ht) during 5 years is listed in ER-1. The coefficient of determination (R2) was found as 0.43 and validated with climate data as shown in the graph in Figure 4.
Figure 4 Relationship between actual and forecasted DF/DHF cases in Sukhothai, Thailand.
ER-1:
Dt = 1408.318+0.151(Rt)-4.368(Tt)-12.798(Ht)
Dt : Number of DF/DHF patients reported in the month t
It was observed that the coefficient of determination (R2) for this model was low for the non-rainy season but high for the rainy season (Figure 4).
Multiple regression analysis was carried out for the occurrence of DF/DHF cases for the rainy season. The Empirical Relationship-2 (ER-2) between DF/DHF cases and rainfall, temperature and humidity factors at time t (Tt, Rt and Ht) was as follows:
ER-2:
Dt = -13.893+1.3444(Tt)-0.276(Ht)+0.377(Rt)
The coefficient of determination (R2) was found as 0.62.
According to the development period from egg to human disease, there is a time lag of about one month that leads to DF/DHF cases occurring during 7 – 45 days. The duration of larvae stages to adult is 7 to 12 days and the lifespan of for female mosquito is about 8 to 15 days [17]. Meantime, the virus develops in the mosquito for a period of 8–10 days. By the time a person infected with dengue virus develops fever the infection is widely disseminated to many people. The virus is found in serum or plasma, in circulating blood cells and in selected tissues, especially those of the immune system, for approximately 2–7 days, roughly corresponding to the period of fever [2]. Thus, DF/DHF cases at time t (in month i.e. May) depends on others factors at time t-1 (one month before t i.e. t-1 or the month April). In this empirical relation, the regression coefficients represent the independent contributions of each variable to the prediction of the dependent variable.
The Empirical Relationship- 3(ER-3) with a one-month time lag as shown below offered a coefficient of determination (R2) as 0.81. This has shown considerable 31 percent increase than ER-2 and 88 percent than ER-1. The results have shown considerable improvement.
ER-3:
Dt = 621.824+0.345(Rt-1)- 0.609(Tt-1)-6.321(Ht-1)
Where,
t = Time (in the unit of month) used as suffix to indicate the month for which the data belong
t-1 = one month before the month t
Therefore, the ER-3 was selected to model the DF/DHF incidence in Sukhothai in as the closest output to the actual data during rainy season and the results were validated with 1998 data (Figure 5).
Figure 5 The relationship between actual and forecasted DF/DHF cases in Sukhothai, Thailand in 1998 during rainy season using ER-3.
Information value computation and analysis
Information values were calculated for each land cover type. Negative values indicate low-risk level and the positive values indicate high-risk level of Dengue. Six possible risk classes were identified; Information-value scores obtained from the quartile of a data set at each range of 0.5 were used as the cut-off levels. Risk classes were designated as very low, low, moderately low, moderately high, high and very high respectively with scores as: less than -1.0, -1.0 to -0.5, -0.5 to 0.0, 0.0 to 0.5, 0.5 to 1.0 and greater than 1.0 respectively (Figure 6).
Figure 6 Histogram between information values and area pixels affected by dengue.
The highest information value was obtained for the Built-up area. This indicated that Built-up area has the maximum influence on the incidence of dengue. The other classes showing negative values indicate lesser influence on dengue epidemics. Two buffer regions of 500 m and 1000 m in the land use/ land cover maps were analysed for the information value (Figure 7). It was done to investigate if the changes of neighbourhood sizes have any effect on the information value or not.
Figure 7 Dengue risk classes using 500 and 1000 m buffer zones.
Information values were clustered in 4 groups: first group represented the high positive values i.e. Built-up (BU) area. Information value in this group represented the highest value that meant built-up category is the highest spatial risk factor in all the districts both for 500 and 1000 m buffer zones. Second group represented the positive relationship with Water Bodies (WB). Information value in this group represented the high values for both buffer regions of 500 and 1000 m in D1, D6 and D7 districts. Third group represented the positive values for the Agriculture areas. Information value in this group represented positive values for both 500 and 1000 m buffer zones that mean the Agriculture category has shown positive influence as a risk factors in D2, D3, D4, D5, D6 and D7 district. Fourth group represented that for all the districts forest (FR) area indicated no risk for dengue, as information values were negative for both 500 and 1000 m buffer zones. The information value obtained using Land use/Land cover risk zone is shown in Figure 8 and Table 3.
Figure 8 Dengue risk zones received for physical environment using information value.
Table 3 Information value computation and analysis for Sukhothai province in physical factors
Physical Factors District Code District Name I-value 500 m Risk Classes500 I-value 1000 m Risk Classes1000
Built-up Area (BU) D1 Muang 0.56 High 0.33 Moderately High
D2 Kong Krai Lat 0.41 Moderately High 0.31 Moderately High
D3 Khili Mat 0.48 Moderately High 0.45 Moderately High
D4 Thung Saliam 1.32 Very High 1.05 Very High
D5 Ban Dan Lan Hoi 1.07 Very High 0.82 High
D6 Si Satchanalai 1.37 Very High 1.14 Very High
D7 Si Samrong 0.62 High 0.41 Moderately High
D8 Sawan Khalok 0.12 Moderately High 0.18 Moderately High
D9 Si Nakhon -0.04 Moderately Low -0.08 Moderately Low
Water Bodies (WB) D1 Muang 0.08 Moderately High 0.17 Moderately High
D2 Kong Krai Lat -2.40 Very Low -0.64 Low
D3 Khili Mat -0.20 Moderately Low -0.34 Moderately Low
D4 Thung Saliam -0.56 Low -0.82 Low
D5 Ban Dan Lan Hoi -0.19 Moderately Low -0.14 Moderately Low
D6 Si Satchanalai 0.95 High 0.87 High
D7 Si Samrong 1.66 Very High 1.13 Very High
D8 Sawan Khalok -0.13 Moderately Low -0.28 Moderately Low
D9 Si Nakhon -1.41 Very Low -1.33 Very Low
Agriculture Area (AGR) D1 Muang -0.02 Moderately Low 0.00 Moderately Low
D2 Kong Krai Lat 0.03 Moderately High 0.02 Moderately High
D3 Khili Mat 0.23 Moderately High 0.22 Moderately High
D4 Thung Saliam 0.19 Moderately High 0.18 Moderately High
D5 Ban Dan Lan Hoi 0.24 Moderately High 0.25 Moderately High
D6 Si Satchanalai 0.19 Moderately High 0.19 Moderately High
D7 Si Samrong 0.04 Moderately High 0.05 Moderately High
D8 Sawan Khalok 0.00 Moderately Low -0.01 Moderately Low
D9 Si Nakhon 0.01 Moderately High 0.00 Moderately Low
Forest Area (FR) D1 Muang -0.67 Low -0.48 Moderately Low
D2 Kong Krai Lat -0.81 Low -0.51 Low
D3 Khili Mat -1.35 Very Low -1.24 Very Low
D4 Thung Saliam -1.38 Very Low -1.01 Very Low
D5 Ban Dan Lan Hoi -1.83 Very Low -1.61 Very Low
D6 Si Satchanalai -1.25 Very Low -1.00 Very Low
D7 Si Samrong -2.30 Very Low -1.82 Very Low
D8 Sawan Khalok -0.40 Moderately Low -0.39 Moderately Low
D9 Si Nakhon -0.16 Moderately Low 0.27 Moderately High
Most of the risk zones areas are located in the moderately risk class both for 500 meter and 1000 m buffer zones. Information values representing the very high and high-risk classes were in built-up area and water body categories as shown in the Figure 8. Risk zones showing the negative or low risk classes are mostly in the forest area. Sithiprasasna and Linthicum have also shown that DHF incidence was poorly correlated with the forest cover in Tak province of Thailand [24]. It confirms that the DF/DHF cases mostly occur in urban and suburban areas. Table 2 shows the physical categories and their Information Values for each of the districts. Very High Risk zone (information value > 1.0) was in D4, D5, and D6 districts. High Risk zone (information values between 1.0 and 0.5) were observed in D7 and D1 districts. The spatial statistical relationship of various land use/land cover classes with dengue-affected areas was quantified in the form of information value and a dengue risk map was generated as shown in the Figure 9.
Figure 9 DF/DHF risk zone map of Sukhothai, Thailand.
Conclusion
This paper has offered some useful information related to the dengue incidence. Analysis of the climatic factors such as rainfall, temperature and humidity with the dengue incidences has revealed that dengue generally occurred when average temperature rose above normal as found during the El-nino in 1998. It also occurred when the rainfall was comparatively lower and humidity was higher than average.
A composite analysis of these three factors with dengue incidences was carried out using multivariate regression analysis. Three empirical models ER-1, ER-2 and ER-3 were evaluated. It was found that these three factors shave significance and can be related to the find expected number of dengue cases. The results have shown significant high coefficient of determination if applied only for the rainy season using empirical relation-2 (ER-2). These results have shown further improvement once a concept of time lag of one month was applied using the ER-3 empirical relation. ER-3 model is most suitable for the Sukhothai province in predicting possible dengue incidence with 0.81 coefficient of determination.
An analysis of physio-environmental factors such as land use/ land cover types with dengue incidence was carried out. The aim of this analysis was not only to find the effect of physio-environmental factors on dengue incidences but also to find influence of these factors in quantitative terms. It was found that built-up areas have highest influence and constitute the highest risk zones. The agriculture areas offered the second level of high-risk influence. Water bodies posed significant risk in only one district. Forest areas almost do not have any influence on the dengue risk zonation. Information value approach applied to develop the dengue risk map of the Sukhothai province. Dengue risk map can be used by the Public Health Department for applying preventive measures to control the dengue outbreak. This map takes into account the physio-environmental factors and also the mobility and the life duration of the Aedes Aegypti mosquitoes. Public Health Department can initiate their effort once the ER-3 predicts a possibility of significant high dengue incidence. This will help in focussing the preventive measures being applied on priority in very high and high-risk zones and save time and money.
Authors' contributions
Authors KN and NKT collaborated intensely on all aspects of the manuscript, from research design to data preparation. KN carried out most of the statistical and GIS analysis using Information Value approach and drafted the manuscript. Both authors read and approved the final manuscript. The manuscript was vastly revised and improved based on the reviewers' comments.
Acknowledgements
We would like to acknowledge the invaluable assistance of Dr. K. Nualchawee, Professor M. Kusanagi, and Dr. P. Parkpian. We are also extremely grateful to Sukhothai Provincial Public Health office, Ministry of Public Health, and Department of Meteorology, Ministry of Information and Communication Technology, Thailand for data and information.
==== Refs
Derouich M Boutayeb A Twizell E A model of dengue fever Biomed Eng Online 2003 19 4 12657162 10.1186/1475-925X-2-4
World Health Organization Dengue haemorrhagic fever: Diagnosis, treatment, prevention and control Geneva 1997 2 World Health Organization: Geneva
World Health Organization Dengue Bulletin 2001 25
Mayer JD The role of spatial analysis and geographic data in the detection of disease causation Soc Sci Med 1986 17 1213 1221 6623127 10.1016/0277-9536(83)90014-X
Gesler W The uses of spatial analysis in medical geography: a review Soc Sci Med 1986 23 963 973 3493536 10.1016/0277-9536(86)90253-4
Twigg L Health based geographical information systems: Their potential examined in the light of existing data sources Soc Sci Med 1990 30 143 155 2305277 10.1016/0277-9536(90)90337-R
Marshal RJ A review of methods for the statistical analysis of spatial patterns of disease JR Statist Soc 1991 A 154 421 441
Scholden HJ de Lepper MJC The benefits of the application of geographical information systems in public and environment health Raprr Trimest Statist Sanit Mond 1991 44 60 170
Walter SD Visual and statistical assessment of spatial clustering in mapped data Stat Med 1993 12 1275 1291 8210826
Hay SI Packer MJ Rogers DJ The impact of remote sensing on the study and control of invertebrate intermediate hosts and vectors for the disease International Journal of Remote Sensing 1997 18 2899 2930 10.1080/014311697217125
Ahearn SC De Rooy C Monitoring the effect of dracunculiasis remediation of agricultural productivity using satellite data International Journal of Remote Sensing 1996 17 917 929
Beck LR Rodrigues MH Dister SW Rodrigues AD Rejmankova E Ulloa A Remote sensing as a landscape epidemiologic tool to identify villages at high risk for malaria transmission Am J Trop Med Hyg 1994 51 271 80 7943544
Hay SI Hay SI, Randolph SE, Rogers DJ An Overview of Remote Sensing and Geodesy for Epidemiology and Public health Application 2000 Oxford: Academic Press 1 35 [Baker JR, Rollinson D(Series Editor): Advances in Parasitology: Remote Sensing and Geographical Information Systems in Epidemiology, Vol.47].
Hay SI Snow RW Rogers DJ From predicting mosquito habitat to malaria seasons using remotely sensed data: practices, problems and perspectives Parasitology Today 1998 4 1306 313
Curran PJ Atkinson PM Foody GM Milton EJ Hay SI, Randolph SE, Rogers DJ Linking Remote Sensing, Land Cover and Disease 2000 Oxford: Academic Press 37 80 [Baker JR, Rollinson D(Series Editor): Advances in Parasitology: Remote Sensing and Geographical Information Systems in Epidemiology, Vol.47]. 10997204
World Health Organization Dengue Bulletin 1999 23
Gubler DJ Kuno G Dengue and dengue hemorrhagic fever 1997 New York, NY: CAB International
Jetten JH Focks DA Changes in the distribution of dengue transmission under climate warming scenarios American Journal of Tropical Medicine and Hygiene 1997
Abraham B Ledolter J Statistical methods for forecasting 1983 New York: Wiley
Yin KL Yan TZ Statistical Prediction models for slope instability of metamorphosed rocks Proceedings of the 5th International Symposium on Landslides, Lausanne 1998 2 1269 1272
Prakash C Landslide Hazard Zonation in Doon Valley using Remote Sensing and GIS Master Thesis 1998 Indian Institute of Technology, Kanpur, India
Suresh KG Rainfall Effect in Landslide hazard Zonation using Remote Sensing and GIS Master Thesis 1999 Indian Institute of Technology, Kanpur, India
Bohra A Prediction Modeling of Dengue Risk based on Socio-cultural and Environmental factors using GIS: Case of Jalor town, Rajasthan, India PhD Thesis 2001 Asian Institute of Technology, School of Advanced Technologies
Sithiprasasna R Linthieum KJ Use of Geographical Information System to study the Epidemiology of Dengue Haemorrhagic Fever in Thailand Dengue Bulletin 1997 21
| 15943863 | PMC1177981 | CC BY | 2021-01-04 16:39:06 | no | Int J Health Geogr. 2005 Jun 8; 4:13 | utf-8 | Int J Health Geogr | 2,005 | 10.1186/1476-072X-4-13 | oa_comm |
==== Front
Int J Health GeogrInternational Journal of Health Geographics1476-072XBioMed Central London 1476-072X-4-161595339410.1186/1476-072X-4-16ResearchVariations in societal characteristics of spatial disease clusters: examples of colon, lung and breast cancer in Japan Fukuda Yoshiharu [email protected] Masahiro [email protected] Keiko [email protected] Takehito [email protected] Health Promotion/International Health, Division of Public Health, Graduate School of Tokyo Medical and Dental University, Yushima 1-5-45, Bunkyo, Tokyo, 113-8519, Japan2005 14 6 2005 4 16 16 28 4 2005 14 6 2005 Copyright © 2005 Fukuda et al; licensee BioMed Central Ltd.2005Fukuda 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
Spatial analyses and ecological studies are essential for epidemiology and public health. The present study combining these two methods was performed to identify spatial clusters of selected types of cancer in Japan and to determine their societal characteristics focusing on homogeneity among clusters.
Results
Spatial clusters of high mortality rates of male colon and lung cancer and of female breast cancer were identified by the spatial scan statistic using Japanese municipal data (N = 3360) from 1993 to 1998 and also municipalities were divided into four societal clusters based on socioeconomic indicators and population density (urban-rich, suburban, rural-poor, and clutter). Five, seven, and four mortality clusters were identified for lung, colon and breast cancer, respectively. For colon and breast cancer, most municipalities of all except one cluster were included in a single societal cluster (urban-rich). The municipalities associated with mortality clusters for lung cancer belonged to various societal clusters.
Conclusion
Increased mortality rates of colon and breast cancer can be explained by homogenous societal characteristics related to urbanisation, although there were exceptional areas with higher mortality rates. The regional variation in lung cancer mortality rate appeared to be due to heterogeneous factors. These findings and the analysis performed in the present study will contribute to both nationwide and region-specific cancer prevention strategies.
==== Body
Background
Health levels vary substantially between different regions, and it is essential to characterise these regional variations and identify areas with an accumulation of health problems for epidemiologic research and to allow appropriate public health policy decisions [1,2]. Recent advancements in technologies, such as geographic information systems (GIS), have allowed the application of not only disease mapping but also spatial analyses, such as spatial clustering and cluster detection, in epidemiological research [3-6]. In this context, clusters are defined as unusual concentrations of health events in both space and time [1].
Ecological studies examining the relationships between regional health levels and various characteristics represent another essential approach in epidemiology and public health. Although such studies have a number of limitations, especially confounding factors and ecological fallacy, factors that may contribute to regional health variations can be identified and hypotheses can be formulated for further research [7,8]. Several ecological studies have demonstrated relationships between mortality and regional characteristics related to the environment, health-related behaviour, and economic and demographic factors in Japan [9-12]. Recent systematic studies using municipal data regarding all causes and cause-specific mortality along with large numbers of societal indicators showed substantial relationships between a region's mortality rate and societal characteristics [13-15].
In general, the relations between health levels and regional characteristics are examined by correlation and regression analyses [8-14]. These methods can reveal factors correlated with regional variations in a specific health issue across study areas. However, if various factors contribute separately to a health issue for different areas, or where there is an exceptional factor contributing to a health issue in a limited area, such analyses would not be effective in identifying the contributing factors and may overlook exceptional factors.
The present study was performed to determine whether areas with a specific health problem have homogeneous regional characteristics or different patterns of characteristics. We first identified spatial clusters of three common types of cancer (colon, lung and breast) using the spatial scan statistic. The societal characteristics of the clusters were then elucidated, focusing on heterogeneity in the characteristics among clusters, using municipal data across Japan.
Results
The result of principle component analysis for seven socioeconomic indicators, to reduce the number of variables and identify dimensional societal factors, was shown in Table 1. Two principle components were identified and they accounted for 78.1% of the total variance inherent in the data. The meaning of these components was considered higher unemployment and overcrowding for the first component and higher income and educational level for the second component. Component scores of the components were assigned to each municipality as societal indices, designated Index 1 and Index 2, respectively, with a mean of 0.0 and standard deviation of 1.0.
Table 1 Result of principle component analysis.
Rotated component Matrix from the principle component analysis of seven societal indicators.
Societal indicator Component 1 Component 2
Unemployment rate (women) 0.88 -0.10
Unemployment rate (men) 0.87 0.18
Number of rooms per household -0.75 -0.32
Dwelling area per capita -0.77 -0.28
Education level (women) 0.19 0.94
Education level (men) 0.23 0.91
Income per capita 0.07 0.89
The results of cluster analysis for the purpose of categorization of municipalities into societal cluster (SC) are shown in Table 2. SC1 was characterized with high Index 2 and high population density; SC2 with moderate Index 1, Index 2 and population density; SC3 with low Index 1, Index 2 and population density; and SC4 with high Index 1 and population density and low Index 2. The map of these societal clusters is shown in Figure 1. Most of municipalities in the metropolitan areas such as Tokyo, Nagoya, and Osaka, and most of municipalities of seat of prefectural government belong to SC1. In generally, SC2 are located surrounding SC1, and SC3 are located in mountain areas. SC4 are separately distributed, including some municipalities in Okinawa prefecture and the central part of Osaka. The characteristics of societal clusters were interpreted as urban-rich, suburban, rural-poor, and clutter, respectively.
Table 2 Characteristics of societal clusters.
Comparison of societal indices and population density among societal clusters (SC).
Variable Societal cluster
SC1 (N = 507) SC2 (N = 1483) SC3 (N = 1246) SC4 (N = 124)
Index 1: high unemployment and overcrowding 0.89 ± 0.66 -0.08 ± 0.63 -0.54 ± 0.74 2.69 ± 1.23
Index 2: high income and educational level 1.56 ± 1.06 -0.06 ± 0.64 -0.45 ± 0.62 -1.13 ± 0.69
Population density (log) 8.04 ± 0.84 5.75 ± 0.63 3.68 ± 0.83 6.44 ± 1.31
Figure 1 Distribution of societal clusters. (a) Map of Japan. (b) Municipalities are classified into four societal clusters (SCs) according to two societal indices (high unemployment and overcorwing and high income and educational level: see Table 1) and popualtion density. The characeritics of clusters are shown in Table 2.
Municipal standardized mortality ratio (SMR) and the results of spatial scan statistic for male colon and lung cancer and female breast cancer are shown in Figures 2, 3 and 4. As shown in Figure 2 (b), the primary cluster for colon cancer (MC1) included 53 municipalities with a relative risk (RR) of 1.14, and was located in the Tokyo metropolitan area. Four additional clusters were also identified: MC2 was located in the northern part of the main island (Honshu Island) and Hokkaido Island (Hokkaido prefecture), MC3 and MC4 were located in Osaka and Nagoya, which are the second and third largest metropolitan areas after the Tokyo area, respectively, and MC5 that included only one city.
Figure 2 SMR and mortality clusters for colon cancer. Municipal standardized mortality ratio (SMR) (a) and mapping of mortality clusters (MC) (b) with higher mortality rates from male colon cancer. The clusters were identified by the spatial scan statistic. N = number of municipalities belonging to each cluster. RR = relative risk.
Figure 3 SMR and mortality clusters for lung cancer. Municipal standardized mortality ratio (SMR) (a) and mapping of mortality clusters (MC) (b) with higher mortality rates from male lung cancer. The clusters were identified by the spatial scan statistic. N = number of municipalities belonging to each cluster. RR = relative risk.
Figure 4 SMR and mortality clusters for breast cancer. Municipal standardized mortality ratio (SMR) (a) and mapping of mortality clusters (MC) (b) with higher mortality rates from female breast cancer. The clusters were identified by the spatial scan statistic. N = number of municipalities belonging to each cluster. RR = relative risk.
The mortality clusters of male lung cancer are illustrated in Figure 3 (b). The primary cluster (MC1) was located in an area including Osaka, with RR of 1.17. A total of six secondary clusters were also identified. The municipalities of MC2 belonged mainly to Hokkaido prefecture, and those of MC3 belonged to Okinawa prefecture consisting of the southern islands. MC4, MC6, and MC7 included the metropolitan areas of Nagoya, Fukuoka and Tokyo, respectively, while MC5 was located in the mountainous area on Shikoku Island.
As shown in Figure 4 (b), of the four mortality clusters identified for female breast cancer, three were located in metropolitan areas: MC1 in Tokyo, MC2 in Osaka and MC4 in Nagoya. The centre of the remaining cluster, MC3, was located in Hokkaido.
RRs of societal indices, population density, and societal clusters for cancer mortality, which were estimated by the hierarchical Poisson regression, are shown in Table 3. Mortality from colon and breast cancers was significantly and positively related to societal indices and population density. Mortality from lung cancer was significantly and negatively related to Index 2. SC2, SC3, and SC4 showed the lower RR compared to SC1 for colon and breast cancers, while SC4 showed the higher RR for lung cancer.
Table 3 Relative risk of societal indices, population density and societal cluster.
Relative risk was estimated by the hierarchical Poisson regression.
Male colon cancer Male lung cancer Female breast cancer
Variable Crude Adjusted a Crude Adjusted a Crude Adjusted a
Societal Index Index 1 1.104 1.062 1.055 1.039 1.124 1.077
Index 2 1.040 1.005* 0.964 0.952 1.097 1.061
Population density (log) 1.064 1.039 1.010 1.017 1.101 1.039
Societal cluster SC1b 1.000 - 1.000 - 1.000 -
SC2 0.903 - 1.011* - 0.827 -
SC3 0.779 - 0.974* - 0.658 -
SC4 0.993* - 1.179 - 0.969 -
aThe model included societal indices and population density. bThe reference for societal clusters. *not significant (p > = 0.05)
Table 4 shows the relationships between mortality clusters and societal clusters. For MC1, MC3, MC4 and MC5 in colon cancer, the dominant societal cluster was SC1. For MC2 the proportion of SC1 was very low (3.5%), and the dominant cluster was SC3 (66.7%). In lung cancer, there were variations in dominant societal clusters. Similar patterns of societal cluster were observed only for MC1 and MC4 with SC1 as the dominant societal cluster and for MC2 and MC5 with SC3. In breast cancer, most municipalities of MC1, MC2 and MC3 consisted of SC1. For MC3, the dominant societal cluster was SC3 (75.0%).
Table 4 Relationship between mortality clusters and societal clusters.
Figures show the number of municipalities and the percentage in parenthesis by mortality cluster.
Mortality cluster a Societal cluster b
SC1 (N = 507) SC2 (N = 1483) SC3 (N = 1246) SC4 (N = 124)
Male colon cancer
MC1 (N = 53) 53 (100.0)
MC2 (N = 228) 8 (3.5) 59 (25.9) 152 (66.7) 9 (3.9)
MC3 (N = 61) 45 (73.8) 7 (11.5) 9 (14.8)
MC4 (N = 27) 25 (92.6) 2 (7.4)
MC5 (N = 1) 1(100.0)
Male lung cancer
MC1 (N = 159) 81 (50.9) 33 (20.8) 33 (20.8) 12 (4.8)
MC2 (N = 289) 8 (2.8) 46 (15.9) 221 (76.5) 14 (4.8)
MC3 (N = 48) 5 (10.4) 6 (12.5) 37 (77.1)
MC4 (N = 54) 28 (51.9) 24 (44.4) 2 (3.7)
MC5 (N = 12) 4 (33.3) 7 (58.3) 1 (8.3)
MC6 (N = 162) 27 (16.7) 94 (58.0) 16 (9.9) 25 (15.4)
MC7 (N = 8) 8 (100.0)
Female breast cancer
MC1 (N = 51) 51 (100.0)
MC2 (N = 68) 57 (83.8) 2 (2.9) 9 (13.2)
MC3 (N = 168) 8 (4.8) 24 (14.3) 126 (75.0) 10 (6.0)
MC4 (N = 15) 15 (100.0)
aMortality cluster (MC) of municipalities with high mortality was identified by the spatial scan statistic: Figures 2-4. bSocietal cluster (SC) of municipalities was classified by cluster analysis with two societal indices and population density (see Table 2).
Comparisons of mortality clusters before and after adjustment for societal indices, population density, and societal clusters are shown in Table 5. After adjustment in colon cancer, MC1, MC3, MC4, and MC5 were not detected or showed a decrease of RR. In lung cancer, only MC2 was not detected after adjustment. In breast cancer, MC1, MC2 and MC4 were not detected or showed a decrease of RR, while MC3 showed an increase of RR after adjustment.
Table 5 Comparisons of crude and adjusted mortality clustering
Mortality clustera Crude Adjusted b Adjusted c
RR p-value RR p-value RR p-value
Male colon cancer
MC1 1.14 <0.001 * 1.10 <0.001
MC2 1.13 <0.001 1.14 <0.001 1.15 <0.001
MC3 1.10 0.002 * *
MC4 1.18 0.015 * *
MC5 1.51 0.02 1.39 0.009 *
Male lung cancer
MC1 1.17 <0.001 1.11 <0.001 1.16 <0.001
MC2 1.12 <0.001 1.08 <0.001 1.10 <0.001
MC3 1.22 <0.001 * *
MC4 1.13 <0.001 1.12 <0.001 1.08 <0.001
MC5 1.66 <0.001 1.61 0.004 1.65 <0.001
MC6 1.08 <0.001 1.13 <0.001 1.09 0.007
MC7 1.13 0.03 1.08 0.02 1.13 <0.001
Female breast cancer
MC1 1.30 <0.001 * 1.22 <0.001
MC2 1.12 <0.001 * *
MC3 1.13 0.01 1.19 <0.001 1.16 <0.001
MC4 1.22 0.045 * *
aMortality cluster (MC) of municipalities with high mortality was identified by the spatial scan statistic: Figures 2-4. bAdjusted for relative risk of societal indices and population density.
cAdjusted for relative risk of societal cluster. *Cluster was not detected.
Discussion
The results of the present study identified spatial clusters with high mortality rates of colon and lung cancer in men, and of breast cancer in women in Japan. The societal characteristics of the municipalities belonging to these clusters were determined by the relationships between mortality clusters and societal clusters. A single dominant societal cluster was detected for colon and breast cancer, although one mortality cluster was exclusive for each cancer. In contrast, we did not detect a dominant societal cluster for lung cancer.
The detection of a single dominant societal cluster for colon and breast cancer, SC1, suggested that there were homogeneous area characteristics for increased mortality due to these types of cancer. This societal cluster had a high Index 2 representing high income and education level and high population density, which were urban characteristics. These findings were consistence with those of a previous study indicating a positive relationship between mortality from these cancers and socioeconomic index of urbanisation [14]. The relationship between mortality from colon and breast cancer and urban residence is plausible considering risk factors of these cancers, such as westernised dietary habits and low birth rate [16].
One mortality cluster for each of colon and breast cancer (MC2 and MC 3, respectively) showed different characteristics from other mortality clusters. It is possible that factors other than those related to urbanisation contributed to the increased mortality in these areas, and further studies are required to elucidate these unique factors. This observation suggests that the factors contributing to the increased mortality in these exceptional areas may be overlooked in conventional ecological studies.
Unlike colon and breast cancer, no dominant societal cluster was observed for lung cancer. The prevalence of smokers was not included in the set of indicators in the present study because municipal data concerning smoking were not available. It is possible that socioeconomic factors used in this study are surrogates of factors related to colon and breast cancers (e.g., dietary habits), while they might not be surrogates of smoking. The higher mortality in Hokkaido prefecture, as identified by MC2, could be explained by the slightly higher smoking rate reported in this area [16]. However, the prefectural data of smoking did not found that other clusters were not related to areas with higher smoking rate [16]. Previous studies showed small variation in male smoking rate and little relationship between smoking rate and regional socioeconomic conditions, and there was no correlation of male smoking rate with lung cancer mortality [17,18]. Thus, it seems that the difference of smoking rate does not thoroughly explain the regional variation in lung cancer mortality rate, although there is no doubt in contribution of smoking to lung cancer, which is the leading cause of cancer deaths in Japanese men [19].
A number of possible contributors to increased mortality from lung cancer in addition to smoking have been reported [20]. Air pollution is an important factor among these possible contributors, and the observation that several clusters of lung cancer were located in metropolitan areas may be explained by the increase in lung cancer due to air pollution. In Okinawa prefecture, local brand cigarettes with a higher tar yield and the prevalence of human papilloma virus infection were suspected to contribute to the increased mortality from lung cancer in this area [21,22]. If multiple factors: i.e., smoking, air pollution and other specific local factors, contribute to the regional variation in lung cancer mortality, it is reasonable that no uniform characteristics of mortality clusters were detected in the present study.
We found a similarity of mortality clusters among three types of cancers. Three metropolitan areas (Tokyo, Osaka, and Nagoya) were detected as mortality clusters for all cancers. Urban areas recently show a decrease of the relative health level in Japan, and cancer mortality largely attributes to the decreased health level among urban populations [13-15]. Mortality from several types of cancers seems to be concurrently increased by factors related to urban areas such as health risk behaviour and fewer attendances in cancer screening [23,24]. On the other hand, our findings suggested that the northern part including lots of rural-poor municipalities (SC3) appeared to be another area with higher mortality from some types of cancer. The possible causes of higher mortality in this area should be carefully investigated focusing on differences from those in urban areas.
Several methodological issues about mortality and societal clusters and their relationship should be mentioned. There are several alternative methods for mortality clustering such as Openshaw's and Begas and Newell's methods [25,26]. Although the spatial scan statistic has been widely applied, some possible limitations remain, especially about setting of maximum spatial cluster size and detecting and meaning of the secondary clusters [25,26]. The comparison of SMR mappings and mortality clusters might suggest that municipalities with higher mortality were not necessarily accumulated with circular shape, and thus the non-circular shaped spatial scan statistic [27] could detect more accurate mortality clusters. Due to the use of mortality, instead of incidence, the result in the present study could be influenced by not only cancer incidence but also regional differences in health care qualities and others. The incidence data from such as cancer registration could detect more accurate disease clusters in restricted local areas [28-30], but the incidence data of cancer across the country was not available in Japan.
The societal indicators used in this study were restricted. We used indicators that were demonstrated previously to be critically associated with health level [13,15,31], although some indicators of potential cancer risks may not have been included, especially with regard to health-related behaviour. Second, the societal clustering of municipalities was an important issue in the present study. In contrast to other countries [32-34], as there are no established area classifications or societal indices representing regional characteristics in Japan, we formulated societal indices and classified municipalities by the principle component analysis and the cluster analysis. Different combinations of indicators may result in different figures of societal clusters. Especially, the principle component analysis has been the subject of a variety of criticisms including sensitivity of indicator selection and meaning of the components extracted [35,36], although it has been used to reduce socioeconomic indicators and to obtain one or a few composite index [29,32,37]. In addition, unlike mortality data, societal data were not treated by spatial statistics. Spatial methods such as using population potential [38] instead population density and data smoothing for unstability in the municipalities with small population could contribute to more accurate societal classifications of municipality.
The relation between societal characteristics and mortality was mainly examined using societal clusters and mortality clusters. Societal indices and population density showed the significant relation to mortality according to types of cancer, and they might be more sensitive than societal clusters. The statistical comparisons of societal indices and population density among mortality clusters showed significant differences for most pairs of mortality clusters (data not shown). Thus, the analyses with these variables appeared to be too sensitive to examine homogeneity and heterogeneity among mortality clusters. Since the number of societal clusters was arbitrary in the cluster analysis, an increase of the number of societal clusters would show more complicated variations in the societal characteristics among mortality clusters. Significantly, in the present study even when simple societal clustering was applied, both heterogeneity and homogeneity in societal characteristics among mortality clusters were observed. In addition, the comparison of mortality clusters before and after adjustment for societal characteristics quantitatively supported these heterogeneity and homogeneity.
Conclusion
The combination of spatial analysis and investigation of the relationships between mortality and societal factors revealed areas in Japan with higher mortality rates and their societal characteristics. The spatial clusters of colon and breast cancer showed homogeneous societal characteristics, with the exception of one cluster. However, the societal characteristics of clusters of lung cancer varied. The homogeneous characteristics of areas with higher mortality rates require strategies across the country or common between higher mortality areas, while exclusive clusters, such as those seen for colon and breast cancer, and variations in societal characteristics for lung cancer imply the need of strategies specific for selected areas with higher mortality.
Methods
Study unit and period
Local public entities in Japan are divided into two categories: the first consists of municipalities (i.e., cities, towns and villages), while the second consists of prefectures. All districts in the country belong to one of the municipalities and fall within the boundaries of one of the prefectures. Tokyo prefecture (Tokyo Metropolis) includes 23 special wards ("ku") in addition to cities, towns and villages. Twelve large cities (cities designated by ordinance), such as Osaka and Nagoya, consist of wards ("ku"). In 1995, there were a total of 3372 municipalities (23 Tokyo special ward cities, 127 wards of 12 cities designated by ordinance, 651 cities, 1994 towns and 577 villages) nested within 47 prefectures [39].
The study was performed from 1993 to 1998 during which time several municipalities were annexed or divided, and therefore the aggregated data from these municipalities could not be used. Thus, the final number of municipalities analyzed in the present study was 3360.
Mortality calculation
In this study, we examined the mortality rates of three high priority cancers: male lung and colon cancer and female breast cancer. Lung and colon cancer were the first and fourth leading causes of cancer death, respectively, in men, and breast cancer was the fourth one in women in the Japanese population in 1995 [20]. The rates of colon and breast cancer have both increased steadily over the last several decades in Japan. Classification was based on the 9th and 10th versions of the International Classification of Diseases (ICD-9 in 1993–94 and ICD-10 in 1995 to 1998): colon cancer, ICD-9 153–154 and ICD-10 C18-C21; lung cancer, ICD-9 162 and ICD-10 C33-C34; and breast cancer, ICD-9 174 and ICD-10 C50 [40,41].
As our focus was on premature mortality, which is more closely related to regional societal characteristics, we examined deaths in the population aged under 75 years old [13]. The numbers of cause-specific deaths by municipality from 1993 to 1998 were compiled. The data regarding deaths in 1995 were excluded to avoid the influence of the Hanshin-Awaji earthquake [14]. Total number of deaths during 5 years was 57,109 for colon cancer, 101,515 for lung cancer, and 32,290 breast cancer. The nationwide age-and cause-specific mortality rates and census municipal age-specific population in 1995 were used as data sources [42]. The aggregated data using macrofiles of the vital statistics were drawn from a database of previous studies [14].
Municipal SMR was calculated and disease mapping was drawn. For calculation of SMR, the hierarchical Poisson regression analysis [13,14,43] was applied since this analysis could correct the unstability in mortality due to heterogeneity of population size: there was marked variation in the population size among municipalities, ranging from a few hundred to a few hundred thousand, and municipalities with a small population showed statistical fluctuation in mortality. The secondary medical care zone (SMCZ), which is defined by prefectural governments for medical care planning according to the Medical Service Law, was used as a higher level. There were 344 SMCZs across Japan in 1995, each of which consisted of neighbouring municipalities and covered a population of 300,000 on average. Bayesian standardized mortality ratio of municipalities was estimated using the iterative generalized least squares (IGLS) and the Markov chain Monte Carlo method [44]. Relative risks (RRs) of societal indices, population density, and societal clusters for cancer morality were estimated using the hierarchical Poisson regression with IGLS. In addition to crude RRs, societal and population density were included in the model to estimate adjusted RRs. For societal clusters, SC1 was used as the reference category. The details of hierarchical Poisson regression are described in previous studies [13,14,44]
Mortality clusters: spatial scan statistic
The spatial scan statistic was used to detect and evaluate the statistical significance of spatial clusters. The details of the spatial scan statistic were reported previously [4-6,36] and SaTScan ver. 4.0.3 was used for the analysis. The numbers of deaths in each municipality were modelled as Poisson distributions. Under the null hypothesis, the expected number of deaths calculated using age-specific national mortality rates and the age-specific municipal population from the 1995 census [42] was proportional to the indirectly age-adjusted population at risk. An infinite number of circles were superimposed on the map, using the municipal centroid as the centre. The municipal centroid (latitude and longitude) was computed with the map of Japan (geographic coordinate system, GRS 1980; ) using ArcGIS 8.3 (ESRI Japan, Tokyo). The radii of the circles were set to vary continuously from zero to a maximum including at most 10% of the total population at risk, to obtain a certain number of potential clusters. The data for an entire circle contained different sets of neighbouring municipalities, and each circle represented a potential mortality cluster. For each circle, the likelihood was calculated for observing the number of deaths occurring within that circle, and the circle with the maximum likelihood was taken as the primary cluster. The distribution of maximum likelihood under the null hypothesis was evaluated using the Monte Carlo hypothesis testing set with 999 simulations. In addition to the primary cluster, the spatial scan statistic identified the secondary clusters, and ordered them according to the likelihood ratio test statistics. In the present study, secondary clusters were identified using no geographical overlap procedure and those with p-values of less than 0.05 were significant. Mortality clusters were mapped using ArcGIS 8.3 (ESRI, Japan).
Societal clusters
Based on the findings of previous studies [13,15,31], seven socioeconomic indicators were chosen as potential factors related to mortality (Table 1). These indicators were obtained and calculated using the System of Social and Demographic Statistics consisting of governmental statistics including mainly census data [42]. Unemployment rate reflected the percentage of unemployed persons aged 15–65 years in the total workforce. Educational level reflected the age-adjusted educational level, using the percentage of those who had graduated from college or a higher level among the population aged 20 and over, and was standardised by nationwide sex-and age-specific populations as for standardisation of age-adjusted mortality rate. Income per capita was calculated by aggregating the annual taxable income per household by municipality, and dividing it by the total municipal population.
To reduce the number of variables and identify dimensional societal factors, the principal component analysis with correlation matrix analysis and varimax rotation was performed. The principle components for which the correlation matrix eigenvalues were more than 1.0 were selected as significant dimensions. The component score for the extracted component was assigned to municipalities as a composite societal index: consequently two indices were obtained as shown in Table 1. Then, municipalities were classified into four societal clusters using the K-means cluster analysis with two societal indices and population density (log-transformed). The principle component analysis and the cluster analysis were performed using SPSS 11.0 (SPSS Inc., Chicago, IL, USA).
Relationships between mortality clusters and societal clusters
The relationships between societal characteristics and mortality clusters identified by the spatial scan statistic were examined by the cross-tabular analysis of mortality clusters and societal clusters.
Furthermore, cluster detections with the spatial scan statistic were performed adjusting for societal indices and population density or societal clusters to determine whether mortality clusters would change before and after adjustment for these variables. The risks of municipalities were calculated using RRs from the hierarchical Poisson regression and used as the adjustment file in the spatial scan statistic [29,44].
Authors' contributions
YF designed the study, analyzed the data, and drafted the article. MU designed the study and interpreted the results. KN helped to interpret the results and edited the draft. TT supervised the data analysis and writing article.
Acknowledgements
This research was supported by Grant-in-Aid for Scientific Researches by the Japan Society for the Promotion of Science (Grant No. 14570326 and 16590497).
==== Refs
Cromley EK McLafferty SL GIS and public health 2002 New York. Guilford Press
Ellioitt P Wakefield JC Best NG Briggs DJ Spatial epidemiology 2000 New York: Oxford University Press
Rosenberg MS Sokal RR Oden NL DiGiovann D Spatial autocorrelation of cancer in Western Europe Eur J Epidemiol 1999 15 15 22 10098991 10.1023/A:1007559728848
Kulldorff M Feuer EJ Miller BA Freedman LS Breast cancer clusters in the northeast United States: a geographic analysis Am J Epidemiol 1997 146 161 170 9230778
Green C Hoppa RD Young TK Blanchard JF Geographic analysis of diabetes prevalence in an urban area Soc Sci Med 2003 57 551 560 12791496 10.1016/S0277-9536(02)00380-5
Jemal A Kulldorff M Devesa SS Hayes RB Fraumeni F A geographic analysis of prostate cancer mortality in the United States, 1970–89 Int J Cancer 2002 101 168 174 12209994 10.1002/ijc.10594
Diez Roux AV Schwartz S Susser E Detels R, McEwen J, Beaglehole R, Tanaka H Ecological variables, ecological studies, and multilevel studies in public health research Oxford textbook of public health 2002 New York: Oxford University Press 493 507
Morgentern H Rothman KJ, Greenland S Ecological study Modern epidemiology 1998 Philadelphia: Lippincott Williams & Wilkins 459 480
Uehara M Takahashi M Hoshuyama T Pan G Feng Y Geographical correlation between ambient UVB level and mortality risk of leukemia in Japan Environ Res 2003 92 78 84 12854686 10.1016/S0013-9351(03)00034-3
Tominaga S Kuroishi T An ecological study on diet/nutrition and cancer in Japan Int J Cancer 1997 Suppl 10 2 6 9209011 10.1002/(SICI)1097-0215(1997)10+<2::AID-IJC2>3.0.CO;2-C
Aihara H Iki M An ecological study of the relations between the recent high suicide rates and economic and demographic factors in Japan J Epidemiol 2003 13 A56 61
Nagata C Ecological study of the association between soy product intake and mortality from cancer and heart disease in Japan Int J Epidemiol 2000 29 832 836 11034965 10.1093/ije/29.5.832
Fukuda Y Nakamura K Takano T Municipal socioeconomic status and mortality in Japan: sex and age difference, and trends of 1973–1998 Soc Sci Med 2004 59 2435 2445 15474199 10.1016/j.socscimed.2004.04.012
Fukuda Y Nakamura K Takano T Cause-specific mortality differences across socioeconomic position of municipalities in Japan, 1973 to 1998: increased importance of injury and suicide to inequality for ages under 75 Int J Epidemiol 2005 34 100 109 15561754 10.1093/ije/dyh283
Fukuda Y Nakamura K Takano T Wide range of socioeconomic factors associated with mortality among cities in Japan Health Promot Int 2004 19 177 187 15128709 10.1093/heapro/dah205
Kanda A Ojima T Miura Y Prefectural variation in alcohol consumption, smoking, physical exercise, and obesity and the trend Koseino Shihyo 2002 49 7 15
Fukuda Y Nakamura K Takano T Socioeconomic pattern of smoking in Japan: income inequality and gender and age differences Ann Epidemiol 2005 15 365 372 15840550 10.1016/j.annepidem.2004.09.003
Asahi S Watanabe M Tajimi M Relationships between prefectural smoking and alcohol consumption rates and cause-specific mortality Koseino Shihyo 2003 50 1 7
Health and Welfare Statitics Association Kokumin ensei no doko 2002 Tokyo: Health and Welfare Statistics Association
Colditz GA Atwood KA Emmons K Harvard report on cancer prevention volume 4: Harvard Cancer Risk Index Cancer Causes Control 2000 11 477 488 10880030 10.1023/A:1008984432272
Wakai K Ohno Y Genka K Smoking habits, local brand cigarettes and lung cancer risk in Okinawa, Japan J Epidemiol 1999 7 99 105 9255031
Nakazato I Hirayasu T Kamada Y Tsuhako K Iwamasa T Carcinoma of the lung in Okinawa, Japan: with special reference to squamous cell carcinoma and squamous metaplasia Pathology International 1997 47 659 672 9361099
Fukuda Y Nakamura K Takano T Accumulation of health risk behaviors is associated with lower socioeconomic status and women's urban residence: a multilevel analysis in Japan BMC Public Health 2005 5 53 15921512 10.1186/1471-2458-5-53
Fukuda Y Nakamura K Takano T Reduced likelihood of cancer screening among women in urban areas and with low socioeconomic status: a multilevel analysis in Japan Public Health
Wakefield JC Kelsall JE Morris SE Elliot P, Wakefield JC, Best NG, Briggs DJ Clustering, cluster detection, and spatial variation in risk Spatial Epidemiology 2000 NY: Oxford University Press 128 152
Waller LA Gotway CA Spatial clustering of health events: regional count data Applied Spatial Statistics for Public Health Data 2004 New Jersey: John Wiley & Sons 200 271
Tango T Takahashi K A flexibly shaped spatial scan statistic for detecting clusters Int J Health Geogr 2005 4 11 15904524 10.1186/1476-072X-4-11
Buntinx F Gey H Lousbergh D Geographical differences in cancer incidence in the Belgian province of Limburg Eur J Cancer 2003 39 2058 2072 12957461 10.1016/S0959-8049(02)00734-7
Sheehan TJ DeChelo LM Kulldorff M Gregorio DI Gershman S Mroszczyk M The geographic distribution of breast cancer incidence in Massachusetts 1988 to adjusted for covariates Int J Health Geogr 2004 3 17 15291960 10.1186/1476-072X-3-17
Kulldorff M Nagarwalla N Spatial disease clusters: detection and inference Stat Med 1995 14 799 810 7644860
Takano T Nakamura K An analysis of health levels and various indicators of urban environments for Healthy Cities projects J Epidemiol Community Health 2001 55 263 270 11238582 10.1136/jech.55.4.263
Krieger N Chen JT Waterman PD Soobader MJ Subramanian SV Carson R Geocoding and monitoring of US socioeconomic inequalities in mortality and cancer incidence: does the choice of area-based measure and geographic level matter?; The Public Health Disparities Geocoding Project Am J Epidemiol 2002 156 471 482 12196317 10.1093/aje/kwf068
House JM Lepkowski JM Williams DR Excess mortality among urban residents: how much, for whom, and why? Am J Public Health 2000 90 1898 1904 11111263
Morris R Carstairs V Which deprivation? A comparison of selected deprivation indexes J Public Health Med 1991 13 318 326 1764290
Armitage P Berrry G Matthews JNS Mutivariate methods Statitical Methods in Medical Research 2002 Oxford: Blackwell Sceince 455 484
Kachigan SK Factor analysis Multvariate Statitical Analysis 1991 New York: Radium 236 260
Folwell K Single measures of deprivation J Epidemiol Community Health 1995 49 S51 S56 8594135
Middleton N Gunnell D Frankel S Whitley E Dorling D Urban-rural differences in suicide trends in young adults: England and Wales, 1981–1998 Soc Sci Med 2003 57 1183 1194 12899903 10.1016/S0277-9536(02)00496-3
Society for self-government of municipalities Municipal handbook 2000 Tokyo: Daiichi Hoki
Ministry of Health and Welfare Vital statistics 1994 1997 Tokyo: Health and Welfare Statistics Association
Ministry of Health, Labour, and Welfare Vital statistics 1995 1997 Tokyo: Health and Welfare Statistics Association
Statistics Bureau, Management and Coordination Agency System of socioeconomic and demographic statistics 2000 Tokyo: Statistical Information Institute for Consulting and Analysis
Leyland AH McLeod A Mortality in England and Wales, 1979–1992 2000 Glasgow: MRC Social & Public Health Sciences Unit
Rasbash J Browne W Goldstein H A user's guide to MLwiN 2002 London: Centre for Multilevel Modeling, Institute of Education, University of London
| 15953394 | PMC1177982 | CC BY | 2021-01-04 16:39:06 | no | Int J Health Geogr. 2005 Jun 14; 4:16 | utf-8 | Int J Health Geogr | 2,005 | 10.1186/1476-072X-4-16 | oa_comm |
==== Front
J Autoimmune DisJournal of Autoimmune Diseases1740-2557BioMed Central London 1740-2557-2-51591670410.1186/1740-2557-2-5ResearchEvaluation of autoantibodies to common and neuronal cell antigens in Chronic Fatigue Syndrome Vernon Suzanne D [email protected] William C [email protected] Division of Viral and Rickettsial Diseases, National Center for Infectious Diseases, Centers for Disease Control and Prevention, Atlanta, Georgia 30333, USA2005 25 5 2005 2 5 5 19 4 2005 25 5 2005 Copyright © 2005 Vernon and Reeves; licensee BioMed Central Ltd.2005Vernon and Reeves; 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.
People with chronic fatigue syndrome (CFS) suffer from multiple symptoms including fatigue, impaired memory and concentration, unrefreshing sleep and musculoskeletal pain. The exact causes of CFS are not known, but the symptom complex resembles that of several diseases that affect the immune system and autoantibodies may provide clues to the various etiologies of CFS. We used ELISA, immunoblot and commercially available assays to test serum from subjects enrolled in a physician-based surveillance study conducted in Atlanta, Georgia and a population-based study in Wichita, Kansas for a number of common autoantibodies and antibodies to neuron specific antigens. Subsets of those with CFS had higher rates of antibodies to microtubule-associated protein 2 (MAP2) (p = 0.03) and ssDNA (p = 0.04). There was no evidence of higher rates for several common nuclear and cellular antigens in people with CFS. Autoantibodies to specific host cell antigens may be a useful approach for identifying subsets of people with CFS, identify biomarkers, and provide clues to CFS etiologies.
chronic fatigue syndromeCFSantibodiesautoantibodies
==== Body
Background
Chronic fatigue syndrome (CFS) is defined as persistent or relapsing fatigue that has occurred for at least 6 months, is not alleviated by rest, and causes substantial reduction in activities. The fatigue cannot be explained by medical or psychiatric conditions and must be accompanied by at least 4 of 8 specified symptoms (unusual post exertional fatigue, impaired memory or concentration, unrefreshing sleep, headaches, muscle pain, joint pain, sore throat, and tender cervical nodes) [1]. There is considerable discrepancy in results between studies from different institutions; so as yet, there are no characteristic signs or laboratory markers of CFS and its pathophysiology has not been elucidated [2].
This lack of diagnostic signs or laboratory markers notwithstanding, many manifestations of CFS resemble those of musculoskeletal and infectious diseases [3]. In large part, the illnesses caused by these diseases reflect immune system activation and there is evidence for immune system dysfunction in some cases of CFS. In particular, antinuclear antibodies (ANA) and other common autoantibodies have been evaluated in people with CFS: unfortunately, with variable results. For example, one study found that 52% of tertiary care- CFS referral-patients had antibodies to nuclear envelope antigens [4] while another study found the same ANA antibody rates in both CFS and controls [5]. Recently, investigators reported that antibodies to the human muscarinic cholinergic receptor 1 may provide a biologic explanation for the cognitive impairment observed in people with CFS [6]. This lack of consensus between studies may in large-part reflect recruitment bias associated with studies of persons enrolled from tertiary referral clinics combined with imprecise evaluation of the illness and inadequate or inappropriate control populations.
We had the opportunity to measure the associations of common autoantibodies and autoantibodies to neuronal cell antigens and CFS in two case control studies; one of primary care patients with CFS who were identified by a physician surveillance network; the other a study of people with CFS identified from the community. The physician surveillance study was conducted 1988 through 1993 in Atlanta, Georgia [7] and the community study 1997 through 2000 and identified subjects with CFS from the general population of Wichita, Kansas [8].
Both studies rigorously classified people as CFS and controls in both studies were enrolled to represent the general population and matched to cases by sex, race, and age [9,10]. The hypothesis of the present study is that the appearance of cell-specific autoimmune antibodies may define subsets of CFS and give clues to the etiology and pathogenesis and may help to explain the neurocognitive symptoms experienced by CFS patients. Secondarily, we wished to evaluate the extent to which patients with CFS who were receiving primary medical care treatment for CFS were similar to people with CFS in the community.
Methods
Study Subjects and samples
Both studies adhered to human experimentation guidelines of the U.S. Department of Health and Human Services and the Helsinki Declaration. The Centers for Disease Control and Prevention (CDC) Institutional Review Board approved study protocols. All participants were volunteers who gave informed consent.
Physician surveillance study
Between 1988 and 1993, the CDC conducted a physician surveillance survey for CFS in primary care patients from Reno, Nevada, Wichita, Kansas, Grand Rapids, Michigan, and Atlanta, Georgia [7]. Patients were classified as CFS according to the 1988 case definition [11]. In 1992, we conducted a case control study of CFS patients and controls in Atlanta by recruiting patients from physician surveillance and sex, race, age matched non fatigued controls identified in the general Atlanta population [9,10]. The case control study classified patients as CFS according to the study collected information concerning several risk factors and blood to measure associations between CFS and laboratory markers. The present study used remaining archived serum samples from 22 CFS patients and 34 age and sex matched controls. All CFS patients met criteria of the current CFS research case definition [1]
Population study participants
Between 1997 and 2000, CDC conducted surveillance of CFS in the general population of Wichita, Kansas [8]. Briefly, the study involved random digit dial surveys to identify people with CFS-like illness and clinically evaluated and classified them according to criteria of the 1994 CFS research case definition [1]. Only 16% of those identified with CFS had been diagnosed or treated for CFS by a physician [12]. The present study used archived serum samples from 37 subjects with CFS and a 57 non-fatigued control subjects
Blood samples
Both the physician surveillance and population study collected blood in BD Vacutainer Serum tubes. The samples were shipped by overnight courier to CDC where they were dispensed into 0.5 ml aliquots and stored at -80°C until testing.
Reagents and Assays
Commercially available kits were used for antibodies to ubiquitous nuclear and cellular autoantigens including dsDNA, ssDNA, Sm, U1-RNP, SS-A/Ro, SS-B/La, Scl-70, and Centromere. Immunoassays were purchased from Helix Diagnostics (West Sacramento, CA) and reagents for western blots were purchased from Diagnostic Products Corporation (Los Angeles, CA). Purified Histone H3, and Histone H4 were purchased from Sigma (St. Louis, MO). and used in ELISA assays that were developed at Scripps. Conventional immunofluorescent antinuclear antibodies and rheumatoid factor tests were performed as described previously [13]. Preparations of microtubule-associated protein 2 (MAP2) and neurofilament triplet (NFT) proteins were purchased from Sigma (St. Louis, MO). The commercially available ELISA assays were performed according to the manufacturers instructions. The ELISA and western blot assays for the neuronal antigens were developed at Scripps and were performed as previously described [14].
Statistical Analysis
Because the subjects are derived from studies that are distinct in design and geographic location, each study was analyzed separately. The distribution of autoantibodies between CFS and non-fatigued controls was compared by Fisher exact probability test. To derive an estimate of confidence, stratified groups were compared by the non-parametric chi square test. To determine associations, subjects were stratified by sex, age, and CFS for all MAP2, NFT and ssDNA. The association of autoantibodes in CFS subjects was compared by grouping by sex, age, age at illness onset, and duration of illness. CFS subjects were stratified by sex, age (<40 years, 40–40 years, >50 years), onset type (gradual versus sudden) and duration of illness (<5 years, >5 years) to determine whether an association with autoantibodies existed.
Results
Although women predominated in both study groups other demographic and clinical characteristics differed and reflected basic differences between patients with CFS who obtain medical care and those in the general population (most of whom have not seen a physician) (Table 1). Of note, CFS cases from physician surveillance were somewhat younger than those identified in the population (mean 39 and 46 years, respectively) and controls were similarly different: those recruited in the physician study had been ill about half as long as those in the community (69 and 128 months, respectively) and were more likely to report sudden onset CFS (36.4%) than those in the general population with CFS (13.5%).
Table 1 Characteristics of subjects evaluated for autoantibodies.
CFS Non-Fatigued
Atlanta Case Control
Subjects (n = 56) 22 34
Female (n = 52) 19 33
Male (n = 4) 3 1
Age Group (yrs)
18–29 (n = 9) 3 6
30–39 (n = 18) 7 11
40–49 (n = 23) 12 11
50–59 (n = 6) 0 6
Mean Age 39 years 38 years
Mean Age Onset 35 years
Onset type
Sudden 8
Gradual 14
Mean Illness Duration 69 months
Number of Subjects Ill
< 5 years 10
>5 years 12
Wichita Population
Subjects (n = 94) 37 57
Female (n = 64) 31 33
Male (n = 30) 6 24
Age Group (yrs)
18–29 (n = 14) 1 13
30–39 (n = 18) 8 10
40–49 (n = 26) 13 13
50–69 (n = 36) 15 21
Mean Age 46 years 42 years
Mean Age Onset 36 years
Onset type
Sudden 5
Gradual 32
Mean Illness Duration 128 months
Number of Subjects Ill
< 5 years 14
>5 years 23
A few CFS subjects in the physician surveillance study aged 18–29 years had antibodies to ssDNA when compared to the same age non-fatigued control group. The mean value for the 3 CFS subjects was 2-fold greater then in the 6 non-fatigued controls (p = 0.038). Among CFS subjects, the 10 who reported being ill for ≤ 5 years had lower levels of autoantibodies to MAP2 (median value of 18, range 12 – 20) compared to the 12 CFS subjects who have been ill for >5 years (median value of 8, range 6 to 10) (p = 0.025). There were no other significant findings in the physician surveillance CFS subjects when stratified by sex or type of illness onset.
In the population-based study, there was a significant difference in the prevalence of autoantibodies to MAP2 between the 30 male subjects (20/30, 67% positive) and the 64 female subjects (19/64, 30% positive) (p = 0.0006). Among the non-fatigued control group, 9 of 33 women (27%) and 19 of 24 (79%) men were positive for antibodies to MAP2 (p = 0.0004). One male CFS subject (16%) was positive for MAP2 antibodies compared to 79% (19/24) male non-fatigued controls (p = 0.04). Among CFS subjects that were ≤ 40 years of age, there was a trend for lower MAP2 antibody levels for those that were ill for ≤ 5 years compared to those ill for >5 years (p = 0.056).
Discussion
CFS is a complex, debilitating illness, which is characterized by at least 6 months of severe persistent or relapsing fatigue and a group of characteristic but nonspecific symptoms. Despite more than a two decades of extensive research, no diagnostic tests exist, and effective control and prevention remain elusive because the cause and pathophysiology of CFS remain unknown. CFS is clinically similar to several rheumatic autoimmune disorders that can be diagnosed and characterized by autoantibody profiles. For this reason, we conducted an exhaustive evaluation of 11 ubiquitous nuclear and cellular autoantigens in addition to two neuronal specific antigens.
The serum samples tested in this study were collected from a physician surveillance study conducted in Atlanta [15] and a population-based community study in Wichita [8]. The physician surveillance study was conducted over 7 years and used approximately 70% of all primary care physicians in Atlanta. All patients were carefully evaluated for unexplained unwellness and fatigue. The community study was a random digit dial survey of 90,000 people (25% of the Wichita, Kansas population). All CFS cases were medically and psychiatrically evaluated and rigorously classified as CFS, other unexplained unwellness, or medically/psychiatrically explained unwellness. The serum evaluated in this study includes carefully evaluated CFS subjects, matched controls and non-fatigued controls from the community. Therefore, the results from this study should be applicable to similarly designed studies.
Very few studies have evaluated the presence of autoantibodies in people with CFS. Those that have tested for the same autoantibodies report discordant results. Konstantinov et al, [4] found high rates of antinuclear antibodies (ANA) in CFS patients while Skowera et al, [5] found no difference in the rate of ANA between CFS patients and controls. One explanation for these discrepancies could be a technical one where laboratories used different reagents and methods resulting in discordant results. Another explanation could be that the CFS subjects were evaluated differently. Rigor in evaluating CFS patients and applying the case definition [1] is pivotal and undoubtedly accounts for much of the variation. The results presented also show little evidence for autoantibodies to ubiquitous nuclear and cellular autoantibodies.
The findings of this study hint that evaluation of certain autoantibodies may give clues to etiology and ongoing pathology in subsets of CFS subjects. A few of the physician surveillance CFS cases from the youngest age category had autoantibodies to ssDNA. Autoantibodies to ssDNA have been associated with both viral and bacterial infection [16,17]. Interestingly, CFS subjects who describe a sudden onset to their illness often report flu-like illness. The fact that antibodies to ssDNA were detected only in this age group may reflect an immune response to infection commonly affecting this age group, such as infectious mononucleosis from Epstein Barr Virus infection.
There was a higher prevalence of autoantibodies to MAP2 in the non-fatigued men in the community study compared to the non-fatigued women. The significance of this finding is not known but highlights the importance of carefully stratifying and controlling for factors that could affect interpretation of results. There did seem to be a slight association of MAP2 autoantibodies with duration of CFS illness. Among CFS subjects in both study populations, those who had been sick longer had higher rates of autoantibodies than those that report shorter duration of illness. MAP2 is a neuron specific cytoskeleton protein. Autoantibodies to MAP2 have been demonstrated in patients with neuropsychiatric systemic lupus erythematosus [14]. While no lesions or loss of central nervous system function is reported in people with CFS, loss of memory, concentration and cognitive impairment are common complaints. Future studies will attempt to associate assessment of these parameters with the presence of MAP2 autoantibodies.
Conclusion
There was no evidence of higher rates of the common autoantibodies in people with CFS. However, certain subsets of CFS subjects that had higher rates of antibodies to microtubule-associated protein 2 (MAP2) and ssDNA. Autoantibodies to specific host cell antigens may be a useful approach to stratify CFS subjects and provide clues to CFS etiologies.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
SDV was instrumental in the design of the experimental approach, analysis, presentation, discussion of these data and manuscript preparation. WCR contributed to the design of the experimental approach, and to the manuscript preparation. Both authors read and approved the final manuscript.
Acknowledgements
We would like to thank Dr. Eng Tan of The Scripps Research Institute for conducting the autoantibody assays. We would also like to thank Dr. Rosane Nisenbaum for assistance with the statistical analysis.
==== Refs
Fukuda K Straus SE Hickie I Sharpe MC Dobbins JG Komaroff A The chronic fatigue syndrome: a comprehensive approach to its definition and study Ann Int Med 1994 121 953 959 7978722
Afari N Buchwald D Chronic fatigue syndrome: a review Am J Psychiatry 2003 60 221 236 12562565 10.1176/appi.ajp.160.2.221
Nisenbaum R Reyes M Unger ER Reeves WC Factor analysis of symptoms among subjects with unexplained chronic fatigue: What can we learn about chronic fatigue syndrome? J Psychosom Res 2004 56 171 178 15016574 10.1016/S0022-3999(03)00039-4
Konstantinov K von Mikecz A Buchwald D Jones J Gerace L Tan EM Autoantibodies to nuclear envelope antigens in chronic fatigue syndrome J Clin Invest 1996 98 1888 1896 8878441
Skowera A Stewart E Davis ET Cleare AJ Unwin C Hull L Ismail K Hossain G Wessely SC Peakman M Antinuclear autoantibodies (ANA) in Gulf War-related illness and chronic fatigue syndrome (CFS) patients Clin Exp Immunol 2002 129 354 358 12165094 10.1046/j.1365-2249.2002.01912.x
Tanaka S Kuratsune H Hidaka Y Hakariya Y Tatsumi KI Takano T Kanakura Y Amino N Autoantibodies against muscarinic cholinergic receptor in chronic fatigue syndrome Int J Mol Med 2003 12 225 230 12851722
Reyes M Gary HE Dobbins JG Randall B Steele L Fukuda K Holmes GP Connell DG Mawle AC Schmid DS Stewart JA Schonberger LB Gunn WJ Reeves WC Surveillance for chronic fatigue syndrome – four US cities, September 1988 through August 1993 MMWR CDC Surveillance Summaries 1997 46 2 13
Reyes M Nisenbaum R Hoaglin DC Unger ER Emmons C Randall B Stewart JA Abbey S Jones JF Gantz N Minden S Reeves WC Prevalence and incidence of chronic fatigue syndrome in Wichita, Kansas Arch Intern Med 2003 163 1530 1536 12860574 10.1001/archinte.163.13.1530
Reyes M Dobbins JG Mawle AC Steele L Gary HE Malani H Schmid S Fukuda K Stewart J Nisenbaum R Reeves WC Risk factors for CFS: a case control study J CFS 1996 2 17 33
Mawle AC Nisenbaum R Dobbins JG Gary HE Stewart JA Reyes M Steele L Schmid DS Reeves WC Immune responses associated with chronic fatigue syndrome: a case-control study J Infect Dis 1997 175 136 141 8985207
Holmes GP Kaplan JE Gantz NM Komaroff AL Schonberger LB Straus SE Jones JF Dubois RE Cunningham-Rundles C Pahwa S Chronic fatigue syndrome: a working case definition Ann Intern Med 1988 108 387 389 2829679
Solomon L Reeves WC Factors influencing the diagnosis of chronic fatigue syndrome by primary health care providers Arch Intern Med 2004 164 2241 2245 15534161 10.1001/archinte.164.20.2241
von Mikecz A Konstantinov K Buchwald DS Gerace L Tan EM High frequency of autoantibodies to insoluble cellular antigens in patients with chronic fatigue syndrome Arthritis Rheum 1997 40 295 305 9041942
Williams RC JrSugiura K Tan EM Antibodies to microtubule-associated protein 2 in patients with neuropsychiatric systemic lupus erythematosus Arthritis Rheum 2004 50 1239 1247 15077307 10.1002/art.20156
Mawle AC Nisenbaum R Dobbins JG Gary HE JrStewart JA Reyes M Steele L Schmid DS Reeves WC Seroepidemiology of chronic fatigue syndrome: a case-control study Clin Infect Dis 1995 21 1386 9 8749620
Pisetsky DS Specificity and immunochemical properties of antibodies to bacterial DNA Methods 1997 11 55 61 8990089 10.1006/meth.1996.0387
Wu YY Hsu TC Chen TY Liu TC Liu GY Lee YJ Tsay GJ Proteinase 3 and dihydrolipoamide dehydrogenase (E3) are major autoantigens in hepatitis C virus (HCV) infection Clin Exp Immunol 2002 28 347 352 11985526 10.1046/j.1365-2249.2002.01827.x
| 15916704 | PMC1177983 | CC BY | 2021-01-04 16:38:03 | no | J Autoimmune Dis. 2005 May 25; 2:5 | utf-8 | J Autoimmune Dis | 2,005 | 10.1186/1740-2557-2-5 | oa_comm |
==== Front
J Inflamm (Lond)Journal of Inflammation (London, England)1476-9255BioMed Central London 1476-9255-2-51592706210.1186/1476-9255-2-5ResearchC-reactive protein does not opsonize early apoptotic human neutrophils, but binds only membrane-permeable late apoptotic cells and has no effect on their phagocytosis by macrophages Hart Simon P [email protected] Karen M [email protected] Shonna M [email protected] Ian [email protected] MRC Centre for Inflammation Research, University of Edinburgh Medical School, Teviot Place, Edinburgh EH8 9AG, UK2005 31 5 2005 2 5 5 25 10 2004 31 5 2005 Copyright © 2005 Hart et al; licensee BioMed Central Ltd.2005Hart 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
It has been reported that C-reactive protein (CRP) binds both leukocyte FcγRIIA (CD32) and the plasma membrane of apoptotic cells. Since FcγRIIA becomes functionally enabled during neutrophil apoptosis, we sought to determine whether CRP bound to apoptotic neutrophils via FcγRIIA.
Methods
We prepared directly labelled CRP and demonstrated that it was essentially free of IgG. We looked for evidence of CRP binding to intact, membrane impermeable apoptotic human neutrophils and to FcγRIIA-transfected Jurkat cells. We examined the functional consequences of incubation with CRP upon phagocytosis of apoptotic cells by human monocyte-derived macrophages.
Results
We could not detect binding of purified soluble CRP to classical early apoptotic human neutrophils or to FcγRIIA-transfected Jurkat cells. In contrast, membrane-permeable late apoptotic neutrophils exhibited strong CRP binding, which comprised both Ca2+-dependent and heparin-inhibitable Ca2+-independent components. However, there was no effect of CRP binding upon phagocytosis of late apoptotic neutrophils by macrophages.
Conclusion
Potential apoptotic cell opsonins such as CRP may bind only to intracellular structures in cells with leaky membranes that have progressed to a late stage of apoptosis.
==== Body
Background
In acute inflammation huge numbers of neutrophils are recruited to sites of tissue injury where they die by undergoing apoptosis [1]. Macrophage clearance of apoptotic neutrophils has been studied extensively under serum-free conditions in vitro, but the presence of opsonins in the inflammatory milieu means it is unlikely that "naked" apoptotic cells would be encountered by macrophages in vivo [2]. We have recently reported that IgG-containing immune complexes bound preferentially to functionally enabled FcγRIIA (CD32) on apoptotic neutrophils [3,4]. It has been proposed that FcγRIIA is also a receptor for the pentraxin C-reactive protein (CRP) [5,6], serum concentrations of which may increase more than 1000-fold during acute inflammation [7,8]. Independently, it was reported that soluble CRP opsonised apoptotic Jurkat cells in vitro [9-11]. In the present study we sought to determine whether binding of CRP to apoptotic neutrophils was mediated by FcγRIIA. To circumvent the inherent difficulties in interpreting the results of antibody binding to Fc receptor-bearing cells, we used directly fluorescein-conjugated purified CRP that was essentially free of contaminating IgG.
Methods
Conjugation of CRP with FITC
1 mg of CRP purified from human plasma (Sigma; Poole, UK) was dissolved in 1 ml deionised water and dialysed against 100 mM sodium bicarbonate pH 8.25. Fluorescein isothiocyanate (FITC; Sigma) was dissolved at 1.5 mg/ml in DMSO and added dropwise to a total volume of 45 μl per ml of protein solution. The mixture was incubated for 2 hours at room temperature in the dark. Unconjugated FITC was removed by exhaustive dialysis.
Assessment of CRP purity
Native and FITC-conjugated CRP were examined under denaturing conditions on a 9% acrylamide gel stained with Coomassie blue. Contamination with human IgG was assessed by comparing Western blots of CRP with known amounts of human IgG (Sigma). Blots were probed with rabbit F(ab')2 anti-human IgG followed by peroxidase-conjugated goat anti-rabbit IgG (DakoCytomation; Ely, UK) and developed by enhanced chemiluminescence (Amersham).
CRP phospholipid binding assay
One micrometer diameter polystyrene microspheres (Polysciences; Warrington, PA) were coated with 1 mg/ml phosphorylcholine-conjugated BSA (Biosearch Technologies; Novato, CA) or BSA (Sigma) in PBS for 1 h at room temperature. Beads were washed and incubated with FITC-conjugated CRP in the presence of 2 mM Ca2+ or 5 mM EDTA. FITC-CRP binding was measured by gating on single beads and analysing fluorescence in the FL1 channel (530 nm) following excitation with an argon laser at 488 nm in a BD FACSCalibur flow cytometer (BD Biosciences, Cowley, Oxford, UK).
CRP binding assay
Human neutrophils were isolated from peripheral blood of healthy volunteers by dextran sedimentation and discontinuous Percoll gradient centrifugation [12]. Genomic DNA extraction and determination of the polymorphism at position 519 in exon 4 of the FcγRIIA gene was performed as previously described [3]. G or A at position 519 leads to either an arginine (R) or histidine (H) amino acid at position 131 in the second Ig-like domain of the FcγRIIA protein. Experiments were performed using cells from donors with each of the genotypes R/R, R/H, and H/H. Neutrophils were aged in culture for 20 h at 37°C/5% CO2 in Iscove's medium (Invitrogen, Paisley, UK) containing 10% autologous serum or FCS. Jurkat cells transfected with human FcγRIIA or control vector were kindly provided by Dr. Eric Brown, University of California, San Francisco, USA [13]. Surface phenotyping using indirect immunofluorescence and flow cytometry confirmed expression of FcγRIIA by the transfected cells, but no expression of FcγRI or FcγRIII was detected. Control Jurkat cells transfected with empty vector did not express any Fc receptors. Cells were washed twice in PBS prior to use. Cell binding assays were performed in 140 mM NaCl pH 7.4, 20 mM HEPES, and either 2 mM CaCl2 or 5 mM EDTA. FITC-CRP was incubated with cells for 30 minutes on ice, then washed twice in binding buffer and incubated with phycoerythrin-conjugated Annexin V (Caltag; Towchester, UK) or 5 μg/ml propidium iodide. In some experiments cells were incubated with 1 mg/ml unfractionated heparin (Sigma) and washed twice prior to incubation with FITC-CRP. Fluorescence was analysed on an Coulter Epics XL flow cytometer (Beckman Coulter, High Wycombe, UK) and/or a BD FACSCalibur flow cytometer.
Immunofluorescence microscopy
Aged neutrophils were labelled with 50 μg/ml FITC-CRP, fixed in 3% paraformaldehyde, permeabilised with 0.1% Triton X-100, counterstained with TO-PRO-3 (Molecular Probes, Leiden, NL), and cytocentrifuged onto glass slides. Visualisation was performed with a Leica TCSNT confocal system (Leica Microsystems, GmBH, Mannheim, Germany).
Flow cytometric cell sorting
Cultured human neutrophils were labelled with 25 μg/ml FITC-CRP and sorted according to FL1 signal intensity using a BD FACSVantage fluorescence activated cell sorter (FACS). Sorted cell populations were checked for purity by flow cytometry, and cell morphology was examined on May-Giemsa-stained cytocentrifuge preparations.
Macrophage phagocytosis of late apoptotic neutrophils
Human neutrophils were labelled with CFDA (CellTracker™Green; Molecular Probes) and incubated at 37°C for 72 h in Iscove's medium containing 10% autologous serum to yield a cell population that contained >70% late apoptotic neutrophils. Aged neutrophils were washed, and incubated with 100 μg/ml CRP or Iscove's medium alone for 30 minutes. Monolayers of 5–8d old human monocyte-derived macrophages in 48 well plates were incubated with 2 × 106 aged neutrophils in Iscove's medium in the absence of serum for 60 minutes at 37°C [14]. The supernatant was aspirated and the macrophages were detached by brief incubation in 0.05% trypsin-0.02% EDTA (Invitrogen) and vigorous pipetting. The percentage of macrophages that had ingested one or more apoptotic neutrophils was determined by flow cytometric analysis as previously described [15].
Statistical analysis
Results are presented as mean ± SEM of at least three independent experiments using cells from different donors. Results were compared using either a paired t-test or repeated measures ANOVA and Tukey-Kramer multiple comparisons test as appropriate, using GraphPad InStat version 3 (GraphPad Software, San Diego, CA).
Results
Purity and functional integrity of FITC-conjugated CRP
Binding studies were performed with human plasma-derived CRP that had been directly conjugated to FITC. Native CRP and FITC-CRP migrated as single bands when examined by SDS-PAGE (Figure 1a). Western blotting demonstrated <0.1% contamination of our CRP preparation with human IgG. (Figure 1b). FITC-CRP was shown to be functionally active by demonstrating Ca2+-dependent binding to phosphorylcholine-coated beads (Figure 1c). Similarly, FITC-CRP bound strongly to cells that had been rendered necrotic by freeze-thawing (Figure 1d).
Figure 1 Purity and functional integrity of FITC-CRP. (a) Native and FITC-conjugated CRP migrated as single bands of approximately 23 kD in SDS-PAGE. (b) Western blot of CRP and known amounts of human IgG on a 9% acrylamide gel under non-reducing conditions. The blot was probed with anti-human IgG-HRP. CRP contained <0.1% IgG (w/w). MW markers in kDa are illustrated. (c) FITC-CRP bound to phosphorylcholine-coated beads in the presence of 2 mM Ca2+ (dark shaded histogram). Binding in 5 mM EDTA (light shaded histogram) and BSA-coated beads (control; open histogram) is also shown. (d) FITC-CRP bound to freeze-thawed necrotic neutrophils (shaded histogram). Binding of FITC-BSA is shown as a control (open histogram).
CRP does not bind to early apoptotic neutrophils
Early apoptotic neutrophils were identified within a population of cultured human neutrophils by their characteristic flow cytometric laser scatter properties, labelling with annexin V, and exclusion of propidium iodide. We found no detectable binding of CRP at concentrations up to and including 100 μg/ml, in presence or absence of 2 mM Ca2+ (Figure 2). The highest concentration of CRP that we used is comparable to serum concentrations recorded during an acute inflammatory response. The lack of CRP binding was observed regardless of FcγRIIA genotype (data not shown).
Figure 2 Lack of CRP binding to early apoptotic human neutrophils. FITC-CRP binding to aged human neutrophils was assessed by dual color flow cytometry using annexin V-PE to identify apoptotic cells. In comparison with buffer alone (a), FITC-CRP did not bind to apoptotic neutrophils at concentrations up to 100 μg/ml (b,c).
Lack of binding of CRP to FcγRIIA
The lack of binding of CRP to non-apoptotic or apoptotic neutrophils suggested that soluble (non-aggregated) CRP was unable to bind to FcγRIIA with significant affinity. For confirmation, Jurkat cells transfected with human FcγRIIA were incubated with FITC-conjugated CRP. There was no evidence of preferential binding of CRP to FcγRIIA-transfected cells compared with those transfected with control vector (Figure 3).
Figure 3 CRP does not bind to FcγRIIA. Jurkat cells transfected with empty vector (-; open symbols) or human FcγRIIA (+; closed symbols) were incubated with FITC-conjugated CRP (circles) or FITC-BSA control (triangles). There was no diffrence in binding between transfected and non-transfected cells.
CRP binds strongly to a subpopulation of cultured human neutrophils
A subpopulation of strongly positive cells was apparent when FITC-CRP binding to ungated cultured neutrophils was analysed. These cells were also positive for annexin V and propidium iodide staining (Figure 4a). Similar results were seen with human peripheral blood lymphocytes that had been induced to undergo apoptosis by exposure to ultraviolet radiation (data not shown). Immunofluorescence microscopy of cultured human neutrophils revealed strong CRP binding to cells that had very little or no residual nuclear staining (Figure 4b). In keeping with the flow cytometric data, there was no detectable binding to classical early apoptotic neutrophils or to non-apoptotic neutrophils. To confirm the morphology of the CRP-positive cells, cultured human neutrophils were incubated with FITC-CRP and sorted in a fluorescence-activated cell sorter (FACS). Light microscopic examination of both cell populations confirmed that the CRPhigh cells were "ghosts" with little or no evidence of nuclear staining, whereas the CRPlow cells comprised a mixture of non-apoptotic and early apoptotic neutrophils (Figure 4c). The CRPhigh neutrophils appear to have progressed to a late stage of apoptosis and undergone "nuclear evanescence" [16,17].
Figure 4 CRP binds to a subpopulation of apoptotic neutrophils. (a) Three color flow cytometry demonstrated that FITC-CRP (50 μg/ml) bound strongly to a subpopulation of apoptotic neutrophils that also stained with propidium iodide (PI). (b) Immunofluorescence microscopy of aged human neutrophils revealed strong CRP binding (green) to a late apoptotic neutrophil (arrow)(left panel). Nuclei have been stained with TO-PRO-3 (blue). The late apoptotic cell has almost no residual nuclear staining (right panel). There was no detectable binding to an early apoptotic neutrophil (arrowhead) or to non-apoptotic neutrophils. (c) Light microscopy of CRPhigh (left panel) and CRPlow (right panel) cells sorted by FACS from a population of aged neutrophils illustrates the ghost-like morphology of CRP-binding apoptotic neutrophils. The CRPlow cells comprise a mixture of non-apoptotic and early apoptotic neutrophils.
Mechanism of CRP binding
CRP binding to phospholipids is dependent on the presence of calcium ions [18], whereas binding to polycationic sites is Ca2+-independent [19]. To determine whether Ca2+ was required for CRP binding to late apoptotic neutrophils we compared FITC-CRP binding in the presence of 2 mM Ca2+ and 5 mM EDTA. In the presence of EDTA, CRP binding was reduced by approximately 50% compared with total CRP binding seen in the presence of Ca2+ (Figure 5). Because it has been reported that heparin binds to necrotic Jurkat cells [20], we sought to identify whether heparin and CRP bound to similar intracellullar sites. Pre-incubation with unfractionated heparin inhibited FITC-CRP binding to late apoptotic human neutrophils in the presence of cations by approximately 50%, and almost abolished residual binding when CRP was incubated in the presence of EDTA (Figure 5), suggesting that CRP bound to a combination of Ca2+-dependent and heparin-inhibitable Ca2+-independent sites.
Figure 5 Additive inhibition of CRP binding by EDTA and heparin. FITC-CRP was incubated with cultured human neutrophils in the presence of 2 mM Ca2+ or 5 mM EDTA, with or without pre-incubation with 1 mg/ml heparin. Binding to late apoptotic neutrophils was assessed by gating on the propidium iodide-positve cell population. The inhibitory effects of EDTA and heparin on CRP binding were additive.
Macrophage phagocytosis of late apoptotic neutrophils
We sought to determine the effect of prior binding of CRP on phagocytosis of late apoptotic neutrophils. Our attempts to sort large numbers of cultured neutrophils by FACS into early and late apoptotic populations were unsuccessful, because during the time required for sorting many early apoptotic neutrophils progressed to late apoptosis, rendering the cells unsuitable for subsequent phagocytosis assays. Thereafter, we prepared neutrophils containing predominantly late apoptotic cells (>70%) by aging neutrophils for 72 h in vitro. Prior incubation of these late apoptotic neutrophils with 100 μg/ml CRP, which resulted in high levels of CRP binding, had no demonstrable effect on their phagocytosis by macrophages (Figure 6).
Figure 6 Macrophage phagocytosis of late apoptotic neutrophils. CRP (100 μg/ml) was allowed to bind to human neutrophils that had been aged for 72 h (>70% late apoptotic) prior to assessment of phagocytosis by human macrophages. Prior incubation with CRP had no effect on the precentage of macrophages that phagocytosed one or more late apoptotic neutrophils.
Discussion
It has been reported that the pentraxins CRP and serum amyloid P are ligands for leukocyte Fcγ receptors [5]. FcγRIIA becomes functionally enabled on early apoptotic human neutrophils [3,4], but we have demonstrated that soluble CRP does not bind to classical early apoptotic human neutrophils in vitro, and we have been unable to demonstrate CRP binding to human FcγRIIA on transfected Jurkat cells. In the present study we have been careful to avoid pitfalls associated with indirect detection of ligand binding to neutrophils by using a preparation of directly labelled CRP that we have shown was essentially free of contaminating IgG, but which was structurally and functionally intact. The failure of CRP to bind FcγRIIA is consistent with the results of Hundt and colleagues who failed to find specific receptors for CRP on human leukocytes [6].
There have been several reports of Ca2+-dependent opsonisation of apoptotic cells by pentraxins [9,11,21]. The lack of CRP binding to classical early apoptotic human neutrophils, which exhibit all the biochemical and surface changes associated with apoptosis yet remain intact and membrane impermeable, raises questions about whether putative opsonins are really able to bind with high affinity to apoptotic cells. In contrast, we demonstrated very strong CRP binding to a subpopulation of aged neutrophils which displayed the characteristics of late apoptotic neutrophils previously reported by Hebert [16] and Ren [17]. It has been recognised that CRP binds to necrotic cells since Kushner and Kaplan demonstrated CRP deposition in necrotic skeletal muscle fibres following typhoid vaccination in vivo [22]. It is not always recognised that induction of apoptosis in many cell types in vitro leads to a significant proportion of membrane-permeable late apoptotic cells [23]. The presence of leaky late apoptotic cell ghosts in a cell population may be overlooked because the lack of nuclear material means that they stain very faintly with May-Giemsa, and in the past late apoptotic cells may have been gated out as "debris" when analysed by flow cytometry. Furthermore, these cells pellet poorly in cytocentrifuge preparations, and this combined with their "invisibility" with May-Giemsa stains means that their prevalence has been underestimated. This clearly has implications for studies of apoptotic cell opsonisation, and much of the published data may reflect binding to intracellular moieties. CRP is not unique in binding to the interior of leaky apoptotic cells, and we have reported a similar phenomenon with the unrelated serum protein thrombospondin [24]. Complement proteins, collectins, and heparin may also bind preferentially to late apoptotic cells [20,25,26]. The precise structures responsible for CRP binding have not been elucidated, but our data suggest that there are both Ca2+-dependent and Ca2+-independent binding sites. Binding to cell membrane phospholipids may account for the Ca2+-dependent component [18], whereas binding to polycations may be responsible for the heparin-inhibitable Ca2+-independent component [19]. The relative paucity of nuclear chromatin in the late apoptotic cells and the absence of nuclear co-localisation seen with fluorescence microscopy means that chromatin binding is unlikely to be responsible [21].
The presence of late apoptotic cells also has implications for studies of the phagocytosis of apoptotic cells, since these cells may be recognised differently from classical early apoptotic cells [17,23]. It is not known whether "opsonins" bound to intracellular components would be accessible for recognition by phagocyte receptors. In the present study we have shown that despite very strong CRP binding to late apoptotic neutrophils, there was no detectable effect on their clearance by macrophages. Ren and colleagues demonstrated that the efficiency of phagocytosis of early- and late apoptotic neutrophils was similar [17], so we think it is unlikely that an effect of CRP on uptake of late apoptotic cells has been masked by baseline uptake of early apoptotic neutrophils in the population of aged cells that we used.
Conclusion
By using a directly labelled pure preparation of CRP we have found no evidence that CRP opsonises classical early apoptotic neutrophils in vitro. Like other proteins and sugars it binds intracellularly to membrane-permeable cells, but has no significant influence their subsequent phagocytosis by macrophages. A precise role for CRP remains to be elucidated.
Abbreviations
CFDA, 5-chloromethylfluorescein diacetate; CRP, C-reactive protein; FACS, fluorescence-activated cell sorter
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
SH designed the study, carried out the binding and phagocytosis experiments, analysed the data, and drafted the manuscript. KA carried out the binding and phagocytosis experiments. SM performed the cell sorting. ID participated in the design and execution of the study and drafted the manuscript. All authors have read and approved the final manuscript.
Acknowledgements
We are grateful to Dr. Eric Brown for providing the FcγRIIA-transfected Jurkat cells, and to Linda Wilson for operating the confocal microscope. This work was funded by a Medical Research Council Clinician Scientist Fellowship (G108/460).
==== Refs
Savill JS Fadok VA Henson PM Haslett C Phagocyte recognition of cells undergoing apoptosis Immunol Today 1993 14 131 136 8385467 10.1016/0167-5699(93)90215-7
Hart SP Smith JR Dransfield I Phagocytosis of opsonized apoptotic cells: roles for 'old-fashioned' receptors for antibody and complement Clin Exp Immunol 2004 135 181 185 14738443 10.1111/j.1365-2249.2003.02330.x
Hart SP Jackson C Kremmel LM McNeill MS Jersmann H Alexander KM Ross JA Dransfield I Specific binding of an antigen-antibody complex to apoptotic human neutrophils Am J Pathol 2003 162 1011 1018 12598333
Hart SP Alexander KM Dransfield I Immune complexes bind preferentially to FcgammaRIIA (CD32) on apoptotic neutrophils, leading to augmented phagocytosis by macrophages and release of proinflammatory cytokines J Immunol 2004 172 1882 1887 14734773
Bharadwaj D Stein MP Volzer M Mold C Du Clos TW The major receptor for C-reactive protein on leukocytes is fcgamma receptor II J Exp Med 1999 190 585 590 10449529 10.1084/jem.190.4.585
Hundt M Zielinska-Skowronek M Schmidt RE Lack of specific receptors for C-reactive protein on white blood cells Eur J Immunol 2001 31 3475 3483 11745367 10.1002/1521-4141(200112)31:12<3475::AID-IMMU3475>3.0.CO;2-1
Kushner I Gewurz H Benson MD C-reactive protein and the acute-phase response J Lab Clin Med 1981 97 739 749 7229508
Pepys MB Hirschfield GM C-reactive protein: a critical update J Clin Invest 2003 111 1805 1812 12813013 10.1172/JCI200318921
Gershov D Kim S Brot N Elkon KB C-Reactive protein binds to apoptotic cells, protects the cells from assembly of the terminal complement components, and sustains an antiinflammatory innate immune response: implications for systemic autoimmunity J Exp Med 2000 192 1353 1364 11067883 10.1084/jem.192.9.1353
Chang MK Binder CJ Torzewski M Witztum JL C-reactive protein binds to both oxidized LDL and apoptotic cells through recognition of a common ligand: Phosphorylcholine of oxidized phospholipids Proc Natl Acad Sci U S A 2002 99 13043 13048 12244213 10.1073/pnas.192399699
Mold C Baca R Du Clos TW Serum amyloid P component and C-reactive protein opsonize apoptotic cells for phagocytosis through Fcgamma receptors J Autoimmun 2002 19 147 154 12419285 10.1006/jaut.2002.0615
Dransfield I Buckle AM Savill JS McDowall A Haslett C Hogg N Neutrophil apoptosis is associated with a reduction in CD16 (FcgRIII) expression J Immunol 1994 153 1254 1263 8027553
Green JM Schreiber AD Brown EJ Role for a glycan phosphoinositol anchor in Fc gamma receptor synergy J Cell Biol 1997 139 1209 1217 9382867 10.1083/jcb.139.5.1209
Hart SP Rossi AG Dransfield I You can lead a macrophage to an inflammatory site but can you make it eat? Biochem Soc Trans 1998 26 650 652 10047799
Jersmann HP Ross KA Vivers S Brown SB Haslett C Dransfield I Phagocytosis of apoptotic cells by human macrophages: Analysis by multiparameter flow cytometry Cytometry 2003 51A 7 15 12500300 10.1002/cyto.a.10005
Hebert MJ Takano T Holthofer H Brady HR Sequential morphologic events during apoptosis of human neutrophils. Modulation by lipoxygenase-derived eicosanoids J Immunol 1996 157 3105 3115 8816421
Ren Y Stuart L Lindberg FP Rosenkranz AR Chen Y Mayadas TN Savill J Nonphlogistic clearance of late apoptotic neutrophils by macrophages: efficient phagocytosis independent of beta 2 integrins J Immunol 2001 166 4743 4750 11254736
Volanakis JE Wirtz KW Interaction of C-reactive protein with artificial phosphatidylcholine bilayers Nature 1979 281 155 157 471064
DiCamelli R Potempa LA Siegel J Suyehira L Petras K Gewurz H Binding reactivity of C-reactive protein for polycations J Immunol 1980 125 1933 1938 6776184
Gebska MA Titley I Paterson HF Morilla RM Davies DC Gruszka-Westwood AM Kakkar VV Eccles S Scully MF High-affinity binding sites for heparin generated on leukocytes during apoptosis arise from nuclear structures segregated during cell death Blood 2002 99 2221 2227 11877301 10.1182/blood.V99.6.2221
Familian A Zwart B Huisman HG Rensink I Roem D Hordijk PL Aarden LA Hack CE Chromatin-independent binding of serum amyloid P component to apoptotic cells J Immunol 2001 167 647 654 11441067
Kushner I KAPLAN MH Studies of acute phase protein. I. An immunohistochemical method for the localization of Cx-reactive protein in rabbits. Association with necrosis in local inflammatory lesions J Exp Med 1961 114 961 974 14460901 10.1084/jem.114.6.961
Devitt A Pierce S Oldreive C Shingler WH Gregory CD CD14-dependent clearance of apoptotic cells by human macrophages: the role of phosphatidylserine Cell Death Differ 2003 10 371 382 12700637 10.1038/sj.cdd.4401168
Hart SP Ross JA Ross K Haslett C Dransfield I Molecular characterization of the surface of apoptotic neutrophils: implications for functional downregulation and recognition by phagocytes Cell Death Differ 2000 7 493 503 10800083 10.1038/sj.cdd.4400680
Gaipl US Kuenkele S Voll RE Beyer TD Kolowos W Heyder P Kalden JR Herrmann M Complement binding is an early feature of necrotic and a rather late event during apoptotic cell death Cell Death Differ 2001 8 327 334 11550084 10.1038/sj.cdd.4400826
Nauta AJ Raaschou-Jensen N Roos A Daha MR Madsen HO Borrias-Essers MC Ryder LP Koch C Garred P Mannose-binding lectin engagement with late apoptotic and necrotic cells Eur J Immunol 2003 33 2853 2863 14515269 10.1002/eji.200323888
| 15927062 | PMC1177984 | CC BY | 2021-01-04 16:36:23 | no | J Inflamm (Lond). 2005 May 31; 2:5 | utf-8 | J Inflamm (Lond) | 2,005 | 10.1186/1476-9255-2-5 | oa_comm |
==== Front
J NeuroinflammationJournal of Neuroinflammation1742-2094BioMed Central London 1742-2094-2-141593509810.1186/1742-2094-2-14ResearchMicroglial inflammation in the parkinsonian substantia nigra: relationship to alpha-synuclein deposition Croisier Emilie [email protected] Linda B [email protected] David T [email protected] Ronald KB [email protected] Manuel B [email protected] Department of Neuropathology, Division of Neuroscience and Mental Health, Imperial College London, and Hammersmith Hospitals Trust, London, UK2 Department of Cellular and Molecular Neuroscience, Division of Neuroscience and Mental Health, Imperial College London, London, UK2005 3 6 2005 2 14 14 26 4 2005 3 6 2005 Copyright © 2005 Croisier et al; licensee BioMed Central Ltd.2005Croisier 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 role of both microglial activation and alpha-synuclein deposition in Parkinson's disease remain unclear. We have tested the hypothesis that if microglia play a primary role in Parkinson's disease pathogenesis, the microglial "activated" phenotype should be associated with histopathological and/or clinical features of the disease.
Methods
We have examined microglial MHC class II expression, a widely used marker of microglial activation, the occurrence of CD68-positive phagocytes and alpha-synuclein immunoreactivity in post-mortem human substantia nigra affected by idiopathic Parkinson's disease (PD). Using semi-quantitative severity ratings, we have examined the relationship between microglial activation, alpha-synuclein deposition, classical neuropathological criteria for PD, subtype of the disease and clinical course.
Results
While we did not observe an association between microglial MHC class II expression and clinical parameters, we did find a correlation between disease duration and the macrophage marker CD68 which is expressed by phagocytic microglia. In addition, we observed a significant correlation between the degree of MHC class II expression and alpha-synuclein deposition in the substantia nigra in PD.
Conclusion
While microglia appeared to respond to alpha-synuclein deposition, MHC class II antigen expression by microglia in the substantia nigra cannot be used as an indicator of clinical PD severity or disease progression. In addition, a contributory or even causative role for microglia in the neuronal loss associated with PD as suggested by some authors seems unlikely. Our data further suggest that an assessment of microglial activation in the aged brain on the basis of immunohistochemistry for MHC class II antigens alone should be done with caution.
==== Body
Introduction
Parkinson's Disease (PD) is a common neurodegenerative disorder with the cardinal clinical features of tremor, rigidity, bradykinesia and loss of postural reflexes. Neuropathologically, the disease is characterized by a marked loss of dopaminergic neurons in the substantia nigra pars compacta (SN) and the presence of alpha-synuclein (aSN)-positive Lewy bodies (LBs) in neurons of this and other brain areas also affected by nerve cell death. An international consensus definition of Lewy body diseases on the basis of molecular as well as morphological criteria, which takes into account aSN status of the brain, has been published recently [1].
The discovery of aSN mutations and gene amplification in some familial forms of PD [2-6]] and the identification of this protein as a major component of Lewy bodies (LBs) in common sporadic PD [7], has spurred interest in the role of aSN in the pathophysiology of PD and other synucleinopathies. However, no direct causal relationship has yet been established between aSN aggregation and the selective neuronal cell death characteristic of PD. LBs are also found incidentally in aged brain in the absence of other pathological features and without a clinical history of parkinsonism or dementia [8]. Attempts have been made to link the clinical progression of PD to the presence of aSN inclusions and an anatomical staging model has been proposed [9], but the latter has been questioned by subsequent studies in which clinical data were also taken into account [10].
Apart from well established morphological criteria, activated microglia can be defined in tissue sections on the basis of the expression of several immune function-related proteins, notably complement receptors and MHC class II antigens (MHCII). Phagocytic activity and cytotoxic properties are usually considered end stages of microglial activation, at which point they are phenotypically indistinguishable from blood-borne macrophages. Activated microglia are associated with a large range of neurological insults from trauma and infection to autoimmune conditions, and their presence represents a common finding also in neurodegenerative disorders [11]. However there is little knowledge about the molecular processes that mediate microglial activation and exactly which biological consequences may result from their enhanced state of "immune alertness" within affected CNS tissue. A transcriptome signature of interferon-gamma activated microglia has been provided recently [12]. Microglial phenotypic changes have also been observed in normal aged individuals [13]. Thus, "microglial senescence" confounds the problem of a definition of microglial activation in disease states, and in neurodegenerative diseases in particular which are often age-related, as no specific causative stimulus has been identified in the process.
While microglia clearly show changes in their phenotypic profile in neurodegeneration, it is by no means clear whether they are actively involved in the progression of PD. Microglia-derived macrophages can be found in the PD SN, and neuromelanin pigment taken up from degenerated dopaminergic nerve cells is characteristically observed in SN phagocytes. In animal models of nigrostriatal degeneration using 6-hydroxydopamine and 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP), inhibition or attenuation of the microglial immune response increases neuronal survival. However, those results have so far not been replicated in clinico-pathological studies, and the simple chemical lesions currently employed in animal studies by all likelihood do not fully reflect the chronic neurodegenerative disease process in humans [14].
In the present study, we independently evaluate the severity of alpha-synuclein deposition and microglial activation identified by immunohistochemical staining in the SN in a large cohort of clinically and pathologically confirmed PD cases. We have studied the microglial response in PD on two levels, by observing MHCII -immunoreactive cells (putatively activated microglia but possibly only senescent cells) and CD68-immunopositive macrophages (corresponding to either phagocytic microglia or cells derived from invading blood-borne macrophages).
Materials and methods
Parkinson's disease cases
37 PD nigrae were evaluated immunohistochemically. 20 cases were provided by the UK Parkinson's Disease Society Tissue Bank at Imperial College London (PDSTB). Additional tissue sections from 17 other cases came from a previous study originally performed at the Institute of Neuropathology, University of Munich, Germany. These Parkinson's cases had been previously diagnosed, neuropathologically screened for confounding pathology, and examined in a study of apoptosis and microglial activation [15]. Archival sections were immunolabelled for alpha-synuclein (see below) and used as a control group to ensure that variation within our PDSTB cohort was within an established range.
Clinical and neuropathological assessment of cases
For the PDSTB cohort, clinical reports were evaluated in detail by an experienced neurologist with a special interest in Parkinson's disease (RKBP). Neuropathological assessment was based on slides provided by the PDSTB for alpha-synuclein, tau and beta-amyloid immunohistochemistry of superior frontal gyrus, the hippocampal region and midbrain as minimum data sets, and screening of the cases for confounding pathology was based on hematoxylin and eosin examination of a standard series of 18 tissue blocks following a standardised dissection procedure [16]. Nine cases showed varying degrees of concurrent Alzheimer's disease (AD)-type pathology (tau-immunopositive tangles and/or beta-amyloid-immunopositive plaques) of isocortical and/or entorhinal type ranging from grades 1–3 . Three cases were excluded based on a final neuropathological diagnosis of AD, progressive supranuclear palsy (PSP), and young-onset familial PD, respectively, leaving a cohort of 17 cases in each the Munich and PDSTB groups (Table 1).
Table 1 PDSTB cases examined.
CASE SEX DIAGNOSIS AAO AAD DD MHCII AVE aSN AVE CD68 AVE
1 m PD, H-T 63 72 9 2 1 1.5
2 f PDD H-T 67 81 14 2 1 1
3 m PD, A-R 57 71 14 2.75 3 2.25
4 f PD, H-T 67 85 18 2.5 2.25 0.5
5 m PD, H-T 57 75 18 2 1.5 2
6 m PD, A-R 78 83 5 1.75 1 2.75
7 f PD, H-T 55 73 18 1.75 2 2.25
8 m PD, H-T 49 77 28 2.5 2.5 0.5
9 f PD, H-T 65 75 10 2 2 2.75
10 m PD, A-R 75 82 7 1 1.75 2
11 m PDD H-T 72 81 9 2 2 1.25
12 m PD, A-R 69 75 6 2 2.5 1.75
13 m PD, A-R 65 83 18 2.5 1.5 1.5
14 m PD, H-T 70 77 7 2 1 1.5
15 m PD, A-R 86 89 3 2.75 2.5 2
16 f PD, H-T 72 83 11 2 1.5 1.25
17 m PD, H-T 66 76 10 2.5 2.5 3
Abbreviations: PD, Parkinson's disease; PDD, Parkinson's disease with dementia; H-T, hemi-tremulous; A-R, akinetic-rigid; AAO, age at disease onset; AAD, age at death; DD, disease duration; AVE, semi-quantitative severity rating, averaged across two observers.
Immunohistochemical evaluation of protein levels
Immunohistochemical reactions were performed using the avidin-biotin complex (ABC)/peroxidase method with mouse monoclonal antibodies anti-human HLA-DP, DQ, DR (clone CR3/43, Dako, dilution 1/100) and anti-alpha-synuclein (Becton-Dickinson, dilution 1/300). For the PDSTB group, additional immunohistochemistry was carried out with anti-CD68 (clone PGM1, Dako, dilution 1/200). Sections were dewaxed in xylene, rehydrated, and endogenous peroxidase activity was blocked by 30 min exposure to 1% hydrogen peroxide in methanol. Antigen unmasking consisted of boiling in 0.01 M EDTA (20 min. at 350 W in microwave) and 100% formic acid treatment (3 min.) prior to incubation with anti-HLA-DP, DQ, DR and anti-alpha-synuclein, respectively. No antigen unmasking was used with anti-CD68. Slides were then incubated in primary antibody diluted in phosphate-buffered saline (PBS) overnight at 4°C. The following day, after washing in PBS, they were incubated in horse-anti-mouse secondary antibody (Vector, dilution 1/200) and finally in ABC complex (Vector, dilution 1/200) each for 1 hour at room temperature. Immunoreactivity was visualised with 3,3'-diaminobenzidine (Vector kit).
After immunohistochemical staining, sections were given semi-quantitative severity ratings for aSN, MHCII, and CD68 immunoreactivity by two investigators (EC and MBG) blinded to case number. The SN was defined as the area extending laterally from the exit of the third nerve, superior to the cerebral peduncle and inferior to the medial lemniscus, ideally at the height of the red nucleus with the presence of melanised neurons or their remnants indicating the main region of interest. The severity ratings were determined across the entire region of SN, based on the density of immunopositive structures, with 0 (none), 1 (mild), 2 (moderate) and 3 (high). For aSN, both intra- and extra-cellular inclusions were considered provided they fell within the immediate area of the substantia nigra. This was particularly relevant in areas of severe neuronal loss, often encountered more laterally, where significant alpha-synuclein pathology could still be observed. The morphological variation in aSN deposition was not assessed, simply the frequency of events. All clearly identifiable aSN-immunoreactive structures, including LBs, neurites, fibrils, and smaller, punctate formations, were considered. For microglial response, severity was judged primarily by immunoreactivity, however morphology was taken into account in that perivascular immunoreactivity was excluded. The thickening of microglial processes increased the apparent density of microglial staining, such that cases undergoing a more intense microglial response were clearly differentiated on the basis of immunoreactivity alone. Morphological features of activated microglia were always noted, however there were no cases for which the severity rating would have changed substantially on the basis of morphological features, ie. cases with low MHCII immunoreactivity but most microglia adopting an amoeboid morphology or cases with high MHCII immunoreactivity but most microglia appearing ramified.
Regression analysis revealed the two sets of ratings from independent observers were highly correlated (p < 0.0001). Ratings were then averaged to generate a severity score for each immunohistochemical stain.
Results
All of the cases examined showed the severe dopaminergic neuronal loss and extra-cellular (free-lying) neuromelanin typical of advanced PD. All were also positive for alpha-synuclein, MHCII, and CD68 (Figure 1). Semi-quantitative ratings for the PDSTB cohort are shown in Table 1. MHCII immunoreactivity was confined to microglial or macrophage-like cells. Most of the CD68-positive cells were of a brain macrophage phenotype, i.e. cells with an enlarged immunoreactive cytoplasm containing lysosomal structures and/or neuromelanin degradation products, shortened and less ramified, stout (compared with typical microglia) cell processes but rarely of the appearance of the full-blown macrophages commonly found in brain infarcts or multiple sclerosis lesions. aSN inclusions were observed in neurons, white matter, and occasionally glial cells and their fine processes. MHCII immunolabelling and the presence of macrophages showed significant variation between cases. Semi-quantitative ratings revealed that despite this inconsistency across the group, aSN deposition and MHCII immunoreactivity were found to correlate within individual cases (p < 0.001) (Figure 2). A General Linear Model test (SPSS) revealed no difference between PDSTB and University of Munich cases (p = 0.01), and the relationship remained statistically significant when we considered only the PDSTB group. We did not find a similar relationship between aSN and CD68 immunoreactivity. Regression analysis revealed no significant statistical link between the two stains.
Figure 1 Immunohistochemical staining of (A) aSN deposition in and around melanised dopaminergic neurons of the SN, (B) microglial expression of MHCII, and (C) macrophage expression of CD68. All images taken at 40X primary magnification.
Figure 2 Case-by-case semi-quantitative severity ratings for MHC class II and alpha-synuclein immunopositivity.
In our cohort of PDSTB cases (n = 17), for which clinical information was available, we then assessed how immunoreactivity may relate to clinical history. Cases were assessed based on clinical subtype (tremor or akinetic-rigid), gender, disease duration (DD), and ages of onset (AAO) and death (AAD). In addition, absence or presence of AD-related pathology was determined following ICDNS criteria . Cases were evaluated as individual data points and in groups. No correlations were found with aSN or MHCII as histological reference points, and no clinical classification seemed to reflect the relationship between aSN deposition and MHCII immunoreactivity in individual cases. A significant difference between CD68 immunoreactivity in PD cases with a DD of 10 years or less (n = 9, mean 7.3 ± 2.4 years) compared to those with a DD greater than 10 years (n = 8, mean 17 ± 5.0 years) was detected when using Student's t-test, with CD68 immunoreactivity significantly higher in cases with a shorter DD (p < 0.05). DD was highly inversely correlated with AAO (p < 0.0005). AAO was also significantly different in the two DD groups according to Student's t-test (p < 0.005) while AAD remained consistent across cases in both the shorter- (n = 9, mean 78.9 ± 5.3) and longer DD groups (n = 8, mean 78.5 ± 5.2). MHCII and aSN immunoreactivity did not show any significant variation across DD, or any other clinically defined PD groups.
Discussion
Our finding of an overall correlation between aSN deposition and MHCII-expressing microglia in the substantia nigra is in line with the finding that both phenotypic changes are associated with neurodegeneration in PD, but it remains unclear whether there is any pathogenetic link. It is perhaps more noteworthy than the correlation between alpha-synuclein deposition and microglial activation that we failed to find any correlation between these parameters and clinical indicators of disease progression.
Studies of multiple system atrophy (MSA), PSP and corticobasal degeneration (CBD) have previously detected a stochastic link between the presence of activated microglia and protein deposition in the neuroanatomic systems specifically affected by the disease and hypothesize that microglial activation may in part be induced by the accumulation of pathological protein in tissue [17,18]. aSN could be one of the pathological substrates that initiate microglial activation. However, there is no evidence that LBs can directly provoke this response [19,20]. LBs may contain complement proteins and chromogranin A [21], which can induce microglial activation in vitro [22], but in post-mortem tissue microglia have not been observed to interact preferentially with these particular LBs [19].
Much has been speculated about the potentially deleterious effects of activated microglia on neuronal survival in PD. Specifically, emphasis has been placed on the production of pro-inflammatory cytokines and reactive oxygen species potentially increasing oxidative stress on surrounding neurons. In various animal and cell culture models, the inhibition of microglial activation has been demonstrated to be neuroprotective in some circumstances [23,24]. However, there is no evidence that microglia initiate neurodegeneration, and their response does not always correlate to active cell death occurring in their microenvironment. The presence of MHCII-immunoreactive microglia in the SN of monkeys one year following chronic administration of MPTP has been interpreted as evidence that the neurodegenerative process was still active and associated with the glial response [25]. However, a comparable experiment demonstrated that although microglia do indeed remain present in the SN, they are absent from the striatum where active neurodegeneration could still be detected [26].
In contrast to the focal activation observed in animal models and acute CNS insults, microglial activation in PD is widespread and not limited to areas of marked cell death [27]. This phenomenon, and the persistence of activation in the SN long after most dopaminergic neurons have been lost, may in part be attributable to the differences in microglial responsiveness between young and aged individuals. The number of MHCII-expressing microglia in the human CNS increases steadily with age independent of disease or trauma [13,28]. In addition, microglia and astrocytes in culture harvested from older rats are more inclined to proliferation and MHCII expression and are less sensitive to transforming growth factor-beta than glia from younger donors [29]. This is in line with the findings of a study on MPTP-treated mice showing that this toxin acts in an age-dependent manner, with protracted microglial activation in older animals [30].
Our failure to find any correlation between MHCII in the parkinsonian nigra and DD, AAO, AAD, gender or predominant motor symptoms raises questions about the significance of microglial MHCII expression as a marker of their involvement in PD and in other chronic CNS diseases especially in the aged brain. Are microglia designated 'active' by the presence of certain proteins necessarily functionally active? Or does microglial MHCII expression serve as a "firewall" against T-cell invasion of already compromised CNS tissue since co-stimulators such as B7 may not be expressed at a sufficient level [31]? The use of MHCII immunoreactivity as a marker of microglial "activation" should be re-evaluated in the light of studies suggesting very long-lasting microglial involvement in chronic and late onset neurological conditions. Microglial "activation" under such conditions may have the quality of a "microglial scar" which would have different functional relevance, and this should be taken into account in the interpretation of neuroimaging studies of activated microglia [32].
Furthermore, it may also be that the presence of activated microglia is a reflection of agonal state rather than indicative of a chronic disease state, and that our microglia results are in part attributable to factors such as hypoxia or infection prior to patient death. This could account for our failure to identify clinical correlates to MHCII expression; however the strong correlation of this expression with the presence of aSN deposition remains unexplained. Another possibility is that the inflammatory response in PD peaks early in the course of the disease, at which time a pathogenetic link between MHCII and clinical progression could be detected. However, by the time of post-mortem evaluation most relevant microglial activity may have stopped. Yet, high levels of CD68 in some cases indicate that microglial phagocytosis can still occur at the time of death. As the presence of tissue macrophages may be considered a sign of ongoing tissue destruction, and macrophages are known to be crucial players in the cytotoxic phase of an inflammatory response, it is of particular interest that they were found to be more prevalent in PD cases with shorter DD. This suggests that microglial phagocytosis may not persist when it is no longer functionally relevant.
Increased extraneuronal neuromelanin and decreased aSN pathology in the SN have been associated with the progression of PD as defined by a staging model based on pathological, rather than clinical criteria [9]. It is possible that as PD progresses from one pathological tier to another, increasing extracellular neuromelanin deposition causes more microglia to adopt a phagocytic phenotype. Neuromelanin has been demonstrated to induce microglial activation in vivo [33], and, since one of the primary functions of activated microglia is the removal of debris produced by necrotic cells and neuromelanin is a readily identifiable component of this debris, the presence of this compound in the SN may very well have a relationship with the presence of phagocytic microglia in the region. However, this pathological staging is not reflective of our disease cohort. All of the cases evaluated for this study were in both clinically and pathologically advanced stages of PD, and would have fallen within the later, more severe proposed tiers. Because the model does not address clinical information regarding disease progression, the distribution of DD across our cases would not affect their pathological staging. Cases with shorter DD may have progressed more rapidly, or they may have remained presymptomatic for a longer period. Our failure to find any relationship between aSN deposition and DD does not contradict observations that LBs decrease as the disease progresses, it merely supports the assertion that our cohort was entirely situated in the most severe stages of the disease. Extraneuronal neuromelanin would be expected, and was observed, throughout our cohort. Our observation that an increase in CD68 immunoreactive, putatively phagocytic microglia is correlated with a shorter DD provides a clinical refinement beyond the scope of a model based exclusively upon pathological observations. Whether microglial phenotype is a direct result of increased neuromelanin deposition does not affect the significance of our finding that it is related to DD.
Regardless of how it is induced, microglial phagocytosis of neural debris is not a rapid process [34,35] and this peculiarity of the brain's intrinsic phagocytes may provide the most straightforward explanation for the persistence of some CD68-positive cells in the SN even after a long disease course. Alternatively, there may be a difference between early- and late-onset cases of PD with respect to their formal pathogenesis, with earlier onset cases having a lower level of phagocytosis throughout the degenerative process. Parkinsonian-type, age-related neurodegeneration, diagnosed as idiopathic PD, observed in late-onset cases may share clinical symptoms with true idiopathic PD in spite of causative, prognostic, or pathogenic differences.
In conclusion, this study demonstrates that throughout the SN, PD cases with relatively high levels of aSN deposition can be expected to contain higher numbers of MHCII positive microglia but there is no correlation with specific clinical subtypes or symptoms. We also report that, unlike CD68-expressing macrophages, neither aSN deposition nor microglial MHCII is indicative of the duration of the disease course. Both aSN deposition and microglial MHCII expression are likely to hold some as yet unknown functional significance in the progression of PD, and their careful localization and characterization throughout the brain will help to shed light on their specific role in the disease process. However, attempts to link alpha-synuclein deposition or microglial activation with the clinical course of PD should be made with caution. Our finding that CD68 immunoreactivity correlates negatively with disease duration suggests that there may be a pathogenic difference between earlier and later-onset PD. Follow-up studies addressing genomic, transcriptomic, and proteomic differences, possible drug interactions and specific clinical correlates are needed.
List of abbreviations
AAD, age at death; AAO, age at onset; AD, Alzheimer's disease; A-R, akinetic-rigid; aSN, alpha-synuclein; DD, disease duration; H-T, hemi-tremulous; LB, Lewy bodies; MHCII, major histocompatibility complex class II; MPTP, 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine; MSA, multiple systems atrophy; PD, Parkinson's disease; PDD, Parkinson's disease with dementia; PDSTB. Parkinson's Disease Society Tissue Bank; PSP, progressive supranuclear palsy; SN, substantia nigra.
Competing interests
The author(s) declare that they have no competing interests
Authors' contributions
MG and EC designed this study. EC did most of the lab work and wrote major parts of the paper. The data analysis was done jointly by EC, MG, RP and LM where indicated. DD played a crucial role in the provision of the necessary case material and contributed to the writing.
Acknowledgements
We would like to thank Dr. Kirsten Goldring, manager of the UK Parkinson's Disease Society Tissue Bank, and Miss Helen C. Cairns and Miss Louisa Djerbib, laboratory of the tissue bank. We are indebted to the brain donors and their families and their support is most gratefully acknowledged. This work was funded in part by a programme grant from the UK Parkinson's Disease Society.
==== Refs
Achim C Auer R Bergeron C Cardozo A Deprez M de Vos R Duyckaerts C Egensperger R Esiri M Frosch MP Giannini C Goebel HH Graeber MB Graham DI Gray F Haltia M Hashizume Y Ikeda K Ironside JW Kreutzberg GW Lantos P Lowe J Ludwin S Matsumoto Y Olsson Y Sasaki A Scheithauer BW Takahashi H Tolnay M Trojanowski JQ Troost D de F Webster H Global democratic consensus on neuropathological disease criteria Lancet Neurol 2002 1 340 12849392 10.1016/S1474-4422(02)00156-4
Polymeropoulos MH Lavedan C Leroy E Ide SE Dehejia A Dutra A Pike B Root H Rubenstein J Boyer R Stenroos ES Chandrasekharappa S Athanassiadou A Papapetropoulos T Johnson WG Lazzarini AM Duvoisin RC Di Iorio G Golbe LI Nussbaum RL Mutation in the alpha-synuclein gene identified in families with Parkinson's disease Science 1997 276 2045 2047 9197268 10.1126/science.276.5321.2045
Chartier-Harlin MC Kachergus J Roumier C Mouroux V Douay X Lincoln S Levecque C Larvor L Andrieux J Hulihan M Waucquier N Defebvre L Amouyel P Farrer M Destee A Alpha-synuclein locus duplication as a cause of familial Parkinson's disease Lancet 2004 364 1167 1169 15451224 10.1016/S0140-6736(04)17103-1
Kruger R Kuhn W Muller T Woitalla D Graeber M Kosel S Przuntek H Epplen JT Schols L Riess O Ala30Pro mutation in the gene encoding alpha-synuclein in Parkinson's disease Nat Genet 1998 18 106 108 9462735 10.1038/ng0298-106
Singleton AB Farrer M Johnson J Singleton A Hague S Kachergus J Hulihan M Peuralinna T Dutra A Nussbaum R Lincoln S Crawley A Hanson M Maraganore D Adler C Cookson MR Muenter M Baptista M Miller D Blancato J Hardy J Gwinn-Hardy K alpha-Synuclein locus triplication causes Parkinson's disease Science 2003 302 841 14593171 10.1126/science.1090278
Zarranz JJ Alegre J Gomez-Esteban JC Lezcano E Ros R Ampuero I Vidal L Hoenicka J Rodriguez O Atares B Llorens V Gomez Tortosa E del Ser T Munoz DG de Yebenes JG The new mutation, E46K, of alpha-synuclein causes Parkinson and Lewy body dementia Ann Neurol 2004 55 164 173 14755719 10.1002/ana.10795
Spillantini MG Schmidt ML Lee VM Trojanowski JQ Jakes R Goedert M Alpha-synuclein in Lewy bodies Nature 1997 388 839 840 9278044 10.1038/42166
Parkkinen L Soininen H Laakso M Alafuzoff I Alpha-synuclein pathology is highly dependent on the case selection Neuropathol Appl Neurobiol 2001 27 314 325 11532162 10.1046/j.0305-1846.2001.00342.x
Braak H Del Tredici K Rub U de Vos RAI Jansen Steur ENH Braak E Staging of brain pathology related to sporadic Parkinson's disease Neurobiol Aging 2003 24 197 211 12498954 10.1016/S0197-4580(02)00065-9
Parkkinen L Kauppinen T Pirttila T Autere JM Alafuzoff I Alpha-synuclein pathology does not predict extrapyramidal symptoms or dementia Ann Neurol 2005 57 82 91 15562510 10.1002/ana.20321
McGeer PL Itagaki S Boyes BE McGeer EG Reactive microglia are positive for HLA-DR in the substantia nigra of Parkinson's and Alzheimer's disease brains Neurology 1988 38 1285 1291 3399080
Moran LB Duke DC Turkheimer FE Banati RB Graeber MB Towards a transcriptome definition of microglial cells Neuregenetics 2004 5 95 108 10.1007/s10048-004-0172-5
Streit WJ Sammons NW Kuhns AJ Sparks DL Dystrophic microglia in the aging human brain Glia 2004 45 208 212 14730714 10.1002/glia.10319
Grunblatt E Mandel S Youdim MB Neuroprotective strategies in Parkinson's disease using the models of 6-hydroxydopamine and MPTP Ann N Y Acad Sci 2000 899 262 273 10863545
Kosel S Egensperger R von EU Mehraein P Graeber MB On the question of apoptosis in the parkinsonian substantia nigra Acta Neuropathol (Berl) 1997 93 105 108 9039456 10.1007/s004010050590
Vonsattel JP Aizawa H Ge P DiFiglia M McKee AC MacDonald M Gusella JF Landwehrmeyer GB Bird ED Richardson EP An improved approach to prepare human brains for research J Neuropathol Exp Neurol 1995 54 42 56 7815079
Ishizawa K Komori T Sasaki S Arai N Mizutani T Hirose T Microglial activation parallels system degeneration in multiple system atrophy J Neuropathol Exp Neurol 2004 63 43 52 14748560
Ishizawa K Dickson DW Microglial activation parallels system degeneration in progressive supranuclear palsy and corticobasal degeneration J Neuropathol Exp Neurol 2001 60 647 657 11398841
Togo T Iseki E Marui W Akiyama H Ueda K Kosaka K Glial involvement in the degeneration process of Lewy body-bearing neurons and the degradation process of Lewy bodies in brains of dementia with Lewy bodies J Neurol Sci 2001 184 71 75 11231035 10.1016/S0022-510X(00)00498-6
Rozemuller AJ Eikelenboom P Theeuwes JW Jansen Steur EN de Vos RA Activated microglial cells and complement factors are unrelated to cortical Lewy bodies Acta Neuropathol (Berl) 2000 100 701 708 11078223 10.1007/s004010000225
Yasuhara O Kawamata T Aimi Y McGeer EG McGeer PL Expression of chromogranin A in lesions in the central nervous system from patients with neurological diseases Neurosci Lett 1994 170 13 16 8041489 10.1016/0304-3940(94)90227-5
Taupenot L Ciesielski-Treska J Ulrich G Chasserot-Golaz S Aunis D Bader MF Chromogranin A triggers a phenotypic transformation and the generation of nitric oxide in brain microglial cells Neuroscience 1996 72 377 389 8737408 10.1016/0306-4522(96)83172-1
Orr CF Rowe DB Halliday GM An inflammatory review of Parkinson's disease Prog Neurobiol 2002 68 325 340 12531233 10.1016/S0301-0082(02)00127-2
Liu B Hong JS Role of microglia in inflammation-mediated neurodegenerative diseases: mechanisms and strategies for therapeutic intervention J Pharmacol Exp Ther 2003 304 1 7 12490568 10.1124/jpet.102.035048
Barcia C Sanchez Bahillo A Fernandez-Villalba E Bautista V Poza Y Poza M Fernandez-Barreiro A Hirsch EC Herrero MT Evidence of active microglia in substantia nigra pars compacta of parkinsonian monkeys 1 year after MPTP exposure Glia 2004 46 402 409 15095370 10.1002/glia.20015
Hurley SD O'Banion MK Song DD Arana FS Olschowka JA Haber SN Microglial response is poorly correlated with neurodegeneration following chronic, low-dose MPTP administration in monkeys Exp Neurol 2003 184 659 668 14769357 10.1016/S0014-4886(03)00273-5
Imamura K Hishikawa N Sawada M Nagatsu T Yoshida M Hashizume Y Distribution of major histocompatibility complex class II-positive microglia and cytokine profile of Parkinson's disease brains Acta Neuropathol (Berl) 2003 106 518 526 14513261 10.1007/s00401-003-0766-2
Mrak RE Griffin ST Graham DI Aging-associated changes in human brain J Neuropathol Exp Neurol 1997 56 1269 1275 9413275
Rozovsky I Finch CE Morgan TE Age-related activation of microglia and astrocytes: in vitro studies show persistent phenotypes of aging, increased proliferation, and resistance to down-regulation Neurobiol Aging 1998 19 97 103 9562510 10.1016/S0197-4580(97)00169-3
Sugama S Yang L Cho BP DeGiorgio LA Lorenzl S Albers DS Beal MF Volpe BT Joh TH Age-related microglial activation in 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP)-induced dopaminergic neurodegeneration in C57BL/6 mice Brain Res 2003 964 288 294 12576189 10.1016/S0006-8993(02)04085-4
Graeber MB Glial inflammation in neurodegenerative diseases 8th Annual Congress of the British Society for Immunology, Harrogate, 5–8 December 2000, Immunology 2000 Oxford: Blackwells Abstract
Cagnin A Gerhard A Banati RB In vivo imaging of neuroinflammation Eur Neuropsychopharmacol 2002 12 581 586 12468021 10.1016/S0924-977X(02)00107-4
Wilms H Rosenstiel P Sievers J Deuschl G Zecca L Lucius R Activation of microglia by human neuromelanin is NF-kappaB dependent and involves p38 mitogen-activated protein kinase: implications for Parkinson's disease FASEB J 2003 17 500 2 12631585
Reichert F Rotshenker S Deficient activation of microglia during optic nerve degeneration J Neuroimmunol 1996 70 153 161 8898724 10.1016/S0165-5728(96)00112-9
Kosel S Egensperger R Bise K Arbogast S Mehraein P Graeber MB Long-lasting perivascular accumulation of major histocompatibility complex class II-positive lipophages in the spinal cord of stroke patients: possible relevance for the immune privilege of the brain Acta Neuropathol (Berl) 1997 94 532 538 9444354 10.1007/s004010050747
| 15935098 | PMC1177985 | CC BY | 2021-01-04 16:38:22 | no | J Neuroinflammation. 2005 Jun 3; 2:14 | utf-8 | J Neuroinflammation | 2,005 | 10.1186/1742-2094-2-14 | oa_comm |
==== Front
Mol CancerMolecular Cancer1476-4598BioMed Central London 1476-4598-4-191591890410.1186/1476-4598-4-19ResearchA transcriptome map of cellular transformation by the fos oncogene Ordway Jared M [email protected] Steven D [email protected] Hong [email protected] Thomas [email protected] Department of Developmental Neurobiology, St. Jude Children's Research Hospital, 332 N. Lauderdale St., Memphis, TN, 38105, USA2 Orion Genomics, 4041 Forest Park Ave., St. Louis, MO, 63108, USA2005 26 5 2005 4 19 19 1 4 2005 26 5 2005 Copyright © 2005 Ordway et al; licensee BioMed Central Ltd.2005Ordway 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 c-fos gene was originally identified as the cellular homolog of the oncogene v-fos carried by the Finkel-Biskis-Jenkins and Finkel-Biskis-Reilly murine osteogenic sarcoma retroviruses. Sustained expression of fos is sufficient to induce cellular transformation in vitro and tumorigenesis in vivo. Fos functions as a component of the AP-1 transcription factor complex to regulate gene transcription and several differentially expressed genes have been identified in cells transformed by fos. We have extended these studies by constructing a cellular system for conditional transformation by v-fos. Using Affymetrix-based DNA microarray technology, we analyzed transcriptional changes over the course of transformation and reversion in an inducible v-fos system.
Results
Microarray analyses of temporal gene expression during the process of v-fos mediated cellular transformation and morphological reversion revealed a remarkably dynamic transcriptome. Of the more than 8000 genes analyzed in this study, 3766 genes were categorized into 18 gene-expression patterns by using self-organizing map analysis. By combining the analysis of gene expression profiles in stably transformed cells with the analysis of sequential expression patterns during conditional transformation, we identified a relatively small cohort of genes implicated in v-fos mediated cellular transformation.
Conclusion
This approach defines a general conditional cell transformation system that can be used to study the endogenous transcription regulatory mechanisms involved in transformation and tumorigenesis. In addition, this study is the first reported analysis of dynamic changes in gene expression throughout experimentally controlled morphological transformation mediated by v-fos.
==== Body
Background
The c-fos proto-oncogene encodes an immediate-early transcription factor that is rapidly and transiently induced by a variety of extracellular stimuli associated with cellular responses such as proliferation, differentiation, apoptosis and neuronal signalling [1]. The c-Fos protein functions by forming leucine zipper dimers with members of the Jun and ATF/CREB families that comprise the transcription factor complexes collectively referred to as AP-1 [2]. The tightly regulated expression and activity of AP-1 family members defines a prototypical mechanism whereby short-term extracellular signals are coupled to appropriate long-term changes in cellular phenotype by selective regulation of gene expression.
The identification of v-fos as the oncogene carried by the Finkel-Biskis-Jenkins and Finkel-Biskis-Reilly murine osteosarcoma retroviruses contributed to the realization that tumorigenic retroviruses harbor viral versions of cellular genes and that these genes can elude the regulatory constraints imposed upon the endogenous gene [3-5]. The viral fos oncogenes contain point mutations and deletions that enhance their transforming potential [6]. However, sustained expression of c-fos is sufficient to induce cellular transformation in vitro and tumorigenesis in vivo [7]. Therefore, fos-induced transformation and tumorigenesis is the consequence of inappropriate fos activity within susceptible cells rather than a gain-of-function mechanism specific to the viral fos oncogene.
Many signal transduction pathways implicated in tumorigenesis functionally converge on activation of c-fos and AP-1, suggesting that inappropriate activation of c-fos contributes to various aspects of tumorigenesis. This contribution involves direct transcriptional regulation of AP-1 target genes and secondary mechanisms of transcriptional regulation. For example, increased expression and activity of Dnmt1, a DNA methyltransferase that methylates CpG dinucleotides [8], is necessary for morphological transformation by c-fos [9]. CpG methylation within promoter regions functions as an epigenetic mark that establishes or maintains transcriptional repression by recruiting chromatin modification machinery [10]. A previous study identified specific genes that are irreversibly silenced in association with DNA hypermethylation in fos-transformed cells [11]. Therefore, during fos-mediated transformation, there is conditional deregulation of target gene expression dependent upon continual oncogene activity, in addition to long-term epigenetic reprogramming of gene expression that can persist even when the direct effects of oncogene activity are suppressed.
Studies of stably transformed cell lines have found gene expression changes associated with fos transformation and have yielded functional data that implicate differentially expressed genes in aspects of oncogenic transformation [12-14]. In the study described here, we took advantage of a conditional cellular system (LacIv-fos) that allows control of v-fos expression and morphologic transformation. This approach refines the analysis of gene expression associated with fos transformation by distinguishing gene expression changes coincident with morphological transformation from those that are potentially associated with clonal variation or phenotypic changes that occur downstream of the transformation process. Comparisons of temporal gene expression patterns during conditional cellular transformation with transcriptome profiles of cells stably transformed by c-fos and v-fos revealed a cohort of genes likely to be critical for induction and maintenance of cellular transformation.
Results
Inducible lacIv-fos system
In the LacIv-fos cell system, the control of FBJ/R v-fos expression is dependent on the presence of isopropyl-b-D-thiogalactopryanoside (IPTG) in the cell culture medium [11]. In the presence of 5 mM IPTG, LacIv-fos cells did not express v-Fos protein detectable by Western blot analysis (Figure 1a). When IPTG was washed away, LacIv-fos cells expressed v-Fos protein with peak expression detected at 72 hours following removal of IPTG and progressive loss of v-Fos protein levels upon re-addition of IPTG. In addition, cells were morphologically transformed within a 72-hour period. Transformation was indicated by an overall change in cell shape that led to a more rounded and light refractory morphology as well as dramatic cytoskeletal alterations (Figure 1b). When IPTG was added back to these transformed cells, v-Fos expression was again repressed and the cells returned to their original morphology within a 72-hour reversion period.
Figure 1 Expression of v-Fos and cell morphology rearrangement in LacIv-fos system during conditional transformation and reversion. (A) Expression of v-Fos protein expression during conditional morphological transformation of LacIv-fos cells. At time zero (t = 0), Fos protein was not detectable in cells. 24 hours after removal of IPTG, Fos protein expression was induced and the immunoreactivity increased over the 72 hour transformation period. After re-addition of IPTG, Fos protein levels rapidly decreased with no detectable signal observed after 72 hours of reversion. (B) Induction of v-Fos expression resulted in dramatic cytoskeletal changes in LacIv-fos cells. Cells were stained with anti-vinculin antibody (green) to detect focal adhesion contact sites and phalloidin (red) was used to track alterations in the actin cytoskeleton. In the presence of IPTG (t = 0), cells have well established focal adhesion sites and coordinated F-actin staining. Seventy-two hours after removal of IPTG (t = 72), cells lacked defined focal adhesion contacts and displayed disorganized F-actin staining
Self-organizing map analysis
The ability to control both v-Fos expression and morphological transformation in vitro provides an opportunity to investigate global gene expression changes relative to the temporal v-Fos cellular transformation and reversion process. RNA was extracted from LacIv-fos cells treated with 5 mM IPTG at time zero (t = 0); from cells at 24, 48 and 72 hours following the removal of IPTG (transforming); and from cells at 24, 48 and 72 hours after the re-addition of 5 mM IPTG (reverting). Previous studies have demonstrated that the addition and removal of IPTG does not itself induce persistent changes in gene expression in the parental 208F rat fibroblast cell line [9,11]. RNA samples were processed and analyzed by hybridization to Affymetrix rat U34A GeneChip microarrays.
Self-organizing map (SOM) analysis is a powerful tool that can be used to categorize gene expression data into groups that share common temporal expression profiles [15-17]. To identify patterns of gene expression, we created an SOM by using the LacIv-fos microarray data obtained at the seven time points throughout v-Fos transformation and reversion (Figure 2a). The data were mapped into 18 groups that provided the greatest pattern distinction among groups as well as the best pattern consistency within groups. Several assumptions were made on the basis of simple visualization of the map. For example, the expression of genes grouped into pattern 1 peaked 24 hours following removal of IPTG, as well as at 24 hours following re-addition of IPTG. This group likely represents genes whose expression was induced within 24 hours following culture media change rather than genes whose expression coincided with v-Fos-mediated transformation. Indeed, genes represented by pattern 1 included serum-responsive genes such as c-Jun and cyclin D1. However, other patterns included genes that were regulated in a manner more consistent with conditional v-Fos transformation and reversion. For example, pattern 8 included genes that were dramatically upregulated during the 3-day transformation process, yet return to baseline levels during the three-day reversion process (Figure 2b). This group included two probe sets specific for FBR v-fos included on the Affymetrix U34A GeneChip. These probe sets give a reliable measurement of FBJ/R v-fos expression and they did not cross-hybridize with endogenous c-fos transcripts [11]. Conversely, group 11 represents genes that were conditionally downregulated specifically during conditional v-Fos transformation (Figure 2c). Plots of the normalized signal values for each probe set included in these groups further confirmed the reliability of the classifications of gene expression profiles within the SOM (Figure 2d-e).
Figure 2 Self-organizing map (SOM) analysis of microarray expression data in the LacIv-fos system during morphological transformation and reversion. (A) Results of SOM analysis of genes depicting co-regulated clusters of genes across the seven time points throughout v-Fos mediated morphological transformation and reversion. Genes whose expression did not change significantly across time points were eliminated by using a variance filter (see Methods). The 3766 probe sets (out of a possible 8799) that passed the variation filter were grouped into 18 clustered patterns. Blue bars at the top of each graph represent the relative number of probe sets included within the SOM bin. (B) A gene-expression heat map of genes within group 8. This bin represents genes dramatically upregulated specifically during the 72 hour v-Fos transformation period (-IPTG). (C) Gene-expression heat map of genes within group 11. This bin represents genes displaying sustained downregulation during the 72 hr v-Fos transformation period (+ITPG). Red signal represents upregulated genes and green represents downregulated genes. (D-E) Normalized expression values of all genes within group 8 (D) or group 11 (E) are plotted to demonstrate the reliability of the representative SOM patterns shown in panel A.
Patterns within the SOM that are consistent with v-Fos transformation and reversion were selected on the basis of the representative expression pattern of each SOM group. For candidate genes whose expression was upregulated, the standard deviation of the mean signal values at a minimum of two of the three time points in the absence of IPTG (transforming) had to be greater than that of the mean signal values at time zero (i.e., the time prior to removal of IPTG from the culture medium). Also, the standard deviation of the mean signal values at a minimum of two of the three final time points in the presence of IPTG (reverting) had to be below that of the mean signal value of the final time point during transformation (i.e., 72 hours after removal of IPTG from the culture medium). For candidate genes whose expression was downregulated, the standard deviation of the mean signal values at a minimum of two of the three transforming time points had to be lower than that of the mean signal values at time zero, and the standard deviation of the mean signal values at a minimum of two of the three time points during reversion had to be higher than that of the final time point during transformation. On the basis of these criteria, patterns 7, 8, 14 and 15 (transforming) and patterns 5, 11, 12, 17 and 18 (reverting) were selected for further analysis. Each of these groups consisted of a large number of genes (Figure 2a). These groups likely included not only genes functionally involved in cellular transformation, but also genes that share similar expression profiles during conditional transformation and reversion yet have no role in the process or maintenance of cellular transformation itself. Interestingly, both patterns 5 and 7 included genes with higher signal-ratio values at the end of morphological reversion (t = 72 hours, reverting) than prior to v-fos induction (t = 0). This difference in transcript levels at the beginning of the time course relative to the final time point may be related to a requirement for a higher level of expression of particular genes during the reversion of morphological transformation than during maintenance of the non-transformed state. Alternatively, this may reflect a type of over-compensation in that expression levels within particular expression pattern groups and transcript levels may not completely return to steady-state within the three-day reversion period.
Stable versus inducible v-fos expression
To identify only the most promising candidate genes, we compared these datasets with gene expression profiles of the parental 208F cells and 208F cells transformed by stable expression of either c-Fos (CMVc-fos cells) or v-Fos (FBJ/R cells). Cells transformed by stable expression of c-Fos or v-Fos exhibit a large number of differentially expressed genes. For example, when differential expression was defined as a 2-fold increase or decrease based on a comparison with expression in 208F cells and when probe sets for expressed sequence tags (ESTs) were excluded, 70 probe sets were scored as upregulated and 104 were scored as downregulated in both CMVc-fos and FBJ/R cells [11]. However, eliminating genes whose expression changes are consistent in both stably and conditionally transformed cell model systems can reduce the level of nonspecific transcriptional variation. Of the 174 differentially expressed probe sets identified in comparisons of both CMVc-fos and FBJ/R cells to 208F cells, 63 (36%) were also categorized into LacIv-fos SOM patterns consistent with increased or decreased expression during conditional v-Fos transformation. Therefore, the majority of gene expression changes associated with stably Fos-transformed cell lines were not recapitulated during conditional v-Fos transformation and reversion. These findings emphasize the value of combining analyses of stable and conditional cellular transformation systems.
To address the reproducibility of gene expression changes during conditional v-Fos transformation, we performed microarray analysis using RNA samples isolated from LacIv-fos cells at time points during a second LacIv-fos transformation and reversion time course experiment. To maintain stringent biological replication of the experiment, these seven time points were treated as an independent experiment rather than as expression values that could be averaged between experiments. Expression patterns of genes differentially expressed in both CMVc-fos and FBJ/R cells relative to 208F cells, as well as conditionally regulated within the LacIv-fos SOM obtained from the first experiment, were compared to expression patterns obtained from the independent LacIv-fos transformation and reversion time course experiment. Of the 63 probe sets representing differentially expression in stably Fos-transformed cells and conditional regulation in the LacIv-fos SOM, 40 (64%) demonstrated a reproducible temporal profile of gene expression in the independent LacIv-fos transformation and reversion time course experiment (Additional file 1 and Figure 3). In addition, a previous study indicated by Northern blot analysis that representative genes (i.e. TGF-β 3 and CAIII) had nearly identical patterns of gene expression when compared to microarray analysis [11]. Therefore, the combination of gene expression analysis of stably transformed cells with the analysis of gene expression in an experimentally controlled conditional transformation system is a useful approach to reduce complicating transcriptional variation inherent in any single cell model system.
Figure 3 Expression of representative genes conditionally regulated in LacIv-fos cells and differentially expressed in both CMVc-fos cells and FBJ/R cells. Expression patterns of selected genes upregulated (A) or downregulated (B) specifically during conditional transformation, as well as in cells stably transformed by either c-Fos or v-Fos. Gene names are indicated above each pair of graphs. Line graphs indicate raw signal values obtained from the two independent LacIv-fos time course experiments. Time points at 24, 48 and 72 hours in the absence of IPTG (transforming) and at 24, 48 and 72 hours after the re-addition of IPTG (reverting) are indicated by arrows below each line graph. Bar graphs indicate raw signal values obtained in microarray analyses of the parental 208F fibroblast cell line (white), the stably transformed c-fos cell line (CMVc-fos; grey) and the stably transformed v-fos cell line (FBJ/R; black).
c-fos versus v-fos gene regulation
Unlike c-fos, v-fos contains several deletions and point mutations that affect its oncogenic potential [18]. Although sustained expression of c-fos is capable of inducing cellular transformation and tumorigenesis, v-fos is a much more potent oncogene [7]. The increased tumorigenic potential of v-Fos suggests that it directly or indirectly influences the expression of genes not affected by c-Fos. Consistent with this hypothesis, gene expression profile analyses to identify genes differentially expressed by a 2-fold margin in FBJ/R cells, but not in CMVc-fos cells, revealed a dramatic increase in the number of transcriptionally upregulated genes in FBJ/R relative to CMVc-fos cells: 337 probe sets were scored as specifically upregulated in FBJ/R cells, whereas only 70 probe sets common to CMVc-fos and FBJ/R cells were considered as increased. In contrast, 61 probe sets were scored as downregulated specifically in FBJ/R cells but 104 probe sets in both CMVc-fos and FBJ/R cells were considered as downregulated [11]. These FBJ/R cell-specific changes in gene expression were compared with the LacIv-fos SOM and the results of the independent LacIv-fos transformation and reversion time course as described previously. These comparisons revealed an additional 38 upregulated and 29 downregulated probe sets that are differentially regulated in v-fos-transformed cells, but not in c-fos transformed cells (Tables 1 and 2, Figure 4).
Table 1 Genes whose expression was upregulated (i.e., increased by a factor≥ 2) in FBJ/R cells (but not in CMVc-fos cells) and was conditionally regulated in LacIv-fos cells.
Upregulated v-Fos specific candidate genes
GenBank Accession No. Gene Pattern
AF100470 ribosome attached membrane protein 4 (RAMP4) 7
AF033027 prenylated SNARE protein Ykt6p (Ykt6) 7
AF036537 homocysteine respondent protein HCYP2 7
AF054618 cortactin isoform C 7
D00092 mRNA for 70 kd mitochondrial autoantigen 7
D30740 14-3-3 protein mRNA for mitochondrial import stimulation factor (MSF) S1 subunit 7
M21476 iodothyronine 5-monodeiodinase (5-MD) 7
M96630 sec61 homologue 7
U49930 ICE-like cysteine protease, Lice (aka Caspase 3) 7
U95161 nuclear protein E3-3 orf2 7
X13722 LDL-receptor 7
X70871 cyclin G 7
X92097 transmembrane protein rnp21.4 7
Y09332 cytosolic peroxisome proliferator-induced acyl-CoA thioesterase 7
AB015432 LAT1 (L-type amino acid transporter 1 8
AF069782 unknown mRNA (aka NAP65) 8
D16479 mitochondrial long-chain 3-ketoacyl-CoA thiolase beta-subunit of mitochondrial trifunctional protein 8
D26393 HK2 gene for type II hexokinase* 8
L12025 tumor-associated glycoprotein E4 (Tage4) 8
L12382 ADP-ribosylation factor 3 8
L19698 GTP-binding protein (ral A) 8
L38644 karyopherin beta 8
M62992 glycoprotein p62 8
U21718 clone C426 intestinal epithelium proliferating cell-associated mRNA 8
U38253 initiation factor eIF-2B gamma subunit (eIF-2B gamma) 8
X82445 C15 mRNA 8
Y00396 c-myc oncogene 8
M58587 interleukin 6 receptor ligand binding chain 14
M65253 transformation-associated protein, 34A (aka MMP10) 14
U17901 phospholipase A-2-activating protein (plap) 14
U35774 cytosolic branch chain aminotransferase 1, cytosolic 14
U92081 epithelial cell transmembrane protein antigen precursor (RTI40) 14
D12498 FGF receptor-1* 15
M74223 VGF mRNA 15
S54008 fibroblast growth factor receptor 1 beta-isoform* 15
S56464 hexokinase II (HK2)* 15
S81025 UDP-galactose:N-acetylglucosamine beta-1,4-galactosyltransferase homolog 15
J05166 band 3 Cl-/HCO3- exchanger, B3RP2 (aka Slc4a2) 15
Table 2 Genes whose expression was downregulated (i.e., decreased by a factor ≥ 2) in FBJ/R cells (but not in CMVc-fos cells) and was conditionally regulated in LacIv-fos cells.
Downregulated v-Fos-specific candidate genes
GenBank Accession No. Gene Pattern
M64780 agrin 5
M38135 cathepsin H (RCHII) 11
U52663 peptidylglycine alpha-amidating monooxygenase (PAM) 11
U75917 clathrin-associated protein 17 (AP17) 11
U75929 SPARC (aka osteonectin)* 11
X05341 3-oxoacyl-CoA thiolase 11
D10026 glutathione S-transferase Yrs-Yrs 12
J02791 acyl coenzyme A dehydrogenase medium chain 12
J03752 glutathione S-transferase 12
M93257 cathechol-O-methyltransferase 12
X05472 Rat 2.4 kb repeat DNA right terminal region 12
X74593 sorbitol dehydrogenase 12
Y09333 mitochondrial very-long-chain acyl-CoA thioesterase 12
AF034218 hyaluronidase (Hyal2) 17
AF065387 vitamin K-dependent gamma-glutamyl carboxylase 17
J02810 prostate glutathione S-transferase 17
J05031 isovaleryl-CoA dehydrogenase (IVD) 17
L01702 protein-tryosine-phosphatase (LRP) 17
U10357 pyruvate dehydrogenase kinase 2 subunit p45 (PDK2) 17
U25651 phosphofructokinase muscle isozyme 17
U75928 SPARC (aka osteonectin)* 17
Y13714 osteonectin (aka SPARC)* 17
D00512 mitochondrial acetoacetyl-CoA thiolase precursor 18
D13921 mitochondrial acetoacetyl-CoA thiolase 18
D16309 cyclin D3 18
M60921 NGF-inducible anti-proliferative putative secreted protein (PC3) 18
S7259 tissue inhibitor of metalloproteinase type 2 18
X95986 CBR gene 18
Figure 4 Expression of representative genes conditionally regulated in LacIv-fos cells and stably regulated in FBJ/R cells, but not CMVc-fos cells. Expression patterns of selected genes upregulated (A) or downregulated (B) specifically during conditional transformation, as well as in cells stably transformed by v-Fos. Gene names are indicated above each pair of graphs. Line graphs indicate raw signal values obtained from the two independent LacIv-fos time course experiments. Time points at 24, 48 and 72 hours in the absence of IPTG (transforming) and at 24, 48 and 72 hours after the re-addition of IPTG (reverting) are indicated by arrows below each line graph. Bar graphs indicate raw signal values obtained in microarray analyses of the parental 208F fibroblast cell line (white), the stably transformed c-Fos cell line (CMVc-fos; grey) and the stably transformed v-Fos cell line (FBJ/R; black).
LacIv-fos web-accessible database
The identification of conditionally regulated genes within this system provides potentially unique opportunities to investigate precise transcriptional regulatory mechanisms of endogenous genes within an experimentally amenable cellular context. Therefore, we assembled these datasets into a web-accessible database equipped with various query functions so that users can mine the data to address questions relevant to their own specific interests (see methods). The initial interface of the database is the LacIv-fos SOM (Figure 2a). Users can select SOM patterns of interest to obtain a list of all probe sets mapped within that common pattern of gene expression. Each probe set is linked to the raw Affymetrix data for both independent LacIv-fos time course experiments. This arrangement provides an opportunity for convenient confirmation of consistent regulation between the independent time course experiments. In addition, users can query the database for information about specific genes by using Genbank accession numbers or by using genes name as keywords. Keyword queries may include the qualifier functions "and" or "or". These queries will provide a list of all probe sets matching the supplied criteria, and each entry is again linked to the raw Affymetrix data obtained from both independent LacIv-fos time course experiments.
Discussion
The process of cellular transformation involves complex alterations of gene expression regulation. This level of complexity raises the challenge of identifying the gene expression changes that are most relevant to mechanisms of tumorigenesis. The challenge is especially evident in studies of primary tumor specimens in which the analyzed material represents the end state of a progressive series of events rather than a snapshot of the process itself. Cell-based model systems of transformation and tumorigenesis offer an opportunity to experimentally simulate steps along the pathway of transformation; however, these systems are subject to their own set of caveats including gene expression changes resulting from variation inherent to cell culture conditions and clonal variation. In the study described here, we attempted to provide a more focused view of relevant gene expression changes associated with oncogenic cellular transformation by analyzing a conditional cell-based model system in which morphological transformation and reversion is experimentally manipulated.
Sustained expression of fos results in the differential expression of a large number of genes [7,9,11]. By analyzing gene expression at time points throughout the conditional cellular transformation and reversion of LacIv-fos cells and by arranging these patterns into SOMs, we have clustered differentially expressed genes into cohorts with common temporal patterns of expression. This type of clustering allows identification of groups of genes with expression patterns that are consistent with a potential functional role in the process of morphological transformation. For example, the exogenous v-fos transcript itself is clustered within the conditionally upregulated group 8. This group also includes the previously reported Fos-regulated target genes such as ezrin [19] and Fra-1 [20,21] genes shown to be strongly upregulated in response to fos-induced transformation [19,21]. Interestingly, independent genes with functionally related roles in cellular transformation fall into different expression pattern bins. For example, ezrin and CD44 form a complex that plays an active role in aspects of tumor progression and metastasis such as tumor-endothelium interactions, cell migration and cell adhesion [22] but the expression patterns of these two genes differ somewhat. The expression of CD44 is more transiently induced (group 7); this characteristic is consistent with features of a gene whose expression is induced in response to serum stimulation. In contrast, ezrin expression is maintained at its maximal level throughout the 3-day morphological transformation period and rapidly returns to its initial expression levels upon repression of exogenous v-fos expression (group 8). These results demonstrate that coordinated yet distinct programs of gene expression regulation can be initiated to affect the expression of genes that function in common mechanisms of cellular transformation.
Our analysis has revealed additional conditionally regulated genes that have not been previously associated directly with Fos or AP-1. For example, PRL-1 is a protein tyrosine phosphatase originally identified as an immediate-early gene in liver regeneration [23]. PRL-1 expression is elevated in cells stably transformed by either c-fos or v-fos, and it is conditionally upregulated during LacIv-fos transformation (group 8). PRL-1 promotes cell migration and invasion in vitro and its expression is elevated in a number of cancer cell lines. PRL-1 is also involved in regulation of progression through mitosis, possibly by modulating spindle dynamics [24,25].
Sustained expression of c-Fos is capable of initiating transformation [7]; however, in our study a comparison of c-Fos transformed cells (CMVc-fos) with v-Fos transformed cells revealed additional differential gene regulation events in cells transformed by v-Fos relative to cells transformed by c-Fos. For example, expression signal values of both c-myc and the Ras GTPase, RalA, were slightly greater in CMVc-fos cells relative to 208F cells. However, expression of these genes was dramatically upregulated (>2fold) in FBJ/R cells. Likewise, PCPE and SPARC expression was slightly downregulated in CMVc-fos cells, but were dramatically repressed in FBJ/R cells. In contrast, two genes involved in enzymatic regulation of the extracellular matrix were differentially expressed only in FBJ/R cells. Expression of the matrix metalloproteinase MMP10 and the metalloproteinase inhibitor TIMP2 was specifically upregulated and downregulated in v-fos-transformed cells, respectively. Interestingly, MMP10 was previously identified as a gene whose expression was upregulated gene in cells stably transformed by FBR-v-fos [14]. The mechanisms responsible for differential expression of these genes specifically in v-fos-transformed cells are currently unknown. Some of these genes may be direct targets of v-Fos, but not of c-Fos, or targets of secondary transcription regulatory factors whose expression is deregulated specifically in v-fos-transformed cells. Alternatively, the expression of some genes may be affected by both c-Fos and v-Fos, but to a higher degree in v-fos-transformed cells.
The identification of 28 conditionally downregulated probe sets specific to v-Fos transformation raises the possibility that loss of expression of a subset of these genes may be due to chromatin modifying mechanisms leading to gene silencing. Included in our analysis are several genes previously implicated as targets for gene silencing via epigenetic repression. For example, the extracellular matrix molecule, SPARC, is downregulated in c-jun-transformed primary rat embryo fibroblasts [26]. Also, in v-jun-transformed chick embryo fibroblasts, reduction in SPARC mRNA levels has been shown to be due in part to the formation of a DNA-Sp1/3-v-Jun chromatin-associated complex [27]. Tissue inhibitor of metalloprotease 2 (TIMP-2), an endogenous inhibitor of MMP-2, has been shown to inhibit invasion and metastasis [28] and overexpression of TIMP-2 inhibited growth and reduced invasive potential in tumor cells [29]. TIMP-2 is subject to aberrant promoter hypermethylation in human cervical cancer cells and increased methylation favors development of primary cervical cancers [30]. Interestingly, treatment of human neuroblastoma cells with the DNA methyltransferase inhibitor 5-azacytidine (5-AzaC) restored TIMP-2 expression and resulted in a reduction of in vitro invasiveness [31]. The precise role of v-fos in mediating gene silencing in the conditional system is not clear; however, our system represents a tool for identifying candidate genes that are subject to epigenetic modification during the process of oncogenesis, as well as a conditional cellular system that can be employed to investigate the temporal mechanisms underlying these regulatory events.
Conclusion
Technological advances in the ability to analyze gene expression profiles on an increasingly global scale have contributed significantly to a more comprehensive view of the complex transcriptional networks that go awry in tumor cells [32,33]. Studies of in vitro cellular model systems of oncogenic transformation have provided a wealth of information relevant to both normal signal transduction pathways and tumorigenic mechanisms. However, the large number of genes differentially expressed in these model systems often complicates the identification of the most promising candidate genes for further study. Our web-accessible database of transcriptional changes detected in the conditional v-fos system provides a powerful tool to identify cohorts of gene candidates associated with specific cellular events during the process of transformation and reversion.
Methods
Cell culture and v-fos transformation and reversion time course
All cells were maintained in DMEM supplemented with 10% fetal calf serum, L-glutamine and penicillin/streptomycin at 37°C in the presence of 5% CO2. Establishment of the stably transformed cell lines, FBJ/R and CMVc-fos, was described previously [9]. Establishment of the LacIv-fos conditional cell line and the conditions required for the transformation/reversion time course were described previously [11]. Briefly, for the v-fos transformation and reversion time course experiment, total RNA was extracted from LacIv-fos cells cultured in the presence of 5 mM IPTG at time zero; 24, 48 and 72 hours after removal of IPTG (transforming); and 24, 48 and 72 hours after the re-addition of 5 mM IPTG (reverting). Total RNA was extracted from cells by using Tripure reagent according to the manufacturer's instructions (Roche). The integrity of all RNA samples was verified by using an Agilent 2100 Bioanalyzer.
Western blot and immunofluorescence analyses
Expression of Fos protein during transformation and reversion was evaluated by Western blotting analysis using an antibody directed against Fos [7]. Whole-cell lysates were prepared in lysis buffer (20 mM Tris pH 7.5, 100 mM NaCl, 0.5% NP-40, protease inhibitor cocktail [Roche]) and separated on a 12% SDS-polyacrylamide gel. For immunofluorescence microscopy, cells were fixed in 4% (w/v) formaldehyde. Fixed cells were incubated with primary mouse monoclonal anti-vinculin antibody (Sigma; dilution of 1:200) and phalloidin-congjugated to Alexa-568 (Molecular Probes). An anti-mouse secondary antibody (dilution of 1:1000) conjugated to Alexa-488 (Molecular Probes) was used to detect vinculin.
Microarray analyses
RNA samples were processed for hybridization without amplification and hybridized to Affymetrix Rat Genome U34A GeneChips that include probe sets representing approximately 7,000 annotated genes and 1,000 EST clones [34]. Labelling and hybridization were preformed as described [35]. GeneChips were scanned using a laser confocal scanner (Agilent Technologies) and images were analyzed using Affymetrix Microarray Suite v.5.0. Datasets were standardized by global scaling of the average fluorescent intensities of all probe sets to a constant target value of 500 for all arrays. Quality control parameters for each hybridization were within MIAME compliant specifications [36]. A variance filter was applied to the dataset to remove data from probe sets representing genes that were not expressed throughout the time course and to standardize expression values for genes whose expression was scored absent at particular time points [37]. Data that were derived from probe sets that reported absent change calls at all seven time points were removed from the analysis. To standardize expression values of genes whose expression was scored as absent at fewer than seven time points, we converted the signal values that corresponded to absent change calls to a value of 1. Genes scored as differentially expressed in CMVc-fos cells and FBJ/R cells relative to 208F cells had a signal log ratio ≥ 1 (increased) or ≤ -1 (decreased), and a change p-value < 0.001 (increased) or > 0.99 (decreased). For visualization of specific profiles of gene expression during the LacIv-fos conditional transformation and reversion time course, signal values obtained at each LacIv-fos time point were plotted by using Microsoft Excel.
SOM analysis
Microarray data that were obtained from the LacIv-fos time course experiments and passed the variance filter were grouped into relative expression pattern bins by the self-organizing map (SOM) program in the GeneMaths software package (version 1.5; Applied Maths, Austin, TX). The SOM analysis was performed by using a 6 × 3 node format to allow optimal representation of gene expression patterns in a reasonably small number of independent bins.
The SOM searchable database
The SOM searchable database has been implemented on an Open Source MySQL 4.0.14 relational database management system. The database has a web interface at . The web application to query and manage the database is driven with a server-side scripting technology JSP. The entries of this database are generated from the tab-delimited output files of the SOM analysis. Data obtained from the SOM analysis can be queried by three approaches. Users can click on a graphical pattern key to view a list of all genes whose expression profiles match the particular self-organization pattern. Each entry in the list contains two parameters associated with the respective gene probe, i.e., the Affymetrix probe set ID number and a description of the gene. The detailed Affymetrix analysis data for a probe set can be obtained by clicking the link of the respective probe set ID number. The corresponding Affymetrix analysis page provides information about the number of oligonucleotide pairs, the signal magnitude, the detection status, and the detection p value at seven different time points derived from 2 independent experiments. Another search method is to query via the GenBank accession number. The query by accession number is based on the fact that each Affymetrix probe set ID number in our current experiments corresponds to a gene accession number. If an accession number in a query matches with any part within a probe set ID number, the entry containing the probe set ID number, its SOM pattern, and gene description is displayed. Hyperlinks are provided for both probe set ID numbers and SOM pattern. By following the link for a probe set ID number, the user can view the detailed Affymetrix analysis data for a selected gene probe; by following the link for an SOM pattern, the user can view the detailed Affymetrix analysis data for all the gene probes within the selected pattern. Also, a keyword-based text search is included in the database. A search for keywords in the paragraph of the gene description field is conducted in a case-insensitive manner. The search-by-keyword method is facilitated by the use of Boolean operations. The search can narrow or broaden quickly the results of the search by combining two or more keyword(s) with only one type of Boolean (and / or) operation.
Authors' contributions
JMO prepared cells, purified RNA, and analyzed the microarray data, SDF performed immunofluoresence microscopy and Western blot analysis, and HR designed the web-accessible database. JMO, SDF, HR, and TC composed the manuscript.
Supplementary Material
Additional File 1
Genes differentially expressed (i.e., expression varied by a factor of at least 2) in CMVc-fos cells and in FBJ/R cells and conditionally regulated in LacIv-fos cells.
Click here for file
Acknowledgements
We thank John J. Morris and the Clinical Applications Core Technology Center at St Jude Children's Research Hospital (SJCRH) for generating microarray data and providing advice about data analysis. We also thank Julia Cay Jones of Scientific Editing at SJCRH for editorial comments and The Hartwell Center at SJCRH for advice on web-site preparation and hosting. This work was supported by the National Institute of Health grant RO1 CA 84139 (TC) and the American Lebanese Syrian Associated Charities (ALSAC).
==== Refs
Morgan JI Curran T Stimulus-transcription coupling in the nervous system: involvement of the inducible proto-oncogenes fos and jun Annu Rev Neurosci 1991 14 421 451 1903243 10.1146/annurev.ne.14.030191.002225
Curran T Reddy EP, Shalka AM, Curran T The fos oncogene The oncogene handbook 1988 Amsterdam: Elsevier 307 325
Finkel MP Biskis BO Jinkins PB Virus induction of osteosarcomas in mice Science 1966 151 698 701 5218049
Curran T Peters G Van Beveren C Teich NM Verma IM FBJ murine osteosarcoma virus: identification and molecular cloning of biologically active proviral DNA J Virol 1982 44 674 682 6292525
Curran T Teich NM Candidate product of the FBJ murine osteosarcoma virus oncogene: characterization of a 55,000-dalton phosphoprotein J Virol 1982 42 114 122 6283132
Curran T Miller AD Zokas L Verma IM Viral and cellular fos proteins: a comparative analysis Cell 1984 36 259 268 6319013 10.1016/0092-8674(84)90219-8
Miao GG Curran T Cell transformation by c-fos requires an extended period of expression and is independent of the cell cycle Mol Cell Biol 1994 14 4295 4310 8196666
Bestor TH Verdine GL DNA methyltransferases Curr Opin Cell Biol 1994 6 380 399 7917329 10.1016/0955-0674(94)90030-2
Bakin AV Curran T Role of DNA 5-methylcytosine transferase in cell transformation by fos Science 1999 283 387 390 9888853 10.1126/science.283.5400.387
Cross SH Bird AP CpG islands and genes Curr Opin Genet Dev 1995 5 309 314 7549424 10.1016/0959-437X(95)80044-1
Ordway JM Williams K Curran T Transcription repression in oncogenic transformation: common targets of epigenetic repression in cells transformed by Fos, Ras or Dnmt1 Oncogene 2004 23 3737 3748 14990994 10.1038/sj.onc.1207483
Hennigan RF Hawker KL Ozanne BW Fos-transformation activates genes associated with invasion Oncogene 1994 9 3591 3600 7970719
Ozanne BW McGarry L Spence HJ Johnston I Winnie J Meagher L Stapleton G Transcriptional regulation of cell invasion: AP-1 regulation of a multigenic invasion programme Eur J Cancer 2000 36 1640 1648 10959050 10.1016/S0959-8049(00)00175-1
Johnston IM Spence HJ Winnie JN McGarry L Vass JK Meagher L Stapleton G Ozanne BW Regulation of a multigenic invasion programme by the transcription factor, AP-1: re-expression of a down-regulated gene, TSC-36, inhibits invasion Oncogene 2000 19 5348 5358 11103936 10.1038/sj.onc.1203927
Kohonen T The self-organizing map Proceedings of the IEEE 1990 9 1464 1480 10.1109/5.58325
Tamayo P Slonim D Mesirov J Zhu Q Kitareewan S Dmitrovsky E Lander ES Golub TR Interpreting patterns of gene expression with self-organizing maps: methods and application to hematopoietic differentiation Proc Natl Acad Sci U S A 1999 96 2907 2912 10077610 10.1073/pnas.96.6.2907
Toronen P Kolehmainen M Wong G Castren E Analysis of gene expression data using self-organizing maps FEBS Lett 1999 451 142 146 10371154 10.1016/S0014-5793(99)00524-4
Curran T Verma IM FBR Murine osteosarcoma virus. I. Molecular analysis and characterization of a 75,000-dalton gag-fos fusion product Virology 1984 135 218 228 6203214 10.1016/0042-6822(84)90132-6
Lamb RF Ozanne BW Roy C McGarry L Stipp C Mangeat P Jay DG Essential functions of ezrin in maintenance of cell shape and lamellipodial extension in normal and transformed fibroblasts Curr Biol 1997 7 682 688 9285722 10.1016/S0960-9822(06)00295-8
Cohen DR Curran T fra-1: a serum-inducible, cellular immediate-early gene that encodes a fos-related antigen Mol Cell Biol 1988 8 2063 2069 3133553
Braselmann S Bergers G Wrighton C Graninger P Superti-Furga G Busslinger M Identification of Fos target genes by the use of selective induction systems J Cell Sci Suppl 1992 16 97 109 1297655
Martin TA Harrison G Mansel RE Jiang WG The role of the CD44/ezrin complex in cancer metastasis Crit Rev Oncol Hematol 2003 46 165 186 12711360
Diamond RH Cressman DE Laz TM Abrams CS Taub R PRL-1, a unique nuclear protein tyrosine phosphatase, affects cell growth Mol Cell Biol 1994 14 3752 3762 8196618
Zeng Q Dong JM Guo K Li J Tan HX Koh V Pallen CJ Manser E Hong W PRL-3 and PRL-1 promote cell migration, invasion, and metastasis Cancer Res 2003 63 2716 2722 12782572
Wang J Kirby CE Herbst R The tyrosine phosphatase PRL-1 localizes to the endoplasmic reticulum and the mitotic spindle and is required for normal mitosis J Biol Chem 2002 277 46659 46668 12235145 10.1074/jbc.M206407200
Mettouchi A Cabon F Montreau N Vernier P Mercier G Blangy D Tricoire H Vigier P Binetruy B SPARC and thrombospondin genes are repressed by the c-jun oncogene in rat embryo fibroblasts EMBO J 1994 13 5668 5678 7988564
Chamboredon S Briggs J Vial E Hurault J Galvagni F Oliviero S Bos T Castellazzi M v-Jun downregulates the SPARC target gene by binding to the proximal promoter indirectly through Sp1/3 Oncogene 2003 22 4047 4061 12821939 10.1038/sj.onc.1206713
Montgomery AM Mueller BM Reisfeld RA Taylor SM DeClerck YA Effect of tissue inhibitor of the matrix metalloproteinases-2 expression on the growth and spontaneous metastasis of a human melanoma cell line Cancer Res 1994 54 5467 5473 7923181
Imren S Kohn DB Shimada H Blavier L DeClerck YA Overexpression of tissue inhibitor of metalloproteinases-2 retroviral-mediated gene transfer in vivo inhibits tumor growth and invasion Cancer Res 1996 56 2891 2895 8674034
Ivanova T Vinokurova S Petrenko A Eshilev E Solovyova N Kisseljov F Kisseljova N Frequent hypermethylation of 5' flanking region of TIMP-2 gene in cervical cancer Int J Cancer 2004 108 882 886 14712492 10.1002/ijc.11652
Cappabianca L Farina AR Tacconelli A Mantovani R Gulino A Mackay AR Reconstitution of TIMP-2 expression in SH-SY5Y neuroblastoma cells by 5-azacytidine is mediated transcriptionally by NF-Y through an inverted CCAAT site Exp Cell Res 2003 286 209 218 12749850 10.1016/S0014-4827(03)00072-7
Harkin DP Uncovering functionally relevant signalling pathways using microarray-based expression profiling Oncologist 2000 5 501 507 11110602 10.1634/theoncologist.5-6-501
Vogelstein B Kinzler KW Cancer genes and the pathways they control Nat Med 2004 10 789 799 15286780 10.1038/nm1087
Affymetrix Website:
Affymetrix Website
Brazma A Hingamp P Quackenbush J Sherlock G Spellman P Stoeckert C Aach J Ansorge W Ball CA Causton HC Gaasterland T Glenisson P Holstege FC Kim IF Markowitz V Matese JC Parkinson H Robinson A Sarkans U Schulze-Kremer S Stewart J Taylor R Vilo J Vingron M Minimum information about a microarray experiment (MIAME)-toward standards for microarray data Nat Genet 2001 29 365 371 11726920 10.1038/ng1201-365
Ross ME Mahfouz R Onciu M Liu HC Zhou X Song G Shurtleff SA Pounds S Cheng C Ma J Ribeiro RC Rubnitz JE Girtman K Williams WK Raimondi SC Liang DC Shih LY Pui CH Downing JR Gene expression profiling of pediatric acute myelogenous leukemia Blood 2004 104 3679 3687 15226186 10.1182/blood-2004-03-1154
| 15918904 | PMC1177986 | CC BY | 2021-01-04 16:36:35 | no | Mol Cancer. 2005 May 26; 4:19 | utf-8 | Mol Cancer | 2,005 | 10.1186/1476-4598-4-19 | oa_comm |
==== Front
Nutr JNutrition Journal1475-2891BioMed Central London 1475-2891-4-201592979010.1186/1475-2891-4-20ResearchUse of plant stanol ester margarine among persons with and without cardiovascular disease: Early phases of the adoption of a functional food in Finland Simojoki Meri [email protected] Riitta [email protected] Antti [email protected] Hannu [email protected] John D [email protected] Joseph K [email protected] Pekka [email protected] Department of Epidemiology and Health Promotion, National Public Health Institute, Helsinki, Finland2 The UKK Institute for Health Promotion, Tampere, Finland3 Department of Forest Resource Management, University of Helsinki, Finland4 International Epidemiology Institute, Rockville, Maryland, USA5 Department of Medicine, Vanderbilt University Medical Center and Vanderbilt-Ingram Cancer Center, Nashville TN, USA2005 1 6 2005 4 20 20 7 1 2005 1 6 2005 Copyright © 2005 Simojoki et al; licensee BioMed Central Ltd.2005Simojoki 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 plant stanol ester margarine Benecol® is a functional food that has been shown to lower effectively serum total and LDL-cholesterol. The purpose of this post-marketing study is to characterize users of plant stanol ester margarine with and without cardiovascular disease.
Methods
A cohort of plant stanol ester margarine users was established based on a compilation of 15 surveys conducted by the National Public Health Institute in Finland between 1996–2000. There were 29 772 subjects aged 35–84 years in the cohort. The users of plant stanol ester margarine were identified by the type of bread spread used.
Results
The plant stanol ester margarine was used as bread spread by 1332 (4.5%) subjects. Almost half (46%) of the users reported a history of cardiovascular disease. Persons with cardiovascular disease were more likely to use plant stanol ester margarine (8%) than persons without cardiovascular disease (3%). Users with and without cardiovascular disease seemed to share similar characteristics.
In particular, they were elderly people with otherwise healthy life-styles and diet. They were less likely smokers, more likely physically active and less likely obese than nonusers. The users reported being in good or average health in general and having used cholesterol-lowering drugs.
Conclusion
Plant stanol ester margarine seems to be used by persons for whom it was designed and in a way it was meant: as part of efforts for cardiovascular disease risk reduction.
Plant stanolmargarinelife-stylecardiovascular diseaseFinland
==== Body
Introduction
Functional foods are becoming increasingly available to consumers worldwide. A food can be regarded as "functional" if it has beneficial effects on one or more target functions in the body, beyond obvious nutritional effects, that are related to either an improved state of health and well-being and/or a reduction of disease risk [1]. As the health effects of long-term use of functional foods are often unknown, there is a need for post-marketing research [2-4].
The plant stanol ester margarine Benecol® (Raisio Group, Raisio, Finland), launched in Finland in 1995, is an example of a functional food that has been shown to lower effectively serum total and LDL-cholesterol levels in several interventions [5,6]. Dietary plant sterols, especially sitostanol, reduce serum cholesterol by inhibiting cholesterol absorption. The present source of plant sterol is extraction from a byproduct inthe refining of vegetable oils or from the oil obtained from pinewoodpulp in papermaking. A tenth of the amount of plant sterols found in these plant sterol margarines occurs naturally in a normal diet. Use of plant sterol ester margarine is recommended, as part of a healthy diet, by the Finnish Society for Internal Medicine for the primary and secondary prevention of coronary heart disease in Finland [7].
Elevated blood cholesterol is one of the main causal risk factors for coronary heart disease [8]. Other factors that affect coronary heart disease include socio-demographic and life-style characteristics such as age, gender, education, region, body-mass index, smoking, physical activity and diet [9]. As many of these factors relate to life-style, the risk can be modified to great extent by changes in health behavior. Indeed, the North Karelia Project in Finland revealed that population rates of cardiovascular diseases can be influenced remarkably by changes in people's dietary habits [8].
The National Public Health Institute in Finland has monitored health behaviors and chronic disease risk factors in the Finnish population regularly since 1972. Two survey systems are especially relevant for the present study: the Finnish Adult Health Behavior Survey [10] and the Chronic Disease Risk Factor Survey (Finrisk) [11]. In these surveys, random samples of the Finnish population are drawn and risk factors are assessed by a self-administered questionnaire and direct measurements. Dietary habits, smoking and physical activity are of special interest in the monitoring. The surveys conducted in 1996 or later include a question on plant stanol ester margarine use.
The health effects, both beneficial and adverse, of long-term use of the plant stanol ester margarine have not been studied. Therefore, a post-marketing study was started at the National Public Health Institute in Finland in 1998 [12]. The study includes several sub-studies where existing survey data as well as health registries in Finland are analyzed. The data on the adoption of the functional food during early post-marketing phase is unique and of importance in relation also to other products of the kind and to the efficacy of dietary counseling.
Our previous study revealed that users of plant stanol ester margarine are elderly people (mean age 59 years), about half of whom have been diagnosed with some form of cardiovascular disease [12]. The purpose of this study is to characterize users of plant stanol ester margarine with and without cardiovascular disease at an early post-marketing phase. The characteristics of interest are socioeconomic background, life-style factors, diet and medication. This information is needed when assessing the efficacy of dietary counseling and the possible health effects of the long-term use of plant stanol ester margarine.
Methods
Fifteen independent health surveys conducted by the National Public Health Institute in Finland between 1996 and 2000 were combined into study cohort (Table 1 in Additional File 1). The main surveys were the annual nationwide Finnish Adult Health Behavior Survey and the Chronic Disease Risk Factor Survey (Finrisk) carried out in five areas of Finland. In each survey, sampling was either entirely random or stratified random with respect to age and geographical area.
In each survey, the subjects completed mailed questionnaires concerning their health and life-style and returned them by mail or in person. In the Finrisk Survey, clinical stage with health examination was included. In the Finrisk Senior Survey, elderly participants aged 65–74 years, who were not able to come to the health examination, were interviewed and examined at their home. The participation rates in the surveys varied between 64–79%.
The users of plant stanol ester margarine were identified on the basis of the following question present in each survey questionnaire: "What type of fat do you usually use on bread?" Choices were no fat at all, low-fat spread, soft margarine, butter-oil mixture, butter, and plant stanol ester margarine Benecol®. The plant stanol ester margarine Benecol® was the only plant sterol margarine available for consumers in Finland during the time the health surveys were conducted.
Subjects who reported myocardial infarction, stroke, hypertension, heart failure, angina pectoris or intermittent claudication were classified as having cardiovascular disease. Subjects who did not report any of these conditions were classified as not having cardiovascular disease.
Overall, 42 692 questionnaires were returned. When duplicate respondents were identified, only the first response was retained, thereby reducing the size of the cohort to 42 406 subjects. Because cardiovascular diseases and the use of plant stanol ester margarine are infrequent in younger age groups, we further excluded respondents aged 15–34 years from the analysis. The final study population consisted of 29 772 subjects aged 35–84 years, 14 262 men and 15 510 women.
Unadjusted prevalence percentages among users with and without cardiovascular disease were computed for age, sex, socioeconomic background, life-style, body mass index, self-perceived health, diet and medication. Odds ratios and 95% confidence limits adjusted for age were calculated using logistic regression models to compare users and nonusers of plant stanol ester margarine. Subjects with and without cardiovascular disease were combined in odds ratio calculations. Data management and statistical analysis were performed using the SAS System (version 6.12) [13].
Results
Overall, the plant stanol ester margarine was used as the bread spread of choice by 1332 (4.5%) subjects. Almost half (46%) of the users and one fourth (25%) of the nonusers reported a history of cardiovascular disease. Table 2 (in Additional File 2)presents the distribution of cardiovascular diseases among users and nonusers of plant stanol ester margarine.
Among subjects with cardiovascular disease, the proportion of users of plant stanol ester margarine was 8%, whereas the proportion among subjects without the disease was 3% (Table 3 in Additional File 3). In general, users with and without cardiovascular disease shared similar characteristics. The use of plant stanol ester margarine was most common among subjects aged 65–74 years. Men were slightly more likely to use plant stanol ester margarine than women. Urban and married people with higher levels of education were represented more among the users than nonusers of plant stanol ester margarine.
Among the users of plant stanol ester margarine, there were less smokers and more physically active people as well as less of those with high levels of stress compared with nonusers (Table 4 in Additional File 4). The users were also less often obese (defined as body mass index over 30 kg/m2).
Self-perceived health was good or average among users of plant stanol ester margarine (Table 5 in Additional File 5). Users of plant stanol ester margarine consumed higher amounts of alcohol compared with nonusers. Substantially greater proportion of users also consumed a healthy diet compared with nonusers.
Users of plant stanol ester margarine used cholesterol-lowering drugs more commonly than nonusers (Table 6 in Additional File 6). The use of hypertensive drugs, diabetic drugs and acetylsalicylic acid for acute myocardial infarction (AMI) prevention, as well as cardiovascular disease drugs, were also more common among users than nonusers of plant stanol ester margarine.
Some continuous variables were evaluated separately for users with and without cardiovascular disease. The mean age was 63 years among the users with cardiovascular disease and 57 years among the users without the disease. The amounts of plant stanol ester margarine on bread were almost the same among users with and without cardiovascular disease: 21 and 20 g per day, respectively.
Blood cholesterol had been measured among 82% and 64%, elevated blood cholesterol had been diagnosed among 85% and 72%, and dietary counseling for lowering of blood cholesterol had been received by 83% and 71%, of users with and without cardiovascular disease, respectively. Among nonusers, the respective proportions were substantially lower. The percentages for blood cholesterol measurement, elevated blood cholesterol, and dietary counseling were 58% and 28%, 48% and 25%, and 52% and 31%, for nonusers with and without cardiovascular disease, respectively.
Discussion
The users of plant stanol ester margarine are a self-selected group of persons who have taken an active interest in their health. They use plant stanol ester margarine as part of a generally healthy life-style and diet. Nevertheless, they commonly have a history of cardiovascular disease or are at risk to have it. Thus plant stanol ester margarine seems to be used by persons for whom it was designed and in a way it was meant: as part of efforts for cardiovascular disease risk reduction.
Users of plant stanol ester margarine appear in this early post-marketing adoption phase to be elderly people. They reported generally good or average health status and took more commonly cholesterol-lowering drugs as compared with nonusers. Overall, users with cardiovascular disease seemed to share similar characteristics compared with users without cardiovascular disease, although users with cardiovascular disease tended to be slightly older.
As the plant stanol ester margarine has been recommended in Finland for the primary and secondary prevention of coronary heart disease [7], we assume that the users with cardiovascular disease attempt to control elevated cholesterol levels and prevent the progress of the disease. Users of plant stanol ester margarine who do not have cardiovascular disease probably also aim to control cholesterol levels and thus prevent cardiovascular disease. Unfortunately, we were unable to directly evaluate this issue, since data on indication for use of plant stanol ester margarine was not collected.
However, the plant stanol ester margarine users both with and without cardiovascular disease obviously had more cholesterol problems than nonusers as they reported more of elevated blood cholesterol and dietary counseling for lowering of blood cholesterol. Furthermore, as the price of the plant stanol ester margarine is 3–5 times higher than other margarines, its use is likely based on clearly perceived need, i.e. the control of blood cholesterol levels.
The use of plant stanol ester margarine may also result from "practical reasons": if one spouse uses it, the other may be more likely to use it too. Previous studies show spouse concordance for physiological and behavioral indicators such as plasma cholesterol and triglyceride, blood pressure, diet, body mass index, smoking and physical exercise [14-17]. The similarity in the characteristics of plant stanol ester margarine users with and without cardiovascular disease might well be explained by family reasons.
Our information on the use of plant stanol ester margarine was based on a single question concerning the type of bread spread usually used. Although we have not validated the question against real use, we assume that the question reliably measures current use and that conclusions can be drawn concerning the characteristics of the users. Information concerning the quantity of the plant stanol ester margarine use was obtained from the Finrisk survey and was shown to be 20 g per day, on an average. A picture from a portion size picture booklet describing the amount of spread on bread was used to improve the validity of the estimation [18]. The quantity should be 20–25 g per day to achieve an optimal cholesterol-lowering effect [6,7], but it is clear from our data that higher levels of use are common.
The age and geographic distributions of the study cohort of 42 406 persons differ slightly from the general Finnish population. The proportion of persons aged 55–64 years is greater (24% versus 20%) and 65–84 years smaller (19% versus 24%) than in the general population. The proportion of persons living in eastern Finland is greater (20% versus 11%) and southern Finland smaller (30% versus 40%) compared with the general population. As the proportion of users is high among less represented groups, i.e. persons aged 65–74 years and those living in southern and thus urban areas of Finland, our cohort probably underestimates the frequency of the use of plant stanol ester margarine in the Finnish population.
Functional foods are usually carefully examined with respect to their health effects and safety before they come to market. These studies often have limitations such as short length of follow-up. Also effects in certain population or patient subgroups might be unknown due to small sample size. Therefore there is a need for post-marketing surveillance. Until now only few post-marketing studies concerning functional foods have been carried out namely with a food additive sweetener aspartame [19] and a non-energy fat substitute olestra [20] as the most notable exceptions.
The LDL-cholesterol reducing effect of plant sterols and stanols may result in reduced heart disease rates. However, they appear to somewhat lower lipid-standardized concentrations of the plasma carotenoids [2,5]. Also the health effects of high daily amounts of plant sterols and stanols are unknown. In our future follow-up studies, we will evaluate possible health effects, both beneficial and adverse, of the long-term use of the plant stanol ester margarine by linking available Finnish health outcome registries to the study cohort [12].
Competing interests
MS in Finland and JDB and JKM in Maryland, USA, have received funding and/or salary from McNeil Consumer Healthcare Company and Raisio Group, Raisio, Finland. However, neither of these organizations is financing publication of this manuscript. None of the authors hold currently any stocks or shares in McNeil Consumer Healthcare Company or Raisio Group.
Authors' contributions
MS, RL and AU conceived the study, MS performed the statistical analysis. JDB and JKM participated to the collection of the cohort, funding, statistical analysis and drafting of the manuscript. HR helped in statistical issues. PP participated in general supervision of the research group, drafting the manuscript and collecting the data base. All authors have read and approved the manuscript.
Supplementary Material
Additional File 1
Population surveys in Finland assembled for plant stanol ester margarine evaluation (Table 1).
Click here for file
Additional File 2
Self-reported cardiovascular diseases among 35–84 year-old users and nonusers of plant stanol ester margarine (Table 2)
Click here for file
Additional File 3
Age, sex, and socioeconomic background among 35–84 year-old users and nonusers of plant stanol ester margarine (Table 3)
Click here for file
Additional File 4
Life-style characteristics and body mass index among 35–84 year-old users and nonusers of plant stanol ester margarine (Table 4)
Click here for file
Additional File 5
Self-perceived health and diet among 35–84 year-old users and nonusers of plant stanol ester margarine. (Table 5)
Click here for file
Additional File 6
Medication among 35–84 year-old users and nonusers of plant stanol ester margarine (Table 6)
Click here for file
Acknowledgements
Funding for this study was provided in part by the International Epidemiology Institute through a grant by the McNeil Consumer Healthcare Company and Raisio Group, Raisio, Finland.
==== Refs
Diplock AT Aggett PJ Ashwell M Bornet F Fern EB Roberfroid MB Scientific concepts of functional foods in Europe: consensus document Br J Nutr 1999 81 S1 S27 10999022
Plat J Kerckhoffs DAJM Mensink RP Therapeutic potential of plant sterols and stanols Curr Opin Lipidol 2000 11 571 76 11086329 10.1097/00041433-200012000-00002
Vainio H Mutanen M Functional foods – blurring the distinction between food and medicine Scand J Work Environ Health 2000 26 178 80 10817384
Raulio S Nurttila A Mannonen L Adding phytosterols and-stanols to food – Modelling the amount received by Finnish adults 2001 10 Helsinki: National Food Agency publications
Law M Plant sterol and stanol margarines and health BMJ 2000 320 861 64 10731187 10.1136/bmj.320.7238.861
Miettinen TA Puska P Gylling H Vanhanen H Vartiainen E Reduction of serum cholesterol with sitostanol-ester margarine in a mildly hypercholesterolemic population N Engl J Med 1995 333 1308 12 7566021 10.1056/NEJM199511163332002
Suomen Yhdistys Sisätautilääkärien Suomen Seura Kardiologinen Suomen Verenpaineyhdistys Kunnallislääkärit-yhdistys Suomen Yhdistys Teollisuuslääketieteen Suomen Sydäntautiliitto Prevention of coronary heart disease in clinical practice [in Finnish] Suom Lääkäril 1996 51 783 802
Puska P Nutrition and mortality: the Finnish experience Acta Cardiol 2000 55 213 20 11041119
Gartside PS Wang P Glueck CJ Prospective assessment of coronary heart disease risk factors: the NHANES I epidemiologic follow-up study (NHEFS) 16-year follow-up J Am Coll Nutr 1998 17 263 69 9627913
Helakorpi S Patja K Prättälä R Uutela A Health Behaviour and Health among Finnish Adult Population, Spring 2001 2001 B16 Helsinki: Publications of the National Public Health Institute
Vartiainen E Jousilahti P Alfthan G Sundvall J Pietinen P Puska P Cardiovascular risk factor changes in Finland, 1972–1997 Int J Epidemiol 2000 29 49 56 10750603 10.1093/ije/29.1.49
Anttolainen M Luoto R Uutela A Boice JD JrBlot WJ McLaughlin JK Puska P Characteristics of users and nonusers of plant stanol ester margarine in Finland: An approach to study functional foods J Am Diet Assoc 2001 101 1365 68 11716321 10.1016/S0002-8223(01)00327-3
The SAS System 1996 Cary, NC: SAS Institute Inc
Barrett-Connor E Suarez L Criqui MH Spouse concordance of plasma cholesterol and triglyceride J Chronic Dis 1982 35 333 40 6896060 10.1016/0021-9681(82)90004-2
Speers MA Kasl SV Freeman DH JrOstfeld AM Blood pressure concordance between spouses Am J Epidemiol 1986 123 818 29 3962964
Lindsted K Kuzma JW Husband-wife diet concordance and changes in dietary practices by surviving spouses of cancer cases Nutr Cancer 1990 13 175 87 2308873
Macken LC Yates B Blancher S Concordance of risk factors in female spouses of male patients with coronary heart disease J Cardiopulm Rehabil 2000 20 361 68 11144042 10.1097/00008483-200011000-00005
Pietinen P Hartman AM Haapa E Räsänen L Haapakoski J Palmgren J Albanes D Virtamo J Huttunen JK Reproducibility and validity of dietary assessment instruments, I. A self-administered food use questionnaire with a portion size picture booklet Am J Epidemiol 1988 128 655 66 2458036
Butchko HH Tschanz C Kotsonis FN Postmarketing surveillance of food additives Regul Toxicol Pharmacol 1994 20 105 18 7838988 10.1006/rtph.1994.1039
Patterson RE Kristal AR Peters JC Neuhouser ML Rock CL Cheskin LJ Neumark-Sztainer D Thornquist MD Changes in diet, weight, and serum lipid levels associated with olestra consumption Arch Intern Med 2000 160 2600 04 10999973 10.1001/archinte.160.17.2600
| 15929790 | PMC1177987 | CC BY | 2021-01-04 16:39:31 | no | Nutr J. 2005 Jun 1; 4:20 | utf-8 | Nutr J | 2,005 | 10.1186/1475-2891-4-20 | oa_comm |
==== Front
Reprod Biol EndocrinolReproductive biology and endocrinology : RB&E1477-7827BioMed Central London 1477-7827-3-231594149010.1186/1477-7827-3-23ResearchStudies on substantially increased proteins in follicular fluid of bovine ovarian follicular cysts using 2-D PAGE and MALDI-TOF MS Maniwa Jiro [email protected] Shunsuke [email protected] Naoki [email protected] Takato [email protected] Graduate School of Biosphere Science, Hiroshima University, Higashi-Hiroshima, Hiroshima 739-8528, Japan2 Preclinical Sciences Department, AstraZeneca KK, Osaka 531-0076, Japan3 Graduate School of Science, Hiroshima University, Higashi-Hiroshima, Hiroshima 739-8526, Japan2005 8 6 2005 3 23 23 2 5 2005 8 6 2005 Copyright © 2005 Maniwa et al; licensee BioMed Central Ltd.2005Maniwa 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 identify substantially increased proteins in bovine cystic follicular fluid (FF) in order to clarify the pathology and etiology of bovine ovarian follicular cysts (BOFC).
Methods
Proteins in normal and cystic FF samples were subjected to two-dimensional polyacrylamide gel electrophoresis (2-D PAGE) and were compared using silver stained gel images with PDQuest image analysis software. Peptides from these increased spots were analyzed by matrix assisted laser desorption/ionization-time of flight mass spectrometry (MALDI-TOF MS), and were identified based on the NCBI database by a peptide mass fingerprinting method.
Results
Comparative proteomic analysis showed 8 increased protein spots present in cystic FF. MS analysis and database searching revealed that the increased proteins in cystic FF were bovine mitochondrial f1-atpase (BMFA), erythroid associated factor (EAF), methionine synthase (MeS), VEGF-receptor, glyceraldehydes 3-phosphate dehydrogenase (GAPDH), heat shock protein 70 (HSP70), β-lactoglobulin (BLG) and succinate dehydrogenase Ip subunit (SD).
Conclusion
Our results suggest that these proteins are overexpressed in BOFC, and that they may play important roles in the pathogenesis of BOFC. Furthermore, these proteins in the FF could be useful biomarkers for BOFC.
==== Body
Background
BOFC is one of the most frequently diagnosed gynecological findings in cattle, and is a major reproductive problem in cows, causing infertility [1-3]. BOFC is generally defined as follicle-like structures of greater than 25 mm in diameter without a corpus luteum in both ovaries [1,4]. There have been several different hypotheses regarding the cause of BOFC, such as inherited factors, high lactation, aging, seasonal effects, nutritional condition, environmental estrogen and stress [1,2,4,5]. The primary cause of BOFC development has not been elucidated because of the variety of histological characteristics [1], various abnormal hormonal patterns [1,4], and differing therapeutic responses [1,4]. However, it is generally believed that cysts are formed in response to an "endocrine imbalance" [3,4,6]. Several hypotheses have been proposed suggesting that gonadotropin releasing hormone (GnRH) and gonadotropin are implicated in the pathogenesis of ovarian cysts. Actually, some authors attribute the disease to alterations in synthesis and release of GnRH [6,7], luteinizing hormone (LH) [7-9] or even to a receptor deficiency at the pituitary level [10,11].
2-D PAGE, originally described by O'Farrell [12] in 1975, is the method of choice for the separation of cell proteins, where proteins are separated in two sequential steps. This technique is an important proteomics tool, using which many protein spots can be visualized, resulting in a global view of a proteome's state [13]. Furthermore, recent developments in 2-D PAGE and mass spectrometry technologies have enabled quantitative analysis of differential proteomics, such as a comparison between normal and disease status, allowing identification of protein markers to characterize a specific physiological or pathological cell or tissue state [14,15] which have been used in the fields of diagnosis and biomarker identification of animal and human diseases [16,17].
As described above, research on BOFC has focused on endocrinological conditions [6,8], but there are few research reports on proteins in the FF of BOFC. Although Mortarino et al. [18] published one short research report about making a 2-D PAGE map of proteins expressed in the FF of BOFC, they used a 2-D PAGE process using FF which had not been depleted of abundant proteins such as albumin or immunoglobulin G. Therefore, it was thought that they overlooked spots of minor, but important proteins. Muranaka et al. [19] also examined protein in FF; the content of total protein in the FF of BOFC was significantly lower than that of the FF of normal follicles.
This study was designed to determine any substantially increased proteins in the FF from BOFC using a differential proteomics technique. The results help to clarify the pathology and etiology of BOCF and will contribute to the discovery of BOFC biomarkers.
Methods
Experimental design
This study consisted of the two following experimental phases; 1) examination of protein sample preparations from FF, and 2) assessment of increased proteins in the FF of BOFC by 2-D PAGE.
Experiment 1: Examination protein sample preparation
Cystic follicles were collected from dairy cows at a local slaughterhouse. FF was aspirated carefully with a 20 mL syringe, and centrifuged at 10,000 × g at 4°C for 30 minutes to eliminate cells and other insoluble materials, and stored at -30°C until processing for protein sample preparation.
Three types of FF protein sample were prepared in order to find an appropriate sample preparation method for 2-D PAGE.
Type A: Non-treatment
Type B: Deplete impurities (salts, lipids, detergent or nucleic acid)
Type C: Deplete both abundant proteins (albumin and IgG) and impurities
A detailed sample preparation method for depletion of both abundant proteins and impurities for Type C is shown in the following section on "Sample preparation". Non-depleted FF was used as Type A. Deplete impurities from FF were used as Type B. Samples depleted of both abundant proteins (albumin and IgG) and impurities were used as Type C. These samples were subjected to 2-D PAGE and silver stain gel images were compared visually.
Details of sample preparation, 2-D PAGE and silver staining are shown in the latter parts of this section.
Experiment 2: Assessment of increased proteins in the FF of BOFC
Ovaries with (n = 4) or without (n = 3) cystic follicles were used in this experiment. A follicular cyst was diagnosed when the follicle was greater than 25 mm in diameter in the absence of a functional corpus luteum in both the right and left ovaries [1]. Follicles of about 10 mm in diameter from ovaries without cystic follicles were used as large normal follicles. About 300 – 1000 μL of FF was aspirated carefully from each follicle with a syringe. After measureming progesterone and estradiol-17β in FF, FFs from cystic follicles were stored individually and those from normal follicles were pooled.
Each follicle was fixed in formalin after aspiration of FF, processed for histological examination and stained with hematoxylin and eosin using a standard method.
The FF protein samples were prepared by the Type C method (depleted of both abundant proteins (albumin and IgG) and impurities). Each sample from cystic follicles with pooled control samples was simultaneously processed for 2-D PAGE twice and gels were visualized by silver staining. The stained gels were scanned using a desktop scanner and scanned images from cystic and normal FFs were analyzed with PDQuest software version 7.1 (Bio-Rad Laboratories Inc., Hercules, USA). Expressions of protein spots on the images from cystic and normal FFs that were processed for 2-D PAGE simultaneously were compared, and a overall comparison of the results from 4 cystic follicle images was made.
Eight protein spots which were increased on cystic FF were excised from the silver stained gels and subjected to mass spectrometry analysis and protein identification.
Details of each experimental process are given in the following section.
Enzyme immunoassay of progesterone and estradiol-17β in follicular fluid
The enzyme immunoassay procedure for progesterone and estradiol-17β was followed as described previously [20]. FF diluted with H2O was extracted with petroleum ether for progesterone or diethylether for estradiol-17β. The decanted ether phase was dried in a glass tube and a buffer (0.05 mol/L boric acid, 0.2% BSA) was added. These reconstituted samples were placed into wells of a microtiterplate that was coated previously with an anti-rabbit IgG antibody (ICN, Pharmaceuticals Inc., USA) followed by the addition of an anti-progesterone antibody and Horseradish peroxidase (HRP) conjugated progesterone for progesterone, and an anti-estradiol-17β antibody (Kambegawa Institute, Tokyo, Japan) and HRP conjugated estradiol-17β (Kambegawa Inst.) for estradiol-17. After 2-h incubation plates were washed and a 3,3',5,5'-Tetramethylbenzidine solution was applied to the substrate followed by reading optical density measurement at a wavelength of 450 nm.
Protein sample preparation
Initially, the total protein content in FF was determined using a commercial protein assay kit (DC Protein Assay, Bio-Rad laboratories Inc., Hercules, USA). After measuring the protein content, high abundant proteins, albumin and immunoglobulin G (IgG) were removed from FF using an Aurum serum protein mini kit (Bio-Rad laboratories, Inc., Hercules, USA) to enable the visualization of low abundant proteins. Then, proteins in depleted FF samples were concentrated by centrifugation using a 3 kDa cut off cellulose membrane filter unit (Microcon YM-3, Millipore, Bedford, USA). Finally, impurities such as salts, lipids, detergent or nucleic acid were removed from concentrated samples using a 2-D Clean-Up Kit (Amersham Biosciences, San Francisco, USA).
2-D PAGE
Fifty μg of protein sample was dissolved in a sample buffer (8 M urea, 0.5% ampholine ph3.5-10, 0.5% Triton X-100, 10 mM dithiothreitol (DTT) and Orange G). The sample buffer, including 50 μg protein, was absorbed into an immobilized pH gradient (IPG) strip (Immobiline Dry Strip, 11 cm, pH range 3–10, Amersham Biosciences, Uppsala, Sweden) overnight. The rehydrated strip was subjected to isoelectric focusing on a MultiPhor II electrophoresis chamber (Amersham Biosciences, Uppsala, Sweden) for a total of 22,651 Vh at 15°C. The focused IPG strip was equilibrated in sodium dodecyl sulphate (SDS) buffer (30% glycerol, 1.0% SDS and 6 M urea in 50 mM Tris-HCl). The first equilibration step was carried out in SDS equilibration buffer with 16 nM DTT, and the second portion of the SDS equilibration buffer contained 240 mM iodoacetamide and bromophenol blue. Both equilibration steps lasted 15 min at room temperature. The equilibrated IPG strip was placed onto an 8–18% gradient polyacrylamide gradient gel (ExcelGel SDS, Amersham Biosciences, Uppsala, Sweden) for second dimensional SDS-PAGE. The SDS-PAGE was performed at 20 mA for 30 min, 50 mA for 5 min and 50 mA for 70 min at 15°C with power supply limitation of 600 V and 30 W using a MultiPhor II electrophoresis chamber. After separation in SDS-PAGE, the proteins on the gel were fixed and visualized by silver staining using a Silver Stain Plus Kit (Bio-Rad laboratories Inc., Hercules, USA).
Mass spectrometry analysis and protein identification
The spot was excised from the silver stained gel and then digested with trypsin using a previously described method [21]. For in-gel digestion, the pieces of gel were immersed in 50% acetonitrile solution for 10 min and this process was repeated several times. After washing with acetonitrile, acetonitrile was removed and the gel pieces were dried in a vacuum centrifuge for 10 min. The gel pieces were shaken with 100 μL of reducing solution (10 mM DTT and 25 mM ammonium bicarbonate) for 60 min at 56°C. Subsequently, the gel pieces were shaken under dark conditions with 100μL of an alkylating solution (55 mM iodoacetamide and 25 mM ammonium bicarbonate) for 45 min. The gel pieces were shaken with 50% acetonitrile and 25 mM ammonium bicarbonate for 10 min twice. The solvent was removed, and the gel pieces were dried again. The gel pieces were then immersed in a digestion solution containing 12.5 μg/mL sequencing grade modified trypsin and 25 mM ammonium bicarbonate and incubated on ice for 10 min. After removing unabsorbed solution, the gel pieces were incubated for 16 h at 37°C. Digested peptides were extracted twice from gel pieces with 50 μL of 50% acetonitrile and 0.1% trifluoroacetic acid. The pooled supernatants were dried in a Speed Vac, and the peptides were dissolved in 10 uL of 50% acetonitrile with0.1% trifluoroacetic acid. MALDI-TOF MS was performed on an Ultraflex time-of-flight instrument (Bruker Daltonics, Billerica, USA) equipped with a nitrogen laser operating at 337 nm. All MALDI-TOF results were obtained in the reflector-positive mode using α-cyano-4-hydroxycinnamic acid (saturated solution in 50 % acetonitrile with 0.1% trifluoroacetic acid) as the matrix. Analytes were prepared by mixing 0.5 μL of peptide sample with 0.5 μL of matrix soln. on a MALDI plate and allowed to air dry at room temperature in a hood before inserting then into the spectrometer. Mass spectra were calibrated with angiotensin II (1046.54 Da), angiotensin I (1296.68 Da), substance P (1347.74 Da), bombesin (1619.82 Da), adrenocorticotropic hormone (ACTH18-39) (2465.20 Da), and insulin (5733.54 Da). All masses are reported as monoisotopic mass values. In addition, peptides derived from trypsin (843.02 and 2212.44 Da) were used for validation when clearly observed. Peptides were identified with the Mascot search program (, Matrix Science, London, UK) against the NCBI database to perform theoretical trypsin digests and search for potentially unmodified tryptic peptides (with up to one missed cleavage) or suspected modified species. Methionine residues were considered as either normal Met or their oxidized form (Met-ox), and cysteine residues were considered to be carbamidomethylated (C-cam) or reacted with acrylamide to produce Cys-S-b-propionamide (C-pam).
Statistical analysis
Differences in the concentration of estradiol-17β and progesterone in follicular fluid, and the ratio of those hormone concentrations between normal and cystic follicles were analyzed by analysis of variance, followed by student's t-test. Differences were considered significant at P < 0.05.
Results
Examination of protein sample preparation
2-D PAGE images of type A (non treatment), type B (depleted of impurities) and type C (depleted of both abundant proteins and impurities) are shown in Fig. 1. Only large streaks and some spots were seen on the gel image of type A. After removing impurities, such as salts, lipids and nuclei acid, the large streaks disappeared, but there were about 15 spots, including a large albumin spot, on the gel image of type B. By removing both abundant proteins and impurities, the image was improved, and about 40 spots were visualized on the gel image of type C.
Figure 1 Silver stained 2-D PAGE images of proteins from bovine ovarian follicular fluid. Sample type A; Non treatment, Sample B; Depleted of impurities (salt, lipids, detergent and nuclei acid), Sample C; Depleted of abundant proteins (albumin and immunoglobulin G) and impurities.
Assessment of increased proteins in the FF of BOFC
On histological examination, ovarian follicles from the ovaries without cysts used as normal controls in this study (Panel A in Figure 2) were intact and were assessed to be large normal healthy follicles. On the other hand, granulosa cell layers were exfoliated and the theca internal of cystic follicles were thinner than those of normal follicles (Panel B in Figure 2), and were assessed as late stage cystic follicles.
Figure 2 Light micrographs of bovine ovarian follicles stained with hematoxylin and eosin. [A] Normal follicle, approximately 10 mm in diameter with granulosa. [B] Cystic follicle, approximately 30 mm in diameter with theca interna, but without granulosa. G: granulosa layer, TI: theca interna, TE: theca externa. Bar = 50 μm...
The ratios of estradiol-17β/progesterone for three of four cystic follicles were below 1, and those of all normal follicles were above 1 (Table 1).
Table 1 Concentrations of estradiol-17β and progesterone, and protein content in the follicular fluid of cystic and normal follicles
Concentration (ng/mL) of Estradiol-17β/progesterone Protein content (mg/mL)
Estradiol-17β Progesterone
Normal (n = 3) 150.3 ± 115.8 57.5 ± 57.9 3.09 ± 2.85 87.6a
Cystic (n = 4) 59.4 ± 108.5 284.7 ± 181.8* 0.80 ± 1.55 83.1 ± 9.7
All data are expressed as mean ± standard error except for protein contents of normal follicles
a Protein content was measured using a pooled sample of three normal follicles.
* Value of cystic FF is significantly different from value of normal FF (p < 0.05)
FFs from normal follicles were generally palish yellow to yellow and those from cystic follicles were dark yellow and fulvescent. This suggested interfusion of blood into the FF of cystic follicles. The total protein contents in FFs are shown in Table 1. Although the number of samples was limited, there was no particular tendency in the contents of proteins, irrespective of cysts.
We compared the differential expression of proteins by 2-D PAGE of FFs from cystic and normal follicles. Silver staining detected about 30–40 protein spots on images from both normal and cystic follicles (Panel A and Bs in Figure 3, respectively). There were only large spots on the image of normal follicles, while small spots of minor proteins were not detected on panel A of Figure 3. On the other hand, there were about 30 spots on the image of cystic follicles, and these included about 10 minor proteins spots on panels B1 and B2 as shown in Figure 3. Most of these spots were clumped in a square area on panels B1 and B2. An image comparison using PDQuest showed 8 increased protein spots were stably present on the all cystic FF gels (Panel C in Figure 3). All 8 spots were digested and forwarded to MALDI-TOF MS analysis for protein identification. Database searching with peptide mass fingerprinting by MALDI-TOF-MS/MS analysis identified the protein from spots No. 1 to 8 as BMFA, EAF, MeS, VEGF-receptor, GAPDH, HSP70, BLG and SD (Table 2), respectively. Although the molecular weight and isoelectric point of HSP70 calculated from the reported primary structure differed from those extrapolated by the position of the spot on the gel, it was confirmed by MS/MS amino acid sequence analysis that the protein from spot No. 6 was HSP70.
Table 2 List of identified proteins from the 8 spots in cystic follicles based on the NCBI database using a peptide mass fingerprinting method.
Spota Identified protein Nominal mass (Mr) Calculated pI value Number of matched peptides Sequence coverage (%)b
1 Bovine mitochondrial f1-atpase 13558 10.09 3 17
2 Erythroid associated factor 10711 4.82 2 26
3 Methionine synthase 14618 6.1 3 35
4 VEGF-receptor (Fragment) 17481 4.52 3 34
5 Glyceraldehydes 3-phosphate dehydrogenase (Fragment) 16813 9.21 3 33
6 Heat-Shock Cognate 70 kd Protein 44 kd Atpase N-Terminal Mutant With Cys 17 Replaced By Lys 41580.5 7.8 4 14
7 β-lactoglobulin 18156 4.84 4 36
8 Succinate dehydrogenase Ip subunit 16176 5.87 4 33
a The numbers correspond to the spot number on the gel shown in Figure 4.
Figure 3 Representative silver stained 2-DPAGE images of proteins from normal follicles [A] and cystic follicles [B1 & B2]. Protein samples from both types of follicles were depleted of both abundant proteins (albumin and IgG) and impurities. Expressions of protein spots on the images from cystic and normal FFs which were processed for 2-D PAGE were simultaneously compared, and the results compared using each image image. Panel C is an enlarged image of square areas in panel B1, which contained some additional protein spots of cystic follicles.
Discussion
It is reported that bovine FF includes abundant protein, such as albumin and IgG [19]. Mortarino et al. [18] carried out 2-D PAGE on the FF of BOFC, which was not depleted of these abundant proteins, so they overlooked some spots of minor, but important proteins. Therefore, the effect of removing abundant protein and impurities, such as salts, lipids, detergent and nucleic acid, from FF on 2-D PAGE was examined in the present study in order to obtain satisfactory 2-D PAGE images for differential proteomics. The quality of the 2-D PAGE images was improved and the number of visualized spots increased by removing abundant proteins, such as albumin and IgG from FF, as well as other body fluids, serum [22,23], cerebrospinal fluid [24] and bile [25], and by removing impediments, such as salts, nucleic acid and lipids, for the isoelectric focusing process [26] in this study. Concomitant to the removal of albumin and IgG, there was a significant enhancement in the staining intensity of several lower abundance protein spots. This indicates that 2-D PAGE using refined FF is a useful experimental tool for research on the physiology of FF in the ovaries of farm and experimental animals.
The sample ovarian follicles used in this study were examined morphologically and endocrinologically. The diameters of cystic follicles were greater than 25 mm by macroscopic observation; the granulosa cell layers were already exfoliated and the theca internals of some cystic follicles were thinner than those of healthy follicles. These observations show that the follicles examined in this study were late stage cystic follicles [1,27-29]. Non-dominant regressing cystic follicles showed lower estradiol-17β concentrations and higher progesterone concentrations compare to normal follicles or dominant cystic follicles, so that the ratio of estradiol-17β/progesterone of non-dominant regressing cystic follicles was below 1 [30]. The ratios of estradiol-17β/progesterone of three of four cystic follicles in this study were below 1 and these values are consistent with the histological examination, i.e. that these follicles were late stage cystic follicles.
This study was designed to identify the specific proteins in the FF of BOFC in order to help clarify the pathology of BOFC using 2-D PAGE. We found 8 additional protein spots on the gel from the cystic follicles compared with the normal control, and these proteins were identified as BMFA, EAF, MeS, VEGF-receptor, GAPDH, HSP70, BLG and SD using the MOLDI-TOF MS technique.
Heat shock proteins (HSPs) are encoded by genes whose expression is substantial during stressful conditions, such as heat shock, inhibition of energy metabolism, exposure to heavy metals, oxidative stress, and inflammation [31,32]. Under these conditions, HSPs increase cell survival by protecting and disaggregating stress-labile proteins. Under no-stress conditions, HSPs have multiple housekeeping functions, such as folding and translocating newly synthesized proteins [32,33]. It was confirmed that HSP70 was clearly expressed in the granulosa cells of human ovaries in vitro [34], and was synthesized by oocytes and cumulus cells in cows [35], and by granulosa cells within follicles in rats [36]. HSPs play an important role in fertilization and early embryonic development [37-39]. Recently, the ability to inhibit apoptosis has become widely recognized as a function of HSP70, and this ability may contribute to its protective effect against cell death [39-42]. HSP70 inhibits apoptosis and this effect may be ascribed to its neutralizing interactions with several proapoptotic effectors [40]. Isobe and Yoshimura [43] reported that in the theca interna of bovine ovarian follicles, a high frequency of apoptosis was noted in the early cystic follicles, whereas this frequency decreased in late cystic follicles. They concluded that the decrease in the rate of cell apoptosis may be responsible for the delay in follicular regression, and that control of apoptosis may be essential for reducing the incidence of cystic follicles. This observation is consistent with our result that HSP70, which decreases apoptotic cell death, was increased in the FF of cystic follicles. Peter [3] emphasized that understanding the mechanisms that regulate granulosa cell growth, differentiation and apoptosis is of great clinical importance because aberrant signalling in any of these pathways is likely to be related to in cystic ovaries. HSP70 might be the key protein related to apoptosis in granulosa cells. It is thought that cows with follicular cysts are exposed to various kinds of stress such as oxidative stress [29], negative energy balance, poor liver function and low circulating insulin-like growth factor-I [44]. Although the relationship between these stresses and the increase in HSP70 in FF in this study is unclear, it is speculated that over-expression of HSP70 decreases apoptosis in the theca interna and leads to delayed regression of cystic follicles which had prevented ovulation due to hormonal imbalance.
VEGF is known as a hypoxia-inducible mitogen for endothelial cells [45], and is also referred to as a vascular permeability factor [46], affecting vasodilatation and capillary permeability and stimulating endothelial cell growth and angiogenesis in vivo. VEGF and its receptors have been found in the ovaries of numerous species including bovine, ovine, porcine and monkey ovaries, and may be important regulators of ovarian angiogenesis [47]. In humans, VEGF is implicated in the etiology of serious reproductive disorders such as polycystic ovary syndrome, and in this syndrome, elevated VEGF may interact with its receptors, such as Flk-1/KDR, in the affected ovaries, preventing granulosa cell apoptosis and the resultant follicle atresia, thus contributing to the growth and persistence of a large number of follicles [45]. Isobe et al. [48] reported that the numbers of positive vessels and areas with von Willebrand factor, an indicator of endothelial damage, were significantly reduced in the theca of follicles in association with the cyst development. This suggested that fewer degenerative changes in the vasculature occur in BOFC, which allows a consistent blood supply and may result in maintaining follicular tissue stability and continuation of FF accumulation. The increase in receptors of VEGF, which is well known as a mitogen for endothelial cells and a vascular permeability factor, in the present study may have been associated with abnormal follicular growth via a steady blood supply and accumulation of FF originating from serum in the BOFC.
BMFA is the water-soluble component of ATP synthase, and the enzyme plays a central role in energy transformation in most living organisms [49]. SD is a key enzyme involved in both the tricarboxylic acid cycle and oxidative phosphorylation as part of the mitochondrial respiratory chain [50]. MeS is an important cellular housekeeping enzyme [51] and methionine is the most limiting amino acid for growing cattle [52]. Therefore, fluctuations in amino acids or energy metabolism in cows with cystic follicles may lead to increased BMFA, SD and MeS. However, there is no specific information on the three proteins or EAF, GDP and BLG in ovarian physiology, follicular development and/or regression, and the exact relationship between cystic follicles and these proteins remains to be clarified.
The majority of proteins identified in the follicular fluid of cystic follicles in the present study were not the type of factors one would expect to find secreted into follicular fluid; they were predominantly intracellular proteins. Isobe and Yoshimura [43] reported that apoptotic sells in the granulosa layer were observed in cystic follicles, and granulosa cell layers were also exfoliated in this study. Therefore, these intracellular proteins might be passed into the FF from the degenerated granulosa layer. On the other hand, the cystic follicular fluids used in this study were dark yellow and fulvescent, suggesting interfusion of blood into the FF of cystic follicles, and EAF was increased in cystic FF. Thus, the possibility that these proteins came via blood cannot be ruled out. Further investigations regarding the source of these proteins and the mechanism by which they entered the FF are needed.
Conclusion
In conclusion, although we mainly examined late stage cysts and it may be only one scene of complicated BOFC formation, we identified 8 increased proteins in the FF of BOFC using 2D-PAGE and the MOLDI-TOF MS technique. Our results suggest that these proteins play important roles in the etiology of BOFC, and that over- expression of HSP70 may contribute to delaying follicular regression by decreasing apoptosis in the theca interna, and over- expression of VEGF-receptors may be associated with abnormal follicular growth and accumulation of FF. Furthermore, the 8 proteins found in the FF could be used as biomarkers for ovarian cysts. Further studies are needed to clarify the pathology and etiology of BOFC including the examination of the source of these proteins, their interaction, the relationship between these proteins and ovarian physiology, and FF from various stages of follicles, including atretic follicles and dominant cysts.
Authors' contributions
JM was responsible for the design, coordination of the study and experiments He performed hormonal measurements, 2D-PAGE, and protein identification, participated in the analysis of data and in drafting the manuscript. SI collaborated in the MOLDI-TOF MS analysis and protein identification. NI collaborated in the hormonal measurements and histological examinations. TT was responsible for design and coordination of the study. He analyzed the data and drafted the manuscript. All authors read and approved the final manuscript.
Acknowledgements
We are grateful to the staff of the Meat Inspection Office of Hiroshima city for supplying the bovine ovaries. We also thank Mr. Matthew Man, AstraZeneca KK, for his English proofreading of the manuscript.
==== Refs
Kesler DJ Garverick HA Ovarian cysts in dairy cattle: a review J Anim Sci 1982 55 1147 1159 6757234
Kawate N Mechanism of ovarian cyst formation in ruminants J Jpn Vet Med Assoc 2001 54 79 84
Peter AT An update on cystic ovarian degeneration in cattle Reprod Dom Anim 2004 39 1 7 10.1046/j.0936-6768.2003.00466.x
Garverick HA Ovarian follicular cysts in dairy cows J Dairy Sci 1997 80 995 1004 9178141
Farin PW Estill CT Infertility due to abnormalities of the ovaries in cattle Vet Clin North Am Food Anim Pract 1993 9 291 308 8348373
Silvia WJ Hatler TB Nugent AM Laranja da Fonseca LF Ovarian follicular cysts in dairy cows: an abnormality in folliculogenesis Domest Anim Endocrinol 2002 23 167 77 12142235 10.1016/S0739-7240(02)00154-6
Cook DL Parfet JR Smith CA Moss GE Youngquist RS Garverick HA Secretory patterns of LH and FSH during development and hypothalamic and hypophysial characteristics following development of steroid-induced ovarian follicular cysts in dairy cattle J Reprod Fertil 1991 91 19 28 1899886
Hamilton SA Garverick HA Keisler DH Xu ZZ Loos K Youngquist RS Salfen BE Characterization of ovarian follicular cysts and associated endocrine profiles in dairy cows Biol Reprod 1995 53 890 898 8547485
Yoshioka K Iwamura S Kamomae H Changes of ovarian structures, plasma LH, FSH, progesterone and estradiol-17 beta in a cow with ovarian cysts showing spontaneous recovery and relapse J Vet Med Sci 1998 60 257 60 9524954 10.1292/jvms.60.257
Odore R Re G Badino P Donn A Vigo D Biolatti B Girardi C Modifications of receptor concentrations for adrenaline, steroid hormones, prostaglandin F2alpha and gonadotropins in hypophysis and ovary of dairy cows with ovarian cysts Pharmacol Res 1999 39 297 304 10208760 10.1006/phrs.1998.0442
Elmore RG Using GnRH, HCG, and anterior pituitary extracts in bovine hormonal therapy Vet Med 1989 84 214 216
O'Farrell PH High resolution two-dimensional electrophoresis of proteins J Biol Chem 1975 250 4007 4021 236308
Gorg A Obermaier C Boguth G Harder A Scheibe B Wildgruber R Weiss W The current state of two-dimensional electrophoresis with immobilized pH gradients Electrophoresis 2000 21 1037 1053 10786879 10.1002/(SICI)1522-2683(20000401)21:6<1037::AID-ELPS1037>3.3.CO;2-M
Thongboonkerd V Proteomics in nephrology: current status and future directions Am J Nephrol 2004 24 360 378 15205555 10.1159/000079148
Blackstock WP Weir MP Proteomics: quantitative and physical mapping of cellular proteins Trends Biotechnol 1999 17 121 127 10189717 10.1016/S0167-7799(98)01245-1
Chambers G Lawrie L Cash P Murray GI Proteomics: a new approach to the study of disease J pathol 2000 192 280 288 11054709 10.1002/1096-9896(200011)192:3<280::AID-PATH748>3.0.CO;2-L
Toda T Satoh M Sugimoto M Goto M Furuichi Y Kimura N A comparative analysis of the proteins between the fibroblasts from Werner's syndrome patients and age-matched normal individuals using two-dimensional gel electrophoresis Mech Ageing Dev 1998 100 133 143 9541134 10.1016/S0047-6374(97)00131-0
Mortarino M Vigo D Maffeo G Ronchi S Two-dimensional polyacrylamide gel electrophoresis map of bovine ovarian fluid proteins Electrophoresis 1999 20 866 869 10344260 10.1002/(SICI)1522-2683(19990101)20:4/5<866::AID-ELPS866>3.0.CO;2-V
Muranaka K Moriyoshi M Nakao T Kawata K Chemical components of bovine follicular fluid and peripheral blood serum J Coll Dairying 1986 11 423 435
Isobe N Nakao T Yamashiro H Shimada M Enzyme immunoassay of progesterone in the feces from beef cattle to monitor the ovarian cycle Anim Reprod Sci 2005 87 1 10 15885436 10.1016/j.anireprosci.2004.08.006
Shevchenko A Wilm M Vorm O Mann M Mass spectrometric sequencing of proteins from silver-stained polyacrylamide gels Anal Chem 1996 68 850 858 8779443 10.1021/ac950914h
Ahmed N Barker G Oliva K Garfin D Talmadge K Georgiou H Quinn M Rice G An approach to remove albumin for the proteomic analysis of low abundance biomarkers in human serum Proteomics 2003 3 1980 1987 14625860 10.1002/pmic.200300465
Sellers TA Yates JR Review of proteomics with applications to genetic epidemiology Genet Epidemiol 2003 24 83 98 12548670 10.1002/gepi.10226
Wenner BR Lovell MA Lynn BC Proteomic analysis of human ventricular cerebrospinal fluid from neurologically normal, elderly subjects using two-dimensional LC-MS/MS J Proteome Res 2004 3 97 103 14998169 10.1021/pr034070r
Kristiansen TZ Bunkenborg J Gronborg M Molina H Thuluvath PJ Argani P Goggins MG Maitra A Pandey A A proteomic analysis of human bile Mol Cell Proteomics 2004 3 715 728 15084671 10.1074/mcp.M400015-MCP200
Hirsch J Hansen KC Burlingame AL Matthay MA Proteomics: current techniques and potential applications to lung disease Am J Physiol Lung Cell Mol Physiol 2004 287 L1 L23 15187006 10.1152/ajplung.00301.2003
Lopez-Diaz MC Bosu WTK A review and an update of cystic ovarian degeneration in ruminants Theriogenology 1992 37 1163 1183 10.1016/0093-691X(92)90173-O
Eyestone WH Ax RL A review of ovarian follicular cysts in cows, with comparisons to the condition in women, rats and rabbits Theriogenology 1984 22 109 125 16725942 10.1016/0093-691X(84)90424-2
Kamomae H Iwamura S Yoshioka K Taniguchi T Sato M Tanaka S Nagasawa S Shiono H Takahashi Y Ohashi T Motoi Y Hiramatsu M Clinico-endocrinological, biochemical and histopathological analyses on ovarian cyst in cow Tohoku J Vet Clin 1995 18 55 64
Calder MD Manikkam M Salfen BE Youngquist RS Lubahn DB Lamberson WR Garverick HA Dominant bovine ovarian follicular cysts express increased levels of messenger RNAs for luteinizing hormone receptor and 3β-hydroxysteroid dehydrogenase Δ4, Δ5 isomerase compared to normal dominant follicles Biol Reprod 2001 65 471 476 11466215
Zylicz M King FW Wawrzynow A Hsp70 interactions with the p53 tumour suppressor protein EMBO J 2001 20 4634 46348 11532927 10.1093/emboj/20.17.4634
Jolly C Morimoto RI Role of the heat shock response and molecular chaperones in oncogenesis and cell death J Natl Cancer Inst 2000 92 1564 1572 11018092 10.1093/jnci/92.19.1564
Helmbrecht K Zeise E Rensing L Chaperones in cell cycle regulation and mitogenic signal transduction: a review Cell Prolif 2000 33 341 365 11101008 10.1046/j.1365-2184.2000.00189.x
Benifla JL Sifer C Bringuier AF Blanc-Layrac G Camus E Madelenat P Feldmann G Induced apoptosis and expression of related proteins in granulosa cells from women undergoing IVF: a preliminary study Hum Reprod 2002 17 916 920 11925381 10.1093/humrep/17.4.916
Edwards JL Hansen PJ Differential responses of bovine oocytes and preimplantation embryos to heat shock Mol Reprod Dev 1997 46 138 145 9021745 10.1002/(SICI)1098-2795(199702)46:2<138::AID-MRD4>3.0.CO;2-R
Yoon SJ Choi KH Lee KA Nitric oxide-mediated inhibition of follicular apoptosis is associated with HSP70 induction and Bax suppression Mol Reprod Dev 2002 61 504 510 11891922 10.1002/mrd.10033
Al-Katanani YM Hansen PJ Induced thermotolerance in bovine two-cell embryos and the role of heat shock protein 70 in embryonic development Mol Reprod Dev 2002 62 174 180 11984827 10.1002/mrd.10122
Kawarsky SJ King WA Expression and localisation of heat shock protein 70 in cultured bovine oocytes and embryos Zygote 2001 9 39 50 11273032 10.1017/S0967199401001058
Matwee C Kamaruddin M Betts DH Basrur PK King WA The effects of antibodies to heat shock protein 70 in fertilization and embryo development Mol Hum Reprod 2001 7 829 37 11517289 10.1093/molehr/7.9.829
Garrido C Schmitt E Cande C Vahsen N Parcellier A Kroemer G HSP27 and HSP70: potentially oncogenic apoptosis inhibitors Cell Cycle 2003 2 579 584 14512773
Gurbuxani S Schmitt E Cande C Parcellier A Hammann A Daugas E Kouranti I Spahr C Pance A Kroemer G Garrido C Heat shock protein 70 binding inhibits the nuclear import of apoptosis-inducing factor Oncogene 2003 22 6669 6678 14555980 10.1038/sj.onc.1206794
Ravagnan L Gurbuxani S Susin SA Maisse C Daugas E Zamzami N Mak T Jaattela M Penninger JM Garrido C Kroemer G Heat-shock protein 70 antagonizes apoptosis-inducing factor Nat Cell Biol 2001 3 839 843 11533664 10.1038/ncb0901-839
Isobe N Yoshimura Y Localization of apoptotic cells in the cystic ovarian follicles of cows: a DNA-end labeling histochemical study Theriogenology 2000 53 897 904 10730978 10.1016/S0093-691X(00)00238-7
Zulu VC Sawamukai Y Nakada K Kida K Moriyoshi M Relationship among insulin-like growth factor-I, blood metabolites and postpartum ovarian function in dairy cows J Vet Med Sci 2002 64 879 885 12419863 10.1292/jvms.64.879
Greenaway J Connor K Pedersen HG Coomber BL LaMarre J Petrik J Vascular endothelial growth factor and its receptor, Flk-1/KDR, are cytoprotective in the extravascular compartment of the ovarian follicle Endocrinology 2004 145 2896 2905 14988387 10.1210/en.2003-1620
Plendl J Angiogenesis and vascular regression in the overy Anat Histol Embryol 2000 29 257 266 11103513 10.1046/j.1439-0264.2000.00265.x
Reynolds LP Redmer DA Expression of the angiogenic factors, basic fibroblast growth factor and vascular endothelial growth factor, in the ovary J Anim Sci 1998 76 1671 1681 9655588
Isobe N Yoshimura Y Nakao T Distribution of immunoreactive von Willebrand factor in the microvascular network of bovine cystc follicle Anim Sci J 2002 73 123 129 10.1046/j.1344-3941.2002.00018.x
Menz RI Walker JE Leslie AG Structure of bovine mitochondrial F(1)-ATPase with nucleotide bound to all three catalytic sites: implications for the mechanism of rotary catalysis Cell 2001 106 331 341 11509182 10.1016/S0092-8674(01)00452-4
Favier J Briere JJ Strompf L Amar L Filali M Jeunemaitre X Rustin P Gimenez-Roqueplo AP Hereditary Paraganglioma/Pheochromocytoma and Inherited Succinate Dehydrogenase Deficiency Horm Res 2005 63 171 179 15795514 10.1159/000084685
Chen Z Chakraborty S Banerjee R Demonstration that mammalian methionine synthases are predominantly cobalamin-loaded J Biol Chem 1995 270 19246 19249 7642596 10.1074/jbc.270.33.19246
Lambert BD Titgemeyer EC Stokka GL DeBey BM Loest CA Methionine supply to growing steers affects hepatic activities of methionine synthase and betaine-homocysteine methyltransferase, but not cystathionine synthase J Nutr 2002 132 2004 2009 12097683
| 15941490 | PMC1177988 | CC BY | 2021-01-04 16:37:13 | no | Reprod Biol Endocrinol. 2005 Jun 8; 3:23 | utf-8 | Reprod Biol Endocrinol | 2,005 | 10.1186/1477-7827-3-23 | oa_comm |
==== Front
Respir ResRespiratory Research1465-99211465-993XBioMed Central London 1465-9921-6-451591890210.1186/1465-9921-6-45ResearchAccelerated decline in lung function in smoking women with airway obstruction: SAPALDIA 2 cohort study Downs Sara H [email protected]ändli Otto [email protected] Jean-Pierre [email protected] Christian [email protected]ünzli Nino [email protected] Margaret W [email protected] Luc [email protected] Robert [email protected] Elisabeth [email protected] Martin [email protected] Roland [email protected] Jean-Marie [email protected] Philippe [email protected] Ursula [email protected] SAPALDIA team 1 Institute of Social and Preventive Medicine, University of Basle, Basle, Switzerland2 Zürcher Höhenklinik, Wald, Switzerland3 Service of Pulmonology, University Hospital Lausanne, CHUV, Switzerland4 Division of Occupational and Environmental Health, University of Southern California, USA5 Division of Pulmonary Medicine, University Hospitals of Geneva, Geneva, Switzerland6 Hôpital intercantonal de la Broye, Payerne, Switzerland7 Hirslanden Klinik Aarau, Switzerland8 Klinik Barmelweid, Aarau, Switzerland9 Centre Valaisan de Pneumologie, Montana, Switzerland2005 26 5 2005 6 1 45 45 28 1 2005 26 5 2005 Copyright © 2005 Downs et al; licensee BioMed Central Ltd.2005Downs 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 was to determine if effects from smoking on lung function measured over 11 years differ between men and women.
Methods
In a prospective population based cohort study (Swiss Study on Air Pollution and Lung Diseases in Adults) current smokers in 1991 (18 – 60 yrs) were reassessed in 2002 (n = 1792). Multiple linear regression was used to estimate effects from pack-years of cigarettes smoked to 1991 and mean packs of cigarettes smoked per day between 1991 and 2002 on change in lung volume and flows over the 11 years.
Results
In both sexes, packs smoked between assessments were related to lung function decline but pack-years smoked before 1991 were not. Mean annual decline in FEV1 was -10.4 mL(95%CI -15.3, -5.5) per pack per day between assessments in men and -13.8 mL(95%CI-19.5,-8.1) in women. Decline per pack per day between 1991 and 2002 was lower in women who smoked in 1991 but quit before 2002 compared to persistent smokers (-6.4 vs -11.6 mL, p = 0.05) but this was not seen in men (-14.3 vs -8.8 mL p = 0.49). Smoking related decline was accelerated in men and women with airway obstruction, particularly in women where decline in FEV1 was three fold higher in participants with FEV1/FVC<0.70 compared to other women (-39.4 vs -12.2 mL/yr per pack per day, p < 0.002).
Conclusion
There are differences in effects from smoking on lung function between men and women. Lung function recovers faster in women quitters than in men. Women current smokers with airway obstruction experience a greater smoking related decline in lung function than men.
==== Body
Background
Cigarette smoking is the most well known risk factor for accelerating lung function decline in adults [1]. Until recently, smoking prevalence and intensity was greater in men than in women; but there is now evidence that women are starting to smoke as much as men [1,2]. At present, there is no consensus whether women are more sensitive to effects from cigarette smoke than men [1-8]. Relatively few large studies have compared effects from smoking between men and women and results have not been consistent between studies. A greater decline in lung function per pack-year in women compared to men was reported by Chen et al but the converse was found by Xu et al [5,6]. Prescott et al reported slightly higher coefficients for decline in FEV1 per pack-year in women compared to men and Connett et al reported similar rates of decline in men and women but a swifter regain in FEV1 in women who quit smoking [4,7].
Smoke related lung damage is characterised by inflammation, airway obstruction and destruction of the lung parenchyma [9,10]. Underlying lung characteristics can vary and there is some evidence that smokers differ in their predisposition to develop predominantly emphysema or bronchitis [11,12]. Predispositions between men and women may vary because of differences in lung morphology that modify the dispersion and deposition of cigarette smoke or differences in homeostatic processes affecting the efficacy of lung clearance and recovery after smoking cessation [13-15].
The strongest effects from smoking have been consistently measured in current smokers and detrimental effects have been shown to reduce with time from cessation [1,16]. The vast majority of surveys of effects from smoking on lung function have focused on FEV1 (forced expiratory volume in one second). However, effects may be also be revealed in other measures of lung function such as FEF25 (forced expiratory flow at 25% of lung volume) that may better reflect damage to the small airways. An argument for not investigating lung flow rates has been the large inter-subject variability in the rates which prejudices against precise estimation of effects in cross-sectional analyses. In a cohort study however, the inter-participant differences become less important because each subject serves as its own control.
The Swiss Study on Air Pollution and Lung Diseases in Adults (SAPALDIA) is a prospective population based cohort study initiated in 1991 and designed to measure long term effects of air pollution and other risk factors on respiratory health [17]. In cross-sectional analyses, associations were found between levels of lung function and exposure to air pollutants as well as to cigarette smoke [18-20]. Eleven years later, participants have been re-examined using the same methodology. Lung volumes and flows as well as smoking history were measured in 1991 and 2002. The aim of the present analysis is to assess the long term effects from smoking on change in lung function and to compare effects between men and women. The sample examined is current smokers at the time of the first assessment (1991). Effects from smoking up to 1991 and between 1991 and 2002 (SAPALDIA 1 and 2) are compared for lung flows (FEF25, FEF50 and FEF75) as well as for FEV1.
Methods
SAPALDIA is a multi-centre population based prospective cohort study. The study participants were recruited from random population samples from local registries of inhabitants from eight areas of Switzerland. The eight areas were chosen to represent the variety of environmental conditions found in Switzerland concerning geography, climate, degree of urbanisation and air pollution. To meet the selection criteria, individuals had to be 18 to 60 years of age by December 1990, local residents for at least three years and Swiss nationals or foreigners with a residence permit. There were 9651 participants in SAPALDIA 1 representing 60% of the eligible sample. Participants in SAPALDIA 1 were invited for re-examination between 2001 and 2003. Methods in SAPALDIA have been described in detail [17,21]. Ethical approval for the study was given by the Regional Ethics Commission for Clinical Medicine (Swiss Academy of Medical Sciences) and each centres' regional ethics committee.
Identical protocols for spirometry that complied with American Thoracic Society recommendations, were followed in SAPALDIA 1 and 2 [22,23]. The highest values for forced vital capacity (FVC) and FEV1 of an acceptable trial were selected. Measures of expiratory flows (FEF75, FEF50 and FEF25) were taken from the flow-volume curve with the highest sum of FVC and FEV1. Participants were requested not to use beta-2-agonists or anticholinergic inhalers four hours prior to and long-acting beta agonists, oral beta-2-agonists, theophyline or oral antimuscarinic medication eight hours prior to the time of appointment of the examination.
Information about smoking and other risk factors was gathered through an interview administered questionnaire based on the European Community Respiratory Health Survey (ECRHS) questionnaire [24]. Three categories of smoking status were derived: current smoker, former smoker and never smoker. Smokers had to have smoked more than 20 packs of cigarettes or more than 360 g of tobacco. 'Current' smokers were smokers who had smoked within the month before the interview. Other participants with validated smoking histories were classified as former smokers. Cumulative cigarette smoking exposure was summarised into two variables: pack-years to SAPALDIA 1 based on responses at SAPALDIA 1 and mean packs of cigarettes smoked per day between SAPALDIA 1 and 2 (packs per day) based on responses at SAPALDIA 1 and 2. Pack-years were calculated as years of smoking multiplied by number of cigarettes per day divided by 20.
Participants were asked not to smoke in the hour before the examination and this was validated in both surveys by measuring carbon monoxide (CO) concentration in exhaled breath using a EC 50 Micro-Smokerlizer Bedfont measuring device.
Statistical analysis
The effect of smoking on mean annual change in FEV1, FEF25, FEF50 and FEF75 between surveys was analysed by multiple linear regression. The main covariates investigated were pack-years to SAPALDIA 1 and mean packs per day smoked between surveys. All analyses were confined to subjects classified as current smokers at SAPALDIA 1.
Separate models analyses were conducted for men and women. Covariates tested in the regression models, other than the variables "pack-year" and "packs per day" included study area, atopy, childhood respiratory infection before age five, maternal smoking, paternal smoking, education, current and ever exposure to dust and fumes at work, age at start of smoking, inhaler or not, pipe smoker, solely non-cigarette smoker, body mass index (BMI), weight at baseline, change in weight, change in BMI between surveys and years since quitting. Predictors in each model are shown in the footnote to table 3. Baseline lung function parameters were included in all models because we could not assume that the change in lung function would be the same for all lung volumes and because we wanted to absorb effects from smoking up to baseline (SAPALDIA 1). However, effects were also re-examined in models that did not contain baseline lung function covariates. For each continuous covariate, we tested whether a linear, quadratic or cubic polynomial best described the relation with change in lung function. Statistical tests for interaction were conducted to determine differences in effects from smoking between men and women and between subjects with FEV1/FVC greater or less than 0.70.
Fourteen men and 11 women were excluded from the regression analyses because they had been classified as current smokers at SAPALDIA 1 and claimed to be never smokers at SAPALDIA 2. Five men and three women were excluded because they claimed to be former smokers at SAPALDIA 2 but had expired CO concentrations greater than 10 ppm (median 15.5, IQR 11.5–20). There were 53 women and 99 men respectively for whom information on pack-years or packs smoked were missing and 74 participants with missing flow data.
Regression diagnostics were conducted to identify influential data. Sensitivity analyses were conducted to examine effects after eliminating points with high leverage or large residuals and to assess possible biases due to missing information about numbers of cigarettes smoked. Analyses were conducted using STATA SE version 8.0 (StataCorp, Texas, 77845 USA).
Results
Of the 3232 current smokers in 1991 at SAPALDIA 1, 1792 (55%) participants provided spirometry and smoking information from both surveys (see Table 1). Smokers who did not participate in SAPALDIA 2 had slightly worse lung function at baseline. Male smokers were more likely than women to show evidence of airway obstruction (FEV1/FVC ratio<0.70) than female smokers and had accumulated more pack-years. The period of follow-up was the same in men and women (median 10.9 years, IQR 10.8–11.0).
Table 1 Characteristics of current smokers in 1991 (SAPALDIA 1)
Participants in SAPALDIA 1 and 2* Participants only in SAPALDIA 1 †
Characteristic measured in 1991 Men (n = 978) Women (n = 814) Men (n = 844) Women (n = 596)
Median age [yrs] (IQR) 41.8 (32.1, 49.6) 38.4 (31.0, 46.6) 40.2 (30.7, 49.2) 39.8 (30.8, 47.3)
Mean height [cm] (SD) 175.2 (6.8) 163.2 (6.6) 174.4 (7.2) 163.4 (6.5)
Mean weight [kg] (SD) 76.7 (11.1) 59.6 (9.5) 75.4 (11.7) 61.8 (13.8)
% atopic 23.2 18.4 22.6 22.5
% severe respiratory infection as an infant 6.2 9.5 5.3 6.7
% basic schooling only 13.8 17.7 22.1 22.1
% exposed to dust & fumes at work 45.5 25.1 48.2 25.2
% mother smoked 13.4 18.3 14.4 22.4
% father smoked 61.3 58.5 63.1 59.9
Median cig/day (IQR) 20 (14, 30) 20 (10, 20) 20 (15, 30) 20 (10, 23)
Median pack-years (IQR) 20.6 (8.3, 36.1) 14.4 (6.4, 25.0) 21.3 (9.1, 38.2) 15.9 (7.2, 26.7)
% inhaler 82.9 88.8 87.9 85.9
Median age started smoking [yrs] (IQR) 18 (16, 20) 18 (16, 20) 18 (16, 20) 18 (16, 20)
Median carbon monoxide [ppm] (IQR) 19 (9, 30) 16 (8, 28) 24 (13, 35) 20 (10, 30)
Mean FVC [mL] (SD) 5238 (861) 3868 (624) 5079 (874) 3745 (631)
Mean FEV1[mL] (SD) 4031 (759) 3099 (552) 3902 (837) 2957 (575)
Mean FEF25[mL/s] (SD) 1535 (793) 1427 (748) 1562 (871) 1325 (734)
Mean FEF50[mL/s] (SD) 4535 (1529) 3807 (1127) 4421 (1641) 3600 (1182)
Mean FEF75[mL/s] (SD) 7876 (2133) 5816 (1479) 7528 (2370) 5536 (1516)
% FEV1/FVC≤0.7 15.9 9.3 19.8 14.1
*With information on spirometry and smoking
† Includes participants in SAPALDIA 2 who provided information on smoking but no spirometry IQR = inter-quartile range
Approximately 30 % of men and women who were current smokers in 1991 had quit by 2002. Men had accumulated more pack-years between surveys but were more likely to have quit smoking (see Table 2). Median years since quitting amongst quitters were 5.5 (IQR 2.1 to 9.7) in men and 4.2 (IQR 2.2 to 9.2) in women. The correlation coefficient between pack-years smoked before 1991 and pack-years smoked between surveys was 0.55 in men and 0.68 in women. Mean weight change was +5.4 kg in men and +5.5 kg in women. Absolute mean declines in lung function were greater in men compared to women except for FEF25. After controlling for mean lung size or flow measured in 1991 differences in change between men and women became less significant except for FEF25. Mean percent change in FEF25 from baseline was -3.4 (5%CI -3.6, -3.2) in women and -2.8 (95%CI -3.0, -2.7) in men. The greatest inter-subject variability in change was observed for FEF75 (forced expiratory flow at 75% of lung volume).
Table 2 Smoking and lung function measured in 2002 (SAPALDIA 2)
Men n = 978 Women n = 814
% new former smokers 31.7 27.2
Median pack-years between 1991 and 2002 (IQR) 9.0 (3.8, 13.5) 7.6 (4.0, 11.0)
% inhaler 84.7 90.3
Mean annual change in:
FVC [mL] (SD) -32.6 (48.9) -20.7 (35.8)
FEV1[mL] (SD) -43.8 (36.6) -34.7 (27.2)
FEF75[mL/s] (SD) -59.1 (167.4) -34.1 (121.6)
FEF50[mL/s] (SD) -78.7 (90.9) -70.1 (71.9)
FEF25[mL/s] (SD) -46.9 (47.3) -51.8 (45.7)
Table 3 Mean annual change in lung function per pack-year to 1991 and per pack of cigarettes smoked per day between 1991 and 2002 in women and men
Term Annual change per pack-year to 1991 Annual change per packs per day between 1991 and 2002
Coef 95%CI p value Coef 95%CI p value
FEV1[mL]
Men (n = 840) Linear -0.1 -0.2, 0.1 0.50 -10.4 -15.3, -5.5 <0.001
Women (n = 735) Linear 1.3 0.7, 2.0 <0.001 -13.8 -19.5, -8.1 <0.001
Quadratic -0.04 -0.06, -0.02 <0.001 na na na
Cubic 0.0004 0.0002, 0.0005 <0.001 na na na
FEF75[mL/s]
Men (n = 798) Linear -0.6 -1.4, 0.1 0.11 -35.1 -57.5, -12.7 0.002
Women (n = 697) Linear -0.2 -1.0, 0.7 0.64 -18.6 -42.7,4.9 0.12
FEF50[mL/s]
Men (n = 825) Linear -0.2 -0.6, 0.2 0.31 -22.4 -34.5, -10.3 <0.001
Women (n = 697) Linear 0.2 -0.3, 0.8 0.31 -31.3 -46.1, -16.5 <0.001
FEF25[mL/s]
Men (n = 823) Linear -0.05 -0.1, 0.1 0.59 -8.9 -14.5, -3.4 <0.002
Women (n = 731) Linear 1.1 0.2, 1.9 0.01 -10.0 -17.3, -2.8 0.007
Quadratic -0.03 -0.6, -0.01 0.01 na na na
Cubic 0.0003 0.00005, 0.0005 0.006 na na na
na: no association with polynomial terms
Covariates other than smoking variables:
Models for Men:
FEV1 FEV1, FEV12, age, height, atopy, weight change, current exposure to dust & fumes at work, study area
FEF75 FEF75, FEF752, age, age2, atopy, maternal smoking, study area
FEF50 FEF50, age, weight, study area
FEF25 FEF25, FEF252, FEF253, age, weight, weight change, study area
Models for Women:
FEV1 FEV1, FEV12, age, age2, height, height2, atopy, weight change, study area
FEF75 FEF75, FEF752, FEF753 age, paternal smoking, study area
FEF50 FEF50, age, weight, weight change, paternal smoking, study area
FEF25 FEF25, age, age2, height, weight, weight change, study area
The adjusted effects from pack-years of cigarettes smoked to 1991 and per pack smoked per day per year between 1991 and 2002 are shown in table 3. Removal of baseline lung function variables from models made only small differences to effect estimates. Point estimates for effects from pack-years were stable in all models for both men and women after removal of influential observations.
In men, there was no association between pack-years to 1991 and change in lung function between 1991 and 2002 (see Table 3). The covariate for pack-year was retained as a linear term in all models however, because it influenced effect estimates for packs per day between 1991 and 2002. Restricting models to participants who were current smokers in both surveys made no difference to the effect estimates for pack-years to 1991.
In contrast to men, significant non-linear relations between annual change in lung function and pack-years smoked to 1991 were found in women (see Tables 3, 4 and 5). The polynomial terms were highly significant. A non-linear relation between pack-years and change in FEF25 was also seen in women that was not seen in men.
Table 4 Mean annual change in FEV1 per pack-year to 1991 and per pack of cigarettes smoked per day between 1991 and 2002 in quitters* and persistent smokers
Term Annual change in mL per pack-year to 1991 Annual change in mL per packs per day between 1991 and 2002
Coef 95%CI p value Coef 95%CI p value
Quitters†
Men (n = 245) Linear -0.1 -0.4, 0.3 0.78 -14.3 -22.6, -6.0 0.001
Women (n = 179) Linear 1.5 0.3, 2.6 0.01 -6.4 -17.6, 4.6 0.25
Quadratic -0.04 -0.08,-0.01 0.01 na na na
Cubic 0.0004 0.0001,0.0006 0.006 na na na
Persistent smokers†
Men (n = 595) Linear -0.1 -0.3, 0.12 0.41 -9.0 -15.4, -2.5 0.006
Women (n = 556) Linear 1.4 0.5, 2.3 0.001 -11.6 -19.8, -3.5 0.005
Quadratic -0.06 -0.08, -0.02 <0.001 na na na
Cubic 0.0006 0.0003, 0.0008 <0.001 na na na
*Quitters were current smokers in 1991 who had stopped smoking by 2002.
† P values for interaction between persistent smoking and quitting with packs per day: men = 0.49 women = 0.05
na: no association with polynomial terms
Effects estimated from regression models with same covariates as in Table 3
Table 5 Mean annual change in FEV1 per pack-year to 1991 and per pack of cigarettes smoked per day between 1991 and 2002 by degree of airway obstruction in 1991
Term Annual change in mL per pack-year to 1991 Annual change in mL per packs per day between 1991 and 2002
Coef 95%CI p value Coef 95%CI p value
FEV1/FVC>0.7†
Men (n = 692) Linear -0.1 -0.3, 0.1 0.48 -8.8 -14.1, -3.5 0.001
Women (n = 659) Linear 1.4 0.8, 2.0 <0.001 -12.2 -18.1, -6.4 <0.001
Quadratic -0.04 -0.06, -0.03 <0.001 na na na
Cubic 0.0004 0.0002, 0.0005 <0.001 na na na
FEV1/FVC<0.7†
Men (n = 127) Linear -0.2 -0.6, 0.2 0.36 -12.9 -25.0, -0.7 0.04
Women (n = 67) Linear -0.9 -2.2, 0.5 0.21 -39.4 -69.1, -9.6 0.01
Quadratic 0.02 0.01, 0.04 0.009 na na na
† P values for interaction between high and reduced FEV1/FVC with packs per day: men = 0.07 women = 0.002
na: no association with polynomial terms
Effects estimated from regression models with same covariates as in Table 3
Lung function decline was strongly associated with packs of cigarettes smoked per day between 1991 and 2002 in both men and women (see Table 3). Point estimates were of a similar magnitude in both sexes with an additional mean annual decline in FEV1 per pack of cigarettes smoked per day of -10.4 mL in men and -13.8 mL in women. Similar patterns to those seen for FEV1, of stronger effects from recent smoking than from past, were found in the change in flows and especially for FEF25 (forced expiratory flow at 25% of lung volume).
Whereas the decline in FEV1 per pack per day smoked between surveys was smaller in women quitters compared to women continuing smokers; there was no difference in effect estimates for men (see Table 4) (p = 0.12 for difference in effect of pack-years in quitters between men and women). Inclusion of a variable for years since quitting into the models for quitters reduced effect estimates for change in FEV1 per pack per day to -13.4 mL/yr (95%CI -1.9 to 2.2) in men and -0.7 mL/yr (-0.6 to 1.9) in women thereby magnifying the difference in effects between men and women quitters (p = 0.07). Difference in effect estimates between women persistent smokers and quitters was also increased by adjusting for years since quitting (p = 0.02).
Decline in lung function per pack per day between surveys was greater in men and women with a reduced FEV1/FVC at baseline compared to other participants (see Table 5 and Figure 1). The difference in effect per packs per day on annual change in FEV1 between women with and without FEV1/FVC <0.70 was three fold and highly significant (p < 0.002). A smaller difference in the effect from cigarette smoking between men with and without FEV1/FVC <0.70 was seen (p = 0.07). The p value for the difference in effects between men and women with a reduced FEV1/FVC was 0.05.
Figure 1 Relation between annual decline in FEV1 and mean packs/day smoked between 1991 and 2002 from regression models reported in Table 3
Discussion
Using quantitative information on cigarettes smoked during two time periods, we show similarities and differences in the effect of smoking on lung function decline between men and women. Decline due to smoking is strongly related to recent exposure and to a similar magnitude in men and women overall. However, we found evidence that the facility for recovery from past smoking is greater in women quitters. In both men and women, effects from recent smoking were more detrimental to lung function in individuals with pre-existing airway obstruction. However, the strongest effects from smoking were seen in the additional decline in FEV1 in women with a reduced FEV1/FVC ratio.
The mean unadjusted changes in lung function over the 11-year follow-up of the SAPALDIA cohort study are broadly consistent with changes reported elsewhere [6,25,26]. Average annual declines in FEV1 in SAPALDIA were higher than those for smokers in the ECRHS although lower than those reported in the Lung Health Study. Mean annual declines in FEV1 were -35 mL in men and -27 mL in women persistent smokers in ECRHS but participants were on average younger than the SAPALDIA participants (mean age 34 years versus 40 at study entry). Mean annual declines of -66 ml and -54 mL in FEV1 were reported for men and women continuing smokers respectively in the Lung Health Study (LHS). However, the LHS participants were older (35–60 years versus 18–60 years), had chronic obstructive airways disease and also reported heavier smoking (mean of 30 cigarettes per day versus 20 in SAPALDIA) at study entry [26]. In the Six Cities study, effect estimates for mean annual change in FEV1 per pack smoked in men (-12.6 ml per pack per day, 95%CI -9.7, -15.5) were similar to the effects we found in men but estimates in women (-7.2 ml per pack per day, 95%CI -9.5, -4.8) were lower than ours. However, the effects of smoking prior to follow-up and smoking during the follow-up period (3–6 years) were summarized into one variable of packs smoked and were not examined separately in the Six Cities study [6].
In both men and women, mean packs smoked per day between 1991 and 2002 were linearly associated with decline in lung function over the time period but pack-years to 1991 were not. The point estimates for annual change in lung function per pack per day in the most recent 11 years were similar in men and women with no evidence of airway obstruction. There was some evidence that men and women with poor lung function and who smoked heavily were less likely to participate in the second survey and this may be part of the explanation why we did not measure a negative effect from earlier pack-years.
Pack-years of cigarettes smoked before the 1991 were associated with non-linear changes in FEV1 and FEF25 between surveys in women. The polynomial terms were highly significant (p < 0.001) implying multiple testing is unlikely to be an explanation. Possibly women, as a group, have a more heterogeneous response to smoking than men. We found evidence that suggested that women both recover more swiftly after quitting smoking than men but also that women with airway obstruction are more vulnerable to effects from smoking.
The distribution of years since quitting amongst men and women quitters was similar and relatively short. Previous studies have also reported attenuated effects from smoking after short periods of cessation [7,27,28]. In the LHS, the attenuation in lung function decline following smoking cessation could be observed after one year [7]. The significance of the difference in effect from packs smoked per day between women quitters and persistent smokers was marginal (p = 0.05), but increased with adjustment for years since quitting; consistent with a recovery effect related to quitting. We did not observe an attenuated effect on lung function decline following smoking cessation in men, which has been reported elsewhere [7,29]. However, the number of quitters in our study was relatively small and in the Tucson Epidemiological Study of Airways Obstructive Disease, recovery in men was restricted to young men with no evidence of airway obstruction [29]. Our data provide additional evidence that facility for recovery following smoking cessation is, overall, greater in women compared to men.
Reliability of reporting habits may differ between sexes and we are unable to assess the possible extent of this bias in our study. However, we attempted to validate smoking status using expired CO and similar proportions of men and women were shown to have reported of inconsistent information about smoking habits. In the ECRHS women were as likely as men to report cough and phlegm but significantly less likely to have airflow obstruction [30]. If perception of decline in lung function is greater in women than in men, susceptible women may be more likely to quit smoking than susceptible men.
Evidence for a larger effect from smoking in men with symptoms of obstructive airway disease was presented by Fletcher and Peto almost 30 years ago [31]. Our data showed that both women and men with a reduced FEV1/FVC experienced an accelerated decline in FEV1 associated with recent cigarette smoking but the effect was significantly greater in women. We are unaware of other studies showing an accelerated effect in lung function decline related to smoking in women with obstructed airways compared to men. However, the female predominance in the Boston Early-Onset COPD Study has been attributed to a higher risk for the development of severe COPD in women compared to men [32]. Men and women persistent smokers in the LHS study have been reported to be approximately equivalent in terms of their percent predicted lung function loss [7]. However, since all LHS participants had evidence of mild to moderate airway obstruction at study entry, we would predict a sex difference in lung function decline per pack of cigarette smoked.
Reduced flow rates are due to a combination of airway narrowing and decreased lung recoil [9]. Predisposition to airway narrowing and decreased lung recoil may vary between men and women given sex differences in lung characteristics [15]. In addition, experimental evidence suggests that the distribution of particle deposition in the airways is likely to be more proximal in women compared to men [14]. Since airway caliber is smaller in women, we could hypothesise that the same reduction in airway diameter would result in a relatively greater impact on the reduction in flow rates in women compared to men [15,33].
An accelerated rate of lung function decline has also been reported amongst asthmatics who smoke [34]. Excluding women with bronchial hyperresponsiveness from our population sample did not reduce the estimates for the effect of packs/day on decline in FEV1 (data not shown). Therefore, although asthmatics may well have experienced an accelerated lung function decline it seems unlikely that asthma per se is an explanation for the different rates of decline between men and women with a reduced FEV1/FVC. Our findings suggest that smokers with pre-existing airway obstruction are likely to experience accelerated lung function decline irrespective of the potential underlying disease.
Conclusion
Some of the inconsistent findings from earlier studies comparing effects from smoking on lung function between men and women may be due to difficulties in separating out effects from recent smoking and past smoking. In both men and women recent smoking is a much stronger predictor of lung function decline than smoking more than 10 years previously. Our findings suggest that women recover faster from past smoking than men, but are particularly susceptible to effects from current smoking when they have existing airways obstruction. This observational study over 11 years suggests that women should have even greater incentives, in terms of lung function, to quit smoking than men.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
SD conducted the analyses and drafted the article. OB, JPZ, CS, NK, MG, LB, RB, EZ, MF, RK JMT, UA and PL contributed to the design of the study, the acquisition of data and the interpretation of data. CS also advised on the conduct of the analyses. All authors contributed to the conception of the research question, made important intellectual contributions during the drafting process and have given approval for the final version.
Acknowledgements
Research support provided by the National Science Foundation of Switzerland (grant no.3265896.01), the Federal Office for Forest, Environment and Landscape, the Federal Office of Public Health, the Cantons Basel-Stadt, Basel-Land, Geneva, Zurich, Ticino, Aargau, Luzern, the Swiss Lung League and the Lung League of Ticino and Zurich. N. Künzli is supported by NIEHS P30 ES07048 and the Hastings Foundation.
The study could not have been conducted without the help of the study participants, technical and administrative support and the medical teams and field workers at the local centres and we acknowledge their indispensable contributions.
The SAPALDIA team includes: Ph. Leuenberger (p) co-dir, U. Ackermann-Liebrich (e) co-dir, J.C. Barthélémy (c), A. Bircher (a), K. Blaser (a), G. Bolognini (p), L. Bayer-Oglesby (exp),O. Brändli (p), R. Bettschart (p), M. Brutsche (p), L. Burdet (p), C. Defila (m), S.H. Downs (e/s), D. Felber Dietrich (c), M. Frey (p), J.M. Gaspoz (c), M.W. Gerbase (p), M. Imboden (g), W. Karrer (p), D. Keidel (s), R. Keller (p), P. Städele-Kessler (s), B. Knöpfli (p), B. Kuna-Dibbert (e), N. Künzli (e/exp), U. Meyer (g), A. Morabia (e), U. Neu (exp), L. Nicod (p), A.P. Perruchoud (p), M. Pons (p), N. Probst Hensch (g), E. Russi (p), C. Schindler (s), J. Schwartz (e), F. Schwarz (p), P. Straehl (exp), JM. Tschopp (p), B. Wüthrich (a), JP. Zellweger (p), E. Zemp-Stutz (e) (a) allergology, (c) cardiology, (e) epidemiology, (exp) exposure, (g) genetic & molecular biology, (p) pneumology, (s) statistics
==== Refs
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 1276 11316667
Ulrik CS Smoking and mortality in women:"smoke like a man, die (at least) like a man" Eur Respir Mon 2003 25 103 117
Marang-van de Mheen PJ Davey Smith G Hart CL Hole DJ Are women more sensitive to smoking than men? Findings from the Renfreww and Paisley study Int J Epidemiol 2001 30 787 792 11511603 10.1093/ije/30.4.787
Prescott E Bjerg AM Andersen PK Lange P Vestbo J Gender difference in smoking effects on lung function and risk of hospitalization for COPD: results from a Danish longitudinal population study Eur Respir J 1997 10 822 827 9150319
Chen Y Horne SL Dosman JA Increased susceptibility to lung dysfunction in female smokers Am Rev Respir Dis 1991 143 1224 1230 2048804
Xu X Dockery DW Ware JH Speizer FE Ferris BGJ Effects of cigarette smoking on rate of loss of pulmonary function in adults: a longitudinal assessment Am Rev Respir Dis 1992 146 1345 1348 1443894
Connett JE Murray RP Buist AS Wise RA Bailey WC Lindgren PG Owens GR Changes in smoking status affect women more than men: results of the lung health study Am J Epidemiol 2003 157 973 979 12777360 10.1093/aje/kwg083
Vollmer WM Enright PL Pedula KL Speizer F Kuller LH Kiley J Weinmann GG Race and gender differences in the effects of smoking on lung function Chest 2000 117 764 772 10713004 10.1378/chest.117.3.764
Brusasco V Pellegrino R Functional classification of COPD Eur Respir Rev 2002 12 284 286
Amin K Ekberg-Jansson A Lofdahl CG Venge P Relationship between inflammatory cells and structural changes in the lungs of asymptomatic and never smokers: a biopsy study Thorax 2003 58 135 142 12554896 10.1136/thorax.58.2.135
Cosio Piqueras MG M.G. C Disease of the airways in chronic obstructive pulmonary disease Eur Respir J Suppl 2001 34 41s 49s 12392034 10.1183/09031936.01.00234601
Clark KD Wardrobe-Wong N Elliott JJ Gill PT Tait NP Snashall PD Patterns of lung disease in a "normal" smoking population: are emphysema and airflow obstruction found together? Chest 2001 120 743 747 11555504 10.1378/chest.120.3.743
Anderson PJ Wilson JD Hiller FC Respiratory tract deposition of ultrafine particles in subjects with obstructive or restrictive lung disease Chest 1990 97 1115 1120 2331906
Kim CS Hu SC Regional deposition of inhaled particles in human lungs: comparison between men and women J Appl Physiol 1998 84 1834 1844 9609774 10.1063/1.368615
Becklake MR Kauffmann F Gender differences in airway behaviour over the human life span Thorax 1999 54 1119 1138 10567633
Scanlon PD Connett JE Waller LA Altose MD Bailey WC Buist AS Smoking cessation and lung function in mild-to-moderate chronic obstructive pulmonary disease. The Lung Health Study Am J Respir Crit Care Med 2000 161 381 390 10673175
Martin BW Ackermann-Liebrich U Leuenberger P Kunzli N Stutz EZ Keller R Zellweger JP Wuthrich B Monn C Blaser K Bolognini G Bongard JP Brandli O Braun P Defila C Domenighetti G Grize L Karrer W Keller-Wossidlo H Medici TC Peeters A Perruchoud AP Schindler C Schoeni MH Villiger B SAPALDIA: methods and participation in the cross-sectional part of the Swiss Study on Air Pollution and Lung Diseases in Adults Soz Praventivmed 1997 42 67 84 9151378 10.1007/BF01318136
Ackermann-Liebrich U Leuenberger P Schwartz J Schindler C Monn C Bolognini G Bongard JP Brandli O Domenighetti G Elsasser S Grize L Karrer W Keller R Keller-Wossidlo H Kunzli N Martin BW Medici TC Perruchoud AP Schoni MH Tschopp JM Villiger B Wuthrich B Zellweger JP Zemp E Lung function and long term exposure to air pollutants in Switzerland. Study on Air Pollution and Lung Diseases in Adults (SAPALDIA) Team Am J Respir Crit Care Med 1997 155 122 129 9001300
Schindler C Kunzli N Bongard JP Leuenberger P Karrer W Rapp R Monn C Ackermann-Liebrich U Short-term variation in air pollution and in average lung function among never-smokers. The Swiss Study on Air Pollution and Lung Diseases in Adults (SAPALDIA) Am J Respir Crit Care Med 2001 163 356 361 11179106
Leuenberger P Schwartz J Ackermann-Liebrich U Blaser K Bolognini G Bongard JP Brandli O Braun P Bron C Brutsche M Passive smoking exposure in adults and chronic respiratory symptoms (SAPALDIA Study). Swiss Study on Air Pollution and Lung Diseases in Adults, SAPALDIA Team Am J Respir Crit Care Med 1994 150 1222 1228 7952544
Ackermann-Liebrich U Kuna-Dibbert B Probst-Hensch NM Schindler C Felber Dietrich D Zemp Stutz E Bayer-Olgesby L Baum F Brändli O Brutsche M Downs SH Keidel D Gerbase MW Imboden M Keller R Knöpfli B Künzli N Nicod L Pons M Städele P Tschopp JM Zellweger JP Leuenberger P Follow-up of the Swiss Cohort study on Air Pollution and Lung Diseases in Adults (SAPALDIA 2) 1991-2003: Methods and characterisation of participants Soz Praventivmed 2005 In press
American Thoracic Society Standardization of spirometry-1987 update Am Rev Respir Dis 1987 136 1285 1298 3674589
American Thoracic Society Standardisation of Spirometry 1994 update Am J Resp Crit Care Med 1995 152 1107 1136 7663792
Burney PG Luczynska C Chinn S Jarvis D The European Community Respiratory Health Survey Eur Respir J 1994 7 954 960 8050554
Chinn S Jarvis D Melotti R Luczynska C Ackermann-Liebrich U Anto J Cerveri I de Marco R Gislasan T Heinrich J Janson C Künzli N Leynaert B Neukirch F Schouten J Sunyer J Svannes C Vermeire P Wjst M Burney P Smoking cessation, lung function, and weight gain: a follow-up study Lancet 2005 365 1629 1635 15885295 10.1016/S0140-6736(05)66511-7
Anthonisen NR Connett JE Murray RP Smoking and lung function of Lung Health Study participants after 11 years Am J Respir Crit Care Med 2002 166 675 679 12204864 10.1164/rccm.2112096
Bosse R Sparrow D Rose CL Weiss ST Longitudinal effect of age and smoking cessation on pulmonary function Am Rev Respir Dis 1981 123 378 381 7224350
Lange P Groth S Nyboe GJ Mortensen J Appleyard M Jensen G Schnohr P Effects of smoking and changes in smoking habits on the decline of FEV1 Eur Respir J 1989 2 811 816 2806504
Sherrill DL Holberg CJ Enright PL Lebowitz MD Burrows B Longitudinal analysis of the effects of smoking onset and cessation on pulmonary function Am J Respir Crit Care Med 1994 149 591 597 8118623
de Marco R Accordini S Cerveri I Corsico A Sunyer J Neukirch F Kunzli N Leynaert B Janson C Gislason T Vermeire P Svanes C Anto JM Burney P An international survey of chronic obstructive pulmonary disease in young adults according to GOLD stages Thorax 2004 59 120 125 14760151 10.1136/thorax.2003.011163
Fletcher C Peto R The natural history of chronic airflow obstruction Br Med J 1977 1 1645 1648 871704
Hersh CP DeMeo DL Al-Ansari E Carey VJ Reilly JJ Ginns LC Silverman EK Predictors of survival in severe, early onset COPD Chest 2004 126 1443 1451 15539711 10.1378/chest.126.5.1443
Mead J Dysanapsis in normal lungs assessed by the relationship between maximal flow, static recoil, and vital capacity Am Rev Respir Dis 1980 121 339 342 7362140
James AL Palmer LJ Kicic E Maxwell PS Lagan SE Ryan GF Musk AW Decline in lung function in the Busselton Health Study The Effects of Asthma and Cigarette Smoking Am J Respir Crit Care Med 2005 171 109 114 15486340 10.1164/rccm.200402-230OC
| 15918902 | PMC1177989 | CC BY | 2021-01-04 16:23:26 | no | Respir Res. 2005 May 26; 6(1):45 | utf-8 | Respir Res | 2,005 | 10.1186/1465-9921-6-45 | oa_comm |
==== Front
Respir ResRespiratory Research1465-99211465-993XBioMed Central London 1465-9921-6-491592979210.1186/1465-9921-6-49ResearchBacterial activity in cystic fibrosis lung infections Rogers Geraint B [email protected] Mary P [email protected] David J [email protected] Peter M [email protected] Valia [email protected] Graeme R [email protected] Kenneth D [email protected] Department of Life Sciences, King's College London, London, UK2 Cystic Fibrosis Unit, Southampton University Hospitals NHS Trust, UK3 Health Protection Agency, Southampton Laboratory, UK4 The Adult Cystic Fibrosis Unit, Mater Adult Hospital, Brisbane, Australia2005 1 6 2005 6 1 49 49 27 8 2004 1 6 2005 Copyright © 2005 Rogers et al; licensee BioMed Central Ltd.2005Rogers 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
Chronic lung infections are the primary cause of morbidity and mortality in Cystic Fibrosis (CF) patients. Recent molecular biological based studies have identified a surprisingly wide range of hitherto unreported bacterial species in the lungs of CF patients. The aim of this study was to determine whether the species present were active and, as such, worthy of further investigation as potential pathogens.
Methods
Terminal Restriction Fragment Length Polymorphism (T-RFLP) profiles were generated from PCR products amplified from 16S rDNA and Reverse Transcription Terminal Restriction Fragment Length Polymorphism (RT-T-RFLP) profiles, a marker of metabolic activity, were generated from PCR products amplified from 16S rRNA, both extracted from the same CF sputum sample. To test the level of activity of these bacteria, T-RFLP profiles were compared to RT-T-RFLP profiles.
Results
Samples from 17 individuals were studied. Parallel analyses identified a total of 706 individual T-RF and RT-T-RF bands in this sample set. 323 bands were detected by T-RFLP and 383 bands were detected by RT-T-RFLP (statistically significant; P ≤ 0.001). For the group as a whole, 145 bands were detected in a T-RFLP profile alone, suggesting metabolically inactive bacteria. 205 bands were detected in an RT-T-RFLP profile alone and 178 bands were detected in both, suggesting a significant degree of metabolic activity. Although Pseudomonas aeruginosa was present and active in many patients, a low occurrence of other species traditionally considered to be key CF pathogens was detected. T-RFLP profiles obtained for induced sputum samples provided by healthy individuals without CF formed a separate cluster indicating a low level of similarity to those from CF patients.
Conclusion
These results indicate that a high proportion of the bacterial species detected in the sputum from all of the CF patients in the study are active. The widespread activity of bacterial species in these samples emphasizes the potential importance of these previously unrecognized species within the CF lung.
Cystic Fibrosisbacterial infections16S rDNAT-RFLP profilingbacterial activity
==== Body
Introduction
The Microbiological analysis of clinical specimens has relied traditionally on cultivation prior to identification. However, recent genetic innovations promise to provide dramatic advances in diagnosis and fresh insights into infections that have previously been considered well characterised. The use of molecular biological approaches in clinical scenarios obviates the requirement for in vitro culture prior to analysis and so removes problems associated with microbial cultivation. In addition, these approaches are ideal for cases such as trauma where predicting the pathogen(s) responsible is challenging.
Molecular biological approaches have gained widespread acceptance for the study of bacterial communities in natural environments [reviewed in [1-3]]. Here, nucleic acids, extracted directly from samples, act as templates for PCR amplification of phylogenetically-informative ribosomal sequences using oligonucleotide primers "universal" for the Domain Bacteria. This means that no prior assumptions are made about the identity of species present. Bacterial community composition can then be assessed through sequence analysis of cloned ribosomal PCR products and by Terminal Restriction Fragment Length Polymorphism (T-RFLP) profiling [4,5].
Here, we focus on bacterial infections within the lungs of cystic fibrosis (CF) patients. There are more than 5,000 registered CF patients in the UK [6]. Although CF patient life expectancy has steadily increased, the mortality rate for patients aged between 26 and 30 years remains at around 50 per 1000 per year [7]. Mortality is primarily determined by repeated infective exacerbations. Ultimately, 80 to 95% of CF patients succumb to respiratory failure brought on by chronic bacterial infection and concomitant airway inflammation [8]. The characterisation of the bacteria present in the CF lung is critical if therapy is to be advanced. Moreover, we suggest that this strategy can benefit a diverse range of clinical scenarios.
Previous molecular biological studies have shown that the level of bacterial diversity in adult CF sputum was much higher than previously recognised [9-12] and that the communities detected were distinct. This contrasts sharply with wisdom informed by traditional screening of sputa that focuses on only a few pathogenic species including Pseudomonas aeruginosa, Staphylococcus aureus and the Burkholderia cepacia complex.
Moreover, many of the species detected were anaerobes – often obligate – from within the genera Bacteroides, Eubacterium, Fusobacterium, Porphyromonas, Prevotella, Rothia and Veillonella. This agreed with earlier studies [13,14] and may be of particular relevance given the increased recognition of the importance of anaerobic growth in CF infections [15]. In addition, many of the species identified, including Abiotrophia adiacens, Mycoplasma salivarium, Ralstonia taiwanensis, Rothia mucilaginosus and Staphylococcus hominis, had not been reported as previously isolated from CF sputum. Although the role of these species in lung pathogenesis has yet to be determined, the first step is to establish whether they are active within the CF lung. Although activity does not necessarily imply pathogenicity, it strongly suggests that further study is warranted.
Here, we assess the extent to which bacteria in sputum sampled from the CF lung were active by exploiting the difference in stability of DNA and RNA. As 16S rRNA is inherently unstable, it can be used to define intact, metabolically active bacterial cells [16] with bacterial metabolic activity inferred from the level of transcription of these sequences [17]. Reverse Transcription Terminal Restriction Fragment Length Polymorphism (RT-T-RFLP), performed in parallel with T-RFLP provides an accepted means of determining relative metabolic activity within communities [18]. T-RFLP and RT-T-RFLP analyse the same genetic sequence. The single difference is that in RT-T-RFLP, complementary DNA (cDNA) copies, generated from 16S rRNA, are used as the template instead of 16S rDNA. Here, we wished to test the hypothesis that the majority of the species identified in the CF lung are active. To do this, two profiling approaches were used to study DNA and RNA extracted directly from the same clinical sample taken from 17 CF patients. Further, the T-RFLP profiles generated from the 17 CF sputa were compared to those generated from sputa obtained from 19 healthy, non-CF individuals.
Materials and methods
Clinical samples and preparation for nucleic acid extraction
Sputum samples were collected from 17 adult CF patients attending Southampton University Hospital, with ethical approval granted by the Southampton Research Ethics Committee (067/01). 12 patients were suffering infective exacerbation at time of sampling, whilst five were considered to be stable. Ten volumes of RNAlater solution (Promega, Southampton, UK) were added to sputum samples immediately following collection following manufacturer's instructions. Prior to nucleic acid extraction, samples were prepared for nucleic acid extraction, using a series of washes and sputasol treatment as described previously [10].
DNA and RNA extraction
All reagents, glassware and plastics used in RNA work were DEPC-treated prior to use. RNA was extracted as follows: 0.75 ml of Tri Reagent (Sigma-Aldrich, Dorset, UK) were added to approximately 0.2 ml of each sample and vortexed for 1 min. Samples were incubated at room temperature for 5 min prior to the addition of 0.2 ml chloroform. Samples were vortexed for 15 sec. and incubated at room temperature for 5 min. Phases were separated by centrifugation at 12,000 × g for 15 min at 4°C.
i) DNA extraction
0.3 ml of 100% ethanol was added to precipitate the DNA from the lower phase. The sample was mixed by inversion, incubated at room temperature for 3 min and centrifuged at 12,000 × g for 5 min at 4°C. The pellet was washed in 0.1 M sodium citrate, 10% ethanol solution (during each wash the pellet was allowed to stand for at least 30 min). Pellets were centrifuged at 12,000 × g for 5 min at 4°C and washed twice in 75% ethanol. The DNA was vacuum dried, with the pellet resuspended in 100 μl H2O and stored at -20°C.
ii) RNA extraction
The upper phase was transferred to a fresh microfuge tube and 0.5 ml of propan-2-ol were added. Samples were incubated for 10 min at room temperature and RNA was pelleted by centrifugation at 12,000 × g for 10 min at 4°C. The supernatant was removed and the RNA pellet washed once in 75% ethanol and re-pelleted by centrifugation at 7,500 × g for 5 min at 4°C. Pellets were air-dried for 10 min, resuspended in 30 μl distilled water and incubated for 10 mins at 55°C. Purified RNA samples were stored as aliquots at -70°C.
Prior to reverse transcription, any residual DNA was removed using DNAseI (Epicentre, Madison, USA) in accordance with the manufacturer's instructions, with PCR amplification controls performed as appropriate.
Reverse transcription
Two universal bacterial primers were used, namely; 8f700 (5'-AGA GTT TGA TCC TGG CTC AG-3') and 920r (5'-CCG TCA ATT CAT TTG AGT TT-3') [5,10]. cDNA was generated from the isolated RNA using 920r and AMV reverse transcriptase (Promega, Southampton, UK) in accordance with the manufacturer's instructions. Double stranded DNA was generated using 1 μl of this cDNA as template in a 50 μl PCR reaction containing both primers (8f700 and 920r). PCR products amplified were verified by Tris-acetate-EDTA (TAE)-agarose gel electrophoresis on 0.8% (wt/vol) TAE-agarose gels stained in ethidium bromide (0.5 mg/L) with images, viewed on a UV transilluminator (Herolab, Wiesloch, Germany), captured by using a Herolab image analyzer with E.A.S.Y STOP win 32 software (Herolab).
DNA quantification
Extracted DNA and restricted PCR products (below) were quantified using a CytoFluor series 4000 multiwell plate reader (PerSeptive Biosystems, Foster City, USA) using the PicoGreen DS DNA quantitation kit (Molecular Probes, Lieden, Netherlands) following the manufacturer's instructions.
T-RFLP amplification and profiling
PCR products for T-RFLP analysis were amplified using primers 8f700 (labelled at the 5'end with IRD700) and 926r from c. 20 ng of extracted DNA as previously described [10]. PCR products (c. 20 ng) were digested to completion using 1 U of the restriction endonuclease CfoI, with c. 0.7 μg of T-RFLP PCR products were separated by length using a LI-COR IR2 automated DNA sequencer again as previously described [10]. The gels were analysed by using GeneimageIR v.3.56 (Scanalytics, Fairfax, USA). When profile data were assessed, only peaks of ≥ 0.1% of the total lane signal were classified as bands for further analysis. The positions of these individual bands were calculated in relation to microSTEP 15 a (700-nm) size marker (Microzone, Lewes, UK). A threshold of ×2 was used as a means of identifying marked differences between T-RFLP and RT-T-RFLP band volumes.
Band Quantification and calling
The IR signal level produced by each band was determined using Phoretix 1D Advanced v.5.10, (Nonlinear Dynamics, Newcastle upon Tyne, UK). Due to the low volumes loaded, small errors could have lead to significant variations in band intensity. To avoid such errors, band volumes were determined as a percentage of the total band volume detected in each profile.
Sputum from non-CF individuals
Sputum production was induced in 19 healthy, non-CF individuals by the inhalation of nebulised saline for 5 min. Following nebulisation, subjects were asked to rinse their mouths thoroughly with water and blow their noses. Expectorated sputum was then collected and subjected to the same DNA extraction, amplification and profiling protocols employed for CF sputum analysis. Individuals were selected at random to provide sputum samples. Individuals that reported either acute or chronic respiratory problems were excluded.
in silico sequence analysis
Published bacterial 16S rRNA gene sequence data, stored at GenBank , were retrieved. MapSort (Wisconsin Package version 10.3; Accelrys) was used to predict the band sizes for T-RFLP analysis. Mapsort, which locates the position of restriction endonuclease recognition motifs in a given sequence, was used to determine the length (in bases) from the 5' end of primer 8f-700IR (see below) to the first cleavage position of the restriction endonuclease CfoI in each 16S rRNA gene sequence. This process was performed on all of the bacterial entries in the Genbank database that spanned the amplified region. In this way, it was possible to predict the length of T-RF bands generated from 853 separate phylotypes (data not shown).
Statistical analysis
For each of the bands that were detected in a T-RFLP or RT-T-RFLP profile unaccompanied by a band in the corresponding RT-T-RFLP or T-RFLP profile, the frequency of unaccompanied detection was compared with the frequency of accompanied detection in the sample set as a whole. The ratio of unaccompanied detection to total detection was multiplied by unaccompanied detection. This was performed because relatively frequently detected bands, which were unaccompanied in a majority of instances, would otherwise appear less significant. A score ≥ 2 was used as a threshold for the identification of bands that differed notably between their detection by the two techniques.
Hierarchical cluster analysis, with the Dice measure and Chi-square test using Yates correction (SPSS for Windows v.10.1, SPSS Inc., Chicago, USA) was used to construct a dendrogram representing level of similarity between the 34 bacterial community T-RFLP and RT-T-RFLP profiles studied here. Further, this process was used to compare the similarity of T-RFLP profiles generated from CF sputa with those generated from healthy, non-CF sputum.
Role of the funding source
The sources of funding of this study were not involved in experimental design and interpretation. No influence was exerted on the decision to publish.
Results
Electrophoretic gel images generated by T-RFLP and RT-T-RFLP profiling from five sputum samples are shown in Figure 1. An example of the identification of individual bands within a region of electrophoretic profile is shown for corresponding areas of a T-RFLP and RT-T-RFLP profiles in Figure 2.
Figure 1 Electrophoretic gel images generated by T-RFLP and RT-T-RFLP. This figure shows the profiles generated from five sputum samples within the sample set. By a process of automated comparison of band positions with those in marker lanes allows their length to be determined and direct comparisons to be made between lanes.
Figure 2 Identification of individual bands within regions of corresponding T-RFLP and RT-T-RFLP profiles. This figure shows regions of profiles as analysed using Phoretix 1D Advanced v.5.10 (Nonlinear Dynamics, Newcastle upon Tyne, UK). In each case, the region of electrophoretic profile is shown (below) next to a trace of relative band intensity. The manual confirmation of correct band identification minimises the inclusion of erroneous peaks.
T-RFLP and RT-T-RFLP profiling
Parallel T-RFLP and RT-T-RFLP analysis was performed on 17 sputum samples. A total of 706 individual T-RF and RT-T-RF bands were detected in this sample set. Of these, 323 were detected by T-RFLP analysis and 383 were detected by RT-T-RFLP analysis. This difference in the number of bands detected was significant (P ≤ 0.001, Chi-square test, Yates correction). The number of bands detected in profiles generated from each individual sample is shown in Table 1.
Table 1 Number of bands detected in T-RFLP and RT-T-RFLP profiles generated from the sample set.
Patient T-RFLP bands RT-T-RFLP bands
1 42 33
2 35 38
3 14 22
4 25 27
5 38 18
6 26 15
7 16 21
8 13 15
9 14 38
10 10 11
11 7 18
12 15 15
13 12 11
14 17 29
15 16 20
16 10 38
17 13 14
Total 323 383
Average 19.0 (± 10.4) 22.5 (± 9.5)
The number of T-RF bands detected above a threshold of 0.1% of the total lane signal volume is shown for both the T-RFLP and RT-T-RFLP profiles generated from each of the 17 samples. Standard deviations for average values are shown in brackets.
The banding positions generated through T-RFLP and RT-T-RFLP analysis were compared for each sample. 178 bands were detected in both a T-RFLP profile and the corresponding RT-T-RFLP profile (356 bands in total), 145 bands were detected in a T-RFLP profile but were absent in the corresponding RT-T-RFLP profile, and 205 bands were detected in an RT-T-RFLP profile but were absent in the corresponding T-RFLP profile.
Where a band of a given length was detected in profiles generated in a sample, it was present in both T-RFLP and RT-T-RFLP profiles in 33.7% of instances. In 27.4% of instances in was detected in the T-RFLP profile alone and in 38.8% of instances it was detected in the RT-T-RFLP profile alone.
The ratio of unaccompanied bands to total bands detected was multiplied by the number of unaccompanied bands detected. This process identified 7 band lengths with a score of ≥ 2 in T-RFLP profiles, compared with 25 band lengths in RT-T-RFLP profiles. The only band from either group whose length corresponded to that of a recognised CF pathogen was 209 bases (B. cepacia complex) had a score of 2.0 in RT-T-RFLP profiles. No pattern was discerned between any of the other bacterial species whose predicted T-RF length corresponded with these bands. No band length had a score ≥ 2 in both T-RFLP and RT-T-RFLP profiles.
The intensity of each of the T-RF bands detected was determined and placed in rank order (descending band volume) (Table 2, see additional file). The five highest rank ordered bands represented 39.5% (± 21.1), 14.4% (± 6.2), 8.6% (± 2.6), 5.8% (± 1.6), and 4.5% (± 2.0) of the total band signal in T-RFLP profiles respectively. For RT-T-RFLP profiles, the top five rank ordered T-RF bands represented 35.2% (± 19.4), 14.4% (± 6.9), 9.2% (± 3.3), 7.0% (± 2.9) and 4.7% (± 1.7), respectively.
Fifty five of the T-RF bands lengths detected by T-RFLP profiling were in the five most intense bands in one or more profile, compared with 53 T-RF band lengths in RT-T-RFLP profiling. T-RF bands of a given length were detected within the top five rank ordered positions of intensity in an average of 1.5 T-RFLP profiles and 1.6 RT-T-RFLP profiles. Of the T-RF band lengths identified in this way, 26 were detected in the top five rank positions in T-RFLP profiles but not RT-T-RFLP profiles, whereas 25 were detected in the top five rank positions in RT-T-RFLP profiles but not in T-RFLP profiles.
In both cases, a T-RF band of 155 bases, corresponding to that produced by P. aeruginosa, was the most frequently detected band in the top five rank ordered positions, being detected in 8.2% and 9.4 % of T-RFLP and RT-T-RFLP profiles respectively. The second most commonly detected T-RF band length was also the same in both profiling approaches – a T-RF band of 78 bases in length was detected in one of top five rank ordered positions in 5.9% and 4.7% of T-RFLP and RT-T-RFLP profiles respectively. Computer-based band length predictions made using published sequence data indicate that a 78 base T-RF band would be consistent with that produced by Ectothiorhodospira mobilis, Methylobacter psychrophilus, Methylomicrobium agile, Methylomonas rfodinarum and Methylomonas rubra.
The prevalence of other bacterial species that have been considered traditionally to be key CF pathogens was determined. In general, it was found that these were not highly represented in either the T-RFLP or RT-T-RFLP profiles. A band corresponding to B. cepacia complex was detected in one T-RFLP profile (patient 7), and one RT-T-RFLP profile (patient 2). These bands represented 2.2% and 3.3% of total band volume respectively. A band corresponding to H. influenzae, representing 1.7% of the total band volume, was detected in a single T-RFLP profile (patient 15). A band consistent with that produced by S. maltophilia was detected in the T-RFLP profiles generated from two patients (patients 1 and 2), representing 1.6% and 5.5% of the total band volume respectively. A band was also detected in the RT-T-RFLP profile from a third sample (patient 5), where it represented 18% of the total band volume. No band of a length corresponding to S. aureus was detected.
Hierarchical cluster analysis
A dendrogram was derived by hierarchical cluster analysis using the Dice measure for all of the T-RFLP and RT-T-RFLP generated profiles (Figure 3). This showed that for the majority (13 of 17) patients studied, the T-RFLP and RT-T-RFLP profiles from the same sample clustered more closely than any other profile. However, profiles were observed in two instances that more closely matched those of other individuals (patients 3 and 9 and patients 1 and 16, Figure 3).
Figure 3 Dendrogram constructed using T-RFLP and RT-T-RFLP profiles generated from the sample set. A dendrogram was constructed using the results of Hierarchical Cluster Analysis (HCA), using Dice measure, of the T-RFLP and RT-T-RFLP profile data. HCA results in the formation of clusters in which profiles are iteratively joined in a descending order of similarity.
Healthy, non-CF samples
Between 18 and 54 individual T-RF bands were detected in the T-RFLP profiles generated from the healthy, non CF sputa. On average 29 (± 10) T-RF bands were detected in each profile.
A total of 556 individual T-RF bands were resolved in the 19 samples analysed. These represented 210 different T-RF band lengths. T-RF bands of a given length were detected in between 1 and 18 of the 19 samples, being detected in 2.65 samples (± 3.67) on average. Of these 210 T-RF band lengths, 81 were detected in both the CF and non CF sputum profiles, 129 were detected in the healthy sample set only, and 114 were detected in the CF only.
Profiles generated from healthy sputum were more similar than profiles generated from CF sputum, with a greater level of overlap between the T-RF bands lengths detected in the profiles generated. In the CF sample set, no T-RF band length was detected in more than 41.1 % of profiles, with 39% of T-RF band lengths detected in a single profile only, and 22.2% of T-RF band lengths detected in two profiles, only. By comparison, only 21.9% and 12.2% of T-RF band lengths were detected in one or two healthy sputum profiles respectively. Further, 14 different T-RF band lengths were detected in 50% or more of healthy profiles and 4 were detected in more that 75% of healthy profiles.
Of the 14 T-RF band lengths detected in 50% or more of healthy sample profiles, 5 were not detected in the CF sample set at all, 3 were detected in a single sample only, two were detected in two profiles, two in three profiles, one in four profiles and one in 6 profiles.
None of the profiles generated from healthy sputa were found to contain T-RF bands of lengths corresponding to the recognised CF pathogens P. aeruginosa, B. cepacia complex, S. aureus, or H. influenzae. A T-RF band of 214 bases was resolved in 5 of the 19 healthy profiles. This is consistent with the T-RF band produced by both five different species (Stenotrophomonas maltophilia, Fusobacterium gonidoformans, Aeromonas hydrophila, Shewanella alga, Vibrio wodanis), of which Stenotrophomonas maltophilia is a recognised CF pathogen.
Of the 209 band lengths detected in the RT-T-RFLP profiles, 118 (56.4%) were not detected in the profiles generated from the healthy sputum sample set whatsoever.
Hierarchical cluster analysis, using Dice similarity measure, was performed on the T-RFLP profiles generated from CF and non-CF samples. The dendrogram that was generated is shown in Figure 4. It was found that there was complete separation of cluster groupings between CF and non-CF samples.
Figure 4 Dendrogram constructed using T-RFLP profiles generated from sputum samples obtained from CF patients and healthy individuals. A dendrogram was constructed using the results of Hierarchical Cluster Analysis (HCA), using Dice measure, of the T-RFLP profile data.
Discussion
This study addresses important questions about the activity of bacteria in infections. Recently, it has been shown that many bacterial species not previously associated with CF lung infections, could be detected when molecular biological approaches were applied to the study of sputa [10,19]. These studies also showed that many species in CF sputum were facultative or obligate anaerobes. This study shows that the majority of these species were metabolically active. Potentially, this has important implications for treatment and it is now critical to determine what impact these species have on lung disease and to identify their clinical significance.
The application to clinical studies of the combined approaches used here is novel. These approaches are robust and have previously been shown to be highly reproducible [data not shown, [10,12]]. Here it was found that, on average, each patient had more than 22 metabolically active bacterial species per sputum sample. This represented a statistically significant difference over species number identified through T-RFLP alone (P ≤ 0.001). This finding makes it important to determine the role of metabolically active bacteria in lung pathogenesis. Some species may emerge as frank pathogens. However, even active bacterial species that might be considered "avirulent" may play an important indirect role; for example, it has been demonstrated that avirulent oropharyngeal flora can cause an upregulation of virulence genes and consequentially pathogenicity of P. aeruginosa [20].
In certain cases, the T-RFLP profile and RT-T-RFLP profile were found by visual comparison to be similar. To provide a more robust analysis, Hierarchical Cluster Analysis (HCA) was used. HCA demonstrated that there was typically greater similarity between profiles generated from individual samples than any other individual. This implies that each patient has an individual collection of typically active bacteria. The lack of HCA clustering according to technique – i.e. discrete groups of T-RFLP profiles and RT-T-RFLP profiles – suggests that the same groups of species are not either "present" or "active" in different individuals. This reinforces the requirement for management to be highly specific for each individual patient, a concept that up to now has been put into practice based on clinical experience without particular scientific backing. Moreover, it may explain the differential response of patients, at apparently similar stages of CF lung disease, to antibiotic regimes.
Marked differences in T-RFLP and RT-T-RFLP profiles were however observed in many cases. For example, in 39.6% of banding positions, a signal was detected in the RT-T-RFLP profile, but not in the corresponding T-RFLP profile. This was not artefactual – there was no significant difference in the overall distribution of the total lane volume in the profiles generated by T-RFLP and RT-T-RFLP profiling. In the case of a signal not being detected in the corresponding T-RFLP profile, it is likely that cells of an individual species were present in low numbers, but exhibited very high metabolic rates. When assessing relative levels of metabolic activity, it should be noted the number of ribosomal gene operons in different bacterial species varies. For example, Rickettsia prowazekii and Mycoplasma pneumoniae have only one ribosomal operon [21,22], whereas Clostridium paradoxum has 15 ribosomal operons [23]. Therefore, in a diverse bacterial community such as is present in the CF lung this variation will influence the apparent abundance of individual bands in T-RFLP profiles. The impact of this phenomenon on T-RFLP profiles has yet to be determined as many bands have no species assignation currently (only ~18% of T-RF band lengths match a band length generated from published sequence data). Moreover, even if all bands were associated with individual species, currently just over one hundred species [24] have been fully sequenced. Iteratively, this will become less problematic as more band-species linkages and genome sequences become available.
In the case of bands not being detected in the corresponding RT-T-RFLP analysis, it is likely that these bacteria were either dead or active at very low metabolic rates. The identification of bands in RT-T-RFLP profiles but absent in the corresponding T-RFLP profile from the same sample therefore suggests that these species are highly active but present in numbers below the threshold of detection. The most interesting such case was the identification of a band corresponding to B. cepacia complex, a known CF pathogen, in RT-T-RFLP profiles but not the corresponding T-RFLP profiles. This suggests that B. cepacia complex is present in relatively low numbers, but is highly metabolically active. Compared to P. aeruginosa, B. cepacia complex infects only a small proportion of CF patients, but its impact on survival is significant [25]. Further, the clinical outcome of CF patients colonised by B. cepacia complex is much poorer following lung transplantation than their non-infected counterparts [26,27]. For these reasons, the detection of strains of B. cepacia complex using these approaches will therefore be carefully monitored in future studies.
The lungs of all individuals are exposed to transient bacteria both that originate both in the oropharyngeal flora and the wider environment. For the detection of such a large number of metabolically active bacterial species in CF sputum to be significant it must be established that they are not due to contamination.
Strenuous attempts were made when processing the sputum samples analysed here to remove bacteria that may have adhered to the sputum bolus during its passage through the upper airways. Further comparisons between the bacterial populations found in CF sputa with those found in mouthwashes obtained from the same patients (data not shown) suggest that there is no significant cross-contamination. Further, here we have analysed the bacterial communities found in sputum obtained from healthy, non-CF individuals. It was found that the majority of the metabolically active bacterial species detected in the CF sputa were not detected at all in non-CF sputa and that bacterial profiles generated from healthy individuals show both a high degree of conservation, and a distinct dissimilarity to those generated from CF sputa. Further, the detection of metabolically active bacterial species in both CF sputum and either CF oropharyngeal community or non-CF sputum is not necessarily an indication of contamination. Bacterial species colonising the CF lung almost certainly derive from either transient bacterial species to which we are all exposed, or to oropharyngeal flora. Therefore, their presence in sputum may well indicate genuine lung colonisation.
Like any other novel approach, data interpretation must be carried out with caution. However, more studies using this and complementary approaches will build upon this characterisation of the complex bacterial community colonising the CF lung. Moreover, this approach can be applied to many other infectious diseases of single or polymicrobial origin. Other diagnostic systems such as gene array applications have clear diagnostic potential [reviewed in [28]]. By their nature however, they can only detect signature sequences for bacteria that have been deposited on the supporting matrix and are, therefore, "closed" systems. The RT-T-RFLP approach that we describe here is an open system where no prior assumptions have been made about the species present in the clinical sample and may, as such, be more flexible. For diagnosis, the RT-T-RFLP system will be developed through the construction of a large database containing digest patterns for individual species. In turn, this will provide clinicians with diagnostic identifications tied to probability value scores. This process translates into more rapid detection – providing results in hours rather than days. This can be equated easily to cost savings in terms of operator time and, where relevant, hospital bed occupancy. Also, there are benefits in terms of more appropriate drug selection. Apart from likely cost savings, this will be particularly important in reducing the impact of long term antibiotic use on patients e.g. renal damage.
The RT-T-RFLP system can offer flexibility in other ways. By switching the PCR primers used, it has the potential to reveal diagnostic information at either the sub-species level, or for fungal or viral infections. Moreover, RT-T-RFLP system can be coupled with Quantitative PCR techniques [29]. With development, RT-T-RFLP has the potential to rapidly assess treatment efficacy by monitoring bacterial viability within clinical samples taken from an individual patient. This monitoring would provide clinicians with data with which to make informed decisions e.g. on dosage modifications or more profound alterations in therapy. Through such studies, this may provide fresh insights into the process of infection in a wide range of respiratory diseases.
Conclusion
Overall, this study has shown that the majority of bacterial species detected and previously reported in CF lung infections are metabolically active. Further, this suggests that the majority of species detected in samples from the CF lung may play important roles in lung infections. The roles of these bacteria in the pathogenic process occurring in the CF lung are therefore worthy of much further investigation. Moreover, this approach may prove very valuable clinically in the study of many other microbial infections.
Authors' contributions
All authors participated in the conception and design of the study. GB Rogers carried out all experimental procedures and data analysis. Work was coordinated by KD Bruce. All authors read and approved the final manuscript.
Supplementary Material
Additional File 1
Table 2. T-RF band lengths and percentages of total lane signal volume. The length in bases and the relative proportion of the total lane signal volume (% total) of each of the T-RF bands detected in the T-RFLP and RT-T-RFLP profiles generated from the samples.
Click here for file
Acknowledgements
We wish to acknowledge Jackie Hunt, Andy Tuck and members of the Adult CF Unit, Southampton General Hospital for their helpful advice and assistance. We wish also to acknowledge the King's College London and the Cystic Fibrosis Trust (CSIG grant) for funding that supported this work.
==== Refs
Horner-Devine MC Carney KM Bohannan BMJ An ecological perspective on bacterial biodiversity Proc Roy Soc London Ser B-Biol Sci 2004 271 113 122 10.1098/rspb.2003.2549
Keer JT Birch L Molecular methods for the assessment of bacterial viability J Microbiol Meth 2003 53 175 183 10.1016/S0167-7012(03)00025-3
Torsvik V Ovreas L Microbial diversity and function in soil: from genes to ecosystems Curr Opin Microbiol 2002 5 240 245 12057676 10.1016/S1369-5274(02)00324-7
Liesack W Stackebrandt E Occurrence of novel groups of the Domain Bacteria as revealed by analysis of genetic material isolated from an Australian terrestrial environment J Bacteriol 1992 174 5072 5078 1629164
Liu WT Marsh TL Cheng H Forney LJ Characterization of microbial diversity by determining Terminal Restriction Fragment Length Polymorphisms of genes encoding 16S rRNA Appl Environ Microbiol 1997 63 4516 4522 9361437
McCormick J Green MW Mehta G Culross F Mehta A Demographics of the UK cystic fibrosis population: implications for neonatal screening Eur J Hum Genet 2002 10 583 90 12357328 10.1038/sj.ejhg.5200850
Lewis PA Morison S Dodge JA Geddes D Coles EC Russell G Littlewood JM Scott MT Survival estimates for adults with cystic fibrosis born in the United Kingdom between 1947 and 1967 Thorax 1999 54 420 422 10212106
Lyczak JB Cannon CL Pier GB Lung infections associated with cystic fibrosis Clin Microbiol Rev 2002 15 194 222 11932230 10.1128/CMR.15.2.194-222.2002
Kolak M Karpati F Monstein HJ Jonasson J Molecular typing of the bacterial flora in sputum of cystic fibrosis patients Int J Med Microbiol 2003 293 309 17 14503795
Rogers GB Hart CA Mason JR Hughes M Walshaw MJ Bruce KD Bacterial diversity in cases of lung infection in cystic fibrosis patients: 16S ribosomal DNA (rDNA) length heterogeneity PCR and 16S rDNA terminal restriction fragment length polymorphism profiling J Clin Microbiol 2003 41 3548 3558 12904354 10.1128/JCM.41.8.3548-3558.2003
van Belkum A Renders NH Smith S Overbeek SE Verbrugh HA Comparison of conventional and molecular methods for the detection of bacterial pathogens in sputum samples from cystic fibrosis patients FEMS Immunol Med Microbiol 2000 27 51 57 10617790 10.1016/S0928-8244(99)00161-3
Rogers GB Carroll MP Serisier DJ Hockey PM Jones G Bruce KD Characterization of Bacterial Community Diversity in Cystic Fibrosis Lung Infections Using 16S rDNA Terminal Restriction Fragment Length Polymorphism Profiling J Clin Microbiol
Brook I Fink R Transtracheal aspiration in pulmonary infection in children with cystic fibrosis Eur J Resp Dis 1983 64 51 57
Jewes LA Spencer RC The incidence of anaerobes in the sputum of patients with cystic fibrosis J Med Microbiol 1990 31 271 274 2325116
Costerton JW Stewart PS Greenberg EP A Common Cause of Persistent Infections Science 1999 284 1318 1322 10334980 10.1126/science.284.5418.1318
Wagner M Assmus B Hartmann A Hutzler P Amann R In situ analysis of microbial consortia in activated sludge using fluorescently labelled, rRNA-targeted oligonucleotide probes and confocal scanning laser microscopy J Microsc 1994 176 181 7 7532718
Cashel M Rudd KE Neidhardt FC, Ingraham JL, Low KB, Magasanik B, Schaechter M, Umbarger HE The stringent response Escherichia coli and Salmonella typhimurium: cellular and molecular biology 1987 2 American Society for Microbiology, Washington, DC 1410 1438
Lüdemann H Arth I Liesack W Spatial Changes in the Bacterial Community Structure along a Vertical Oxygen Gradient in Flooded Paddy Soil Cores Appl Environ Microbiol 2000 66 754 762 10653747 10.1128/AEM.66.2.754-762.2000
Coenye T Goris J Spilker T Vandamme P LiPuma JJ Characterization of unusual bacteria isolated from respiratory secretions of cystic fibrosis patients and description of Inquilinus limosus gen. nov., sp nov J Clin Microbiol 2002 40 2062 2069 12037065 10.1128/JCM.40.6.2062-2069.2002
Duan K Dammel C Stein J Rabin H Surette MG Modulation of Pseudomonas aeruginosa gene expression by host microflora through interspecies communication Mol Microbiol 2003 50 1477 1491 14651632 10.1046/j.1365-2958.2003.03803.x
Anderson SG Zomorodipour A Winkler HH Kurland CG Unusual organisation of the rRNA genes in Rickettsia prowazekii J Bacteriol 1995 177 4171 4175 7608097
Bercovier H Kafri O Sela S Mycobacteria possess a surprisingly small number of ribsomal RNA genes in relation to the size of their genome Biochem Biophys Res Commun 1986 136 1136 1141 3013168 10.1016/0006-291X(86)90452-3
Rainey FA Ward-Rainey NL Janssen PH Hippe H Clostridium paradoxum DSM 7308(T) contains multiple 16S rRNA genes with heterogenous intervening sequences Microbiology 1996 142 2087 2095 8760921
Centre for Biological Sequence Analysis, Technical University of Denmark (accessed July 19th, 2004)
Liou TG Adler FR Fitzsimmons SC Cahill BC Hibbs JR Marshall BC Predictive 5-year survivorship model of cystic fibrosis Am J Epidemiol 2001 153 345 352 11207152 10.1093/aje/153.4.345
Aris RM Routh JC LiPuma JJ Heath DG Gilligan PH Lung transplantation for cystic fibrosis patients with Burkholderia cepacia complex: survival linked to genomovar type Am J Respir Crit Care Med 2001 164 2102 2106 11739142
LiPuma JJ Burkholderia cepacia complex: a contraindication to lung transplantation in cystic fibrosis? Transpl Infect Dis 2001 3 149 160 11493397 10.1034/j.1399-3062.2001.003003149.x
Raoult D Fournier PE Drancourt M What does the future hold for Clinical Microbiology? Nature Rev Microbiol 2004 2 151 159 15040262 10.1038/nrmicro820
Harris KA Hartley JC Development of broad-range 16S rDNA PCR for use in the routine diagnostic Clinical Microbiology service J Med Microbiol 2003 52 685 691 12867563 10.1099/jmm.0.05213-0
| 15929792 | PMC1177990 | CC BY | 2021-01-04 16:23:26 | no | Respir Res. 2005 Jun 1; 6(1):49 | utf-8 | Respir Res | 2,005 | 10.1186/1465-9921-6-49 | oa_comm |
==== Front
Respir ResRespiratory Research1465-99211465-993XBioMed Central London 1465-9921-6-561594638110.1186/1465-9921-6-56ResearchTGF-β1 induces human alveolar epithelial to mesenchymal cell transition (EMT) Kasai Hidenori [email protected] Jeremy T [email protected] Roger M [email protected] Takashi [email protected] Zhi [email protected] Teijin Biomedical Laboratory, Medical Research Council Technology, 1–3 Burtonhole Lane, London, NW7 1AD, UK2 Biosciences Research Institute, University of Salford, Greater Manchester M5 4WT, UK3 Renal Medicine Section, Faculty of Medicine, Imperial College London, Hammersmith Hospital, London, W12 0NN, UK4 Institute for Bio-Medical Research, Teijin Pharma Ltd, Tokyo, 191-8512, Japan2005 9 6 2005 6 1 56 56 8 12 2004 9 6 2005 Copyright © 2005 Kasai et al; licensee BioMed Central Ltd.2005Kasai 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
Fibroblastic foci are characteristic features in lung parenchyma of patients with idiopathic pulmonary fibrosis (IPF). They comprise aggregates of mesenchymal cells which underlie sites of unresolved epithelial injury and are associated with progression of fibrosis. However, the cellular origins of these mesenchymal phenotypes remain unclear. We examined whether the potent fibrogenic cytokine TGF-β1 could induce epithelial mesenchymal transition (EMT) in the human alveolar epithelial cell line, A549, and investigated the signaling pathway of TGF-β1-mediated EMT.
Methods
A549 cells were examined for evidence of EMT after treatment with TGF-β1. EMT was assessed by: morphology under phase-contrast microscopy; Western analysis of cell lysates for expression of mesenchymal phenotypic markers including fibronectin EDA (Fn-EDA), and expression of epithelial phenotypic markers including E-cadherin (E-cad). Markers of fibrogenesis, including collagens and connective tissue growth factor (CTGF) were also evaluated by measuring mRNA level using RT-PCR, and protein by immunofluorescence or Western blotting. Signaling pathways for EMT were characterized by Western analysis of cell lysates using monoclonal antibodies to detect phosphorylated Erk1/2 and Smad2 after TGF-β1 treatment in the presence or absence of MEK inhibitors. The role of Smad2 in TGF-β1-mediated EMT was investigated using siRNA.
Results
The data showed that TGF-β1, but not TNF-α or IL-1β, induced A549 cells with an alveolar epithelial type II cell phenotype to undergo EMT in a time-and concentration-dependent manner. The process of EMT was accompanied by morphological alteration and expression of the fibroblast phenotypic markers Fn-EDA and vimentin, concomitant with a downregulation of the epithelial phenotype marker E-cad. Furthermore, cells that had undergone EMT showed enhanced expression of markers of fibrogenesis including collagens type I and III and CTGF. MMP-2 expression was also evidenced. TGF-β1-induced EMT occurred through phosphorylation of Smad2 and was inhibited by Smad2 gene silencing; MEK inhibitors failed to attenuate either EMT-associated Smad2 phosphorylation or the observed phenotypic changes.
Conclusion
Our study shows that TGF-β1 induces A549 alveolar epithelial cells to undergo EMT via Smad2 activation. Our data support the concept of EMT in lung epithelial cells, and suggest the need for further studies to investigate the phenomenon.
==== Body
Background
Idiopathic pulmonary fibrosis (IPF), the most common pulmonary fibrotic disorder, is a progressive and lethal disease of unknown etiology whose pathogenesis uniquely features the presence of fibroblastic foci in the parenchyma of the lungs [1]. These are comprised of aggregates of mesenchymal cells including fibroblasts and cells which exhibit phenotypic features of myofibroblasts, α-smooth muscle actin (αSMA) expression, increased mitogenic capacity, and enhanced extracellular matrix (ECM) production. The number of fibroblastic foci correlates with worsening lung function, progression of IPF and a poor prognosis [2]. According to the recent epithelial/fibroblastic model of IPF pathogenesis it is considered that fibroblastic foci underlie areas of unresolved epithelial injury and are sites where activated fibroblasts and myofibroblasts migrate, proliferate and synthesize ECM proteins [3]. However, the cellular origins of the mesenchymal phenotypes in fibroblast foci remain unclear.
It is now well recognized from many studies that a number of key growth factors are responsible for driving the process of fibrogenesis [4]. For example, transforming growth factor-beta1 (TGF-β1), interleukin-1 beta (IL-1β), and tumor necrosis factor-alpha (TNF-α) are able to induce the characteristic motility, proliferation and ECM synthesis observed in mesenchymal cells with a myofibroblast-like phenotype from fibroblastic foci. In general though, it is levels of TGF-β1 that best correlate with the extent of fibrosis and myofibroblast-like cell induction [5] and TGF-β1 continues to be regarded as the most important of the growth factors involved in pulmonary fibrogenesis [6]. For example, the biologically active form of TGF-β1 was aberrantly expressed in the epithelial cells lining honeycomb cysts within the lung of patients with IPF [7,8]. An increased level of TGF-β1 was found in BAL fluid derived from patients suffering from IPF [8]. Furthermore, overexpression of TGF-β1 in lung tissue induced prolonged pulmonary fibrosis in an animal model [9].
Recent evidence from studies of other fibrotic disorders, including renal [10,11] and liver fibrosis [12], supports a view that TGF-β1 may play a novel role in pulmonary fibrogenesis by promoting alveolar epithelial cell transition to form mesenchymal cells with a myofibroblast-like phenotype [10-14]. This process, termed epithelial-mesenchymal transition (EMT), occurs widely under both physiologic and pathologic conditions, for example during normal wound healing [13] and renal fibrosis [10,11]. Very recently it was reported that TGF-β1 induced type II alveolar epithelial cells isolated from rat lung to undergo EMT [15]. Epithelial cells are polarised, and display cytokeratin filaments and membrane-associated junctions. During EMT membrane-associated adherens junctions and desmosomes are dissociated, whilst at the same time or shortly after, cytoskeletal rearrangement takes place and mRNA for intermediate filament proteins is increased, facilitating the cell adopting a mesenchymal phenotype [14]. E-cadherin (E-cad) is an epithelial cell transmembrane protein whose extracellular domain interacts with that of an E-cad molecule expressed by an adjacent cell. It has a critical role in establishing firm adhesion, maintaining cell polarity and epithelial tightness [16]. The cadherin complex suppresses the dissociation of epithelial cells, and thus, the crucial step of EMT is the downregulation of E-cad [14].
To begin to understand the role of EMT in the development of fibroblastic foci in IPF, we have examined whether TGF-β1 can induce EMT in a human lung epithelial cell line (A549). A549 cells retain important characteristics of alveolar type II epithelial cells and have been employed in numerous studies as a valuable tool for studying promoter activity [17], apoptosis [18], and alveolar epithelial cell DNA damage [19]. In this study we demonstrate that TGF-β1 induces EMT in human type II alveolar epithelial cells through the activation of Smad2, and that this transition is accompanied by functional changes that are relevant to the progression of lung fibrosis.
Methods
Materials
Recombinant human TGF-β1, human IL-1β; and TNF-α were purchased from R&D systems (Minneapolis, MN). A mouse monoclonal antibody against human E-cad was from BD Transduction Laboratory (Oxford, UK). Mouse monoclonal anti-human fibronectin EDA+ splice form (Fn-EDA) and anti-human cytokeratin-19 were purchased from Abcam Ltd (Cambridge, UK). Goat anti-actin (C-11), Smad2 (S-20) and connective tissue growth factor (CTGF) polyclonal antibodies were supplied by Santa Cruz Biotechnology (Santa Cruz, CA). Rabbit anti-Erk1/2 and rabbit anti-phosphorylated Erk1/2 polyclonal antibodies were purchased from New England Biolabs (Hertfordshire, UK). Both biotinylated and non-biotinylated goat anti-human collagen type I, type III, and type IV polyclonal antibodies, and their standard proteins were supplied by Chemicon International (Temecula, CA). Rabbit anti-mouse immunoglobulin-HRP, rabbit anti-goat immunoglobulin-HRP and goat anti-rabbit immunoglobulin HRP were from DAKO Ltd (Cambridgeshire, UK). Rabbit anti-goat IgG-fluorescein (FITC) was purchased from Vector Labs (Perterborough, UK). Alkaline phosphatase-conjugated anti-Extravidin was supplied by Sigma (Dorset, Poole). PD-098059 and U-0126 were supplied by Merck Bioscience Ltd (Nottingham, UK). SMARTpool SMAD2 and non-specific control pool were purchased from Dharmacon, Inc. (Chicago, IL).
Cell culture
Human type II alveolar epithelial cells (A549) were supplied from ATCC/LGC Promochem (Middlesex, UK). Cells were maintained in low glucose-DMEM containing 10% FBS, 2 mM L-glutamine, 100 U/ml penicillin and 100 ug/ml streptomycin at 37°C in a humidified 5% CO2 atmosphere. Confluent cultures of cells were maintained in serum-free DMEM containing 0.1% BSA for 24 h prior to stimulation with cytokines. The cells were incubated with several concentrations of the cytokines for the periods indicated. In experiments using inhibitors of MAPK/Erk kinase (MEK), the cells were preincubated for 1 h with PD-098059 or U-0126 (up to 10 μM) before treatment with exogenous TGF-β1. In small interfering RNA (siRNA)-dependent gene silencing experiments, after transfection the cells were stimulated with 5 ng/ml of TGF-β1 in serum free 0.1% BSA/DMEM for 48 h. The cells were then harvested and lysed. In experiments testing collagen expression, the cells were incubated in the presence of 5 ug/ml of L-ascorbic acid. For immunocytochemical staining, the cells were maintained throughout the experiment in culture medium without serum deprivation.
SDS-PAGE and Western blot
The cells were scraped and lysed in M-PER (Pierce, Cheshire, UK) containing a protease inhibitor cocktail (Roche Diagnostic, East Sussex, UK). Cell suspensions were cleared by centrifugation at 13,000 g for 15 min at 4°C. Total protein concentration was measured using the BCA protein assay kit (Pierce) with bovine serum albumin as the standard protein. Equal amounts of protein were loaded for each lane of 10% SDS PAGE gels, followed by electrophoresis, and protein transfers to Hybond ECL membranes (Amersham, Buckinghamshire, UK), as described previously [14]. After the transfer, membranes were blocked with 5% skimmed milk and then probed with primary antibodies for 1 h at room temperature. After washing, the membranes were probed with appropriate peroxidase-conjugated secondary antibodies. After further extensive washing, the immunoblots were visualized by ECL (Amersham) and the band densities for each phenotype marker were quantified using QuantityOne Software (Bio-Rad) after scanning with a GS-710 Calibrated Imaging Densitometer (Bio-Rad). Results were expressed as a ratio of band density to total actin. For Western blotting of phosphorylated proteins, the cells were scraped and lyzed in RIPA buffer containing a protease inhibitor cocktail (Roche), plus 1 mM sodium-orthovanadate and 1 mM NaF. Western blot was performed according to manufacturer's instructions for each antibody. Changes in levels of phosphorylated proteins were assessed with reference to the respective non-phosphorylated proteins.
Reverse Transcription PCR (RT-PCR)
Total RNA was isolated using Trizol (Invitrogen, Carlsbad, CA). A one-step RT-PCR was performed using the SuperScript One-Step RT-PCR kit (Invitrogen). The target transcript was reverse transcribed at 50°C for 30 min. The PCR products for type III collagen and GAPDH were amplified using 35 cycles (initial denaturation at 94°C/2 min followed by PCR amplification, 94°C/15 s, 60°C/30 s, and 68°C/1 min). For type I collagen, the PCR product was amplified using 35 cycles (initial denaturation at 94°C/2 min followed by PCR amplification, 94°C/30 s, 58°C/30 s, and 68°C/1 min). PCR products were visualized on a GelDoc 1000 system (Bio-Rad) and semi-quantified using a scanning densitometer. The following primers were used for amplification of target transcripts; human collagen type I forward; 5'-ACGTCCTGGTGAAGTTGGTC-3', human collagen type I reverse; 5'-ACCAGGGAAGCCTCTCTCTC-3', human collagen type III forward; 5'-AGCCTCATTAGTCCTGATGGTTCTCG-3', human collagen type III reverse; 5'-CTTCTCAGCACTAGAATCTGTCCACC-3', human GAPDH forward; 5'-GGGCTGCTTTTAACTCTGGT-3', human GAPDH reverse; 5'-TGGCAGGTTTTTCTAGACGG-3'.
RNA Interference and Transfection
Dharmacon's SMARTpool SMAD2 siRNA reagent includes a pool of 4 SMARTselection-designed synthetic Smad2 siRNA duplexes, together with a non-specific control pool of siRNA as a negative control. Transfection of these pooled 21-nucleotide siRNA duplexes was carried out using Lipofectamine 2000 (Invitrogen), following the manufacturer's instructions. Cells were seeded at 5 × 104 cells/well of 24 wells plate and then incubated in normal medium without antibiotics for overnight to reach 80–90% confluence. siRNA-transfection reagent complexes were prepared as described by the manufacturer. Briefly 1 μl/well Lipofectamine 2000 was mixed with either 50 pmol/well Smad2 or negative control siRNA. FITC conjugated dsRNA (Block-iT, Invitrogen) was used as positive control. The cells were then transfected with siRNA/Lipofectamine complexes in Opti-MEM (Invitrogen) and incubated for 24 h at 37°C in a CO2 incubator. Following incubation, transfection efficiency was evaluated under fluorescence microscopy.
Immunocytochemistry
Cells were plated into a Lab-TeK Chamber Slide at a density of 2 × 104 cells/well and grown until approximately 80% confluent, after which they were treated with TGF-β1 at 5 ng/ml for 72 h, as described above. The cells were washed, fixed in ice-cold methanol for 10 min, and then soaked in PBS containing 2% TritonX-100 for 5 min to increase their permeability to antibodies. These cells were exposed to antibodies (1:20) for 1 h. After washing, bound primary antibodies were detected using appropriate FITC-conjugated secondary antibodies (1:50). Images were collected using an Eclipse TE2000-S microscope system (Nikon UK Ltd, Surrey) and Image-Pro Plus (Media Cybernetics UK, Berkshire).
Gelatin zymography for matrix metalloproteinases (MMPs) expression
Conditioned media were concentrated approximately 10-fold using ultra-centrifugation. After measuring protein concentration, equal amounts of samples were mixed with an equal volume of 2 × non-reducing SDS PAGE sample buffer. The samples were applied to a 10% (w/v) polyacrylamide gel impregnated with 2 mg/ml gelatin (Sigma). After electrophoresis, SDS was removed from the gel by washing 3 times for 10 min in 2.5% Triton X-100 solution. Then the gels were incubated overnight with gentle shaking at 37°C in buffer (50 mM Tris-HCl (pH7.6), 10 mM CaCl2, 50 mM NaCl, 0.05% Brij35), after which, the gel was stained with 0.25% Coomassie blue R250 in 40% methanol and 10% acetic acid for 2 h at room temperature, and subsequently destained with a 40% methanol-10% acetic acid solution until the bands became clear.
ELISA
Culture media were analyzed for collagens type I, type III and type IV using sandwich ELISA [14]. Each purified human collagen was used as a standard. Culture media were incubated in ELISA plates in which wells had been coated with non-biotinylated primary antibodies. Following addition of biotinylated antibodies, the plates were washed and reacted with alkaline phosphatase-conjugated anti-Extravidin. P-nitrophenyl phosphate substrate tablets were used to detect alkaline phosphatase activity and the product was measured at 405 nm using a micro-plate reader (Bio-Rad).
Statistical analysis
Data are presented as the mean ± SD of at least three independent experiments. For statistical analysis, an unpaired t-test was used for pair-wise comparisons and ANOVA with Dunnett's post test for other data. Statistical analysis was performed using commercial statistical software Prism Version 4.0 (GraphPad, Software Inc., San Diego, CA). P values less than 0.05 were considered as statistically significant.
Results
TGF-β1 induces alveolar epithelial cells to undergo EMT
We first determined the optimum concentrations and time required for TGF-β1 to initiate EMT in cultures of the alveolar type II epithelial cell line, A549. The expression of the epithelial phenotype markers, E-cad and cytokeratin 19, and of the mesenchymal phenotype markers, Fn-EDA and vimentin, were determined following treatment of A549 cells with various concentrations (0.01–10 ng/ml) of TGF-β1 for 24, 48 and 72 h (Figure 1). Changes in cell morphology were also assessed under phase contrast light microscopy (Figure 2).
Figure 1 Expression changes of EMT-related markers in A549 cells. (A) A549 cells were incubated with up to 10 ng/ml of TGF-β1 in the absence of serum for up to 72 h. Expression of the epithelial marker E-cadherin is down-regulated by TGF-β1 stimulation in a concentration-and time-dependent manner. Expression of Fn-EDA, which is a mesenchymal marker, is up-regulated by TGF-β1 in parallel with the down regulation in the epithelial marker. The same amounts of total protein are loaded in each lane. (B) Densitometric analysis of band intensities for each EMT related marker was performed at 48 h. Each bar represents mean ± SD of three independent experiments. * P < 0.05 and ** P < 0.01.
Figure 2 Morphological changes induced by TGF-β1. A549 cells were incubated with 5 ng/ml of TGF-β1 for 48 h. (A) Untreated A549 cells show a pebble-like shape and cell-cell adhesion is clearly observed. (B) TGF-β1-treated cells show a decrease in cell-cell contacts and adopt a more elongated morphological shape (magnification of 200×).
TGF-β1 significantly (P < 0.01) decreased E-cad expression in a concentration-and time-dependent manner (Figure 1A). Concentrations as low as 1 ng/ml of TGF-β1 induced up to 70% loss of E-cad expression within 48 h (Figure 1B). However, the extent of cytokeration 19 suppression was not as profound as for E-cad. Only high concentrations of TGF-β1 resulted in a measurable decrease (Figure 1B). In parallel with the marked decrease in the E-cad epithelial marker, TGF-β1 significantly (P < 0.01) induced expression of the mesenchymal marker Fn-EDA, in a concentration-and time-dependent manner (Figure 1A). But the level of vimentin expression was not as profound as for Fn-EDA. Our data also suggested that the de novo expression of Fn-EDA might occur earlier than E-cad suppression (Figure 1A).
In addition to the changes in the phenotypic markers expressed in A549 cells after TGF-β1 stimulation, the cells also underwent morphological changes on exposure to the growth factor (Figure 2). A549 cells cultured in the absence of TGF-β1 maintained a classic cobblestone epithelial morphology and growth pattern (Figure 2A), but after stimulation with 5 ng/ml of TGF-β1 for 48 h, the cells adopted a more fibroblast-like morphology and reduced their cell-cell contact (Figure 2B).
Both IL-1β and TNF-α have been suggested to play a role in fibroblast/myofibroblast motility, proliferation and ECM synthesis [4], and IL-1β induces kidney epithelial cells to undergo EMT [20,21]. We therefore tested whether these cytokines had similar properties to TGF-β1 in inducing lung alveolar epithelial cells to form mesenchymal-like cells. Up to 20 ng/ml of IL-1β had no effect on the expression of any of the molecular markers examined, whilst TNF-α, at 20 ng/ml concentration, resulted in a ~20% fall in the expression of E-cad (Figure 3). Neither TNF-α nor IL-1β induced the expression of Fn-EDA. These results suggested that alveolar EMT might be induced by aberrant expression and activation of TGF-β1 rather than by IL-1β or TNF-α.
Figure 3 Comparison of EMT-related marker expression in response to TGF-β1, IL-1β and TNF-α treatments. A549 cells were incubated with TGF-β1, IL-1β and TNF-α at the indicated concentrations for 48 h. Only TGF-β1 decreases E-cadherin expression concomitant with increasing Fn-EDA expression. In contrast, TNF-α only slightly decreases E-cadherin expression and IL-1β has no influence on EMT-related marker expression. Equal amounts of total protein are loaded in each lane.
TGF-β1-induced EMT is accompanied by other changes relevant to fibrogenesis
The results presented above suggest that lung epithelial cells take on a mesenchymal-like phenotype in response to TGF-β1. To confirm this, we examined the expression of collagen type I and type III by A549 cells in the presence and absence of TGF-β1, as these fibrillar collagens are characteristically synthesized by fibroblastic-type cells. We compared the production of these fibrillar collagens with that of type IV collagen which is characteristically synthesized by epithelial cells making basement membrane. Cells were treated with several concentrations of TGF-β1, and the effects on collagen expression assessed by RT-PCR; protein synthesis of collagens was also examined by ELISA and immunocytochemical staining. Figure 4 shows that TGF-β1 significantly (P < 0.01) stimulated the expression of collagens type I and type III, as detected by both RT-PCR (Figure 4A) and immuno-staining of the cell layer (Figure 4B). Protein levels of collagens type I and type III secreted into the culture medium were too low to be measured quantitatively by ELISA. Interestingly, the secretion of collagen type IV into the medium was increased in the presence of TGF-β1 (Figure 5). Concentrations of TGF-β1 as low as 0.1 ng/ml significantly induced secretion of collagen type IV as compared with control cells, the increased secretion level reaching a plateau at 1 ng/ml of TGF-β1 (Figure 5).
Figure 4 TGF-β1 induces the expression of collagens type I and type III. A549 cells were incubated with 5 ng/ml of TGF-β1 in the presence of 5 μg/ml of L-ascorbic acid for 72 h. (A) mRNA expression of collagens type I and type III was detected using RT-PCR. Densitometric analysis was performed. The changes of expression level are expressed as fold increase compared to the control. Each bar represents the mean ± SD of three independent experiments. ** P < 0.01. (B) Protein expression of collagens type I and type III was detected by immunocytochemical staining. Panels (a) and (b) represent collagen type I and panels (c) and (d) represent collagen type III expression, respectively. TGF-β1 induces collagen type I and type III expression in A549 as shown in panel (b) and (d).
Figure 5 TGF-β1 increases collagen type IV secretion in A549. A549 cells were incubated with several concentrations of TGF-β1 in the presence of 5 μg/ml of L-ascorbic acid for 72 h. Concentrations of collagen type IV in conditioned media were determined using ELISA. TGF-β1 increases collagen type IV in a concentration-dependent manner. Each bar was expressed as the mean ± SD of four independent experiments. ** P < 0.01.
CTGF acts in concert with TGF-β1 and is thought to have a significant role in promoting and maintaining fibrogenesis [22]. Thus, we investigated whether the expression of CTGF in A549 was affected by TGF-β1 treatment. Western analysis of both A549 cell lysates (Figure 6) and culture medium (data not shown) indicated that treatment with TGF-β1 at concentrations >1 ng/ml upregulated the expression of native CTGF (36–38 kDa) in a time-dependent manner. An additional smaller immunoreactive CTGF species was also detected and may correspond to a CTGF breakdown product which is frequently found in cell cultures [23].
Figure 6 TGF-β1 induces CTGF expression in A549. A549 cells were incubated for the times indicated with the concentrations of TGF-β1 shown. Cell lysates were used as a source for Western blot analysis of CTGF expression. Up-regulation of CTGF expression is observed with 1 ng/ml and higher concentrations of TGF-β1. This phenomenon parallels the altered expression of EMT related markers. Representative blots are shown from three independent experiments.
MMPs expression was examined using gelatin zymography to check whether A549 cells adopt characteristics necessary for cell migration in response to TGF-β1 treatment. As shown in Figure 7, A549 cells express gelatinases with molecular weights consistent with an identity of MMP-2 and MMP-9. MMP-2 was the main gelatinase expressed and TGF-β1 treatment up-regulated MMP-2 expression in a concentration-dependent manner. Basal MMP-9 expression was low and TGF-β1 had almost no effect on it.
Figure 7 Effect of TGF-β1 on MMPs expression in A549. Gelatin zymography was performed using the conditioned media that were harvested after 48 h TGF-β1 treatment (0.1 to 5 ng/ml). The samples were applied without reduction to a 10% polyacrylamide gel containing gelatin, and proteolytic activity was demonstrated by digestion of the gelatin and clearing of the gel.
EMT is associated with TGF-β1 signalling through the Smad pathway rather than via MAP Kinases
TGF-β1 signaling involves both the Smads and MAP kinases pathways [24]. The phosphorylation of Erk1/2 and Smad2 was examined at various time points after adding TGF-β1 (5 ng/ml) to A549 cells. The phosphorylation of Erk1/2 was increased slightly at 5 min after stimulation, and this effect lasted for at least 4 h without alteration of total Erk1/2 protein (Figure 8A). The TGF-β1-induced phosphorylation of Erk1/2 was completely suppressed in the presence of the MEK inhibitors, PD98059 (data not shown) or U0126 (Figure 8A). However MEK inhibitors had little or no effect on TGFβ1-induced changes in the expression of EMT markers over 48 h (Figure 8B).
Figure 8 Activation of Smad and ERK1/2 pathways by TGF-β1. A549 cells were pre-incubated in the presence or the absence of 10 μM of U0126, a potent MEK inhibitor, for 1 h prior to TGF-β1 stimulation. 5 ng/ml of TGF-β1 was used as a stimulus. TGF-β1 activates Erk1/2 and Smad2 pathways within 5 min after stimulation. The MEK inhibitor blocks Erk1/2 phosphorylation, but does not influence Smad2 phosphorylation. Equal amounts of total protein are loaded in each lane.
Interestingly, 5 ng/ml of TGF-β1 induced phosphorylation of Smad2 within 5 min of stimulation, and the level of Smad2 phosphorylation reached a maximum between 30–60 min after treatment and remained elevated for the duration of the experiment without affecting total Smad2 expression (Figure 8A). Co-incubation with either of the MEK inhibitors, PD98059 (data not shown) or U0126, had no effect on the TGF-β1 mediated Smad2 phosphorylation (Figure 8A). Taken together, these data indicate that rapid and sustained phosphorylation of Smad2 is associated with TGF-β1-induced EMT events and that TGF-β1-induced Erk1/2 signalling pathways are less likely to be involved in the EMT of A549 cells.
siRNA-mediated Smad2 gene silencing inhibits TGF-β-mediated EMT
In order to confirm whether Smad2 is involved in TGF-β1-mediated EMT, siRNAs were used to silence Smad2 gene expression in A549 cells. Transfection efficiency, using FITC-conjugated dsRNA oligomers, was approximately 100 % (data not shown). Western analysis revealed that total Smad2 protein expression was significantly (P < 0.01) depleted by Smad2 siRNA gene silencing, while pooled negative control siRNA had no detectable effect on Smad2 protein expression (Figure 9). Using a phospho-specific antibody against Smad2, we also tested for the effect of Smad2 siRNA on phosphorylated Smad2 levels after TGF-β1 treatment. In the presence of TGF-β1 phosphorylated Smad2 levels were significantly (P < 0.01) diminished by Smad2 siRNA (Figure 9). Most importantly, Smad2 siRNA restored the decreased expression of E-cad and cytokeration 19 induced by TGF-β1 treatment. However, the expression of Fn-EDA and vimentin were only slightly suppressed by Smad2 siRNA (Figure 9). Nevertheless, our data suggest that the activation of Smad2 signaling pathway is involved in TGF-β1-mediated EMT in A549 cells.
Figure 9 Effect of Smad2 siRNA on TGF-β1 induced EMT in A549 cells. Pooled synthetic siRNA duplexes targeting different regions of Smad2 were transfected into A549 cells at 50 pmol per well. 24 h after transfection, cells were stimulated with 5 ng/ml of TGF-β1 in serum free 0.1% BSA/DMEM for a further 48 h prior to harvest. Equal amounts of lysates were resolved by SDS-PAGE and analyzed by Western blotting for expression of proteins.
Discussion
Fibroblastic foci in the IPF lung parenchyma are characterized by vigorous replication of mesenchymal cells, including fibroblasts and myofibroblasts, and subsequent abnormal deposition of ECM proteins [25]. However, precise mechanisms responsible for the formation of fibroblastic foci are unknown and their cellular origins are unclear. It is recognized though, that injury to the alveolar epithelium precedes their formation and that epithelial cells may therefore contribute to their formation, perhaps by their responses to, or production of, key fibrogenic mediators including TGF-β1 [3]. Thus, one possible role of TGF-β1 in IPF is to induce or promote epithelial cells in injured alveolar epithelium to undergo EMT [15], as is evident from studies of other fibrotic disorders such as renal fibrosis [10,11,14]. Cells that have undergone EMT would then contribute to the development of fibroblastic foci and to their abnormal ECM production. The results of our present in vitro study support the concept underpinning this hypothesis; namely that TGF-β1 can induce alveolar EMT in human lung epithelial cells via the Smad2 pathway.
Epithelial phenotype is determined by the specific range of proteins expressed by the epithelial cell and the nature of its environment [14,26,27]. E-cad is an epithelial cell transmembrane protein with conserved cadherin repeats in the extracellular domain. In the presence of Ca2+, the extracellular domain binds to that of E-cad on an adjacent epithelial cell to form tight cell-cell adhesion and to suppress the dissociation of epithelial cells from their location. Certain morphogenic and/or environmental cues, such as local expression of TGF-β1, result in the loss of epithelial cell polarity, adherens junctions, tight junctions, desmosomes, cytokeratin intermediate filaments and, subsequently, rearrangement of F-actin stress fibers and the development of filopodia and lamelopodia [10,11]. Several in vitro studies have demonstrated that addition of TGF-β1 to cultured human epithelial cells from organs other than lung induces them to downregulate E-cad expression and to become mesenchymal cells, resembling myofibroblasts, via EMT [27-29].
In the present study, we investigated the potential for alveolar EMT by examining the expression of phenotypic markers in A549 cells. Although submerged monolayer culture of A549 may not completely mimic the pulmonary epithelium, these cells are still widely used for studies on the role of human alveolar type II epithelial cells as they retain features and metabolic properties characteristic of type II cells [17,19,30]. The present study demonstrates that low doses of TGF-β1 induce A549 cells to lose expression of epithelial phenotypic markers, such as E-cad and cytokeratin 19 expression, similar to our previously reported findings in kidney epithelial EMT [14]. Concurrently A549 cells gain a mesenchymal phenotype with de novo expression of Fn-EDA and increased expression of vimentin. In addition to these classic mesenchymal markers, these newly formed mesenchymal cells expressed the fibrillar collagens type I and type III which are likely to contribute to excessive accumulation of ECM in fibrotic tissue. Moreover, these cells expressed an elevated level of MMP-2 after EMT which could contribute to breakdown of basement membranes and facilitate migration. Our data is consistent with the findings made in in kidney tubular epithelial cells where MMP-2 and MMP-9 were up-regulated by TGF-β1 in parallel with changes in EMT markers [27,31]. It remains uncertain whether these fibroblast-like cells possess migratory and/or invasive capabilities, but TGF-β1 induced MMP-2 expression in A549 supports this notion. Clearly further investigation of cell migration in TGF-β1-mediated alveolar EMT in lung fibrosis is warranted.
We have yet to confirm our results in primary human type II alveolar epithelial cells because cultures of these cells rapidly adopt fibroblast-like morphology, even in the absence of TGF-β1. Therefore, we cannot rule out at this stage the possibility that the responses of A549 cells to TGF-β1 are unique to this human cell line. However, a recent report indicates that rat primary type II alveolar epithelial cells undergo EMT in vitro in response to treatment of TGF-β1 [15], suggesting EMT is likely to be a phenomenon common to all type II alveolar epithelial cells.
In addition to TGF-β1, many other cytokines, including IL-1β and TNF-α, have been suggested to play a role in IPF [4]. We examined whether, like TGF-β1, IL-1β and TNF-α also convert A549 to fibroblast-like cells. It is widely accepted that IL-1β induces EMT of renal epithelial cells through a TGF-β1-dependent mechanism [20,21]. However, our results showed that both TNFα and IL-1β failed to induce alveolar epithelial cells to undergo EMT, possibly due to the differences in the cell types investigated. Similarly A549 cells failed to respond to stimulation with 30% (v/v) activated PBMC-conditioned medium (aPBMC-CM) (H. Kasai, unpublished data), a stimulus which induces renal epithelial cells to undergo EMT-mediated conversion to myofibroblasts [14,32,33].
We observed that A549 EMT was not accompanied by expression of αSMA, even during longer periods (7 days) of exposure to TGF-β1 (data not shown). Although αSMA, together with vimentin and desmin, is often used to classify myofibroblasts [34,35], the expression of these cellular markers is varied and dependent on cell types and culture conditions [36].
TGF-β1 regulates various cell functions, such as cell proliferation, cell differentiation, apoptosis, cell adhesion/motility, ECM production, and its association with pulmonary fibrosis is well known [37]. In vivo studies have demonstrated increased TGF-β1 gene expression and protein secretion in the lungs of animals [38] and humans with fibrotic diseases [7,8,39]. Furthermore, transient overexpression of active TGF-β1 in rat lung resulted in severe interstitial and pleural fibrosis characterized by extensive deposition of ECM proteins, and by the emergence of cells with the myofibroblast phenotype [9]. A time-dependent production of endogenous TGF-β1 from rat alveolar epithelial cells after exogenous TGF-β1 treatment was also noted by Yao et al [15], suggesting initial EMT induced by TGF-β1 may result in further EMT induced by endogenous TGF-β1 production in an autocrine or paracrine manner. Based on these studies and our data it is tempting to speculate about a role for TGF-β1-mediated EMT in pulmonary fibrogenesis. Indirectly, our data showing TGF-β1-mediated induction of collagen type I and type III expression, which are present in fibrotic lesions in vivo [40], and the expression of Fn-EDA and vimentin which represent cellular markers for myofibroblasts, supports such a role for EMT.
TGF-β1 exerts its effects through heteromeric receptor complexes composed of type I and type II serine/threonine receptors. Upon ligand binding, the type II receptor phosphorylates the type I receptor inducing its kinase activity [41]. TGF-β1 activity may be transduced along the Smads pathway immediately downstream of the receptor complex in a variety of cell systems and also via JNK, p38 MAPK and Erk pathways [37,41,42]. Thus, we examined the intracellular signalling pathway involved in TGF-β1-mediated alveolar EMT. Although phosphorylation and activation of p38 MAPK and Erk1/2 were observed in some lung fibroblasts [43], our data suggested that the signalling pathway involved in alveolar EMT was likely to be a Smad2-dependent pathway since the MEK inhibitors, PD98059 and U0126, failed to reverse TGF-β1-induced phenotypic modulation of A549 cells, whereas Smad2 siRNA attenuated the loss of E-cad and cytokeratin 19 induced by TGF-β1. Smad2 phosphorylation has been noted in EMT processes for several cell types including breast and renal epithelial cells [44-46].
Transition from the alveolar epithelial phenotype to the mesenchymal phenotype initiated by TGF-β1 was accompanied by elevated expression of CTGF. CTGF induction is mediated through a TGF-β1 response element in the CTGF promoter and its mediation of at least some of the activities attributed to TGF-β1 is well recognized [4,22], as is its involvement in the maintenance of fibrogenesis. In lung for example, Allen et al reported that the mRNA level of CTGF in BALF cells of patients with IPF was significantly higher than that in healthy control subjects [46]. Furthermore, expression of CTGF was found in both interstitial fibroblasts and type II alveolar epithelial cells in patients with IPF [47]. CTGF secreted by alveolar epithelial cells and myofibroblasts responding to TGF-β1 may also act as a paracrine factor for lung fibroblasts and is known to be a critical intermediate for the synthesis of connective tissue proteins stimulated by TGF-β1, but not by other fibrogenic cytokines [48,49]. CTGF may therefore play a role in mediating the expression of collagens by EMT cells. Most recently, investigations with mesangial cells showed that CTGF interacts with the TrkA receptor, triggering events which lead to the induction of a transcriptional repressor, TIEG [50]. It was proposed that since this suppresses negative regulation of the TGF-β-Smad signaling pathway by repressing Smad7 expression, it leads to enhanced TGF-β1 signaling [51]. It is presently unclear whether CTGF plays a similar role in alveolar epithelial responses to TGF-β1, or in EMT cells derived from them. However, in spite of uncertainties surrounding its precise role, CTGF remains implicated in the pathogenesis of many fibrotic disorders [4], and is likely to contribute significantly to fibrogenesis in the lung.
Conclusion
Our findings show that TGF-β1 induces an EMT-like process in A549 alveolar epithelial cells, most likely by activation of the Smad2 signaling pathway. These data provide evidence to support the concept that human lung epithelial cells can undergo EMT and indicate a need for further studies. In particular, the expression profile associated with alveolar epithelial cells that have undergone EMT indicates a potential role for EMT in pulmonary fibrogenesis.
List of abbreviations
IPF = Idiopathic pulmonary fibrosis; αSMA = α-smooth muscle actin; ECM = extracellular matrix; TGF-β1 = transforming growth factor-beta1; IL-1β = interleukin-1 beta; TNF-α = tumor necrosis factor-alpha; EMT = epithelial mesenchymal transition; E-cad = E-cadherin; Fn-EDA = fibronectin EDA+ splice form; CTGF = connective tissue growth factor; MEK = MAPK/Erk kinase; RT-PCR = Reverse Transcription PCR; MMP = matrix metalloproteinase; siRNA = small interfering RNA
Authors' contributions
HK carried out the cellular and biochemical studies and participated in drafting the manuscript. JA, RM and TK participated in the design of the study and drafted the manuscript. ZZ conceived the study, and participated in its design and coordination, and in drafting and finalizing the manuscript. All authors read and approved the final manuscript.
Acknowledgements
This work was supported by the MRC Technology and Teijin Pharma Ltd.
==== Refs
Selman M King TE Pardo A Idiopathic pulmonary fibrosis: prevailing and evolving hypotheses about its pathogenesis and implications for therapy Ann Intern Med 2001 134 136 151 11177318
King TE JrSchwarz MI Brown K Tooze JA Colby TV Waldron JA JrFlint A Thurlbeck W Cherniack RM Idiopathic pulmonary fibrosis: relationship between histopathologic features and mortality Am J Respir Crit Care Med 2001 164 1025 1032 11587991
Selman M Pardo A Idiopathic pulmonary fibrosis: an epithelial/fibroblastic cross-talk disorder Respir Res 2002 3 3 11806838 10.1186/rr175
Allen JT Spiteri MA Growth factors in idiopathic pulmonary fibrosis: relative roles Respir Res 2002 3 13 11806848 10.1186/rr162
Gauldie J Kolb M Sime PJ A new direction in the pathogenesis of idiopathic pulmonary fibrosis? Respir Res 2002 3 1 11806836 10.1186/rr158
Kelly M Kolb M Bonniaud P Gauldie J Re-evaluation of fibrogenic cytokines in lung fibrosis Curr Pharm Des 2003 9 39 49 12570673 10.2174/1381612033392341
Khalil N O'Connor RN Unruh HW Warren PW Flanders KC Kemp A Bereznay OH Greenberg AH Increased production and immunohistochemical localization of transforming growth factor-beta in idiopathic pulmonary fibrosis Am J Respir Cell Mol Biol 1991 5 155 162 1892646
Khalil N Parekh TV O'Connor R Antman N Kepron W Yehaulaeshet T Xu YD Gold LI Regulation of the effects of TGF-beta 1 by activation of latent TGF-beta 1 and differential expression of TGF-beta receptors (T beta R-I and T beta R-II) in idiopathic pulmonary fibrosis Thorax 2001 56 907 915 11713352 10.1136/thorax.56.12.907
Sime PJ Xing Z Graham FL Csaky KG Gauldie J Adenovector-mediated Gene Transfer of Active Transforming Growth Factor-β1 Induces Prolonged Severe Fibrosis in Rat Lung J Clin Invest 1997 100 768 776 9259574
Kalluri R Neilson EG Epithelial-mesenchymal transition and its implications for fibrosis J Clin Invest 2003 112 1776 1784 14679171 10.1172/JCI200320530
Liu Y Epithelial to mesenchymal transition in renal fibrogenesis: pathologic significance, molecular mechanism, and therapeutic intervention J Am Soc Nephrol 2004 15 1 12 14694152 10.1097/01.ASN.0000106015.29070.E7
Desmouliere A Darby IA Gabbiani G Normal and pathologic soft tissue remodeling: role of the myofibroblast, with special emphasis on liver and kidney fibrosis Lab Invest 2003 83 1689 1707 14691287 10.1097/01.LAB.0000101911.53973.90
Desmouliere A Factors influencing myofibroblast differentiation during wound healing and fibrosis Cell Biol Int 1995 19 471 476 7640660 10.1006/cbir.1995.1090
Nightingale J Patel S Suzuki N Buxton R Takagi KI Suzuki J Sumi Y Imaizumi A Mason RM Zhang Z Oncostatin M, a cytokine released by activated mononuclear cells, induces epithelial cell-myofibroblast transdifferentiation via Jak/Stat pathway activation J Am Soc Nephrol 2004 15 21 32 14694154 10.1097/01.ASN.0000102479.92582.43
Yao HW Xie QM Chen JQ Deng YM Tang HF TGF-beta1 induces alveolar epithelial to mesenchymal transition in vitro Life Sci 2004 76 29 37 15501477 10.1016/j.lfs.2004.06.019
Vleminckx K Kemler R Cadherins and tissue formation: integrating adhesion and signaling Bioessays 1999 21 211 220 10333730 10.1002/(SICI)1521-1878(199903)21:3<211::AID-BIES5>3.0.CO;2-P
Higashimoto Y Keicho N Elliott WM Hogg JC Hayashi S Effect of adenovirus E1A on ICAM-1 promoter activity in human alveolar and bronchial epithelial cells Gene Expr 1999 8 287 297 10947078
Maeyama T Kuwano K Kawasaki M Kunitake R Hagimoto N Matsuba T Yoshimi M Inoshima I Yoshida K Hara N Upregulation of Fas-signalling molecules in lung epithelial cells from patients with idiopathic pulmonary fibrosis Eur Respir J 2001 17 180 189 11334117 10.1183/09031936.01.17201800
Upadhyay D Bundesmann M Panduri V Correa-Meyer E Kamp DW Fibroblast growth factor-10 attenuates H2O2-induced alveolar epithelial cell DNA damage: role of MAPK activation and DNA repair Am J Respir Cell Mol Biol 2004 31 107 113 14975937 10.1165/rcmb.2003-0064OC
Vesey DA Cheung CWY Cuttle L Endre Z Gobe G Johnson DW Interleukin-1β induces human proximal tubule cell injury, α-smooth muscle actin expression and fibronectin production Kidney Int 2002 62 31 40 12081561 10.1046/j.1523-1755.2002.00401.x
Fan J Huang X Ng Y Nikolic-Paterson DJ Mu W Atkins RC Lan HY Interleukin-1 induces tubular epithelial-myofibroblast transdifferentiation through a transforming growth factor-β1-dependent mechanism in vitro Am J Kidney Dis 2001 37 820 831 11273883
Bonniaud P Margetts PJ Kolb M Haberberger T Kelly M Robertson J Gauldie J Adenoviral gene transfer of connective tissue growth factor in the lung induces transient fibrosis Am J Respir Crit Care Med 2003 168 770 778 12816739 10.1164/rccm.200210-1254OC
Wahab NA Yevdokimova N Weston BS Roberts T Li XJ Brinkman H Mason RM Role of connective tissue growth factor in the pathogenesis of diabetic nephropathy Biochem J 2001 35 77 87 11563971 10.1042/0264-6021:3590077
Bottinger EP Bitzer M TGF-beta signaling in renal disease J Am Soc Nephrol 2002 13 2600 2610 12239251 10.1097/01.ASN.0000033611.79556.AE
Gross TJ Hunninghake GW Idiopathic pulmonary fibrosis N Engl J Med 2001 345 517 525 11519507 10.1056/NEJMra003200
Ewing CM Ru N Morton RA Robinson JC Wheelock MJ Johnson KR Barrett JC Isaacs WB Chromosome 5 suppresses tumorigenicity of PC3 prostate cancer cells: correction with re-expression of alpha-catenin and restoration of E-cadherin function Cancer Res 1995 55 4813 4817 7585512
Yang J Liu Y Dissection of key events in tubular epithelial to myofibroblast transition and its implications in renal interstitial fibrosis Am J Pathol 2001 159 1465 1475 11583974
Miettinen PJ Ebner R Lopez AR Derynck R TGF-beta induced transdifferentiation of mammary epithelial cells to mesenchymal cells: involvement of type I receptors J Cell Biol 1994 127 2021 2036 7806579 10.1083/jcb.127.6.2021
Bhowmick NA Ghiassi M Bakin A Aakre M Lundquist CA Engel ME Arteaga CL Moses HL Transforming growth factor-beta1 mediates epithelial to mesenchymal transdifferentiation through a RhoA-dependent mechanism Mol Biol Cell 2001 12 27 36 11160820
Foster KA Oster CG Mayer MM Avery ML Audus KL Characterization of the A549 cell line as a type II pulmonary epithelial cell model for drug metabolism Exp Cell Res 1998 243 359 366 9743595 10.1006/excr.1998.4172
Strutz F Zeisberg M Ziyadeh FN Yang CQ Kalluri R Muller GA Neilson EG Role of basic fibroblast growth factor-2 in epithelial-mesenchymal transformation Kidney Int 2002 61 1714 1728 11967021 10.1046/j.1523-1755.2002.00333.x
Sharmila P Takagi K Suzuki J Imaizumi A Kimura T Mason RM Kamimura T Zhang Z RhoGTPase activation is a key step in renal epithelial mesenchymal transdifferentiation J Am Soc Nephrol 2005 16 7 1977 1984 15901767 10.1681/ASN.2004110943
Healy E Leonard M Madrigal-Estebas L O'Farrelly C Watson AJ Ryan MP Factors produced by activated leukocytes alter renal epithelial cell differentiation Kidney Int 1999 56 1266 1269 10504471 10.1046/j.1523-1755.1999.00694.x
Mermall V Post PL Mooseker MS Unconventional myosins in cell movement, membrane traffic, and signal transduction Science 1998 279 527 533 9438839 10.1126/science.279.5350.527
Powell DW Mifflin RC Valentich JD Crowe SE Saada JI West AB Myofibroblasts. I. Paracrine cells important in health and disease Am J Physiol 1999 277 C1 C9 10409103 10.1111/j.1469-7793.1999.001af.x
Powell DW Mifflin RC Valentich JD Crowe SE Saada JI West AB Myofibroblasts. II. Intestinal subepithelial myofibroblasts Am J Physiol 1999 277 C183-C201 10444394
Derynck R Feng XH TGF-β receptor signaling Biochim Biophys Acta 1997 1333 F105 F150 9395284
Hoyt DG. Lazo JS Alterations in pulmonary mRNA encoding procollagens, fibronectin, and transforming growth factor-β precedes bleomycin-induced pulmonary fibrosis in mice J Pharmacol Exp Ther 1988 246 765 771 2457084
Khalil N O'Conner RN Flanders KC Umruh H TGF-β1, but not TGF-β2 or TGF-β3, is differentially present in epithelial cells of advanced pulmonary fibrosis: an immunohistochemical study Am J Respir Cell Mol Biol 1996 14 131 138 8630262
Raghu G Masta S Meyers D Narayanan AS Collagen synthesis by normal and fibrotic human lung fibroblasts and the effect of transforming growth factor-beta Am Rev Respir Dis 1989 140 95 100 2751176
Whitman M Smads and early developmental signaling by the TGFβ superfamily Genes Dev 1998 12 2445 2462 9716398
Atfi A Djellou S Chastre E Davis R Gespach C Evidence for a role of Rho-like GTPase and stress-activated protein kinase/c-Jun-N-terminal kinase (SAPK/JNK) in transforming growth factorβ-mediated signaling J Biol Chem 1997 272 1429 1432 8999807 10.1074/jbc.272.40.24731
Hashimoto S Gon Y Takeshita I Matsumoto K Maruoka S Horie T Transforming growth factor-β1 induces phenotypic modulation of human lung fibroblasts to myofibroblast through a c-Jun-NH2-terminal kinase-dependent pathway Am J Respir Crit Care Med 2001 163 152 157 11208641
Piek E Moustakas A Kurisaki A Heldin CH ten Dijke P TGF-(beta) type I receptor/ALK-5 and Smad proteins mediate epithelial to mesenchymal transdifferentation in NMuMG breast epithelial cells J Cell Sci 1999 112 4557 4568 10574705
Rhyu DY Yang Y Ha L Lee GT Song JS Uh ST Lee HB Role of reactive oxygen species in TGF-beta1-induced mitogen-activated protein kinase activation and epithelial-mesenchymal transition in renal tubular epitheliasl cells J Am Soc Nephrol 2005 16 667 675 15677311 10.1681/ASN.2004050425
Allen JT Knight RA Bloor CA Spiteri MA Enhanced insulin-like growth factor binding protein-related protein 2 (Connective tissue growth factor) expression in patients with idiopathic pulmonary fibrosis and pulmonary sarcoidosis Am J Respir Cell Mol Biol 1999 21 693 700 10572066
Pan LH Yamauchi K Uzuki M Nakanishi T Takigawa M Inoue H Sawai T Type II alveolar epithelial cells and interstitial fibroblasts express connective tissue growth factor in IPF Eur Respir J 2001 17 1220 1227 11491168 10.1183/09031936.01.00074101
Grotendorst GR Connective tissue growth factor: a mediator of TGF-beta action on fibroblasts Cytokine Growth Factor Rev 1997 8 171 179 9462483 10.1016/S1359-6101(97)00010-5
Moussad EE Brigstock DR Connective tissue growth factor: what's in a name? Mol Genet Metab 2000 71 276 292 11001822 10.1006/mgme.2000.3059
Wahab NA Weston BS Mason RM Connective tissue growth factor (CTGF, CCN2) interacts with and activates the tyrosine kinase receptor TrkA J Am Soc Nephrol 2005 16 340 351 15601748 10.1681/ASN.2003100905
Wahab NA Mason RM Connective tissue growth factor and renal diseases: some answers, more questions Curr Opin Nephrol Hypertens 2004 13 53 58 15090860
| 15946381 | PMC1177991 | CC BY | 2021-01-04 16:23:26 | no | Respir Res. 2005 Jun 9; 6(1):56 | utf-8 | Respir Res | 2,005 | 10.1186/1465-9921-6-56 | oa_comm |
==== Front
Respir ResRespiratory Research1465-99211465-993XBioMed Central London 1465-9921-6-581595525210.1186/1465-9921-6-58ResearchInhibition or knock out of Inducible nitric oxide synthase result in resistance to bleomycin-induced lung injury Genovese Tiziana [email protected] Salvatore [email protected] Paola Rosanna [email protected] Marco [email protected] Emanuela [email protected] Maria Angela [email protected] Giuseppina [email protected] Elisa [email protected] Nunzio [email protected] Achille P [email protected] Carlo [email protected] Department of Clinical and Experimental Medicine and Pharmacology, Torre Biologica, Policlinico Universitario, 98123 Messina, Italy2 Department of Internal and Specialistic Medicine, Section of Respiratory Diseases, University of Catania, Catania, Italy3 Department of Experimental and Clinical Pharmacology, University of Catania, Catania, Italy2005 14 6 2005 6 1 58 58 17 2 2005 14 6 2005 Copyright © 2005 Genovese et al; licensee BioMed Central Ltd.2005Genovese 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 the present study, by comparing the responses in wild-type mice (WT) and mice lacking (KO) the inducible (or type 2) nitric oxide synthase (iNOS), we investigated the role played by iNOS in the development of on the lung injury caused by bleomycin administration. When compared to bleomycin-treated iNOSWT mice, iNOSKO mice, which had received bleomycin, exhibited a reduced degree of the (i) lost of body weight, (ii) mortality rate, (iii) infiltration of the lung with polymorphonuclear neutrophils (MPO activity), (iv) edema formation, (v) histological evidence of lung injury, (vi) lung collagen deposition and (vii) lung Transforming Growth Factor beta1 (TGF-β1) expression.
Methods
Mice subjected to intratracheal administration of bleomycin developed a significant lung injury. Immunohistochemical analysis for nitrotyrosine revealed a positive staining in lungs from bleomycin-treated iNOSWT mice.
Results
The intensity and degree of nitrotyrosine staining was markedly reduced in tissue section from bleomycin-iNOSKO mice. Treatment of iNOSWT mice with of GW274150, a novel, potent and selective inhibitor of iNOS activity (5 mg/kg i.p.) also significantly attenuated all of the above indicators of lung damage and inflammation.
Conclusion
Taken together, our results clearly demonstrate that iNOS plays an important role in the lung injury induced by bleomycin in the mice.
==== Body
Background
Pulmonary fibrosis is a progressive interstitial lung disease of unknown etiology. Pulmonary fibrosis is characterized by inflammatory cell infiltration, fibroblast proliferation, and excessive deposition of extracellular matrix proteins in the lung parenchyma [1,2]. The disease most commonly affects middle-age adults, although infants and children are also affected. Various studies have also indicated that the treatment with bleomycin during cancer chemotherapy in humans also induces interstitial fibrosis [3,4].
Nitric oxide (NO) is a pleiotropic mediator, which acts in a variety of physiological and pathophysiological processes [5-8]. NO is produced from the oxidation of L-arginine by the enzyme NO synthase [9,10] which occurs in three major isoforms; two are constitutive (endothelial and neuronal, indicated with cNOS), and one is inducible (macrophagic). The constitutively expressed enzyme (cNOS) are calcium-dependent, release NO under physiological condition in various cells, including endothelial cells and neurons, and NO released by these isoform are involved in the regulation of blood pressure in organ blood flow distribution, in the inhibition of the adhesion and activation of platelets and polymorphonuclear granulocytes and in neuronal transmission. The inducible isoform of NOS (iNOS) is calcium-independent and can be induced by proinflammatory agents, such as endotoxins (bacterial lipopolysaccharide, LPS), interleukin-1β, tumor necrosis factor-α (TNF-α) and interferon-γ (INF-γ), in endothelial and smooth-muscle cells, in macrophages and in other cell types [5-9]. Enhanced formation of NO following the induction of iNOS has been implicated in the pathogenesis of shock and inflammation [5].
Although the severity and duration of inflammation may dictate the timing and extent of NOS expression, it is now evident that the up-regulation of NOS can modulate inflammation [9-11]. Pharmacological inhibition of iNOS or genetic inactivation of NOS (iNOS knockout mice) attenuates the activation of the transcription factors nuclear factor kappa B (NF-κB) and signal transducer and activator of transcription-3 (STAT-3), and increases Granulocyte Colony-Stimulating Factor (G-CSF) messenger RNA levels in the tissue. Thus, induced nitric oxide, in addition to being a "final common mediator" of inflammation, is essential for the up-regulation of the inflammatory response. Furthermore, it has been recently suggested that some of the cytotoxic effects of NO are tightly related to the production of peroxynitrite, a high-energy oxidant deriving by the rapid reaction of NO with superoxide [12-14]. The resulting oxidative stress may cause cell death and tissue damage that characterize a number of human disease states like neurological disorders and stroke, inflammatory bowel disease, arthritis, toxic shock and acute reperfusion injuries [15-18]. Thus peroxynitrite, and not NO, has been proposed to be the ultimate cytotoxic species in many conditions acting through some mechanisms including the initiation of lipid peroxidation, the inactivation of a variety of enzymes (e.g. MnSOD) and the depletion of glutatione. Moreover, peroxynitrite is also able to induce DNA damage [19,20] resulting in inactivation of the nuclear enzyme PARS, in depletion of nicotinamide adenine dinucleotide (NAD+) and adenosine triphosphate (ATP) and lastly in cell death [21]. The realization of the cytotoxic potential of NO and peroxynitrite made it important to seek for pharmacological approaches, in order to neutralize NO and peroxynitrite-induced damage by inhibiting iNOS. The role of iNOS in pathologic condition have induced the development of selective iNOS inhibitors like GW274150 [(S)-2-Amino-(1-iminoethylamino)-5-thioheptanoic acid]. This molecule is a novel NOS-inhibitor (sulphur-substituted acetamine amono acid), which acts in competition with L-arginine and has a very high degree of selectivity for iNOS when compared to either eNOS (> 300-fold) or nNOS (> 100-fold) [22]. In addition GW274150 is a long acting (5 hours half life in rats) iNOS inhibitor and is also able to inhibit LPS-mediated increase in plasma NO2- NO3- levels 14 h after single intraperitoneal dose (ED50 3 mg kg-1) [23]. The inhibition of iNOS activity caused by GW274150 is NADPH-dependent and develops very slowly, but is rapidly reversible and recent studies reports the role of this iNOS selective inhibitors in reducing organ injury in hemorrhagic shock, in collagen induced arthritis and in renal ischemia/reperfusion [24-26]. In addition recently we have demonstrated that GW274150 treatment significantly reduced acute lung injury in an experimental model of carrageenan induced pleurisy [27]. Therefore the aim of this study was to investigate the role of iNOS in a model of lung injury induced by bleomycin administration using iNOSKO mice and iNOSWT mice. In addition, we have investigated the effects of the systemic administration of GW274150 in iNOSWT mice subjected to bleomycin-induced lung injury. In particular, we have investigated the effect of the genetic or pharmacological inhibition of iNOS on the bleomycin induced (i) loss of body weight, (ii) PMN lung infiltration [myeloperoxidase (MPO) activity], (iii) lung tissue edema [wet/dry ratio], (iv) lipid peroxidation, (v) the nitration of tyrosine residues (an indicator of the formation of peroxynitrite), (vi) lung damage (histology), (vii) lung collagen deposition and (viii) lung TGF-β1 expression.
Materials and methods
Animals
Male CD mice (25–35 g; Harlan Nossan; Italy) were housed in a controlled environment and provided with standard rodent chow and water. Animal care was in compliance with Italian regulations on protection of animals used for experimental and other scientific purpose (D.M. 116192) as well as with the EEC regulations (O.J. of E.C. L 358/1 12/18/1986).
Experimental groups
Mice were randomly allocated into the following groups: (i) iNOSWT + BLEO group. Mice were subjected to bleomycin-induced lung injury (N = 30), (ii) iNOSKO + BLEO group. Mice were subjected to bleomycin-induced lung injury (N = 30), (iii) iNOSWT +saline group. Sham-operated group in which identical surgical procedures to the BLEO group was performed, except that the saline was administered instead of bleomycin, (iv) iNOSKO+saline group. Identical to iNOSWT +saline group, except for the use of iNOSKO mice. GW274150 group. Same as the iNOSWT + BLEO group but iNOSWT mice were administered with GW274150 (5 mg/kg) i.p. bolus 30 min after the administration of BLEO and every 24 h starting from day 1 (N = 30), (v) Sham+ GW274150 group. Identical to iNOSWT +saline group, except for the administration of GW274150 (5 mg/kg) i.p. bolus 30 min after the administration of BLEO and every 24 h starting from day 1 (N = 30). In another sets of studies, following bleomycin administration, the various groups of mice (N = 20 for each group) were observed for 15 days in order to determine survival differences. The dose of GW274150 used here has previously been reported by us to reduce the tissue injury caused by inflammation [26].
Induction of lung injury by bleomycin
Mice received a single intratracheal instillation of saline (0.9%) or saline containing bleomycin sulphate (1 mg/kg body weight) in a volume of 50 μl and were killed after 15 days by pentobarbitone overdose.
Measurement of fluid content in lung
The wet lung weight was measured after careful excision of extraneous tissues. The lung was exposed for 48 h at 180°C and the dry weight was measured. Water content was calculated by subtracting dry weight from wet weight.
Histological examination
Lung biopsies were taken 15 days after injection of bleomycin. Lung biopsies were fixed for 1 week in 10% (w/v) PBS-buffered formaldehyde solution at room temperature, dehydrated using graded ethanol and embedded in Paraplast (Sherwood Medical, Mahwah, NJ, USA). After embedding in paraffin, the sections were prepared and stained by H&E or by trichrome stain. All sections were studied using light microscopy (Dialux 22 Leitz). The severity of fibrosis was semi quantitatively assessed according to the method proposed by Ashcroft and co-workers [28]. Briefly, the grade of lung fibrosis was scored on a scale from 0 to 8 by examining section randomly chosen fields per sample at a magnification of ×100. Criteria for grading lung fibrosis were as follows: grade 0, normal lung; grade 1, minimal fibrous thickening of alveolar or bronchiolar walls; grade 3, moderate thickening of walls without obvious damage to lung architecture; grade 5, increased fibrosis with definite damage to lung structure and formation of fibrous bands or small fibrous masses; grade 7, severe distortion of structure and large fibrous areas; grade 8, total fibrous obliteration of fields.
Collagen Protein Measurement
Total lung collagen content was measured by means of Sircol Soluble Collagen Assay (Biocolor, Newtownabbey, Northern Ireland), an assay based on a modification of the sirius red method, as recommended by the manufacturer. Briefly, after the sacrifice, mice lungs were explanted and homogenized. Samples were then incubated at 4°C for 2 h and centrifuged at 15,000 × g. Supernatants (20 μl) were diluted 5 times in lysis buffer, added to 1 mL of Sircol Dye Reagent and then mixed for 30 minutes at room temperature in a mechanical shaker. The collagen-dye complex was precipitated by centrifugation at 10000 × g for 10 min. The unbound dye solution was then carefully removed. The precipitated complex was resuspended in 1 mL of alkali reagent. The obtained solution was finally placed in a 96 wells flat bottomed plate and evaluated in a plate reader (absorbance = 540 nm). Obtained values were then compared to the standard curve as recommended to obtain absolute collagen content. Shown data represent the mean collagen content, expressed as μg/μl of lung homogenates (± SE), of at least 4 independent experiments.
Immunohistochemical localization of nitrotyrosine
Tyrosine nitration, an index of the nitrosylation of proteins by peroxynitrite and/or ROS, was determined by immunohistochemistry as previously described [29]. At the end of the experiment, the tissues were fixed in 10% (w/v) PBS-buffered formaldehyde and 8 μm sections were prepared from paraffin embedded tissues. After deparaffinization, endogenous peroxidase was quenched with 0.3% (v/v) hydrogen peroxide in 60% (v/v) methanol for 30 min. The sections were permeablized with 0.1% (w/v) Triton X-100 in PBS for 20 min. Non-specific adsorption was minimized by incubating the section in 2% (v/v) normal goat serum in PBS for 20 min. Endogenous biotin or avidin binding sites were blocked by sequential incubation for 15 min with biotin and avidin (DBA, Milan, Italy), respectively. Sections were incubated overnight with anti-nitrotyrosine polyclonal antibody (1:500 in PBS, v/v). Sections were washed with PBS, and incubated with secondary antibody. Specific labeling was detected with a biotin-conjugated goat anti-rabbit IgG and avidin-biotin peroxidase complex (DBA, Milan, Italy). In order to confirm that the immunoreactions for the nitrotyrosine were specific some sections were also incubated with the primary antibody (anti-nitrotyrosine) in the presence of excess nitrotyrosine (10 mM) to verify the binding specificity. In this situation no positive staining was found in the sections indicating that the immunoreactions were positive in all the experiments carried out.
Myeloperoxidase activity
Myeloperoxidase (MPO) activity, an indicator of polymorphonuclear leukocyte (PMN) accumulation, was determined as previously described [30]. At the specified time following injection of bleomycin, lung tissues were obtained and weighed, each piece homogenized in a solution containing 0.5% (w/v) hexadecyltrimethyl-ammonium bromide dissolved in 10 mM potassium phosphate buffer (pH 7) and centrifuged for 30 min at 20,000 × g at 4°C. An aliquot of the supernatant was then allowed to react with a solution of tetramethylbenzidine (1.6 mM) and 0.1 mM hydrogen peroxide. The rate of change in absorbance was measured spectrophotometrically at 650 nm. MPO activity was defined as the quantity of enzyme degrading 1 μmol of peroxide/min at 37°C and was expressed in milliunits per g of wet tissue.
Thiobarbituric acid-reactant substances measurement
Thiobarbituric acid-reactant substances measurement, which is considered a good indicator of lipid peroxidation, was determined, as previously described [27], in the lung tissues. At the specified time following injection of bleomycin lung tissues were homogenized in 1.15% KCl solution. An aliquot (100 μl) of the homogenate was added to a reaction mixture containing 200 μl of 8.1% SDS, 1500 μl of 20% acetic acid (pH 3.5), 1500 μl of 0.8% thiobarbituric acid and 700 μl distilled water. Samples were then boiled for 1 h at 95°C and centrifuged at 3,000 × g for 10 min. The optical density at 650 nm (OD650) was measured using ELISA microplate reader (SLT- Labinstruments Salzburg, Austria). Thiobarbituric acid-reactant substances were calculated by comparison with OD650 of standard solutions of 1,1,3,3-tetramethoxypropan 99% malondialdehyde bis (dymethyl acetal) 99% (Sigma, Milan). The absorbance of the supernatant was measured by spectrophotometry at 650 nm.
Bronchoalveolar Lavage (BAL)
Seven days after bleomycin or saline solution instillation, mice were euthanized and the trachea was immediately cannulated with an I.V. polyethylene catheter (Neo Delta Ven 2, delta Med, Viadana, Italy) equipped with a 24-gauge needle on a 1 mL syringe. Lungs were lavaged once with 0.5 ml D-PBS (GIBCO, Paisley, U.K.). In >95% of the mice, the recovery volume was over 0.4 ml. The BAL fluid was spun at 800 rpm, the supernatant was removed and the pelleted cells were collected. Total BAL cells were enumerated by counting on a hemocytometer in the presence of trypan blue. Cytospins were prepared from resuspended BAL cells.
Cytospins of BAL cells were made by centrifuging 50,000 cells onto microscope slides using a Shandon Cytospin 3 (Shandon, Astmoore, U.K.). Slides were allowed to air dry and were then stained with Diff-Quick Stain Set (Diff-Quick; Baxter Scientific, Miami, FL). A total of 400 cells were counted from randomly chosen high power microscope fields for each sample. The differential percentage was multiplied by the total leukocyte number per mL to derive the absolute number of monocyte/macrophages, neutrophils, lymphocytes and eosinophils.
TGF-β1 western blot analysis
Immediately after sacrifice, lungs were removed, thoroughly washed, frozen and stored at -80°C until protein extraction. Thawed tissues were washed in PBS and homogenized with an Ultra-Turrax T25 tissue grinder in 400 μl of 10 mM cold Tris homogenization buffer containing 5 mM EDTA, 1% Tryton-X100, 1 mM phenylmethylsulfonylfluoride, 25 μg/ml leupeptin and 0.5% aprotinin (all from Sigma-Aldrich). After homogenization, samples were incubated at 4°C for 2 h, centrifuged for 10 min at 15,000 × g and the supernatant was processed for protein concentration according to the method of Bradford [31]. Samples were diluted in sample buffer and boiled for 5 min. Electrophoresis was performed in 15% sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (40 mA/h) using 60–80 μg of cell proteins per lane. After separation, proteins were transferred onto a nitrocellulose membrane (Hybond ECL, Amersham Biosciences Europe GmbH, Milan, Italy) for 2 h at room temperature using a transblot semidry transfer cell. After blocking, the membranes were incubated with a monoclonal mouse anti-TGF-β1 (0.8 μg/ml; Chemicon, Temecula, CA) overnight at 4 °C. Membranes were then thoroughly washed and incubated with HRP-conjugated secondary antibody. Specific bands were visualized using the SuperSignal chemiluminescent detection system (Pierce Biotechnology Inc., Rockford, IL). The same membranes were washed with a stripping solution containing 0.2 M glycine, 0.1% SDS, 1% Tween-20 and re-blotted with mouse anti-β-actin (1:250; Sigma-Aldrich) overnight at 4°C and processed for signal detection as described above.
TGF-β1 expression was normalized for β-actin and data are expressed as the percent increase of TGF-β1 expression in bleomycin-treated mice vs. control animals.
Materials
Unless otherwise stated, all compounds were obtained from Sigma-Aldrich Company Ltd. (Poole, Dorset, U.K.). All other chemicals were of the highest commercial grade available. All stock solutions were prepared in non-pyrogenic saline (0.9% NaCl; Baxter, Italy, UK).
Statistical evaluation
All values in the figures and text are expressed as mean ± standard error of the mean (SEM) of N observations. For the in vivo studies N represents the number of animals studied. In the experiments involving histology or immunohistochemistry, the figures shown are representative of at least three experiments performed on different experimental days. The results were analyzed by one-way ANOVA followed by a Bonferroni post-hoc test for multiple comparisons. A P-value of less than 0.05 was considered significant. Statistical analysis for survival data was calculated by Fisher's exact probability test. For such analyses, p < 0.05 was considered significant. Mann-Withney U test was used to compare the percent increase of TGF-β1 in iNOS-/- and GW274150-treated animals versus iNOS wild type mice; a p < 0.05 was considered significant.
Results
The development of bleomycin-induced lung injury is attenuated in iNOSKO mice
Histological examination of lung sections revealed significant tissue damage (Fig 1B,B1 Table 1). Thus, when compared to lung sections taken from saline-treated animals (Fig. 1A,A1 Table 1), histological examination of lung sections of iNOSWT mice treated with bleomycin characterized by extensive inflammatory infiltration by neutrophils, lymphocyte and plasma cells extending through the lung epithelial (Fig. 1B,B1 Table 1), fibrosis (Fig. 1B1 Table 1) and granulomas in perivascular region (Fig. 1B1). The absence or inhibition of iNOS in mice (animals with the iNOSKO phenotype or iNOSWT mice treated with GW274150) significantly prevented lung inflammation induced by bleomycin administration (Fig. 1C,C1, D,D1 respectively). Furthermore, the injection of bleomycin in iNOSWT mice elicited an inflammatory response characterized by the accumulation of water in lung as an indicator of fluid content, (Fig. 2) and neutrophils infiltration in the lung tissues (Fig. 3). The absence or inhibition of iNOS in mice (animals with the iNOSKO phenotype or iNOSWT mice treated with GW274150) significantly reduced the fluid content and the neuthrophil infiltration (Figs. 2, 3).
Figure 1 Effect of iNOS inhibition on lung injury. H&E stain: × 250. A: Saline control, normal lung architecture. B: Bleomycin alone in iNOSWT mice, extensive inflammation with inflammatory cells infiltration and fibrosis. C: Bleomycin in iNOSKO mice patchy areas of inflammation with minimal fibrosis. D: Bleomycin in iNOSWT mice plus GW274150 patchy areas of inflammation with minimal fibrosis. Comparable sections of mouse lung stained with trichrome: A1: saline control: normal lung architecture; B1; Bleomycin alone in iNOSWT mice, extensive areas of collagen; C1: Bleomycin in iNOSKO mice minimal collagen; D1: Bleomycin in iNOSWT mice plus GW274150 minimal collagen. Figure is representative of at least 3 experiments performed on different experimental days.
Table 1 Histological Scoring of lung fibrosis
SHAM + vehicle Bleomycin + iNOSWT Bleomycin + iNOSKO Bleomycin + iNOSWT + GW274150
Lung fibrosis score ND 5.1 ± 0.11* 1.5 ± 0.08° 1.4 ± 0.10°
The above parameters were evaluated at 15 days after bleomycin administration. *p < 0.01 versus sham. °p < 0.01 represents significant reduction of the various parameters in the group in which iNOS was inhibited or absent.
Figure 2 Effect of genetic or pharmacological inhibition of iNOS on edema in the lung. The injection of bleomycin in iNOSWT mice elicited an inflammatory response characterized by the accumulation of water in lung as an indicator of edema. The genetic or pharmacological inhibition of iNOS significantly reduced the edema formation. Data are means ± s.e. means from 10 mice for each group. *p < 0.01 versus sham. °p < 0.01 represents significant reduction of the various parameters in the group in which iNOS was inhibited or absent.
Figure 3 Effect of genetic or pharmacological inhibition of iNOS on myeloperoxidase activity in the lung. Myeloperoxidase (MPO) activity in the lungs of bleomycin-treated iNOSWT mice were significantly increased in comparison to sham-operated mice. The genetic or pharmacological inhibition of iNOS significantly reduced the bleomycin-induced increase in MPO activity. Data are means ± s.e. means from 10 mice for each group. *p < 0.01 versus sham. °p < 0.01 represents significant reduction of the various parameters in the group in which iNOS was inhibited or absent.
iNOSKO and GW274150 treated mice show a reduced collagen production in response to bleomycin
iNOSWT mice exposed to bleomycin showed a significant increase of lung collagen content after 7 days if compared to sham mice: from 1.23 ± 0.39 μg/μl to 3.62 ± 0.33 μg/μl, p < 0.001. iNOSKO and GW274150 treated iNOSWT mice that underwent bleomycin tracheal instillation did not show such an increase of lung collagen content (Fig 4). These animals, when exposed to bleomycin, showed a reduced collagen lung deposition if compared to iNOSWT mice: 0.94 ± 0.12 and 2.18 ± 0.17 μg/μl vs. 3.62 ± 0.33 μg/μl, (p < 0.001 and p < 0.01 respectively, Fig. 4)
Figure 4 Data represent the mean collagen content, expressed as μg/μl of lung homogenates (+/- SE), of at least 4 independent experiments. *p < 0.01 versus sham. °p < 0.01 represents significant reduction of the various parameters in the group in which iNOS was inhibited or absent.
Nitrotyrosine formation and lipid peroxidation
Immunohistochemical analysis of lung sections obtained from bleomycin-treated iNOSWT mice revealed a positive staining for iNOS manly localized in plasma cell and lymphocytes (Fig. 5B). In contrast, no staining for iNOS was found in the lungs of bleomycin-treated iNOSKO mice (Fig. 5C) and in the lung from bleomycin-injected iNOSWT mice treated with GW274150 (Fig. 5D). Staining was absent in lung tissue obtained from the sham group (Fig. 5A). All iNOSWT mice, who were treated with bleomycin, exhibited a substantial increase in the lung thiobarbituric acid-reactant substances levels (index of lipid peroxidation) (Fig. 6). The absence or inhibition of iNOS in mice (animals with the iNOSKO phenotype or iNOSWT mice treated with GW274150) significantly attenuate the increase in thiobarbituric acid-reactant substances lung levels caused by bleomycin in the lung (Fig. 6). There was no increase in lung thiobarbituric acid-reactant substances level in sham-operated animals (Fig 6).
Figure 5 Immunohistochemical localization of nitrotyrosine in the lung. No positive staining was observed in the lung section for sham-treated mice (A). After bleomycin injection in iNOSWT mice, positive staining for nitrotyrosine (B) was localized mainly in nuclei of inflammatory cells. There was a marked reduction in the immunostaining in the lungs of bleomycin-treated iNOSKO mice (C) and in the lungs of bleomycin-treated iNOSWT mice which received GW274150 (D). Original magnification: 150×. This figure is representative of at least 3 experiments performed on different experimental days.
Figure 6 Effect of genetic or pharmacological inhibition of iNOS on Thiobarbituric acid-reactant substances, a good indicator of lipid peroxidation, in the lung. Thiobarbituric acid-reactant substances levels in the lungs of bleomycin-treated iNOSWT mice were significantly increased in comparison to sham-operated mice. The genetic or pharmacological inhibition of iNOS significantly reduced the bleomycin-induced increase in Thiobarbituric acid-reactant substances levels. Data are means ± s.e. means from 10 mice for each group. *p < 0.01 versus sham. °p < 0.01 represents significant reduction of the various parameters in the group in which iNOS was inhibited or absent.
Effect of INOS inhibition on the on changes of body weight and survival rate
In iNOSWT mice, the severe lung injury caused by bleomycin administration was associated with a significant loss in body weight (Fig. 7). The absence or inhibition of iNOS in mice (animals with the iNOSKO phenotype or iNOSWT mice treated with GW274150) significantly attenuate the loss in body weight (Fig. 7). The survival of animals was monitored for 15 days. Bleomycin-treated iNOSWT mice developed severe lung injury and 60% of these animals died within 15 days after bleomycin administration (Fig 8). In contrast, none of the iNOSKO mice as well as the iNOSWT which had been treated with GW274150 died (Fig. 8).
Figure 7 Effect of genetic or pharmacological inhibition of iNOS on body weight. Body weight was recorded immediately before bleomycin administration and daily for all the experimental period. The genetic or pharmacological inhibition of iNOS significantly prevents the loss of body weight. Data are means ± s.e. means from 10 mice for each group. *p < 0.01 represents significant reduction of the various parameters in the group in which iNOS was inhibited or absent.
Figure 8 Effect of genetic or pharmacological inhibition of iNOS on bleomycin-induced mortality. Survival is significantly improved in iNOSKO and iNOSWT GW274150-treated mice in comparison to the high mortality rate of the bleomycin-treated iNOSWT mice. Data are means ± s.e. means from 20 mice for each group. *p < 0.01 represents significant reduction of the various parameters in the group in which iNOS was inhibited or absent.
Bronchoalveolar Lavage
Instillation of saline solution produced no significant increase in leukocyte numbers in BAL fluid of iNOSKO and GW274150 treated iNOSWT mice compared to the sham wild type group (2.48 ± 0.45 and 1.77 ± 0.24 vs. 1.69 ± 0.37 cells × 105/mL ± SE). Bleomycin instillation in iNOSWT mice produced a significant increase of inflammatory cells compared to sham iNOSWT mice (8.93 ± 0.53 vs 1.69 ± 0.37 cells × 105/mL ± SE, p < 0.001) (Fig. 9). iNOSKO and GW274150 treated iNOSWT mice that underwent to bleomycin tracheal instillation did not show such an increase of BAL total cellularity as compared to bleomycin iNOSWT mice group (2.05 ± 0.35 and 2.83 ± 0.41 vs. 8.92 ± 0.53 cells × 105/mL ± SE, p < 0.001). Differential cell counts showed a similar profile across all of the sham groups. In bleomycin treated iNOSWT mice it was evident an increase of monocytes (6.48 ± 0.39 vs. 1.57 ± 0.36 cells × 105/mL ± SE, p < 0.001), lymphocytes (1.49 ± 0.20 vs. 0.19 ± 0.08 cells × 105/mL ± SE, p < 0.001) and neutrophils (0.94 ± 0.20 vs. 0.10 ± 0.04 cells × 105/mL ± SE, p < 0.001) if compared to sham wild type mice. iNOSKO and GW274150 treated iNOSWT mice that underwent to bleomycin tracheal instillation did not show any increase of BAL inflammatory cells (Fig. 9). In these mice monocytes (1.55 ± 0.31 and 2.40 ± 0.41 vs. 6.48 ± 0.39 cells × 105/mL ± SE, p < 0.001), lymphocytes (0.11 ± 0.03 and 0.27 ± 0.1 vs. 1.49 ± 0.2 cells × 105/mL ± SE, p < 0.001) and neutrophils (0.38 ± 0.18 and 0.16 ± 0.07 vs. 0.94 ± 0.20 cells × 105/mL ± SE, p < 0.05 and p < 0.001 respectively) were significantly reduced compared to bleomycin treated iNOSWT group (Fig. 9). Eosinophils did not show any statistically significant difference among all groups.
Figure 9 Effect of genetic or pharmacological inhibition of iNOS on bleomycin-induced total and differential cellularity of bronchoalveolar lavage (BAL). (A) Total BAL cellularity for sham and bleomycin treated mice. (B) Differential cells counts for macrophages, lymphocytes, neutrophils and eosinophils per milliliter of BAL fluid are shown. Data, expressed as means ± s.e., are representative of 10 mice for each group. * p < 0.001 vs. sham, °p < 0.001 vs. bleomycin treated iNOSWT, # p < 0.001 vs. sham, ● p < 0.001 vs. bleomycin treated iNOSWT, ◆ p < 0.05 vs. bleomycin treated iNOSWT.
TGF-β1 western blot analysis
TGF-β1 was expressed in all sham groups undergoing intra-tracheal saline instillation as detected by western blot analysis (data not shown). Exposure of iNOSWT mice to bleomycin produced a remarkable increase of TGF-β1 expression (247.9 + 34% of control). In contrast, iNOSKO and GW274150-treated iNOSWT mice subjected to intra-tracheal bleomycin instillation exhibited only a slight increase of TGF-β1 expression, 129.4 ± + 21.4% and 120.1 ± 19.4% for iNOSKO and GW274150-treated iNOSWT mice, respectively, that yielded statistical significance when compared to iNOSWT (p < 0.05 and p < 0.001) (Fig. 10).
Figure 10 Western blot analysis of TGF-β1 expression in wild type (WT), iNOS-/-(KO) and GW274150-treated (GW) mice exposed to vehicle (C) or bleomycin (bleo) for 7 days. Representative blots are reported in a. In b, data, normalized for β-actin expression, represent mean the bleomycin induced TGF-β1 percent increase of control groups (+/- SE) of 7 independent experiments. *p < 0.05 and **p < 0.001 vs iNOSWT mice.
Discussion
Pulmonary fibrosis is a common response to various insults to the lung and it is the end-point of a numerous and heterogeneous group of disorders known as interstitial lung diseases (ILD) that are characterized by chronic inflammation and progressive fibrosis of the pulmonary interstitium: alveolar walls (including epithelial cells and capillaries), septae, and the perivascular, perilymphatic, and peribronchiolar connective tissues [32]. While the pathogenesis is incompletely understood, a growing body of evidence suggests two different pathogenic routes for developing pulmonary fibrosis. The inflammatory pathway, where a shift to the so-called T-helper 2 type cytokine networks is critical, and the epithelial pathway represented by idiopathic pulmonary fibrosis, by far the most aggressive ILD. Both routes may trigger a number of cytokines/growth factors inducing fibroblast migration/proliferation and phenotype change to myofibroblasts, with a consequent accumulation of extracellular matrix [33]. In addition, various evidences have point out an important role for IL-6 and IL-11 in the pulmonary fibrosis [34,35].
The common pathologic features in ILD, include the fibrosis of the interstitium, involve collagen, elastic and smooth muscle elements, architectural remodeling and chronic inflammation of the interstitium (ie, variable increases in lymphocytes, neutrophils, plasma cells, macrophages, eosinophils, and mast cells), hyperplasia of type II cells and hyperplasia of endothelial cells [32]. Intratracheal instillation of the antitumour agent BLM is the most commonly used animal model for pulmonary fibrosis [36]. The bleomycin-oxygen complex is thought to bind to DNA and lead to the efficient cleavage of the phosphodiester-deoxyribose backbone and the generation of ROS [37] In the presence of oxygen and a reducing agent in fact, the ferrous ion-BLM complex becomes activated and functions mechanically as a ferrous oxidase, transferring electrons from ferrous ion to molecular oxygen to produce ROS that cause scission of DNA [38,39]. Therefore, earlier reports [40,41] point out that the pathogenesis of BLM-induced fibrosis, at least in part, is mediated through the generation of reactive oxygen species (ROS) which cause the peroxidation of membrane lipids and DNA damage. If, that perspective is true, then antioxidant therapy may prevent the lung fibrosis caused by BLM and may prevent other diseases related with interstitial pulmonary fibrosis. Because BLM administration results in increased lipid peroxidation (LPO) and alters activities of antioxidant enzymes in bronchoalveolar lavage fluids (BALFs) and lung tissue [42,43], in previous studies [44,45] some natural or synthetic antioxidants have been used to protect against BLM oxidative lung toxicity both in vivo and also in vitro. In addition to ROS, an overproduction of nitric oxide (NO) due to the expression of the inducible isoform of NO synthase (iNOS) also plays important role in various models of inflammation [5,9,46]. Pharmacological inhibitors of NOS, and also ablation of the gene for iNOS has been shown to reduce the development of the inflammatory response [47-49]. In addition, various studies have point out that iNOS play an important role in the pulmomary fibrosis induced by bleomycin [50-52]. Therefore, some study have also demonstrated that non-selective and partial selective iNOS inhibitor like aminoguanidine exert beneficial effect against lung injury induced by bleomycin [53-57]. However, frequently a misunderstanding is the definition of an inhibitor as selective for, e.g. iNOS versus eNOS, and then ignoring its non-selectivity for nNOS or completely distinct enzyme targets. An interesting example is aminoguanidine that is a partially selective for iNOS versus eNOS [58], while the selectivity over nNOS is minimal. Moreover it has a wide range of other effects, inhibiting advanced glycosylation end-product formation, diamine oxidase and polyamine metabolism [59,60], catalase [61] and having anti-oxidant effects [62,63]. For this reason aminoguanidine should not be described as a selective inhibitor.
In contrast, GW274150 is a novel, potent and selective inhibitor of iNOS activity and previous studies have demonstrated its protective effect in organ injury in hemorrhagic shock, in renal ischemia and reperfusion and in a model of collagen-induced arthritis [24-26]. This study provides the first evidence of the protective role of GW274150 in an experimental model of lung fibrosis. Here we demonstrate that the lack of iNOS gene as well as the pharmacological inhibition of iNOS (GW274150 treatment) reduces: i) the development of bleomycin-induced lung injury, (ii) the infiltration of the lung with inflammatory cells, (iii) the degree of nitrosative stress in the lung and (iv) the mortality rate. All of these findings support the view that NO plays an important role in the degree of inflammation and lung fibrosis caused by bleomycin in the mice. We report in the present study a reduction of the tissue damage in the lung of bleomycin-treated iNOSKO mice as well as bleomycin-treated iNOSWT mice which received the treatment with GW274150.
Two of the most recognized fibrosis markers, lung collagen deposition and TGF-β1 expression, were significantly reduced in these animals. Evidence from animal models and human studies suggests that TGF-β1 plays a central role in a variety of fibroproliferative disorders, including pulmonary fibrosis. TGF-β1 plays a critical role in the pathogenesis of lung fibrosis through stimulation of collagen and fibronectin production in fibroblasts [64], as well as through inhibition of biosynthesis of proteases that degrade the extracellular matrix [65]. TGF-β1 promotes wound healing [66] and its presence has been shown to be increased in bleomycin -induced lung fibrosis [67,68] and particularly in lung macrophages [69]. In addition it has been shown that during the course of pulmonary fibrosis the increase of TGF-β1 mRNA precede the increase of type I and type III procollagen mRNAs The secretion of biologically active TGF-β1 by alveolar macrophages is transiently elevated in bleomycin -induced pulmonary inflammation, whereas latent (L)-TGF-β1 secretion remains elevated for a prolonged length of time and it is probable that the extent of inflammation and fibrosis in this model depends on the quantity of active TGF-β1 available [70].
In BAL of iNOSWT animals, that underwent bleomycin instillation we observed, a strong increase of inflammatory cells such as macrophages and neutrophils. This could justify, at least in part, the significantly higher TGF-β1 expression and the increased lung collagen content in these mice. Bleomycin-treated iNOSKO mice as well as bleomycin-treated iNOSWT mice which received the treatment with GW274150 showed both significantly reduced TGF-β1 expression as well as lung collagen deposition, together with a significantly reduced inflammatory cells presence in the BAL. Furthermore we report in the present study in the lung tissue of bleomycin-treated iNOSKO mice as well as bleomycin-treated iNOSWT mice which received the treatment with GW274150 a significant reduction of leukocyte infiltration as assessed by the specific granulocyte enzyme MPO. Neutrophils recruited into the tissue can contribute to tissue destruction by the production of reactive oxygen metabolites, granule enzymes, and cytokines that further amplify the inflammatory response by their effects on macrophages and lymphocytes [71]. Furthermore, we found that the tissue damage induced by bleomycin in vehicle-treated mice was associated with an intense immunostaining of nitrotyrosine formation also suggesting that a structural alteration of the lung had occurred, most probably due to the formation of highly reactive nitrogen-derivatives. Recent evidence indicates, that bleomycin is a well-known cause of intracellular oxidative stress, several findings in this study suggest that extracellular oxidative stress may also play a role in the pathogenesis of bleomycin-induced lung injury [72,73]. Therefore, in this study we clearly demonstrate that the genetic and pharmacological inhibition of iNOS prevent the formation of peroxynitrite. Nitrotyrosine formation, along with its detection by immunostaining, was initially proposed as a relatively specific marker for the detection of the endogenous formation "footprint" of peroxynitrite [74]. There is, however, recent evidence that certain other reactions can also induce tyrosine nitration; e.g., the reaction of nitrite with hypoclorous acid and the reaction of myeloperoxidase with hydrogen peroxide can lead to the formation of nitrotyrosine [75]. Increased nitrotyrosine staining is considered, therefore, as an indication of "increased nitrative stress" rather than a specific marker of the generation of peroxynitrite. From the present data, we cannot determine the mechanism of tyrosine nitration: inhibition of NOS by NOS inhibitor would inhibit both NO formation (and thus, reduce the generation of peroxynitrite) as well as it would suppress nitrite formation (and thereby attenuate the peroxidase dependent mechanisms of tyrosine nitration). Nevertheless, we can certainly conclude from the current data that the absence of iNOS significantly reduced tyrosine nitration in vivo.
Conclusion
In conclusion, this study demonstrates that the degree of inflammation and fibrosis caused by injection of bleomycin is significantly attenuated in iNOSKO mice as well as in the iNOSWT mice treated with GW274150. These findings support the view that the induction of iNOS contributes to the extension of inflammation in the model of bleomycin-induced lung fibrosis used here. Finally, we provide the first evidence that GW274150 causes a substantial reduction of lung fibrosis in the mice and our findings suggest that interventions, which may reduce the generation or the effects of iNOS, may be useful in conditions associated with local or systemic inflammation.
Acknowledgements
The authors would like to thank Giovanni Pergolizzi and Carmelo La Spada for their excellent technical assistance during this study, Mrs Caterina Cutrona for secretarial assistance and Miss Valentina Malvagni for editorial assistance with the manuscript.
==== Refs
Gross TJ Hunninghake GW Idiopathic pulmonary fibrosis N Engl J Med 2001 345 517 25 11519507 10.1056/NEJMra003200
Crouch E Pathobiology of pulmonary fibrosis Am J Physiol 1990 259 L159 84 2221080
Crystal RG Bitterman PB Rennard SI Hance AJ Keogh BA Interstitial lung diseases of unknown cause: disorders characterized by chronic inflammation of the lower respiratory tract N Engl J Med 1984 310 154 66 6361560
Sleijfer S Bleomycin-induced pneumonitis Chest 2001 120 617 24 11502668 10.1378/chest.120.2.617
Nathan C Nitric oxide as a secretary product of mammalian cells FASEB J 1992 6 3051 3064 1381691
Dinerman J Molecular mechanism of nitric oxide production. Potential relevance to cardiovascolar disease Circ Res 1993 73 217 222 7687202
Szabò C Alterations in the production of NO in various forms of circulatory shock New Horiz 1995 3 3 32
Southan G Szabò C Selective pharmacological inhibition of distinct NOS isoforms Bioch Pharm 1996 51 383 394 10.1016/0006-2952(95)02099-3
Moncada S Palmer RMJ NO: physiology, patophysiology and pharmacology Pharm Rev 1991 43 109 142 1852778
Moncada S Higgs A The L-arginine NO pathway New Eng J of Med 2002 329 2002 2012 7504210 10.1056/NEJM199312303292706
Cuzzocrea S Effect of inhibitors of nitric oxide in animal models and future directions for therapy in inflammatory disorders Current Med Chem Anti-inflammatory & Anti-allergy agent 2004 3 261 270 10.2174/1568014043355302
Crow JP Beckman JS Reactions between nitric oxide, superoxide, and peroxynitrite: footprints of peroxynitrite in vivo Adv Pharmacol 1995 34 17 43 8562432
Pryor WA Squadrito GL The chemistry of peroxinitrite: a product from reaction of NO with superoxide Am J Physiol 1995 268 699 722
Beckman JS Beckman TW Chen J Marshall PA Freeman BA Apparent hydroxil radical production by peroxynitrite: implications for endothelial injury from NO and superoxide Proc Natl Acad Sci USA 1990 87 1620 1624 2154753
Cuzzocrea S Riley DP Caputi AP Salvemini D Antioxidant therapy: a new pharmacological approach in shock, inflammation, and ischemia/reperfusion injury Pharmacol Rev 2001 53 135 159 11171943
Rao AV Balachandran B Role of oxidative stress and antioxidants in neurodegenerative diseases Nutr Neurosci 2002 5 291 309 12385592 10.1080/1028415021000033767
Iuliano L The oxidant stress hypothesis of atherogenesis Lipids 2001 36 41 44
Yamada T Grisham MB Role of neutrophil-derived oxidants in the pathogenesis of intestinal inflammation Klin Wochenschr 1991 69 988 994 1665888 10.1007/BF01645144
Inoue S Kawanishi S Oxidative DNA damage induced by simultaneous generation of nitric oxide and superoxide FEBS Lett 1995 371 86 88 7664890 10.1016/0014-5793(95)00873-8
Salgo MG Peroxinitrite causes apoptosis in rat thymorytes Bioch Bioph Res Commun 1995 215 1111 1118 10.1006/bbrc.1995.2578
Szabò C Role of poly(ADP-ribose)synthetase in inflammation Eur J Pharmacol 1998 350 1 19 9683009 10.1016/S0014-2999(98)00249-0
Alderton WK Cooper CE Knowles RG Nitric oxide synthases: structure, function and inhibition Biochem J 2001 357 593 615 11463332 10.1042/0264-6021:3570593
Young RJ Beams RM Carter K Clark HA Coe DM Chambers CL Davies PI Dawson J Drysdale MJ Franzman KW French C Hodgson ST Hodson HF Kleanthous S Rider P Sanders D Sawyer DA Scott KJ Shearer BG Stocker R Smith S Tackley MC Knowles RG Inhibition of inducible nitric oxide synthase by acetamidine derivatives of hetero-substituted lysine and homolysine Bioorg Med Chem Lett 2000 10 597 600 10741561 10.1016/S0960-894X(00)00055-X
Mcdonald MC Izumi M Cuzzocrea S Thiemermann C A novel, potent and selective inhibitor of the activity of inducible nitric oxide synthase (GW274150) reduces the organ injury in hemorrhagic shock J Physiol Pharmacol 2002 53 555 569 12512692
Chatterjee PK Patel NS Sivarajah A Kvale EO Dugo L Cuzzocrea S Brown PA Stewart KN Mota-Filipe H Britti D GW274150, a potent and highly selective inhibitor of iNOS, reduces experimental renal ischemia/reperfusion injury Kidney Int 2003 63 853 865 12631066 10.1046/j.1523-1755.2003.00802.x
Cuzzocrea S Chatterjee PK Mazzon E Mcdonald MC Dugo L Di Paola R Serraino I Britti D Caputi AP Thiemermann C Beneficial effects of GW274150, a novel, potent and selective inhibitor of iNOS activity, in a rodent model of collagen-induced arthritis Eur J Pharmacol 2002 453 119 129 12393067 10.1016/S0014-2999(02)02338-5
Dugo L Marzocco S Mazzon E Di Paola R Genovese T Caputi AP Cuzzocrea S Effects of GW274150, a novel and selective inhibitor of iNOS activity, in acute lung inflammation Br J Pharmacol 2004 141 6 979 87 14769784 10.1038/sj.bjp.0705683
Ashcroft T Simpson JM Timbrell V Simple method of estimating severity of pulmonary fibrosis on a numerical scale J Clin Pathol 1988 41 467 470 3366935
Cuzzocrea S Ianaro A Wayman NS Mazzon E Pisano B Dugo L Serraino I Di Paola R Chatterjee PK Di Rosa M The cyclopentenone prostaglandin 15-deoxy-delta-(12,14)-PGJ2 attenuates the development of colon injury caused by dinitrobenzene sulphonic acid in the rat Br J Pharmacol 2003 138 678 88 12598422 10.1038/sj.bjp.0705077
Mullane KM Kraemer R Smith B Myeloperoxidase activity as a quantitative assessment of neutrophil infiltration into ischemic myocardium J Pharmacol Methods 1985 14 157 67 2997548 10.1016/0160-5402(85)90029-4
Bradford M A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of proteindye binding Anal Biochem 1976 72 248 54 942051
Green FH Overview of pulmonary fibrosis Chest 2002 122 334S 339S 12475811 10.1378/chest.122.6_suppl.334S
Pardo A Selman M Molecular mechanisms of pulmonary fibrosis Front Biosci 2002 7 d1743 61 12133818
Hardie WD Le Cras TD Jiang K Tichelaar JW Azhar M Korfhagen TR Conditional expression of transforming growth factor-alpha in adult mouse lung causes pulmonary fibrosis Am J Physiol Lung Cell Mol Physiol 2004 286 L741 L749 14660483 10.1152/ajplung.00208.2003
Tang W Geba GP Zheng T Ray P Homer RJ Kuhn C 3rdFlavell RA Elias JA Targeted expression of IL-11 in the murine airway causes lymphocytic inflammation, bronchial remodeling, and airways obstruction J Clin Invest 1996 98 2845 53 8981933
Chandler DB Hyde DM Giri SN Morphometric estimates of infiltrative cellular changes during the development of bleomycin-induced pulmonary fibrosis in hamsters Am J Pathol 1983 112 170 7 6192723
Hecht SM Bleomycin: new perspectives on the mechanism of action J Nat Prod 2000 63 158 168 10650103 10.1021/np990549f
Burger RM Projan SJ Horwitz SB Peisach J The DNA cleavage mechanism of iron-bleomycin. Kinetic resolution of strand scission from base propenal release J Biol Chem 1986 261 15955 15959 2430964
Arslan SO Zerin M Vural H Coskun A The effect of melatonin on bleomycin-induced pulmonary fibrosis in rats J Pineal Res 2002 32 21 5 11841596 10.1034/j.1600-079x.2002.10796.x
Slosman DO Costabella PM Roth M Werlen G Polla BS Bleomycin primes monocytes-macrophages for superoxide production Eur Respir J 1990 3 772 8 1702065
Goodman MT Hernandez B Wilkens LR Lee J Le Marchand L Liu LQ Franke AA Kucuk O Hsu TC Effects of beta-carotene and alpha-tocopherol on bleomycin-induced chromosomal damage Cancer Epidemiol Biomarkers Prev 1998 7 113 7 9488585
Giri SN Chen ZL Younker WR Schiedt MJ Effects of intratracheal administration of bleomycin on GSH-shuttle enzymes, catalase, lipid peroxidation, and collagen content in the lungs of hamsters Toxicol Appl Pharmacol 1983 71 132 41 6195772 10.1016/0041-008X(83)90052-2
Karam H Hurbain-Kosmath I HousSET B Antioxidant activity in alveolar epithelial type 2 cells of rats during the development of bleomycin injury Cell Biol Toxicol 1998 14 13 22 9538941 10.1023/A:1007460320109
Ikezaki S Nishikawa A Enami T Furukawa F Imazawa T Uneyama C Fukushima S TakahASHI M Inhibitory effects of the dietary antioxidants butylated hydroxyanisole and butylated hydroxytoluene on bronchioloalveolar cell proliferation during the bleomycin-induced pulmonary fibrosing process in hamsters Food Chem Toxicol 1996 34 327 35 8641658 10.1016/0278-6915(96)00127-5
Venkatesan N Punithavathi V Chandrakasan G Curcumin protects bleomycin-induced lung injury in rats Life Sci 1997 61 PL51 8 9250724 10.1016/S0024-3205(97)00443-8
Cuzzocrea S Zingarelli B Gilard E Hake P Salzman AL Szabó C Anti-inflammatory effects of mercaptoethylguanidine, a combined inhibitor of nitric oxide synthase and peroxynitrite scavenger, in carrageenan-induced models of inflammation Free Rad Biol Med 1998 24 450 459 9438558 10.1016/S0891-5849(97)00280-3
Salvemini D Wang ZQ Wyatt P Bourdon DM Marino MH Manning PT Currie MG Nitric oxide: a key mediator in the early and late phase of carrageenan-induced rat paw inflammation Br J Pharmacol 1996 118 829 838 8799551
Tracey WR Nakane M Kuk J Budzik G Klinghofer V Harris R Carter G The nitric oxide synthase inhibitor, L-NG-monomethylarginine, reduces carrageenan-induced pleurisy in the rat J Pharmacol Exp Ther 1995 273 1295 1299 7540689
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 11 7539113 10.1038/375408a0
El-Khatib AS Possible modulatory role of nitric oxide in lung toxicity induced in rats by chronic administration of bleomycin Chemotherap 2002 48 244 51 10.1159/000066767
Jang AS Lee JU Choi IS Park KO Lee JH Park SW Park CS Expression of nitric oxide synthase, aquaporin 1 and aquaporin 5 in rat after bleomycin inhalation Intensive Care Med 2004 30 489 95 14767587 10.1007/s00134-003-2129-9
Davis DW Weidner DA Holian A McConkey DJ Nitric oxide-dependent activation of p53 suppresses bleomycin-induced apoptosis in the lung J Exp Med 2000 192 857 69 10993916 10.1084/jem.192.6.857
de Rezende MC Martinez JA Capelozzi VL Simoes MJ Beppu OS Protective effect of aminoguanidine in a murine model of pulmonary fibrosis induced by bleomycin Fundam Clin Pharmacol 2000 14 561 7 11206706
Hu J Xu Q Li B The effect of aminoguanidine, a nitric oxide synthase inhibitor, on bleomycin-induced lung injury in rats Zhonghua Jie He He Hu Xi Za Zhi 1999 22 51 3 11812359
Yildirim Z Turkoz Y Kotuk M Armutcu F Gurel A Iraz M Ozen S Aydogdu I Akyol O Effects of aminoguanidine and antioxidant erdosteine on bleomycin-induced lung fibrosis in rats Nitric Oxide 2004 11 156 65 15491848 10.1016/j.niox.2004.07.011
Giri SN Biring I Nguyen T Wang Q Hyde DM Abrogation of bleomycin-induced lung fibrosis by nitric oxide synthase inhibitor, aminoguanidine in mice Nitric Oxide 2002 7 109 18 12223180 10.1016/S1089-8603(02)00102-7
Chen XL Huang SS Li WB Wang DH Wang XL Inhibitory effect of aminoguanidine on bleomycin-induced pulmonary toxicity in rat Acta Pharmacol Sin 2001 22 711 5 11749843
Alderton WK Cooper CE Knowles RG NO Synthases: Structure, Function and Inhibition Biochem J 2001 357 593 615 11463332 10.1042/0264-6021:3570593
Bieganski T Kusche J Lorenz W Hesterberg R Stahlknecht CD Feussner KD Distribution and Properties of Human Intestinal Diamine Oxidase and Its Relevance for the Histamine Catabolism Biochim Biophys Acta 1983 756 196 203 6403048
Nilsson BO Kockum I Rosengren E Inhibition of Diamine Oxidase Promotes Uptake of Putrescine From Rat Small Intestine Inflamm Res 1996 45 513 518 8912017
Ou P Wolff SP Aminoguanidine: a Drug Proposed for Prophylaxis in Diabetes Inhibits Catalase and Generates Hydrogen Peroxide in Vitro Biochem Pharmacol 1993 46 1139 1144 8216363 10.1016/0006-2952(93)90461-5
Giardino I Fard AK Hatchell DL Brownlee M Aminoguanidine Inhibits Reactive Oxygen Species Formation, Lipid Peroxidation, and Oxidant-Induced Apoptosis Diabetes 1998 47 1114 1120 9648836
Yildiz G Demiryurek AT Sahin-Erdemli I Kanzik I Comparison of Antioxidant Activities of Aminoguanidine, Methylguanidine and Guanidine by Luminol-Enhanced Chemiluminescence Br J Pharmacol 1998 124 905 910 9692775 10.1038/sj.bjp.0701924
Fine A Goldstein RH The effect of transforming growth factor on cell proliferation and collagen by fibroblasts J Biol Chem 1987 262 3897 3902 3493244
Sporn MB Roberts AB Wakefield LM De Crombrugghe B Some recent advances in the chemistry and biology of transforming growth factor-beta J Cell Biol 1987 105 1039 1045 3308901 10.1083/jcb.105.3.1039
Pierce GF Mustoe TA Lingelbach J Masakowski VR Gramates P Deuel TF Transforming growth factor reverses the glucocorticoid-induced wound-healing deficit in rats: possible regulation in macrophages by platelet-derived growth factor Proc Natl Acad Sci USA 1989 86 2229 2233 2928327
Hoyt DG Lazo JS Alterations in pulmonary mRNA encoding procollagens, fibronectin and transforming growth factor precede bleomycin induced pulmonary fibrosis in mice J Pharmacol Exp Ther 1988 246 765 771 2457084
Raghow R Irish P Kang AH Coordinate regulation of transforming growth factor gene expression and cell proliferation in hamster lungs undergoing bleomycin-induced pulmonary fibrosis J Clin Invest 1989 84 1836 1842 2480367
Khalil N Bereznay O Sporn M Greenberg AH Macrophage production of transforming growth factor and fibroblast collagen synthesis in chronic pulmonary inflammation J Exp Med 1989 170 727 737 2475572 10.1084/jem.170.3.727
Khalil N Corne S Whitman C Yacyshyn H Plasmin regulates the activation of cell-associated latent TGF-1 secreted by rat alveolar macrophages after in vivo bleomycin injury Am J Respir Cell Mol Biol 1996 15 252 259 8703482
Chatham WW Swaim R Frohsin H JrHeck LW Miller EJ Blackburn WD Jr Degradation of human articular cartilage by neutrophils in synovial fluid Arthritis Rheum 1993 36 51 8 8424836
Bowler RP Nicks M Warnick K Crapo JD Role of extracellular superoxide dismutase in bleomycin-induced pulmonary fibrosis Am J Physiol Lung Cell Mol Physiol 2002 282 L719 26 11880297
Chen XL Li WB Zhou AM Ai J Huang SS Role of endogenous peroxynitrite in pulmonary injury and fibrosis induced by bleomycin A5 in rats Acta Pharmacol Sin 2003 24 697 702 12852838
Beckman JS Oxidative damage and tyrosine nitration from peroxynitrite Chem Res Toxicol 1996 9 836 44 8828918 10.1021/tx9501445
Eiserich JP Hristova M Cross CE Jones AD Freeman BA Halliwell B Van der Vliet A Formation of nitric oxide-derived inflammatory oxidants by myeloperoxidase in neutrophils Nature 1998 391 393 397 9450756 10.1038/34923
| 15955252 | PMC1177992 | CC BY | 2021-01-04 16:23:26 | no | Respir Res. 2005 Jun 14; 6(1):58 | utf-8 | Respir Res | 2,005 | 10.1186/1465-9921-6-58 | oa_comm |
==== Front
Respir ResRespiratory Research1465-99211465-993XBioMed Central London 1465-9921-6-661598751910.1186/1465-9921-6-66ResearchGreen tea polyphenol extract attenuates lung injury in experimental model of carrageenan-induced pleurisy in mice Di Paola Rosanna [email protected] Emanuela [email protected]à Carmelo [email protected] Tiziana [email protected] Marta [email protected] Raffaela [email protected] Hisanory [email protected] Salvatore [email protected] Department of Clinical and Experimental Medicine and Pharmacology, Torre Biologica, Policlinico Universitario, Messina, Italy2 Biochemistry Division, Department of Neuroscience and Vision, University of Verona, Verona, Italy2005 29 6 2005 6 1 66 66 21 4 2005 29 6 2005 Copyright © 2005 Di Paola et al; licensee BioMed Central Ltd.2005Di Paola 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.
Here we investigate the effects of the green tea extract in an animal model of acute inflammation, carrageenan-induced pleurisy. We report here that green tea extract (given at 25 mg/kg i.p. bolus 1 h prior to carrageenan), exerts potent anti-inflammatory effects in an animal model of acute inflammation in vivo.
Injection of carrageenan (2%) into the pleural cavity of mice elicited an acute inflammatory response characterized by fluid accumulation in the pleural cavity that contained many neutrophils (PMNs), an infiltration of PMNs in lung tissues and increased production of nitrite/nitrate, tumour necrosis factor alpha. All parameters of inflammation were attenuated by green tea extract treatment. Furthermore, carrageenan induced an up-regulation of the adhesion molecule ICAM-1, as well as nitrotyrosine and poly (ADP-ribose) synthetase (PARS) formation, as determined by immunohistochemical analysis of lung tissues. Staining for the ICAM-1, nitrotyrosine, and PARS was reduced by green tea extract.
Our results clearly demonstrate that treatment with green tea extract exerts a protective effect and offers a novel therapeutic approach for the management of lung injury.
green tea extractcarrageenan-induced pleurisyneutrophils infiltrationlung injury
==== Body
Introduction
The role of oxyradical formation in various forms of inflammation is well established [1] Reactive oxygen species (ROS) are associated with the inflammatory response and frequently they contribute to the tissue damaging effects of inflammatory reactions [2-4]. ROS formation and degradation are key components of the metabolism of aerobic organisms. Certain levels of ROS are required for normal cell functions, but if in surplus, they will cause oxidative stress [5-7]. ROS like superoxide, hydrogen peroxide and lipid hydroperoxides can regulate the activities of several kinases, transcription factors, cell death machinery and proteins such as COX-2 and iNOS [8,9].
Recent data demonstrate that the expression of the inducible isoform of nitric oxide (NO) synthase also plays important pathogenetic roles in various models of inflammation [10-12]. Peroxynitrite, a cytotoxic oxidant species formed from the reaction of NO and superoxide [13], may mediate part of the oxidative injury associated with simultaneous production of NO and oxyradicals. Peroxynitrite formation has been demonstrated in various inflammatory disorders [14,15] and in circulatory shock [16]. Peroxynitrite is a potent oxidant, and therefore it is conceivable that endogenous antioxidant mechanisms counteract its toxicity. In in vitro studies, it has been established that antioxidants such as glutathione, ascorbic acid, and alpha-tocopherol are scavengers of peroxynitrite and inhibitors of its oxidant capacity [17,18].
Green tea – a minimally processed product of the same plant that gives us black and oolong teas – is rich in powerful antioxidant compounds called polyphenols. The polyphenols found in tea are more commonly known as flavanols or catechins and comprise 30–40 percent of the extractable solids of dried green tea leaves. The main catechins in green tea are epicatechin, epicatechin-3-gallate, epigallocatechin, and epigallocatechin-3- gallate (EGCG), with the latter being the highest in concentration. Green tea polyphenols have demonstrated significant antioxidant, anticarcinogenic, anti-inflammatory, thermogenic, probiotic, and antimicrobial properties in numerous human, animal, and in vitro studies [19,20]. Recently it has been showed that green tea polyphenols inhibited tumour necrosis factor-alpha induction in macrophages by attenuating nuclear factor-kβ NF-Kβ) activation [21]. Similarly [22] showed that EGCG inhibits lipopolysaccaride (LPS) – stimulated nitric oxide production and inducibile nitric oxide synthase gene expression in peritoneal macrophages by decreasing NF-κβ activation. These studies provide significant evidence that green tea polyphenols have anti-inflammatory effects. Lung inflammation is characterised by T-cell rich infiltrates and enhanced expression of pro-inflammatory cytokines. The signalling pathway of IFN-γ, secreted by type-1 helper lymphocyte (Th-1), lead to the activation of signal transducer and activator of transcription-1 (STAT-1) [23]. Moreover, IFN-γ is involved in the induction of iNOS and ICAM-1 gene expression by the activation of STAT-1 transcription factor [24,25]. Thus, upregulation of STAT-1 activity could play a key role in the pathogenesis of carrageenan-induced pleurisy. STAT-1 are activated by phosphorylation on conserved tyrosine and serine residues by the Janus kinases (JAKs) and MAP kinase families respectively, which allow the STAT-1 to dimerise and translocate to the nucleus and there by regulate gene expression [23]. Previously, we demonstrated, in some epithelial cell cultures, the inhibitory effect of EGCG on iNOS induction by preventing STAT-1 phosphorylation and activation [26].
In this study we investigated the role of Green tea extract in rodent model carrageenan-induced pleurisy.
This experimental model has been widely used to investigate the pathophysiology of acute inflammation and also to evaluate the efficacy of drugs in inflammation. Injection of carrageenan into the pleural space leads to pleurisy, infiltration by polymorphonuclear leukocytes (PMN), and lung injury. In this study, we have investigated the effect of the green tea on: PMN infiltration [myeloperoxidase (MPO) activity]; STAT-1 activity (by EMSA), up-regulation of ICAM-1 (by immunohistochemistry); the nitration of tyrosine residues (an indicator of the formation of peroxynitrite) (by immunohistochemistry) and lung damage (histology).
Materials and methods
Animals
Mice (4–5 weeks old, 20–22 g) were purchased from Jackson Laboratories (Harlan Nossan, Italy). The animals were housed in a controlled environment and provided with standard rodent chow and water. Animal care was in compliance with Italian regulations on protection of animals used for experimental and other scientific purposes (D.M. 116192) as well as with the EEC regulations (O.J. of E.C. L 358/1 12/18/1986).
Green tea extract
Green tea extract (GTE) was a kind gift of Indena (Milano, Italy), and it was defined by the producer as having a polyphenolic content of 75 ± 5% with the major constituent being epigallocatechin-3-gallate at 62% and the minor ones being epicatechin-3-gallate, epigallocatechin and epicathechin.
Carrageenan-induced pleurisy
Mice were anaesthetised with isoflurane and submitted to a skin incision at the level of the left sixth intercostals space. The underlying muscle was dissected and saline (0.1 ml) or saline containing 2%λ-carrageenan (0.1 ml) was injected into the pleural cavity. The skin incision was closed with a suture and the animals were allowed to recover. At 4 h after the injection of carrageenan, the animals were killed by inhalation of CO2. The chest was carefully opened and the pleural cavity rinsed with 1 ml of saline solution containing heparin (5 U/ml) and indomethacin (10 μg/ml). The exudate and washing solution were removed by aspiration and the total volume measured. Any exudate, which was contaminated with blood, was discarded. The amount of exudate was calculated by subtracting the volume injected (1 ml) from the total volume recovered. The leukocytes in the exudate were suspended in phosphate-buffer saline (PBS) and counted with an optical microscope in a Burker's chamber after Blue Toluidine staining.
Experimental groups
Mice were randomly allocated into the following groups: (i) CAR + saline group. Mice were subjected to carrageenan-induced pleurisy (N = 10), (ii) Green Tea group Same as the CAR + saline group but Green Tea (25 mg/kg i.p) were administered 1 h prior to carrageenan (N = 10), (iii) Sham+saline group. Sham-operated group in which identical surgical procedures to the CAR group was performed, except that the saline was administered instead of carrageenan (N = 10), (iv) Sham + Green Tea group. Same as the Sham+saline group but Green Tea (25 mg/kg i.p) were administered 1 h prior to carrageenan (N = 10). The doses of Green Tea used here to reduce acute lung injury have been reported by us to reduce the tissue injury caused by ischemia-reperfusion in the gut (dose-response curve study) (Muià et al submitted 2005).
Determination of myeloperoxidase activity
Myeloperoxidase (MPO) activity, an indicator of polymorphonuclear leukocyte (PMN) accumulation, was determined as previously described [27]. At 4 h after intrapleural injection of carrageenan lung tissues, were obtained and weighed. Each piece of tissue was homogenised in a solution containing 0.5% hexa-decyl-trimethyl-ammonium bromide dissolved in 10 mM potassium phosphate buffer (pH 7) and centrifuged for 30 min at 20,000 × g at 4°C. An aliquot of the supernatant was then allowed to react with a solution of tetra-methyl-benzidine (1.6 mM) and 0.1 mM H2O2. The rate of change in absorbance was measured spectrophotometrically at 650 nm. MPO activity was defined as the quantity of enzyme degrading 1 μmol of peroxide min at 37°C and was expressed in mill units per gram weight of wet tissue.
Measurement of TNF-α levels
TNF-α levels were evaluated in the exudates at 4 h after the induction of pleurisy by carrageenan injection. The assay was carried out by using a colorimetric, commercial ELISA kit (Calbiochem-Novabiochem Corporation, USA).
Measurement of nitrite/nitrate
Nitrite/nitrate (NOx) production, an indicator of NO synthesis, was measured in pleural exudate. At the first nitrate in the supernatant was incubated with nitrate reductase (0.1 U/ml) and NADPH (1 mM) and FAD (50 μM) at 37°C for 15 min. Then followed another incubation with LDH (100 U/ml) and sodium pyruvate (10 mM) for 5 min. The nitrite concentration in the samples was measured by the Griess reaction, by adding 100 μl of Griess reagent (0.1% naphthylethylenediamide dihydrochloride in H2O and 1% sulphanilamide in 5% concentrated H2PO4; vol. 1: 1) to 100 μl samples. The optical density at 550 nm (OD550) was measured using ELISA microplate reader (SLT- Lab instruments Salzburg, Austria). Nitrate concentrations were calculated by comparison with OD550 of standard solutions of sodium nitrate prepared in saline solution.
Immunohistochemical localisation of ICAM-1, PAR and Nitrotyrosine
At 4 h after carrageenan administration, the lungs were fixed in 10% buffered formaldehyde and 8 μm sections were prepared from paraffin embedded tissues. After deparaffinization, endogenous peroxidase was quenched with 0.3% H2O2 in 60% methanol for 30 min. The sections were permeabilized with 0.1% Triton X-100 in PBS for 20 min. Non-specific adsorption was minimised by incubating the section in 2% normal goat serum in phosphate buffered saline for 20 min. Endogenous biotin or avidin binding sites were blocked by sequential incubation for 15 min with avidin and biotin. The sections were then incubated overnight with primary anti-ICAM-1 antibody (1:500), with 1:1000 dilution of primary antinitrotyrosine antibody (DBA), and anti-PAR antibody (1:500) or with control solutions. Controls included buffer alone or non-specific purified rabbit IgG.
To confirm that the immunoreaction for the nitrotyrosine was specific, some sections were also incubated with the primary antibody (anti-nitrotyrosine) in the presence of excess nitrotyrosine (10 mM) to verify the binding specificity. To verify the binding specificity for PARS, sections were also incubated with only the primary antibody (no secondary) or with only the secondary antibody (no primary). In these situations, no positive staining was found in the sections, indicating that the immunoreaction was positive in all the experiments carried out.
Immunocytochemistry photographs (n = 5) were assessed by densitometry. The assay was carried out by using Optilab Graftek software on a Macintosh personal computer (CPU G3-266).
Histological examination
Lung biopsies were taken at 4 h after injection of carrageenan. The biopsies were fixed for 1 wk in buffered formaldehyde solution (10% in PBS) at room temperature, dehydrated by graded ethanol and embedded in Paraplast (Sherwood Medical, Mahwah, N.J.). Tissue sections (thickness 7 μm) were deparaffinized with xylene, stained with trichromic Van Gieson, and studied using light microscopy (Dialux 22 Leitz). Blood was passed on the slide, fixed at 37°C, stained with May Grunward-Giensa, and studied using light microscopy.
Electrophoretic mobility shift assay
The lung samples have been collected in liquid nitrogen and stored at -80°C until use. Nuclear extracts have been prepared according to [28] in the presence of 10 μg/ml leupeptin, 5 μg/ml antipain and pepstain, and 1 mM PMSF (Sigma-Aldrich Company, Milan, Italy). Protein concentration in the nuclear extracts was determined using the method of [29]. Ten μg of nuclear extract have been incubated at room temperature for 20 min with (2–5 × 104cpm) of 32P-labeled double stranded oligonucleotide, containing the STAT-1 binding site (sis-inducible factor-binding recognition element, SIE/M67) from the c-Fos promoter (5'-GTCGACATTTCCCGTAAATCG-3'), the PARP-1 binding site (5'-TTCCTTGCCCCTCCCATTTTTC-3') from the Reg promoter [30] or the SP-1 consensus sequence (5'GGGGCGGGGC-3', Santa Cruz Biotechnology, CA) in a 15 μl of binding reaction buffer (20 mM HEPES, pH 7.9, 50 mM KCl, 10% glycerol, 0.5 mM DTT, 0.1 mM EDTA, 2 μg poly(dI-dC), 1 μg salmon sperm DNA). Products have been fractioned on a non denaturing 5% polyacrilamide gel in TBE (Tris-Borate-EDTA buffer, 0.5X). The intensity of the retarded bands has been measured with a Phosphorimager (Molecular Dynamic, Milan, Italy).
Materials
Unless otherwise stated, all compounds were obtained from Sigma-Aldrich Company (Milan, Italy). Primary monoclonal ICAM-1 (CD54) for immunoistochemistry was purchases by Pharmingen. Reagents and secondary and nonspecific IgG antibody for immunohistochemical analysis were from Vector Laboratories InC. Primary monoclonal anti-poly (ADP-ribose) antibody was purchased by Alexis. All other chemicals were of the highest commercial grade available. All stock solutions were prepared in non pyrogenic saline (0.9% NaCl; Baxter Healthcare Ltd., Thetford, Norfolk, U.K.).
Data analysis
All values in the figures and text are expressed as mean ± standard error (s.e.m.) of the mean of n observations. For the in vivo studies n represents the number of animals studied. In the experiments involving histology or immunohistochemistry, the figures shown are representative of at least three experiments performed on different experimental days. The results were analysed by one-way ANOVA followed by a Bonferroni post-hoc test for multiple comparisons. A p-value less than 0.05 was considered significant.
Results
Effects of green tea extract in carrageenan-induced pleurisy
Histological examination of lung sections revealed significant tissue damage (Fig. 1B). Thus, when compared with lung sections taken from saline-treated animals (Fig. 1A), histological examination of lung sections of mice treated with carrageenan showed oedema, tissue injury (Fig 1B), and infiltration of the tissue with neutrophils (PMNs) (Fig. 1B1). GTE significantly reduced the degree of injury as well as the infiltration of PMNs (Fig. 1C). Furthermore, injection of carrageenan into the pleural cavity of mice elicited an acute inflammatory response characterized by the accumulation of fluid (oedema) that contained large amounts of PMNs (Fig. 2A,B). Oedema and PMNs infiltration in pleural cavity were attenuated by the i.p. injection of GTE (Figs. 2A,B).
Figure 1 Effect of GTE on lung injury. When compared with lung sections taken from control animals (A), lung sections from carrageenan-treated mice (B) demonstrate interstitial haemorrhage and polymorphonuclear leukocyte accumulation (B1). Lung sections from a carrageenan-treated mice that received GTE (C) exhibit reduced interstitial haemorrhage and a lesser cellular infiltration. Figure is representative of all the animals in each group.
Figure 2 Effect of GTE on carrageenan-induced inflammation. The increase in volume exudate (A) and accumulation of polymorphonuclear cells (PMNs, B) in pleural cavity 4 h after carrageenan injection was inhibited by GTE. Data are means ± SEM of 10 mice for each group. *P < 0.01 vs. sham. °P < 0.01 vs. carrageenan.
Effect of green tea extract on TNF-α levels
The levels of TNF-α were significantly elevated in the exudates from vehicle-treated mice at 4 h after carrageenan administration (Fig. 3). In contrast, the levels of this pro-inflammatory cytokine was significantly lower in carrageenan-treated mice treated with GTE (Fig. 3). No significant increased of TNF-α levels was observed in the exudates of sham-operated mice.
Figure 3 Effect of GTE on TNF-α, level. Pleural injection of carrageenan caused by 4 h an increase in the release of the pro-inflammatory cytokines, tumour necrosis factor alpha (TNF-α). GTE significantly inhibited TNF-α. Data are means ± SEM of 10 mice for each group. *P < 0.01 vs. sham. °P < 0.01 vs. carrageenan.
Effects of green tea extract on MPO activity
The above histological pattern of lung injury appeared to be correlated with the influx of leukocytes into the lung tissue. Therefore, we investigate the role of GTE on the neutrophils infiltration by measurement of the activity of myeloperoxidase. Myeloperoxidase activity was significantly elevated (p < 0.001) at 4 h after carrageenan administration in vehicle-treated mice (Fig. 4). In Mice treated with green tea extract lung myeloperoxidase activity was significantly reduced (p < 0.01) in comparison to those of vehicle-treated mice (Fig. 4).
Figure 4 Effect of GTE on myeloperoxidase (MPO) activity in the lung. Within 4 h, pleural injection of carrageenan led to an increase in neutrophil accumulation in the lung (as measured by MPO activity) GTE treatment significantly inhibited neutrophil infiltration. Data are means ± SEM of 10 mice for each group. *P < 0.01 vs. sham. °P < 0.01 vs. carrageenan.
Effects of green tea extract on the expression of adhesion molecule (ICAM-1)
Staining of lung tissue sections obtained from saline-treated mice with anti-ICAM-1 antibody showed specific staining along bronchial epithelium, demonstrating that ICAM-1 is constitutively expressed (data not shown). At 4 h after carrageenan injection, the staining intensity substantially increased along the bronchial epithelium (see arrows, Fig. 5A, 6). Sections from GTE-treated mice did not reveal any up-regulation of constitutively expressed ICAM-1, which was normally expressed in the epithelium (see arrows, Fig. 5B, 6). To verify the binding specificity for ICAM-1, some sections were also incubated with only the primary antibody (no secondary). In these situations, no positive staining was found in the sections, indicating that the immunoreaction was positive in all the experiments carried out.
Figure 5 Immunohistochemical localization of ICAM-1 in the lung. Section obtained from carrageenan-treated mice showed intense positive staining for ICAM-1 (A, see arrows). The degree of bronchial epithelium (see arrows) staining for ICAM-1 (B) was markedly reduced in tissue section obtained from GTE-treated mice. Figure is representative of all the animals in each group.
Figure 6 Typical Densitometry evaluation. Densitometry analysis of immunocytochemistry photographs (n = 5) for ICAM-1, Nitrotyrosine and PAR from lung was assessed. The assay was carried out by using Optilab Graftek software on a Macintosh personal computer (CPU G3-266). Data are expressed as % of total tissue area. ND: not detectable. *P < 0.01 vs. sham. °P < 0.01 vs. carrageenan.
Effects of green tea extract on nitric oxide production
The levels of NOx were significantly (P < 0.01) increased in the exudate from carrageenan-treated mice (Fig. 7). In contrast, levels of NOx were significantly lower in the exudate of mice treated with GTE (Fig. 7).
Figure 7 Effect of GTE on NO production. Nitrite and nitrate concentrations in pleural exudate at 4 h after carrageenan administration. Nitrite and nitrate levels in carrageenan-treated mice was significantly increased vs. sham group. GTE treatment significantly reduced the carrageenan-induced elevation of nitrite and nitrate levels. Data are means ± SEM of 10 mice for each group. *P < 0.01 vs. sham. °P < 0.01 vs. carrageenan.
Effects of green tea extract on nitrotyrosine and PARS
At 4 h after carrageenan injection, lung sections were taken in order to determine the immunohistological staining for nitrotyrosine or PARS. Sections of lung from saline-treated mice did not reveal any immunoreactivity for nitrotyrosine or PARS within the normal architecture (data not shown). A positive staining for nitrotyrosine (Fig. 6, 8A) and PARS (Fig. 6, 8C) was localized primarily in the vessels and in the bronchial epithelium. GTE reduced the staining for both nitrotyrosine (Fig. 6, 8B) and PARS (Fig. 6, 8D). Therefore, no differences between groups were shown for SP-1 DNA binding activity (data not shown). It was also shown the DNA binding capacity of PARP-1 to the promoter sequence of the Reg gene [30]. The retarded bands of the carraggeenan-treated mice were reduced in comparison to those of vehicle-treated or GTE pre-treated mice (Fig. 9A, B)
Figure 8 Immunohistochemical localization for nitrotyrosine and PARS in the lung. Immunohistochemistry for nitrotyrosine (A) and PARS (C) show positive staining along the vessels and in the bronchial epithelium from a carrageenan-treated mice. The intensity of the positive staining for nitrotyrosine (B) and PARS (C) was significantly reduced in the lung from GTE-treated mice. Figure is representative of all the animals in each group.
Figure 9 Effect of GTE on PARP-1 activation. (A) DNA binding activity of PARP-1 in sham operated, carrageenan-treated (CAR) and GTE pre-treated mice (CAR + GTE). Nuclear extracts (10 μg) from lung sample were incubated with a 32P-labeled double-stranded oligonucleotide containing binding sequence for PARP-1 and separated by nondenaturing PAGE. The specificity of the retarded bands was demonstrated by competition with 100-fold excess of specific unlabeled oligonucleotide (not shown). (B) The intensity of retarded bands (measured by phosphoimager) in carrageenan-treated mice was significantly increased vs. sham group. GTE treatment significantly reduced the carrageenan-induced elevation of PARP-1 activity. Data are means ± SEM of 5 mice for each group. *P < 0.01 vs. sham. °P < 0.001 vs. carrageenan.
Effects of green tea extract on the STAT-1 activation
To examine the molecular mechanisms responsible for mediating the anti-inflammatory effects of GTE we measured, by EMSA, the changes in activation of the transcription factors STAT-1 and SP-1. DNA-binding activity of STAT-1 was significantly elevated at 4 h after carrageenan administration in vehicle-treated mice (Fig. 10A, B). In Mice treated with green tea extract lung STAT-1 activity was similar to those of sham-operated group and significantly reduced in comparison to those of vehicle-treated mice (Fig. 10A, B).
Figure 10 Effect of GTE on STAT-1 activation. (A) DNA binding activity of STAT-1 in sham operated, carrageenan-treated (CAR) and GTE pre-treated mice (CAR + GTE). Nuclear extracts (10 μg) from lung sample were incubated with a 32P-labeled double-stranded oligonucleotide containing binding sequence for STAT-1 and separated by nondenaturing PAGE. The specificity of the retarded bands was demonstrated by competition with 100-fold excess of specific unlabeled oligonucleotide (not shown). (B) The intensity of retarded bands (measured by phosphoimager) in carrageenan-treated mice was significantly increased vs. sham group. GTE treatment significantly reduced the carrageenan-induced elevation of STAT-1 activity. *P < 0.01 vs. sham. °P < 0.001 vs. carrageenan.
Discussion
Polyphenols are the most significant group of tea components, especially the catechin group of the flavonols. The major tea catechins are EGCG, EGC, ECG, EC, (+)-gallocatechin, and (+)-catechin.
Many biological functions of tea polyphenols have been studied [31], including anti-inflammatory, antioxidative [32-34], antimutagenic [35], and anticarcinogenic [36] effects.
This study provides the evidence that pretreatment of mice with green tea extract attenuates 1) the development of carrageenan-induced pleurisy, 2) the infiltration of the lung with PMNs (histology and MPO activity), 3) the degree of lung injury (histology) caused by injection of carrageenan. All of these findings support the view that green tea extract attenuates the degree of acute inflammation in mice. What, then, is the mechanism by which green tea extract reduces acute inflammation?
The generation of oxidative and nitrosative species, which exert their effects both directly and indirectly, is a important contributor to inflammatory injury. The terms oxidative and nitrosative refer to the formation of reactive oxygen species (ROS), such as superoxide (O2.-), hydrogen peroxide (H2O2), and hydroxyl radicals, and reactive nitrogen species (RNS), such as nitric oxide (NO), peroxynitrite (ONOO-), and nitrogen dioxide.
Oxidants are generated as a result of the inflammatory response by phagocytic cells, such as mononuclear cells. Oxidants that are generated in excess of antioxidant defences or that are lacking in antioxidant defences can result in severe pulmonary inflammation leading to acute lung injury Additionally, NO plays a multifaceted role in mediating inflammatory processes [37]. Potential sources of NO in the lungs include expression of iNOS, activated neutrophils, [38] alveolar type-IIcells [39] endothelial cells [40] and airway cells [41]. It has been demonstrated that levels of NO2-, increase markedly during acute and chronic inflammation [42]. Recent study had demonstrated that green tea polyphenols inhibit NO production in peritoneal exudate (macrophage) cells [43] and EGCG inhibits lipopolysaccharide (LPS)-induced NO production and iNOS gene expression in isolated peritoneal macrophages by decreasing NF-KB activation [22]. In agreement with these observations in this study we shown that the treatment with green tea extract in vivo reduce NO formation.
Simultaneous generation of NO. and O2.- favours the production of a toxic reaction product, peroxynitrite anion (ONOO-)[13] and this product may account for some of the deleterious effects associated with NO. production.
The pro-inflammatory and cytotoxic effects of ONOO- are numerous [44]. Peroxynitrite nitrosates tyrosine residues in proteins and nitrotyrosine formation has been used as a marker for the detection of the endogenous formation of peroxynitrite [45]. Using nitrotyrosine as a marker for the presence of ONOO- has been challenged by the demonstration that other reactions can also induce tyrosine nitration; e.g., the reaction of nitrite with hypochlorous acid and the reaction of myeloperoxidase with hydrogen peroxide can lead to the formation of nitrotyrosine [46]. Thus, increased nitrotyrosine staining is considered, as an indicator of "increased nitrosative stress" rather than a specific marker of the generation of peroxynitrite [47]. We have found that nitrotyrosine is indeed present in lung sections taken after carrageenan injection and that green tea extract reduced the staining in these tissues.
ROS and peroxynitrite produce cellular injury and necrosis via several mechanisms including protein denaturation, and DNA damage ROS produce strand breaks in DNA that trigger energy-consuming DNA repair mechanisms and activate the nuclear enzyme PARS, resulting in the depletion of its substrate NAD in vitro and a reduction in the rate of glycolysis. As NAD functions as a cofactor in glycolysis and the tricarboxylic acid cycle, NAD depletion leads to a rapid fall in intracellular ATP. This process has been termed the 'PARS suicide hypothesis'. There is recent evidence that the activation of PARS may also play an important role in inflammation [48,49].
We demonstrate here that green tea extract treatment reduced the activation of PARS during carrageenan-induced pleurisy in the lung. In light of the role of PARS in inflammation, it is possible that PARS inhibition by green tea extract accounts for the anti-inflammatory response.
Besides attenuating ONOO- production and PARS activation, green tea extract also reduced the development of oedema, neutrophil accumulation and had an overall protective effect on the degree of lung injury as assessed by histological examination.
A possible mechanism by which green tea extract attenuates PMNs infiltration is by down-regulating adhesion molecules ICAM-1
The activation and expression of adhesion molecules allows for the adhesion, conformational change, and extravasation (emigration) of the neutrophil that may induce local injury and participate in the orchestration of systemic inflammation and all of its consequences. ICAM-1 in particular plays a role in inflammatory processes and in the T cell-mediated host defense system. Within the endothelium, ICAM-1 has an important role in migration of leukocytes to sites of inflammation, enabling the firm adhesion and diapedesis of leukocytes. In accordance with these findings, we observed that green tea extract reduce the upregulation the surface expression of ICAM-1 on endothelial cells and prevents the infiltration of neutrophils at inflamed sites.
Lung inflammation is associated with enhanced expression of proinflammatory cytokines which serve as intercellular signals that recruit cells and modulate cell function. Cytokines produced predominantly by activated macrophages and lymphocytes mediate many inflammatory processes [49]. These proinflammatory cytokines include interferon-γ (IFN-γ), interleukin-1 (IL-1), tumour necrosis factor-α (TNFα) and chemokines (e.g. interleukin-8 [IL-8], macrophage chemotactic and activating factor [MCAF]). Though all of these cytokines play important roles in the evolving inflammatory response, TNFα appears to be a critical mediator of the inflammatory cascade. Numerous studies show that TNFα rises rapidly following acute trauma/inflammation/infection and that blocking TNFα activity reduces injury [50-52]. Recently it has been shown that, EGCG inhibits okadaic acid-induced TNFα production and gene expression in BALB/3T3 cells [53]. In agreement with these observations we have found in this study that green tea pre-treatment reduce exudates level of TNF-alpha. Moreover cytokines activate, several signaling pathways, leading to the downstream activation of NF-κB transcription factors. In this study we have placed our attention on STAT-1. Signal transducers and activators of transcription (STAT) factors are a family of cytoplasmic transcription factors that mediate intracellular signaling initiated at cytokine cell surface receptors and transmitted to the nucleus. Recently, it has been reported that the polyphenolic agent epigallocatechin-3-gallate (EGCG), a major constituent of green tea, is a potent inhibitor of STAT-1 phosphorylation and activation [26]. The JAK/STAT pathway has been shown to be essential for human and murine iNOS expression [24] and in ICAM-1 induction [25]. We demonstrate here that green tea extract pretreatment reduced STAT-1 activation in carrageenan-treated mice. The fall in STAT-1 activity may further account for the repression of ICAM-1 and iNOS lung expression with decrease in nitrite/nitrate concentration in the pleural exudate. Thus, the down-regulation of STAT-1 action provides a molecular basis for the anti-inflammatory effect of GTE.
Data generated from the present study indicate that green tea extract cause a substantial reduction of carraggenan induced-pleurisy in the mice suggesting toxicity from oxygen metabolites, released by stimulated neutrophils, macrophages, and other cells, as one of the significant mechanisms of lung injury. These findings support the potential use of green tea extract as therapeutic agents in the therapy of conditions associated with acute inflammation.
Acknowledgements
This study was supported by grant from PRIN 2003. The authors would like to thank Giovanni Pergolizzi and Carmelo La Spada for their excellent technical assistance during this study, Mrs Caterina Cutrona and Deborah A. Schaller for secretarial assistance and Miss Valentina Malvagni for editorial assistance with the manuscript.
==== Refs
Cuzzocrea S Zingarelli B Gilad E Hake P Salzman AL Szabo C Protective effect of melatonin in carrageenan-induced models of local inflammation: relationship to its inhibitory effect on nitric oxide production and its peroxynitrite scavenging activity J Pineal Res 1997 23 2 106 16 9392449
Pawliczak R The role of radical oxygen species in airway inflammation Pol Merkuriusz Lek 2003 14 493 6 14524253
Cuzzocrea S McDonald MC Filipe HM Costantino G Mazzon E Santagati S Caputi AP Thiemermann C Effects of tempol, a membrane-permeable radical scavenger, in a rodent model of carrageenan-induced pleurisy Eur J Pharmacol 2000 390 209 22 10708726 10.1016/S0014-2999(99)00910-3
Leiro J Alvarez E Arranz JA Laguna R Uriarte E Orallo F Effects of cis resveratrol on inflammatory murine macrophages: antioxidant activity and down-regulation of inflammatory genes J Leukoc Biol 2004 75
Los M Schenk H Hexel K Baeuerle PA Droge W Schulze-Osthoff K IL-2 gene expression and NF-kappa B activation through CD28 requires reactive oxygen production by 5-lipoxygenase EMBO J 1995 14 3731 40 7641692
Turpaev KT Reactive oxygen species and regulation of gene expression Biochemistry (Mosc) 2002 67 281 92 11970728 10.1023/A:1014819832003
Adcock IM Brown CR Kwon O Barnes PJ Oxidative stress induces NF kappa B DNA binding and inducible NOS mRNA in human epithelial cells Biochem Biophys Res Commun 1994 199 1518 24 7511898 10.1006/bbrc.1994.1403
Schenk H Klein M Erdbrugger W Droge W Schulze-Osthoff K Distinct effects of thioredoxin and antioxidants on the activation of transcription factors NF-kappa B and AP-1 Proc Natl Acad Sci USA 1994 91 1672 6 8127864
Mathy-Hartert M Deby-Dupont GP Reginster JY Ayache N Pujol JP Henrotin YE Regulation by reactive oxygen species of interleukin-1beta, nitric oxide and prostaglandin E(2) production by human chondrocytes Osteoarthritis Cartilage 2002 10 547 55 12127835 10.1053/joca.2002.0789
Moncada S Palmer RMG Higgs EA Nitric oxide: physiology, pathophysiology and pharmacology Pharmacol Rev 1991 43 109 142 1852778
Nathan C Nitric oxide as a secretory product of mammalian cells FASEB J 1996 6 3051 3064 1381691
Cuzzocrea S Zingarelli B Gilard E Hake P Salzman AL Szabó C Anti-inflammatory effects of mercaptoethylguanidine, a combined inhibitor of nitric oxide synthase and peroxynitrite scavenger, in carrageenan-induced models of inflammation Free Rad Biol Med 1998 24 450 459 9438558 10.1016/S0891-5849(97)00280-3
Beckman JS Beckman TW Chen J Marshall PA Freeman BA Apparent hydroxyl radical production by peroxynitrite: implications for endothelial injury from nitric oxide and superoxide Proc Natl Acad Sci U S A 1990 87 4 1620 4 2154753
Halliwell B Oxygen radicals, nitrite oxide and human inflammatory joint disease Ann Rheum Dis 1995 54 505 510 7632097
Salvemini D Wang ZQ Wyatt P Bourdon DW Marino MH Manning PT Currie MG Nitric oxide: a key mediator in the early and late phase of carrageenan-induced rat paw inflammation Br J Pharmacol 1996 118 829 838 8799551
Cuzzocrea S Zingarelli B Costantino G Szabó A Salzman AL Caputi AP Szabó C Beneficial effects of 3-aminobenzamide, an inhibitor of poly (ADP-ribose) synthetase in a rat model of splanchnic artery occlusion and reperfusion Br J Pharmacol 1997 121 1065 1074 9249240
Radi R Beckman JS Bush KM Freeman BA Peroxynitrite oxidation of sulfhydryls. The cytotoxic potential of superoxide and nitric oxide J Biol Chem 1991 266 4244 4250 1847917
Karoui H Hogg N Frejaville C Tordo P Kalyanaraman B Characterization of sulfur-centered radical intermediates formed during the oxidation of thiols and sulfite by peroxynitrite. ESR-spin trapping and oxygen uptake studies J Biol Chem 1996 271 6000 6009 8626383 10.1074/jbc.271.11.6000
Alschuler L Grean Tea: Healing tonic Am J Natur Med 1998 5 28 31
Graham HN Green tea composition, consumption, and polyphenol chemistry Prev Med 1992 21 334 350 1614995 10.1016/0091-7435(92)90041-F
Yang F de Villiers WJ McClain CJ Varilek GW Green tea polyphenols block endotoxin-induced tumor necrosis factor-production and lethality in a murine model J Nutr Dec 1998 128 12 2334 40
Lin YL Lin JK (-)-Epigallocatechin-3-gallate blocks the induction of nitric oxide synthase by down-regulating lipopolysaccharide-induced activity of transcription factor nuclear factor-KB Mol Pharm 1997 52 465 472
Shuai K Schindler C Prezioso RV Darnell JE Jr Activation of transcription by IFN-γ: tyrosine phosphorylation of a 91-kD DNA binding protein Science 1992 258 1808 1812 1281555
Gao J Morrison DC Parmely TJ Russel SW Murphy WJ An interferon-γ activated site (GAS) is necessary for the full expression of the mouse iNOS gene in response to interferon-γ and lipopolysaccharide J Biol Chem 1997 272 1226 1230 8995425 10.1074/jbc.272.2.1226
Roy J Audette M Tremblay MJ Intercellular adhesion molecule-1(ICAM-1) gene expression in human T cells is regulated by phosphotyrosyl phosphatase activity. Involvement of NF-kB, Ets, and palindromic interferon-gamma-responsive element-binding site J Biol Chem 2001 276 14553 61 11278281 10.1074/jbc.M005067200
Menegazzi M Tedeschi E Dussin D De Prati AC Cavalieri E Mariotto S Suzuki H Anti-interferon gamma action of epigallocatechin-3-gallate mediated by specific inhibition of STAT1 activation FASEB J 2001 15 1309 1311 11344123
Mullane KM Kraemer R Smith B Myeloperoxidase activity as a quantitative assessment of neutrophil infiltration into ischemic myocardium J Pharmacol Meth 1985 14 157 167 10.1016/0160-5402(85)90029-4
Schreiber E Matthias P Muller MM Schaffner W Rapid detection of octamer binding proteins with miniextracts prepared from a small number of cell Nucleic Acid Res 1989 17 6419A 2771659
Bradford M A rapid and sensitive method for the quantitation of microgram quantities of protein using the principle of protein-dye binding Anal Biochem 1976 72 248 254 942051
Akiyama T Takasawa S Nata K Kobayashi S Abe M Shervani NJ Ikeda T Nakagawa K Unno M Matsuno S Okamoto H Activation of Reg gene for insulin-producing β-cel regeneration: poly(ADP-ribose) polymerase binds Reg promoter and regulates the transcription by autopoly(ADP-ribosyl)ation PNAS 2001 98 48 53 11134536 10.1073/pnas.240458597
Yang CS Wang ZY Tea and cancer J Natl Cancer Inst 1993 85 1038 1049 8515490
Lin YL Juan IM Chen YL Liang YC Lin JK Composition of polyphenols in fresh tea leaves and associations of their oxygen-radical-absorbing capacity with antiproliferative actions in fibroblast cells J Agric Food Chem 1996 44 1387 1394 10.1021/jf950652k
Ho CT Chen Q Shi H Zhang KQ Rosen RT Antioxidative effect of polyphenol extract prepared from various Chinese teas Prev Med 1992 21 520 525 1409493 10.1016/0091-7435(92)90059-Q
Katiyar SK Agarwal R Zaim MT Mukhtar H Protection against N-nitrosodiethylamine and benzo(a)pyrene-induced forestomach and lung tumorigenesis in A/J mice by green tea Carcinogenesis 1993 14 849 855 8504476
Shiraki M Hara Y Osawa T Kumon H Nakauama T Kawakishi S Antioxidative and antimutagenic effects of theaflavins from black tea Mutat Res 1994 323 29 34 7508563 10.1016/0165-7992(94)90041-8
Huang MT Ho CT Wang ZY Ferraro T Finnegan-Olive T Lou YR Mitchell JM Laskin JD Newmark H Yang CS Conney AH Inhibitory effect of topical application of a green tea polyphenol fraction on tumor initiation and promotion in mouse skin Carcinogenesis 1992 13 947 954 1600615
Haddad IY Pitt BR Matalon S Fishman AP Nitric oxide and lung injury Pulmonary diseases and disorders 1996 McGraw-Hill New York, NY 337 346
Fierro IM Nascimento-DaSilva V Arruda MA Induction of NOS in rat blood PMN in vivo and in vitro: modulation by tyrosine kinase and involvement in bactericidal activity J Leukoc Biol 1999 65 508 514 10204580
Kooy NW Royall JA Agonist-induced peroxynitrite production from endothelial cells Arch Biochem Biophys 1994 310 352 359 8179319 10.1006/abbi.1994.1178
Kobzik L Bredt DS Lowenstein CJ Nitric oxide synthase in human and rat lung: immunocytochemical and histochemical localization Am J Respir Cell Mol Biol 1993 9 371 377 7691109
Hickman-Davis JM Lindsey JR Zhu Surfactant protein A mediates mycoplasmacidal activity of alveolar macrophages Am J Physiol 1998 274 L270 L277 9486213
Farrell AJ Blake DR Palmer RMJ Moncada S Increased concentrations of nitrite in synovial fluid and serum samples suggest increased nitric oxide synthesis in rheumatic diseases Ann Rheum Dis 1992 51 1219 1222 1466599
Chan MMY Ho CT Huang HI Effects of three dietary phytochemicals from tea, rosemary and turmeric on inflammation-induced nitrite production Cancer Lett 1995 96 23 29 7553604 10.1016/0304-3835(95)03913-H
Squadrito GL Pryor WA The formation of peroxynitrite in vivo from nitric oxide and superoxide Chem Biol Interact 1995 96 2 203 6 Review 7728908 10.1016/0009-2797(94)03591-U
Beckman JS Oxidative damage and tyrosine nitration from peroxynitrite Chem Res Toxicol 1996 9 5 836 44 Review 8828918 10.1021/tx9501445
Eiserich JP Hristova M Cross CE Jones AD Freeman BA Halliwell B van der Vliet A Formation of nitric oxide-derived inflammatory oxidants by myeloperoxidase in neutrophils Nature 1998 391 393 7 9450756 10.1038/34923
Cuzzocrea S Caputi AP Zingarelli B Peroxynitrite-mediated DNA strand breakage activates poly (ADP-ribose) synthetase and causes cellular energy depletion in carrageenan-induced pleurisy Immunology 1998 93 96 101 9536124 10.1046/j.1365-2567.1998.00409.x
Cuzzocrea S Zingarelli B Gilard E Hake P Salzman AL Szabó C Protective effects of 3-aminobenzamide, an inhibitor of poly (ADP-ribose) synthase in carrageenan-induced models of local inflammation Eur J Pharmacol 1998 342 67 76 9544794 10.1016/S0014-2999(97)01417-9
Brennan FM Feldmann M Cytokines in autoimmunity Curr Opin Immunol 1996 8 872 877 8994869 10.1016/S0952-7915(96)80018-5
Beutler B Cerami A Cachectin and tumour necrosis factor as two sides of the same biological coin Nature 1986 320 584 588 3010124 10.1038/320584a0
Machleidt T Kramer B Adam D Neumann B Schutze S Wiegmann K Kronke M Function of the p55 tumor necrosis factor receptor "death domain" mediated by phophatidylcholine-specific phospholipase C J Exp Med 1996 184 725 733 10.1084/jem.184.2.725
Pfeffer K Matsuyama T Kundig TM Wakeham A Kishihara K Shahinian A Wiegmann K Ohashi PS Kronke M Mak TW Mice deficient for the 55 kd tumor necrosis factor receptor are resistant to endotoxin shock, yet succumb to L. monocytogenes infection Cell 1993 73 457 467 8387893 10.1016/0092-8674(93)90134-C
Suganuma M Okabe S Sueoka E Iida N Komori A Kim SJ Fujiki H A new process of cancer prevention mediated through inhbition of tumor necrosis factor αexpression Cancer Res 1996 56 3711 3715 8706012
| 15987519 | PMC1177993 | CC BY | 2021-01-04 16:23:26 | no | Respir Res. 2005 Jun 29; 6(1):66 | utf-8 | Respir Res | 2,005 | 10.1186/1465-9921-6-66 | oa_comm |
==== Front
Respir ResRespiratory Research1465-99211465-993XBioMed Central London 1465-9921-6-691600884010.1186/1465-9921-6-69ResearchRelation of exaggerated cytokine responses of CF airway epithelial cells to PAO1 adherence Kube Dianne M [email protected] David [email protected] Pamela B [email protected] Department of Pediatrics, Case Western Reserve University School of Medicine, BRB 8th floor, 2109 Adelbert Rd. Cleveland, OH 44106, USA2005 11 7 2005 6 1 69 69 19 4 2005 11 7 2005 Copyright © 2005 Kube et al; licensee BioMed Central Ltd.2005Kube 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 many model systems, cystic fibrosis (CF) phenotype airway epithelial cells in culture respond to P. aeruginosa with greater interleukin (IL)-8 and IL-6 secretion than matched controls. In order to test whether this excess inflammatory response results from the reported increased adherence of P. aeruginosa to the CF cells, we compared the inflammatory response of matched pairs of CF and non CF airway epithelial cell lines to the binding of GFP-PAO1, a strain of pseudomonas labeled with green fluorescent protein. There was no clear relation between GFP-PAO1 binding and cytokine production in response to PAO1. Treatment with exogenous aGM1 resulted in greater GFP-PAO1 binding to the normal phenotype compared to CF phenotype cells, but cytokine production remained greater from the CF cell lines. When cells were treated with neuraminidase, PAO1 adherence was equalized between CF and nonCF phenotype cell lines, but IL-8 production in response to inflammatory stimuli was still greater in CF phenotype cells. The polarized cell lines 16HBEo-Sense (normal phenotype) and Antisense (CF phenotype) cells were used to test the effect of disrupting tight junctions, which allows access of PAO1 to basolateral binding sites in both cell lines. IL-8 production increased from CF, but not normal, cells. These data indicate that increased bacterial binding to CF phenotype cells cannot by itself account for excess cytokine production in CF airway epithelial cells, encourage investigation of alternative hypotheses, and signal caution for therapeutic strategies proposed for CF that include disruption of tight junctions in the face of pseudomonas infection.
Cystic FibrosisInflammationPseudomonas aeruginosaIL-8NeuraminidaseTight junctions
==== Body
Background
Chronic infection of the lung with Pseudomonas aeruginosa and the inflammatory response it stimulates cause much of the morbidity and nearly all the mortality in CF patients. Since the inflammatory response can be reduced pharmacologically in CF patients without allowing infection to increase and with benefit to the patient [1], and since infants and young children with CF have interleukin-8 (IL-8) and neutrophil count in BAL fluid significantly in excess of that observed for non-CF children with comparable bacterial burden [2,3], many investigators have concluded that the inflammatory response is excessive and deleterious in the CF lung [reviewed in [4]]. Though the cellular origin of the excessive inflammatory response in CF is not fully established, in vivo mouse CFTR complementation data suggest that the airway epithelium plays a substantive role in driving excess inflammation [5]. In many, but not all, model systems, CF airway epithelial cells respond to P. aeruginosa or its products with increased IL-8 and/or IL-6 production compared to non-CF cells [4,6-11]. In addition, in some, but not all, model systems binding of P. aeruginosa to CF airway epithelial cells is in excess of its binding to non-CF cells [12-16]. Taken together, these data have been interpreted to mean that the excess cytokine responses in CF epithelium are due to increased stimulus applied at the cell surface by elevated bacterial adherence in the CF phenotype cells [15].
Our prior studies, in two separate cell model systems, have shown that there is increase in available asialoGM1 (aGM1), which binds to P. aeruginosa pilin and flagellin and serves as a major ligand for this organism, on the CF member of the cell pair [17-19]. In these same cell pairs, there is an increased response of IL-8, IL-6, and granulocyte macrophage colony stimulating factor (GM-CSF) to a laboratory strain of P. aeruginosa, PAO1, in the CF member of the pair [6]. However, these studies did not directly address the relationship between PAO1 binding and the cytokine response. In order to test the hypothesis that the cytokine response of CF phenotype airway epithelial cells to PAO1 can be attributed solely to increased pseudomonas adherence, we took several approaches. First, we determined whether cytokine responses and P. aeruginosa adherence changed in parallel with increasing amounts of added PAO1. Second, we manipulated the cells to alter surface receptor access to P. aeruginosa. We incubated CF and non-CF cells with exogenous aGM1 to increase the binding sites for P. aeruginosa, and treated them with neuraminidase to add or expose more desialylated binding sites. We then compared cytokine production and binding of PAO1 in the altered cell preparations. In the cell lines that form tight junctions, we increased access to native P. aeruginosa binding sites on the basolateral surface by disrupting tight junctions [20], then tested the ability of the treated cells to respond to PAO1. Our results indicate that excess cytokine responses in CF airway epithelial cells do not correlate well with adherence of P. aeruginosa, and suggest that the excess cytokine response cannot result solely from the increased adherence of P. aeruginosa.
Methods
Cell lines
pCEP and pCEP-R cell lines
The development and maintenance of this matched pair of human tracheal epithelial cells derived from SV40 transformed human tracheal epithelial cells (9HTEo-, kindly provided by Dieter Gruenert, University of Calif, San Francisco) have been described previously and these methods were followed here [6,18,21].
16HBE-14o- AS and S cell lines
The development and maintenance of these cell lines have been described previously and these were the methods used [6,22].
Bacteria
The laboratory isolate PAO1 and its GFP derivative strain were kindly provided by Alice Prince, Columbia University, NY, and were grown as previously described [6].
PAO1 Binding Assay
Green fluorescent protein (GFP)-PAO1 at 109 CFU were incubated with cell monolayers of pCEP or pCEP-R cells for 1 hr. Cells were washed with Hanks buffered salt solution (HBSS), lysed, and GFP fluorescence quantitated by fluorimeter. Serial dilutions of GFP-PAO1 were used to assess the change in GFP-PAO1 binding over a range of concentrations.
Stimulation of cytokine production by P. aeruginosa
These studies were performed as previously described [6]. Briefly, 9/HTEo- cells, pCEP and pCEP-R, were plated at a density of 1 × 106 cells per well on vitrogen-coated 24-well plates, and the sense and antisense clones of 16HBEo- cells were plated at density of 1 × 106 cells per 12 mm Millicell HA filter. Eighteen to 24 hr before the experiment, cells were switched to serum-free media, because PAO1 is serum-sensitive. Washed bacterial aliquots (0.5 ml/well) were incubated for 60 min with the confluent monolayers of epithelial cells at 37°C. Non-treated control wells were processed similarly with HBSS alone. For polarized 16HBE-14o- cells on filters, PAO1 and other treatments were applied to the apical surface only. As a positive control, cells were stimulated for 1 hr with IL-1β (100 ng/ml) and tumor necrosis factor (TNF)-α (100 ng/ml), (Sigma, St. Louis, MO). Cell monolayers were washed 3 times in Hanks Buffered Salt Solution (HBSS), then incubated for 24 hr in 0.5 ml serum-free cell culture medium containing 100 μg/ml gentamicin. Media were collected and analyzed for IL-8 and IL-6 by enzyme linked immunoadsorbant assay (ELISA), and normalized to the protein concentration of the lysed cells.
Glycophospholipid addition and fluorescence microscopy
250 μg (or 5 μl of 10 mg/ml stock in dimethylsulfoxide (DMSO)) monosialoganglioside (GM1) or gangliotetraosyl ceramide (aGM1) (Matreya, Inc), was added in 0.195 ml of serum-free media for 1 hr with gentle rocking to pCEP and pCEP-RF cells. Cells were then washed twice with HBSS, and PAO1 was applied as above. Immunofluorescence was performed by incubating the cells with a 1:1000 dilution of rabbit polyclonal anti-aGM1 (Wako Pure Chemical Industries Ltd, Osaka, Japan) in phosphate buffered saline (PBS) with 0.1% bovine serum albumin (BSA), for 1 hr at 37°C, followed by two washes with PBS, and fixation with 4% paraformaldehyde (PFA) for 1 hr, and washed with PBS. Monolayers were then incubated with FITC- conjugated goat anti-rabbit antibody (Jackson Immunoresearch Laboratories Inc.) diluted 1:100 PBS with 0.1% BSA for 1 hr at room temperature, washed with PBS and fixed again with 4% PFA for 20 minutes. Cells were mounted under coverslip with Fluoromount G antifade (Southern Biotechnology Associates, Inc, Birmingham, AL) and visualized by fluorescence microsopy using a fluorescein filter set.
FITC-Peanut Agglutinin (PNA, which binds to aGM1), or FITC-Maakia Amurensis lectin (MAL I, which recognizes sialic acid in α2,3 linkages to GlcNAC), at 100 μg in 300 ml PBS, was incubated with cells for 30 minutes after fixation in 4% PFA, washed with PBS and fixed in methanol for 10 minutes. Cells were mounted under coverslip and visualized by epifluorescent microscopy with a Zeiss 100 Axiovert, 40X water immersion objective, NA 0.75, and FITC filter set. Fluorescent-conjugated lectins were purchased from Vector Laboratories, Burlingame, CA.
Neuraminidase treatment
Neuraminidase from Clostridium perfringens, which removes sialic acid in α2,3, α2,6 or α2,8- linkages (5 U) or Salmonella typhimurium neuraminidase, which preferentially removes α2,3- linked sialic acid residues, (22.5 U) (Sigma, St. Louis, MO), was added to 200 μl serum-free media per well for 1 hr prior to PAO1 exposure. Effectiveness of treament was assessed qualitatively by immunofluorescent microscopy of fixed cells as described above.
Treatments to disrupt tight junctions
The integrity of junctional complexes was diminished in two ways: first, by calcium chelation by incubating 16HBEo- monolayers with 30 mM EGTA in PBS buffer, for 60 min, or second, by overnight incubation with 250 μg of a monoclonal mouse E-cadherin antibody (Zymed Laboratories, San Francisco, CA) in 0.5 ml serum-free media.
Transepithelial Resistance
Transepithelial resistance (TER) of cell monolayers grown on transwell filters was measured with a Millicell-ERS resistance system (Millipore, Bedford, MA) meter and STX-2 Electrodes (World Precision Instruments, Inc). Electrodes were equilibrated in cell culture media at room temperature, and measurements made with one electrode placed inside the insert and the other outside in the basolateral media. Baseline resistance of filters alone was determined. The TER of the polarized monolayers on filters was determined prior to treatments, immediately following treatment, and then at the final 24 hr time point.
Cytotoxicity Assays
To quantify cytotoxity of treatments, the concentration of lactate dehydrogenase (LDH) released from cells into the medium was measured using materials purchased from Sigma Chemical Co. (St. Louis, MO) at the same time point as was used for measuring cytokines.
Statistics
Results are expressed as mean ± standard error of the mean (SEM). All experiments reported were repeated on at least three separate occasions, and each individual cytokine experiment was performed in triplicate wells, except as specified in the legends of Figures 4 and 6. To combine multiple experiments of the 9/HTEo- cell lines, the secreted cytokine concentration (pg/mg protein) of 109 CFU of PAO1-stimulated pCEP-R cells at 24 hr. was set to 100% for each experiment, and other concentrations are expressed relative to this value. Most analysis was performed by t-test, some by ANOVA, using Sigma Plot software (SPSS, Inc., Chicago, IL). Results were considered significant when p ≤ 0.05.
Figure 4 Lectin binding to 9HTEo- cell pairs following treatment with neuraminidase. A. FITC-PNA binding to pCEP (a-c) and pCEP-RF (d-f) cells before (a,d) and after treatment with Clostridium perfringens neuraminidase (b,e) or Salmonella typhimurium neuraminidase (c,f). Binding is similar to pCEP and pCEP-R cells after treatment. Micrographs are representative of two separate experiments. B. FITC-MALI binding to pCEP (a-c) and pCEP-RF (d-f) cells before (a,d) and after treatment with C. perfringens (b,e), or S. typhimurium neuraminidase (c,f).
Figure 6 Treatments that disrupt tight junctions increase the PAO1-stimulated IL-8 response, but not the TNF-α/IL-1β stimulated response of CF-phenotype cells. 16HBEo- Sense (open bars) and Antisense (black bars) monolayers on filters were pretreated for 60 minutes with 30 mM EGTA prior to 1 hr. stimulation with 109 CFU PAO1/(EGTA, n = 5 independent experiments, each with triplicate wells), or an overnight incubation with 250 μg monoclonal antibody to E-Cadherin (ECAD, n = 3 independent experiments, each with triplicate wells), and the IL-8 (A, C) and IL-6 (B, D) response measured 24 H later by ELISA. The IL-8 response to PAO1 was significantly (*) increased in the 16HBE-Antisense cells following pretreatment with E-Cadherin antibody (p = 0.034) or EGTA (p < 0.001). The 16HBEo-AS cells produced significantly more IL-8 than their sense congeners (p < 0.05). There was a significant (*) reduction in the IL-6 (p = 0.05) and IL-8 (p = 0.041) response to TNF-α/IL-1β after overnight incubation to the E-cadherin antibody (n = one experiment of triplicate wells). There is a significant increase of IL-8 in response to PAO1 prior to treatment in the CF phenotype cells compared to normal (p = 0.001).
Results
Binding of GFP-PAO1 to the cell lines
Our prior data indicate that for both the 16HBEo- AS and S cell pairs, and for the 9HTEo- pCEP and pCEP-R cell pairs, IL-8 and IL-6 production increased with addition of increasing amounts of PAO1 over the range of 107 to 109 organisms [6]. Figure 1 illustrates the changes in GFP-PAO1 binding with increasing concentrations of bacteria. For the 16 HBEo- cells, PAO1-GFP binding also increased with added PAO1 from 107 to 109 CFU/mL, but for 9HTEo- cells, binding increased from 106 to 108 CFU/mL but did not increase not further with 109 CFU/mL, even though the cytokine responses did. Binding of GFP-PAO1 was similar in untreated 16HBEo- sense (S) and antisense (AS) cell lines, at all concentrations, and in untreated 9HTEo-pCEP and pCEP-R cell lines, at all concentrations (Figure 1). Therefore, the previously reported increase in available aGM1 in the CF member of the pairs, confirmed below, was not necessarily associated with increased GFP-PAO1 binding, and increased cytokine production was not invariably associated with increased binding of GFP-PAO1.
Figure 1 Binding of GFP-PAO1 to airway epithelial cells. GFP-PAO1 was added to cultured cells for one hour at 37°C, washed, and the cultures lysed and fluorescence determined (and expressed in arbitrary units). A and B, 9HTEo- cells, C, 16 HBEo- cells. For the 9HTEo- cells, binding appears to saturate at about 108 organisms/well (A) but for 16 HBEo- cells, binding increases with increasing dose of bacteria over the range tested (C). The 9HTEo- cells change GFP-PAO1 binding with addition of a GM1 or GM1, or with neuraminidase treatment (B) (*, significantly different from no treatment, p < 0.05), but the 16 HBEo- cells do not (C).
Providing additional P. aeruginosa binding sites by addition of asialoGM1
Others report that exogenous aGM1 is incorporated into the cell membrane and provides additional binding sites for P. aeruginosa [23]. We therefore incubated our cell lines with exogenous a GM1 and measured cell-associated aGM1, GFP-PAO1 binding, and cytokine responses. Incubation of the 9/HTEo- cell lines with aGM1 resulted in increased cell-associated aGM1, as demonstrated both by specific antibody binding, and by binding of PNA, a lectin which recognizes aGM1 (Figure 2). There was no change in LDH release (Table 1). Prior to treatment, as reported previously [19], the 9HTEo-pCEP-R cells displayed more aGM1 than the 9HTEo-pCEP cells (Figure 2A vs E for aGM1 and C vs G for PNA), but following treatment, the two cell lines had similar aGM1 antibody fluorescence and PNA fluorescence (Figure 2, B vs F and D vs H). Prior to treatment, binding of GFP-PAO1 to the two cell types is equivalent, (Figure 1, Table 1). After aGM1 incubation, both cell lines showed increased GFP-PAO1 binding, but more so in the non-CF than the CF phenotype cells (Table 1). Untreated CF phenotype cells had increased IL-8 and IL-6 production in response to PAO1 compared to normal, as previously reported [6]. Following incubation, although aGM1 and PAO1 binding increased in the normal cells, cytokine production did not, but IL-8 production by the CF phenotype cells showed a statistically significant increase (Figure 3). As a control, the cells were loaded with GM1, which is less efficient in binding PAO1. Following GM1 preincubation, despite the increase in PAO1-GFP binding (Table 1), there was a significant decrease in production of both IL-8 and IL-6 by the CF phenotype cell line (Figure 3), possibly because more PAO1 was bound at sites that do not initiate an inflammatory signal. No changes in cytokine response to TNF-α/IL-1β occurred following incubation with aGM1 or GM1 (data not shown).
Figure 2 Exogenous aGM1 is incorporated into 9HTEo-pCEP and pCEP-R. pCEP (a-d) and pCEP-R (e-h) cells were incubated with either vehicle (a,c,e,g) or aGM1 (b,d,f,h) and fluorescent staining with either antibody to aGM1 and FITC secondary (b,f) or FITC- conjugated PNA (D,H) was performed. In the untreated state, there is more binding of antibody to aGM1 or PNA to pCEP-R than pCEP cells. After incubation with aGM1, fluorescence patterns are similar for pCEP and pCEP-R cells with antibody to aGM1 (b,f) and PNA (d,h). Micrographs are representative of 3 separate experiments.
Table 1 Binding of PAO1 and LDH release by 9HTEo- cell lines
pCEP pCEP-R
LDH (U/ml) GFP -PA01 (RFU) LDH (U/ml) GFP-PA01 (RFU)
PA 8.83 ± 2.5 0.3483 ± 0.12 4.64 ± 0.7* 0.3745 ± 0.123
PA + EGTA 3.77 ± 0.9 ND 2.82 ± 0.0 ND
PA + aGM1 2.36 ± 0.5 0.7764 ± 0.037† 3.3 ± 0.5 0.5489 ± 0.085*†
PA + GM1 2.83 ± 0.0 0.6929 ± 0.049† 1.89 ± 0.0 0.4598 ± 0.090*
PA + C.p. ND 0.6350 ± 0.091† ND 0.4359 ± 0.079*
No treatment 4.79 ± 1.0 ND 2.44 ± 0.9 * ND
EGTA alone 2.36 ± 0.5 ND 4.24 ± 0.5 ND
aGM1 alone 3.3 ± 0.5 ND 2.83 ± 0.9 ND
GM1 alone 2.83 ± 0.9 ND 3.3 ± 0.5 ND
*Different from pCEP congener, p < 0.05
† Different from no treatment, p < 0.05
Figure 3 IL-8 and IL-6 responses to PAO1 or no stimulation, with or without preincubation with aGM1 or GM1. *, different from no treatment, p < 0.05.
In polarized epithelial cell lines (16BHBEo-), the addition of aGM1 or GM1 did not increase GFP-PAO1 binding (Table 2, Figure 1), nor alter the proinflammatory cytokine response to P. aeruginosa or TNF-α/IL-1β (data not shown).
Table 2 Transepithelial resistance, LDH release and PAO1 binding to 16HBEo- cell lines.
Sense AntiSense
Resistance (Ω*cm2) LDH (u) GFP-PA01 (RFU) Resistance (Ω*cm2) LDH (u) GFP-PA01 (RFU)
PA alone 229.0 ± 13.8 4.08 ± 0.4 0.6056 ± 0.078 217.85 ± 8.0 2.12 ± 0.2* 0.5753 ± 0.021
PA + EGTA 136.14 ± 6.1 2.42 ± 0.3 ND 142.43 ± 5.9 2.78 ± 0.6 ND
PA + ECAD 126.0 ± 2.3 2.95 ± 0.3 ND 131.0 ± 1.3 2.99 ± 0.4 ND
PA + aGM1 ND ND 0.6456 ± 0.051 ND ND 0.7437 ± 0.165†
PA + C.p. ND ND 0.5851 ± 0.031 ND ND 0.5154 ± 0.031
No treatment 316.54 ± 10.5 4.31 ± 0.7 ND 225.72 ± 2.9 1.24 ± 0.1* ND
EGTA alone 213.58 ± 15.2 3.22 ± 0.0 ND 202.4 ± 4.3 0.86 ± 0.0 ND
ECAD alone 147.186 ± 14.4 3.73 ± 0.0 ND 151.09 ± 6.2 1.07 ± 0.2 ND
* Different from Sense congener, p < 0.05
†Different from no aGM1, p < 0.05
Providing additional P. aeruginosa binding sites by enzymatic removal of sialic acid
C. perfringens neuraminidase removes sialic acid in the α2,3, α2,6, and α2,8 linkages. S. typhimurium neuraminidase attacks the α2,3 linkage preferentially. Increases in binding of PNA, which recognizes aGM1, are evident for both cell lines following neuraminidase treatment (Figure 4A, panels B,C,E,F). This increase may result from relief of steric hindrance to binding to existing sites, since some investigators find that the final sialic acid residue is not removed by their action. Nevertheless, MAL I, a lectin which recognizes sialic acid in terminal α2,3 linkages, shows visible decrease in surface binding in the non-CF phenotype cells that have been treated with C. perfringens neuraminidase (Figure 4B, panels B and E), but the change in MAL I fluorescence after S. typhimurium neuraminidase is less clear (Figure 4B, panels E and F). Clostridium perfringens neuraminidase treatment significantly increased GFP-PAO1 binding on the non-CF cell line (Table 1), but the cytokine responses to PAO1 or TNF-α/IL-1β did not increase in the non-CF cells (Figure 4). There was no significant increase in GFP-PAO1 binding in the CF phenotype cells, even though they showed increased IL-8 and IL-6 responses following treatment with the broad spectrum neuraminidase of C. perfringens. Following treatment with the more specific S. typhimurium enzyme only IL-8 was increased (Figure 5).
Figure 5 Neuraminidase treatment alters cytokine responses in 9HTEo- cell lines. IL-8 (A) and IL-6 (B) responses to 109 CFU PAO1 or TNF-α/IL-1β are shown. For 9HTEo-pCEP cells, only IL-8 secretion increased, and only following treatment with C. perfringens neuraminidase (C.p.), not with the enzyme from S. typhimurium. However, 9HTEo-pCEP-R cells showed increased IL-8 response to PAO1 following treatment with either enzyme and IL-6 response to C.p. neuraminidase. Three separate experiments were performed, each with triplicate wells. (*, different from untreated samples, p < 0.05).
C. perfringens neuraminidase treatment did not alter either the IL-8 response or PAO1 binding in the polarized cell lines. However, the IL-6 response of the CF phenotype line was reduced (data not shown).
Exposure of basolateral receptors to P. aeruginosa
We expected that disrupting the tight junctions in the monolayer would permit PAO1, applied to the apical surface, to access basolateral receptors that were not available when the monolayer was intact [23], and thereby would increase the cytokine response to PAO1. The tight junctions in both the Sense (control) and Antisense-treated (CF phenotype) 16 HBEo- cell lines were disrupted by treatment with EGTA or antibodies to E-cadherin, as shown by the decrease in transepithelial resistance following these treatments (Table 2). Incubation of the filters without disrupting agents for the time course of the experiment did not alter transepithelial resistance. When incubation with the disrupting agents was combined with P. aeruginosa exposure, the transepithelial resistance fell even further, to approximate that of the filters alone. The non CF phenotype cell lines (pCEP and 16HBEo- Sense) show both a greater transepithelial resistance and a greater amount of lactate dehydrogenase in the medium at baseline than CF phenotype cell lines (pCEP-R and 16HBEo- AntiSense) (Tables 1 and 2). Apoptosis is reported to be reduced in CF versus non-CF cell lines [24,25], which may account for the lesser release of LDH. However, none of the treatments that alter PA receptor availability further disrupted the integrity of cellular membranes or increased LDH release (Table 1).
Although the disruptive treatments had similar effects on resistance in CF and nonCF phenotype cells, the cytokine response to PAO1 increased with disruption of tight junctions only in the CF phenotype cells. There was no increase in cytokine production following TNF-α/IL-1β stimulation: in fact in one sample a small decrease was seen (Figure 6). In order to test whether the EGTA treatment, in and of itself, altered cytokine production by airway epithelial cell lines, we treated non-polarized 9HTEo- cell lines with EGTA in the same manner as it was applied to the 16HBEo- cells. There was a slight but statistically significant decrease in IL-6 in response to PAO1 production by the CF phenotype cells, but no other changes (data not shown). Since in the polarized cells, EGTA pretreatment resulted in increased cytokine production in the CF cell line, and if anything, EGTA treatment of nonpolarized cells produced no such increase, we ascribe the increases in the polarized cells to disruption of the tight junctions and not to some nonspecific effect of EGTA.
Discussion
In some model systems, CF airway epithelial cells produce more IL-8 and/or IL-6 than non CF cells in response to P. aeruginosa, and in some model systems, CF cell surfaces bind the organism to a greater extent than normal [6-16]. The studies reported here were designed to test the hypothesis that increased binding sites for PAO1 result in increased stimulus and increased cytokine production in response to PAO1 in airway epithelial cells (Figure 7). The hypothesis was not supported. Surprisingly, although aGM1 was increased on the CF phenotype cells studied here under basal conditions, GFP-PAO1 binding was not, so the increased cytokine responses of CF phenotype cells to PAO1 in the basal state [6] cannot be attributed solely to increased PAO1 adherence. Moreover, increasing the binding of PAO1 to non-polarized normal airway epithelial cell lines (9HTEo-pCEP), either by adding aGM1 or by cleaving sialic acid at the cell surface, does not change the cytokine responses to PAO1. CF phenotype cells (9HTEo-pCEP-R) still respond to PAO1 with greater cytokine release than their matched normal counterparts, despite significantly less PAO1 adherence than normal phenotype cells. For matched polarized cell lines (16HBEo-), there was little change in PAO1 binding from adding aGM1 or cleaving sialic acid at the cell surface in either the CF or the non CF line, and little change in the cytokine response to PAO1. However, when the basolateral surface was made available for PAO1 binding by disruption of the tight junctions, cytokine responses to PAO1 increased only in the CF phenotype cells.
Figure 7 Cartoon comparing CF and non-CF epithelial cell responses to P. aeruginosa and illustrating two hypotheses to explain the increased cytokine response from CF airway epithelial cells. Bacterial adherence to the cell stimulates an intracellular signaling cascade. CF cells produce more IL-8 and IL-6 than non-CF cells. In the first hypothesis, increased bacterial adherence to the CF cell leads to increased signal, with consequent increase in IL-8 and IL-6 secretion. In the second hypothesis, the CF cell responds to each binding event with amplification of the signal compared to non-CF cells, and increased IL-8 and IL-6 secretion.
It is likely that there are multiple ligands for PAO1 on airway epithelial cells. Two that have been identified are aGM1 and CFTR itself [18,24], and it is likely that GM1 is a weak binding site as well. Thus, it is possible that GFP-PAO1 adheres more to increased aGM1 binding sites on the CF cells (which apparently signal for inflammatory mediators) but may adhere less at other sites, perhaps at CFTR itself, making it appear that adherence has little relation to cytokine response when in fact a only a subset of pseudomonas receptors is responsible for the increased response. Nevertheless, attempts to increase aGM1 directly did not produce the expected changes in the cytokine responses of non-CF cell lines, but did enhance the responses of the CF cell lines. Adding exogenous aGM1 effectively equalized surface aGM1 in both normal and CF cell lines as measured by antibodies to aGM1 or PNA lectin binding, and actually increased GFP-PAO1 binding to the non CF relative to the CF cell line. Were P. aeruginosa binding to aGM1 the principal determinant of the pro-inflammatory cytokine response, one would expect that the response of the normal cell lines under these conditions would equal or exceed that of the CF phenotype cell lines. However, the CF cell lines still produced more IL-8 in response to PAO1. Although one could argue that increasing aGM1 in this manner might produce binding sites that are not connected to the signaling machinery, others have shown that exogenous aGM1 incorporates into cellular membranes, increases the binding of P. aeruginosa, and augments biological responses to P. aeruginosa, including cytotoxicity, internalization and the apoptotic response [23]. Moreover, the CF cell line treated in this manner did augment its cytokine response to PAO1 (but not to another stimulus, TNF-α/IL-1β, eliminating the possibility of a generalized increase in cytokine production). In contrast, GM1 preincubation, which also resulted in increased PAO1 binding, did not increase the IL-8 response to PAO1: in this case, the increase in binding of P. aeruginosa in the GM1-incubated cells is probably not coupled to a proinflammatory signaling cascade. Association of PAO1 with a non-signaling GM1 ligand could block access to aGM1 receptors, actually reducing the response.
Another attempt to alter access to aGM1 binding sites also did not reveal association between binding and inflammatory response. Treating the cells with the broad spectrum neuraminidase from C. perfringens resulted in significantly increased lectin binding sites and IL-8 response to PAO1 (but not to TNF-α/IL-1β) in both CF phenotype and non-CF phenotype 9HTEo- cells. However, the excess cytokine response of the CF cells was preserved following neuraminidase treatment, despite equalizing apparent binding of PAO1. Neuraminidase from S. typhimurium, which preferentially removes sialic acid in the α2,3 linkage, produced a significant increase in IL-8 production in response to PAO1 only in the CF phenotype cells.
We made no attempt to assess with lectin binding or specific antibody the nature of the basolateral binding sites revealed in polarized cultures by disruption of tight junctions. Others have shown that allowing access to basolateral receptors greatly increases P. aeruginosa binding, cytotoxicity, internalization, and apoptosis independent of CFTR [23,26,27]. Nevertheless, opening tight junctions did not enhance PAO1-stimulated cytokine production in the non-CF cell line, whereas it did in the CF congener.
Conclusion
The data presented here indicate that the increased cytokine responses in CF airway epithelial cells to P. aeruginosa cannot be attributed solely to increased adherence of the organism. There are several implications of this finding. First, these data focus attention on alternative hypotheses to explain the increased inflammatory response of the CF airway epithelial cell (Figure 7). Our data make the hypothesis that increased pseudomonas binding entirely accounts for the increased inflammatory response of CF epithelium [15] much less likely. Alternatively, there may be increased amplification of the signal from the bacterium in CF cells to account for the increased response. Considerable attention has been paid to the excess activation of NF-κB in CF epithelial cells. Some investigators find that there is activation of this transcription factor even in the unstimulated state in CF epithelial cells, and others find that it is activated to excess only under conditions of stimulation. This pivotal transcription factor could account for a panoply of abnormalities, including the excess cytokine production documented here, but also increased release of MMP-9 and reduced apoptosis of CF airway epithelial cells. Others have proposed that in CF there is failure of anti-inflammatory control mechanisms such as IL-10, NO, or transcription factors that compete with NF-κB for helicases, or there may be subtle abnormalities in both the pro- and anti-inflammatory arms of the cascade [reviewed in [4]]. A second caveat raised by our data is that disrupting tight junctions in the CF epithelium can markedly increase inflammatory responses to P. aeruginosa, even if the bacteria are applied only briefly and the cells are given opportunity to recover. Moreover, the combination of disrupting agents and P. aeruginosa produced complete loss of the electrophysiologic barrier in a manner that EDTA or anti-E-cadherin did not. These observations signal caution for therapeutic strategies that propose to access the basolateral surface of CF airway epithelial cells by disrupting the tight junctions in vivo [28,29], especially in patients already infected with P. aeruginosa.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
DMK designed the experiments, conducted many of them, prepared the figures, and drafted the paper. DF conducted some of the experiments, including stimulations of cells for cytokine measurements and determinations of transepithelial resistance. PBD contributed to experimental design and data interpretation and wrote and edited the final version of the paper. All authors read and approved the final manuscript.
Acknowledgements
The authors thank Alice Prince for the PAO1 strains, Aura Perez for the 9/HTEo- cell lines, Ute Sontich for the 16HBEo- cells lines, and our CF Center's Inflammatory Mediator Core for their expert technical services. This work was supported by NIH P50HL60293, and P30 DK27651
==== Refs
Konstan MW Byard PJ Hoppel CL Davis PB Effect of high-dose ibuprofen in patients with cystic fibrosis N Engl J Med 1995 332 13 848 54 7503838 10.1056/NEJM199503303321303
Muhlebach MS Reed W Noah TL Quantitative cytokine gene expression in CF airway Pediatr Pulmonol 2004 37 5 393 399 15095321 10.1002/ppul.20010
Muhlebach MS Stewart PW Leigh MW Noah TL Quantitation of inflammatory responses to bacteria in young cystic fibrosis and control patients Am J Respir Crit Care Med 1999 160 186 191 10390398
Chmiel JF Davis PB State of the Art: Why do the lungs of patients with cystic fibrosis become infected and why can't they clear the infection? Respir Res 2003 4 1 8 14511398 10.1186/1465-9921-4-8
Oceandy D McMorran BJ Smith SN Schreiber R Kunzelmann K Alton EW Hume DA Wainwright BJ Gene complementation of airway epithelium in the cystic fibrosis mouse is necessary and sufficient to correct the pathogen clearance and inflammatory abnormalities Hum Mol Genet 2002 11 9 1059 67 11978765 10.1093/hmg/11.9.1059
Kube D Sontich U Fletcher D Davis PB Proinflammatory cytokine responses to P. aeruginosa infection in human airway epithelial cell lines Am J Physiol Lung Cell Mol Physiol 2001 280 L493 L502 11159033
DiMango E Ratner AJ Bryan R Tabibi S Prince A Activation of NF-kappaB by adherent Pseudomonas aeruginosa in normal and cystic fibrosis respiratory epithelial cells J Clin Invest 1998 101 2598 2605 9616231
Joseph T Look D Ferkol T NF-kappaB activation and sustained IL-8 gene expression in primary cultures of cystic fibrosis airway epithelial cells stimulated with Pseudomonas aeruginosa Am J Physiol Lung Cell Mol Physiol 2005 288 3 L471 9 15516493 10.1152/ajplung.00066.2004
Becker MN Sauer MS Muhlebach MS Hirsh AJ Wu Q Verghese MW Randell SH Cytokine secretion by cystic fibrosis airway epithelial cells Am J Respir Crit Care Med 2004 169 5 645 53 14670800 10.1164/rccm.200207-765OC
Escotte S Tabary O Dusser D Majer-Teboul C Puchelle E Jacquot J Fluticasone reduces IL-6 and IL-8 production of cystic fibrosis bronchial epithelial cells via IKK-beta kinase pathway Eur Respir J 2003 21 4 574 81 12762338
Tabary O Zahm JM Hinnrasky J Couetil JP Cornillet P Guenounou M Gaillard D Puchelle E Jacquot J Selective up-regulation of chemokine IL-8 expression in cystic fibrosis bronchial gland cells in vivo and in vitro Am J Pathol 1998 153 3 921 30 9736040
Saiman L Prince A Pseudomonas aeruginosa pili bind to asialoGM1 which is increased on the surface of cystic fibrosis epithelial cells J Clin Invest 1993 92 1875 1880 8104958
Zar H Saiman L Quittell L Prince A Binding of Pseudomonas aeruginosa to respiratory epithelial cells from patients with various mutations in the cystic fibrosis transmembrane regulator J Pediatr 1995 126 230 233 7531240
Scheid PL Kempster U Griesenbach JC Davies A Dewar PP Weber WH Colledge MJ Evans D Geddes M Alton EW Inflammation in cystic fibrosis airways: relationship to increased bacterial adherence Eur Respir J 2001 17 27 35 11307750 10.1183/09031936.01.17100270
Davies J Dewar A Bush A Pitt T Gruenert D Geddes DM Alton EW Reduction in the adherence of Pseudomonas aeruginosa to native cystic fibrosis epithelium with anti-asialoGM1 antibody and neuraminidase inhibition Eur Respir J 1999 13 565 570 10232427 10.1183/09031936.99.13356599
Davies JC Stern M Dewar A Caplen NJ Munkonge FM Pitt T Sorgi F Huang L Bush A Geddes DM Alton EW CFTR gene transfer reduces the binding of Pseudomonas aeruginosa to cystic fibrosis respiratory epithelium Am J Respir Cell Mol Biol 1997 16 6 657 63 9191467
Bryan R Kube D Perez A Davis P Prince A Overproduction of the CFTR R domain leads to increased levels of asialoGM1 and increased Pseudomonas aeruginosa binding by epithelial cells Am J Respir Cell Mol Biol 1998 19 269 277 9698599
McNamara N Khong A McKemy D Caterina M Boyer J Julius D Basbaum C ATP transduces signals from ASGM1, a glycolipid that functions as a bacterial receptor Proc Natl Acad Sci USA 2001 98 9086 9091 11481474 10.1073/pnas.161290898
Kube D Adams L Perez A Davis PB Terminal sialylation is altered in airway cells with impaired CFTR-mediated chloride transport Am J Physiol Lung Cell Mol Physiol 2001 280 L482 L492 11159032
Lee A Chow D Haus B Tseng W Evans D Fleiszig S Chandy G Machen T Airway epithelial tight junctions and binding and cytotoxicity of Pseudomonas aeruginosa A J Physiol 1999 277 L204 L217
Perez A Risma KA Eckman EA Davis PB Overexpression of R domain eliminates cAMP-stimulated Cl- secretion in 9/HTEo- cells in culture Am J Physiol 1996 271 L85 L92 8760136
Rajan S Cacalano G Bryan R Ratner AJ Sontich CU van Heerckeren A Davis PB Prince A Pseudomonas aeruginosa induction of apoptosis in respiratory epithelial cells: analysis of the effects of cystic fibrosis transmembrane conductance regulator dysfunction and bacterial virulence factors Am J Respir Cell Mol Biol 2000 23 304 312 10970820
Comolli JC Waite LL Mostov KE Engel JN Pili binding to asialo-GM1 on epithelial cells can mediate cytotoxicity or bacterial internalization by Pseudomonas aeruginosa Infect Immun 1999 67 3207 3214 10377092
Cannon CL Kowalski MP Stopak KS Pier GB Pseudomonas aeruginosa-induced apoptosis is defective in respiratory epithelial cells expressing mutant cystic fibrosis transmembrane conductance regulator Am J Respir Cell Mol Biol 2003 29 2 188 197 12878584 10.1165/rcmb.4898
Gottlieb RA Dosanjh A Mutant cystic fibrosis transmembrane conductance regulator inhibits acidification and apoptosis in C127 cells: possible relevance to cystic fibrosis Proc Natl Acad Sci USA 1996 93 8 3587 3591 8622979 10.1073/pnas.93.8.3587
Plotkowski MC De Bentzmann S Pereira SH Zahm JM Bajolet-Laudinat O Roger P Puchelle E Pseudomonas aeruginosa internalization by human epithelial respiratory cells depends on cell differentiation, polarity, and junctional complex integrity Am J Respir Cell Mol Biol 1999 20 880 890 10226058
Fleiszig SM Evans DJ Do N Vallas V Shin S Mostov KE Epithelial cell polarity affects susceptibility to Pseudomonas aeruginosa invasion and cytotoxicity Infect Immun 1997 65 7 2861 7 9199460
Parsons DW Grubb BR Johnson LG Boucher RC Enhanced in vivo airway gene transfer via transient modification of host barrier properties with a surface-active agent Hum Gene Ther 1998 9 18 2661 72 9874264
Coyne CB Kelly MM Boucher RC Johnson LG Enhanced epithelial gene transfer by modulation of tight junctions with sodium caprate Am J Respir Cell Mol Biol 2000 23 602 609 11062138
| 16008840 | PMC1177994 | CC BY | 2021-01-04 16:23:26 | no | Respir Res. 2005 Jul 11; 6(1):69 | utf-8 | Respir Res | 2,005 | 10.1186/1465-9921-6-69 | oa_comm |
==== Front
Virol JVirology Journal1743-422XBioMed Central London 1743-422X-2-491592708410.1186/1743-422X-2-49ResearchPeptide inhibitors of dengue virus and West Nile virus infectivity Hrobowski Yancey M [email protected] Robert F [email protected] Scott F [email protected] Department of Microbiology and Immunology, Tulane University Health Sciences Center, New Orleans, Louisiana 70112 USA2 Graduate Program in Cellular and Molecular Biology, Tulane University, New Orleans, LA 70112 USA3 Biotechnology Program, Florida Gulf Coast University, Fort Myers, FL 33965 USA2005 1 6 2005 2 49 49 20 5 2005 1 6 2005 Copyright © 2005 Hrobowski et al; licensee BioMed Central Ltd.2005Hrobowski 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.
Viral fusion proteins mediate cell entry by undergoing a series of conformational changes that result in virion-target cell membrane fusion. Class I viral fusion proteins, such as those encoded by influenza virus and human immunodeficiency virus (HIV), contain two prominent alpha helices. Peptides that mimic portions of these alpha helices inhibit structural rearrangements of the fusion proteins and prevent viral infection. The envelope glycoprotein (E) of flaviviruses, such as West Nile virus (WNV) and dengue virus (DENV), are class II viral fusion proteins comprised predominantly of beta sheets. We used a physio-chemical algorithm, the Wimley-White interfacial hydrophobicity scale (WWIHS) [1] in combination with known structural data to identify potential peptide inhibitors of WNV and DENV infectivity that target the viral E protein. Viral inhibition assays confirm that several of these peptides specifically interfere with target virus entry with 50% inhibitory concentration (IC50) in the 10 μM range. Inhibitory peptides similar in sequence to domains with a significant WWIHS scores, including domain II (IIb), and the stem domain, were detected. DN59, a peptide corresponding to the stem domain of DENV, inhibited infection by DENV (>99% inhibition of plaque formation at a concentrations of <25 μM) and cross-inhibition of WNV fusion/infectivity (>99% inhibition at <25 μM) was also demonstrated with DN59. However, a potent WNV inhibitory peptide, WN83, which corresponds to WNV E domain IIb, did not inhibit infectivity by DENV. Additional results suggest that these inhibitory peptides are noncytotoxic and act in a sequence specific manner. The inhibitory peptides identified here can serve as lead compounds for the development of peptide drugs for flavivirus infection.
==== Body
Introduction
Enveloped viruses utilize membrane-bound fusion proteins to mediate attachment and entry into specific target host cells. During the virion assembly process, newly synthesized envelope proteins are targeted to the endoplasmic reticulum and golgi apparatus where initial folding and post-transcriptional processing occurs, including multimerization, glycosylation, and proteolysis. This initial folding and processing is required to achieve a conformation where the proteins are held in a metastable state prior to virion release. Post virion release, the multimeric envelope proteins are poised to undergo structural rearrangement leading to fusion of the virion and the new target cell lipid bilayer membranes. Depending on the virus system, the rearrangement trigger can take the form of specific receptor binding, multiple receptor binding, decreased pH following receptor mediated endocytosis, or a combination of triggers.
The prototypic viral envelope fusion protein, the hemagglutinin of influenza virus, contains short alpha helical domains in the trimeric virion configuration. In response to receptor binding and decreased pH, the short helices rearrange with adjoining sequences to produce a longer helix, thus exposing an N-terminal fusion peptide that is believed to interact directly with the target cell membrane. This is followed by a hinge-like bending of the entire complex to adjoin and fuse the two lipid membranes [2,3]. The structural rearrangements that result in extrusion of the fusion peptide and subsequent collapse involve alterations in packing between regions both within individual fusion proteins as well as between monomeric subunits in the trimeric structures. Several disparate viruses, including arenaviruses, coronaviruses, filoviruses, orthomyxoviruses, paramyxoviruses and retroviruses, encode similar proteins that together are classified as class I fusion proteins. These class I viral fusion proteins vary in length and sequence, but are similar in overall structure [4,5].
Qureshi et al. (1990) demonstrated that a peptide from one of the two extended helical domains of the HIV-1 transmembrane protein can block virion infectivity. Subsequently, the FDA approved anti-HIV-1 drug Fuzeon™ (aka DP178, T-20, enfuvirtide) and other N- and C-helix inhibitory peptides were developed [6,7]. These results have greatly motivated the search for other HIV-1 inhibitory peptides [8,9]. Additional peptide mimics of the fusion proteins of other retroviruses, and of orthomyxoviruses, paramyxoviruses, filoviruses, coronaviruses, and herpesviruses have also been identified and shown to inhibit viral entry [10-18]
The envelope fusion proteins of several virus types, including the flaviviruses and alphaviruses, have a structure distinct from class I viral fusion proteins. The envelope glycoprotein (E) of the flavivirus tick-borne encephalitis virus (TBEV) consists of three domains: a structurally central amino terminal domain (domain I), a dimerization domain (domain II) and a carboxyl terminal Ig-like domain (domain III), all containing predominantly beta sheet folds [19]. The primary sequence of E1, the fusion protein of Semliki Forest virus, an alphavirus, revealed a remarkable fit to the scaffold of TBEV E [20] suggesting the existence of a second class of viral fusion proteins. The dengue virus (DENV) E protein has also been shown to have a class II structure [21]. Recent studies of flavivirus virions and proteins by cryoelectron microscopy and crystal structure analysis have lead to a greatly increased understanding of the function of these class II viral envelope proteins, including the structural rearrangements they undergo during maturation, triggering and fusion [21-28].
The flaviviruses, which include DENV, West Nile virus (WNV), yellow fever virus, Japanese encephalitis virus (JEV), and TBEV, among others, are transmitted between vertebrate hosts by insect vectors. The most serious manifestations of DENV infection are dengue hemorrhagic fever (DHF) and dengue shock syndrome (DSS). There are four serotypes of DENV (1–4), which together cause an estimated 50 million human infections per year [29], and each can cause DF, DHF or DSS. Because of the phenomenon of antibody-dependent enhancement (ADE), or other immune phenomena, protection against one DENV serotype increases the risk of DHF or DSS when the individual is exposed to another serotype [30-32]. Cross-reactive, but non-neutralizing antibodies can mediate entry of DENV into macrophages, dendritic cells and other viral target cells via Fc receptors, increasing virus titers and thus pathology. Multivalent DENV vaccines have shown some promise in humans [32-39] and in nonhuman primate studies [40,41], but face several obstacles. Antiviral drugs, which target each of the four serotypes of DENV without enhancing pathogenesis of any serotype, are urgently needed. The recent introduction of WNV in the United States further highlights the public health challenges posed by flaviviruses. No effective vaccine or antiviral drug therapy is currently available against either DENV or WNV.
Although there are many differences between the structures of class I and class II viral fusion proteins, we hypothesized that they function through a similar membrane fusion mechanism involving rearrangements of domains, and that peptides mimicking portions of class II viral fusion proteins would inhibit virion fusion and entry steps thereby serving as lead compounds for the development of antivirals. We used a physio-chemical algorithm, the Wimley-White interfacial hydrophobicity scale [1] in combination with known structural data to predict regions of the DENV and WNV E proteins that may play roles in protein-protein rearrangements or bilayer membrane interactions during the entry and fusion process. Several of these peptides specifically inhibit DENV or WNV infection.
Results
Identification of Flavivirus inhibitory peptides
The domains that precede the transmembrane anchors of most class I fusion proteins are not highly hydrophobic, however, they usually contain a cluster of aromatic amino acids and display a tendency to partition into bilayer membranes, as revealed by analyses using the experimentally-determined Wimley-White interfacial hydrophobicity scale (WWIHS) [42]. Fuzeon's corresponding sequence overlaps the aromatic pre-anchor domain of HIV-1 TM. Synthetic peptides corresponding to other domains of class I viral fusion proteins with significant WWIHS scores may also inhibit viral infectivity [43]. Previously, we suggested that peptide drugs analogous to Fuzeon might be developed for HCV and other members of the Flaviviridae [44]. DENV E contains five domains with significant WWIHS scores (Fig. 1, WWIHS sequences in black). These include the fusion peptide domain, a portion of subdomain IIb, the pre-anchor stem region following domain III, and the transmembrane domain. Sequences with high WWIHS scores are similarly located in the X-ray structures of WNV E and alphavirus (SFV and Sindbis virus – SINV) E1, and potentially also in the putative class II fusion proteins of hepatitis C virus (HCV), pestiviruses and bunyaviruses [44,45]. Regions with high WWIHS scores are predicted to play a role in protein-protein interactions during structural rearrangements or protein-lipid interactions during bilayer fusion, and we predicted that synthetic peptides corresponding to these regions may have the potential to inhibit flavivirus infectivity.
Figure 1 Diagramatic structure of DENV envelope protein showing inhibitory peptide regions. Grey lines: dicysteine linkages. Black stick figures: N-glycosylation sites. Regions with significant Wimley-White interfacial hydrophobicity scale scores were predicted with MPeX (Boxed in left depiction; black in right depiction). Sequences of DENV peptides and the location of WNV homologs are indicated.
To test this hypothesis, an initial set of synthetic peptides representing sequences of DENV E and WNV E with significant WWIHS scores was synthesized and screened for the ability to inhibit plaque formation by these flaviviruses (Table 1). Peptides corresponding to the transmembrane domain were not tested because this region is not exposed during the entry process. Initial assays for inhibitory activity were performed using the highest concentration of each peptide that could be obtained in aqueous solution with a maximum of 1% DMSO (between 29 and 128 μM). Plaque reduction in which the inhibitor is removed after virus adsorption is the most stringent test of an antiviral agent. Prior to initiating these studies, we developed a new immunoplaque assay for DENV and WNV. Approximately 200 focus forming units (FFU) of either WNV or DENV were preincubated with each of the peptides and used to infect monolayers of LLCKM-2 monkey kidney epithelial cells. The number of resulting viral foci was determined from three experiments and normalized to a no-peptide control to calculate the percent inhibition. Our screening of this initial set detected several peptides that were able to inhibit infection by DENV or WNV (Table 1, Fig. 2). Peptides similar in sequence to domains with a significant WWIHS scores, including domain II (IIb) (WN53 and WN83), and stem domain (DN59), were found to have inhibitory activity.
Table 1 Initial peptides synthesized and tested for inhibitory activity.
Peptide Sequence Locationa Concentration (μM) % Inhibition +/- SD
DN80 MVDRGWGNHAGLFGKGSIV 386–400 49.9 17 +/- 10
DN57 AWLVHTQWFLDLPLPWLPGADTQGSNWI 485–503 30.6 7 +/- 4
DN81 AWLVHRQWFLDLPLPWLPG 485–512 42.6 25 +/- 8
DN59 MAILGDTAWDFGSLGGVFTSIGKALHQVFGAIY 692–724 29.0 93 +/- 2
WN82 VVDRGWGNGAGLFGKGSID 396–410 52.5 4 +/- 13
WN53 TFLVHREWFMDLNLPWSSAGSTVWR 500–518 98.7 56 +/- 5
WN83 TFLVHREWFMDLNLPWSSA 500–524 128.0 70 +/- 2
a numbering from the beginning of the E polyprotein in either DENV or WNV.
Figure 2 Plaque inhibition assay. (A) Preincubation of DENV with neutralizing antiserum reduces plaque number by 74%. (B) Preincubation of DENV with a non-inhibitory peptide shows no reduction in plaques. (C) Preincubation of DENV with one of the inhibitory peptides (DN59) shows a greater than 95% reduction in plaques.
Determination of 50% inhibitory concentrations
Dose-response curves were determined for the most potent of the peptides WN53 and WN83 against WNV and also for peptide DN59 against DENV (Fig. 3). The WN53 peptide showed a maximum inhibitory activity against WNV of 56.0 +/- 3.0% (mean +/- SD) at 99 μM. The inhibitory activity decreased with decreasing concentration with a 50% inhibitory concentration (IC50) at roughly 10 μM. The WN83 peptide showed a maximum inhibitory activity against WNV of 70.0 +/- 3.0% at 128 μM. The inhibitory activity decreased with decreasing concentration with an IC50 of roughly 10 μM. The DN59 peptide showed a maximum inhibitory activity against DENV of 100.0 +/- 0.5% at 20 μM. The inhibitory activity decreased with decreasing concentration with an IC50 of at roughly 10 μM. DENV stem peptide 59 (DN59) and WNV peptides 53 and 83 (WN53, WN83) reproducibly inhibited infectivity at low μM concentrations.
Figure 3 Dose-response curves for WN53, WN83 and DN59 peptides. (A) Increasing concentrations of peptide WN53 produce a corresponding increase in WNV inhibition with an IC50 in the 10μ range. (B) Increasing concentrations of peptide WN83 produce a corresponding increase in WNV inhibition with an IC50 in the 10 μM range. (C) Increasing concentrations of peptide DN59 also produce a corresponding increase in DNV inhibition with an IC50 in the 10 μM range. All measurements were made in triplicate, with mean +/- SD shown.
Specificity of peptide inhibitory activity
The DN59 peptide matches a pre-anchor domain sequence that is highly conserved among insect-transmitted flaviviruses. DN59 inhibited infection by DENV (>99% inhibition of plaque formation at a concentrations of <25 μM). Cross-inhibition of WNV fusion/infectivity (>99% inhibition at <25 μM) was also reproducibly demonstrated with DN59 (Fig. 6). However, WNV inhibitory peptide WN83 did not inhibit infectivity by DENV.
Figure 6 Inhibitory effect of peptides WN53 and DN59 alone and in combination. WN53 and DN59 peptides were tested alone (■ DN59 alone, ● WN53 alone) or together (◆) with WNV and inhibitory activity at three concentrations was measured (mean of three trials +/- SD shown). DN59 and WN53 together have an effect intermediate between the two peptides alone.
To determine if these peptides specifically inhibit infectivity of the viruses for which they were designed, the WN53, WN83 and DN59 peptides were tested for inhibitory effects against Sindbis virus (SINV), an alphavirus that encodes a class II fusion protein. None of the peptides showed a statistically significant effect against SINV infectivity (Fig. 4). Peptides with the same amino acid composition as WN83 and DN59, but a scrambled sequence (Scrambled WN83: VATWHLDWSREFPLFLMNS; Scrambled DN59: YFIDTSGAIWGASHLTGVLFDFMGIQGGAVLAK) were synthesized and tested for the ability to inhibit infection by WNV and DENV respectively. Neither scrambled peptide significantly inhibited infection by these viruses (Fig. 5). These results provide evidence that the action of these inhibitory peptides not due to general inactivation of enveloped virions and is sequence specific.
Figure 4 Effect of DENV and WNV specific peptides against another virus. Peptides DN59, WN83 and WN53 at 100 μg/ml concentrations were tested for inhibitory activity against the alphavirus SINV in a similar plaque reduction assay. Results are shown as the mean of three trials +/- SEM. None of the peptides showed a statistically significant inhibitory effect against SINV (ANOVA, p = 0.705, with Dunnett's posthoc test).
Figure 5 Effect of scrambled peptide order on inhibitory function. Scrambled versions of peptides DN59 and WN83 at 100 μg/ml concentrations were tested for inhibitory effect against DENV and WNV, respectively. Scrambled versions of the peptides showed no inhibitory activity compared to the original DN59 and WN83 peptides.
Peptide toxicity
It is possible that inhibitory peptides induce cellular alterations or toxicity that can block flavivirus entry or other steps in the replication cycle. To address this possibility, LLCMK-2 monkey kidney epithelial cell monolayers were exposed to 100 mg/ml concentrations of WN53, WN83 and DN59 peptides for 24 hrs, and cell viability was assayed with an MTT assay. No statistical difference was observed between the viability of control cells versus cells exposed to the peptides or DMSO (Fig. 7). This result suggests that these inhibitory peptides are not blocking infectivity via effects on host cell metabolism or viability.
Figure 7 Toxicity of inhibitory peptides in cell culture. MTT assays for cell viability were performed after 24 hr incubation of cells with 100 μg/ml of peptides WN83, WN53, and DN59 (mean from three experiments +/- SEM). No statistically significant differences in cell viability were observed (ANOVA p = 0.672, with Dunnett's posthoc test).
Non-synergistic activity of combined peptides
When added in combination, peptides that block entry at different steps or that target different domains may produce greater inhibition of DENV-2 infectivity than either peptide alone. Synergistic or antagonistic effects are also possible, if a peptide that alters protein-protein interactions allows greater or lesser access to E domains targeted by another peptide. Since the WN53, WN83 and DN59 peptides all inhibited WNV entry, the possibility of antagonistic or synergistic function was examined by testing WN53 and DN59 alone or in combination at three concentrations (5, 10 and 20 μM). At all three concentrations, the peptide combination was more effective than WN53 alone, but less effective than DN59 alone. This indicates that the activity of the WN53 peptide has an antagonistic effect on the function of the DN59 peptide (Fig. 6).
Discussion
Synthetic peptides corresponding to sequences in DENV and WNV E proteins were identified that inhibited infectivity of these viral pathogens of major public health importance. The inhibitory effects of these peptides were dose dependent with IC50s in the range of 10 μM. Several of the most potent of these peptides showed no inhibitory activity against SINV, an alphavirus that possesses a class II viral fusion protein with a similar overall structure as flavivirus E. Scrambled peptides with the same amino acid composition as the inhibitory peptides, but with a different primary sequence, failed to inhibit DENV and WNV infection. None of the DENV or WNV inhibitory peptides induced gross cytopathic effects, killing of cultured cells or showed evidence of in vitro cellular toxicity. These results indicate that these inhibitory peptides function through a sequence-specific mechanism and are not merely cytotoxic.
Membrane fusion by both class I and II viral fusion proteins is initiated by interaction of the fusion peptide with the target cell membrane. In class I viral fusion proteins, a subsequent rearrangement of a trimer of the proteins, each with two α helices, to form a six-helix bundle brings the viral and cell membranes into closer proximity. Inhibitors of viruses with class I fusion proteins, such as Fuzeon™ that mimic a portion of one or the other of the two α helices, interfere with a step proximal to six-helix bundle formation possibly by forming an inactive aggregate with the opposite helix. Recent studies indicate that after insertion of the fusion peptide, class II viral fusion proteins likewise undergo rearrangements. In this case, intraprotein interactions may occur between the stem domain and domains I, II and/or III [21-28,46]. According to this model, the viral and cellular membranes are brought closer by interactions of the stem with other portions of E, resulting in bilayer fusion. DN59, WN53 and WN83 peptides may interfere with the intramolecular interactions between the stem and other portions of class II viral fusion proteins, a possibility suggested previously [23,24,46].
Two of the inhibitory peptides (WN53 and WN83) are designed from overlapping regions of the E protein domain I/II junction and are specifically inhibitory against WNV. Recently, other investigators have hypothesized that small molecule inhibitors to this domain I/II junction region might be developed. Modis et al (2003; 2004) predicted that interactions near this region (the k-l loop) that are involved in the rotational changes between these domains might be blocked by small molecule inhibitors. However, our similar peptides designed from the analogous region of the DENV E protein (D57 and D81) failed to inhibit DENV infectivity.
The possibility that WN53, WN83 and DN59 interact with some target cell surface component to exert their inhibitory effects cannot be ruled out. However, the majority of flavivirus neutralizing antibodies that appear to be involved in receptor blocking bind to domain III, and soluble domain III itself can block flavivirus entry, apparently through competition for cellular receptors [47-51]. In contrast, the domains that correspond to WN53, WN83 and DN59 peptides, IIb and the pre-anchor stem, appear to be involved in structural rearrangements during fusion, rather than direct interactions with cellular receptors. Interestingly, previous observations indicate that some monoclonal antibodies block virion entry at a post attachment step, indicating that they may interfere with conformational changes necessary for fusion [52]. That possibility that some antibodies can gain access to regions important for conformational changes and block these changes suggests that these inhibitory peptide regions might be candidates for novel vaccine designs that either utilize the inhibitory peptide regions directly as antigens, or target the regions that interact with the inhibitory peptides. Further studies are needed to define the exact mechanism of inhibition of these DENV and WNV peptides, and the specific nature or location of their interactions with viral targets.
The DN59 peptide is inhibitory against DENV as well as WNV. The corresponding pre-anchor region is highly conserved between DENV and WNV as well as among other flaviviruses (Table 2) and probably functions in a similar manner during entry of all flaviviruses [53]. Thus, DN59 or similar peptides may act as broad-spectrum flavivirus inhibitors. Other flaviviruses considered potential bioterrorism agents, including JEV, Kyasanur Forest disease virus and TBEV, may also be inhibited by DN59, a DN59 derivative, or by an analogous peptide. Unlike proposed DENV vaccines, which must be multivalent (ie. simultaneously effective against each of the four DENV serotypes because of the phenomenon of antibody dependent enhancement), peptide drugs targeting the highly conserved stem conserved motifs in flavivirus E may demonstrate cross-strain efficacy.
Table 2 Alignment of preanchor domain sequences from representative flaviviruses.
Virusa Pre-anchor sequenceb Locationc
DENV-1 AILGDTAWDFGSIGGVFTSVGKLIHQIFGTA 693–723
DENV-2 AILGDTAWDFGSLGGVFTSIGKALHQVFGAI 693–723
DENV-3 AILGDTAWDFGSVGGVLNSLGKMVHQIFGSA 691–721
DENV-4 AILGETAWDFGSVGGVLTSLGKAVHQVFGSV 692–722
WNV AVLGDTAWDFGSVGGVFTSVGKAVHQVFGGA 709–739
YFV AVMGDTAWDFSSAGGFFTSVGKGIHTVFGSA 695–725
SLEV AVLGDTAWDFGSIGGVFTSIGKALHQVFGGA 707–737
JEV AALGDTAWDFGSIGGVFNSIGKAVHQVFGGA 712–742
TBEV TVIGEHAWDFGSAGGFLSSIGKAVHTVLGGA 694–724
OHFV TVLGEHAWDFGSTGGFLSSIGKALHTVLGGA 694–724
KFV TVVGEHAWDFGSVGGMLSSVGKALHTAFGAA 695–725
POWV SVVGEHAWDFGSVGGVLSSVGKAIHTVLGGA 693–723
aDENV: dengue virus; WNV: West Nile virus; YFV: yellow fever virus; SLEV: St. Louis encephalitis virus; JEV: Japanese encephalitis virus; TBEV: Tick-borne encephalitis virus; OHFV: Omsk hemorrhagic fever virus; KFV: Kyasanur Forest virus; POWV: Powassan virus
b * = identical or chemically similar amino acids in every sequence
cnumbering from the beginning of the E polyprotein.
IC50s in the μM range have been considered promising for class I viral fusion protein inhibitor development [54,55]. Thus, the peptides identified here can serve as lead compounds that may be developed as peptide drugs against the four serotypes of DENV, WNV and potentially other flaviviruses. We anticipate that such peptide inhibitors may be as successful as the HIV-1 inhibitory peptide Fuzeon™. Unlike persistant HIV infections, immune responses against DENV and other flaviviruses are capable of clearing the viruses in individuals that survive the initial infection. By reducing the viral load during the initial stages of infection, it may be possible to extend the window of time during which an immune response could arise, and thus enable more individuals to control, eliminate and survive infections by these agents. Evidence for the ability to therapeutically intervene in flavivirus-induced diseases has been demonstrated with the recent observation that administration of neutralizing antibodies against WNV can be curative, even after symptom initiation [56]. Development of resistant mutants will be a concern, but should be a less problematic than in the case of long-term treatment of persistent retroviral infections.
It is worth noting that the HIV inhibitor Fuzeon™, was initially identified using a predictive strategy without the availability of structural data [6,7,57]. The fact that we developed these peptides using a predictive algorithm validates our approach as well as the accuracy of the flavivirus E protein structural data. A similar approach may be useful for the large number of other viruses with class II envelope fusion proteins with or without known structures.
Materials and methods
Design and synthesis of peptides
Sequences of DENV and WNV E with positive Wimley-White interfacial hydrophobicity scale scores were determined using the program Membrane Protein eXplorer [1]. After consideration of the known secondary structures for several subdomains of E, selected peptides were synthesized by solid-phase conventional N-α-9-flurenylmethyloxycarbonyl chemistry (Genemed Synthesis, San Francisco, CA). Peptides were purified by reverse-phase high performance liquid chromatography and confirmed by amino acid analysis and electrospray mass spectrometry. Peptide stock solutions were prepared in 20% (v/v) dimethyl sulfoxide (DMSO): 80% (v/v) H20, and concentrations determined by absorbance of aromatic side chains at 280 nm. Scrambled peptides sequences were obtained by drawing from a hat.
Viruses and Cells
DENV strain New Guinea-2 and WNV strain Egypt 101 were obtained from R. Tesh at the World Health Organization Arbovirus Reference Laboratory at the University of Texas at Galveston. DENV and WNV were propagated in the African green monkey kidney epithelial cell line, LLCKM-2, a gift of K. Olsen at Colorado State University. Sindbis virus (SINV) containing the enhanced green fluorescent protein (EGFP) protein expression cassette was obtained from K. Ryman at Louisiana State University at Shreveport and was propagated in baby hamster kidney cells. All cells were grown in Dulbecco's modified eagle medium (DMEM) with 10% (v/v) fetal bovine serum (FBS), 100 U/ml penicillin G and 100 mg/ml streptomycin, at 37°C with 5% (v/v) CO2.
Viral plaque reduction assays
LLCKM-2 target cells were seeded at a density of 3 × 105 cells in each well of a 6-well plate 48 h prior to infection. Approximately 200-focus forming units (FFU) of DENV, WNV, or SINV/EGFP were incubated with or without peptides in serum-free DMEM for 1 h at rt. Virus/peptide or virus/control mixtures were allowed to infect confluent LLCKM-2 monolayers for 1 h at 37°C, after which time the medium was removed and the cells were washed once with phosphate buffered saline (PBS) and overlaid with fresh DMEM/10% (v/v) FBS containing 0.85% (w/v) SeaPlaque Agarose (Cambrex Bio Science, Rockland, ME). Cells were then incubated at 37°C with 5% CO2 for 1 day (Sindbis virus), 3 days (WNV) or 6 days (DNV). Sindbis virus infections were quantified by directly counting green fluorescing foci. Cultures infected with DENV were fixed with 10% formalin overnight at 4°C and permeablized with 70% (v/v) ethanol prior to immunostaining and visualization using a human polyclonal anti-flavivirus antibody (a gift of V. Vorndam, CDC, San Juan) followed by horseradish peroxidase (HRP) conjugated goat anti-human immunoglobulin (Pierce, Rockford, IL) and AEC chromogen substrate (Dako, Carpinteria, CA). WNV plaques were similarly visualized using a mouse anti-WNV antibody (Chemicon, Temecula, CA) and an HRP conjugated goat anti-mouse antibody (Dako, Carpinteria, CA).
Toxicity assay
Peptide cytotoxicity was measured using the TACS™ MTT cell proliferation assay (R&D systems, Minneapolis, MN) according to the manufacturer's instructions.
Acknowledgements
The authors would like to thank S. Isern, B. Sainz, W. Wimley and W. Gallaher for helpful discussions and technical assistance.
==== Refs
White SH Snider C Jaysinghe S Kim J Membrane Protein Explorer version 2.2a http://blancobiomoluciedu/mpex/ 2003
Carr CM Kim PS A spring-loaded mechanism for the conformational change of influenza hemagglutinin Cell 1993 73 823 832 8500173 10.1016/0092-8674(93)90260-W
Wilson IA Skehel JJ Wiley DC Structure of the haemagglutinin membrane glycoprotein of influenza virus at 3 A resolution Nature 1981 289 366 373 7464906 10.1038/289366a0
Gallaher WR Similar structural models of the transmembrane glycoproteins of Ebola and avian sarcoma viruses Cell 1996 85 1 2 8620528 10.1016/S0092-8674(00)81248-9
Gallaher WR Ball JM Garry RF Griffin MC Montelaro RC A general model for the transmembrane proteins of HIV and other retroviruses AIDS Res Hum Retroviruses 1989 5 431 440 2788443
Wild CT Shugars DC Greenwell TK McDanal CB Matthews TJ Peptides corresponding to a predictive alpha-helical domain of human immunodeficiency virus type 1 gp41 are potent inhibitors of virus infection Proc Natl Acad Sci U S A 1994 91 9770 9774 7937889
Wild C Greenwell T Matthews T A synthetic peptide from HIV-1 gp41 is a potent inhibitor of virus-mediated cell-cell fusion AIDS Research & Human Retroviruses 1993 9 1051 1053 8312047
Pozniak A HIV fusion inhibitors J HIV Ther 2001 6 91 94 11740409
Sodroski JG HIV-1 entry inhibitors in the side pocket Cell 1999 99 243 246 10555140 10.1016/S0092-8674(00)81655-4
Lambert DM Barney S Lambert AL Guthrie K Medinas R Davis DE Bucy T Erickson J Merutka G Petteway SRJ Peptides from conserved regions of paramyxovirus fusion (F) proteins are potent inhibitors of viral fusion Proc Natl Acad Sci U S A 1996 93 2186 2191 8700906 10.1073/pnas.93.5.2186
Young JK Li D Abramowitz MC Morrison TG Interaction of peptides with sequences from the Newcastle disease virus fusion protein heptad repeat regions J Virol 1999 73 5945 5956 10364347
Watanabe S Takada A Watanabe T Ito H Kida H Kawaoka Y Functional importance of the coiled-coil of the Ebola virus glycoprotein J Virol 2000 74 10194 10201 11024148 10.1128/JVI.74.21.10194-10201.2000
Bosch BJ van der Zee R de Haan CA Rottier PJ The coronavirus spike protein is a class I virus fusion protein: structural and functional characterization of the fusion core complex J Virol 2003 77 8801 8811 12885899 10.1128/JVI.77.16.8801-8811.2003
Bultmann H Brandt CR Peptides containing membrane-transiting motifs inhibit virus entry J Biol Chem 2002 277 36018 36023 12130644 10.1074/jbc.M204849200
Markosyan RM Bates P Cohen FS Melikyan GB A study of low pH-induced refolding of Env of avian sarcoma and leukosis virus into a six-helix bundle Biophys J 2004 87 3291 3298 15339808 10.1529/biophysj.104.047696
Rapaport D Ovadia M Shai Y A synthetic peptide corresponding to a conserved heptad repeat domain is a potent inhibitor of Sendai virus-cell fusion: an emerging similarity with functional domains of other viruses Embo J 1995 14 5524 5531 8521809
Okazaki K Kida H A synthetic peptide from a heptad repeat region of herpesvirus glycoprotein B inhibits virus replication J Gen Virol 2004 85 2131 2137 15269351 10.1099/vir.0.80051-0
Yao Q Compans RW Peptides corresponding to the heptad repeat sequence of human parainfluenza virus fusion protein are potent inhibitors of virus infection Virology 1996 223 103 112 8806544 10.1006/viro.1996.0459
Rey FA Heinz FX Mandl C Kunz C Harrison SC The envelope glycoprotein from tick-borne encephalitis virus at 2 A resolution Nature 1995 375 291 298 7753193 10.1038/375291a0
Lescar J Roussel A Wien MW Navaza J Fuller SD Wengler G Rey FA The Fusion glycoprotein shell of Semliki Forest virus: an icosahedral assembly primed for fusogenic activation at endosomal pH Cell 2001 105 137 148 11301009 10.1016/S0092-8674(01)00303-8
Kuhn RJ Zhang W Rossmann MG Pletnev SV Corver J Lenches E Jones CT Mukhopadhyay S Chipman PR Strauss EG Baker TS Strauss JH Structure of dengue virus: implications for flavivirus organization, maturation, and fusion Cell 2002 108 717 725 11893341 10.1016/S0092-8674(02)00660-8
Zhang Y Corver J Chipman PR Zhang W Pletnev SV Sedlak D Baker TS Strauss JH Kuhn RJ Rossmann MG Structures of immature flavivirus particles Embo J 2003 22 2604 2613 12773377 10.1093/emboj/cdg270
Bressanelli S Stiasny K Allison SL Stura EA Duquerroy S Lescar J Heinz FX Rey FA Structure of a flavivirus envelope glycoprotein in its low-pH-induced membrane fusion conformation Embo J 2004 23 728 738 14963486 10.1038/sj.emboj.7600064
Modis Y Ogata S Clements D Harrison SC Structure of the dengue virus envelope protein after membrane fusio Nature 2004 427 313 319 14737159 10.1038/nature02165
Modis Y Ogata S Clements D Harrison SC A ligand-binding pocket in the dengue virus envelope glycoprotein Proc Natl Acad Sci U S A 2003 100 6986 6991 12759475 10.1073/pnas.0832193100
Mukhopadhyay S Kim BS Chipman PR Rossmann MG Kuhn RJ Structure of West Nile virus Science 2003 302 248 14551429 10.1126/science.1089316
Mukhopadhyay S Kuhn RJ Rossmann MG A structural perspective of the flavivirus life cycle Nat Rev Microbiol 2005 3 13 22 15608696 10.1038/nrmicro1067
Rey FA Dengue virus envelope glycoprotein structure: new insight into its interactions during viral entry Proc Natl Acad Sci U S A 2003 100 6899 6901 12782795 10.1073/pnas.1332695100
WHO Dengue and dengue haemorrhagic fever 2002 http://www.who.int/mediacentre/factsheets/fs117/en/,
Hotta H Sanchez LF Takada H Homma M Kotani S Enhancement of dengue virus infection in cultured mouse macrophages by lipophilic derivatives of muramyl peptides Microbiol Immunol 1985 29 533 541 4046890
Halstead SB Pathogenesis of dengue: challenges to molecular biology Science 1988 239 476 481 3277268
Guy B Chanthavanich P Gimenez S Sirivichayakul C Sabchareon A Begue S Yoksan S Luxemburger C Lang J Evaluation by flow cytometry of antibody-dependent enhancement (ADE) of dengue infection by sera from Thai children immunized with a live-attenuated tetravalent dengue vaccine Vaccine 2004 22 3563 3574 15315835 10.1016/j.vaccine.2004.03.042
Sabchareon A Lang J Chanthavanich P Yoksan S Forrat R Attanath P Sirivichayakul C Pengsaa K Pojjaroen-Anant C Chokejindachai W Jagsudee A Saluzzo JF Bhamarapravati N Safety and immunogenicity of tetravalent live-attenuated dengue vaccines in Thai adult volunteers: role of serotype concentration, ratio, and multiple doses Am J Trop Med Hyg 2002 66 264 272 12139219
Edelman R Wasserman SS Bodison SA Putnak RJ Eckels KH Tang D Kanesa-Thasan N Vaughn DW Innis BL Sun W Phase I trial of 16 formulations of a tetravalent live-attenuated dengue vaccine Am J Trop Med Hyg 2003 69 48 60 14740955
Kanesa-Thasan N Sun W Ludwig GV Rossi C Putnak JR Mangiafico JA Innis BL Edelman R Atypical antibody responses in dengue vaccine recipients Am J Trop Med Hyg 2003 69 32 38 14740953
Kanesa-thasan N Sun W Kim-Ahn G Van Albert S Putnak JR King A Raengsakulsrach B Christ-Schmidt H Gilson K Zahradnik JM Vaughn DW Innis BL Saluzzo JF Hoke CHJ Safety and immunogenicity of attenuated dengue virus vaccines (Aventis Pasteur) in human volunteers Vaccine 2001 19 3179 3188 11312014 10.1016/S0264-410X(01)00020-2
Bhamarapravati N Sutee Y Live attenuated tetravalent dengue vaccine Vaccine 2000 18 Suppl 2 44 47 10821973 10.1016/S0264-410X(00)00040-2
Barrett AD Current status of flavivirus vaccines Ann N Y Acad Sci 2001 951 262 271 11797783
Sun W Edelman R Kanesa-Thasan N Eckels KH Putnak JR King AD Houng HS Tang D Scherer JM Hoke CHJ Innis BL Vaccination of human volunteers with monovalent and tetravalent live-attenuated dengue vaccine candidates Am J Trop Med Hyg 2003 69 24 31 14740952
Guirakhoo F Pugachev K Zhang Z Myers G Levenbook I Draper K Lang J Ocran S Mitchell F Parsons M Brown N Brandler S Fournier C Barrere B Rizvi F Travassos A Nichols R Trent D Monath T Safety and efficacy of chimeric yellow Fever-dengue virus tetravalent vaccine formulations in nonhuman primates J Virol 2004 78 4761 4775 15078958 10.1128/JVI.78.9.4761-4775.2004
Guirakhoo F Pugachev K Arroyo J Miller C Zhang ZX Weltzin R Georgakopoulos K Catalan J Ocran S Draper K Monath TP Viremia and immunogenicity in nonhuman primates of a tetravalent yellow fever-dengue chimeric vaccine: genetic reconstructions, dose adjustment, and antibody responses against wild-type dengue virus isolates Virology 2002 298 146 159 12093182 10.1006/viro.2002.1462
Wimley WC White SH Experimentally determined hydrophobicity scale for proteins at membrane interfaces Nat Struct Biol 1996 3 842 848 8836100 10.1038/nsb1096-842
Giannecchini S Bonci F Pistello M Matteucci D Sichi O Rovero P Bendinelli M The membrane-proximal tryptophan-rich region in the transmembrane glycoprotein ectodomain of feline immunodeficiency virus is important for cell entry Virology 2004 320 156 166 15003871 10.1016/j.virol.2003.12.001
Garry RF Dash S Proteomics computational analyses suggest that hepatitis C virus E1 and pestivirus E2 envelope glycoproteins are truncated class II fusion proteins Virology 2003 307 255 265 12667795 10.1016/S0042-6822(02)00065-X
Garry CE Garry RF Proteomics computational analyses suggest that the carboxyl terminal glycoproteins of Bunyaviruses are class II viral fusion protein (beta-penetrenes) Theor Biol Med Model 2004 1 10 15544707 10.1186/1742-4682-1-10
Gibbons DL Vaney MC Roussel A Vigouroux A Reilly B Lepault J Kielian M Rey FA Conformational change and protein-protein interactions of the fusion protein of Semliki Forest virus Nature 2004 427 320 325 14737160 10.1038/nature02239
Beasley DW Barrett AD Identification of neutralizing epitopes within structural domain III of the West Nile virus envelope protein J Virol 2002 76 13097 13100 12438639 10.1128/JVI.76.24.13097-13100.2002
Volk DE Beasley DW Kallick DA Holbrook MR Barrett AD Gorenstein DG Solution structure and antibody binding studies of the envelope protein domain III from the New York strain of West Nile virus J Biol Chem 2004 279 38755 38761 15190071 10.1074/jbc.M402385200
Chen Y Maguire T Marks RM Demonstration of binding of dengue virus envelope protein to target cells J Virol 1996 70 8765 8772 8971005
Bhardwaj S Holbrook M Shope RE Barrett AD Watowich SJ Biophysical characterization and vector-specific antagonist activity of domain III of the tick-borne flavivirus envelope protein J Virol 2001 75 4002 4007 11264392 10.1128/JVI.75.8.4002-4007.2001
Wu KP Wu CW Tsao YP Kuo TW Lou YC Lin CW Wu SC Cheng JW Structural basis of a Flavivirus recognized by its neutralizing antibody: Solution structure of the domain III of the Japanese Encephalitis virus envelope protein J Biol Chem 2003
Hung SL Lee PL Chen HW Chen LK Kao CL King CC Analysis of the steps involved in Dengue virus entry into host cells Virology 1999 257 156 167 10208929 10.1006/viro.1999.9633
Allison SL Stiasny K Stadler K Mandl CW Heinz FX Mapping of functional elements in the stem-anchor region of tick-borne encephalitis virus envelope protein E J Virol 1999 73 5605 5612 10364309
Liu S Xiao G Chen Y He Y Niu J Escalante CR Xiong H Farmar J Debnath AK Tien P Jiang S Interaction between heptad repeat 1 and 2 regions in spike protein of SARS-associated coronavirus: implications for virus fusogenic mechanism and identification of fusion inhibitors Lancet 2004 363 938 947 15043961 10.1016/S0140-6736(04)15788-7
Sia SK Carr PA Cochran AG Malashkevich VN Kim PS Short constrained peptides that inhibit HIV-1 entry Proc Natl Acad Sci U S A 2002 99 14664 14669 12417739 10.1073/pnas.232566599
Oliphant T Engle M Nybakken GE Doane C Johnson S Huang L Gorlatov S Mehlhop E Marri A Chung KM Ebel GD Kramer LD Fremont DH Diamond MS Development of a humanized monoclonal antibody with therapeutic potential against West Nile virus Nat Med 2005 11 522 30 15852016 10.1038/nm1240
Qureshi NM Coy DH Garry RF LA H Characterization of a putative cellular receptor for HIV-1 transmembrane glycoprotein using synthetic peptides AIDS 1990 4 553 558 1974767
| 15927084 | PMC1177995 | CC BY | 2021-01-04 16:38:59 | no | Virol J. 2005 Jun 1; 2:49 | utf-8 | Virol J | 2,005 | 10.1186/1743-422X-2-49 | oa_comm |
==== Front
BMC AnesthesiolBMC Anesthesiology1471-2253BioMed Central London 1471-2253-5-71593875710.1186/1471-2253-5-7Research ArticleHypothermia in a surgical intensive care unit Abelha Fernando J [email protected] Maria A [email protected] Aida M [email protected] Nuno M [email protected] Cristina C [email protected] Department of Anesthesia and Intensive Care, Hospital de São João, Porto, Portugal2 Biostatistics and Medical Informatics Department, Faculty of Medicine at the University of Porto, Portugal2005 6 6 2005 5 7 7 15 1 2005 6 6 2005 Copyright © 2005 Abelha et al; licensee BioMed Central Ltd.2005Abelha 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
Inadvertent hypothermia is not uncommon in the immediate postoperative period and it is associated with impairment and abnormalities in various organs and systems that can lead to adverse outcomes. The aim of this study was to estimate the prevalence, the predictive factors and outcome of core hypothermia on admission to a surgical ICU.
Methods
All consecutive 185 adult patients who underwent scheduled or emergency noncardiac surgery admitted to a surgical ICU between April and July 2004 were admitted to the study. Tympanic membrane core temperature (Tc) was measured before surgery, on arrival at ICU and every two hours until 6 hours after admission. The following variables were also recorded: age, sex, body weight and height, ASA physical status, type of surgery, magnitude of surgical procedure, anesthesia technique, amount of intravenous fluids administered during anesthesia, use of temperature monitoring and warming techniques, duration of the anesthesia, ICU length of stay, hospital length of stay and SAPS II score. Patients were classified as either hypothermic (Tc ≤ 35°C) or normothermic (Tc> 35°C). Univariate analysis and multiple regression binary logistic with an odds ratio (OR) and its 95% Confidence Interval (95%CI) were used to compare the two groups of patients and assess the relationship between each clinical predictor and hypothermia. Outcome measured as ICU length of stay and mortality was also assessed.
Results
Prevalence of hypothermia on ICU admission was 57.8%. In univariate analysis temperature monitoring, use of warming techniques and higher previous body temperature were significant protective factors against core hypothermia. In this analysis independent predictors of hypothermia on admission to ICU were: magnitude of surgery, use of general anesthesia or combined epidural and general anesthesia, total intravenous crystalloids administrated and total packed erythrocytes administrated, anesthesia longer than 3 hours and SAPS II scores. In multiple logistic regression analysis significant predictors of hypothermia on admission to the ICU were magnitude of surgery (OR 3.9, 95% CI, 1.4–10.6, p = 0.008 for major surgery; OR 3.6, 95% CI, 1.5–9.0, p = 0.005 for medium surgery), intravenous administration of crystalloids (in litres) (OR 1.4, 95% CI, 1.1–1.7, p = 0.012) and SAPS score (OR 1.0, 95% CI 1.0–1.7, p = 0.014); higher previous temperature in ward was a significant protective factor (OR 0.3, 95% CI 0.1–0.7, p = 0.003). Hypothermia was neither a risk factor for hospital mortality nor a predictive factor for staying longer in ICU.
Conclusion
The prevalence of patient hypothermia on ICU arrival was high. Hypothermia at time of admission to the ICU was not an independent factor for mortality or for staying longer in ICU.
==== Body
Background
Hypothermia is defined as a core body temperature of less than 35°C and may be classified as mild (32°C – 35°C), moderate (28°C – 32°C) and severe (<28°C) [1-4].
Hypothermia promotes impairment and abnormalities in various organs and systems that can lead to a decrease in oxygen release into the tissues [5] and include depression in myocardial contractility, peripheral vasoconstriction [6,7] ventilation-perfusion mismatch, increased blood viscosity and shifts to the left in the oxyhemoglobin-dissociation curve [8,9]. Hypothermia also reduces platelet function [10-12] and decreases the activation of the coagulation cascade [13] and may result in coagulopathy [14,15].
Inadvertent core hypothermia is often found in the immediate postoperative period especially in high risk surgical patients [9,16]. Anesthesia impairs central thermoregulation, allowing re-distribution of body heat [17,18]. Cool ambient temperatures and high volume fluid administration accelerate loss of heat to the environment. The initial response to cold stress is to generate and conserve temperature via the activation of the sympathetic nervous system. In the postoperative period, hypothermia is often complicated by shivering and peripheral vasoconstriction. The consequences of shivering include an increase in cardiac and systemic energy demand, raising oxygen consumption and carbon dioxide production and an increase in cardiac work [19,20]. Randomized control trials have proven that mild hypothermia can lead to an increased incidence of surgical wound infection [21], myocardial ischemia and cardiac arrhythmias [22], impaired coagulation and immune response [11,23]; other complications include enhanced anesthesia drug effects [24,25], shivering [20,26], prolonged recovery room stays and delayed discharge from the post anesthesia care and from hospital [27].
On the basis of review of the literature, the most important risk factors for the development of hypothermia in the perioperative period include age [28,29], sex [9,30,31], body weight and body surface area [16,32], preoperative body temperature [16,33] history of diabetic neuropathy [34], emergency surgery [30], ASA physical status [16], surgical procedure in which major body cavities or major vessels are exposed [9,16], anesthetic technique [31] warming method [35], amount of intravenous replacement [30], duration of anesthesia or surgery [9,16,30,31] and ambient operating room temperature [16,28].
With the progress in surgery and anesthesia more advanced technology was introduced, procedures are using more complex techniques, lasting longer and sometimes are made with larger incisions. These procedures are done in older and severely ill patients who present significant underlying medical problems and have sometimes undergone previous extensive surgery. The choice of anesthesia technique is also changing and combined epidural and general anesthesia are being used more often. These are reasons that may increase risk of intraoperative and immediate postoperative hypothermia. Whereas effective warming measures are available [35] and more information about the adverse effects of hypothermia has been reported, the incidence of hypothermia at the time of admission to the surgical ICU is still frequent [22,29,30]. Re-identifying significant predictive factors would help in decreasing this incidence. Such information could be helpful to prevent unnecessary risks and adverse outcomes and could decrease the current frequent incidence of hypothermia.
The purpose of this study was to estimate the prevalence of hypothermia on admission to a surgical ICU and retrospectively identify their clinical predictors. A second objective was to prospectively evaluate the outcome measured in terms of ICU length of stay (LOS) and hospital mortality.
Methods
The protocol was approved by our institutional review board, and written consent was obtained from the patients. All consecutive postoperative patients, aged 18 years or older, who underwent scheduled or emergency noncardiac surgery, admitted to a nine bed surgical ICU of a tertiary care hospital during a three month period between April and July 2004, were eligible for the study. Temperature of the operating room was not controlled for every patient because the air conditioning system is automatically settled to a temperature room from 20°C to 22°C. This fact was confirmed by a pilot study in which operating room temperature was measured using thermocouples placed near patients but away from any heat-generating equipment.
Core temperature (Tc) was measured by an infrared tympanic membrane thermometer (ThermosScan® Type 6014 Pro 3000, WelchAllyn, with an accuracy of ± 0.03°C in the range of 20°C–42.2°C – Welch Allyn Medical Products, NY, USA) [36,37]. This instrument was maintained and calibrated in accordance with the manufacturer's guidelines. These tympanic membrane temperatures were obtained before surgery in the ward, on arrival at ICU and every two hours until 6 hours after admission by experienced nurses trained to measure tympanic membrane temperatures.
The following clinical variables were recorded on admission to the ICU: age, sex, body weight and height, preoperative body temperature, ASA physical status, emergency or scheduled surgery, magnitude of surgical procedure as major (surgery in which body cavities or major vessels are exposed to ambient temperature such as major abdominal, thoracic, major vascular, thoracic spine surgery with instrumentation, or hip arthroplasty), medium (surgery in which body cavities are exposed to a lesser degree such as appendectomy), and minor surgery (superficial surgery), anesthesia technique, amount of intravenous crystalloids, colloids, packed erythrocytes and fresh frozen plasma administered during surgery, use of temperature monitoring and warming techniques, and duration of the anesthesia.
The LOS and the mortality in ICU and in hospital were also recorded for all patients, as well as the Simplified Acute Physiology Score II (SAPS II) was calculated [38].
The prevalence of core hypothermia on admission to the ICU and its 95% Confidence Interval (CI) were calculated by using the cut off point of Tc 35.0°C. That was used to classify patients as either hypothermic (Tc<35°C on admission to ICU) or normothermic (Tc ≥ 35°C on admission to ICU). All hypothermic patients were treated with passive and active external rewarming measures that include synthetic covers and a convective air-warming system using heated air blanket units. The two groups were compared to assess the relationship between each clinical predictor and core hypothermia using univariate analysis performed by simple binary logistic regression with an odds ratio (OR) and its 95% CI and independent sample t test, χ2 test or Fisher's exact test. A multiple regression binary logistic with forward conditional elimination was used to examine covariate effects of each factor on core hypothermia, ICU LOS and hospital mortality and to calculate OR and their 95% CI. Covariates with a univariate p < 0.1 in the respective univariate analysis were entered in these models. In the model for ICU LOS and hospital mortality the categorical variable temperature on admission was also entered. A two-sided significance level of 0.05 was used for all analyses.
Quantitative variables are presented as mean ± SD. All analyses were performed using SPSS for Windows (version 12.0, Chicago, IL).
Results
All 185 patients (Table 1) were included in the study. The mean ( ± SD) admission Tc was 34.7°C ± 1.0°C (range, 32.1°C–38.2°C, first quartile, 34.1°C, third quartile 35.4°C). Prevalence of core hypothermia on ICU admission was 57.8% (95% CI, 54.2%-61.6%). Ten patients were admitted with a Tc <33.0°C.
Table 1 Patient baseline characteristics (n = 185)
Variable mean ± SD or median and range or number (%)
Age Median 67, mean 66.0 ± 12.6 (range, 25 – 94)
<65 / ≥ 65 79 / 106
Male / Female 112 (60.5) / 73 (39.5)
Body mass index (Kg/m2) 25.4 ± 5.9 (range, 15.6 – 57.6)
Previous temperature in the ward (°C) 36.37 ± 0.49 (range, 35.00 – 38.60)
ASA Physical status
I 5 (2.7)
II 72 (38.9)
III 90 (48.6)
IV 18 (9.7)
Emergency surgery 29 (15.7)
Magnitude of surgery
Minor 36 (19.5)
Medium 45 (24.3)
Major 104 (56.2)
General anesthesia 158 (85.4)
Regional anesthesia 19 (10.3)
Combined anesthesia 7 (3.8)
Temperature monitoring 22 (11.9)
Warming technique 81 (43.8)
Intravenous crystalloids (L.) 2.85 ± 1.65 (range 0.20 – 10.50)
Intravenous colloids (L.) 0.09 ± 0.26 (range 0 – 1.500)
Packed erythrocytes (Units) 0.7 ± 1.3 (range, 0 – 7)
Fresh frozen plasma (Units) 0.2 ± 0.9 (range, 0 – 9)
Duration of anesthesia (min.) 218 ± 108 (range, 44 – 660)
> 180 min. 94 (50.8)
Temperature on admission 34.69 ± 1.02 (32.1 – 38.2)
≤ 35 107 (57.8)
Temperature 2 hours after admission 35.34 ± 0.85 (32.4 – 38.3)
≤ 35 48 (25.9)
Temperature 4 hours after admission 35.90 ± 0.80 (32.5 – 38.4)
≤ 35 16 (8.8)
Temperature 6 hours after admission 36.12 ± 0.73 (33.4 – 38.5)
≤ 35 9 (5)
Score of Acute Physiologic system (SAPS II) 24.4 ± 14.0 (range 3–74)
Length of ICU stay (days) median 0.92; percentile 25, 0.79; percentile 75, 2.11 range (0.08–82)
> 2 days 47 (25.4)
Length of hospital stay (days) median 15; percentile 25, 10; percentile 75, 29.5 range 1–170
Mortality in the ICU 14 (7.60)
Mortality in the Hospital 29 (15.70)
The prevalence of hypothermia two hours, four hours and six hours after admission was respectively 26,0% (95% CI, 25.6%-26.3%), 8.8% (95% CI, 8.6%-9.1%) and 5,0% (95% CI, 4.8%-5,1%).
According to univariate analysis (Table 2), age, sex, body weight or body mass index, amount of intraoperative colloids and plasma volume, ASA physical status, emergency surgery, surgery longer than 180 minutes were not predictive of core hypothermia at ICU admission.
Table 2 Univariate Analysis of categorical and continuous predictors of core hypothermia
Variable Hypotermic / non-hypothermic (n° or mean ± sd) Odds ratio (95% CI) p-value
Age (years)
< 65 44/35 1
≥ 65 63/43 0.9 (0.5–1.6) 0.611
Gender
Female 39/34 1
Male 68/44 0.7 (0.4 – 1.4) 0.327
Body weight (Kg) 69.19 ± 16.79 / 68.49 ± 16.06 1,0 (1.0 – 1.0) 0.771
Body mass index (Kg/m2) 25.30 ± 5.77 / 25.58 ± 6.14 1,0 (0.9 – 1.0) 0.750
Previous temperature in the ward (°C) 36.28 ± 0.44 / 36.49 ± 0.54 0.4 (0.2 – 0.8) 0.007
ASA Physical status
I 3/2 1
II 43/29 1.5 (0.2 – 11.2) 0.693
III 52/38 1.5 (0.5 – 4.2) 0.457
IV 9/9 1.4 (0.5 – 3.8) 0.544
Emergency surgery 18/11 0.8 (0.4 – 1.8) 0.616
Magnitude of surgery
Minor 11/25 1
Medium 26/19 3.1 (1.2 – 7.8) 0.016
Major 70/34 4.7 (2.1 – 10.6) < 0.001
General anesthesia 93/65 6,0 (1.9 – 18.8) 0.002
Regional anesthesia 4/15 1
Combined anesthesia 6/1 22.5 (2.1–244.8) 0.011
Temperature monitoring 19/3 0.2 (0.1–0.7) 0.009
Warming technique 55/26 0.5 (0.3 – 0.9) 0.015
Intravenous crystalloids (L.) 3.21 ± 1.75 / 2.37 ± 1.35 1.4 (1.2 – 1.8) 0.001
Intravenous colloids (L.) 0.08 ± 0.26 / 0.10 ± 0.26 0.7 (0.2 – 2.2) 0.565
Packed erythrocytes (Units) 0.92 ± 1.45 / 0.37 ± 0.81 1.6 (1.1 – 2.2) 0.006
Fresh frozen plasma (Units) 0.27 ± 0.81 / 0.19 ± 1.08 1.1 (0.8 – 1.6) 0.575
Duration of anesthesia (min.)
≤ 180 min. 46/45 1
> 180 min. 61/33 1.8 (1.0 – 3.3) 0.049
Score of Acute Physiologic system (SAPS II) 26.45 ± 14.95 / 21.51 ± 12.14 1.0 (1.0 – 1.1) 0.020
Temperature monitoring (OR 0.2, 95% CI 0.1–0.7, p = 0.009), use of warming techniques (OR 0.5, 95% CI 0.3–0.9, p = 0.015) and higher previous body temperature (OR 0.4, 95% CI 0.2–0.8, p = 0.007) were significant protective factors against core hypothermia.
Significant independent predictors of hypothermia on admission were the magnitude of surgery (OR 3.1, 95% CI 1.2–7.8, p = 0.016 for medium surgery; OR 4.7, 95% CI 2.1–10.6, p < 0.001 for major surgery), use of general anesthesia or combined epidural and general anesthesia (OR 6.0, 95% CI, 1.9–18.8, p = 0.002 for general anesthesia; OR 22.5, 95% CI 2.1–244.8, p = 0.011 for combined epidural and general anesthesia), amount of intravenous crystalloids (OR 1.4, 95% CI 1.2–1.8, p = 0.001) and number of total units of packed erythrocytes (OR 1.6, 95% CI 1.1–2.2, p = 0.006), anesthesia longer than 3 hours (OR, 1.8; 95% CI, 1.0–3.3, p= 0.049) and SAPS II scores (OR 1.0, 95% CI 1.0–1.1, p = 0.020).
Table 3 displays the results of the multiple regression binary logistic analysis. In this model higher preoperative body temperature (OR 0.3, 95% CI 0.1–0.7, p = 0.003) is seen as a significant protective factor against core hypothermia. In this analysis the higher the SAPS II of the patient is, the higher the risk of core hypothermia (OR 1.0, 95% CI 1.0–1.7, p = 0.014). The magnitude of surgery (OR 3.9, 95% CI, 1.4–10.6, p = 0.008 for major surgery; OR 3.6, 95% CI, 1.5–9.0, p = 0.005 for medium surgery) and total intravenous crystalloids (in litres) (OR 1.4, 95% CI, 1.1–1.7, p = 0.012) were considerably significant risk factors.
Table 3 Predictors of core hypothermia by multiple logistic regression
Variable Odds ratio (95% CI) p-value
Magnitude of surgery
Medium 3.6 (1.5 – 9.0) 0.005
Major 3.9 (1.4 – 10.6) 0.008
Intravenous crystalloids (L) 1.4 (1.1 – 1.7) 0.012
Previous temperature in the ward 0.3 (0.1 – 0.7) 0.003
Score of Acute Physiologic system (SAPS II) 1.0 (1.0 – 1.7) 0.014
Model adjusted to: previous temperature in the ward, magnitude of surgical procedure, type of anesthesia, use of temperature monitoring, use of warming technique, total intravenous crystalloids, total packed erythrocytes, duration of anesthesia and SAPS II score
No significant relationship between hypothermia at ICU admission and LOS in hospital or in ICU was to be found.
As can be seen in table 1, ICU LOS varied from 0.1 to 82 days with median of 0,9 days (percentile 25, 0.8 days and percentile 75, 2.1 days). The percentage of patients who stayed in ICU longer than 2 days were 25.4% (n = 47).
Multiple regression logistic analysis was used to examine covariate effects of each factor on ICU LOS (table 4). In this analysis the regression model included temperature on admission and all variables that showed statistical significance in the univariate analysis made for predictors of hypothermia and for mortality. This analysis showed that significant risk factors for staying longer in ICU were SAPS II (OR 1.1, 95% CI 1.1–1.1, p < 0.001), ASA physical status (OR 3.4, 95% CI 1.3–9.0, p = 0.012 for ASA III/IV patients) and amount of intravenous crystalloids administered (OR 1.3, 95% CI 1.0–1.7, p = 0.023).
Table 4 Predictors of ICU LOS longer than two days by multiple logistic regression
Variable Odds ratio (95% CI) p-value
ASA III/IV 3.4 (1.3 – 9.0) 0.012
SAPS II 1.1 (1.1 – 1.1) <0.001
Intravenous crystalloids (L.) 1.3 (1.0–1.7) 0.023
Model adjusted to: temperature on admission, previous temperature in the ward, magnitude of surgery, type of anesthesia, use of temperature monitoring, use of warming technique, total intravenous crystalloids, total packed erythrocytes, duration of anesthesia, BMI, ASA and SAPS II score.
Fourteen (7.6%) patients died in ICU and 29 (15.7%) died during their hospitalization.
According to univariate analysis (table 5), age, sex, anesthesia technique, use of a perioperative warming technique, temperature monitoring, duration of anesthesia or surgery were not independent risk factors for mortality in the hospital, as temperature was not a risk factor on admission, neither at two, four and six hours after arrival at ICU.
Table 5 Univariate Analysis of categorical and continuous predictors of mortality in the hospital
Variable nonsurvivors / survivors n° or mean ± SD Odds ratio (95% CI) p-value
Age (years)
< 65 14/65 1
≥ 65 15/91 1.3 (0.6 – 2.9) 0.510
Gender
Male 19/93 0.8 (0.3 – 1.8) 0.551
Female 10/63
Body weight (Kg) 63.2 ± 11.8 / 69.8 ± 16.9 1.0 (0.9 – 1.0) 0.044
Body mass index (Kg/m2) 23.1 ± 3.83 / 25.8 ± 6.1 0.9 (0.8 – 1.0) 0.019
ASA Physical status
I/II 8/ 69 1
III/IV 21/87 2.1 (0.8 – 5.0) 0.10
Emergency surgery 13/16 7.1 (2.9 – 17.4) <0.001
Magnitude of surgery
Minor 2/34 1
Medium 11/34 3.1 (0.7 – 14.2) 0.146
Major 16/88 5.5 (1.1 – 26.7) 0.034
Temperature monitoring 5/17 1.7 (0.6 – 5.1) 0.337
Warming technique 10/71 1.6 (0.7 – 3.6) 0.274
Intravenous crystalloids (L.) 2.88 ± 1.59 / 2.85 ± 1.66 1.0 (0.8 – 1.3) 0.904
Intravenous colloids (L.) 0.16 ± 0.33 / 0.08 ± 0.24 2.4 (0.7 – 8.7) 0.174
Packed erythrocytes (Units) 0.83 ± 1.14 / 0.66 ± 1.27 1.1 (0.8 – 1.5) 0.508
Fresh frozen plasma (Units) 0.17 ± 0.54 / 0.25 ± 0.99 0.9 (0.5 – 1.5) 0.683
Duration of anesthesia (min.) 204 ± 112 / 221 ± 107
≤ 180 min. 16/75 1
> 180 min. 13/81 1.3 (0.6 – 3.0) 0.484
Temperature on admission
≤ 35 20/87 1
> 35 9/69 0.6 (0.2 – 1.3) 0.190
Score of Acute Physiologic system (SAPS II) 41.41 ± 17.86 / 21.20 ± 10.54 1.1 (1.1 – 1.1) <0.001
Length of ICU stay (days)
≤ 2 days 10/128 1
> 2 days 19/28 8.7 (3.7 – 20.7) <0.001
Length of hospital stay (days) 45.7 ± 41.8 / 21.8 ± 21.9 1.0 (1.0 – 1.0) <0.001
Statistically significant independent risk factors for hospital mortality were low body weight (OR 1.0, 95% CI 0.9–1.0, p = 0.044) and low body mass index (OR 0.9, 95% CI 0.8–1.0, p = 0.019), emergency surgery (OR 7.1, 95% CI 2.9–17.4, p < 0.001), major surgery (OR 5.5, 95% CI 1.1–26.7, p = 0.034), high SAPS II scores (OR 1.1, 95% CI 1.1–1.1, p < 0.001), longer stay in ICU (OR 8.7, 95% CI 3.7–20.7, p < 0.001 for ICU LOS longer than 2 days) and in the hospital (OR 1.0, 95% CI 1.0–1.0, p < 0.001).
The multiple regression logistic analysis (table 6) showed that considerably significant factors predicting death in the hospital were higher SAPS scores (OR 1.1, 95% CI 1.0–1.1, p < 0.001) and ICU LOS (OR 1.2, 95% CI 1.1–1.3, p = 0.003). This analysis showed that these were the factors that more significantly predicted death in the in-hospital setting.
Table 6 Predictors of mortality by multiple logistic regression
Variable Odds ratio (95% CI) p-value
SAPS II 1.1 (1.0 – 1.1) <0.001
Length of ICU stay (days) 1.2 (1.1 – 1.3) 0.003
Model adjusted to: temperature on admission to the ICU, BMI, body weight, magnitude of surgery, type of surgery, SAPS II score, ICU LOS and hospital LOS.
Discussion
Only noncardiac surgical patients were enrolled in the study. Since none of them underwent a neurosurgical procedure, hypothermia as a therapeutic tool had not been applied.
The prevalence of hypothermia on arrival at ICU was important and proved higher than in other studies [6,9,16,29,30] although not all of them have chosen the same defining criteria to consider patients hypothermic.
In this study, warmer preoperative body temperature is considered a significant protective factor for hypothermia what was also demonstrated previously [16,33]. This result supports the suggestion that efforts should be done to increase body temperature before surgery. In this context is the idea of increasing operating room temperature to prevent heat loss. Various studies have showed that increasing operating room temperature could prevent core hypothermia [16,39], however, when the operating room is too warm, that is not a comfortable working environment and may even increase the risk of infection [21,40].
Previous studies have shown that older patients [9,30,31] and those with higher ASA physical status [16] had an increased risk for hypothermia. Our study could not confirm these results. We think that may be explained because these patients had more often benefited from any perioperative method of warming and monitoring. In fact, monitoring temperature and the use of techniques to warm patients have demonstrated as protective measures according to our results.
We found that as measured by SAPS II score the more severely ill the patients, the more the probability of arriving hypothermic at ICU.
Although the objective of this study was exploring preadmission clinical factors to hypothermia, we included SAPS in the list of predictive values because this way we could have a measure of severity of illness in the patients admitted to ICU. Measured after admission to ICU, this score reflects not only the pre admission physiologic alterations but also variables like age and co morbidities that were not altered with admission. Only with this premises should the predictive value of SAPS II be evaluated.
Hypothermia is a common occurrence of postoperative period particularly if major body cavities are exposed for long periods of time and other important contributors like the use of intraoperative large amount of intravenous fluids, prolonged operative time and use of general anesthesia or combined anesthesia (general and loco regional anesthesia) are present. In our study, the magnitude of surgery was an independent risk factor for core hypothermia what is in agreement with other studies [9,16] and like other studies [9,16,30,31] that found that duration of anesthesia or surgery are risk factors for hypothermia, we concluded that anesthesia lasting longer than 3 hours was a predictor of core hypothermia with statistical significance. If patients undergoing major procedures with long lasting anesthesia became hypothermic more often, they were probably not warmed actively as we could suspect from the active warming rate and rate of temperature monitoring of this set of patients.
We concluded that the amount of intravenous intraoperative crystalloids was a significant risk factor for developing hypothermia on arrival on ICU and that was already identified in previous studies [9,30]. In fact the infusion of crystalloid solutions at room temperature may significantly contribute to intraoperative hypothermia because warming fluids to core temperature requires body heat. Several studies had demonstrated that infusion of warmed fluids helps in the prevention of hypothermia and reduces the incidence of postoperative shivering [41,42]. A limitation of our protocol was the absence of data about how often had been used warmed intraoperative fluids and what methods were used to warm them.
We found that general anesthesia added significant intraoperative risks for core hypothermia which is in accordance with previous reports showing a frequent incidence of hypothermia with general anesthesia [43,44]. Our results showed a closer relationship concerning the incidence of core hypothermia in the presence of combined epidural anesthesia and general anesthesia and those patients in whom this technique was used were significantly hypothermic at the admission to ICU.
Inadvertent hypothermia during anesthesia is by far the most common perioperative thermal disturbance; this results from a combination of impaired thermoregulation and exposure to the cold environment of the operating room [5]. Prevention of hypothermia reduces anesthesia-related morbidity; so body temperature should be carefully monitored, regardless of anesthetic technique [45] and active warming should be used more frequently because it is a simple and effective technique to avoid hypothermia. In our hospital, and as we could demonstrate with this study, warming rates were very low and the rate of temperature monitoring was even lower, predisposing patients to hypothermia. Like Sessler stated "the minor and major complications of hypothermia are thus well documented. In some patients mild hypothermia is likely to be dangerous. In others it will be uncomfortable and slow recovery". Like him we propose that intraoperative core hypothermia should be avoided and we think the proposed management guidelines should always be observed [46].
Unlike found in an older study [29], we could not show any relationship between hypothermia on arrival at ICU and longer staying in ICU.
We could confirm the results of others about severity of illness of patients as measured by ASA physical status and SAPS II as predictors of prolonged stays in ICU [47,48] and the same happens with the greater amount of intraoperative intravenous fluid administration [16].
Overall mortality in our study was 7,6% in ICU and 15,7% in hospital, values that are within the range expected from the corresponding standard mortality ratios of SAPS II [38,49]. Hypothermia on admission to ICU was not an independent factor for mortality.
Body mass index, ASA physical status, type of surgery, major surgery, SAPS II score and ICU and hospital LOS were found as statistically significant clinical predictors of death, according to the univariate analysis; whereas according to multiple logistic regression analysis, only the severity of illness of patients as measured by SAPS II and longer ICU stay were confirmed as significantly predictors of death.
A limitation of our protocol is that core temperatures were estimated from the aural canal using an infrared thermometer. Infrared measurements may introduce a degree of variability that could be avoided with carefully positioned thermocouples [37]. However, previous observations state that infrared thermometers are very accurate for determining a patient's temperature when used by those who routinely perform thermometry in hospitalized patients and they can provide accurate estimative of core body temperature [50,51].
A controversial point of our study is the considered definition of hypothermia as Tc < 35°C. Others have chosen the same value of core temperature or even a lower one as definition of hypothermia [3,6,9]. Meanwhile many approaches have used 36°C as the cut off point between hypothermia and normothermia in ICU [52] and studies have considered it in their methodology [16,53].
Improved awareness about perioperative hypothermia and its complications with a strictly adherence to the recommendations about temperature monitoring and thermal management guidelines [46] should play a role in decreasing adverse outcome. It may be advantageous to take steps in the way to decrease prevalence of hypothermia and conduct a study to validate this hypothesis.
Conclusion
The prevalence of patient hypothermia on arrival at intensive care is very high but, in our results, is not an independent factor for mortality or for staying longer in ICU. The magnitude of surgery, the severity of disease, longer anesthesia periods, the technique of combined epidural and general anesthesia, large amounts of fluids in the operative setting, not using temperature monitoring and a technique to warm patients tend to be risk factors for hypothermia.
All effort to prevent hypothermia should be done, including the more frequently use of body temperature monitoring and methods to warm patients.
Competing interests
None of the authors have any financial or other relationship that might influence the objectivity when performing the study or preparing the manuscript.
Authors' contributions
FA participated in conception, design, acquisition of the data, analysis of the data, statistical analysis, critical revision of the manuscript and supervision.
MC participated in conception, design, acquisition of the data, analysis of the data and critical revision of the manuscript.
CS participated in analysis of the data, statistical analyses and drafting of the manuscript.
AN and NL participated in acquisition of the data, analysis of the data and revision of the manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
==== Refs
Danzl DF Pozos RS Accidental Hypothermia N Engl J Med 1994 331 1756 1760 7984198 10.1056/NEJM199412293312607
Paton BC Accidental hypothermia Pharmacol Ther 1983 22 331 377 6361803 10.1016/0163-7258(83)90008-6
Vassal T Benoit-Gonin B Carrat F Guidet B Maury E Offenstadt G Severe Accidental Hypothermia Treated in an ICU – prognosis and outcome Chest 2001 120 1998 2003 11742934 10.1378/chest.120.6.1998
Kempainen RR Brunette DD The evaluation and management of accidental hypothermia Respir Care 2004 49 192 205 14744270
Sessler DI Mild Perioperative Hypothermia N Engl J Med 1997 336 1730 1737 9180091 10.1056/NEJM199706123362407
Frank SM Beattie C Christopherson R Norris EJ Perler BA Williams GM Gottlieb SO Unintentional hypothermia is associated with postoperative myocardial ischemia Anesthesiology 1993 78 468 476 8457047
Frank SM Higgins MS Breslow MJ Fleisher LA Gorman RB Sitzmann JV Raff H Beattie C The catecholamine, cortisol, and hemodynamic responses to mild perioperative hypothermia Anesthesiology 1995 82 83 93 7832339 10.1097/00000542-199501000-00012
Ream AK Reitz BA Silverberg GS Temperature correction of PCO2 and pH in estimating acid-base status Anesthesiology 1982 56 41 44 6797331
Bush HL Hydo LJ Fischer E Fantini GA Silane MF Barie PS Hypothermia during elective abdominal aortic aneurysm repair: the high price of avoidable morbidity J Vasc Surg 1995 21 392 400 7877221
Ferrel JE JrMartin GS Platelet tyrosine-specific protein phosphorylation is regulated by trombin Mol Cell Biol 1988 8 3603 3610 2464741
Valeri CR Feingold H Cassidy G Ragno G Khuri S Altschule MD Hypothermia-induced reversible platelet dysfunction Ann Surg 1987 205 175 181 3813688
Michelson AD MacGregor H Barnard MR Kestin AS Roher MJ Valeri CR Irreversible inhibition of human platelet activation by hypothermia in vivo and in vitro Thromb Haemost 1994 71 633 640 7522354
Roher MJ Natale AM Effect of hypothermia on the coagulation cascade Crit Care Med 1992 20 1042 1405
Winkler M Akca O Birkenberg B Hetz H Scheck T Arkilic CF Kabon B Marker E Grubl A Czepan R Greher M Goll V Gottsauner-Wolf F Kurz A Sessler DI Aggressive warming reduces blood loss during hip arthroplasty Anesth Analg 2000 91 978 984 11004060 10.1097/00000539-200010000-00039
Schmied H Kurz A Sessler DI Kozek S Reiter A Mild hypothermia increases blood loss and transfusion requirements during total hip arthroplasty Lancet 1996 47 289 292 8569362 10.1016/S0140-6736(96)90466-3
Kongsayreepong S Chainbundit C Chadpaibool J Komoltri C Suraseranivongse S Predictor of Core Hypothermia and Surgical Intensive Care Unit Anesth Analg 2003 96 826 833 12598269 10.1213/01.ANE.0000048822.27698.28
Annadata RS Sessler DI Tayefeh F Kurz A Dechert M Desflurane slightly increases the sweating threshold, but produces marked, nonlinear decreases in the vasoconstrition and shivering thresholds Anesthesiology 1995 83 1205 1211 8533913 10.1097/00000542-199512000-00011
Sessler DI Perioperative heat balance Anesthesiology 2000 92 578 596 10691247 10.1097/00000542-200002000-00042
Frank SM Higgins MS Breslow MJ Fleisher LA Gorman RB Sitzmann JV Raff H Beattie C The catecholamine, cortisol, and hemodynamic responses to mild perioperative hypothermia: a randomized clinical trial Anesthesiology 1995 82 83 93 7832339 10.1097/00000542-199501000-00012
Bilotta F Pietropaoli P La Rosa I Spinelli F Rosa G Effects of shivering prevention on haemodynamic and metabolic demands in hypothermic postoperative neurosurgical patients Anaesthesia 2001 56 514 519 11412155 10.1046/j.1365-2044.2001.02057.x
Kurz A Sessler DI Lenhardt R Perioperative normothermia to reduce the incidence of surgical-wound infection and shorten hospitalization: Study of Wound Infection and Temperature Infection and Temperature Group N Engl J Med 1996 334 1209 1215 8606715 10.1056/NEJM199605093341901
Frank SM Fleisher LA Breslow MJ Higgins MS Olson KF Kelly S Beattie C Perioperative maintenance of normothermia reduces the incidence of morbid cardiac events: a randomized trial JAMA 1997 277 1127 1134 9087467 10.1001/jama.277.14.1127
Wenisch C Narzt E Sesseler DI Mild intraoperative hypothermia reduces production of reactive oxygen intermediates by polymorphonuclear leukocytes Anesth Analg 1996 82 810 816 8615502 10.1097/00000539-199604000-00023
Leslie K Sessler DI Bjorksten AR Moayeri A Mild hypothermia alters propofol pharmacokinetics and increases the duration of action of atracurium Anesth Analg 1995 80 1007 14 7726398 10.1097/00000539-199505000-00027
Heier T Caldwell JE Sessler DI Miller RD Mild intraoperative hypothermia increases duration of action and spontaneous recovery of vecuronium blockade during nitrous oxide-isoflurane anesthesia in humans Anesthesiology 1991 74 815 819 1673591
Macário A Weinger M Truong P Lee M Which clinical anaesthesia outcomes are both common and important to avoid? The perspective of a panel of expert anesthesiologists Anest Analg 1999 88 1085 1091 10.1097/00000539-199905000-00023
Lenhardt R Marker E Goll V Tschernich H Kurz A Sessler DI Narzt E Lackner F Mild intraoperative hypothermia prolongs postanesthetic recovery Anesthesiology 1997 87 1318 1323 9416715 10.1097/00000542-199712000-00009
Frank SM Beattie C Christopherson R Norris EJ Rock P Parker S Kimball AW Jr Epidural versus general anesthesia, ambient operating room temperature, and patient age as predictors of inadvertent hypothermia Anesthesiology 1992 77 252 257 1642343
Berry JM Garrett K Clifton GL Postoperative hypothermia in a tertiary care hospital Anesthesiology 1999 91 A1177
Slotman GJ Jed EH Burchard KW Adverse effect of hypothermia in postoperative patients Am J Surg 1985 149 495 501 3985289
Roberts S Bolton DM Stoller ML Hypothermia associated with percutaneous nephrolithotomy Urology 1994 44 832 835 7985311 10.1016/S0090-4295(94)80166-5
Kurz A Sessler DI Narzt E Lenhardt R Lackner F Morphometric influences on intraoperative core temperature changes Anesth Analg 1995 80 562 567 7864426 10.1097/00000539-199503000-00023
Just B Trevien V Delva E Leinhardt A Prevention of intraoperative hypothermia by preoperative skin-surface warming Anesthesiology 1993 79 214 218 8251019
Kitamura A Hoshino T Kon T Ogawa R Patient with diabetic neuropathy are at risk of a greater intraoperative reduction in core temperature Anesthesiology 2000 92 1311 1318 10781276 10.1097/00000542-200005000-00019
Hynson JM Sessler DI Intraoperative warming therapies: a comparison of three devices J Clin Anesth 1992 4 194 199 1610573 10.1016/0952-8180(92)90064-8
Amoateng-Adjepong Y Del Mundo J Manthous CA Accuracy of an infrared tympanic thermometer Chest 1999 115 1002 1005 10208200 10.1378/chest.115.4.1002
Panzer O Ghazanfari N Sessler DI Yucel Y Greher M Akca O Donner A Germann P Kurz A Shivering and shivering-like tremor during labor with and without epidural analgesia Anesthesiology 1999 90 1609 1616 10360859 10.1097/00000542-199906000-00017
Le Gall JR Lemeshow S Saulnier F A new simplified acute physiology score (SAPSII) based on a European/North American multicenter study JAMA 1993 270 2957 2963 8254858 10.1001/jama.270.24.2957
El-Gamal N El-Kassabany N Frank SM Amar R Khabar HA El-Rahmany HK Okasha AS Age-related thermoregulatory differences in a warm operating room environment (approximately 26°C) Anest Analg 2000 90 694 698 10.1097/00000539-200003000-00034
Mills SJC Holland DJ Hardy AE Operative field contamination by the sweating surgeon Aust N Z J Surg 1999 70 837 839 11167569 10.1046/j.1440-1622.2000.01999.x
Camus Y Delva E Cohen S Lienhart A The effects of warming intravenous fluids on intraoperative hypothermia and postoperative shivering during prolonged abdominal surgery Acta Anaesthesiol Scand 1996 40 779 82 8874562
Smith CE Desai R Glorioso V Cooper A Pinchak AC Hagen KF Preventing hypothermia: convective and intravenous fluid warming versus convective warming alone J Clin Anesth 1998 10 380 5 9702617 10.1016/S0952-8180(98)00049-X
Matsukawa T Kurz A Sessler DI Bjorksten AR Merrifield B Cheng C Propofol linearly reduces the vasoconstriction and shivering thresholds Anesthesiology 1995 82 1169 1180 7741292 10.1097/00000542-199505000-00012
Hynson JM Sessler DI Moayeri A McGuire J Absence of nonshivering thermogenesis in anesthetized adult humans Anesthesiology 1993 79 695 703 8214747
Cattaneo CG Frank SM Hesel TW El-Rahmany HK Kim LJ Tran KM The accuracy and precision of body temperature monitoring methods during regional and general anesthesia Anesthesiology 1996 84 1327 1331 8669673
Sessler DI A Proposal for New Temperature Monitoring and Thermal Management Guidelines Anesthesiology 1998 89 1298 1300 9822041 10.1097/00000542-199811000-00061
Wolters U Wolf T Stutzer H Schroder T ASA classification and perioperative variables as predictors of postoperative oucome Br J Anaesth 1996 77 217 222 8881629
Stricker K Rothen HU Takala J Resource use in ICU: short-vs. long-term patients Acta Anaesthesiol Scand 2003 47 508 515 12699506 10.1034/j.1399-6576.2003.00083.x
Moreno R Morais P Outcome prediction in intensive care: results of a prospective, multicentre Portuguese study Intensive Care Med 1997 23 177 186 9069003 10.1007/s001340050313
Imamura M Matsukawa T Ozaki M Sessler DI Nishiyama T Kumazawa T The accuracy and precision of four infrared aural canal thermometers during cardiac surgery Acta Anaesthesiol Scand 1998 42 1222 1226 9834809
Rotello LC Crawford L Terndrup TE Comparison of infrared ear thermometer derived and equilibrated rectal temperatures in estimating pulmonary artery temperatures Crit Care Med 1996 24 1501 1516 8797622 10.1097/00003246-199609000-00012
Sessler DI Robinstein EH Moaeyri BA Physiologic responses to mild perianesthetic hypothermia Anesthesiology 1991 75 594 610 1928769
Insler SR O'Connor MS Leventhal MJ Nelson DR Starr NJ Association between postoperative hypothermia and adverse outcome after coronary artery bypass surgery Ann Thorac Surg 2000 70 175 181 10921704 10.1016/S0003-4975(00)01415-6
| 15938757 | PMC1180426 | CC BY | 2021-01-04 16:28:05 | no | BMC Anesthesiol. 2005 Jun 6; 5:7 | utf-8 | BMC Anesthesiol | 2,005 | 10.1186/1471-2253-5-7 | oa_comm |
==== Front
BMC CancerBMC Cancer1471-2407BioMed Central London 1471-2407-5-611594904710.1186/1471-2407-5-61Research ArticlePhase II trial of weekly 24-hour infusion of gemcitabine in patients with advanced gallbladder and biliary tract carcinoma von Delius Stefan [email protected] Christian [email protected] Ewert [email protected] Martina [email protected] Roland M [email protected] Florian [email protected] Department of Internal Medicine II, Technical University of Munich, Klinikum rechts der Isar, Ismaninger Str. 22, 81675 München, Germany2005 12 6 2005 5 61 61 15 2 2005 12 6 2005 Copyright © 2005 von Delius 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
Patients with advanced gallbladder and biliary tract carcinoma face a dismal prognosis, as no effective palliative chemotherapy exists. The antitumor effect of gemcitabine is schedule-dependent rather than dose-dependent. We evaluated the activity of a prolonged infusion of gemcitabine in advanced gallbladder and biliary tract carcinomas.
Methods
Nineteen consecutive eligible patients were enrolled. All patients were required to have histologically confirmed diagnosis and measurable disease. Gemcitabine was infused over 24 hours at a dose of 100 mg/m2 on days 1, 8, and 15. Treatment was repeated every 28 days until progression of disease or limiting toxicity. Tumor response was evaluated every second course by computed tomography (CT) scans.
Results
Eighteen patients were evaluable for response. A total of 89 cycles of therapy were administered. One partial response was observed (6%; 95% confidence interval (CI): 0–27%) and ten additional patients had stable disease for at least two months (disease control rate 61%; 95% CI: 36–83%). The therapy was well tolerated, with moderate myelosuppression as the main toxicity. The median time to tumor progression and median overall survival was 3.6 months (95% CI 2.6–4.6 months) and 7.5 months (95% CI 6.5–8.5 months), respectively.
Conclusion
Weekly 24-hour gemcitabine at a dose of 100 mg/m2 is well tolerated. There was a relatively high rate of disease control for a median duration of 5.3 months (range 2.8–18.8 months). However, the objective response rate of this regimen in gallbladder and biliary tract carcinomas was limited.
==== Body
Background
Adenocarcinomas of the gallbladder and the biliary tract are relatively uncommon in the western world, with approximately 5,000 cases of gallbladder carcinoma and 2,500 cases of cholangiocarcinoma annually in the USA [1,2]. Worldwide, the highest prevalence of gallbladder cancer is seen in Israel, Mexico, Chile, Japan, and among Native American women, particularly those living in New Mexico. Incidence of cholangiocarcinoma is highest in Israel, Japan, among Native Americans, and in Southeast Asia, where it can reach 87 per 100,000 [3].
Biliary tract tumors can occur anywhere in the hepatobiliary system. Cholangiocarcinoma may be further classified as intrahepatic or extrahepatic (hilar and distal bile duct) carcinomas. Hilar tumors, so called Klatskin's tumors, usually require partial liver resection for cure, whereas distal tumors may require pancreatectomy.
Gallbladder carcinoma has been linked to several risk factors. Gallstones, especially larger than 3 cm, chronic inflammation, bacterial infection, polyps, calcified (porcelain) gallbladder, ulcerative colitis, high energy and total carbohydrate intake, and high body mass index in women increase risk. Risk factors for cholangiocarcinoma are well understood, although most patients lack identifiable risks. Patients with primary sclerosing cholangitis, congenital choledochal cysts (most likely due to associated inflammation and bacterial infection), ulcerative colitis, exposure to carcinogens including cigarette smoke, hepatolithiasis, liver infection, hepatitis C, and biliary parasites carry an elevated risk [3].
Due to the lack of characteristic early symptoms, a definitive diagnosis of gallbladder and biliary tract carcinoma is often established at an advanced stage, and prognosis of patients with advanced tumors remains dismal. Even in patients undergoing aggressive surgery, the general outcome has been disappointing. Five-year survival rates are between 5–10% [1]. Median survival of patients with advanced-stage disease is in the range of only a few months. Due to the lack of randomized phase III studies there is no standard regimen for palliative chemotherapy of gallbladder and biliary tract carcinomas. Depending on the patient's general condition best supportive care, a clinical trial, 5-FU, or gemcitabine is recommended according to guidelines of the National Comprehensive Cancer Network.
Gemcitabine is phosphorylated by deoxycytidine kinase into the active metabolite gemcitabine triphosphate (GemTP). The rate-limiting step in the formation of GemTP is the phosphorylation of gemcitabine to the monophosphate by deoxycytidine kinase [4]. The diphosphate is a potent inhibitor of ribonucleotide reductase, an action that reduces deoxynucleotide pools. Decreased cellular concentrations of deoxycytidine triphosphates permit a more rapid phosphorylation of gemcitabine and decrease the metabolic clearance of gemcitabine nucleotides. As a consequence of this, the active nucleotide forms of gemcitabine are effectively accumulated to high concentrations in the cell [5].
According to the rate-limited activation of gemcitabine a schedule-dependent antitumor effect of gemcitabine was found in various in vivo model systems in preclinical studies [6] suggesting slower infusion rates in the clinical setting. Furthermore, the treatment of mice bearing the colon 26–10 murine colon carcinoma was considerably more effective by applying a weekly 24-hour schedule [7].
Based on the results of a pilot study [8] and a phase II study in patients with pancreatic adenocarcinoma [9] we started a phase II trial of weekly 24-hour infusion of gemcitabine in patients with advanced gallbladder and biliary tract carcinoma.
Methods
Patients
To be eligible for the trial, patients were required to have histologically confirmed, irresectable or metastatic carcinoma of the gallbladder or biliary tract. Patients had to have measurable disease on a computed tomography (CT) scan or magnetic resonance imaging (MRI) and had to have a performance status (WHO) ≤ 2. Patients were allowed to have received one previous chemotherapy, but no previous treatment with gemcitabine. They also met the following laboratory criteria: serum bilirubin value ≤ two times upper normal limit, leukocyte count ≥ 3.0 × 109/L, platelet count ≥ 100 × 109/L, and serum creatinine ≤ 2.0 mg/dL.
Patients with active infections, unstable cardiovascular conditions, brain metastases, or other serious medical illnesses were excluded from this trial.
Informed consent was obtained from all patients included in the study, which was approved by local ethics committees.
Treatment Plan
Therapy was administered on an outpatient basis and patients were premedicated with appropriate antiemetics.
Patients received gemcitabine (100 mg/m2) as a 24-hour intravenous (i.v.) infusion. Treatments were repeated every week for three consecutive weeks followed by one week of rest; a full cycle consisted of four weeks.
Doses were modified according to the following criteria: The weekly gemcitabine dose was reduced by 25% for patients whose leukocyte count was ≤ 2.5–3.0 × 109/mL, whose platelet count was ≥ 75–100 × 109/L, or if grade 2 nonhematologic toxicity was present. The weekly dose was reduced by 50% for patients whose leukocyte count was ≤ 2.0–2.5 × 109/L or whose platelet platelet count was ≥ 50–75 × 109/L. The weekly dose was omitted if the leukocyte count was less than 2.0 × 109/L, platelet count was less than 50 × 109/L, or grade 3 nonhematologic toxicity was present.
Patient Evaluation
Weekly blood counts were obtained to determine the level of myelosuppression. Before each cycle a full clinical evaluation noting the performance status and a physical exam were performed and a complete chemistry panel was obtained. In addition, CA19-9 levels were measured every eight weeks.
The antitumor response was evaluated every eight weeks by CT or MRI, unless signs of progression were evident. In terms of response RECIST criteria [10] were applied, since in most cases of these tumors interpretation of bi-dimensional measurements is complicated by associated inflammation and/or necrosis, as well as anatomical structures in the vicinity. Complete response was defined as the disappearance of all signs and symptoms of disease. Partial response was defined as a decrease of > 30 % of the sum of the largest diameters of target (= measurable) lesions without appearance of new lesions or progression of non-target (= evaluable) lesions. To be assigned a status of response, changes in tumor measurement were confirmed by repeat assessment that was performed no less than 4 weeks after the criteria for response were first met. Stable disease was defined as no sufficient shrinkage to qualify for partial response or less than a 20% increase in the sum of the largest diameters of target lesions without appearance of new lesions or progression of non-target lesions. Progressive disease was defined as a 20% increase in the sum of the largest diameters of target lesions or as appearance of new lesions or as progression of non-target lesions.
The response was validated by radiologists independent of the study.
Disease control was defined as the absence of tumor progression (i.e. complete and partial response and stable disease) for at least two months.
Statistical Methods
The primary end point of this trial was objective response. Secondary end points included progression free survival, overall survival, and toxicity.
Simon's optimal two-stage design was used for calculation of the sample size [11]. Early stopping rules were provided to allow for study discontinuation in the event of lack of efficacy. Sufficient responses (greater than 2 in 18 evaluable patients) were required to trigger the second phase of enrollment for another 17 for a maximum of 35 patients. 95% confidence interval was calculated by the method of Clopper and Pearson. Correlations are nonparametric by the method of Spearman. Progression free survival and overall survival were calculated from the start of chemotherapy until progression, death or last follow-up by the Kaplan-Meier method using SPSS, version 12.0 software.
Results
A total of 19 patients – eight men and eleven women – were entered onto this trial. Their pre-treatment characteristics are listed in Table 1. The median age was 63 years (range 30–83 years, 14 patients younger than 75 years). All patients had adenocarcinomas. There were three patients with gallbladder carcinoma, all of whom were women. Four patients (21%) had relapse of disease after surgery with incomplete resections in two patients. Two patients (11%) had been previously treated with a combination of cyclophosphamide, leucovorin, 24-hour fluorouracil and tamoxifen [12]. All other patients were chemotherapy-naïve. Seven patients (37%) required biliary tract decompression by endoscopic or percutaneous stenting because of obstructive jaundice before the start of chemotherapy. Baseline CA19-9 levels were recorded in 17 patients and were elevated in 14 of them. One female patient was not evaluable regarding response due to early discontinuation of therapy because of withdrawal of consent.
Toxicity
A total of 89 cycles were administered during the trial (median 4 cycles; range 1–19 cycles). Overall, weekly 24-hour gemcitabine was well tolerated (Table 2). No treatment-related deaths occurred. Myelosuppression was the major toxicity. Among all 19 patients grade 3/4 neutropenia occurred in four patients (21%) with no episode of febrile neutropenia. One patient developed grade 4 thrombocytopenia with epistaxis and needed a platelet transfusion. No grade 4 and only one case of grade 3 nonhematologic toxicity effects were noted. Nausea and emesis, when present, usually were mild. Grade 3 mucositis was present in one patient. Allergic urticaria following gemcitabine infusion was seen in one patient, but was manageable by antiallergic medication during the further course.
Response and Survival
No complete response was observed. One of the 18 evaluable patients (6%; 95%CI: 0–27%) with a gallbladder carcinoma had a partial response. Ten additional patients had stable disease for a median duration of 5.3 months (range 2.8–18.8 months) resulting in a disease control rate of 61% (95%CI: 36–83%). Two of these patients had gallbladder carcinoma, four intrahepatic cholangiocarcinoma, and four extrahepatic cholangiocarcinoma. Seven (39%, 95%CI: 17–64%) patients had stable disease after 4 cycles of chemotherapy.
Median progression free survival was 3.6 months (95% CI 2.6–4.6 months) with a 3 and 6-month progression free survival rate of 61% and 23%, respectively. Median overall survival for all 19 patients was 7.5 months (95% CI: 6.5–8.5 months) with a 6 and 12-month survival rate of 73% and 34%, respectively (Figure 1). There was a significant correlation between progression free and overall survival (r = 0.54, p = 0.016).
In fourteen of seventeen patients (82%) assessed baseline CA19-9 levels were elevated (median 581, mean 98,849, range 115–1,244,000; normal < 32 U/mL). In eleven patients follow-up values were obtained. Ten of these eleven patients had elevated baseline levels. There was a significant negative correlation (r = -0.65, p = 0.032) between relative changes of CA19-9 during chemotherapy compared with baseline and progression free survival.
Discussion
In this phase II trial of a prolonged infusion of gemcitabine for advanced gallbladder and biliary tract carcinoma we only found a 6% objective response rate (95%CI: 0–27%). According to the early stopping rules accrual was terminated after enrollment of 19 patients, thereof 18 evaluable for response. Disease control for at least two months was observed in 11 patients (61%, 95%CI: 36–83%) for a median duration of 5.3 months (range 2.8–18.8 months). As expected there was a significant correlation of progression free and overall survival. Toxicity was generally mild with myelosuppression as the major toxicity.
Only relatively small phase II trials have assessed the efficacy and toxicity profiles of chemotherapy regimens in the palliative treatment of gallbladder cancer and cholangiocarcinoma. Due to tumor-specific complications such as obstructive jaundice with impaired hepatic metabolism and biliary excretion toxicity profiles of chemotherapy regimens may be different in gallbladder and biliary tract carcinoma compared to other cancers. Single-agent and multiagent regimens have yielded modest results in patients with advanced biliary carcinomas. Overall response rates and disease control rates are about 25% (mostly 13–35%, range 0–64%) and about 65% (mostly 55–72%, range 19–94%), respectively [13].
Among several different new anticancer drugs, gemcitabine has generated particular interest. Considered as standard treatment for advanced pancreatic carcinoma [14] and given the histogenetic affinity between the pancreas and the biliary tract, a number of phase II trials of gemcitabine with differing results were conducted in patients with advanced gallbladder and biliary tract cancers. Patients mostly received gemcitabine at a dose of 1000 mg/m2 over 30 minutes.
Possible options to further improve the therapeutic efficacy of gemcitabine monotherapy include modifications of the administration schedule. Tempero et al. [15] found improved activity of gemcitabine 1500 mg/m2 as a fixed dose rate infusion of 10 mg/m2/min in comparison to high dose gemcitabine (2200 mg/m2) infusion over 30 minutes in pancreatic cancer. Pharmacokinetic analysis in this study also revealed significantly higher concentrations of intracellular gemcitabine triphosphate with the prolonged infusion. However, grade 3–4 hematologic toxicity occurred in up to almost 50% in the fixed dose rate arm. In contrast, in our trial of 24-hour gemcitabine at a dose of 100 mg/m2 toxicity was mild.
Median survival achieved by our patients was 7,5 months (95% CI: 6.5–8.5) and therefore in the range of most other studies investigating gemcitabine as single agent for biliary tract cancer. Progression free survival (3.6 months; 95% CI 2.6–4.6) was similar to other published data of phase II trials of gemcitabine in biliary tract cancer [Table 3]. Nevertheless, with only small objective response observed our results do not support the use of gemcitabine as continuous infusion in this setting.
Conclusion
In conclusion, weekly 24-hour infusion of gemcitabine at a dose of 100 mg/m2 in patients with advanced gallbladder and biliary tract carcinoma can be safely administered in an outpatient setting. A relatively high disease control rate was observed in this trial. However, in terms of objective response antitumor activity was only marginal. Therefore, gemcitabine as continuous infusion seems not to be superior to gemcitabine administered in short term.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
SvD carried out the care of the patients, collection and interpretation of the data, and drafting of the manuscript. CL conceived of the study, and participated in its design and coordination. ES-F participated in its design and coordination. MM participated in patient care and coordination. RMS helped to draft the manuscript and to interpret the results. FE conceived of the study, participated in its design and coordination and carried out the statistical analysis. All authors read and approved the final manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Figures and Tables
Figure 1 Probability of progression free survival (dotted line) and overall survival for all 19 patients enrolled.
Table 1 Patient characteristics of all patients enrolled.
Total no. of patients (evaluable for response) 19 (18)
Sex (female/male) 11/8
Age, median (range) 63 (30–83)
Performance status (WHO)
0 1 (5%)
1 12 (63%)
2 6 (32%)
Tumor site
Gallbladder 3 (16%)
Intrahepatic bile ducts 9 (47%)
Extrahepatic bile ducts 7 (37%)
Site of metastasis
Liver 9 (47%)
Lung 2 (11%)
Lymph nodes 5 (26%)
Peritoneum 2 (11%)
Recurrence after previous surgical resection 4 (21%)
Pre-treatment requiring stent or percutaneous transhepatic drainage because of obstructive jaundice 7 (37%)
Pre-treatment chemistries
Total bilirubin (mg/dl)
Median 0,7
Range 0,3–2,3
Gamma-glutamyltransferase (<66 U/La)
Median 159
Range 34–671
Aspartate aminotransferase (10–50 U/la)
Median 29
Range 8–145
CA19-9 (<32 U/mLa)
Median 387
Range 4-1244000
a normal levels
Table 2 Treatment-related toxicities of all 19 patients enrolled.
Number of patients (%)
NCI-CTC Grade 1 2 3 4
Leukopenia 2 (11) 6 (32) 3 (16) 0
Neutropenia 3 (16) 1 (5) 2 (11) 2 (11)
Anemia 0 1 (5) 0 0
Thrombocytopenia 3 (16) 3 (16) 0 1 (5)
Diarrhea 1 (5) 1 (5) 0 0
Nausea/emesis 5 (26) 2 (11) 0 0
Stomatitis 3 (16) 2 (11) 1 (5) 0
NCI-CTC: National Cancer Institute Common Toxicity Criteria
Table 3 Selection of phase II studies of gemcitabine in advanced biliary tract cancer.
Study Number of assessable patients Response SD PFS (months) OS (months)
Metzger et al (16) 13 1 (8%) 11 (85%) 7 16
Verderame et al (17) 4 1 (25%) 3 (75%) 11 NR
Gebbia et al (18) 18 4 (22%) 5 (28%) 3 8
Penz et al (19) 32 7 (22%) 14 (44%) 6 12
Lin et al (20) 24 3 (13%) 8 (33%) 3 7
Eng (21) 14 0 (0%) 2 (13%) 2 5
Tsavaris et al (22) 30 9 (30%) 11 (37%) 7 14
Park et al (23) 23 6 (26%) 8 (35%) 8 13
Abbreviations: NR, not reported; OS, overall survival; PFS, progression-free survival; SD, stable disease.
==== Refs
De Groen PC Gores GJ LaRusso NF Gunderson LL Nagorney DM Biliary tract cancers N Engl J Med 1999 341 1368 78 10536130 10.1056/NEJM199910283411807
Gores GJ Cholangiocarcinoma: current concepts and insights Hepatology 2003 5 961 9 10.1053/jhep.2003.50200
Rajagopalan V Daines WP Grossbard ML Kozuch P Gallbladder and biliary tract carcinoma: A comprehensive update, Part 1 Oncology (Huntingt) 2004 18 889 96 15255172
Grunewald R Kantarjian H Keating MJ Abbruzzese J Tarassoff P Plunkett W Pharmacollogically directed design of the dose rate and schedule of 2, 2-dideoxy-fluorocytidine (Gemcitabine) administration in leukemia Cancer Res 1990 50 6823 6 2208147
Plunkett W Huang P Searcy CE Gandhi V Gemcitabine: Preclinical pharmacology and mechanisms of action Semin Oncol 1996 23 3 15 8893876
Braakhuis BJ Ruiz van Haperen VW Boven E Veerman G Peters GJ Schedule-dependent antitumor effect of gemcitabine in in vivo model systems Semin Oncol 1995 22 42 6 7481844
Veerman G Ruiz van Haperen VW Vermorken JB Noordhuis P Braakhuis BJ Pinedo HM Peters GJ Antitumor activity of prolonged as compared with bolus administration of 2, 2-difluorodeoxycytidine in vivo against murine colon tumors Cancer Chemother Pharmacol 1996 38 335 42 8674156 10.1007/s002800050492
Eckel F Lersch C Assmann G Schulte-Frohlinde E Toxicity of a 24-hour infusion of gemcitabine in biliary tract and pancreatic cancer: A pilot study Cancer Invest 2002 20 180 5 11901537 10.1081/CNV-120014881
Eckel F Schmelz R Erdmann J Mayr M Lersch C Phase II trial of a 24-hour infusion of gemcitabine in previously untreated patients with advanced pancreatic adenocarcinoma Cancer Invest 2003 21 690 4 14628426 10.1081/CNV-120023767
Therasse P Arbuck SG Eisenhauer EA Wanders J Kaplan RS Rubinstein L Verweij J van Glabbeke M van Oosterom AT Christian MC Gwyther SG New guidelines to evaluate the response to treatment in solid tumors J Natl Cancer Inst 2000 92 205 16 10655437 10.1093/jnci/92.3.205
Simon R Optimal two-stage designs for phase II clinical trials Controlled Clinical trials 1989 10 1 10 2702835 10.1016/0197-2456(89)90015-9
Eckel F Lersch C Assmann G Schulte-Frohlinde E Phase II trial of low-dose cyclophosphamide, leucovorin, high-dose 5-fluorouracil 24-hour continuous infusion and tamoxifen in advanced biliary tract cancer Ann Oncol 2000 11 762 3 10942070 10.1023/A:1008368611993
Daines WP Rajagopalan V Grossbard ML Kozuch P Gallbladder and biliary tract carcinoma: A comprehensive update, Part 2 Oncology (Huntingt) 2004 18 1049 59 15328897
Burris HA 3rdMoore MJ Andersen J Green MR Rothenberg ML Modiano MR Cripps MC Portenoy RK Storniolo AM Tarassoff P Nelson R Dorr FA Stephens CD Von Hoff DD Improvements in survival and clinical benefit with gemcitabine as first-line therapy for patients with advanced pancreas cancer: a randomized trial J Clin Oncol 1997 15 2403 13 9196156
Tempero M Plunkett W Ruiz van Haperen V Hainsworth J Hochster H Lenzi R Abbruzzese J Randomized phase II comparison of dose-intense gemcitabine: Thirty-minute infusion and fixed dose rate infusion in patients with pancreatic adenocarcinoma J Clin Oncol 2003 21 3402 8 12885837 10.1200/JCO.2003.09.140
Metzger J Sauerbruch T Ko Y A phase II trial of gemcitabine in gallbladder and biliary tract carcinomas Onkologie 1998 21 232 4 10.1159/000026821
Verderame F Mandina P Abruzzo F Scarpulla M Di Leo R Biliary tract cancer: our experience with gemcitabine treatment Anticancer Drugs 2000 11 707 8 11129732 10.1097/00001813-200010000-00006
Gebbia V Giuliani F Maiello E Colucci G Verderame F Borsellino N Mauceri G Caruso M Tirrito ML Valdesi M Treatment of inoperable and/or metastatic biliary tree carcinomas with single-agent gemcitabine or in combination with levofolinic acid and infusional fluorouracil: results of a multicenter phase II study J Clin Oncol 2001 19 4089 91 11600613
Penz M Kornek GV Raderer M Ulrich-Pur H Fiebiger W Lenauer A Depisch D Krauss G Schneeweiss B Scheithauer W Phase II trial of two-weekly gemcitabine in patients with advanced biliary tract cancer Ann Oncol 2001 12 183 6 11300321 10.1023/A:1008352123009
Lin MH Chen JS Chen HH Su WC A phase II trial of gemcitabine in the treatment of advanced bile duct and periampullary carcinomas Chemotherapy 2003 49 154 8 12815209 10.1159/000070622
Eng C Ramathan RK Wong MK Remick SC Dai L Wade-Oliver KT Mani S Kindler HL A Phase II trial of fixed dose rate gemcitabine in patients with advanced biliary tree carcinoma Am J Clin Oncol 2004 27 565 9 15577433 10.1097/01.coc.0000135924.94955.16
Tsavaris N Kosmas C Gouveris P Gennatas K Polyzos A Mouratidou D Tsipras H Margaris H Papastratis G Tzima E Papadoniou N Karatzas G Papalambros E Weekly gemcitabine for the treatment of biliary tract and gallbladder cancer Invest New Drugs 2004 22 193 8 14739669 10.1023/B:DRUG.0000011797.09549.53
Park JS Oh SY Kim SH Kwon HC Kim JS Jin-Kim H Kim YH Single-agent gemcitabine in the treatment of advanced biliary tract cancers: a phase II study Jpn J Clin Oncol 2005 35 68 73 15709089 10.1093/jjco/hyi021
| 15949047 | PMC1180427 | CC BY | 2021-01-04 16:03:05 | no | BMC Cancer. 2005 Jun 12; 5:61 | utf-8 | BMC Cancer | 2,005 | 10.1186/1471-2407-5-61 | oa_comm |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.